MUTANT ANTIBODIES AND CONJUGATION THEREOF

Abstract

The present invention relates to a polypeptide comprising 7 β-strands A, B, C, D, E, F, and G sequentially connected together by connecting chains of amino acids, and a first α-helix sequentially located on the EF chain between β-strands E and F, wherein the β-strands are arranged so as to form a first β-sheet comprising β-strands A, B, D, and E, and a second β-sheet comprising β-strands C, F and G, said first and second β-sheets being covalently bonded together so as to form a first Ig domain; wherein the EF chain between β-strands E and F comprises the sequence X.sup.1-X.sup.2-X.sup.3-X.sup.4-K.sup.5H.sup.6 (SEQ ID NO:98), and X.sup.1, X.sup.3 and X.sup.4 are each independently any amino acid residue, characterized in that X.sup.2 is selected from the group consisting of A, G, I, V, L, R, S, T, Q, P, N, M, H, W, and pharmaceutically acceptable salts, stereoisomers, tautomers, solvates, and prodrugs thereof

Claims

1.-12. (canceled)

13. A polypeptide comprising a mammalian antibody constant light domain comprising the residues K and H at positions corresponding to positions 80 and 81 of SEQ ID NO:6 when said antibody constant domain is aligned with the sequence of SEQ ID NO:6, and characterized in that the antibody constant domain further comprises a residue selected from the group consisting of A, G, I, L, R, S, T, P, at a position corresponding to position 77 of SEQ ID NO:6, and a residue selected from the group consisting of D, E, N, and Q at a position corresponding to residue 43 of SEQ ID NO:6, and wherein the K located at a position corresponding to residue 80 of SEQ ID NO:6 comprises a side chain with an F-amino group covalently attached to a linker, and pharmaceutically acceptable salts, stereoisomers, tautomers, solvates, and prodrugs thereof.

14. The polypeptide as claimed in claim 13, further comprising a residue selected from the group consisting of V, I and L at a position corresponding to residue 42 of SEQ ID NO:6.

15. (canceled)

16. The polypeptide as claimed in claim 13, wherein the residues at positions corresponding to residues 42 and 43 of SEQ ID NO:6 are not in α-helical formation.

17. (canceled)

18. (canceled)

19. (canceled)

20. (canceled)

21. The polypeptide as claimed in claim 13, wherein the mammalian antibody constant domain is a humanized or human domain.

22. The polypeptide as claimed in claim 13, wherein the constant domain is connected to an antibody variable domain.

23. (canceled)

24. The polypeptide as claimed in claim 13, wherein the linker comprises a formula selected from the group consisting of X.sup.1-Y.sup.1-Z.sup.1, X.sup.1-Φ-Y.sup.1-Z.sup.1, and X.sup.1-Y.sup.1-Φ-Z, wherein Φ is a cleavable group, X.sup.1 is a group covalently connectable to at least one Effector Moiety, Y.sup.1 is a linear or branched connecting chain, and Z is a group covalently connected to the ε-amino group of the side chain of K that is located at a position corresponding to residue 80 of SEQ ID NO:6.

25. The polypeptide as claimed in claim 24, wherein the cleavable group Φ is present, and is of the formula ##STR00137## wherein the wavy line and parallel line each indicate a point of attachment to either the X.sup.1, Y.sup.1 or Z.sup.1 group as appropriate.

26. The polypeptide as claimed in claim 24, wherein the linker is selected from the group consisting of: ##STR00138## ##STR00139## ##STR00140## ##STR00141## wherein m, n, j and k are each independently a range whose lower limits are selected from the group consisting of 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, and 20, and whose upper limit is selected from the group consisting of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, and 30, and wherein the overall length of the linker does not exceed 200 atoms.

27. (canceled)

28. (canceled)

29. The polypeptide as claimed in claim 24, wherein the Effector Moiety is a therapeutic agent, protein, peptide, nucleic acid, aptamer, small molecule, protein agonist, protein antagonist, metabolic regulator, hormone, toxin, growth factor, or diagnostic agent.

30. The polypeptide as claimed in claim 29, wherein the Effector Moiety is a toxin, and comprises the formula: ##STR00142## or a pharmaceutically acceptable salt or solvate thereof, wherein, independently for each occurrence, W.sup.2 is ##STR00143##  R.sup.11 is ##STR00144## Y.sup.2 is —C.sub.2-C.sub.20 alkylene-, —C.sub.2-C.sub.20 heteroalkylene-; —C.sub.3-C.sub.8 carbocyclo-, -arylene-, —C.sub.3-C.sub.8heterocyclo-, —C.sub.1-C.sub.10alkylene-arylene-, -arylene-C.sub.1-C.sub.10alkylene-, —C.sub.1-C.sub.10alkylene-(C.sub.3-C.sub.8carbocyclo)-, —(C.sub.3-C.sub.8carbocyclo)-C.sub.1-C.sub.10alkylene-, —C.sub.1-C.sub.10alkylene-(C.sub.3-C.sub.8heterocyclo)- or —(C.sub.3-C.sub.8 heterocyclo)-C.sub.1-C.sub.10alkylene-; Z.sup.2 is ##STR00145## 10 R.sup.12 is hydrogen, C.sub.1-C.sub.8 alkyl or C.sub.1-C.sub.8 haloalkyl;  R.sup.13A and R.sup.13B are either of the following: (i) R.sup.13A is hydrogen, C.sub.1-C.sub.8 alkyl, C.sub.1-C.sub.8 haloalkyl, C.sub.3-C.sub.8 carbocyclyl, C.sub.1-C.sub.10 heterocyclyl, aryl, heteroaralkyl, aralkyl or halogen; and R.sup.13B is C.sub.1-C.sub.8 alkyl, C.sub.1-C.sub.8 haloalkyl, C.sub.3-C.sub.8 carbocyclyl, C.sub.1-C.sub.10 heterocyclyl, aryl, heteroaralkyl or aralkyl or halogen; or (ii) R.sup.13A and R.sup.13B taken together are C.sub.2-C.sub.8 alkylene or C.sub.1-C.sub.8 heteroalkylene; R.sup.14A and R.sup.14B are either of the following: (i) R.sup.14A is hydrogen, C.sub.1-C.sub.8 alkyl, C.sub.1-C.sub.8 haloalkyl, C.sub.3-C.sub.8 carbocyclyl, C.sub.1-C.sub.10 heterocyclyl, aryl, heteroaralkyl or aralkyl; and R.sup.14B is hydrogen, C.sub.1-C.sub.8 alkyl, C.sub.1-C.sub.8 haloalkyl, C.sub.3-C.sub.8 carbocyclyl, C.sub.1-C.sub.10 heterocyclyl, aryl, heteroaralkyl or aralkyl; or (ii) R.sup.14A and R.sup.14B taken together are C.sub.2-C.sub.8 alkylene or C.sub.1-C.sub.8 heteroalkylene; R.sup.15 is ##STR00146## C.sub.1-C.sub.10 heterocyclyl, C.sub.3-C.sub.8 carbocycly and C.sub.6-C.sub.14 aryl optionally substituted with 1, 2, 3, 4 or 5 groups independently selected from the group consisting of —C.sub.1-C.sub.8 alkyl, —C.sub.1-C.sub.8 alkyl-N(R′).sub.2, —C.sub.1-C.sub.8 alkyl-C(O)R′, —C.sub.1-C.sub.8 alkyl-C(O)OR′—O—(C.sub.1-C.sub.8 alkyl), —C(O)R′, —OC(O)R′, —C(O)OR′, —C(O)N(R′).sub.2, —NHC(O)R′, —S(O).sub.2R′, —S(O)R′, —OH, halogen, —N.sub.3, —N(R′).sub.2, —CN, —NHC(═NH)NH.sub.2, —NHCONH.sub.2, —S(═O).sub.2R′ and —SR′, wherein each R′ is independently selected from the group consisting of hydrogen, C.sub.1-C.sub.8 alkyl and unsubstituted aryl, or two R′ can, together with the nitrogen to which they are attached, form a C.sub.1-C.sub.10 heterocyclyl; or R.sup.15 is ##STR00147## optionally substituted with 1, 2, 3, 4 or 5 groups independently selected from the group consisting of C.sub.1-C.sub.8 alkyl, —C.sub.1-C.sub.8 alkyl-N(R′).sub.2, —C.sub.1-C.sub.8 alkyl-C(O)R′, —C.sub.1-C.sub.8 alkyl-C(O)OR′, —O—(C.sub.1-C.sub.8 alkyl), —C(O)R′, —OC(O)R′, —C(O)OR′, —C(O)N(R′).sub.2, —NHC(O)R′, —S(O).sub.2R′, —S(O)R′, —OH, halogen, —N.sub.3, —N(R′).sub.2, —CN, —NHC(═NH)NH.sub.2, —NHCONH.sub.2, —S(═O).sub.2R′, —SR′ and arylene-R′, wherein each R′ is independently selected from the group consisting of hydrogen, C.sub.1-C.sub.8 alkyl, C.sub.1-C.sub.8heterocyclyl, C.sub.1-C.sub.10alkylene-C.sub.3-C.sub.8heterocyclyl and aryl, or two R′ can, together with the nitrogen to which they are attached, form a C.sub.1-C.sub.10 heterocyclyl; R.sup.16 is hydrogen, —C.sub.1-C.sub.8 alkyl, —C.sub.2-C.sub.8 alkenyl, —C.sub.2-C.sub.8 alkynyl or —C.sub.1-C.sub.8 haloalkyl; R.sup.22 is hydrogen, C.sub.1-C.sub.4 alkyl, C.sub.1-C.sub.10 heterocyclyl or C.sub.6-C.sub.14 aryl; R.sup.23 is C.sub.1-C.sub.10 heterocyclyl; and R.sup.17 is independently selected for each occurrence from the group consisting of F, Cl, I and Br; R.sup.20 is -aryl, —C.sub.1-C.sub.10alkylene-aryl, where aryl on R.sup.10 comprising aryl is substituted with [R.sup.17].sub.h; h is 5; and X is O or S; provided that when R.sup.13A is hydrogen X is S.

31. The polypeptide as claimed in claim 29, wherein the Effector Moiety is a toxin, and is selected from the group consisting of ##STR00148## ##STR00149##

32. (canceled)

33. (canceled)

34. (canceled)

35. (canceled)

36. (canceled)

37. The polypeptide as claimed in claim 13, wherein the CL domain is a kappa domain (CLκ).

38. (canceled)

39. (canceled)

40. The polypeptide as claimed in claim 3, wherein the CL domain is a lambda domain (CLλ).

41. (canceled)

42. (canceled)

43. The polypeptide as claimed in claim 13, comprising a sequence selected from the group consisting of SEQ ID NO:37, SEQ ID NO:38, SEQ ID NO:39, SEQ ID NO:40, SEQ ID NO:41, SEQ ID NO:42, SEQ ID NO:44, SEQ ID NO:46, SEQ ID NO:47, SEQ ID NO:49, SEQ ID NO:50, SEQ ID NO:51, SEQ ID NO:52, SEQ ID NO:54, SEQ ID NO:60, SEQ ID NO:61, SEQ ID NO:77, SEQ ID NO:78, SEQ ID NO:94, SEQ ID NO:95, SEQ ID NO:96, SEQ ID NO:97, SEQ ID NO:104, SEQ ID NO:105, SEQ ID NO:106, SEQ ID NO:119, SEQ ID NO:120, SEQ ID NO:121, SEQ ID NO:122, SEQ ID NO:127, SEQ ID NO:128, SEQ ID NO:129, SEQ ID NO:134, SEQ ID NO:135, SEQ ID NO:140, SEQ ID NO:141, SEQ ID NO:142, SEQ ID NO:143, SEQ ID NO:144, SEQ ID NO:148, SEQ ID NO:149, SEQ ID NO:150, SEQ ID NO:151, SEQ ID NO:156, SEQ ID NO:158, SEQ ID NO:159, SEQ ID NO:160, SEQ ID NO:161, SEQ ID NO:163, SEQ ID NO:165, SEQ ID NO:167, SEQ ID NO:168, SEQ ID NO:169, SEQ ID NO:236, SEQ ID NO:237, SEQ ID NO:238, SEQ ID NO:239, SEQ ID NO:240, SEQ ID NO:241, and SEQ ID NO:254, or a polypeptide at least about 85% identical thereof.

44. A composition comprising the polypeptide as claimed in claim 13, wherein at least about 70% of the Linker in the composition or sample is conjugated to the ε-amino group of the side chain of K that is located at a position corresponding to residue 80 of SEQ ID NO:6.

45. A composition comprising multiples of the polypeptide as claimed in claim 13, wherein at least about 70% of the polypeptides comprises a Linker covalently attached to the ε-amino group of the side chain of K that is located at a position corresponding to residue 80 of SEQ ID NO:6.

46. An antibody, or antigen binding portion thereof, wherein the antibody comprises the polypeptide as claimed in claim 13.

47. The antibody as claimed in claim 46, wherein the antibody is a full length antibody, Fab, Fab′, F(ab′).sub.2, V.sub.H, diabody, or minibody.

48. The antibody as claimed in claim 13, wherein the antibody comprises VH and VL domains from an antibody selected from the group consisting of h38C2, rituximab, cetuximab, infliximab, adalimumab, natalizumab, omalizumab, ranibizumab, trastuzumab and palivizumab.

49.-58. (canceled)

Description

DETAILED DESCRIPTION OF FIGURES

[0275] FIG. 1A, 1B, 1C, and 1D: Intact molecular weight analysis of MAC by mass spectrometry demonstrates that multiple peptides are attached to the anti-IGF1R antibody 2.12.1.fx. FIG. 1A: mass spectrometry data of anti-IGF1R antibody 2.12.1.fx. FIG. 1B-1D: mass spectrometry data of MAC-2, showing replicate experiments of 3 individual lots.

[0276] FIGS. 2A, 2B, 2C, 2D, 2E, 2F, 2G, and 2H: Mass spectrometry data of 2.12.1.fx (IGF1R) and 3 lots of MAC-2 (MAC) where the disulfide bonds have been reduced. FIG. 2A: Mass spectrometry data of 2.12.1.fx (IGF1R), light chain. FIG. 2B: Mass spectrometry data of 2.12.1.fx (IGF1R), heavy chain. FIG. 2C: mass spectrometry data of light chain of MAC-2, lot-1. FIG. 2D: mass spectrometry data of heavy chain of MAC-2, lot-1. FIG. 2E: mass spectrometry data of light chain of MAC-2, lot-2. FIG. 2F: mass spectrometry data of heavy chain of MAC-2, lot-2. FIG. 2G: mass spectrometry data of light chain of MAC-2, lot-3. FIG. 2H: mass spectrometry data of heavy chain of MAC-2, lot-3.

[0277] FIG. 3A: Amino acid sequence of light chain of antibody 2.12.1.fx with chymotrypsin cleavage sites noted with bullets. Chymotryptic fragments that contain a Lys residue (site of potential conjugation) are labeled by number from the N-terminus. The Y15 fragment of the light chain is underlined. FIG. 3B: Amino acid sequence of heavy chain of antibody 2.12.1.fx with chymotrypsin cleavage sites noted with bullets. Chymotryptic fragments that contain a Lys residue (site of potential conjugation) are labeled by number from the N-terminus.

[0278] FIG. 4A: Mass spectrometry data of a conjugated lysine-containing peptide: light chain Y15, showing mass spectrometry data for unconjugated anti-IGF1R antibody 2.12.1.fx (IGF1r) and MAC-2 (MAC), as well as a representation of the Y15 fragment. FIG. 4B: Mass spectrometry data of un-conjugated light chain Y15 fragment, showing mass spectrometry data for unconjugated anti-IGF1R antibody 2.12.1.fx (IGF1r) and MAC-2 (MAC), as well as a representation of the Y15 fragment.

[0279] FIG. 5A: The selected ion LCMS chromatogram data for the tryptic fragment of 2.12.1.fx. FIG. 5B: The selected ion LCMS chromatogram data for the tryptic fragment when LC-K.sup.188 is modified with ABP of MAC-2.

[0280] FIG. 6A: The selected ion LCMS chromatogram data for the tryptic fragment of 2.12.1.fx. FIG. 6B: The selected ion LCMS chromatogram data for the tryptic peptide when LC-K.sup.190 is modified with ABP of MAC-2.

[0281] FIG. 7A: Mass spectra of intact MAC-2. FIG. 7B: Mass spectra of reduced heavy chain for MAC-2. FIG. 7C: Mass spectra of reduced light chain for MAC-2.

[0282] FIG. 8A: Amino acid sequence alignment of the variable domains of m38c2, h38c2, and human germlines. Framework regions (FR) and complementarity determining regions (CDR) are defined according to Kabat et al. Asterisks mark differences between m38c2 and h38c2 or between h38c2 and the human germlines.

[0283] FIG. 8B: Amino acid sequence alignment of murine constant light chain kappa region (mCLκ), human constant light chain kappa region (hCLκ), and human constant light chain lambda region (hCLλ). Differences between mCLκ and hCLκ; and between hCLκ and hCLλ; are shown as asterisks, and conserved substitutions are shown as crosses. β-strands A-G are underlined. The turn between β-strands A and B and the α-helix between β-strands E and F are each indicated in italics. Di-sulfide bond-forming cysteines between the first β-sheet (made up of β-strands ABDE; single underline) and the second β-sheet (made up of β-strands CGF, double underline) are indicated by §. Known polymorphic loci in the human sequences are indicated in bold.

[0284] FIG. 9A: Binding ELISA data for HER2 receptor binding of trastuzumab and trastuzumab-[CLκ-D.sup.185A] conjugation products to [PEG.sub.5-K.sup.11-SEQ:27]. FIG. 9B: Binding ELISA data for HER2 receptor binding of trastuzumab and trastuzumab-[CLκ-D.sup.185A] conjugation products to MMAD toxin.

[0285] FIG. 10A: representation of a constant Ig domain showing the 7 β-strands forming the two β-sheets. FIG. 10B: close up of the α-helix between β-sheets E and F.

[0286] FIG. 11. Crystal structure-based minimized ribbon representation of CLκ, showing the halo-phenyl ester reactive ‘binding site’ (small jacks) within the overall steric ‘binding pocket’ created by the 3D structure. B-strands are labeled.

[0287] FIG. 12. Crystal structure-based minimized ribbon representation of the CLκ ‘binding pocket’, showing CLκ-D.sup.77 and CLκ-D.sup.43 (as a stick model) in the hydrogen bond with CLκ-H.sup.81 Nε or Nδ, and atomic distances in Å.

[0288] FIGS. 13A and 13B. Crystal structure-based minimized ribbon representation of the CLκ and CLκ-D.sup.77A mutant ‘binding pockets’. The distance between carbonyl oxygen of CLκ-D.sup.43 and Nδ of CLκ-H.sup.81 differs by 1Å between the CLκ and CLκ-D.sup.77A mutant, pointing to the predominance of catalytically active CLκ-H.sup.81 tautomer Nδ in the CLκ-D.sup.77A mutant. In addition, the modeling identifies a clear increase in the overall size of the pocket in CLκ-D.sup.77A mutant, which is represented by the figure. FIG. 13C. Crystal structure-based ribbon representation of the CLκ (grey) and CLκ-D.sup.77A mutant (black) ‘binding pockets’ superimposed over each other. In the CLκ-D.sup.77A mutant, CLκ-H.sup.81 is shifted toward CLκ-D.sup.43 by the hydrogen bond interactions, as CLκ-D.sup.77A is unable to form a hydrogen bond with Nε of CLκ-H.sup.81.

[0289] FIG. 14. Crystal structure-based minimized ribbon representation of the CLκ ‘binding pocket’. The binding site is depicted as a small jacks. The π electron stacking interactions with CLκ-H.sup.81 are shown, maintaining the imidazole ring at the optimum position in relation to the incoming halo-phenyl ester substrate.

[0290] FIGS. 15A and 15B. Crystal structure-based minimized ribbon representation of the CLκ and CLλ ‘binding pockets. FIG. 15A shows the 7-electron interactions between CLκ-V.sup.42 and CLκ-H.sup.81, assisting in maintaining the CLκ-H.sup.81 imidazole ring and the Nε electron pair at the plane needed for nucleophilic attack during catalytic reaction. The distance between the center of CLκ-H.sup.81 imidazole ring and each of the hydrogen atoms on the CLκ-V.sup.42 are 2.8Å allowing for strong interactions. In FIG. 15B, CLλ-A.sup.49 is shown at a distance of 4.2Å from CLλ-H.sup.82 (identified as H81 in the figure for the purposes of clarity of comparison). This distance is modeled as likely too far to have a significant influence on the position or tautomeric form of CLλ-H.sup.82.

[0291] FIG. 16. Sequence alignment of hCHλ1, hCHλ2, hCHλ3, hCLκ and hCLλ. β-strands of the CLκ are indicated as underlined regions. α-helices are indicated in italics.

[0292] FIG. 17. Crystal structure based alignment of sequences of hCHλ1, hCHλ2, hCHλ3, hCLκ and hCLλ according to minimized 3D homology. β-strands are indicated as boxed regions, α-helices are indicated within wavy scrolls. Key residues corresponding to CLκ-V.sup.42, CLκ-D.sup.43, CLκ-D.sup.77, CLκ-K.sup.80, and CLκ-H.sup.81 are identified with rectangular dotted-line boxes extending vertically between sequences. The crystal structure modeling of the domains that generated this alignment suggested a short break in the D β-strand in the hCHλ1, hCHλ2, hCHλ3, and hCLλ domains. Other modeling and crystal structure analysis indicates that most, if not all antibody constant domains comprise 7 β-strands, and the D β-strand is contiguous. The two D β-strands modeled herein have accordingly been indicated as D′ and D″. FIGS. 18B-31B, depicting CLλ, and CHγ domains, are orientated such that the D β-strand is on the lower left of the structure pointing downwards, and lying against the E β-strand. The D′ and D″ β-strands together can be seen to occupy approximately the same relative position as the CLκ D β-strand (β-strands D and D′, D″ labelled in FIG. 18 as a point of reference).

[0293] FIGS. 18A and 18B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCLλ (B) domains.

[0294] FIGS. 19A and 19B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and WT-hCHγ1 (B) domains.

[0295] FIGS. 20A and 20B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and WT-hCHγ1 (B) domains, showing the sidechain location of significant residues.

[0296] FIGS. 21A and 21B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ1-m1-D44 mutant (B) domains, showing the sidechain location of significant residues.

[0297] FIGS. 22A and 22B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ1-T78K/Q79H/CD loop swap mutant (B) domains, showing the sidechain location of significant residues.

[0298] FIGS. 23A and 23B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and WT-hCHγ2 (B) domains.

[0299] FIGS. 24A and 24B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and WT-hCHγ2 (B) domains, showing the sidechain location of significant residues.

[0300] FIG. 25A and FIG. 25B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ2m mutant (B) domains, showing the sidechain location of significant residues.

[0301] FIGS. 26A and 26B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ2m-D.sup.82A mutant (B) domains, showing the sidechain location of significant residues.

[0302] FIGS. 27A and 27B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and WT-hCHγ3 (B) domains.

[0303] FIGS. 28A and 28B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and WT-hCHγ3 (B) domains, showing the sidechain location of significant residues.

[0304] FIGS. 29A and 29B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ3m mutant (B) domains, showing the sidechain location of significant residues.

[0305] FIGS. 30A and 30B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ3m CD/EF mutant (B) domains, showing the sidechain location of significant residues.

[0306] FIGS. 31A and 31B. Crystal structure-based minimized ribbon representation of the hCLκ (A) and hCHγ3m CD/EF-m2 mutant (B) domains, showing the sidechain location of significant residues.

[0307] FIGS. 32A and 32B. Alignment of CHγ1, CHγ2, and CHγ3 with their respective mutants and proposed mutant, together with CLκ and CLλ.

EXAMPLES

[0308] In order that this invention may be better understood, the following examples are set forth. These examples are for purposes of illustration only and are not to be construed as limiting the scope of the invention in any manner.

Description of Conjugation Additions (CA), Average CA

[0309] Conjugation additions (CA) are measured on an antibody scaffold using intact mass measurement by mass spectrometry. Upon conjugation, the overall mass of the intact product increases by the mass and number of additions of the conjugated peptide, toxin, etc. If multiple additions occur then a distribution of conjugate forms is observed in a mass spectra and the observed signal intensity of each conjugate form gives a quantitative measurement. This analysis is routinely presented in a table by listing each CA form as a percentage of all the observed CA forms. Average CA (e.g.: the overall number of CA present on a scaffold) is an additional value that describes the average conjugate load. An example is provided below of 2 conjugated drug products. Example 1 has an even distribution of CA with an average CA=2.00. Example 2 has a distribution that heavily favors the presence of 2 CA with a minimal amount of other conjugation forms. The average CA is similar between these examples (2.13 vs. 2.00); however, Example 1 is a more heterogeneous product comprised of more conjugation forms, while Example 2 is a more homogeneous product that contains mostly 2 CA.

TABLE-US-00003 TABLE 1 Average 0 CA % 1 CA % 2 CA % 3 CA % 4 CA % CA Example 1 11 22 33 22 11 2.00 Example 2 2 4 77 12 5 2.13

[0310] A similar analytical treatment is also possible on these antibody scaffolds after disulfide bonds have been reduced to generate free light chains and heavy chains. Measurement of the intact mass of conjugated light/heavy chains can provide information about the location of CA on these respective subunits.

Explanation of Directional Conjugation to CLκ-K″

[0311] To determine site specific attachment of drug conjugates to an antibody scaffold, a peptide map is produced. Peptide maps are the analysis of a protein sequence in detail to characterize peptide produced following a proteolytic digestion of the conjugated drug product. Once the protein is digested then the resulting peptides are analyzed by reversed phase liquid chromatography with mass spectrometry detection (RPLC/MS). The presence of conjugate additions on discreet amino acid residues is observed as a corresponding mass shift compared to the un-conjugated peptide. This process has been repeated on multiple antibodies conjugated with multiple conjugate additions using PFP reactive esters to target lysine residues. (Data is presented below, and also in WO2012007896, whose contents are herein incorporated entirely). In these studies, the following observations were consistent: 1—conjugate additions were observed more frequently on the LC than the HC, 2-CLκ-K.sup.80 is the specific preferred residue that is modified, 3—multiple other locations are also modified on both the LC and HC; however, each alternative site is modified at a low level. To summarize, halo-phenyl ester conjugation results in preferred modification of CLκ-K.sup.80 and additional conjugation is distributed at a low level across multiple residues. For this reason, the conjugation process is generally optimized to result in high % CA values for 2 conjugate additions because this promotes a product that is fully conjugated at a single location on each LC. While elevated % CA levels of 0-1 conjugates result in preferable CLκ-K.sup.80 modification, these conjugate forms also represent a significant amount of un-reacted scaffold. Products that display average CA values significantly greater than 2 suggest the presence of conjugate additions that are not targeted at discreet residues. When HC and LC are reduced and analysed separately, the % value of 1 CA indicates the % conjugation on single LC species, and is thus reliable indicator of the efficiency of conjugation to the CLκ-K.sup.80 residue.

Example 1 Exemplary Synthesis of Peptides Used in the Invention

[0312] ##STR00066##

[0313] Scheme 1: Solid phase synthesis of a peptide chain using Fmoc chemistry (exemplified with a typical Ang2-binding peptide (ABP) SEQ ID NO:27. TFA/water/phenol/triisopropylsilane (90:4:4:2). Rink Amide Resin. Steps for SPPS using Fmoc chemistry: (i) Fmoc removal with 20% piperidine/DMF, (ii) Amino acid coupling; HBTU:Amino acid:HOBt:NMM ratio relative to resin amine loading is 5:5:5:20. Solvent used was NMP, (iii) Repeat steps for each amino acid coupling. X=acid-labile side chain protecting group. Completed assembly of fully-protected, resin-bound peptide:

Synthesis of Peptide-Thiol-Linker Compounds

[0314] ##STR00067##

[0315] An Ang2-binding peptide (ABP; SEQ ID NO:27) (284 mg, 0.1 mmol) was dissolved in dimethylformamide (0.5 ml) with stirring. Separately, S-Trityl-mercaptopropionic acid (MPA, 62 mg, approx 0.125 mmol), HBTU (48 mg, 0.125 mmol) and N-methylmorpholine (0.025 ml, 0.25 mmol) were stirred in DMF (0.5 ml) for 5 min until dissolved. The ABP solution and activated MPA solutions were mixed together for 2 hrs. Progress of the reaction was monitored by LCMS. After 2 hrs, the solution was slowly added to ice-cold ether (40 ml) to precipitate the ABP-S-trityl-MPA product. The white precipitate was collected by filtration then dried. The solid residue was then dissolved in a solution of trifluoroacetic acid in dichloromethane (1:10, 10 ml), with triisopropylsilane (TIPS) added (0.050 ml) and stirred for 1 hr. The solution was evaporated under reduced pressure to a light-yellow oil then the crude thiol peptide precipitated by the addition of ice-cold ether. The product was collected by centrifugation and dried in vacuo. The residue was dissolved in 50% aqueous acetonitrile then lyophilized to yield the crude thiol peptide (approx 80% pure by HPLC analysis). The crude thiol peptide was purified by semi-preparative HPLC to yield 145 mg of SEQ ID NO:27-K.sub.SH.sup.11.

Generation of Ang-2-Binding-Peptide-Thiol Intermediates

[0316] Peptide chain assembly was conducted on a 0.1 mmol scale. The resin used was Fmoc-Rink-PL resin (150 mg, 0.67 mmol/g substitution). Standard Fmoc chemistry protocols were used to assemble the peptide. Fmoc removal was with 20% piperidine/DMF for 3×5 min. and all resin washing steps used DMF. To incorporate the amino acids, a single coupling step was employed for each residue, using HBTU/HOBt/NMM activation, for a 2 hr period. The Linking Residue (Kai) was incorporated as Fmoc-Lys(N.sup.ε-mercaptopropionate-S-Trt)-OH. Upon chain assembly, the N-terminal Fmoc group was removed and the peptidoresin capped by acetylation. The final resin was washed with DCM and dried overnight in vacuo.

[0317] Acidolytic removal of protecting groups and cleavage of the peptide from the resin was achieved using a cocktail of TFA/water/dithiothreitol/triisopropylsilane (ratio 90:4:4:2, 5 ml) for 2 hrs. The solution was filtered from the resin and the resin washed with another 5 ml of neat TFA. The combined filtrates were evaporated to a syrup then addition of ice-cold ether precipitated a white powder. The powder was collected by centrifugation then dissolved in 50% aqueous acetonitrile (20 ml), frozen and lyophilized overnight.

[0318] A preparative HPLC column was pre-equilibrated with dilute aqueous TFA and acetonitrile. The crude ABP-thiol intermediates (i.e. ABP with K.sub.SH as linking residue) was dissolved in DMF (3 ml), then adsorbed onto the column and eluted by applying a gradient of acetonitrile in dilute TFA. Fractions were collected automatically by mass (M=1465). Elution from the column was monitored by UV, the fractions obtained were analyzed by analytical RP-HPLC.

Example 2 Conjugation Strategies

[0319] 5 different conjugation strategies were considered for conjugating peptides to antibodies (exemplary structures are shown using SEQ ID NO:27-K.sub.SH.sup.11 and 2.12.1.fx) (full details are provided in the Examples of PCT/US2011/053092, filed 11 Jul. 2011, whose contents are hereby incorporated entirely). Briefly, NHS esters, maleimide, squarate esters, AZD and halo-phenyl esters were all investigated as potential mechanisms to develop directional conjugation to antibodies.

##STR00068##

[0320] NHS esters suffered from the problems of slowly converting to a free acid form, where the NHS ester is converted to an inactive carboxyl. It was concluded that although some success was obtained with NHS esters, it appeared that the aqueous lability of the resulting NHS-ester may limit their application in subsequent conjugation reactions. Further tests of NHS-PEG2-MAL are shown below (comprising Z* group Z13).

[0321] The ethyl squarates conjugate well to free thiols but poorly to free amines on proteins and antibodies unless the pH is above 9.

[0322] In general, the maleimide-activated peptides did not conjugate well to proteins or antibodies which lack either an endogenous thiol (derived from a free cysteine side chain) or a thiol introduced by other chemical means, e.g. via Traut's reagent.

[0323] AZD reacted slowly with antibody amino groups, and attempts to increase the pH to 7-9 yielded low levels of conjugation and high levels of AZD hydrolysis (in order to increase the nucleophilic tendency of the antibody surface lysines by decreasing their charge, as the pKa of surface lysines is about 9.1 to 11.2).

Example 3 Synthesis of Pentafluorophenyl Esters (PFP)

[0324] ##STR00069##

[0325] The present invention also provides for the use of pentafluorophenyl (PFP; Z*=Z1) esters to form relatively stable Effector Moiety-linker complexes. This method has several advantages over other approaches in that the PFP group can be introduced in solution easily from a stable activated peptide product, which itself can be purified using standard HPLC methods with little PFP ester hydrolysis observed.

[0326] The present invention provides a synthetic route whereby an activated ester group, such as PFP, can be coupled directly to a side chain lysine on the peptide by either a chemoselective reaction (using thiol/maleimide chemistry) or by using a bis-active ester reagent, which forms an amide with the peptide side chain but leaves the other end as the active ester.

[0327] In some embodiments, the strategy may be a bis-acid PEG with each acid activated as a PFP ester. In organic solutions, with some base present, the end of the bis-PFP linker reacted with the N-ε-amino side chain of lysine in the required tether position to form a stable amide linkage, while the other end maintained the other PFP group. One potential problem with this strategy is the possibility of forming peptide dimers, where a peptide would add to each of the PFP moieties present at each end of the linker. In some aspects, the present invention overcomes this additional problem by altering the stoichiometry and addition of the respective peptide and bis-PEG-PFP linker. One solution provided by the invention is to have an excess of the bis-Pfp linker in solution and slowly add the peptide in solution, such that an excess of linker over peptide is always present. By having a ratio of between about 3.7:1 to about 4.3:1, or in some embodiments, a ratio of about 4:1, of linker over peptide, the required PFP-activated peptide can be synthesized with no dimer present. The synthesis scheme for [PFP-PEG.sub.5-K.sup.11-SEQ:27] is shown below in Scheme 3.

Synthesis of [PFP-PEG.SUB.5.-K11-SEQ:27]

[0328] ##STR00070##

[0329] Bis-dPEG.sub.5-acid (1 mmol, 338 mg) was dissolved in anhydrous dichloromethane (5 ml) then pentafluorophenol (2 mmol, 368 mg) was added, along with dicyclohexycarbodiimide (1 mmol, 208 mg). The solution was stirred overnight at RT. After this time, the fine white dicyclohexylurea side-product was filtered off and the filtrate evaporated to dryness to give a pale yellow light oil. Analysis by TLC and HPLC indicated a pure product with correct MS=670. The product was used in the next step without further purification. The product is stable for several months at −20° C.

[0330] SEQ ID NO:27 (730 mg) was dissolved in anhydrous dimethylformamide (8 ml) and N-methylmorpholine (0.05 ml) added. An aliquot of neat bis-dPEG.sub.5-OPfp reagent (0.5 ml) was placed in a glass vial (20 ml). With vigorous stirring, the SEQ ID NO:27/NMM solution was added in 4×2 ml aliquots to the bis-dPEG.sub.5-OPfp reagent over 2 hr, then the final mixture stirred for a further 1 hr. Progress of the conversion to [PFP-PEG.sub.5-K.sup.11-SEQ:27] product was monitored by analytical HPLC. At the end of the reaction, the solution was filtered and directly purified by semi-preparative HPLC on a 1″ C8 column. The purest fractions (>95% by analytical HPLC) were combined and lyophilized to give 400 mg (48% yield) of final [PFP-PEG5-K.sup.11-SEQ:27] peptide-linker product. A similar mechanism can be used to generate [PFP-PEG2-MAL-K.sub.SH.sup.11-SEQ:27] (see Scheme 4).

Synthesis of [PFP-PEG.SUB.2.-K.SUB.SH..SUP.11.-SEQ:27]

[0331] ##STR00071##

[0332] Maleimide-dPEG.sub.2-acid (328 mg, 1 mmol, Quanta Biodesign), pentafluorophenol (0.103 ml, 1 mmol, PFP) and dicyclohexylcarbodiimide (206 mg, 1 mmol, DCC) were dissolved in dry DCM (10 ml) and stirred for 1 hr at RT. The fine white precipitate (DCU side-product) that formed was removed by filtration and the filtrate evaporated to dryness in vacuo. The product was obtained as a fine white powder in high yield (490 mg, quantitative). Purity was >95% by analytical HPLC; MS showed [M+H].sup.+=495.

[0333] A sample (30-40 mg) of SEQ ID NO:27-K.sub.SH.sup.11 was dissolved in anhydrous DMF (2 ml). Maleimide-PEG2-PFP (20 mg) was added along with N-methylmorpholine (5 mL). The reaction was stirred and monitored at RT by HPLC to follow the time-course of product formation. The complete conversion of starting peptide to PFP-activated product was observed within the first 2 hrs. The solution was filtered and the product peak directly isolated by semi-preparative HPLC. In each case, the product was isolated in approximately 40% yield after lyophilization.

Example 4 Antibody Conjugation

[0334] The MAC-1 and MAC-2 exemplary antibody-Effector Moiety conjugates were made by conjugating the antibody 2.12.1.fx (SEQ ID NO:1 and SEQ ID NO:2) with an Ang2 binding peptide (SEQ ID NO:27). MAC-1 comprises 2.12.1.fx conjugated to [PFP-PEG2-MAL-K.sub.SH.sup.11-SEQ:27] to yield 2.12.1.fx-[PEG2-MAL-K.sub.SH.sup.11-SEQ:27] and MAC-2 comprises 2.12.1.fx conjugated to [PFP-PEG.sub.5-K.sup.11-SEQ:27] to yield 2.12.1.fx-[PEG5-K.sup.11-SEQ:27].

Generation of MAC-1

[0335] ##STR00072##

Generation of MAC-2

[0336] ##STR00073##

[0337] The number of peptide conjugations per 2.12.1.fx antibody in a sample of each MAC was calculated (see Table 2).

TABLE-US-00004 TABLE 2 Conjugation profile of MAC-1 and MAC-2. Conjugation Additions (CA) (%) 0 1 2 3 4 Avg CA MAC-1 3 26 42 25 3 1.97 MAC-2 2 20 47 26 5 2.12

Example 5 Optimizing Conditions for PFP-Based Conjugation

[0338] A series of assays were run to establish optimal reaction conditions for directed conjugation. At the end of each reaction conjugation, the reaction was quenched with a succinate and glycine buffer, lowering the pH to approximately 5.5 and quenching any free peptide or peptide/linker. MAC-2 analysis was conducted by measuring the intact molecular weight (MW) of the MAC using electrospray time-of-flight mass spectrometry detection following protein separation from salts and excipients through a size exclusion chromatography column.

Temperature

[0339] 2.12.1.fx antibody was adjusted to 18 mg.Math.ml.sup.−1 at pH 7.7 with a phosphate buffer to a final concentration of 0.06M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 4.3:1 and allowed to react for 2 hrs at 18, 22, or 25° C. Results are presented in Table 3.

TABLE-US-00005 TABLE 3 Reaction temperature in 0.06M phosphate at 4.3:1 peptide:antibody. CA (%) Temp 0 1 2 3 4 Avg CA 18 C. 1 16 51 23 8 2.21 22 C. 3 15 57 21 5 2.11 25 C. 2 12 53 25 7 2.24

Reaction pH

[0340] 2.12.1.fx antibody was adjusted to 18 mg.Math.ml.sup.−1 at pH 6.5, 6.75, 7.0, 7.25, 7.5, 7.75, or 8.0 with a phosphate buffer to a final concentration of 0.06M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT. The results are presented in Table 4.

TABLE-US-00006 TABLE 4 pH in 0.06M sodium phosphate buffer at 4.3:1 peptide:antibody. CA (%) pH 0 1 2 3 4 Avg CA 6.5 7 42 41  9 0 1.51 6.75 3 31 52 12 3 1.83 7.0 3 24 53 16 4 1.94 7.25 2 18 54 22 5 2.12 7.5 2 12 57 23 7 2.23 7.75 3 15 55 22 6 2.15 8.0 1 14 52 29 4 2.21

[0341] 2.12.1.fx was adjusted to 2 mg.Math.ml.sup.−1 at pH 7.0, 7.5 and 8.0 with a HEPES buffer to a final concentration of 0.02M. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in DMSO to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 5:1 and allowed to react overnight at RT. The results are presented in Table 5. The level of conjugation decreased above pH 8.0

TABLE-US-00007 TABLE 5 pH in 0.02M HEPES Buffer at 5:1 peptide:antibody. ABP Additions (%) pH 0 1 2 3 4 Avg CA 7 2 21 41 28 4 2.03 7.5 3 22 44 26 5 2.08 8 9 30 42 17 2 1.73

Duration of Conjugation Reaction

[0342] 2.12.1.fx was adjusted to 18 mg.Math.ml.sup.−1 at pH 7.7 with a phosphate buffer to a final concentration of 0.06M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 4.3:1 and allowed react for 30, 60, 120, 180, 240, 300, or 2400 mins at room temperature (Table 6).

TABLE-US-00008 TABLE 6 Duration of conjugation reaction in 0.06M sodium phosphate at 4.3:1 peptide:antibody. Time CA (%) (mins) 0 1 2 3 4 Avg CA 30 6 38 44 13 0 1.64 60 1 22 52 21 3 2.02 120 0 15 50 29 6 2.24 180 1 12 51 31 5 2.28 240 1 9 51 33 5 2.33 300 1 9 50 35 5 2.35 2400 1 10 48 35 6 2.35

Molar Ratio of Peptide to Protein

[0343] 2.12.1.fx was adjusted 18 mg.Math.ml.sup.−1 to pH 7.5 with a HEPES buffer to a final concentration of 0.2M HEPES. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 1, 2, 3, 4, and 5:1 (Table 7), and reacted for at least 2 hrs at RT, but the high concentration of HEPES buffer resulted in decreased conjugation.

TABLE-US-00009 TABLE 7 Molar ratio of peptide to protein 1:1-5:1 in 0.2M HEPES. CA (%) Peptide:2.12.1.fx 0 1 2 3 4 5 6 7 Avg CA 1:1 80 20 0 0 0 0 0 0 0.20 2:1 60 35 5 0 0 0 0 0 0.45 3:1 39 49 12 0 0 0 0 0 0.73 4:1 27 51 19 3 0 0 0 0 0.98 5:1 11 47 37 5 0 0 0 0 1.36

[0344] 2.12.1.fx was adjusted 18 mg.Math.ml.sup.−1 to pH 7.7 with a phosphate buffer to a final concentration of 0.06M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 5, 7, 10, 12, and 15:1 (Table 8) and allowed to react for 2 hrs at RT to generate a MAC with a higher level of conjugation.

TABLE-US-00010 TABLE 8 Molar ratio of peptide to protein 7:1-15:1 in 0.06M sodium phosphate. CA (%) Peptide:2.12.1.fx 0 1 2 3 4 5 6 7 Avg CA  7:1 1 1 29 39 17 10 2 0 3.06 10:1 1 1 18 33 25 19 3 0 3.49 12:1 3 1 11 22 26 26 8 3 3.92 15:1 1 2 9 19 23 32 12 3 4.22

[0345] To further optimize the molar ratio of 2.12.1.fx and [PFP-PEG5-K.sup.11-SEQ:27], 2.12.1.fx was adjusted 18 mg.Math.ml.sup.−1 to pH 7.7 with a phosphate buffer to a final concentration of 0.06 M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx antibody at a molar ratio of 2.5, 2.8, 3.1, 3.4, 3.7, 4.0, 4.3, or 4.6:1 (Table 9) and allowed to react for 2 hrs at RT.

TABLE-US-00011 TABLE 9 Molar ratio of peptide to protein 2.5:1- 4.6:1 in 0.06M sodium phosphate. CA (%) Peptide:2.12.1.fx 0 1 2 3 4 5 6 7 Avg CA 2.5:1 14 53 30 4 0 0 0 0 1.25 2.8:1 10 45 37 8 0 0 0 0 1.43 3.1:1 7 39 45 8 0 0 0 0 1.53 3.4:1 5 40 44 11 0 0 0 0 1.61 3.7:1 4 25 51 15 5 0 0 0 1.92 4.0:1 2 26 55 15 2 0 0 0 1.89 4.3:1 1 24 55 16 4 0 0 0 1.98 4.6:1 2 19 56 19 5 0 0 0 2.08

[0346] 2.12.1.fx was adjusted to 2 mg.Math.ml.sup.−1 at pH 7.0 with a HEPES buffer to a final concentration of 0.02M. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in DMSO to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 5, 6, 7, 8, 10:1 and allowed to react overnight at RT. The results are presented in Table 10.

TABLE-US-00012 TABLE 10 Molar ratio of peptide to protein 5:1-10:1 in 0.02M HEPES. CA (%) Peptide:2.12.1.fx 0 1 2 3 4 5 6 7 Avg CA 5:1 2 21 49 24 4 0 0 0 2.07 6:1 2 15 42 32 9 0 0 0 2.31 7:1 1 11 34 42 13 0 0 0 2.57 8:1 0 9 32 42 16 1 0 0 2.68 10:1  0 4 21 47 25 4 0 0 3.07

Conjugation Profile of 2.12.1.Fx at Various Protein Concentrations

[0347] The conjugation profiles of 2.12.1.fx with [PFP-PEG5-K.sup.11-SEQ:27] at various concentrations were analyzed. 2.12.1.fx was concentrated to >50 mg/mL, diluted to the desired concentration with 20 mM sodium acetate, 200 mM trehalose pH 5.5, and spiked with 60 mM sodium phosphate pH 7.7. [PFP-PEG5-K.sup.11-SEQ:27] was resuspended with 50% propylene glycol and mixed with the antibody at a 4.3:1 molar ratio and allowed to react overnight at RT. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple conjugation sites are observed as distinct signals separated by the mass difference of a peptide. Relative quantitation of multiple conjugation species is performed by measuring the signal magnitude. Table 11 shows the conjugation profile of 2.12.1.fx with peptide at various concentrations of antibody. At antibody concentrations 10 mg/mL to 50 mg/mL, the conjugation occurs at a distribution between 0-5 addition with an average of 1.8 or greater additions. At antibody concentrations 0.5 to 5 mg/mL, the conjugation occurs at a distribution between 0-3 additions with an average of 1.5 or less additions.

TABLE-US-00013 TABLE 11 Effect of antibody concentration. Antibody Concentration CA (%) (mg/ml) 0 1 2 3 4 5 Avg CA 0.5 65 32 3 — — — 0.37 1 44 44 12 — — — 0.67 5 10 41 40 8 — — 1.45 10 3 30 47 17 2 1 1.87 15 1 24 51 20 3 1 2.02 20 1 16 57 22 2 1 2.11 30 2 20 55 20 3 1 2.04 40 2 21 53 22 2 0 2.04 50 2 19 50 24 4 1 2.11

Reaction Buffer Selection

[0348] 2.12.1.fx was adjusted to 18 mg.Math.ml.sup.−1 at pH 7.7 with a sodium carbonate, sodium borate, or sodium phosphate buffer to a final concentration of 0.05M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 1, 2, 3, 4, or 5:1 and allowed to react for 2 hrs at RT. The low reaction pH resulted in the reduced level of conjugation (Table 12).

TABLE-US-00014 TABLE 12 Buffer and pH alterations. CA (%) Buffer 0 1 2 3 4 Avg CA 50 mM sodium carbonate pH 7.4 2 24 48 26 0 1.98 50 mM sodium borate pH 7.0 1 17 45 31 5 2.20 50 mM sodium phosphate pH 7.0 10 48 38 4 0 1.36

[0349] 2.12.1.fx was adjusted to 18 mg.Math.ml.sup.−1 at pH 7.5, 7.7 and 8.0 with a sodium borate and sodium phosphate buffer to a final concentration of 0.04 M. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1 and added to 2.12.1.fx at a molar ratio of 4.3:1, and reacted for 2 hrs at RT (Table 13).

TABLE-US-00015 TABLE 13 Buffer and pH alterations. CA (%) Buffer 0 1 2 3 4 Avg CA Phosphate, pH 7.5 1 21 53 21 3 2.02 Phosphate, pH 7.7 0 15 50 29 6 2.26 Phosphate, pH 8.0 1 14 52 29 4 2.21 Borate, pH 7.5 46 44 10 0 0 0.64 Borate, pH 7.7 22 51 23 4 0 1.09 Borate, pH 8.0 1 17 48 30 4 2.19

[0350] 2.12.1.fx was adjusted to 18 mg.Math.ml.sup.−1 at pH 7.7 with a phosphate buffer to a final concentration of 0.04 M, 0.06 M, or 0.08 M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT. The results are presented in Table 14.

TABLE-US-00016 TABLE 14 Concentration of phosphate. Concentration (mM) of CA (%) phosphate at pH 7.7 0 1 2 3 4 Avg CA 40 2 23 54 16 4 1.95 60 2 28 51 15 4 1.91 80 2 29 51 13 4 1.86

Effect of Buffer Constituents on Conjugation

[0351] Propylene glycol: 2.12.1.fx was adjusted to 18 mg.Math.ml.sup.−1 at pH 7.7 with a phosphate buffer to a final concentration of 0.06 M sodium phosphate. [PFP-PEG5-K11-SEQ:27] was reconstituted in a propylene glycol solution to 20 mg.Math.ml.sup.−1 (5% propylene glycol in the conjugation reaction). [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 4.3:1 and spiked with an additional 0 to 15% propylene glycol (final propylene glycol percentage of 5, 10, 15, and 20%) and allowed to react for 2 hrs at RT. The results are presented in Table 15.

TABLE-US-00017 TABLE 15 Percent of propylene glycol in 0.06M sodium phosphate. CA (%) Percent (%) Propylene Glycol 0 1 2 3 4 Avg CA 5 2 18 55 20 5 2.08 10 2 20 53 21 5 2.09 15 2 23 49 20 5 2.01 20 4 23 50 19 4 1.96

[0352] NaCl: 2.12.1.fx was adjusted to 2 mg.Math.ml.sup.−1 at pH 7.0 with a HEPES buffer to a final concentration of 0.02M in the presence and absence of 0.14M NaCl. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in DMSO to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 5:1 and allowed to react overnight at RT. The level of conjugation decreases in the presence of NaCl (Table 16).

TABLE-US-00018 TABLE 16 Concentration of sodium chloride in 0.02M HEPES. Concentration of sodium ABP Additions (%) chloride (mM) 0 1 2 3 4 Avg CA 0 2 21 41 28 4 2.03 0.14 9 34 42 14 1 1.64

[0353] HEPES: 2.12.1.fx was adjusted to 2 mg.Math.ml.sup.−1 at pH 7.0 with a HEPES buffer to a final concentration of 0.2 M and 0.02 M. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in 50% propylene glycol to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 5:1 and allowed to react 2 hrs at RT. The results are presented in Table 17. The level of conjugation is reduced at 0.2M HEPES buffer.

TABLE-US-00019 TABLE 17 HEPES concentration. ABP Additions (%) Concentration of HEPES (mM) 0 1 2 3 4 Avg CA 0.02 2 35 47 16 0 1.77 0.2 21 49 26 4 0 1.13

[0354] DMSO: 2.12.1.fx was adjusted to 15 mg.Math.ml.sup.−1 at pH 7.7 with sodium phosphate buffer to a final concentration of 0.06 M and DMSO was added to a final concentration of 30%. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1. [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx at a molar ratio of 4:1 and allowed to react for 2 hrs at RT. The results are presented in Table 18.

TABLE-US-00020 TABLE 18 DMSO in 0.06M sodium phosphate. ABP Additions (%) Percent of DMSO 0 1 2 3 4 Avg CA 0 3 28 49 14 6 1.92 30 8 28 32 22 10 1.98

Discussion of Conjugation Reaction Parameters

[0355] When the molar ratio of Effector Moiety (in this example, a peptide) to antibody is reduced below about 3.5:1, the level of conjugation is decreased, as seen in Table 9. Alternatively, Table 10 shows that increasing the molar ratio results in an increased level of conjugation. Increasing the number of peptides per antibody generally decreases the binding efficiency of the antibody (in this case 2.12.1 fx) to its antigen (in this case the IGF1R receptor), therefore the molar ratio of peptide to antibody was optimized to maximise both antibody-antigen, and peptide-cognate binding.

[0356] It was also found that varying the conjugation buffer can alter the conjugation pattern. Amine-containing excipients are less preferable in general as they can react with the PFP group. Buffers such as carbonate and borate can be used for conjugation but were avoided as their pKa (boric acid with a pKa ˜9 and carbonate with two pKa of ˜6 and ˜11) were far from the conjugation pH of 7.7 that was identified as optimal for MAC-1 and MAC-2 (Table 12). The level of conjugation is not only dependent on the chemical conditions of the reaction but also based on time. After 2 hrs, most of the PFP-activated peptide had reacted with the antibody or the PFP Z* has hydrolyzed (Table 6).

[0357] The PFP-activated peptide/linker reacted quickly with lysine side chain amino groups. Conjugation was performed at pH 6.5 to 8 in phosphate buffer to increase the nucleophilic tendency of the antibody surface lysines by decreasing their charge (the pKa of lysines on the surface proteins is about 9.1 to 11.2) as shown in Tables 4 and 5.

[0358] Optimal conditions for conjugation of MAC-1 and MAC-2 are described as follows: 2.12.1.fx antibody was adjusted to pH 7.7 with a phosphate buffer to a final concentration of 0.06M sodium phosphate. [PFP-PEG5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg.Math.ml.sup.−1 (final propylene glycol concentration in reaction is 10%). [PFP-PEG5-K.sup.11-SEQ:27] was added to 2.12.1.fx antibody at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT. The reaction was quenched with a succinate and glycine buffer, lowering the pH to approximately 6.0 and quenching any free peptide. In some aspects, the reaction may be concentrated and peptide-related species (such as peptides where the linker was hydrolyzed by reaction with water solvent) and other elements of the reaction mixture (such as PFP) may be removed by diafiltration, for example, using a 50 kDa membrane or size exclusion chromatography into a succinate, glycine, sodium chloride, and trehalose buffer, pH 5.5 at 30 mg/ml.

[0359] The conjugation conditions listed above were varied to determine the range of each process parameter. Parameter ranges were set based on variability that may occur during the conjugation and/or were expanded until greater than 10% change in species population was observed. Table 19 summarizes the parameters that result in similar conjugation profiles for MAC-2.

TABLE-US-00021 SEQ ATCC Date of Material ID NO: Accession No. Deposit hCLk-Km(3)-D77A 37 PTA-13394 Dec. 12, 2012 h38C2-[LC-D185A] 254 PTA-13395 Dec. 12, 2012

Example 6 Location of Conjugated Peptides on Antibody

[0360] The MAC-2 drug product molecule consists of a distribution of 1-4 [PEG5-K.sup.11-SEQ:27] molecules attached to the 2.12.1.fx antibody. This was determined by measuring the intact molecular weight (MW) of MAC-2 using electrospray, time-of-flight mass spectrometry detection following protein separation from salts and excipients through a size exclusion chromatography column. Mass spectrometry data that demonstrated the MW of the 2.12.1.fx and 3 lots of MAC-2 are shown in FIG. 2. FIG. 1A shows 2.12.1.fx before conjugation. This is a uniform molecule that displays a single MW. The MAC-2 lots display a distribution of conjugated peptides to 2.12.1.fx; between 1-4 conjugation additions (CA) are observed. The relative amount of each form is consistent between lots and the most common form in each lot has 2 peptides (SEQ ID NO:27) attached to each individual 2.12.1.fx antibody.

[0361] By reducing disulfide bonds in the 2.12.1.fx antibody, light and heavy chains are observed separately. Disulfide reduction is performed by treating the intact 2.12.1.fx antibody with 20 mM tris(2-carboxyethyl) phosphine (TCEP). The resulting mixture of heavy and light chains is analyzed for intact molecular weight as described above. The data shown in FIG. 2 provides evidence toward the location of the ABP on 2.12.1.fx. The majority of light chain (>65%) in the MAC-2 lots are conjugated. Most of the conjugated light chain contains 1CA. 2CA is also observed at a lower level. Almost all observed heavy chain (>90%) is unmodified, which suggests that very few of the conjugated peptides are located on the heavy chain.

[0362] Peptide mapping was used to determine the precise location of conjugation. The procedure was as follows: an aliquot of MAC-2 was denatured with 8M Guanidine-hydrochloride, disulfide bonds were reduced with TCEP, and the resulting cysteine sulphydryls were alkylated with Iodoacetamide. This treated protein sample was then digested with the protease chymotrypsin (1:125 protease:MAC ratio by weight). The resulting chymotryptic peptides were then detected individually by mass spectrometry after separation through a C8 liquid chromatography column. With this technique, MAC-2 was digested by chymotrypsin on the heavy and light chains into fragments at the locations noted in the sequence (with bullets) in FIG. 3. Liquid chromatography-mass spectrometry (LC-MS) detection of the MW of each peptide was then used to determine which Lysine residues are modified by a conjugated peptide. If a fragment was modified by attachment of conjugated peptide, its MW was shifted accordingly.

[0363] Fragments Y1, Y6, Y9, Y10, Y20, Y25, Y26, Y29, Y32, Y33, Y34, Y37, Y40 and Y43 of the heavy chain contain Lys residues. Of these, peptide conjugation was detected at Y6, Y10, Y25, Y33, and Y37. Fragments Y3, Y10, Y11, Y12, Y13, Y14, Y15, and Y16 of the light chain contain Lys residues. Of these, conjugation was detected at Y3, Y13, and Y15.

[0364] The light chain fragment referred to as Y15 (the 15th chymotryptic fragment on the light chain from the N-terminus) was found to be conjugated based on the data shown in FIG. 4. The MW of the modified Y15 fragment in MAC was clearly detected. In the un-conjugated 2.12.1.fx sample, there was no evidence of modified Y15 fragment. The unmodified Y15 fragment was observed in both MAC-2 and 2.12.1.fx. The magnitude of this fragment is higher in the 2.12.1.fx sample because this entire fragment is present in the un-modified form. As this fragment is conjugated in MAC-2, the observed level of un-modified Y15 decreases, which is seen in FIG. 4 as a peak with a smaller area.

[0365] The amount of conjugation of [SEQ:27-K.sup.11-PEG5] observed on light chain fragment Y15 in MAC-2 is estimated by measuring the decreased peak area of un-modified Y15. After normalizing the signal intensity such that unconjugated 2.12.1.fx showed 100%, 3 independent lots of MAC-2 showed 17%, 27% and 22% unconjugated Y15 fragments respectively.

[0366] The observed magnitude of Y15 in the MAC samples was normalized to the magnitude of Y15 in the 2.12.1.fx sample. Between 75-85% of the Y15 fragments are determined as modified in MAC-2. Considering that MAC-2 contains mostly 1-2 conjugation additions, this suggests that most of the conjugation in MAC-2 is located at one of the 2 K residues of light chain fragment Y15 (LC-K.sup.188 or LC-K.sup.190). The location of fragment Y15 in relation to the sequence of 2.12.1.fx is shown in FIG. 3.

[0367] Trypsin enzymatic digestion was used to discriminate between LC-K.sup.188 and LC-K.sup.190 (trypsin has specificity for the C-terminus of K and R). As trypsin does not digest conjugated K residues, the enzymatic digestion generates different peptide lengths, depending on which K residue is conjugated. Examination of LCMS data from MAC-2 that was digested with trypsin provides evidence that the peptide attaches specifically to LC-K.sup.188. No evidence of modified LC-K.sup.190 was observed.

[0368] MAC-2 was reduced with TCEP and denatured with guanidine hydrochloride as described above. The protein concentration was adjusted to 2 mg/ml and the pH to 7.8 with Tris digestion buffer. Purified trypsin was added at a 1:125 protease:MAC ratio by weight and incubated at 30° C. for 4 hrs. Samples were stored at −20° C. until analyzed by LCMS. Fragment samples were separated on a C18 reversed phase column using water/acetonitrile+0.1% TFA mobile phases. Detection of fragments was monitored both by UV 214 nm and ESI-TOF mass spectrometry. All data analysis was performed using MassLynx software.

[0369] The formation of fragments upon trypsin digestion of MAC-2 depends on the site of peptide conjugation. Lysines are the targeted residue for conjugation. Data shown in FIGS. 1-4 indicates that the predominant site of peptide binding is either LC-K.sup.188 or LC-K.sup.190. The scheme below shows the trypsin digestion reactions that would occur upon conjugation at either 2.12.1.fx-[LC-K.sup.188] or 2.12.1.fx-[LC-K.sup.190].

##STR00074##

[0370] The chemical structures of the two potential digestion fragments in question are as follows:

##STR00075##

[0371] FIG. 5 shows the selected ion LCMS chromatogram data for the trypsin peptide when LC-K.sup.188 is conjugated to the peptide. FIG. 6 shows the selected ion LCMS chromatogram data for the trypsin fragment when LC-K.sup.190 is modified with a conjugated peptide. These data suggest that only LC-K.sup.188 alone is conjugated; this situation results in a significant signal that is detected in MAC-2 but is absent in the 2.12.1.fx control experiment. The results from modification at LC-K.sup.190 do not provide any data that is unique compared to the negative control.

[0372] In contrast to what may be expected, the peptide/linker appears to preferentially decorate LC-K.sup.188 of the light chain of 2.12.1.fx. This has the surprising advantage that the Fc portion of the 2.12.1.fx antibody is unaffected. Tests show that the resulting PK of MAC-2 is approximately equal to the PK of unconjugated 2.12.1.fx. Promiscuous, non-specific conjugation to multiple sites on an antibody can result in a product with lower PK. The directional conjugation of the invention, exemplified by MAC-1 and MAC-2, provide the advantage of minimizing some of the possible deleterious effects that can be caused by promiscuous, non-specific conjugation, including lower PK. LC-K.sup.188 is the same residue as CLκ-K.sup.80 (i.e. K.sup.80 of SEQ ID NO:6), as the Light Chain (LC) comprises the variable region as well as the constant light kappa chain (CLκ).

[0373] To establish the reproducibility of the process, the experiment was repeated. MAC-2 was diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple conjugation sites are observed as distinct signals separated by the mass difference of a conjugated peptide/linker. Relative quantitation of multiple conjugation species is performed by measuring the signal magnitude. FIG. 7A shows a representative spectrum of MAC-2; the calculations used for quantitation are shown in Table 20. The average conjugation addition for the intact MAC-2 is calculated as 2.11 using the following formula: SUMPRODUCT (Number of Conjugation Additions (CA), Percent per CA). This example demonstrates conjugation of peptides occurring as a distribution between 0-4 peptide additions with the largest form being 2 peptide additions and the average number of peptide additions is 2.11. Replicate analysis by multiple individuals demonstrates that the profile of conjugation is consistent and reproducible.

TABLE-US-00022 TABLE 20 Weighted average of conjugation additions: 2.11. Conjugation additions Predicted mass Intensity Percent 0 149210 1615  1% 1 152350 20533 17% 2 155490 69395 56% 3 158630 27708 22% 4 161770 4818  4% 124069 100% 

[0374] The extent of peptide conjugation was examined separately on the light and heavy chains of 2.12.1.fx. MAC-2 was denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. FIGS. 7B and 7C show a representative spectrum of each chain; the calculation used for quantitation are shown in Table 21. The average conjugation additions (Avg CA) for the reduced heavy chain MAC-2 is calculated as 0.14 and the Avg CA for the reduced light chain MAC-2 is calculated at 0.86 using the following formula: SUMPRODUCT (Number of Conjugation Additions (CA), Percent per CA). These data demonstrate that the location of conjugation is higher on the light chain; the most abundant form on the light chain contains one peptide addition and the light chain contains an average of 0.86 peptide additions. Conjugation on the heavy chain is observed at a significantly lower level. Replicate analysis of this experiment by multiple individuals demonstrates that the profile of conjugation is consistent and reproducible.

TABLE-US-00023 TABLE 21 Peptide mapping characterization of MAC-2 identifying specific location of conjugation. Mass Conj. (Da) Additions Species Intensity Percent Avg CA 51020 0 HC 102093 86% 54165 1 HC + (1x) ABP-1 16204 14% Total HC 118297 100%  0.14 23584 0 LC 19752 21% 26729 1 LC + (1x) ABP-1 68757 72% 29874 2 LC + (2x) ABP-2 6561  7% Total LC 95070 100%  0.86

[0375] MAC-2 was reduced with dithiothreitol and cysteine residues were alkylated by carboxymethylation with iodoacetamide. Chymotrypsin was used for proteolytic digestion. Digested fragments in solution were analyzed using liquid chromatography mass spectrometry (LCMS). Individual fragments were separated over a C18 HPLC column and their accurate mass is measured in a Quadrupole Time-of-Flight (Q-ToF) mass spectrometer. The resulting fragment mass was used to identify unmodified fragments or fragments modified with a conjugated peptide. This experiment was interpreted by focusing on chymotryptic fragments that contain a lysine residue, as these were possible sites for peptide conjugation. Table 22 shows a listing of all such fragments. Blank entries are fragments that are not detected using this technique. Detected fragments that are observed with a peptide modifier are considered potential sites of conjugation.

The table entries for Table 17 are explained below:
Fragment number: Chymotrypsin fragment numbering from the N-terminus; joined fragments (i.e. Y1-2) indicate a missed cleavage site.
Start/End: Numbering of the fragment location from the N-terminus.
Peptide Mass (Da): Theoretical mass of the fragment listed in Daltons.
Retention Time (Control/Analyte): Time of chromatographic retention/elution in the LCMS fragment mapping experiment.
MS Signal Intensity (Control/Analyte): Magnitude of observed signal observed by MS.
Mass Error-ppm (Control/Analyte): Comparison of theoretical vs. observed mass of the fragment; values >10, and especially closer to zero (0) demonstrate better mass accuracy.
Modifiers: Potential covalent additions to the fragment; peptide-antibody binding fragment of Lys residue, CAM-carboxymethylation of Cysteine residue.
Asterisks indicate the modified (e.g. conjugated) version of the respective fragment. Pep indicates a conjugated peptide.

[0376] Directional conjugation of a peptide to the Y15 fragment is demonstrated by quantitating the conjugation level. The following analysis was performed on each of the peptide fragments that were observed having conjugation during the peptide mapping experiment of the 2.12.1.fx reference product. The ratio of observed signal intensity for the unmodified peptide in the non-conjugated control (2.12.1.fx antibody scaffold—no conjugation) compared to the conjugated reference product (MAC-2) is shown in Table 23. The unmodified signal is used because a direct comparison of the same peptide signal is possible in each sample. For example, an unconjugated peptide would be expected to have the same observed signal intensity in the control vs. product samples resulting in a ratio of one (1). Conjugation would result in a decrease in the observed amount of unmodified peptide in the product sample which would be indicated by a ratio greater than one (1). The data in Table 23 was further normalized to correct for sample and experimental variation between the control and product. Table 23 demonstrates that light chain peptide Y15 is conjugated at a significantly higher level than each of the other conjugated peptides. This suggests that conjugation occurs in a directional manner and is not randomly distributed across K residues.

TABLE-US-00024 TABLE 22 Peptide mapping characterization of MAC-2 heavy chain reference product. Peptide Fragment Mass Retention Time MS Signal Intensity Mass Error (ppm) Number Start End (Da) Control Analyte Control Analyte Control Analyte Modifiers Y1 1 27 2617.3533 Y1-2 1 29 2865.4695 Y5-6 34 47 1657.8398 Y6 37 47 1253.688 19.2 19.2 516640 583534 1.9 −1.1 Y6-7 37 50 1602.8518 22.1 22.1 26537 37988 −1.6 −2.2 Y6-7* 37 50 3295.7017 21.8 6316 −19.4 Pep(1) Y8-9 51 68 1931.9337 16.5 16.5 60894 85742 −2.2 0.4 Y9 61 68 878.461 11.3 11.3 376224 412997 0 −1 Y9-10 61 80 2241.1501 Y10 69 80 1380.6997 13.3 13.3 261813 299847 −1.1 0.7 Y10* 69 80 3073.5498 23.4 6350 −8.7 Pep (1) Y10-011 69 94 2972.4661 Y19-20 111 157 4748.2773 Y20 116 157 4160.0405 Y20-21 116 166 5202.5527 Y20-21* 116 166 5316.5957 34.1 6445 0.5 CAM(2) Y24-25 202 245 4702.2109 Y25 207 245 4151.9722 Y25* 207 245 4437.0796 20.9 20.9 1495322 1800079 1.1 −3.1 CAM(5) Y25* 207 245 6129.9297 24.4 6652 −4.5 CAM(5) Pep(1) Y25-26 207 279 7985.9092 Y26 246 279 3851.9478 Y26-27 246 281 4152.0698 Y28-29 282 300 2245.1128 Y29 283 300 2082.0493 14.6 14.6 20665 16662 −0.6 −3.8 Y29-30 283 304 2531.2405 Y31-32 305 323 2241.1907 Y32 318 323 722.3599 7.9 7.9 93966 96639 0.1 2.6 Y32 318 323 722.3599 17.7 18.4 37943 12802 11.4 30.6 Y32 318 323 722.3599 18.4 11761 23.8 Y32-33 318 353 4028.188 Y33 324 353 3323.8386 20 5422 3.1 Y33* 324 353 3380.8601 19.7 19.7 2196329 2497507 −2.5 −3.1 CAM(1) Y33* 324 353 5073.71 24 5973 1.3 CAM(1) Pep(1) Y33-34 324 376 5883.1577 Y34 354 376 2577.3293 Y34-35 354 385 3637.8159 Y34-35* 354 385 3694.8374 33 32.9 10095 20682 1.9 −2.4 CAM(1) Y36-37 386 408 2527.0808 Y37 396 408 1394.6388 19.6 19.6 62942 71902 −0.9 −0.4 Y37-38 396 409 1541.7072 25.1 25.1 827336 878570 0 −1.9 Y37-38* 396 409 3234.5571 29.7 7749 −5.3 Pep(1) Y39-40 410 421 1494.8195 Y40 412 421 1218.672 15.8 15.8 77917 88243 −0.3 −1.6 Y40-41 412 427 1891.9905 20.3 20.3 107513 149676 0.2 −2 Y42-43 428 450 2525.1792 Y43 441 450 1016.5502 Y2-3 36 49 1688.9725 16.2 16.2 145374 170451 −1.7 −2.6 Y2-3* 36 49 3381.8225 24.2 7192 −9.2 Pep(1) Y3 37 49 1525.9093 15.5 15.5 331068 393638 −2.7 −2.9 Y3* 37 49 3218.7593 24 28193 −9 Pep(1) Y3-4 37 62 2882.6355 Y9-10 88 116 3244.729 Y10 99 116 1871.0992 Y10-11 99 139 4331.335 Y11 117 139 2478.2463 22.8 47035 −5.9 Y11-12 117 148 3635.8445 Y12 140 148 1175.6088 Y12-13 140 173 3886.8245 Y13 149 173 2729.2263 13.1 13.1 1140556 1218022 −1.1 0.1 Y13* 149 173 4422.0762 21.4 8424 −6.5 Pep(1) Y13-14 149 186 4095.9243 Y14 174 186 1384.7086 Y14-15 174 192 2169.1318 Y15 187 192 802.4337 7.5 7.5 275639 62720 −1.9 −0.2 Y15* 187 192 2495.2837 20.9 936267 −9.8 Pep(1) Y15-16 187 209 2574.29 Y16 193 209 1789.8668 18.7 5400 4.4 Y16* 193 209 1846.8883 18.1 18.1 169490 235914 −1.7 −2.5 CAM(1) Y16-17 193 214 2349.0842 17.8 9211 0.1

TABLE-US-00025 TABLE 23 Directional conjugation of peptide to Y15 fragment on the light chain. Unmodified Intensity Ratio: Control/Analyte- Fragment normalized Light Y3  1.000 Light Y13 1.112 Light Y15 5.218 Heavy Y6  0.831 Heavy Y10 1.038 Heavy Y25 0.988 Heavy Y33 1.045 Heavy Y37 1.120

Example 7 Demonstration of Potency of MAC Products

[0377] Full details of in vitro and in vivo assays of MAC-1 and MAC-2 are provided in the Examples of PCT/US2011/053092 (WO2012/007896). Ang2-h38C2-IgG1 was used as a control in certain examples. The generation and structure of the Ang2-h38C2 is fully described as compound 43 in WO2008056346, whose contents is incorporated herein, with particular reference to aspects referring to the generation of compound 43. Briefly, the structure is as follows:

##STR00076##

wherein the linker is covalently attached to the ε-amino group of HC-K.sup.99 (K.sup.93 according to Kabat numbering) of the combining site of the antibody and the antibody is h38C2-IgG1 (SEQ ID NO:64 and 65) (SEQ ID NO:189 and SEQ ID NO:190 of WO2008/056346).

[0378] In summary, MAC-1 and MAC-2 were able to bind Ang2 and prevent its binding to Tie2 as shown in an Ang2 competition assay, and both MAC-1 and MAC-2 have similar activity as the parental anti-IGF1R antibody (2.12.1.fx) for competing with IGF1 for IGF1R binding (Table 24). Surprisingly, in comparison with Ang2-h38c2, MAC-1 and MAC-2 both showed an increase in ability to competitively bind Ang2. Therefore, conjugation of limited Ang2 peptides does not appear to change the innate binding and inhibition of the antibody, and may in some cases improve the Effector Moiety activity.

[0379] The MACs were tested for the ability to downregulate IGF1R levels on a human colon carcinoma cell line Colo205. Cells were treated for 3 hrs in culture with titration of MAC compounds. Cells were collected and IGF1R surface expression determined by flow cytometry. The percentage of IGF1R downregulated as compared to negative control hIgG2 was determined (Table 23).

TABLE-US-00026 TABLE 24 Ability of MAC-1 and MAC-2 to bind & modulate IGF1R and Ang2. IGF1R % IGF1R Ang2 IGF1R phosphorylation down- IC.sub.50 (nM) IC.sub.50 (nM) IC.sub.50 (nM) regulated MAC-1 0.092 ± 0.049 5.1 ± 1.1 150.7 ± 59.6 43 ± 5 MAC-2 0.057 ± 0.022 6.1 ± 1.1  91.4 ± 40.2 50 ± 5 2.12.1.fx nd 3.8 ± 0.8  48.7 ± 14.0 48 ± 3 antibody Ang2-h38c2- 0.582 ± 0.242 nd nd nd IgG1

[0380] It was also demonstrated that conjugating 2 peptides per antibody was ideal in terms of effecting IGF1R autophosphorylation and downregulation and that conjugating more or less than 2 peptides per antibody lessens the ability of the MAC to effect these functions.

[0381] To assess the effect of the number of peptides per antibody on the ability of 2.12.1.fx to modulate IGF1R activity, 2 samples of MAC-1 were prepared where the reaction conditions were set to provide either reduced conjugation (MAC-1 low) or increased conjugation (MAC-1 high) (Table 25). The samples were analysed for the ability to downregulate and phosphorylate IGF1R (Table 25). There is a significant difference in the ability of the MAC-1 high as compared with MAC-1 low to effectively modulate the IGF1R pathway. Conjugation of greater than about 2 peptides per antibody limits the functional activity of the MAC to both inhibit IGF1R autophosphorylation and induce IGF1R downregulation, compared to conjugation of about 2 or less peptides per antibody. Therefore, in order to efficiently modulate 2 different biological pathways in one bifunctional entity, conjugation of about 2 peptides per antibody may be ideal (depending on peptide's and target's pharmacokinetic profile).

TABLE-US-00027 TABLE 25 Analysis of MAC-1-High and MAC-1 Low. Phosphory- Ang2 % IGF1R lation IC.sub.50 down- IGF1R CA (%) Avg (nM) regulated IC.sub.50 (nM) 0 1 2 3 4 5 CA MAC-1 0.103 32 ± 1 12.8 14 42 32 12 0 0 1.42 Low MAC-1 0.035  9 ± 2 >300 0 4 19 41 32 5 3.18 High 2.12.1.fx nd 36 ± 3 3.5 Ang2- 0.252 nd nd h38c2- IgG1

Example 8 In Vivo Pharmacokinetics

[0382] PK studies were conducted using male Swiss Webster mice and 2 male Cynomolgus monkeys (Macaca fascicularis). Full details of PK studies are provided in the Examples of PCT/US2011/053092. In mouse, MAC-1 and MAC-2 demonstrated similar residence time as the parental anti-IGF1R antibody with 13 phase half-lives of 383-397 hrs. The MAC-1 and MAC-2 Ang2 binding capability demonstrated similar residence time as Ang2-h38c2 with T½ of 105-120 hrs in mouse in single dose IV studies. In cynomolgus monkey, MAC-2 demonstrated a slightly shorter residence time as the parental anti-IGF1R antibody with T½ of 100.4 hrs. The MAC-2 Ang2 binding capability demonstrated similar residence time as Ang2-h38c2 with T1/2 of 97.8 hrs.

TABLE-US-00028 TABLE 26 Single-dose PK of IV administered MACs at 10 mg/mkg in mouse and cynomolgus monkey. α-IGF1R antibody dosed at 10 mg/kg in mouse, and 5 mg/kg in monkey. Mouse β-T ½ (hr) Monkey T ½ (hr) Compound (mg .Math. Kg.sup.−1) Ang2 IGF1R Ang2 IGF1R Ang2-h38c2, (10) 95.2 — 95.3 — α-IGF1R antibody, (10), (5) — 390 — 146.4 MAC-1, (10) 105 383 NT NT MAC-2, (10) 120 397 97.8 100.4 NT: not tested.

Example 9 In Vivo Pharmacology

[0383] The anti-tumour activity of MAC-2 was evaluated in the Colo205 (human colon adenocarcinoma) or MDA-MB-435 (melanoma) xenograft model. Full details of tumour studies are provided in the Examples of PCT/US2011/053092 (WO2012/007896). Weekly administration of Ang2-h38c2 or anti-IGF1R antibody (2.12.1.fx) inhibited Colo205 tumour growth. Combination of weekly administered Ang2-h38c2 and anti-IGF1R antibody showed an additive benefit on inhibiting Colo205 tumour growth. Weekly administration of MAC-2 alone showed similar benefit as the combination. In a separate study, MAC-2 dose-dependently inhibited Colo205 tumour growth and final tumour weights.

[0384] At day 28, tumour microvessel density after compound treated was significantly reduced (˜42%) by MAC-2 (10 mg/kg, once weekly) in comparison with the Vehicle-treated group confirming the anti-angiogenic activity of the MAC-2 treatment.

[0385] To investigate whether MAC-2 targets both Ang2 and IGF1R in vivo, the effects of MAC-2 on Ang2 and IGF1R expression levels were assessed in 2 independent Colo205 xenograft tumors treated with Vehicle, Ang2-h38c2, IGF1R antibody (2.12.1.fx) or MAC-2 (dose response ranging from 0.3 mg/kg to 10 mg/kg). The results showed that Ang2 and IGF1R immunoreactivity was significantly reduced by MAC-2 treatment in a dose-dependent manner (1, 3 and 10 mg/kg) in comparison with the Vehicle-treated group. The effect of MAC-2 on IGF1R levels was similar to that observed for an IGF1R antagonizing antibody. In addition, the levels of phosphorylated IGF1R were reduced in tumours from MAC-2 treated animals. These data demonstrate that MAC-2 treatment affects both Ang2 and IGF1R pathways in Colo205 xenograft model. MAC-2 treatment did not affect body weight gain and mice appeared to be in good health throughout the studies. The anti-tumor efficacy of MAC-2 was also evaluated in an MDA-MB-435 melanoma xenograft model. Weekly administration of MAC-2 (3 and 20 mg/kg IP) resulted in a significant 40% reduction (day 67) in tumor growth in the MDA-MB-435 model. Thus, MAC-2 demonstrates significant anti-tumor efficacy in 2 different human xenograft tumor models.

Example 10 Peptide Conjugation Profile of Various Antibodies

[0386] The conjugation profiles of several different antibodies with peptides were analyzed, using SEQ ID NO:27 and PEG5 as an exemplary peptide and linker respectively. All antibodies tested were human or fully humanized IgG antibodies with well defined and characterized antigen interactions. hAbATest comprises a CLλ (hIL22: SEQ ID NOs:136 and 137), whereas 2.12.1.fx, mAbKTest1 (an IgG2 anti-Alk1 antibody, as disclosed in U.S. Pat. No. 7,537,762, incorporated herein by reference), h38C2-IgG1 (SEQ ID NO:64 and 65) and h38C2-IgG2 (SEQ ID NO:64 and 66) each comprise CLκ. Each of the antibodies were buffer exchanged into 20 mM HEPES, pH 7.0 and concentrated to 5-20 mg/mL. [PFP-PEG5-K.sup.11-SEQ:27] was resuspended with 50% propylene glycol and mixed with the relevant antibody at a 4.3:1 molar ratio and allowed to react for at least 2 hrs at RT. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple peptide conjugation sites are observed as distinct signals separated by the mass difference of a bound peptide. Relative quantitation of multiple peptide conjugation species is performed by measuring the signal magnitude. Table 22 shows the peptide conjugation profile of various antibodies

[0387] For antibodies containing a CLκ, peptide conjugation occurs at a distribution between 0-4 peptide additions with the largest form being 2 to 3 peptide additions. In contrast, for the CLλ comprising antibody, hAbATest, conjugation of the peptide occurs at a distribution between 0-4 peptides additions with the largest form being 1 to 2 peptide additions.

[0388] The extent of peptide conjugation was examined separately on the light and heavy chains. Each sample was denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. The peptide conjugation profile on the light and heavy chain of various antibodies is shown in Table 27. On 2.12.1.fx and hAbKTest1, the data demonstrate that the location of conjugation is higher on the light chain; the most abundant form on the light chain contains 1 peptide addition. Conjugation on the heavy chain is observed at a significantly lower level. On h38C2-IgG1 and h38C2-IgG2, comparable levels of conjugation are observed on the light and heavy chain, with a slight conjugation preference on the light chain. On a CLλ containing antibody (hAbATest; comprising SEQ ID NOs:136 and 137)), the majority of the conjugation occurs on the heavy chain with a low level of conjugation observed on the light chain.

TABLE-US-00029 TABLE 27 Conjugation profile of various antibodies. Light Chain Heavy chain CA (%) Avg % CA % CA Antibody 0 1 2 3 4 CA 0 1 2 0 1 2 2.12.1.fx 1 15 53 26 5 2.2 20 70 9 84 16 0 hAbλTest 10 37 37 11 6 1.66 95 5 0 74 22 4 hAbκTest1 7 10 35 27 14 2.55 11 74 14 87 13 0 h38C2 IgG1 1 3 28 55 13 2.75 49 46 4 70 30 0 h38C2 IgG2 4 6 31 44 2.6 61 35 4 73 27 0
Each of the antibodies 2.12.1.fx, hAbATest and hAbKTest1 was assessed after the conjugation process to determine the effect of the conjugation additions on the ability of the antibody scaffold to retain its receptor binding (compared to native mAb) (Table 28). The results show that the directional conjugation of peptides to the test antibodies did not appear to alter the antibody binding.

TABLE-US-00030 TABLE 28 Antibody binding to respective native antigen before and after conjugation. Antigen binding Antibody Native (IC.sub.50, nM) After conjugation (IC.sub.50, nM) 2.12.1.fx 3.2 5.7 hAbλTest 0.4 1.7 hAbκTest1 59 53

Example 11 Peptide Conjugation Profile of an IgG2-K Antibody

[0389] The conjugation profile of an IgG2 K antibody (hABκTest2) with a 39-mer peptide was analyzed (SEQ ID NO:164). The antibody was concentrated to 8 mg/mL and buffered exchanged into 40 mM HEPES pH 8.0. The peptide was resuspended with 100% DMSO and mixed with the antibody at a 5.0:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple peptide conjugation sites are observed as distinct signals separated by the mass difference of a peptide. Relative quantitation of multiple peptide conjugation species is performed by measuring the signal magnitude. Table 29 shows the peptide conjugation profile of hAbKTest2 with the 39-mer peptide. The conjugation of peptide occurs at a distribution between 0-4 CA with an average of 2.03 CA, and is consistent with directional conjugation on the CLκ-K.sup.80.

TABLE-US-00031 TABLE 29 Conjugation profile of 39-mer peptide and hAbκTest2. Antibody % CA Avg scaffold Binding Peptide 0 1 2 3 4 CA hAbκTest2 39-mer peptide 1 22 53 18 5 2.03

[0390] In a separate experiment, the 39-mer peptide was conjugated to h38C2-IgG2 with MAL-PEG2-PFP as described above, at different molar concentrations. In addition, binding of the cognate receptor for the 39-mer peptide was assayed. The results (Table 30) shown are consistent with directional conjugation at CLκ-K.sup.80. Moreover, increasing the average number of peptides per antibody did not substantially increase overall binding to the target. This demonstrates that in certain scenarios, increasing the conjugation per antibody may not increase target binding, demonstrating one of the advantages of the invention; control of the number of peptides conjugating per antibody can help achieve the maximum target binding per unit peptide.

TABLE-US-00032 TABLE 30 Conjugation profile of 39-mer peptide and H38C2-IgG2. 39-mer peptide: CA (%) Avg # Peptide target: h38C2-IgG2 mole ratio 0 1 2 3 Conjugates EC50 (nM)   2:1 57 32 10 0 0.52 0.99 2.5:1 19 56 25 0 1.06 1.06   4:1 20 25 35 20 1.55 1.01   5:1 0 16 45 40 2.26 0.82

Example 12 Conjugation of Biotin to 2.12.1.fx Fab

[0391] The conjugation profile of the Fab region of 2.12.1.fx (SEQ ID NOs:4 and 64) with PFP-Biotin was analyzed. The antibody Fab was concentrated to 20 mg/mL and buffered exchanged into 20 mM sodium acetate+200 mM trehalose, pH 5.5 and spiked with 60 mM sodium phosphate pH 7.7. PFP-Biotin was resuspended with 100% DMSO and mixed with the antibody at successive molar ratios and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms. This technique measures the molecular weight of each protein form; multiple conjugation sites are observed as distinct signals separated by the mass difference of a conjugated peptide. Relative quantitation of multiple conjugation species is performed by measuring the signal magnitude. Table 31 shows the conjugation profile of 2.12.1.fx Fab with PFP-Biotin at molar ratios. The conjugation of occurs at a distribution between 0-2 additions as the molar ratio increases. The lower number of molecules per antibody was consistent with earlier results, based on the molar ratio used. This demonstrates the flexibility of the process to control the amount of conjugation by altering reaction parameters.

TABLE-US-00033 TABLE 31 Conjugation profile of Biotin to 2.12.1.fx Fab. [00077] Biotin-PFP Binding Peptide:Antibody % CA Avg Peptide Molar Ratio 0 1 2 3 CA Biotin-PFP 1:1 54 46 — — 0.46 Biotin-PFP 1.5:1   42 51  7 — 0.65 Biotin-PFP 2:1 34 55 10 — 0.76 Biotin-PFP 3:1 28 55 17 — 0.88 Biotin-PFP 4:1 21 46 26 8 1.21

Example 12 Conjugation of Biotin to h38C2-IgG1

[0392] The antibody h38C2-IgG1 was adjusted to 20 mg/mL with HEPES buffer pH 7.5 to a final concentration of 0.02 M. Biotin-PFP was reconstituted in water to 10 mg/mL and added to h38C2-IgG1 at a molar ratio of 5:1 and allowed to react at room temperature for 2 hrs. The unreacted PFP-Biotin was removed by size exclusion chromatography and buffer exchanged into a histidine, glycine, and sucrose buffer pH 6.5. The samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. Table 32 shows the conjugation profile of h38C2-IgG1 with Biotin-PFP. Conjugation of h38C2-IgG1 occurs at a distribution between 0-3 CA with an average of 1.1 conjugations. Increased conjugation would be possible following optimization of the reaction conditions. The reactivity of VH-K.sup.99 (K.sup.93 according to Kabat numbering) on h38C2-IgG1 was confirmed to be >95% when reacted with the catalytic antibody test compound CATC-1, and analyzed via reversed phase chromatography.

TABLE-US-00034 TABLE 32 Conjugation of Biotin and h38C2-IgG1. [00078] CATC-1 Antibody 0 1 2 3 Avg CA h38C2-IgG1 16 61 20 3 1.1

Example 13 Conjugation profile of 2.12.1.fx and CLκ-K.SUP.80., CLκ-K.SUP.82 .mutants

[0393] Based on peptide mapping, there are 2 Lys in Y15 fragment. In order to distinguish the active conjugation site, CLκ-K.sup.80 and CLκ-K.sup.82 were mutated to R respectively or in combination. Mutants of the test antibody, 2.12.1.fx, were generated following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The mutated mAbs were transiently expressed in HEK 293 cells, and purified using Protein A affinity column. The purified mAbs were characterized using MS. SEQ ID NOs:12, 13 and 14 show the mutant CLκ sequences.

[0394] The antibody was buffer exchanged to 0.02M HEPES buffer pH 7.5 or 6.5 at 2 mg/mL. If the pH was 6.5, the antibody was then spiked with 60 mM sodium phosphate pH 7.7. [PFP-PEG5-K.sup.11-SEQ:27] was resuspended with 50% propylene glycol and mixed with the protein at a 4.3:1 molar ratio and allowed to react overnight at RT. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography—mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple conjugation sites are observed as distinct signals separated by the mass difference of a conjugated protein. Relative quantitation of multiple protein conjugation species is performed by measuring the signal magnitude. Table 33 shows the conjugation profile of unmodified 2.12.1.fx, 2.12.1.fx-[CLκ-K.sup.80R] (CLκ: SEQ ID NO:12), 2.12.1.fx-[CLκ-K.sup.82R] (CLκ: SEQ ID NO:13), and 2.12.1.fx-[CLκ-K.sup.80R-K.sup.82R] (CLκ: SEQ ID NO:14). CLκ-K.sup.80R mutant showed reduced conjugation. CLκ-K.sup.82R had similar conjugation as the unconjugated 2.12.1.fx. The conjugation of MAC-2 was lower than observed in other assays due using a combination HEPES/phosphate buffer.

TABLE-US-00035 TABLE 33 Conjugation profile of 2.12.1.fx, K.sup.80 and R.sup.82 mutants. CA (%) LC SEQ ID NO: Mutants 0 1 2 3 4 Avg CA  6 MAC-2 14 49 31 5 1 1.29 12 K80R 82 14 4 0 0 0.22 13 K82R 11 46 36 6 0 1.37 14 K80R/K82R 51 37 9 3 0 0.63

Example 14 Elucidation of Directional Conjugation Mechanism on K.SUP.80

[0395] CLκ-H.sup.81 side chain is very close to the ε-amino group of CLκ-K.sup.80. Since His is often involved in proton transfer reactions, CLκ-H.sup.81 is very likely required for CLκ-K.sup.80 conjugation. In order to study the role of CLκ-H.sup.81 in CLκ-K.sup.80 site specific conjugation, the imidazole ring was eliminated by a CLκ-H.sup.81A mutation. CLκ-D.sup.43A and CLκ-D.sup.43A/H.sup.81A mutants were made to study the role of CLκ-D.sup.43 in site specific conjugation and the combined effect of CLκ-D.sup.43 and CLκ-H.sup.81.

[0396] Mutants were generated following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The mutated mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS. The following 2.12.1.fx I CLκ mutants were generated: CLκ-D.sup.43A (SEQ ID NO:15), CLκ-K.sup.80A (SEQ ID NO:16), CLκ-H.sup.81A (SEQ ID NO:17), CLκ-K.sup.82A (SEQ ID NO:18) and CLκ-D.sup.43A/H.sup.81A (SEQ ID NO:19).

[0397] Each of the antibodies was buffer exchanged to 20 mM sodium acetate, 200m trehalose pH 5.5 at 20 mg/ml. The proteins were then spiked with 60 mM sodium phosphate pH 7.7. [PFP-PEG5-K.sup.11-SEQ:27] was resuspended with 50% propylene glycol and mixed with the antibody at a 4.3:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple conjugation sites are observed as distinct signals separated by the mass difference of a conjugated peptide. Relative quantitation of multiple conjugation species is performed by measuring the signal magnitude.

[0398] Table 34 shows the conjugation profile of 2.12.1.fx, 2.12.1.fx-[CLκ-D.sup.43A], 2.12.1.fx-[CLκ-K.sup.80A], 2.12.1.fx-[CLκ-H.sup.81A], 2.12.1.fx-[CLκ-K.sup.82A], and 2.12.1.fx-[CLκ-D43A/H.sup.81A] mutants. All the mutants showed reduced average conjugation level compared to the unmodified 2.12.1.fx antibody, except for CLκ-K.sup.80A, which maintained directional conjugation.

[0399] The extent of conjugation was examined separately on the light and heavy chains. Each sample was denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. The conjugation profile on the light and heavy chain of 2.12.1.fx and mutants are shown in Table 34. All the mutants listed in the table showed reduced conjugation level on light chain compared to the unmodified 2.12.1.fx except CLκ-K.sup.80A. The heavy chain conjugation level of the mutants was at the similar level as the unmodified 2.12.1.fx. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 34.

TABLE-US-00036 TABLE 34 Conjugation profile of MAC-2 and K.sup.80A, D.sup.43 and H.sup.81 mutants. LC SEQ CA (%) Avg LC CA% LC Avg HC CA % HC Avg 1 LC ID NO: Mutants 0 1 2 3 4 CA 0 1 2 CA-LC 0 1 2 CA-HC WT %  6 MAC-2 1 15 53 26 5 2.2 23 69 8 0.85 86 14 0 0.14 15 D.sup.43A 17 38 31 14 0 1.41 68 30 1 0.33 79 21 0 0.21 43 16 K.sup.80A 56 31 10 4 0 0.61 89 11 0 0.11 91 9 0 0.09 16 17 H.sup.81A 34 44 17 6 0 0.95 89 11 0 0.11 78 22 0 0.22 16 18 K.sup.82A 9 7 31 37 16 2.42 8 77 15 1.06 83 17 0 0.17 111 19 D.sup.43A/ 34 39 18 9 0 1.02 83 17 0 0.17 87 13 0 0.13 25 H.sup.81A

Example 15 Lambda/Kappa Substitution

[0400] The CLλ in hAbλTest1 (SEQ ID NOs:136 and 137) was substituted with CLκ.sub.□ to determine whether this increased the level, directionality and/or control of CL-specific conjugation. The CLλ/CLκ.sub.□ domain substitution hybrid constructs were generated using overlap PCR. The VLλ and CLκ were PCR amplified using hAbλTest and a κ mAb light chain as templates separately. These 2 PCR products were mixed as templates; hAbλTest1 forward primer and LCLκ.sub.□ reverse primer were used in overlap PCR reaction to amplify the full length hAbλTestVL/CLκ.sub.□ DNA. The hybrid antibody constructs were transiently expressed in HEK 293 cells, and purified using Protein A affinity column. The purified antibodies were characterized using MS. The hAbλTest CLκ.sub.□ hybrid bound to its cognate ligand similarly to the native mAb (hAbλTest) (Table 35). SEQ ID NOs:59, 60 and 61 are the light chain constant regions from hAbλTest, hAbλTest-λκ (with λJ), and hAbλTest-λκJ (with κJ).

TABLE-US-00037 TABLE 35 Antibody: Antigen binding of lambda/Kappa substitution. LC SEQ Inhibition of IL22 binding to hAbλTest1 Mutants ID NO: antigen (IC.sub.50, nM) hAbλTest (CONTROL) 59 0.4 hAλTest-λκ 60 0.3 hAbλTest-λκJ 61 0.3

Example 16 hAbλTest1 Mutants: Motif Modification

[0401] To establish whether the short motif “KH” was sufficient for MAC formation in the corresponding region of the CLλ, a mutant with simple sequence switch of residues CLλ.sup.81/82 in hAbλTest to place a histidine beside K.sup.80 was made, hence “K.sup.80S.sup.81H.sup.82” became “K.sup.80H.sup.81S.sup.82”. Mutants were generated following protocols described in QuickChange site-directed mutagenesis kit (Stratagene). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The mutated antibody constructs were transiently expressed in HEK 293 cells, and purified using Protein A affinity column. The purified antibodies were characterized using MS. The hAbλTest-[CLλ-S.sup.81H/H.sup.82S] (CL: SEQ ID NO:62) mutant bound to its ligand as well as the parent hAbλTest antibody did (Table 36).

TABLE-US-00038 TABLE 36 hAbλTest-S.sup.81H/H.sup.82S. hAbλTest1 Mutants LC SEQ ID NO: Ligand binding (IC.sub.50, nM) hAbλTest (CONTROL) 59 0.3 hAbλTest-S.sup.81H/H.sup.82S 62 0.4

Example 17 Conjugation Profile of hAbλTest1 Mutants

[0402] Each antibody (hAbλTest, hAbλTest-AK, hAbλTest-λκJ and hAbλTest-[CLλ-S.sup.81H/H.sup.82S]) was buffer exchanged to 20 mM sodium acetate, 200m trehalose pH 5.5 at 20 mg/ml. The proteins were then spiked with 60 mM sodium phosphate pH 7.7. [PFP-PEG5-K.sup.11-SEQ:27] was resuspended with 50% propylene glycol and mixed with the antibody at a 4.3:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple peptide conjugation sites are observed as distinct signals separated by the mass difference of a peptide. Relative quantitation of multiple peptide conjugation species is performed by measuring the signal magnitude. Table 37 shows the overall level of conjugation has been increased in the 2 LC-switched hybrids (λκ and λκJ—the former includes a λ J fragment, the latter includes a κ J fragment). The conjugation level increases over the hAbλTest control's average CA, going from 1.66 to 2.19 (λκ) and 2.53 (λκJ) respectively. The mutant had little effect compared to the native sequence, suggesting that “KH” motif alone is not sufficient for MAC formation.

[0403] The extent of peptide conjugation was examined separately on the light and heavy chains (Table 37). Each sample was denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. In the reduced analyses, the LC of native hAbλTest has only 5% 1CA but this jumps dramatically to 58% 1CA for hAbλTest-λκ and 63% 1CA for hAbλTest-λκJ. The LC switch had little effect on the level of HC conjugation, which remained fairly constant (except for λκJ, where HC conjugation increased moderately). Again, the mutant had little effect compared to the native sequence, suggesting that “KH” motif alone is not sufficient for MAC formation. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 37.

TABLE-US-00039 TABLE 37 Conjugation profile of hAbATest mutants. LC SEQ Avg Avg hAbλTest ID CA (%) Avg LC CA% CA- HC CA % CA- 1LC Mutants NO: 0 1 2 3 4 CA 0 1 2 LC 0 1 2 HC WT % hAbλTest 59 10 37 37 11 6 1.66 95 5 0 0.05 74 22 4 0.3 hAbλTest- 60 3 18 43 29 7 2.19 42 58 0 0.58 78 22 0 0.22 1160 λκ hAbλTest- 61 2 11 34 36 17 2.53 33 63 4 0.71 64 36 0 0.36 1260 λκJ hAbλTest- 62 7 34 37 16 6 1.79 82 18 0 0.18 79 21 0 0.21 360 S.sup.81H/H.sup.82S

[0404] The receptor binding attributes of these conjugated forms was also assessed to determine the effect of conjugation with [PFP-PEG.sub.5-K.sup.11-SEQ:27] on the ability of the conjugated antibodies to still bind to their ligand (Table 38).

TABLE-US-00040 TABLE 38 Antibody: Antigen binding of lambda at antibodies. SEQ ID NO: 27 conjugated LC SEQ Inhibition of IL22 binding to hAbλTest1 Mutants ID NO: antigen (IC.sub.50, nM) hAbλTest 59 1.7 hAbλTest-λκ 60 1.5 hAbλTest-λκJ 61 1.6 hAbλTest1-S.sup.81H/H.sup.82S 62 1.6

Example 18 MAC Generation Using Different Leaving Groups

[0405] To investigate if the degree of activation and/or structure of the active ester leaving group was important in defining the directional conjugation effect, a series of alternatively activated ester analogs of [PFP-PEG.sub.2-MAL-K.sub.SH.sup.11-SEQ:27] were synthesized. The distribution of the conjugate product was examined by MS of the intact conjugates, and the degree of peptide addition to both the light and heavy chains were also determined by MS following reduction of the intact conjugate and separation of the light and heavy chains.

[0406] The structure and designations of the alternatively activated esters are shown below. The alternatively activated peptides were synthesized using the same strategies and methods shown above. Briefly, each activated group was incorporated into a MAL-PEG.sub.2-Z* linker, where Z* represented the new leaving group replacing PFP. To synthesize the above compounds, a sample (30-40 mg) of the purified ABP-thiol peptide (i.e. ABP with K.sub.SH as linking residue) was dissolved in anhydrous DMF (2 ml). MAL-PEG.sub.2-Z* (20 mg) was added along with N-methylmorpholine (5 mL). The reaction was stirred and monitored at RT by HPLC to follow the time-course of product formation. The complete conversion of starting peptide to activate-ester linked ABP product was observed within 2-6 hrs. The solution was filtered and the product peak directly isolated by semi-preparative HPLC. The products were isolated in yields ranging from approximately 30-50%, after lyophilization.

[0407] The conjugation reactions were carried out under the standard conditions. Briefly, the 2.12.1.fx antibody solution was prepared by diluting the 2.12.1.fx solution with sodium phosphate, pH 7.7 to a final concentration of 0.06M. Separately, the peptide solution was prepared by dissolving the peptide to 20 mg/ml in propylene glycol, then diluting this solution to 10 mg/ml with water. For the conjugation reaction, the peptide and antibody solutions were mixed at a 4:1 molar ratio for the prescribed period. For the time-course studies, samples of the conjugation solution were quenched at various time points by mixing a sample of the conjugation reaction with a solution of 40 mM succinic acid, 200 mM glycine, pH 4.0 (1:1, v/v). Time-course of the conjugation reactions were followed by HPLC. SEQ ID NO:27 was used as an exemplary peptide.

TABLE-US-00041 TABLE 39 Reactive esters—intact conjugation at 24 hrs. [00079] Z*—PEG.sub.2—MAL—K.sub.SH.sup.11—SEQ: 27 Z8 Z9 Z2 Z3 Z4 Z5 Z6 Z7 NH5 2HI Z11 Z12 Z1 2,3,4 2,3,6 2,3,6 2,6 2,4 5,7 NB2, 3 1,3 Z10 2,6- 1 CA PFP TFP TFP TCP DCP DCN DCQ DCI D 4NP DFP NAP 0 3 32 17 100 81 38 73 34 20 41 50 100 1 34 45 43 0 19 45 25 40 36 42 39 0 2 51 20 30 0 0 16 2 18 31 15 11 0 3 12 3 11 0 0 2 0 5 12 3 0 0

[0408] Table 39 shows the final product distribution of the intact conjugates 24 hrs after initiation of the conjugation reaction. The results show that some of esters did not react at all (Z4, Z12), others reacted sluggishly (e.g. Z5), while several gave profiles approaching that of PFP (Z1) (e.g. Z3).

Conjugation Kinetics

[0409] The rates of addition over time for each of the final conjugates are shown in Tables 39, 40, 41, and 42. OCA represents underivatized 2.12.1.fx antibody, whereas 1, 2 or 3CA represents additions of 1, 2 or 3 peptides to the 2.12.1.fx antibody at each of the time periods examined.

TABLE-US-00042 TABLE 40 Conjugation kinetics of different Z* groups yielding 0 CA. Z8 0CA Z2 Z3 Z4 Z5 Z6 Z7 NH5 time Z1 2,3,4 2,3,6 2,3,6 2,6 2,4 5,7 NB2,3 Z9 Z10 Z11 Z12 (hr) PFP TFP TFP TCP DCP DCN DCQ DCI 2HI 1,3 D 4NP 2,6-DFP 1 nap  0 84 97 94 100 100 100 100 95 95 96 100 100  1 5 83 58 100 100 95 96 43 24 79 93 100  2 4 75 40 100 100 89 93 42 20 67 88 100  4 4 62 27 100 96 81 88 40 20 54 79 100 24 3 32 17 100 81 38 73 34 20 41 50 100

TABLE-US-00043 TABLE 41 Conjugation kinetics of different Z* groups yielding 1 CA. Z2 Z3 Z4 Z5 Z6 Z7 Z8 Z9 Z11 1CA Z1 2,3,4 2,3,6 2,3,6 2,6 2,4 5,7 NH5 2HI Z10 2,6- Z12 time (hr) PFP TFP TFP TCP DCP DCN DCQ NB2,3 DCI 1,3 D 4NP DFP 1 nap  0 16 3 6 0 0 0 0 5 5 5 0 0  1 38 17 36 0 0 5 4 39 39 21 8 0  2 37 25 45 0 0 11 7 39 38 29 12 0  4 33 34 43 0 4 19 12 42 39 37 21 0 24 34 45 43 0 19 45 25 40 36 42 39 0

TABLE-US-00044 TABLE 42 Conjugtion kinetics of different Z* arouos yielding 2 CA. Z8 Z2 Z3 Z4 Z5 Z6 Z7 NH5 Z9 2CA Z1 2,3,4 2,3,6 2,3,6 2,6 2,4 5,7 NB2,3 2HI Z10 Z11 Z12 time (hr) PFP TFP TFP TCP DCP DCN DCQ DCI 1,3 D 4NP 2,6-DFP 1 nap  0 0 0 0 0 0 0 0 0 0 0 0 0  1 49 0 6 0 0 0 0 15 27 0 0 0  2 50 0 14 0 0 0 0 16 30 4 0 0  4 52 4 25 0 0 0 0 15 29 9 0 0 24 51 20 30 0 0 16 2 18 31 15 11 0

TABLE-US-00045 TABLE 43 Conjugation kinetics of different Z* groups yielding 3 CA. Z8 Z2 Z3 Z4 Z5 Z6 Z7 NH5 Z9 3CA Z1 2,3,4 2,3,6 2,3,6 2,6 2,4 5,7 NB2,3 2HI Z10 Z11 Z12 time (hr) PFP TFP TFP TCP DCP DCN DCQ DCI 1,3 D 4NP 2,6-DFP 1 nap  0 0 0 0 0 0 0 0 0 0 0 0 0  1 8 0 0 0 0 0 0 3 11 0 0 0  2 10 0 2 0 0 0 0 3 12 0 0 0  4 12 0 5 0 0 0 0 4 12 0 0 0 24 12 3 11 0 0 2 0 5 12 3 0 0

Light and Heavy Chain Distribution

[0410] The extent of peptide conjugation for each of the alternatively activated esters was examined separately on the light and heavy chains. Each sample was denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. The peptide conjugation profile on the light and heavy chain of 2.12.1.fx and mutants are shown in Table 44. Almost all of the activated peptides listed in the table showed reduced conjugation level on light chain compared to the compound using PFP (Z1), except 2,3,6-trifluorophenyl (Z3), which showed a similar level of conjugation. Activated esters derived from N-hydroxysuccinimide (NHS), i.e. N-Hydroxyl-5-norbornene-2,3-dicarboxylic acid imide and 2-hydroxyl-isoindoline-1,3-dione (Z8 and Z9) showed a greater propensity for heavy chain derivatization.

TABLE-US-00046 TABLE 44 Summary of activated ester results. Z* Z* Z* Time course of conjugation adducts Reduced conjugation # Name Structure [separate 24 hr expt in bold] at 24 hr  1 Penta Fluoro Phenyl [00080] CA 0 1 2 3  0  84  16  0  0  1  5  38  49  8  2  4  37  50  10  4  4  33  52  12  24  3  34  51  12 LC 30 LC + 1CA 64 L + 2CA 7 HC  94 HC + 1CA  6 HC + 2CA —  2 2,3,4- trifluoro- phenyl [00081] CA 0 1 2 3  0  97  3  0  0  1  83  17  0  0  2  75  25  0  0  4  62  34  4  0  24  32  45  20  3 LC 59 LC + 1CA 41 L + 2CA — HC  94 HC + 1CA  6 HC + 2CA —  3 2,3,6- trifluoro- phenyl [00082] CA 0 1 2 3  0  94  6  0  0  1  58  36  6  0  2  40  45  14  2  4  27  43  25  5  24  17  43  30  11 LC 30 LC + 1CA 64 L + 2CA 7 HC  90 HC + 1CA 10 HC + 2CA —  4 2,3,6- trichloro- phenyl [00083] CA 0 1 2 3  0 100  0  0  0  1 100  0  0  0  2 100  0  0  0  4 100  0  0  0  24 100  0  0  0 LC 95 LC + 1CA  5 L + 2CA — HC 100 HC + 1CA — HC + 2CA —  5 2,6 dichloro- phenyl [00084] CA 0 1 2 3  0 100  0  0  0  1 100  0  0  0  2 100  0  0  0  4  96  4  0  0  24  81  19  0  0 LC 89 LC + 1CA 11 L + 2CA — HC 100 HC + 1CA — HC + 2CA —  6 2,4 DiCl Napthalene [00085] CA 0 1 2 3  0 100  0  0  0  1  95  5  0  0  2  89  11  0  0  4  81  19  0  0  24  38  45  16  2 LC 66 LC + 1CA 34 L + 2CA — HC  95 HC + 1CA  5 HC + 2CA —  7 5,7- dichloro- quinolin- 8-yl [00086] CA 0 1 2 3  0 100  0  0  0  1  96  4  0  0  2  93  7  0  0  4  88  12  0  0  24  73  25  2  0 LC 92 LC + 1CA  8 L + 2CA — HC  95 HC + 1CA  5 HC + 2CA —  8 N- Hydroxyl- 5- norbornene- 2,3- dicarboxylic [00087] CA 0 1 2 3  0  95  5  0  0  1  43  39  15  3  2  42  39  16  3  4  40  42  15  4  24  38  40  18  5 LC 77 LC + 1CA 23 L + 2CA — HC  82 HC + 1CA 18 HC + 2CA — acid imide  9 2- hydroxyl- isoindoline- 1,3-dione [00088] CA 0 1 2 3  0  95  5  0  0  1  24  39  27  11  2  20  38  30  12  4  20  39  29  12  24  20  36  31  12 LC 70 LC + 1CA 30 L + 2CA — HC  50 HC + 1CA 50 HC + 2CA — 10 4-nitro- phenyl [00089] CA 0 1 2  0  96  5  0  1  79  21  0  2  67  29  4  4  54  37  9  24  41  42  15 LC 68 LC + 1CA 32 L + 2CA — HC  92 HC + 1CA  8 HC + 2CA — 3  0  0  0  0  3 11 2,6- difluoro- phenyl [00090] CA 0 1 2 3  0 100  0  0  0  1  93  8  0  0  2  88  12  0  0  4  79  29  0  0  24  50  39  11  0 12 1- naphthyl [00091] CA 0 1 2 3  0 100  0  0  0  1 100  0  0  0  2 100  0  0  0  4 100  0  0  0  24 100  0  0  0

Example 19

[0411] Further examples of alternatively activated esters are shown in Table 45. The time-course of conjugation of several analogs of PFP esters were examined. By decreasing the number and position of the fluorine groups in PFP, less reactive active ester forms can be synthesized and investigated. 2,3,5,6-tetrafluorophenyl ester and 2,4,6-trifluorophenyl ester were both tested after conjugation to [PEG.sub.2-MAL-K.sub.SH.sup.11-SEQ:27]. 1-hydroxyl-pyrrolidine-2,5-dione (NHS) was conjugated to [PEG.sub.5-K.sup.11-SEQ:27].

##STR00092##

[0412] After 2 hrs conjugation, these less activated forms gave lower overall conjugation to 2.12.1.fx than PFP. NHS group also showed lower overall conjugation. NHS and PFP-containing peptides were conjugated to 2.12.1.fx. The reduced forms were analyzed to see the distribution at 2 hrs. PFP showed a much greater propensity for light chain derivatization (77% overall to LC, only 6% to heavy) compared to 1-hydroxyl-pyrrolidine-2,5-dione (NHS) (31% overall to LC, but 34% overall to heavy).

TABLE-US-00047 TABLE 45 Alternatively activated esters—further examples. Active esters—reduced analysis of Name Structure CA at 2 hr conjugation at 2 hr  1 Penta Fluoro Phenyl [00093] CA 0 1 2 3 4 2 hr  3 40 42 14  1 LC 23 LC + 1CA 72 LC + 2CA 5 HC 94 HC + 1CA  6 HC + 2CA 0 13 1- hydroxyl- pyrrolidine- 2,5-dione (NHS) [00094] CA 0 1 2 3 4 2 hr 18 44 24 12  3 LC 70 LC + 1CA 28 LC + 2CA 3 HC 66 HC + 1CA 31 HC + 2CA 3 14 2,3,5,6- tetra- fluoro- phenyl [00095] CA 0 1 2 3 4 2 hr 21 44 29  5  2 15 2,4,6- trifluoro phenyl [00096] CA 0 1 2 3 4 2 hr 80 27  2  0  0

[0413] Compounds Z1-Z15 represent a variety of different structural types of active ester. It is enlightening to consider the series of fluorinated aromatic active esters, which have a different number and pattern of substitution of fluorine atoms around the aromatic ring (compounds Z1, Z2, Z3, Z11, Z14 and Z15) and consider how their structure influences their reactivity and propensity for protein derivatization. The kinetics of the antibody-conjugation of these derivatives can be conveniently compared at the 2 hr time-point, when the pentafluorophenyl (Z1) reaction has gone to completion. With an increasing level of fluorine substitution around the ring, there is an increasing level of overall conjugation and a concomitant decrease in unreacted antibody. The rate of reaction is directly related to the pKa of the fluorinated phenol leaving group, with the most acidic phenols giving higher reaction rates. The rates of conjugation are Z1>Z14>Z3>Z15>Z2>Z11. The subtle effects of the fluorine substitution patterns can be seen by comparing compounds Z2, Z3 and Z15.

[0414] The structure of the active ester also significantly affected the directionality of the conjugation reaction. In general, the fluorinated aromatic esters showed a marked propensity towards light chain derivatization (principally CLκ-K.sup.80 as previously mentioned). In contrast, several esters based on N-hydroxysuccinimide derivatives (Z8, Z9 and Z13) showed less preference, with often greater levels of heavy chain derivatization observed.

Example 20

[0415] The rate of conjugation between MAC-1 (PEG.sub.2-MAL-mercaptopropionyl linker between the peptide and PFP activating group) and MAC-2 (straight-chained PEG.sub.5 linker between the peptide and PFP activating group) was assessed. Table 46 compares these activated peptides to 2.12.1x. The results show that the activated peptides behave very similarly in terms of the rate and extent of derivatization, despite their slightly different linker structures.

TABLE-US-00048 TABLE 46 Comparison of conjugation between MAC-1 and MAC-2. MAC-2 MAC-1 Intact time (min) OCA 1CA 2CA 3CA 4CA OCA 1CA 2CA 3CA 4CA   0 72 27 1 0 0 82 18 1 0 0  10 26 56 17 1 0 29 49 20 2 0  20 13 53 29 5 0 15 47 33 5 0  30 9 51 32 8 1 9 43 40 8 0  40 7 45 39 9 1 8 41 41 8 2  50 6 43 39 11 1 7 41 42 9 2  60 5 41 40 11 2 6 36 45 11 2  70 4 40 40 14 2 6 35 46 11 2  80 3 38 44 14 2 5 36 47 10 2  90 4 37 45 13 1 6 35 46 12 2 100 4 40 41 13 2 6 35 46 11 2 110 3 40 42 14 1 6 34 46 12 3 120 4 37 44 13 1 5 35 46 12 2

Example 21 Effect of Linker Length

[0416] The effect on the final conjugate distribution profile of having different lengths of linker was examined. Compounds were synthesized with different PEG length linkers joining the peptide to the PFP group. The results for the addition to 2.12.1.fx of 0, 1, 2, 3 and 4 peptides are summarized in Table 47. Overall, changing the length of the PEG linker had generally little effect on the distribution of conjugates obtained.

TABLE-US-00049 TABLE 47 Effect of linker length. Y.sup.1 = [00097]   CA (%) n 0 1 2 3 4 2 8 39 44 8 0 3 6 34 47 10 2 5 4 37 44 13 1 7 4 35 49 11 0 9 3 28 49 19 2 13 3 32 54 10 0 17 6 37 51 7 0 21 4 43 45 5 2 25 11 44 38 7 0 [00098]

Example 22 Conjugation of Alternative Peptide Sequences

[0417] To confirm the applicability of the invention across other peptide sequences, SEQ ID NO:80 and SEQ ID NO:81 (Test-peptides-1, and -2) were conjugated. SEQ ID NOs:80 and 81 were conjugated with [PFP-PEG.sub.5] and then the 2.12.1.fx under conditions previously optimized for reaction with [PFP-PEG.sub.5-K.sup.11-SEQ:27]. The results of analysis of the conjugation profile and LC/HC conjugation are shown in Table 48. SEQ ID NO:80 and SEQ ID NO:81 both showed directional conjugation to the light chain. On further analysis of the LC/HC distributions, similar profiles to that of MAC-2 were observed, with around 70% LC derivatization and less than 10% on the HC.

TABLE-US-00050 TABLE 48 Conjugation profile of SEQ ID NO 80 and SEQ ID NO: 81. LC % CA HC % CA SEQ % CA LC LC HC HC ID NO: 0 1 2 3 4 LC +1 +2 HC +1 +2 27 2 24 55 17 3 24 65 11 91 9 — 80 11 39 43 8 0 32 68 — 95 5 — 81 8 35 48 10 0 29 71 — 94 6 — [00099]  [00100]

Example 23 Summary of Peptide Conjugation Analysis

[0418] Peptide mapping experiments were performed on a range of protein/conjugate combinations for the purpose of confirming the important parameters that lead to directional conjugation at CLκ-K.sup.80 on antibody light chains. Table 49 lists the results of the peptide mapping experiments performed. For each study parameter, the peptide mapping procedure described earlier was used. “***” indicates a high level of directional conjugation to CLκ-K.sup.80. “**” and to a lesser extent, “*”, indicates directional conjugation is still observed, but may show differences, such as slower reaction conditions, less overall conjugation, or averaging at one light chain only, and so may be more suitable to special circumstances, such as generating MACs with between 0.5 and 1.5 peptide per antibody (for example). “-” indicates that these reaction conditions did not appear favorable towards directional conjugation at CLκ-K.sup.80.

[0419] As CLκ-K.sup.80 was observed in MAC-2 to be the location of directional conjugation, peptide mapping studies on alternative parameters focused on this location. Detailed peptide mapping data for each study parameter is not included, but significant conjugation levels at other K residues was not observed, and observations of other MACs were consistent with directional conjugation at CLκ-K.sup.80.

[0420] CLκ-K.sup.80R and CLκ-K.sup.80A mutations of 2.12.1.fx resulted in the loss of directional conjugation at this site; suggesting an essential role for this specific residue. CLκ-K.sup.82R, and CLκ-K.sup.82A mutations did not hinder directional conjugation to CLκ-K.sup.80, and may even enhance it. Of the other study parameters examined, at least a portion of the sub-type of light chain constant region was observed to have a significant impact on directional conjugation; at least a portion of the light chain sub-type kappa was determined to be necessary. Conjugation onto a CLλ sub-type (using an exemplary A containing antibody, hAbλTest1), did not demonstrate directional conjugation. When the CLλ of hAbλTest1 was mutated to a CLκ, directional conjugation at K.sup.80 was recovered.

TABLE-US-00051 TABLE 49 Summary of directional conjugation at CL.sub.K-K.sup.80. Mutations/ Differences Directional Antibody LC Vs MAC1/2 SEQ ID NO Linker Z* conjugation 2.12.1.fx κ 27 PEG.sub.2-MAL PFP * * * 2.12.1.fx κ 27 PEG.sub.5 PFP * * * 2.12.1.fx Fab κ 27 PEG.sub.5 PFP * * * h38C2-IgG1 κ 27 PEG.sub.5 PFP * * * h38C2-IgG2 κ 27 PEG.sub.5 PFP * * * hAbλTest λ K.sup.80SH 27 PEG.sub.5 PFP — hAbκTest1 κ 27 PEG.sub.5 PFP * * * hAbκTest3 κ 39-mer PEG.sub.5 PFP * * * hAbλTest λκ 27 PEG.sub.5 PFP * * * hAbλTest λκJ 27 PEG.sub.5 PFP * * * 2.12.1.fx κ K.sup.80R 27 PEG.sub.5 PFP — 2.12.1.fx κ K.sup.82R 27 PEG.sub.5 PFP * * * 2.12.1.fx κ K.sup.80R/R.sup.82R 27 PEG.sub.5 PFP — 2.12.1.fx κ D.sup.43A 27 PEG.sub.5 PFP * * 2.12.1.fx κ K.sup.80A 27 PEG.sub.5 PFP — 2.12.1.fx κ H.sup.81A 27 PEG.sub.5 PFP — 2.12.1.fx κ H.sup.82A 27 PEG.sub.5 PFP * * * 2.12.1.fx κ D.sup.43A/H.sup.81A 27 PEG.sub.5 PFP — hAbλTest1 λ S.sup.81H/H.sup.82S 27 PEG.sub.5 PFP — 2.12.1.fx κ 39-mer PEG.sub.5 PFP * * * 2.12.1.fx κ 80 PEG.sub.5 PFP * * * 2.12.1.fx κ 81 PEG.sub.5 PFP * * * h38C2-IgG2 κ 39-mer PEG.sub.5 PFP * * * 2.12.1.fx Fab κ biotin PEG.sub.5 PFP * * * 2.12.1.fx κ 27 PEG.sub.2-MAL PFP * * * 2.12.1.fx κ 27 PEG.sub.2-MAL 2,3,4 TFP (2) * * 2.12.1.fx κ 27 PEG.sub.2-MAL 2,3,6 TFP (3) * * 2.12.1.fx κ 27 PEG.sub.2-MAL 2,3,6 TCP (4) — 2.12.1.fx κ 27 PEG.sub.2-MAL 2,6 DCP (5) — 2.12.1.fx κ 27 PEG.sub.2-MAL 2,4 DCN (6) * 2.12.1.fx κ 27 PEG.sub.2-MAL 5,7 DCQ (7) — 2.12.1.fx κ 27 PEG.sub.2-MAL NH-5-N2,3DI (8) * 2.12.1.fx κ 27 PEG.sub.2-MAL 2Hi1,3 DIO (9) * 2.12.1.fx κ 27 PEG.sub.2-MAL 4NP (10) * * 2.12.1.fx κ 27 PEG.sub.2-MAL 2,6 DFP (11) * * 2.12.1.fx κ 27 PEG.sub.2-MAL NAP (12) — 2.12.1.fx κ 27 PEG.sub.2-MAL 1HP 2,5D (13) * 2.12.1.fx κ 27 PEG.sub.2-MAL 2,3,5,6 TFP (14) * * 2.12.1.fx κ 27 PEG.sub.2-MAL 2,4,6 TFP (15) * * 2.12.1.fx κ 27 PEG.sub.2-MAL Squarate * 2.12.1.fx κ 27 PEG.sub.2-MAL AZD * 2.12.1.fx κ 27 PEG.sub.2-17 PFP * * * 2.12.1.fx κ 27 PEG.sub.17-21 PFP * * 2.12.1.fx κ 27 PEG.sub.25 PFP * *

Example 24 Examination of CLκ-D.SUP.77

[0421] Residues geographically close to the CLκ-K.sup.80 in the 3-D structure were examined. Initial Crystal structure analysis suggested the possibility that CLκ-D.sup.77 could form a salt bridge with CLκ-K.sup.80, which could have an impact on the CLκ-K.sup.80 directional conjugation. In order to study the effect of CLκ-D.sup.77 on conjugation to CLκ-K.sup.80, CLκ-D.sup.77 was mutated to CLκ-A.sup.77 on the 2.12.1.fx antibody to create 2.12.1.fx-[CLκ-D.sup.77A] (CLκ of SEQ ID NO:37). The CLκ-D.sup.77A mutation was generated on an antibody light chain following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). The mutation was introduced by oligonucleotide primers and confirmed by DNA sequencing. 2.12.1.fx-[CLκ-D.sup.77A] was transiently expressed in HEK 293 cells, and purified using Protein A affinity column. The purified mAbs were characterized using MS.

[0422] 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A] (1 mg reaction size) were adjusted to 18 mg/ml to pH 7.7 with a phosphate buffer to a final concentration of 0.06M sodium phosphate. The exemplary test peptide-linker pair [PFP-PEG.sub.5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg/ml. [PFP-PEG.sub.5-K.sup.11-SEQ:27] was added to antibody at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT. The conjugated product was diluted to 2 mg/ml and analyzed as an intact conjugated protein by SEC-MS to determine the number and quantitation of conjugate forms of the protein. Relative quantitation of multiple peptide-linker conjugation species was performed by measuring the signal magnitude.

[0423] Table 50 compares the conjugation profile of [2.12.1.fx]-[PEG.sub.5-K.sup.11-SEQ:27] and [2.12.1.fx-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ:27]. The conjugation profile of [2.12.1.fx]-[PEG.sub.5-K.sup.11-SEQ:27] occurs as a distribution between 0-4 peptide additions with the largest form being 2 peptide additions and the average number of peptide additions is 2.16. The profile changes when the residue CLκ-D.sup.77 is mutated to CLκ-A.sup.77 in the scaffold protein; the average number of peptide additions rises to 2.38 and significantly less single peptide addition is observed. This result suggests that the single point mutation CLκ-D.sup.77A has the effect of increasing the overall conjugation to the scaffold. In both conditions replicate analysis (n=3) demonstrates that the conjugations profiles observed are reproducible.

[0424] The extent of peptide conjugation was examined separately on the light and heavy chains of 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A]. The produced MACs were denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. Table 50 demonstrates that the average conjugation is higher on the light chain of 2.12.1.fx-[CLκ-D.sup.77A] than 2.12.1.fx; the average conjugate addition value for 2.12.1.fx-[CLκ-D.sup.77A] is 1.15 compared to 0.85 for 2.12.1.fx. In addition, unconjugated light chain is undetected in 2.12.1.fx-[CLκ-D.sup.77A]. Conjugation on the heavy chain is observed at a significantly lower level. The majority of observed heavy chain for both 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A] is unconjugated; this is especially true in the case of 2.12.1.fx-[CLκ-D.sup.77A] heavy chain. These results suggest that the CLκ-D.sup.77A mutation alters the light chain to make it significantly more susceptible to conjugation. Replicate analysis of this experiment by multiple scientists is shown in Table 50 which demonstrates that the profile of conjugation is consistent and reproducible.

TABLE-US-00052 TABLE 50 Analysis of conjugation of [PEG.sub.5-K.sup.11-SEQ: 27] to Abs 2.12.1.fx (WT) and 2.12.1.fx-[CLK-D77A]. Rep = replicate. AvReps is the average of the results of the three replicate experiments, with the standard deviation shown beneath (StdDev). Ab % CA shows % conjugations additions per antibody, followed by the average CA per antibody. Reduced light and heavy chain analysis also shown, with respective average CA per chain. CLκ-D.sup.77A shows 123% rate of LC % 1CA species compared to native CLκ. Ab % CA Avg. LC % CA Avg. HC % CA Avg. Ab 0 1 2 3 4 CA 0 1 2 CA 0 1 CA WT Rep1 1 17 56 22 4 2.11 21 72 7 0.86 86 14 0.14 Rep2 2 14 54 25 5 2.18 25 67 8 0.83 85 15 0.15 Rep3 2 13 55 25 6 2.2 23 69 8 0.85 85 15 0.15 Av.Rep 2 15 55 24 5 2.16 23 69 8 0.85 85 15 0.15 StdDev 1 2 1 2 1 0.05 2 3 1 0.02 1 1 0.01 D.sup.77A Rep1 5 4 54 30 8 2.32 0 84 16 1.16 94 6 0.06 Rep2 6 4 43 36 12 2.44 0 85 15 1.15 94 6 0.06 Rep3 3 3 56 31 7 2.37 0 86 14 1.14 94 6 0.06 AvRep 5 4 51 32 9 2.38 0 85 15 1.15 94 6 0.06 StdDev 2 1 7 3 3 0.06 0 1 1 0.01 0 0 0.00

Example 25 Peptide Mapping Characterization of 2.12.1.Fx-[CLκ-D.SUP.77.A]-[SEQ:27-K.SUP.11.-PEG5] Heavy and Light Chain Reference Product

[0425] 2.12.1.fx-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ:27] conjugated antibody was reduced with dithiothreitol and cysteine residues were alkylated by carboxymethylation with iodoacetamide. Chymotrypsin was used for proteolytic digestion. Digested fragments in solution were analyzed using LCMS. Individual fragments were separated over a C18 HPLC column and their accurate mass was measured in a Q-Tof mass spectrometer. The resulting fragment mass was used to identify unmodified fragments or fragments modified with a [PEG.sub.5-K.sup.11-SEQ:27] conjugation group. This experiment was interpreted by focusing on chymotryptic fragments that contain a lysine residue and are therefore possible sites for peptide conjugation. Tables 51 and 52 list of all such fragments on the heavy chain and light chains respectively. Blank entries are fragments that were not detected using this technique. Detected fragments that were observed with a [PEG.sub.5-K.sup.11-SEQ:27] modifier are considered potential sites of peptide conjugation.

[0426] 2 [PEG.sub.5-K.sup.11-SEQ:27]-conjugated fragments were detected using LCMS peptide mapping of the 2.12.1.fx-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ:27] product. Both of these conjugated fragments were present on the light chain of the 2.12.1.fx-[CLκ-D.sup.77A] antibody. In comparison, 8 fragments conjugated to [PEG.sub.5-K.sup.11-SEQ:27] were detected in 2.12.1.fx-[PEG.sub.5-K.sup.11-SEQ:27].

[0427] Overall, these results suggest that conjugation levels in the CLκ-D.sup.77A mutant are elevated at fewer conjugation sites, possibly suggesting increased conjugation specificity relative to the unmutated antibody. Further, structural analysis has shown that the CLκ-D.sup.77 residue is in close proximity (<10 As) to the identified major conjugation site CLκ-K.sup.80. It was speculated that an electrostatic interaction, possibly a salt bridge, could exist between the carboxylic acid of CLκ-D.sup.77 and the primary amine of CLκ-K.sup.80. The CLκ-D.sup.77A mutation would disrupt such an electrostatic interaction, resulting in the reactive amine on CLκ-K.sup.80 being more exposed and susceptible to conjugation with the reactive esters of the invention. Although subsequent analysis coupled with sophisticated modeling helped build a more complete picture of the reaction site, and indicated that CLκ-D.sup.77 exerted its effect primarily through its interaction with CLκ-H.sup.81, the initial hypothesis of an interaction between CLκ-D.sup.77 and CLκ-K.sup.80 was helpful in underlining the significance of the CLκ-D.sup.77 residue.

[0428] The observed conjugation sites in the 2.12.1.fx-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ:27] product are light chain chymotrypsin fragments Y3 and Y15. Analysis of the signal intensities for these fragments suggests that fragment Y15, which carries the CLκ-K.sup.80 residue, is the primary conjugation site. Fragment Y15 is only observed as an [PEG.sub.5-K.sup.11-SEQ:27]-modified fragment at a very high signal intensity (1118572 counts, Table 52), whilst the unmodified form of Y15 is not observed, suggesting that all or nearly all of fragment Y15 exists in the modified form. Fragment Y3 is observed in both the [PEG.sub.5-K.sup.11-SEQ:27]-modified and unmodified forms; unmodified Y3 signal intensities in 2.12.1.fx-[PEG.sub.5-K.sup.11-SEQ:27] and 2.12.1.fx-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ:27] are within 15%. [PEG.sub.5-K.sup.11-SEQ:27]-modified Y3 is observed at a relatively low level (9737 counts, Table 52).

The table entries for Tables 51 and 52 are explained below:
Fragment number: Chymotrypsin fragments numbering from the N-terminus; joined fragments (ie—Y1-2) indicate a missed cleavage site.
Start/End: Numbering of the fragment location from the N-terminus.
Fragment Mass (Da): Theoretical mass of the fragment listed in Daltons.
Retention Time (Control/Analyte): Time of chromatographic retention/elution in the LCMS peptide mapping experiment.
MS Signal Intensity (Control/Analyte): Magnitude of observed signal observed by MS.
Mass Error-ppm (Control/Analyte): Comparison of theoretical vs. observed mass of the peptide fragment; values closer to zero (0) demonstrate better mass accuracy. The control protein for Retention Time, MS signal intensity and Mass Error is 2.12.1.fx-[CLκ-D.sup.77A] and the analyte protein in each case is 2.12.1.fx-[CLκ-D.sup.77A]+[PEG.sub.5-K.sup.11-SEQ:27].
Modifiers: Potential covalent additions to the fragment; [PEG.sub.5-K.sup.11-SEQ:27]- antibody binding peptide of Lysine residue, CAM-carboxymethylation of Cysteine residue.

TABLE-US-00053 TABLE 51 Peptide mapping characterization of 2.12.1.fx-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ: 27] heavy chain referenceproduct. Fragment Fragment Retention Time MS Signal Intensity Mass Error (ppm) Number Start End Mass (Da) Control Analyte Control Analyte Control Analyte Modifiers Y1 1 27 2617.3533 Y1-2 1 29 2865.4695 Y5-6 34 47 1657.8398 Y6 37 47 1253.688 19.2 19.2 516640 548267 1.9 0 Y6-7 37 50 1602.8518 22.1 22.1 26537 31229 −1.6 −2 Y8-9 51 68 1931.9337 16.5 16.5 60894 82459 −2.2 0.5 Y9 61 68 878.461 11.3 11.4 376224 403402 0 −0.3 Y9-10 61 80 2241.1501 Y10 69 80 1380.6997 13.3 13.3 261813 286406 −1.1 0.3 Y10-011 69 94 2972.4661 Y19-20 111 157 4748.2773 Y20 116 157 4160.0405 Y20-21 116 166 5202.5527 Y20-21* 116 166 5316.5957 34.1 34.1 6445 8275 0.5 −5.7 CAM(2) Y24-25 202 245 4702.2109 Y25 207 245 4151.9722 Y25* 207 245 4437.0796 20.9 20.9 1495322 1771622 1.1 0.7 CAM(5) Y25-26 207 279 7985.9092 Y26 246 279 3851.9478 Y26-27 246 281 4152.0698 Y28-29 282 300 2245.1128 Y29 283 300 2082.0493 14.6 14.6 20665 18618 −0.6 −0.2 Y29-30 283 304 2531.2405 Y31-32 305 323 2241.1907 Y32 318 323 722.3599 7.9 7.9 93966 81618 0.1 3 Y32 318 323 722.3599 17.7 18.4 37943 11371 11.4 27.1 Y32 318 323 722.3599 18.4 11761 23.8 Y32-33 318 353 4028.188 Y33 324 353 3323.8386 20 5422 3.1 Y33* 324 353 3380.8601 19.7 19.7 2196329 2374835 −2.5 −3.8 CAM(1) Y33-34 324 376 5883.1577 Y34 354 376 2577.3293 Y34-35 354 385 3637.8159 Y34-35* 354 385 3694.8374 33 33 10095 10026 1.9 −2 CAM(1) Y36-37 386 408 2527.0808 Y37 396 408 1394.6388 19.6 19.6 62942 65871 −0.9 −1.6 Y37-38 396 409 1541.7072 25.1 25.1 827336 874876 0 −3.8 Y39-40 410 421 1494.8195 Y40 412 421 1218.672 15.8 15.8 77917 78774 −0.3 0.2 Y40-41 412 427 1891.9905 20.3 20.3 107513 150305 0.2 0.1 Y42-43 428 450 2525.1792 Y43 441 450 1016.5502

TABLE-US-00054 TABLE 52 Peptide mapping characterization of 2.12.1.fx-[CLκ-D.sup.77A]- [PEG.sub.5-K.sup.11-SEQ: 27] light chain reference product. Peptide Fragment Mass Retention Time MS Signal Intensity Mass Error (ppm) Number Start End (Da) Control Analyte Control Analyte Control Analyte Modifiers Y2-3 36 49 1688.9725 16.2 16.2 145374 172582 −1.7 −1.1 Y3 37 49 1525.9093 15.5 15.5 331068 390175 −2.7 −2.4 Y3* 37 49 3218.7593 24 9737 −6.8 [SEQ: 27-K11- PEG5](1) Y3-4 37 62 2882.6355 Y9-10 88 116 3244.729 Y10 99 116 1871.0992 Y10-11 99 139 4331.335 Y11 117 139 2478.2463 22.8 61217 −1.7 Y11* 117 139 2535.2678 21.5 5127 −16.8 CAM(1) Y11-12 117 148 3635.8445 Y12 140 148 1175.6088 Y12-13 140 173 3886.8245 Y13 149 173 2729.2263 13.1 13.1 1140556 1153543 −1.1 −1.8 Y13-14 149 186 4051.9346 Y14 174 186 1340.7188 Y14-15 174 192 2125.1418 Y15 187 192 802.4337 7.5 275639 −1.9 Y15* 187 192 2495.2837 20.9 1118572 −6.3 [SEQ: 27-K11- PEG5](1) Y15-16 187 209 2574.29 Y16 193 209 1789.8668 18.7 5400 4.4 Y16* 193 209 1846.8883 18.1 18.1 169490 246823 −1.7 −2.2 CAM(1) Y16-17 193 214 2349.0842 17.8 9211 0.1

Example 26 Examination of CLκ-D.SUP.77 .Mutations

[0429] CLκ-D.sup.77 residue of 2.12.1.fx antibody was mutated to each of the other 18 amino acids in addition to the CLκ-D.sup.77A mutation. The CLκ-D.sup.77G (SEQ ID NO:38), CLκ-D.sup.77L (SEQ ID NO:40), CLκ-D.sup.77S (SEQ ID NO:49), CLκ-D.sup.77E (SEQ ID NO:53), and CLκ-D.sup.77R (SEQ ID NO:54) and mutants were generated following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The other 13 mutants on the CLκ-D.sup.77 site (CLκ-D.sup.77V (SEQ ID NO:39), CLκ-D.sup.771 (SEQ ID NO:41), CLκ-D.sup.77P (SEQ ID NO:42), CLκ-D.sup.77F (SEQ ID NO:43), CLκ-D.sup.77W (SEQ ID NO:44), CLκ-D.sup.77Y (SEQ ID NO:45), CLκ-D.sup.77H (SEQ ID NO:46), CLκ-D.sup.77M (SEQ ID NO:47), CLκ-D.sup.77C (SEQ ID NO:48), CLκ-D.sup.77T (SEQ ID NO:50), CLκ-D.sup.77Q (SEQ ID NO:51), CLκ-D.sup.77N (SEQ ID NO:52), CLκ-D.sup.77K (SEQ ID NO:55)) were generated following protocols described in Quick PCR Cloning Kit (BPS Bioscience). Mutations were introduced by oligonucleotide primers and cloned to a modified p2.12.1.fxP4 vector (Invitrogen) cut with BgIII and NheI. Insert DNA were confirmed by DNA sequencing. The mutated mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0430] 2.12.1.fx and 2.12.1.fx mutants were adjusted 18 mg/ml to pH 7.7 with a phosphate buffer to a final concentration of 0.06 M sodium phosphate. [PFP-PEG.sub.5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg/ml. The peptide/linker was added to antibody at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT.

[0431] Table 53 describes the overall conjugation profile of the CLκ-D.sup.77 mutants. CLκ-D.sup.77C aggregated due to an introduction of a free cysteine, and the results were not interpretable. Mutations CLκ-D.sup.77W, CLκ-D.sup.77M, CLκ-D.sup.77H, CLκ-D.sup.77Q, CLκ-D.sup.77N, and CLκ-D.sup.77V did not change the overall conjugation profile compared to wild-type 2.12.1.fx. Mutations CLκ-D.sup.77F, CLκ-CLκ-D.sup.77K, CLκ-D.sup.77Y, and CLκ-D.sup.77E decreased the overall level of conjugation. Mutations CLκ-D.sup.77P, CLκ-D.sup.771, CLκ-D.sup.77T, CLκ-D.sup.77R, CLκ-D.sup.77L, CLκ-D.sup.77S, and CLκ-D.sup.77G increased the level of conjugation.

[0432] Analysis of the reduced LC and HC showed that only the mutations CLκ-D.sup.77F, CLκ-D.sup.77K, CLκ-D.sup.77Y, CLκ-D.sup.77E, and CLκ-D.sup.77C resulted in reduced levels of conjugation on the light chain. The levels of conjugation on the light chain increased (and conjugation on the heavy slightly decreased) for CLκ-D.sup.77M, CLκ-D.sup.77H, CLκ-D.sup.77Q, CLκ-D.sup.77N, CLκ-D.sup.77W and CLκ-D.sup.77V. Mutations CLκ-D.sup.77P, CLκ-D.sup.77I, CLκ-D.sup.77T, CLκ-D.sup.77R, CLκ-D.sup.77L, CLκ-D.sup.77S, and CLκ-D.sup.77G increased the level of conjugation on the light chain by reducing the level of unconjugated light chain. CLκ-D.sup.77K increased the level of 2 conjugates on the light chain due to the introduction of another lysine, a potential conjugation site.

TABLE-US-00055 TABLE 53 Conjugation analysis of CLκ-D.sup.77 mutants to alternative amino acids (data is separated within the table according to the protein amount used during the conjugation reaction, or because the conjugation reaction was set up at a different time). The decreased level of conjugation for the tests run at 0.5 and 0.25 mg/ml was due to the low levels of the antibody. Ab % CA shows % conjugations additions per antibody, followed by the average CA per antibody. Reduced light and heavy chain analysis also shown, with respective average CA per chain. The % of 1-LC % relative to the respective WT run is shown in the right column: for example, D.sup.77Q, 1-LC % value of 81 is 126% of the respective WT 1-LC % of 69 for that experimental run. All samples were tested with 1 mg Ab, except the run including D.sup.77M, D.sup.77F and D.sup.77H (0.5 mg), and the run including D.sup.77W and D.sup.77C (0.25 mg). Ab % CA Avg. LC % CA Avg. HC % CA Avg. 1LC Ab 0 1 2 3 4 CA 0 1 2 CA 0 1 CA WT % WT 1 16 54 24 4 2.14 19 69 13 0.94 88 12 0.12 D.sup.77R 6 12 41 26 15 2.33 8 87 5 0.96 94 6 0.06 126 D.sup.77L 5 5 53 28 10 2.33 1 88 11 1.11 91 9 0.09 127 D.sup.77E 8 27 39 18 8 1.90 32 57 10 0.78 84 16 0.78 82 D.sup.77S 5 6 47 29 12 2.36 0 92 8 1.08 94 6 0.06 133 D.sup.77G 7 7 44 28 13 2.34 0 91 9 1.09 92 8 0.08 131 WT 2 24 50 21 3 1.99 24 66 10 0.85 90 10 0.1 D.sup.77Q 2 10 78 10 1.96 15 81 4 0.88 96 4 0.04 122 D.sup.77P 3 4 63 24 7 2.26 5 88 7 1.02 96 4 0.04 133 D.sup.77K 9 30 36 17 8 1.84 34 45 21 0.86 93 7 0.07 68 D.sup.77N 3 8 79 11 1.98 11 83 6 0.94 96 4 0.04 125 D.sup.77Y 28 41 19 11 1.14 54 42 4 0.51 89 11 0.11 63 D.sup.77V 6 10 63 21 1.98 13 79 9 0.96 95 5 0.05 119 D.sup.77I 3 4 63 24 7 2.27 6 85 9 1.03 93 7 0.07 128 D.sup.77T 3 11 59 18 9 2.17 8 88 4 0.97 95 5 0.05 133 WT 4 30 44 20 3 1.88 28 63 10 0.82 89 11 0.11 D.sup.77M 4 15 67 14 1.90 16 81 3 0.87 96 4 0.04 132 D.sup.77F 21 42 24 13 1.29 54 42 4 0.49 86 14 0.14 66 D.sup.77H 5 16 60 19 1.92 16 75 9 0.93 94 6 0.06 119 WT 5 36 47 12 1.66 30 60 10 0.8 90 10 0.1 D.sup.77W 8 29 46 18 1.73 20 75 4 0.84 91 9 0.09 125 D.sup.77C 60 37 2 0.42 95 5 0.05 61

Example 27 Effects of Other Mutations to the CHκ Region on Conjugation

[0433] In addition to CLκ-D.sup.77A, other residues within 10A distance to CLκ-K.sup.80 were mutated to alanine: CLκ-K.sup.41A (SEQ ID NO:20), V.sup.42A (SEQ ID NO:21), CLκ-D.sup.43A (SEQ ID NO:56), CLκ-N.sup.44A (SEQ ID NO:22), CLκ-L.sup.48A (SEQ ID NO:23), CLκ-Q.sup.47A (SEQ ID NO:24), CLκ-S.sup.48A (SEQ ID NO:25), CLκ-N.sup.80A (SEQ ID NO:26), CLκ-L.sup.73A (SEQ ID NO:28), CLκ-S.sup.74A (SEQ ID NO:29), CLκ-K.sup.75A (SEQ ID NO:30), CLκ-Y.sup.78A (SEQ ID NO:31), CLκ-E.sup.79A (SEQ ID NO:32), CLκ-H.sup.81A (SEQ ID NO:33), CLκ-V.sup.83A (SEQ ID NO:34), CLκ-Y.sup.84A (SEQ ID NO:35), and CLκ-R.sup.103A (SEQ ID NO:36) were also mutated to Ala. The data of CLκ-D.sup.43A and CLκ-H.sup.81A are discussed in Example 14.

[0434] The L.sup.73A mutant was introduced to 2.12.1.fx CLκ using the three way ligation method. A primer specific to the 5′ end of 2.12.1.fx-LC (2.12.1.fx.LC.FOR: SEQ ID NO:85) and a reverse primer containing the desired L.sup.73A mutation (L181A.REV: SEQ ID NO:88) were used to PCR the first half of the 2.12.1.fx-LC using 2.12.1.fx-LC DNA as the PCR template. This PCR fragment was then digested using restriction enzymes BgIII and BsaI. A forward primer containing CLκ-L.sup.73A mutation (L181A.FOR: SEQ ID NO:87) paired with the reverse primer specific to the 3′ end of 2.12.1.fx-LC (2.12.1.fx.LC.REV: SEQ ID NO:86) were used to PCR amplify the second half of 2.12.1.fx-LC DNA fragments carrying mutation using 2.12.1.fx-LC DNA as the PCR template. This PCR fragment was then digested using restriction enzymes BsaI and NheI. The two restriction enzyme digested PCR fragments were ligated with a modified p2.12.1.fxP4 plasmid (Invitrogen®) cut with BgIII and NheI. The insert sequence was confirmed by DNA sequencing. 2.12.1.fx-[CLκ-12.sup.3A] (i.e. comprising SEQ ID NO:28) was transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0435] The CLκ-V.sup.42A and CLκ-K.sup.75A mutants were generated by overlap PCR. Mutations were introduced by oligonucleotide primers. Primer specific to the 5′ end of 2.12.1.fx-LC (2.12.1.fx.LC.FOR) paired a reverse primer carrying the desired mutation, and a forward primer carrying the desired mutation paired with the reverse primer specific to the 3′ end of 2.12.1.fx light chain (2.12.1.fx.LC.FOR) were used to PCR amplify 2.12.1.fx-LC DNA fragments using 2.12.1.fx-LC as template. These two PCR products were mixed as templates; 2.12.1.fx-LC forward primer and reverse primer were used in overlap PCR reaction to amplify the full length 2.12.1.fx-LC DNA with desired mutation. The PCR was then digested with restriction enzyme BgIII and NheI. The digested PCR was ligated with a modified p2.12.1.fxP4 plasmid (Invitrogen®) cut with BgIII and NheI. The insert sequence was confirmed by DNA sequencing. The mutated mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0436] The other mutants were generated on 2.12.1.fx-LC following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The mutated mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0437] 2.12.1.fx and 2.12.1.fx mutants (1 mg reaction size) were adjusted 18 mg/ml to pH 7.7 with a phosphate buffer to a final concentration of 0.06 M sodium phosphate. [PFP-PEG.sub.5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg/ml. The peptide/linker was added to the antibody at a molar ratio of 4.3:1 and allowed to react for 2 hrs at room temperature.

[0438] Table 54 compares the conjugation profile of 2.12.1.fx-[PEG.sub.5-K.sup.11-SEQ:27] with 2.12.1.fx-[CLκ-mutants]-[PEG.sub.5-K.sup.11-SEQ:27]. The conjugation profile of 2.12.1.fx-[PEG.sub.5-K.sup.11-SEQ:27] occurs as a distribution between 0-4 peptide additions with the largest form being 2 peptide additions. The profile changes when the residues are mutated to Ala in the scaffold protein; the average number of [PEG.sub.5-K.sup.11-SEQ:27] additions either decreased (CLκ-V.sup.42A, CLκ-L.sup.46A, CLκ-S.sup.74A, CLκ-Y.sup.78A and CLκ-Y.sup.84A) or increased (CLκ-Q.sup.47A, CLκ-N.sup.50A and CLκ-D.sup.77A/E.sup.79A double mutants) compared to their corresponding 2.12.1.fx-[PEG.sub.5-K.sup.11-SEQ:27] controls. When comparing the conjugation profile of CLκ-D.sup.77A/E.sup.79A with CLκ-E.sup.79A, the significant increase of average [PEG.sub.5-K.sup.11-SEQ:27] additions to the antibody is mainly contributed by the CLκ-D.sup.77A mutation.

[0439] The extent of [PEG.sub.5-K.sup.11-SEQ:27] conjugation was examined separately on the light and heavy chains of 2.12.1.fx and 2.12.1.fx-[CLκ-mutants]. The MACs were denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each. Table 54 demonstrates that the 1CA on the light chain of 2.12.1.fx-[CLκ-Q.sup.47A] and —[CLκ-N.sup.50A] are higher than 2.12.1.fx. The average CAs are 0.84 and 0.85 compared to 0.78 of the 2.12.1.fx antibody. Both CLκ-Q.sup.47A and CLκ-N.sup.80A mutants have over 70% 1CA compared to the 59% 1CA of the 2.12.1.fx wild type antibody. In addition, the unconjugated light chain levels of these two mutants were reduced from 31% of the wild type antibody to 22% and 19%. The V.sup.42A had reduced level of light chain conjugation. The average light chain CA is 0.45 with 59% unconjugated light chain and 37% 1CA. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 54.

[0440] Conjugation of CLκ-V.sup.42A, CLκ-Q.sup.47A and CLκ-N.sup.80A to [PEG.sub.5-K.sup.11-SEQ:27] were repeated, and the results are shown at the bottom of the table. The elevated levels of light chain conjugation in CLκ-Q.sup.47A and CLκ-N.sup.80A and reduced light chain conjugation in CLκ-V.sup.42A were confirmed by both intact and reduced LC-MS analysis. Overall, the conjugation data suggests that CLκ-V.sup.42, CLκ-D.sup.43 and CLκ-H.sup.81 all have an impact on for PFP directional conjugation at CLκ-K.sup.80.

TABLE-US-00056 TABLE 54 Conjugation analysis of 2.12.1.fx variants conjugated to [SEQ: 27-K.sup.11- PEG5], showing amino acid mutants within10 Ås of light chain K.sup.80. Ab % CA shows % conjugations additions per antibody, followed by the average CA per antibody. Reduced light and heavy chain analysis also shown, with respective average CA per chain. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 53. AB % CA Avg. LC % CA Avg. HC % CA Avg. 1LC Ab 0 1 2 3 4 CA 0 1 2 CA 0 1 CA WT % WT 5 40 43 10 2 1.63 31 59 9 0.78 90 10 0.1 K.sup.41A 6 32 49 10 2 1.70 28 65 7 0.78 90 10 0.1 110 V.sup.42A 26 45 21 8 1.11 59 37 4 0.45 90 10 0.1 63 N.sup.44A 9 37 40 10 4 1.6 31 63 6 0.75 93 7 0.07 107 L.sup.46A 14 42 35 9 1.4 35 62 3 0.68 95 5 0.05 105 Q.sup.47A 6 27 53 9 4 1.78 22 73 5 0.84 95 5 0.05 124 S.sup.48A 6 34 46 9 4 1.72 34 59 7 0.73 93 7 0.07 100 N.sup.50A 4 23 56 13 4 1.89 19 76 4 0.85 95 5 0.05 129 L.sup.73A 12 36 33 13 7 1.69 41 55 3 0.62 87 13 0.13 93 S.sup.74A 11 42 39 7 1.43 35 61 5 0.7 96 4 0.04 103 K.sup.75A 12 38 41 9 1.48 35 62 3 0.68 93 7 0.07 105 D.sup.77A- 3 7 71 15 4 2.1 5 87 8 1.04 95 5 0.05 147 E.sup.79A V.sup.83A- 9 35 39 11 6 1.7 33 61 6 0.73 86 14 0.14 103 WT 2 23 49 22 4 2.03 34 60 5 0.71 81 19 0 E.sup.79A 6 13 51 24 6 2.12 21 73 6 0.86 85 15 0 122 R.sup.103A 4 18 51 20 6 2.06 23 70 7 0.83 82 18 0 117 WT 1 17 51 25 5 2.16 14 77 9 0.96 84 16 0.16 D.sup.77A 4 6 60 23 7 2.23 0 86 13 1.13 92 8 0.08 112 V.sup.42A 13 28 29 21 9 1.86 49 42 9 0.61 86 14 0.14 55 Q.sup.47A 6 7 58 21 9 2.2 8 85 7 0.99 91 9 0.09 110 N.sup.50A 7 8 51 25 9 2.2 10 81 9 0.99 90 10 0.1 105

Example 28 Analysis of CLκ-D.SUP.43.A and CLκ-H.SUP.81.A Mutants

[0441] In order to determine whether the charge, hydrogen bond or the size of CLκ-D.sup.43 are important to the CLκ-K.sup.80 directional conjugation, CLκ-D.sup.43 was mutated to CLκ-D.sup.43E (SEQ ID NO:107), CLκ-D.sup.43N (SEQ ID NO:108) and CLκ-D.sup.43L (SEQ ID NO:109) respectively. The mutants were generated on 2.12.1.fx antibody light chain following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The mutated mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0442] Similarly, in order to assess the role of CLκ-H.sup.81, the following mutant versions of the test antibody 2.12.1.fx were assessed: CLκ-H.sup.81N (SEQ ID NO:110), CLκ-H.sup.81Q (SEQ ID NO:111), CLκ-H.sup.81Y (SEQ ID NO:112), CLκ-H.sup.81W (SEQ ID NO:113) and CLκ-H.sup.81F (SEQ ID NO:114).

[0443] 2.12.1.fx antibody and 2.12.1.fx-[CLκ-mutant]antibodies (1 mg reaction size) were adjusted 18 mg/ml to pH 7.7 with a phosphate buffer to a final concentration of 0.06 M sodium phosphate. [PFP-PEG.sub.5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg/ml. The peptide/linker was added to antibody at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT.

[0444] CLκ-D.sup.43N has the similar overall conjugation and light chain levels (Table 55) compared to the wild type antibody. CLκ-D.sup.43E and CLκ-D.sup.43L showed reduced overall conjugation level light chain conjugation level.

[0445] CLκ-H.sup.81N, CLκ-H.sup.81Q, CLκ-H.sup.81Y, CLκ-H.sup.81W and CLκ-H.sup.81F mutants showed reduced overall conjugation level light chain conjugation level, suggesting that the imidazole ring is required for the PFP directional conjugation. The conjugation reaction does not involve the 7-stacking interaction nor the H-bonds formed with Nε2 or Nδ11 of the imidazole ring.

TABLE-US-00057 TABLE 55 2.12.1.fx variants conjugated to [PEG.sub.5-K.sup.11-SEQ: 27] (WT is 2.12.1.1fx). Ab % CA shows % conjugations additions per antibody, followed by the average CA per antibody. Reduced light and heavy chain analysis also shown, with respective average CA per chain. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 53. Ab % CA Avg LC % CA Avg HC % CA Avg 1LC Ab 0 1 2 3 4 CA 0 1 2 CA 0 1 2 CA WT % WT 2 23 49 22 4 2.03 34 60 5 0.71 81 19 0 0.19 D.sup.43N 7 24 39 21 8 2.00 34 59 7 0.73 84 16 0 0.16 98 D.sup.43L 14 33 34 19 0 1.57 55 40 5 0.50 74 26 0 0.26 67 D.sup.43E 9 33 37 17 4 1.75 44 48 8 0.63 74 26 4 0.34 80 H.sup.81N 32 41 20 7 0 1.04 84 16 0 0.16 70 26 5 0.35 27 H.sup.81Q 29 40 21 11 0 1.14 82 17 1 0.19 73 27 0 0.27 28 H.sup.81Y 27 40 24 8 0 1.14 80 20 0 0.20 70 26 4 0.35 33 H.sup.81W 29 45 19 8 0 1.05 85 15 0 0.15 69 26 5 0.36 25 H.sup.81F 13 41 30 15 0 1.48 79 21 0 0.21 70 30 0 0.30 35

Example 29 2.12.1.Fx-[CLκ-D.SUP.n.A] Conjugation Using Different Reactive Esters

[0446] 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A] were conjugated to [PEG.sub.5-K.sup.11-SEQ:27] using different reactive esters (see Examples 18 and 19) (results shown in Table 56). For all of the different activated esters, the 2.12.1.fx-[CLκ-D.sup.77A] mutant gave a higher level of intact average CA upon conjugation compared to the wt 2.12.1.fx. Another clear trend was that the level of 0 and 1 CA in the wild type 2.12.1.fx was markedly decreased in the 2.12.1.fx-[CLκ-D.sup.77A] mutant for each of the activated esters, and that the level of 2 CA was increased in each case for most activated esters, except for Z9.

[0447] The results of the reduced LC/HC analyses showed a further obvious trend comparing the 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A] conjugation results. In each case, the degree of underivatized LC decreased, substantially in some cases. This was accompanied by a concomitant increase in the level of 1 CA on the LC, again for each different active ester, so that overall the average amount of derivatization on the LC increased. The general trend for the LC was that the amount of 1CA increased by the amount that OCA decreased, as the amount of 2CA present in each case was essentially unchanged.

[0448] In considering the HC, another trend was apparent in that the already low amount of 1CA derivatization for each active ester was further decreased. The outlier in this trend was Z9, the only non-phenolic ester. This ester shows little of the directional conjugation effect towards CLκ-K.sup.80 compared to the other phenolic esters and the levels of both LC and HC derivatization are similar, with only a minor improvement in directionality imparted by the 2.12.1.fx-[CLκ-D.sup.77A] mutant. Overall, the 2.12.1.fx-[CLκ-D.sup.77A] mutant provides clear evidence of improved directional LC conjugation compared to native 2.12.1.fx fora range of activated esters

TABLE-US-00058 TABLE 56 Analysis of 2.12.1.fx.-[CLκ-D.sup.77A] conjugation to [PEG.sub.5-K.sup.11-SEQ: 27] using different reactive esters (see Examples 26 and 27). Ab % CA shows the overall % of conjugation additions per antibody, with reduced light chain and heavy chain analysis also shown (LC % CAN, HC % CA). and Δ indicates the difference between the WT and D.sup.185A mutant results for Ab % CA. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 53. Ab % CA Avg LC % Avg. Avg. 1LC Z# Ab 0 1 2 3 4 CA Δ 0 1 2 CA 0 1 2 CA WT % 1 WT 1 17 51 25 5 2.16 0.07 14 77 9 0.96 84 16 0.16 D.sup.77A 4 6 60 23 7 2.23 0 86 13 1.13 92 8 0.08 112 2 WT 15 43 33 9 1.36 0.72 49 49 3 0.54 84 16 0.16 D.sup.77A 7 14 53 19 8 2.09 16 81 3 0.87 90 10 0.1 33 10 WT 23 44 26 7 1.18 0.73 60 38 2 0.42 81 19 0.19 D.sup.77A 6 27 43 18 6 1.91 29 67 4 0.74 89 11 0.11 176 3 WT 7 32 39 18 5 1.82 0.43 44 48 7 0.63 79 21 0.21 D.sup.77A 6 11 46 25 12 2.25 12 79 9 0.97 90 10 0.1 165 6 WT 15 46 30 9 1.34 0.71 41 52 7 0.66 88 12 0.12 D.sup.77A 6 21 44 20 9 2.05 18 73 8 0.9 93 7 0.07 140 9 WT 7 29 35 21 8 1.94 0.26 57 34 9 0.51 56 34 10 0.54 D.sup.77A 10 19 29 27 15 2.20 47 44 9 0.63 61 32 7 0.46 129

Example 30 Trastuzumab (Herceptin®) Conjugation

[0449] In order to confirm that the improved directional conjugation to CLκ-K.sup.80 caused by CLκ-D.sup.77 mutation can be applied to other antibodies comprising CLκ, D.sup.77A mutation was also inserted to the CLκ of trastuzumab (hTrast). Trastuzumab light chain and heavy chain DNA were synthesized based on the amino acid sequences on Drug Bank, Accession Number DB00072 (BIOD00098, BTD00098).

[0450] hTrast-[CLκ-D.sup.77A] mutant was generated in two steps. First, D.sup.77A mutation was generated on an antibody light chain following protocols described in QuickChange site-directed mutagenesis kit (Stratagene®). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing. The VL of trastuzumab was ligated with the CL of the antibody with D.sup.77A mutation. Primer pair TRAST.VL.FOR (SEQ ID NO:89) and TRAST.VL.REV (SEQ ID NO:90) were used to amplify trastuzumab VL. The PCR fragment was digested with BgIII and BsaI. Primer pair TRAST.CL.D185A.FOR (SEQ ID NO:91) and TRAST.CL.D185.A.REV (SEQ ID NO:92) were used to amplify CL with D.sup.77A mutation. The resulting PCR fragment was digested with BsaI and NheI. Restriction enzyme digested PCR fragments were ligated with a modified p2.12.1.fxP4 plasmid (Invitrogen®) cut with BgIII and NheI. The insert sequence was confirmed by DNA sequencing. The mutated mAb was transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAb was characterized using MS.

[0451] Trastuzumab and hTrast-[CLκ-D.sup.77A] (1 mg reaction size) were adjusted 18 mg/ml to pH 7.7 with a phosphate buffer to a final concentration of 0.06 M sodium phosphate. [PFP-PEG.sub.5-K.sup.11-SEQ:27] was reconstituted in a propylene glycol solution to 10 mg/ml. The peptide/linker was added to antibody at a molar ratio of 4.3:1 and allowed to react for 2 hrs at RT.

[0452] Table 57 compares the conjugation profile of trastuzumab-[PEG.sub.5-K.sup.11-SEQ:27] with hTrast-[CLκ-D.sup.77A]-[PEG.sub.5-K.sup.11-SEQ:27]. The conjugation profile of trastuzumab -[PEG.sub.5-K.sup.11-SEQ:27] occurs as a distribution between 0-4 peptide additions with the average number of peptide additions being 1.75. The profile changes following the D.sup.77A mutation; the average number of peptide additions rises to 2.18 and significantly less overall levels of 0 and 1 peptide addition is observed. This result suggests that the single point mutation CLκ-D.sup.77A has the effect of increasing the overall conjugation to the scaffold, as seen in the test antibody 2.12.1.fx.

[0453] The reduced light and heavy chain analysis demonstrates that the average conjugation is higher on the light chain of hTrast-[CLκ-D.sup.77A] than unmodified trastuzumab; the average light chain conjugate addition value for hTrast-[CLκ-D.sup.77A] is 1.01 compared to 0.70 for trastuzumab. In addition, unconjugated light chain is significantly reduced in hTrast-[CLκ-D.sup.77A]. Conjugation on the heavy chain is observed at a significantly lower level. The majority of observed heavy chain for both trastuzumab and hTrast-[CLκ-D.sup.77A] is unconjugated; this is especially true in the case of hTrast-[CLκ-D.sup.77A] heavy chain. These results suggest that the CLκ-D.sup.77A mutation alters the light chain to make it significantly more susceptible to conjugation.

TABLE-US-00059 TABLE 57 Analysis of conjugation of [PEG.sub.5-K.sup.11-SEQ: 27] to Abs trastuzumab (WT) and hTrast-[CLκ-D.sup.77A]. Rep = replicate. AvReps is the average of the results of the three replicate experiments, with the standard deviation shown beneath (StdDev). Ab % CA shows % conjugations additions per antibody, followed by the average CA per antibody. Reduced light and heavy chain analysis also shown. Ab Ab % CA Avg. LC % CA Avg. HC % CA Avg. WT 0 1 2 3 4 CA 0 1 2 CA 0 1 CA Rep1 6 33 42 15 5 1.80 35 65 ND 0.65 85 15 0.15 Rep2 7 34 39 14 6 1.79 26 74 ND 0.74 84 16 0.16 Rep3 9 38 38 11 5 1.65 30 70 ND 0.7 82 18 0.18 Av. Rep 7 35 40 13 5 1.75 30 70 0.70 84 16 0.16 StdDev 2 3 2 2 1 0.08 5 5 0.05 2 2 0.02 D.sup.77A 5 92 3 0.99 94 6 0.06 Rep1 4 5 66 20 6 2.19 1 94 5 1.04 91 9 0.09 Rep2 3 5 65 21 6 2.22 3 95 2 0.99 94 6 0.06 Rep3 3 8 69 17 4 2.12 3 94 3 1.01 93 7 0.07 AvRep 3 6 67 19 5 2.18 2 2 2 0.03 2 2 0.02 StdDev 1 2 2 2 1 0.05 5 92 3 0.99 94 6 0.06

Example 31 Conjugation of Trastuzumab with MMAD

[0454] Table 58 compares the conjugation profile of trastuzumab -[PEG.sub.5-MMAD] (Auristatin derivative) with hTrast-[CLκ-D.sup.77A]-[PEG.sub.5-MMAD].

##STR00101##

[0455] The conjugation profile of trastuzumab-[PEG.sub.5-MMAD] occurs as a distribution between 0-4 conjugations per antibody with the largest form being 2 conjugations and the average number of conjugations is 1.65. When CLκ-D.sup.77A is mutated, the average number of conjugations rises to 2.00 and significantly less overall levels of 0 and 1 MMAD addition is observed. This result suggests that the single point mutation CLκ-D.sup.77A has the effect of increasing the overall conjugation to the scaffold and that this technology is applicable to an antibody toxin conjugation model.

[0456] Reduced heavy and light chain analysis demonstrates that the average conjugation is higher on the light chain of hTrast-[CLκ-D.sup.77A] than unmodified trastuzumab; the average light chain conjugate addition value for hTrast-[CLκ-D.sup.77A] is 0.88 compared to 0.56 for trastuzumab. In addition, unconjugated light chain is significantly reduced in hTrast-[CLκ-D.sup.77A]. Conjugation on the heavy chain is observed at a significantly lower level. The majority of observed heavy chain for both trastuzumab and hTrast-[CLκ-D.sup.77A] is unconjugated; this is especially true in the case of hTrast-[CLκ-D.sup.77A] heavy chain. These results suggest that the CLκ-D.sup.77A mutation alters the light chain to make it significantly more susceptible to conjugation.

TABLE-US-00060 TABLE 58 Analysis of [PEG.sub.5-MMAD] conjugation to trastuzumab (WT) and hTrast-[CLκ-D.sup.77A], also showing reduced LC and HC analysis. Ab % CA shows % conjugations additions per antibody, followed by the average CA per antibody. Reduced light and heavy chain analysis also shown. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 53. hTrast conjugated Percent CA Avg. LC % CA Avg. HC % CA Avg. 1LC to MMAD 0 1 2 3 4 CA 0 1 2 CA 0 1 CA WT % WT 8 35 42 13 2 1.65 45 54 1 0.56 76 23 0.23 D.sup.77A 1 16 65 16 1 2.00 11 88 0.88 84 14 0.14 163

Example 32 Ability of Conjugated Trastuzumab to Bind Targets

[0457] The ability of trastuzumab and hTrast-[CLκ-D.sup.77A], unconjugated and conjugated to either [PEG.sub.5-K.sup.11-SEQ:27] or [PEG.sub.5-MMAD] and to bind to the Her2 receptor was studied using a Her2 binding ELISA assay. Half well ELISA plates were coated with 1 ug/ml of Fc-ErbB2 fusion protein in PBS and incubated at 4° C. overnight. Plates were washed 3 with KPL wash buffer and subsequently blocked with Superblock for 1 hr at RT. 10× serial dilutions of samples were prepared in Superblock, with a top concentration of 100 μg/ml. Samples were added to wells and plates were incubated for 1 hr at RT. Plates were washed 3× with KPL wash buffer. Bound samples were detected by incubating with a 1:1000 dilution of anti-human Fab-HRP secondary antibody for 1 hr at RT. Plates were again washed 3× with KPL wash buffer and HRP was detected with TMB substrate. The reaction was stopped with 2M H.sub.2SO.sub.4 and OD was measured at 450 nm on a Spectramax plate reader.

[0458] FIGS. 9A and 9B demonstrate that commercial trastuzumab, trastuzumab generated from the available sequence, and hTrast-[CLκ-D.sup.77A], as well as trastuzumab and hTrast-[CLκ-D.sup.77A] when bound to either of [PEG.sub.5-K.sup.11-SEQ:27] or [PEG.sub.5-MMAD] each display similar Her2 binding characteristics. These result suggest that conjugation, primarily at CLκ-K.sup.80, does not significantly interfere with the receptor binding function of the native antibody.

Example 33 Comparison of PFP and NHS Conjugation Strategies

[0459] Trastuzumab was conjugated to [PEG.sub.5-MMAD] using two separate strategies: directional conjugation to CLκ-K.sup.80 using PFP ester (Z1) as the Z* group (generating trastuzumab-[5PEG-MMAD]), or NHS (Z13, generating trastuzumab-[MMAD].sub.n), which resulted in a wider conjugation pattern across the antibody, and dosed to rats to compare the tolerability of the antibody drug conjugates. Both conjugates were given as 10, 30 and 100 mg/kg single bolus doses. All animals dosed at 10, and 30 mg/kg doses of both conjugates during the one week study period survived without significant body weight loss. However, the 100 mg/kg dose group showed a clear difference between random conjugation (Z13) and site selective conjugation to CLκ-K.sup.80 (Z1). Greater than 50% of the animals in 100 mg/kg dose of the random conjugate (NHS conjugation) died within the one week study period while all animals in the 100 mg/kg dose of the site selective conjugate (PFP conjugation) survived without significant body weight loss (Table 59). This may suggest that preferential conjugation at CLκ-K.sup.80 may provide a more reliable mechanism for conjugation of Effector Moieties then traditional ‘random’ approaches, as conjugation on multiple surface lysine residues may give rise to Effector Moieties that have less reliable cleavage and degradation patterns.

TABLE-US-00061 TABLE 59 Site selective conjugation of toxin improves the tolerability of the antibody drug conjugates. Comparison of tolerability of trastuzumab -[PEG.sub.5-MMAD] in rats after conjugation using Z1 and Z13 as Z* groups. Conjugation Dose % Conjugate type (mg/Kg) Survival Trastuzumab-[PEG.sub.5- PFP 10 100 MMAD] 30 100 100 100 Trastuzumab-[MMAD].sub.n NHS 10 100 30 100 100 50

Example 34 h38C2 Conjugated with Toxin and Cleavable Linker

[0460] A targeting peptide was conjugated to the combining site of a CLκ-D.sup.77A mutated version of catalytic antibody h38C2 (HC=SEQ ID NO:65 and LC=SEQ ID NOs:37 and 67) using a linker of the formula P-Q-W as herein described, with a β-lactam group as the W group to form a covalent attachment with the side chain of K.sup.99 of SEQ ID NO:65. This conjugated antibody was then further conjugated with the PFP-activated ester of an exemplary Auristatin-based toxin attached to a valine-citrulline p-aminobenzyl carbamate cleavable linker ([PFP-PEG.sub.2-ValCitABC-TOXIN]).

##STR00102##

[0461] A distribution of conjugates was observed with primarily 2-3 toxins per antibody scaffold (Table 60). In vitro cytotoxicity assays of this conjugate demonstrated potent anti-proliferative effects in AU565 cell lines (IC.sub.50=0.4 nM) and OVCAR5 cell lines (IC.sub.50=0.2 nM).

TABLE-US-00062 TABLE 60 Conjugation profile of h38C2 conjugated to [PFP-PEG.sub.2-ValCitABC-TOXIN]. Conjugation Additions (CA) (%) 0 1 2 3 4 5 Avg CA PFP-PEG.sub.2-ValCitABC-TOXIN 4 8 31 29 18 10 2.8

Example 35

[0462] A structural analog of PFP, with a trifluoromethyl group replacing the para-fluorine atom, was used to make a derivative Z* group; 2,3,5,6-tetrafluoro-4-(trifluoromethyl)phenyl (Z16):

##STR00103##

This was used to generate a Z16-PEG.sub.5 linker, which was conjugated to the test peptide SEQ ID NO:27, and tested in conjugation with the test antibody 2.12.1.fx. In contrast to other Z* groups tested, this derivative gave a conjugate with a slightly higher level of conjugation for the native 2.12.1.fx compared to 2.12.1.fx-[CLκ-D.sup.77A], but both levels of intact conjugation were higher for the conjugate using the Z16 group than the corresponding PFP (Z1) analogs (around 10% increase in overall average CA compared to Z1). Using Z16, both native 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A] conjugations showed an equivalent level of 1CA, and this was lower than that observed for native 2.12.1.fx conjugated with [PFP-PEG.sub.5-K.sup.11-SEQ:27]: for other conjugations, the level of 1CA is typically reduced in the D.sup.77A compared to the corresponding native antibody. Overall, the results suggest that the leaving group D16 is more reactive than Z1 (PFP) (Table 61).

[0463] The Z16 leaving group shows roughly equivalent derivatization for both the native 2.12.1.fx and 2.12.1.fx-[CLκ-D.sup.77A] antibodies and the amount of underivatized LC is small in both cases. Again the overall level of LC and HC derivatization is increased using Z*16 compared to Z1. The leaving group Z16 appears a more reactive ester than PFP, but it is possible that the CF.sub.3 group is providing an additional interaction near the CLκ-K.sup.80 region that is also driving reactivity and preferential derivatization of the LC

TABLE-US-00063 TABLE 61 Analysis of 2.12.1.fx and 2.12.1.fx.-[CLκ-D.sup.77A] conjugation to [PEG.sub.5-K.sup.11-SEQ: 27] using different reactive Z* groups Z1 and Z16. Ab % CA shows the overall % of conjugation additions per antibody, with reduced light chain and heavy chain analysis also shown (LC % CA, HC % CA). Δ indicates the difference between the WT and D.sup.185A mutant results for Ab % CA. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 53. Avg HC % Avg Ab % CA Avg LC % CA CA- CA CA- 1LC Z* 2.12.1.fx 0 1 2 3 4 CA Δ 0 1 2 LC 0 1 HC WT % 1 WT 1 17 51 25 5 2.16 14 77 9 0.96 84 16 0.16 1 D.sup.77A 4 6 60 23 7 2.23 0.07 0 86 13 1.13 92 8 0.08 112 16 WT 4 4 44 36 12 2.48 5 83 12 1.07 81 19 0.19 16 D.sup.77A 9 9 31 33 18 2.43 0.05 1 79 20 1.2 89 11 0.11 95

Example 36 Synthesis of Toxin 0101

Experimental for Toxin 0101 (#54 in the Schematic)

[0464] Preparation of 2-Methylalanyl-N-[(3R,4S,5S)-3-methoxy-1-{(2S)-2-[(1R,2R)-1-methoxy-2-methyl-3-oxo-3-{[(1S)-2-phenyl-1-(1,3-thiazol-2-yl)ethyl]amino}propyl]pyrrolidin-1-yl}-5-methyl-1-oxoheptan-4-yl]-N-methyl-L-valinamide (#54)

##STR00104##

[0465] Step 1. Synthesis of N-[(9H-fluoren-9-ylmethoxy)carbonyl]-2-methylalanyl-N-R3R,4S,5S)-3-methoxy-1-{(2S)-2-[(1R,2R)-1-methoxy-2-methyl-3-oxo-3-{[(1S)-2-phenyl-1-(1,3-thiazol-2-yl)ethyl]amino}propyl]pyrrolidin-1-yl}-5-methyl-1-oxoheptan-4-yl]-N-methyl-L-valinamide (#53). According to general procedure D (below), from #32 (2.05 g, 2.83 mmol, 1 eq.) in dichloromethane (20 mL, 0.1 M) and N,N-dimethylformamide (3 mL), the amine #19 ((2R,3R)-3-methoxy-2-methyl-N-[(1S)-2-phenyl-1-(1,3-thiazol-2-yl)ethyl]-3-[(2S)-pyrrolidin-2-yl]propanamide, trifluoroacetic acid salt) (2.5 g, 3.4 mmol, 1.2 eq.), HATU (1.29 g, 3.38 mmol, 1.2 eq.) and triethylamine (1.57 mL, 11.3 mmol, 4 eq.) was synthesized the crude desired material, which was purified by silica gel chromatography (Gradient: 0% to 55% acetone in heptane), producing #53 (2.42 g, 74%) as a solid. LC-MS: m/z 965.7 [M+H.sup.+], 987.6 [M+Na.sup.+], retention time=1.04 mins; HPLC (Protocol A): m/z 965.4 [M+H.sup.+], retention time=11.344 mins (purity >97%), .sup.1H NMR (400 MHz, DMSO-d.sub.6), presumed to be a mixture of rotamers, characteristic signals: δ 7.86-7.91 (m, 2H), [7.77 (d, J=3.3 Hz) and 7.79 (d, J=3.2 Hz), total 1H], 7.67-7.74 (m, 2H), [7.63 (d, J=3.2 Hz) and 7.65 (d, J=3.2 Hz), total 1H], 7.38-7.44 (m, 2H), 7.30-7.36 (m, 2H), 7.11-7.30 (m, 5H), [5.39 (ddd, J=11.4, 8.4, 4.1 Hz) and 5.52 (ddd, J=11.7, 8.8, 4.2 Hz), total 1H], [4.49 (dd, J=8.6, 7.6 Hz) and 4.59 (dd, J=8.6, 6.8 Hz), total 1H], 3.13, 3.17, 3.18 and 3.24 (4 s, total 6H), 2.90 and 3.00 (2 br s, total 3H), 1.31 and 1.36 (2 br s, total 6H), [1.05 (d, J=6.7 Hz) and 1.09 (d, J=6.7 Hz), total 3H]. Step 2. Synthesis of 2-methylalanyl-N-[(3R,4S,5S)-3-methoxy-1-{(2S)-2-[(1R,2R)-1-methoxy-2-methyl-3-oxo-3-{[(1S)-2-phenyl-1-(1,3-thiazol-2-yl)ethyl]amino}propyl]pyrrolidin-1-yl}-5-methyl-1-oxoheptan-4-yl]-N-methyl-L-valinamide (#54)

[0466] According to general procedure A (below), from #53 (701 mg, 0.726 mmol) in dichloromethane (10 mL, 0.07 M) was synthesized the crude desired material, which was purified by silica gel chromatography (Gradient: 0% to 10% methanol in dichloromethane). The residue was diluted with diethyl ether and heptane and was concentrated in vacuo to afford #54 (406 mg, 75%) as a white solid. LC-MS: m/z 743.6 [M+H.sup.+], retention time=0.70 minutes; HPLC (Protocol A): m/z 743.4 [M+H.sup.+], retention time=6.903 minutes, (purity >97%); .sup.1H NMR (400 MHz, DMSO-d.sub.6), presumed to be a mixture of rotamers, characteristic signals: δ [8.64 (br d, J=8.5 Hz) and 8.86 (br d, J=8.7 Hz), total 1H], [8.04 (br d, J=9.3 Hz) and 8.08 (br d, J=9.3 Hz), total 1H], [7.77 (d, J=3.3 Hz) and 7.80 (d, J=3.2 Hz), total 1H], [7.63 (d, J=3.3 Hz) and 7.66 (d, J=3.2 Hz), total 1H], 7.13-7.31 (m, 5H), [5.39 (ddd, J=11, 8.5, 4 Hz) and 5.53 (ddd, J=12, 9, 4 Hz), total 1H], [4.49 (dd, J=9, 8 Hz) and 4.60 (dd, J=9, 7 Hz), total 1H], 3.16, 3.20, 3.21 and 3.25 (4 s, total 6H), 2.93 and 3.02 (2 br s, total 3H), 1.21 (s, 3H), 1.13 and 1.13 (2 s, total 3H), [1.05 (d, J=6.7 Hz) and 1.10 (d, J=6.7 Hz), total 3H], 0.73-0.80 (m, 3H).

[0467] General Procedure A: Fmoc removal using diethylamine. To a solution of the Fmoc-containing compound in dichloromethane was added an equal volume of diethylamine. Reaction progress was monitored by LC-MS (or HPLC or TLC); the reaction was usually completed within three hours. Solvents were removed in vacuo, and the residue was azeotroped three times with heptane. The residue was then diluted with dichloromethane and a small amount of methanol before being reduced down onto silica and purified by chromatography on silica gel, eluting with methanol in dichloromethane to afford the desired material.

[0468] General Procedure D: coupling with O-(7-azabenzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HATU). To a stirring solution of the amine (1 eq.) and acid (1.1 eq.) in dichloromethane: N,N-dimethylformamide (4:1, 0.3 M in amine) was added HATU (1.2 eq.) followed by Et.sub.3N (3 eq.). Reaction progress was monitored by LC-MS (or HPLC or TLC), the reaction was usually completed within three hours. Solvents were removed in vacuo. The residue was azeotroped three times with heptane and was diluted with a small amount of ethyl acetate before being reduced down onto silica and purified by silica gel or reverse phase chromatography.

Example 37 Preparation of MAC Comprising mAb Hu08 and Toxin 0101

[0469] Hu08 is a human anti-IL-13Rα2 antibody, and is described fully in U.S. 61/723,545, whose contents are herein incorporated by reference.

A mutant version of hu08, comprising the CLκ-D.sup.77A mutation was generated according to standard protocols (hu08-[CLκ-D77A]). Toxin-0101 (#54: Example 36) was conjugated with a cleavable linker to form the structure:

##STR00105##

and then the toxin-linker was conjugated to hu08-[CLκ-D77A] according to the techniques described herein. The compound hu08-vc-0101 was generated, resulting in non-specific conjugation of multiples of Toxin-0101 on the antibody hu08 via the linker-Toxin-0101 species:

##STR00106##

In Vitro Cytotoxicity Assay

[0470] Cell lines expressing the IL-13Rα2 antigen and a negative control cell line were cultured with increasing concentrations of hu08-[CLκ-D.sup.77A]. After four days, viability of cultures were assessed. IC.sub.50 values were calculated by logistic non-linear regression and are presented as ng Ab/mL.

[0471] The data demonstrate that hu08-vc0101 and hu08-[CLκ40, -D.sup.77A] were both effective against both of the IL-13Rα2 positive cell lines tested (PC3MM2 and A375), having an IC.sub.50 ranging from 2.5 to 7.9 ng Ab/mL (Table 62). Neither hu08-vc0101 nor hu08-[CLκ-D.sup.77A] were active against the IL-13Rα2 negative cell line, H460, and the non-IL-13Rα2 binding control, hIgG8.84-vc0101, was not active against any of the cell lines tested.

TABLE-US-00064 TABLE 62 In vitro cytotoxicity assay of hu08-[CLκ-D.sup.77A] -0101. IC.sub.50 (ng Ab/mL) ADC Toxin: Ab PC3MM2 A375 H460 hu08-vc-0101 3.2 2.5 3.8 >400000 hu08-[CLκ-D.sup.77A] -0101 1.9 4.9 7.9 >400000 hIgG8.8-vc-0101 3.7 >400000 >400000 >400000

Subcutaneous Xenograft Models of Cys Mutant ADCs

[0472] Female, athymic (nude) mice were injected s.c. with PC3MM2 tumor cells. Mice with staged tumors, approximately 0.1 to 0.3 g (n=8 to 10 mice/treatment group) were administered intravenously q4d×4 with normal saline (vehicle) or MAC-0001. Compounds were dosed based on Ab content. Tumors were measured at least once a week and their size (mm2+/−□SEM) is calculated as mm.sup.3=0.5×(tumor width.sup.2)×(tumor length). The data in Table 63 indicate that hu08-[CLκ-D.sup.77A]-0101 inhibits the growth of PC3MM2 xenografts.

TABLE-US-00065 TABLE 63 Dose PC3MM2 xenograft, tumor volume (mm.sup.3 ± SEM) single Day Day Day Day Day Day Day Day Day Day ADC dose 0 5 8 12 15 20 30 41 55 77 Vehicle 0 325 ± 590 ± 782 ± 1140 ± GT GT GT GT GT GT 9 41 79 142 hu08-[CLκ- 1.5 328 ± 393 ± 352 ± 432 ± 556 ± 732 ± GT GT GT GT D.sup.77A]-0101 49 96 106 124 236 305 hu08-vc- 1.5 333 ± 431 ± 281 ± 299 ± 362 ± 450 ± 956 ± GT GT GT 0101 12 40 25 32 47 58 166 GT = group terminated due to large tumor size

Example 38 Trastuzumab MMAD Conjugate Activities

[0473] Three trastuzumab conjugates were made: trastuzumab-[5PEG-MMAD], hTrast-[CLκ-D.sup.77A]-[5PEG-MMAD], and trastuzumab-(MMAD).sub.n, where MMAD was connected to a 5PEG linker with Z13 (NHS) as leaving group, and conjugated to trastuzumab without directional conjugation techniques, resulting in non-specific conjugation of MMAD to trastuzumab surface lysines (see Examples 30-33)

##STR00107##

[0474] The three mAb conjugates were evaluated in an exploratory toxicity study in rats in which animals received single intravenous bolus doses of each ADC at 0 (vehicle), 10, 30, and 100 mg/kg (5 male rats/group) and were then observed for 14 days. Toxicology evaluation included daily clinical observations, weekly body weight measurements and clinical pathology evaluation on the day of necropsy. Animals were euthanized on day 15 and selected tissues were collected for microscopic examination. In addition, blood samples were collected from all animals at approximately 0.0833 (5 min), 6, 24 (day 2), 48 (day 3), 72 (day 4), 96 (day 5), 168 (day 8), and 312 (day 14) hrs post-dose and analyzed for antibody-conjugate and trastuzumab antibody concentrations.

Results

[0475] The plasma exposures (based on AUC) were overall similar for all 3 conjugates at any given dose. The AUC(0-312) of hTrast-[CLκ-D.sup.77A]-[5PEG-MMAD], trastuzumab-[5PEG-MMAD] and trastuzumab-(MMAD).sub.n at 100 mg/kg were 177000, 174000 and 139000 ng.Math.h/mL, respectively. hTrast-[CLκ-D.sup.77A]-[5PEG-MMAD] and trastuzumab-[5PEG-MMAD] were clinically well tolerated at all doses. However, trastuzumab-(MMAD).sub.n administration at 100 mg/kg was associated with marked clinical signs and premature euthanasia of ⅕ rats on Day 8. Other rats from this group had decreased skin turgor and decreased body weight gain.

[0476] Clinical pathology changes were overall similar with hTrast-[CLκ-D.sup.77A]-[5PEG-MMAD] and trastuzumab-[5PEG-MMAD] and included in particular mild decreases in red blood cell (RBC) mass (RBC count, hemoglobin and/or hematocrit) at 0 or ≥30 mg/kg and minimal increases in aspartate aminotransferase (AST) at 100 mg/kg. The RBC mass changes were more pronounced with trastuzumab-(MMAD).sub.n and were associated with decreased erythroid cellularity in the bone marrow. Other noteworthy trastuzumab-(MMAD).sub.n-related clinical pathology changes at 100 mg/kg included moderate decreases in platelet counts and mild increases in ALT, AST, ALP and total bilirubin.

[0477] Microscopic findings were overall similar for hTrast-[CLκ-D.sup.77A]-[5PEG-MMAD] and trastuzumab-[5PEG-MMAD] and included alveolar histiocytosis/inflammation in the lung, degeneration of small bile ducts in the liver and increased tingible body macrophages (containing cell debris) in the bone marrow. Increased mitoses in several tissues and single cell necrosis in the cornea were considered pharmacologically mediated effects of tubulin inhibition resulting in mitotic arrest and apoptosis. By contrast, trastuzumab-(MMAD).sub.n administration was associated with more pronounced microscopic tissue alterations, which included tubular degeneration/necrosis and glomerulopathy in the kidney; single cell necrosis, bile duct degeneration and hyperplasia and centrilobular necrosis/fibrosis in the liver; alveolar histiocytosis/inflammation in the lung; increased tangible body macrophages, degeneration/decreased numbers of hematopoietic cells and osteolysis in the bone marrow; decreased marginal zone cellularity in the spleen. Of special note, centrilobular fibrosis in the liver in association with disruption of the normal lobular architecture was consistent with a reparative change suggesting earlier, more extensive treatment-related hepatocellular damage. In addition, pharmacologically mediated increased mitoses and/or single cell necrosis were observed in several tissues.

[0478] In summary, hTrast-[CLκ-D.sup.77A]-[5PEG-MMAD] and trastuzumab-[5PEG-MMAD] were well tolerated at all doses (10, 30, and 100 mg/kg) and demonstrated overall similar toxicity profiles. Trastuzumab-(MMAD).sub.n administration led to premature mortality at 100 mg/kg and was associated with significant target organ toxicities in the liver, kidney, lung and bone marrow in particular.

Example 39 Modeling of CLκ and CLλ

[0479] FIG. 10A depicts an Ig fold of a constant light domain containing a 3-stranded β-sheet packed against a 4-stranded β-sheet. The fold is stabilized by hydrogen bonding between the β-strands of each β-sheet, by hydrophobic bonding between residues of opposite β-sheets in the interior, and by a disulfide bond between the β-sheets. The 3-stranded β-sheet comprises β-strands C, F, and G, and the 4-stranded β-sheet has β-strands A, B, E, and D. The letters A through G denote the sequential positions of the β-strands along the amino acid sequence of the Ig fold. Linking each β-strand with the subsequent β-strand is an amino acid connecting chain that may or may not comprise a turn (A/B) or α-helix (E/F) (FIG. 10B).

[0480] FIG. 8B plots the secondary structures along the primary sequence of the mouse and human CLκ, and the human CLλ. The EF connecting chain between β-strands E and F of the CLκ and CLλ have identical secondary structure and consist of a 5-6 residue α-helical region (CLκ-K.sup.75-K.sup.80, and CLλ-P.sup.78-S.sup.81), followed by a 2-3 amino-acid turn (CLκ-H.sup.81-K.sup.82 and CLλ-H.sup.82-S.sup.84) (FIGS. 8B and 10). The CD connecting chain brings the side chain of structurally equivalent aspartic acids (CLκ-D.sup.53, CLλ-D.sup.45) to the vicinity of the EF chain (approximately 3.5 Å), and allows CLκ-D.sup.53 interact with the imidazole ring of CLκ-H.sup.81.

[0481] Modeling of the CLλ and comparison with the CLκ suggests that CLλ-D.sup.45 cooperates with the imidazole ring of CLλ-H.sup.82 in the same manner, and that mutating CLλ-S.sup.81 to CLλ-K.sup.81 recreates a similar local environment as found on the CLκ, and so allows directional conjugation to the CLλ.

[0482] Accordingly, the present invention also provides for CLλ domains comprising one of the following mutations: CLλ-S.sup.81K, and CLλ-K.sup.80x/S.sup.81 K, wherein x is any amino acid except P, K, R or H, wherein the numbering is according to SEQ ID NO:93. In some aspects, the invention provides for novel CLλ domains comprising K.sup.81, or x.sup.80/K.sup.81, wherein x is one of G, A, I, L, V, S, T, M, N, Q, F, Y, W, D, or E. In some aspects, the invention also provides for a CLλ domain comprising a sequence selected from the group consisting of SEQ ID NO:94, and SEQ ID NO:95.

[0483] Modeling also suggests that CLλ-E.sup.77 would be available to form a salt bridge with either of CLλ-K.sup.80 or CLλ-K.sup.81 in the same manner as CLκ-D.sup.77 appears to form a salt bridge to CLκ-K.sup.80. Thus, mutating CLλ-E.sup.77 to any of R, L, S, G, Q, P, N, V, I, T, and M is likely to facilitate directional conjugation at either of CLλ-K.sup.80CLλ when S.sup.81Δ (i.e. deletion of S.sup.81), or CLλ-K.sup.81. Accordingly, the present invention also provides for a CLλ domain comprising a sequence selected from the group consisting of SEQ ID NO:96 and SEQ ID NO:97.

Example 40 Modeling of CLκ-K.SUP.80 .Conjugation Mechanism

Structure and Sequence Description

[0484] Crystal structures of the Fab domain of 2.12.1.fx and h38C2-[CLκ-D.sup.77A] were used to model the specificity of CLκ-K.sup.80 reactivity towards a halo-phenol (such as PFP)/ester-mediated conjugation. Experience with conjugations across multiple and varied antibodies, as well as early modeling analysis, indicated that modeling only needed be focused on the CLκ, as the remainder of the antibody appeared to exert very little, if any, influence on the mechanism of conjugation.

Computational Approach

[0485] The goal of the computational calculations was to clarify the critical attributes of CLκ-K.sup.80 and interacting residues that preferentially bias this site towards the PFP-ester conjugation reaction. After the 3-D coordinates of CLκ and CLκ-D.sup.77A were selected from the crystal structure (based on those coordinates that contained those residues already identified as being relevant to directional conjugation), the coordinates were subjected to standard computational protocols, such as protein preparation, attachment of hydrogen atoms, and force field parameters assignment.

[0486] Hydrogen atoms were assigned to all amino acid atoms on the respective 3D CLκ domain according to the calculated pKa values and pI (protein ionization potential). pKa is the value of a protonation state of a given titratable amino acid at the neutral pH (negative logarithm of hydrogen concentration) taking in to account the influence of amino acids in the protein chain (Spassov; A fast and accurate computational approach to protein ionization. Protein Science 2008, 17, 1955-1970). The results of these calculations are shown in Table 64.

TABLE-US-00066 TABLE 64 pKa values of amino acid residues in CLκ Titratable residues Calculated pKa Standard pKa CLκ-D.sup.14 3.315 3.65 CLκ-D.sup.43 3.324 3.65 CLκ-D.sup.59 3.667 3.65 CLκ-D.sup.62 4.411 3.65 CLκ-D.sup.77 3.307 3.65 CLκ-E.sup.15 3.93 4.25 CLκ-E.sup.35 3.906 4.25 CLκ-E.sup.53 4.604 4.25 CLκ-E.sup.57 4.284 4.25 CLκ-E.sup.79 3.636 4.25 CLκ-E.sup.89 4.244 4.25 CLκ-Y.sup.32 10.276 10.07 CLκ-Y.sup.65 11.847 10.07 CLκ-Y.sup.78 11.844 10.07 CLκ-Y.sup.84 14 10.07 CLκ-H.sup.81 7.22 6.0 CLκ-H.sup.90 6.126 6.0 CLκ-K.sup.18 11.057 10.53 CLκ-K.sup.37 10.761 10.53 CLκ-K.sup.40 10.505 10.53 CLκ-K.sup.61 10.852 10.53 CLκ-K.sup.75 10.952 10.53 CLκ-K.sup.80 11.119 10.53 CLκ-K.sup.82 10.451 10.53 CLκ-K.sup.99 10.706 10.53 CLκ-R.sup.34 13.428 12.48 CLκ-R.sup.103 14 12.48

[0487] Hydrogen atoms are added accordingly to the pKa value for all titratable amino acids and for the remaining amino acids, according to the atomic valence number. The calculations indicated that CLκ-H.sup.81 is unprotonated at pH 7.0-7.4 (physiological pH). This observation is consistent with proposed mechanism for catalytic reaction, where CLκ-H.sup.81 acts as a nucleophilic catalyst. In general, conjugation would be expected to decrease with as the pH goes below about pH 6.5-7.0, as CLκ-H.sup.81 would be protonated. As the pH increases above about 7.4, a greater overall level of conjugation would be predicted, as other residues (especially lysines) become more reactive and contribute to the overall conjugation reaction, consequently we would expect that the directional effect would be increasingly lost as the pH rises. This correlates with what was observed in the pH study shown in Example 4, Table 4.

Minimization

[0488] CHARMm [Chemistry at HARvard Macromolecular Mechanics] is an energy minimization technique, and was used to bring the 3-D structures to the equilibrium position and find the best geometrical position for its atomic structure. CHARMm was used at the first step with the SMART Minimizer with 1000 steps of Steepest Descent minimization with a RMS gradient tolerance of 3[Kcal/(mol*Å)], followed by Conjugate Gradient minimization with a RMS gradient of 0.01. For the energy change, a tolerance of 3[Kcal/(mol*Å)] was applied to the average gradient during a cycle of minimization. The Steepest Descent method takes the molecule to the nearest minimum and the Conjugated Gradient improves the final conformation obtained. Momany Rone charges were used, (as described in Momany & Rone; Validation of the general purpose QUANTA 3.2/CHARMm force field. Comp. Chem. 1992, 13, 888-900.) The minimized structure of CLκ differed by 1Å when compared to the un-minimized structure.

Complex Between CLκ Domains and PFP-PEG.SUB.2

[0489] The respective complexes between a PFP ester and the CLκ and CLκ-D.sup.77A domains was built in silico to better understand each interaction. Structural analysis of the 3D CLκ domains by the Accelrys protocol Define and Edit Binding Site (Discovery Studio software version DS3.5) revealed a region that could be termed a ‘binding pocket’, located between the CD and EF connecting chains, and underneath CLκ-K.sup.80, as shown in FIG. 11. This pocket is facilitated by the amino acids located at CLκ-K.sup.75-E.sup.79 (EF α-helix on the EF connecting chain), CLκ-K.sup.80-K.sup.82 (EF loop on the EF connecting chain) and CLκ-V.sup.42 (CD loop on the CD connecting chain) giving the shape and the electrostatic properties required by the catalytic reaction.

[0490] To measure the pocket size, the Define and Edit Binding Site protocol was used on the previously minimized CLκ structure. This protocol requires defining the CLκ domain as a receptor to use on it the space filling method to calculate the cavity size and assess its suitability for binding of a molecule with a particular size, such as PFP. The binding site of a receptor can be represented in many ways, for example a sphere or a list of residues surrounding this sphere. To define a binding site, the receptor is first mapped to a grid. Grid points within a given distance of the receptor atoms are marked as occupied by the receptor, and thus undesirable as locations for ligand atoms. Two methods exist to identify a binding site. The first uses an “eraser” algorithm to identify sites based on the shape of the receptor. The second uses the volume occupied by a known ligand already positioned in an active site (Venkatachalam et al Flt: a novel method for the shape-directed rapid docking of ligands to protein active sites. J. Mol. Graph. Model 2003, 21, 289-307).

[0491] As a result of protocol execution, binding sites are identified as a set of points located on a grid that encompasses the molecule under consideration. This definition permits measurement of the size and shape of the binding site, which allows for its qualification as a binding site. PFP-PEG2 was docked manually to the site surrounded by amino acids experimentally identified as important for the catalytic reaction. These amino acids were CLκ-K.sup.80, CLκ-H.sup.81, CLκ-D.sup.43 and other amino acids within 10 Å distance from CLκ-K.sup.80 (shown in Table 54).

[0492] The initial complex between each CLκ domain and a PFP ester was minimized using the QM/MM hybrid method, applying the CHARMm force field for the CLκ domain and QM calculations for the PFP ester. The PFP ester was minimized prior to the placement using first CHARMm then the QM/MM approach in which the PFP ester is treated as a QM system. The structure was minimized using the CHARMm minimization protocol described above.

[0493] For the minimizations of the complex, a QM/MM hybrid protocol was used in which quantum mechanical (QM) calculations gave information about the electron densities change upon interaction with WT or mutated protein allowing us to capture the influence of the surrounding environment on the PFP ester electron densities and its susceptibility to the conjugation reaction. The QM/MM protocol is a hybrid method where the molecular system is divided into two regions: first, the central region PFP ester, to be treated by a QM calculation and second, the outer region CLκ domain, treated by molecular mechanics (MM) methods.

[0494] QM treatment brings a higher level of theory, enabling the modeling of additional phenomena compared to traditional forcefield techniques, for example, where a chemical reaction occurs or polarization effects play an important role. The remaining bulk of the structure is described using a forcefield CLκ. The way in which the two regions are allowed to interact and how the total QM/MM energy is evaluated define the specific QM/MM protocol employed. Energy calculated in this method is composed from three basic parts: the QM energy of the PFP ester (E.sub.QM), CD-kappa (E.sub.MM), and the interaction energy between these two systems (EQ.sub.MMM). The formula E.sub.tot=E.sub.QM+E.sub.MM+E.sub.QM/MM is used.

[0495] As far as the coulomb interaction between the electronic density of the QM region and the forcefield point charges is concerned, QM/MM methods in this application always employ electronic embedding. This means that the forcefield atom's partial charges enter the QM calculation as an external potential, thereby polarizing the QM electronic density relative to a QM calculation in vacuo, and giving rise to an electrostatic interaction energy between the QM density and the point charges. Contrary to the handling of the electrostatic interactions in QM/MM methods, the van der Waals interactions are treated entirely at the classical level. This means that appropriate forcefield parameters must be determined for all atoms in the simulation. The van der Waals QM/MM interaction energy (and forces) is a part of the CHARMm simulation server energy and is listed as a separate term in the output file.

Calculations

[0496] The total effects observed from in vitro experiments during and after the conjugation can be modeled by computational techniques using stepwise approximations to describe such phenomenon as:

[0497] 1. PFP binding pocket size—binding of the PFP ester and direction.

[0498] 2. Protein stability—especially, the stability of a PFP binding pocket.

[0499] 3. Tautomerization of the imidazole ring of CLκ-H.sup.81.

[0500] 4. Directional PFP placement.

[0501] 5. Initial interaction between the PFP ester and CLκ.

[0502] 6. Reactivity of each of the catalytic amino acids.

Modeling Results

[0503] Identification of the binding pocket in the CLκ domain permitted modeling and docking-placement of the PFP ester. After the PFP ester was modeled as docked into the respective CLκ binding site, the 3-D structure of the complex was minimized using a hybrid method QM/MM. In this approach, the PFP ester was defined as a QM system and the CLκ domain was treated by a MM method. After minimization of the complex, a network of intermolecular interactions between the CLκ and the PFP ester was revealed. This network was formed by intermolecular hydrogen bonds, hydrophobic interactions, and 7-electron stacking. All of these forces together appear to be involved in formation of a network responsible for the directional placement and conjugation of the PFP ester to CLκ-K.sup.80. The minimized structure of the complex indicates that CLκ-H.sup.81 is an important catalytic amino acid.

[0504] It is known that the imidazole moiety in histidine sidechains can serve as a catalytic amino acid in enzymatic reactions, with the imidazole acting as a nucleophile and forming an acyl imidazole during the catalytic transition state [J Phys Chem B. 2011 Oct. 20; 115(41):11895-901. Epub 2011 Sep. 23]. It is also known that the imidazole ring of histidine can undergo tautomerization, depending on whether the Nδ or Nε atom bonds with H:

##STR00108##

[0505] Spatial positioning of CLκ-H.sup.81 in the PFP binding pocket points to the requirement of Nδ tautomeric form of imidazole ring CLκ-H.sup.81 for nucleophilic attack on carbonyl carbon of the PFP ester. It also indicates a requirement for the electron lone pair at Nε to be in the same plane as the carbonyl carbon of PFP ester group. This is possible when the imidazole ring of CLκ-H.sup.81 is in the Nδ tautomeric form (see Scheme II). The tautomeric equilibrium can be controlled by the hydrogen bond interactions with the neighboring hydrogen acceptor amino acids. There are two aspartic acid residues, CLκ-D.sup.77 and CLκ-D.sup.43, in the vicinity of CLκ-H.sup.81 and both of them appear to control the tautomerization state of the CLκ-H.sup.81 imidazole ring.

##STR00109##

[0506] FIG. 12 depicts the 3-D arrangement of CLκ-H.sup.81, CLκ-D.sup.43 and CLκ-D.sup.77, accompanied by the distances between the oxygen atoms of the aspartic acids and nitrogen atoms of the imidazole ring. Hydrogen bonding between CLκ-D.sup.43 and CLκ-H.sup.81 promote and stabilize the catalytically-active tautomeric form NO, while the hydrogen bond with CLκ-D.sup.77 will promote the catalytically-inactive form of CLκ-H.sup.81 NE. In addition to the hydrogen bonds formed between CLκ-H.sup.81 Nδ and CLκ-D.sup.77 Oδ (negatively charged), CLκ-H.sup.81 Nε and CLκ-D.sup.77 Oδ, CLκ-H.sup.81 Nδ can form a hydrogen bond with the carbonyl oxygen of the CLκ-D.sup.77 backbone. This last interaction stabilizes the catalytically-active conformation of CLκ-K.sup.81 NO, rendering it capable of catalysis. It is known from the mutational analysis and computational work discussed herein that the optimal conformation for CLκ-K.sup.81 reactivity involves hydrogen bonding with CLκ-D.sup.43 via Nδ, as this stabilizes an active conformation capable of the hydrogen Hε transfer from CLκ-K.sup.80 Nε. The two other CLκ-K.sup.81 tautomeric forms will therefore most likely be less active or inactive in the catalytic reaction.

Modeling the Effect of CLκ-D.SUP.77.A Mutation

[0507] As predicted from the WT model, in the [CLκ-D.sup.77A] domain, CLκ-H.sup.81 forms a hydrogen bond with CLκ-D.sup.43 via the imidazole Nδ atom, exposing electron pair at NE for the nucleophilic attack on carbonyl carbon of the ester group in the PFP-PEG.sub.2 molecule (Scheme III). FIGS. 13A, 13B and 13C depict a comparison of crystal structure modeled CLκ and CLκ-D.sup.77A domains, and illustrate the shift in spatial position of CLκ-H.sup.81 in the CLκ-D.sup.77A mutant.

##STR00110##

Mutations to CLκ-D.SUP.77

[0508] Modeling suggests that CLκ-D.sup.77 mutations would have a significant impact on the conjugation rate of CLκ-K.sup.80. A series of QM/MM calculations were conducted on the CLκ domain complexed with PFP, where CLκ-D.sup.77 was mutated to all other natural amino acids (CLκ-D.sup.77C was excluded).

[0509] Modeling and experimental analysis suggests that CLκ-D.sup.77 may bind to PFP and change the unfavorable electron density on the ester atoms of PFP moiety involved in the catalytic reaction. The CLκ-D.sup.77A mutation would then improve the conjugation rate by allowing better distribution of the electron densities on the reacting atoms of PFP ester.

[0510] Analysis of the calculation results performed on the mutants of CLκ-D.sup.77 indicate that the best performing mutants (in terms of conjugation efficiency) are likely to be those small hydrophobic amino acids unable to form hydrogen bonds to the PFP moiety. Modeling data therefore suggests the mutations CLκ-D.sup.77G, CLκ-D.sup.77P, CLκ-D.sup.77M, CLκ-D.sup.77L, CLκ-D.sup.77I, CLκ-D.sup.77A, and CLκ-D.sup.77V would improve directional conjugation. As seen in Example 26, this is borne out by experimental testing. All CLκ .sup.77 mutants with hydrophobic amino acids have no impact on the tautomeric equilibrium of CLκ-H.sup.81, thus CLκ-H.sup.81 remains in the Nδ tautomeric form and consequently conjugation rates improve.

[0511] Modeling also suggest that most hydrophilic amino acids are likely to result in higher protein stability than CLκ-D.sup.77 according to by QM/MM calculations; and points to the mutations CLκ-D.sup.77S and CLκ-D.sup.77T in particular, and also CLκ-D.sup.77Q, CLκ-D.sup.77N, CLκ-D.sup.77H and CLκ-D.sup.77R as potentially improving directional conjugation. As seen in Example 26, this is also borne out by experimental testing.

[0512] In addition to the interaction energies and the modifications of the charge distribution, the strongest interaction and the lower conjugation rates were observed for CLκ-D.sup.77F and CLκ-D.sup.77Y mutants. The model indicates that the side chains of these aromatic amino acids indtude directly into the binding pocket, and are likely to be involved in 7-electron stacking interactions with the PFP group and therefore change the directional placement of PFP moiety. In addition, aromatic amino acids can sterically hinder PFP, preventing a similar binding mode to that possible in the WT protein thus adversely impacting the directional placement of the PFP moiety and resulting in a decreased conjugation rate. Interestingly, the model suggests that the side chain of CLκ-D.sup.77W occupies a slightly different spatial position to the aromatic side chains of CLκ-D.sup.77F and CLκ-D.sup.77Y, with the CLκ-D.sup.77W indole group sitting outside the binding pocket and not predicted to interfere with the spatial access of a halo-phenyl ester to the binding pocket. The model also predicts that a conservative mutation, such as CLκ-D.sup.77E is likely to have a similar effect on the tautomeric form of CLκ-H.sup.81, but due to the larger side-chain, create additional steric interference in the binding pocket. Mutations to CLκ-D.sup.77C or CLκ-D.sup.77K are also predicted to interfere with the reactivity of the site.

[0513] CLκ-D.sup.77 favors the presence of the hydrogen atom on Nε of the imidazole ring, while CLκ-D.sup.77A favors maintaining the hydrogen at the position Nδ. Spatial distribution of amino acids in the site indicates the important function Nδ tautomeric form of CLκ-H.sup.81 in the catalytic reaction enabling nucleophilic attack by Nε at the carbonyl carbon of the ester group, further facilitating directional conjugation.

Example 41 Predictive Modeling Based on Model of CLκ and CLκ-D.SUP.77.A

[0514] The impact of mutations on protein stability and interaction strength with the PFP ester were analyzed to identify the most relevant and reactive mutants of CLκ. For this computational experiment, key amino acids within a 10A distance from Cα carbon of CLκ-K.sup.80 were selected and analysed for (except CLκ-K.sup.80 and CLκ-H.sup.81, given the unique requirement for each of the residues).

[0515] For many or most of the mutations modeled for PFP conjugation with the CLκ or CLκ-D.sup.77A, it will be understood that in certain applications, it may appropriate to substitute or retain a residue that would reduce PFP conjugation in the CLκ or CLκ-D.sup.77A, as it may be desirable to reduce or increase the pocket size; such as where larger or smaller halo-phenols are used as Z1 groups, or where the precise geometry of a specific immunoglobulin domain appears to merit such a feature.

CLκ-K.SUP.75

[0516] This residue does not appear to have any direct impact on the conjugation reactivity of CLκ-K.sup.80 in the native or CLκ-D.sup.77A mutant by QM/MM calculations.

CLκ-A.SUP.76

[0517] CLκ-A.sup.76 is located at the beginning of the α-helix, and lies and above the plane that contains the hydrogen-bonded carboxyl group of CLκ-D.sup.77 and the hydroxyl group of CLκ-S.sup.74. This location allows amino acids with a large sidechain to interact with CLκ-D.sup.77 and they may have a positive impact of the conjugation reaction, especially those with hydrophilic groups capable of hydrogen bond formation. Most amino acid substitutions at this residue would be expected to have little effect on conjugation, specifically CLκ-S.sup.74A, CLκ-S.sup.74D, CLκ-S.sup.74E, CLκ-S.sup.74I, CLκ-S.sup.74L, CLκ-CLκ-S.sup.74F, CLκ-S.sup.74W and CLκ-S.sup.74V. Other residues that could provide hydrogen bonding opportunities would be expected to enhance conjugation, namely CLκ-S.sup.74R, CLκ-S.sup.74N, CLκ-S.sup.74Q, CLκ-S.sup.74H, and CLκ-S.sup.74K, and to a lesser degree CLκ-S.sup.74S, CLκ-S.sup.74T and CLκ-S.sup.74Y. While residues that disrupt the α-helix may have a somewhat negative effect on directional conjugation to CLκ-K.sup.80, such as CLκ-S.sup.74G and CLκ-S.sup.74P, directional conjugation is unlikely to be abrogated, merely reduced. Introduction of a cysteine would present a risk of the potential to form aggregates in expression.

CLκ-Y.SUP.78

[0518] CLκ-Y.sup.78 is located on the α-helix, facing the opposite direction to the binding pocket. CLκ-Y.sup.78 makes a number of hydrophobic interactions with surrounding amino acids and supports the CLκ structure. Accordingly, smaller sidechains (Ala, Ser, Thr and Val) or those that affect the stability or formation of the α-helix (Gly, Pro) will be more likely to adversely affect the conjugation reaction to CLκ-K.sup.80, although as these mutations are not predicted to directly interfere with the CLκ-K.sup.80 reactivity, such mutations may not necessarily abrogate directional conjugation. Additionally, CLκ-Y.sup.78 appear to interact with the sidechain of CLκ-R.sup.103, thus hydrophobic or negatively-charged sidechains would be expected to favor conjugation to facilitate this interaction (Asn, Asp, Gln, Glu, Phe and Trp). Other sidechains would not be expected to effect the conjugation reaction (Arg, His, Ile, Leu, Lys and Met).

CLκ-E.SUP.79

[0519] Modeling suggests that the side chain of CLκ-E.sup.79 is pointing in the opposite direction from the binding site. In addition, CLκ-E.sup.79 appears to form a salt bridge with sidechain of CLκ-K.sup.75. Based on the modelling, it is postulated that most amino acid substitutions in this position would have little effect on the conjugation reaction (Asn, Asp, Gln, His, Met, Phe, Ser, Thr, Tyr and Trp). Small hydrophilic or charged residues would likely favour the conjugation reaction (Ala, Arg, Ile, Leu, Lys and Val), while those that affect the stability or formation of the α-helix (Gly, Pro) may adversely affect the conjugation reaction to CLκ-K.sup.80, without necessarily abrogating directional conjugation.

CLκ-V.SUP.83

[0520] This residue does not appear to have a direct impact on the conjugation reaction with CLκ-K.sup.80 by experimental data and QM/MM calculations.

TABLE-US-00067 TABLE 65 Summary of effect of mutations on certain CLκ residues on directional conjugation to CLκ-K.sup.80. A.sup.76 D.sup.77 Y.sup.78 E.sup.79 G 0 +++ −− − P 0 +++ −− − A WT ++ − + V 0 + − + L 0 ++ 0 + I 0 ++ 0 + M 0 +++ 0 0 F 0 −−− + 0 W 0 + + 0 Y + −−− WT 0 T + +++ − 0 S + +++ − 0 N ++ ++ + 0 Q ++ ++ + 0 D 0 WT + 0 E 0 −− + WT H ++ ++ 0 0 R ++ ++ 0 + K ++ −− 0 +

Additional Modeling

CLκ-V.SUP.42

[0521] CLκ-V.sup.42 is positioned at the end of β-strand C and the beginning of the CD loop. From a careful examination of the structure, it was observed that CLκ-V.sup.42 is located underneath the imidazole ring of CLκ-H.sup.81 (FIGS. 14 and 15). To help determine the nature of the interactions between CLκ-H.sup.81 and CLκ-V.sup.42, the x, y,z coordinates of the centroid for the imidazole ring were calculated, followed by measurement of the distance between the calculated centroid and the hydrogen atom Hδ of the CLκ-V.sup.42. This distance of 2.9 Å indicates direct interactions between the π-electrons on the imidazole ring and the hydrogen Hδ of CLκ-V.sup.42. This interaction therefore appears to strongly influence the optimal positioning of CLκ-H.sup.81 in the pocket and for the tautomeric equilibrium shift to the Nδ tautomer. Thus, while not essential, the presence of CLκ-V.sup.42 exerts a positive influence on directional conjugation. These analyses are borne out by experimental data, where it was found that CLκ-V.sup.42A mutation caused a decrease in the conjugation rate (Table 54). Modeling suggests that CLκ-V.sup.421 is likely to be able to assist in positioning the CLκ-H.sup.81 imidazole ring to favour directional conjugation, although the slightly larger side chain will reduce the overall binding pocket size. In many circumstances, this reduction of pocket size may not have an appreciable effect on the directional conjugation mechanics. CLκ-V.sup.42L may also be able to assist with positioning the CLκ-H.sup.81 imidazole ring.

CLκ-D.SUP.43

[0522] CLκ-D.sup.43 is located on the CD loop. In native CLκ, CLκ-D.sup.43 appears to interact with the backbone of CLκ-H.sup.81 and CLκ-K.sup.82, contributing to protein stability. No hydrogen bonds between Hδ of CLκ-H.sup.81 and the carboxylic group of CLκ-D.sup.43 were identified on the crystal structure of CLκ. This may suggest that CLκ-D.sup.43 exerts only a minimal influence on the tautomeric equilibrium of the Nδ catalytically active form of CLκ-H.sup.81. However, experimental analysis with 2.12.1.fx-[CLκ-D.sup.43A] mutants showed that mutating this residue had a significant inhibitory effect on directional conjugation. Taken together with the overall model, it is likely that CLκ-H.sup.81 alternates between the catalytically active and inactive form, forming H-bonds between CLκ-D.sup.43 and CLκ-D.sup.77 as it does so, and that removing the CLκ-D.sup.43 residue eliminates one of the forces pushing CLκ-H.sup.81 towards the active Nδ tautomer.

[0523] Analyses of the CLκ-D.sup.77A crystal structure revealed a well-defined hydrogen bond between hydrogen Nδ and the CLκ-D.sup.43 carboxylic group. This suggests that one of the most significant effects of the CLκ-D.sup.77A mutant is in stabilization of CLκ-H.sup.81 catalytically favorable tautomeric form NO, as shown in FIGS. 13 and 14. Finally, proximity of CLκ-D.sup.43 to CLκ-H.sup.81 would allow CLκ-D.sup.43 to participate in the catalytic reaction as a hydrogen atom recipient in the last step of the putative reaction mechanism (Scheme III). Thus the role of CLκ-D.sup.43 in the native CLκ and CLκ-D.sup.77A mutant, appears to be in stabilization of catalytically active Nδ tautomeric form of CLκ-H.sup.81 and participation in the catalytic reaction as the hydrogen atom recipient.

[0524] The modeling analyses strongly suggests that binding of the PFP molecule in native CLκ is controlled by the tautomeric forms of CLκ-H.sup.81, and the affinity of this binding is in turn controlled by the state of the tautomeric equilibrium constant. The presence of Nδ tautomeric form allows the lone pair on Nε to nucleophilic attack the PFP carbonyl (Scheme III).

[0525] CLκ-D.sup.43 is involved in the catalytic reaction by being a hydrogen ion acceptor (Scheme III). CLκ-D.sup.43 was mutated in silico to all 19 amino acids and in the final series of calculated mutants, the WT protein (comprising CLκ-D.sup.43) creates the highest interaction energy with the CLκ-PFP ester complex, with CLκ-D.sup.43N also predicted to be of similar chain length and be able to form a hydrogen bond to the CLκ-H.sup.81 NO, thus aiding the catalytically-active tautomeric form Nδ of CLκ-H.sup.81. Other residues likely to be acceptable substitutions are CLκ-D.sup.43E, CLκ-D.sup.43Q and CLκ-D.sup.43S, these being able to form the desired H-bond with CLκ-H.sup.81, but likely to have chain lengths either slightly too large or small to be of optimal size for CLκ conjugation with PFP, but may be better suited to either optimizing a non-CLκ immunoglobulin domain for directional conjugation with PFP, or optimizing an Ig domain for conjugation with a different halo-phenyl ester.

CLκ-N.SUP.44

[0526] CLκ-N.sup.44 is located on the CD loop. Its sidechain is pointing outwards, away from the PFP binding pocket, and appears not to have any role in, nor influence on, the conjugation reaction. QM/MM calculations predicted that polar amino acids and those with a negatively-charged sidechain may enhance protein stability due to the putative interactions with CLκ-K.sup.41, also located outside of the pocket.

CLκ-L.SUP.46

[0527] A CLκ-L.sup.46A mutation had a relatively neutral impact on the conjugation rate, based on experimental data, as well as from our QM/MM calculations. CLκ-L.sup.46A is located outside of the PFP binding pocket, but it may facilitate the pocket shape due to the size of its sidechain. Being located outside of the pocket, it can interact with amino acids located outside of the pocket as well. Mutations to large and hydrophilic sidechains may increase protein stability due to the interaction with CLκ-Q.sup.39 and support the shape of the PFP pocket according to QM/MM calculations.

CLκ-Q.SUP.47

[0528] A CLκ-Q.sup.47A mutation improved the conjugation rate. This mutation is likely exerting an effect by increasing the size of the PFP binding site, thus impacting complex formation and protein stability in a positive way according to modeling calculations. Amino acids with larger sidechains will likely impact conjugation in a slightly negative way, according to QM/MM calculations. Thus, mutations to CLκ-Q.sup.47A, CLκ-Q.sup.47G, CLκ-Q.sup.47V, CLκ-Q.sup.471, CLκ-Q.sup.47L, CLκ-Q.sup.47T, CLκ-Q.sup.47S, CLκ-Q.sup.47N, CLκ-Q.sup.47D, CLκ-Q.sup.47H, CLκ-Q.sup.47P, or CLκ-Q.sup.47E will likely be beneficial or neutral; whereas mutations to CLκ-Q.sup.47W, CLκ-Q.sup.47F, CLκ-Q.sup.47Y, or CLκ-Q.sup.47K may have a somewhat negative impact on PFP conjugation to a CLκ domain. As before, it will be understood that in certain applications, it may be desirable to reduce the pocket size; such as where smaller halo-phenols are used as Z1 groups, or where the precise geometry of a specific immunoglobulin domain appears to merit such a feature.

CLκ-S.SUP.48

[0529] Mutation of CLκ-S.sup.48A improved the conjugation rates according to experimental data: modeling suggest that the reason for this is most likely due to the change of electrostatic properties of the pocket. Hydrophobic amino acids at this position will therefore likely have a positive impact on the directional conjugation. Consequently, the following mutations may be especially favoured: CLκ-S.sup.48A, CLκ-S.sup.48G, CLκ-S.sup.48V, CLκ-S.sup.48I, CLκ-S.sup.48L, CLκ-S.sup.48P, and CLκ-S.sup.48M. Other mutations are likely to be tolerated.

Example 42 Double and Triple Mutations of CLκ-D.SUP.77.A

[0530] Antibody 2.12.1.fx was used to test the effect of further mutations to the CLκ region. As before, numbering of residues is according to their location within the CLκ (SEQ ID NO:6, for example). CLκ-D.sup.43 and CLκ-H.sup.81 were each mutated to Ala in on 2.12.1.fx-[CLκ-D.sup.77A] antibody to understand the conjugation mechanisms. Mutations were generated on 2.12.1.fx light chain following protocols described in QuickChange site-directed mutagenesis kit (Stratagene). The antibodies 2.12.1.fx, 2.12.1.fx-[CLκ-D.sup.77A] (CLκ comprising SEQ ID NO:37), 2.12.1.fx-[CLκ-D.sup.43A] (CLκ comprising SEQ ID NO:15), 2.12.1.fx, 2.12.1.fx-[CLκ-D.sup.43A/D.sup.77A] (CLκ comprising SEQ ID NO:127), 2.12.1.fx-[CLκ-D.sup.77A/H.sup.81A] (CLκ comprising SEQ ID NO:128), and 2.12.1.fx-[CLκ-D.sup.43A/D.sup.77A/H.sup.81A] (CLκ comprising SEQ ID NO:129), were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0531] The expressed antibodies were buffer exchanged to 20 mM sodium acetate, 200 mM trehalose pH 5.5 at 20 mg/ml. The antibody solutions were then spiked with 60 mM sodium phosphate pH 7.7. [PEG.sub.5-K.sup.11-SEQ:27] (ABP) was resuspended with 50% propylene glycol and mixed with the protein at a 4.3:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple ABP conjugation sites are observed as distinct signals separated by the mass difference of an ABP. Relative quantitation of multiple ABP conjugation species is performed by measuring the signal magnitude. Results of % CA are shown in Table 66.

TABLE-US-00068 TABLE 66 Intact conjugation analysis of 2.12.1.fx-[CLκ-mutants] with ABP. CLκ % CA Avg SEQ ID 0 1 2 3 4 CA 2.12.1.fx 6 1 15 53 26 5 2.2 D.sup.77A 37 2 3 56 32 7 2.39 D.sup.43A 15 17 38 31 14 0 1.41 D.sup.43A/D.sup.77A 127 4 21 46 21 7 2.06 D.sup.77A/H.sup.81A 128 33 38 20 10 0 1.07 D.sup.43A/D.sup.77A/H.sup.81A 129 25 40 28 7 0 1.18

[0532] The extent of ABP conjugation was examined separately on the light and heavy chains of 2.12.1.fx and 2.12.1.fx-[CLκ mutants]. MACs were denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each (Table 67).

TABLE-US-00069 TABLE 67 Reduced heavy, light chain conjugation analysis of 2.12 1.fx-[CLκ- mutants] with ABP. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 67. CLκ Avg Avg SEQ LC % CA CA- HC % CA CA- 1LC ID 0 1 2 LC 0 1 HC WT % 2.12.1.fx 6 23 69 8 0.85 86 14 0.14 D.sup.77A 37 1 92 7 1.06 97 3 0.03 133 D.sup.43A 15 68 30 1 0.33 79 21 0.21 43 D.sup.43A/D.sup.77A 127 24 64 12 0.88 90 10 0.1 93 D.sup.77A/H.sup.81A 128 72 28 0 0.28 85 15 0.15 41 D.sup.43A/D.sup.77A/H.sup.81A 129 76 20 4 0.28 87 13 0.13 29

[0533] These results are consistent with the His tautomer hypothesis. In the CLκ-D.sup.77A mutant, CLκ-D.sup.43 stabilizes the catalytically active tautomer of CLκ-H.sup.81, which in turn allows CLκ-K.sup.80 to be more receptive and reactive to a PFP-ester. Conversely, in the CLκ-D.sup.43A mutant, CLκ-D.sup.77 stabilizes the inactive tautomer of CLκ-H.sup.81, therefore leading to a reduction in the directional conjugation observed at CLκ-K.sup.80. In the double mutant CLκ-D.sup.43A/D.sup.77A, there are no interactions between CLκ-H.sup.81 and either of CLκ-D.sup.43 or CLκ-D.sup.77, and accordingly, the double mutant acts more like the WT CLκ.

Example 43 Rabbit CLκ Analysis

[0534] Rabbit antibody light chain kappa region (rCLκ) has the same 3D structure as that of other immunoglobulins. rCLκ has Asp at position 151 (kabat number), Ser at position 188, and His at 189, (rCLκ-D.sup.43, rCLκ-S.sup.80, rCLκ-H.sup.81). It was postulated that a rCLκ-S.sup.80K mutant may create the reaction site for PFP directional conjugation. To validate this hypothesis, two trastuzumab rabbit chimera antibodies were constructed. mAb “rTrast” (rabbit trastuzumab) comprises the VL and VH domains of trastuzumab (SEQ ID NOs:75 and 72 respectively) fused to the rCLκ and rabbit constant heavy chain (rCH) (SEQ ID NOs:130 and 131 respectively), to generate the full length rTrast-LC (SEQ ID NO:132) and rTrast-HC (SEQ ID NO:133).

[0535] rTrast-[rCLκ-S.sup.80K] comprises the VL and VH domains of trastuzumab (SEQ ID NOs:75 and 72 respectively) fused to the rCLκ-S.sup.80K (SEQ ID NO:134) and rabbit constant heavy chain (rCH) (SEQ ID NO:131), to generate the full length rTrast-LC-[rCLκ-S.sup.80K] (SEQ ID NO:135) and rTrast-HC (SEQ ID NO:133).

[0536] Rabbit IgG heavy chain and kappa1 light chain were PCRed from plasmids pFUSE-CHIg-rG and pFUSE2ss-CLIg-rk1 (Invivogen) respectively with ends overlapping with trastuzumab variable domains and vector. Trastuzumab VH and VL were PCRed from synthetic genes with ends overlapping with vector and rabbit constant domains. PCRs were mixed with a modified pCEP4 vector (Invitrogen) cut with BgIII and NheI following protocols described in Quick PCR Cloning Kit (BPS Bioscience). Insert DNA were confirmed by DNA sequencing. rCLκ-S.sup.80K mutation was generated following protocols described in Quick PCR Cloning Kit (BPS Bioscience). Mutation was introduced by oligonucleotide primers and cloned to a modified pCEP4 vector (Invitrogen) cut with BgIII and NheI. Insert DNA were confirmed by DNA sequencing.

[0537] The chimeric mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS. Each of the antibodies was buffer exchanged to 20 mM sodium acetate, 200 mM trehalose pH 5.5 at 20 mg/ml, and then spiked with 60 mM sodium phosphate pH 7.7. [PEG.sub.5-K.sup.11-SEQ:27] (ABP) was resuspended with 50% propylene glycol and mixed with the protein at a 4.3:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple ABP conjugation sites are observed as distinct signals separated by the mass difference of an ABP. Relative quantitation of multiple ABP conjugation species is performed by measuring the signal magnitude.

[0538] The extent of ABP conjugation was examined on the light of antibodies. MACs were denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each (Table 68).

TABLE-US-00070 TABLE 68 Intact conjugation and reduced light chain analysis of rabbit chimera conjugated with ABP. % CA Avg LC % CA Avg 1LC 0 1 2 3 CA 0 1 2 CA-LC WT % rTrast 58 24 19 0 0.61 56 38 6 0.5 rTrast-[rCLκ-S.sup.80K] 10 25 54 11 1.657 18 73 8 0.89 192

[0539] The rCLκ-S.sup.80K mutation significantly increased total conjugation (0.61 to 1.657) and 1CA on the light chain (38% to 73%). This result suggests that a directional conjugation site can be created on CLκ chains from species other than human, as long as the critical residues, CLκ-K.sup.80H.sup.81, are present on the immunoglobulin. The % of 1-LC % relative to the respective WT run is shown in the right column, as described in Table 68.

Example 44 Lambda Chain

[0540] As demonstrated in Example 10 above, hCLλ does not demonstrate directional conjugation with PFP esters. hCLλ shares sequence identity with hCLκ at hCLκ-D.sup.43 (CLλ-D.sup.48) and hCLκ-H.sup.81 (CLλ-H.sup.82), and has serine in place of hCLκ-K.sup.80 (CLλ-S.sup.81). It was postulated that a CLλ-S.sup.81K mutant may enable PFP conjugation to at the CLλ-S.sup.81K residue.

[0541] Comparing the crystal structures of CLκ and CLλ (FIGS. 17 & 18), CLκ-V.sup.42 is located at the bottom of the putative PFP binding pocket, whereas the corresponding residue in CLλ is CLλ-A.sup.44. As discussed above, CLκ-V.sup.42A mutations have a negative impact on the directional conjugation, likely owing to the ability of CLκ-V.sup.42 to stabilize the orientation of the imidazole ring of CLκ-H.sup.81. Modeling suggested that a CLλ-A.sup.44V mutation should exert a similar effect on CLλ-H.sup.82.

[0542] As in Example 10, and Examples 15-17, a monoclonal anti-human IL22 antibody (hIL22) was used as an exemplary CLλ comprising antibody. hIL22 comprised SEQ ID NOs:136 and 137 (hIL22-LC and hIL22-HC respectively), and variable light chain comprising SEQ ID NO:138 (hIL22-CLλ-VL).

[0543] A number of mutant versions of hIL22 were generated, to assess the effects of various CLλ mutations on directional conjugation. All hIL22 mutant antibodies comprised SEQ ID NO:137 (hIL22-HC), and SEQ ID NO:138 (hIL22-CLλ-VL).

[0544] hIL22-[LKJ] comprised the A/K swap as described in Example 15, and comprised a CL of SEQ ID NO:61. hIL22-[CLλ-S.sup.81K] comprised the single residue swap of CLλ-S.sup.81K, and comprised SEQ ID NO:140. hIL22-[CLλ-Q.sup.78A/S.sup.81K] comprised a double mutation in the loop, and comprised SEQ ID NO:141. hIL22-[CLλ-A.sup.44V/S.sup.81K] comprised the CLλ-S.sup.81K residue swap, and also a CLλ-A44V mutation at the bottom of the “binding pocket”, and comprised SEQ ID NO:142. hIL22-[CLλ-A.sup.44V/Q.sup.78A/S.sup.81K] comprised both loop mutations CLλ-Q.sup.78A and CLλ-S.sup.81 K, as well as the “binding pocket” CLλ-A.sup.44V mutation, and comprised SEQ ID NO:143. All the point mutations were generated on hIL22-LC following protocols described in QuickChange site-directed mutagenesis kit (Stratagene). Mutations were introduced by oligonucleotide primers and confirmed by DNA sequencing.

[0545] hIL22-[CLλ-λ.sup.76-84/145] comprised inserting SEQ ID NO:145 in place of the CLλ E-F loop, located from CLλ-P.sup.76 inclusive through to CLλ-S.sup.83 inclusive. SEQ ID NO:145 comprises the sequence KAAYEKHKV, which corresponds to the [CLκ-D.sup.77A] E-F loop (i.e. between β-strands E and F) from [CLκ-D.sup.77A]K.sup.75 inclusive through to [CLκ-D.sup.77A]K.sup.82 inclusive.

[0546] hIL22-[CLλ-λ.sup.76-84/145] was generated by overlap PCR. Mutations were introduced by oligonucleotide primers. Primer specific to the 5′ end of hIL22-LC paired with a reverse primer encoding SEQ ID NO:145, and a forward primer encoding SEQ ID NO:145, paired with the reverse primer specific to the 3′ end of 1L22-LC were used to PCR amplify DNA fragments carrying CLκ E-F loop using 1L22-LC as template. These two PCR products were mixed as templates; 1L22-LC forward primer and reverse primer were used in overlap PCR reaction to amplify the full length 1L22-LC DNA with SEQ ID NO:145. The PCR was then digested with restriction enzyme BgIII and NheI. The digested PCR was ligated with a modified pCEP4 plasmid (Invitrogen) cut with BgIII and NheI.

[0547] The mutated mAbs were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0548] Each of the antibodies was buffer exchanged to 20 mM sodium acetate, 200 mM trehalose pH 5.5 at 20 mg/mi, and then spiked with 60 mM sodium phosphate pH 7.7. [PEG.sub.5-K.sup.11-SEQ:27] (ABP) was resuspended with 50% propylene glycol and mixed with the protein at a 4.3:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/mi and the extent of ABP conjugation was examined separately on the light and heavy chains of antibodies. The MACs were denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each (Table 69).

TABLE-US-00071 TABLE 69 Reduced heavy, light chain conjugation analysis of lambda antibody and mutants. CL LC % CA HC % CA SEQ Ave Ave Ab conjugated with ABP ID 0 1 2 LC 0 1 2 HC 2.12.1.fx 6 15 79 6 0.91 93 7 0 0.07 2.12.1.fx-CLκ-D.sup.77A 37 0 92 8 1.08 98 2 0 0.02 Trastuzumab-WT 76 32 66 2 0.7 83 17 0 0.17 Trastuzumab-CLκ-D.sup.77A 77 0 95 5 1.05 90 10 0 0.1 hIL22 136 100 0 0 0 76 22 2 0.26 hIL22-[LKJ] 61 55 45 0 0.45 79 21 0 0.21 hIL22-[CLλ-S.sup.81K] 140 69 31 0 0.31 80 20 0 0.2 hIL22-[CLλ-Q.sup.78A/S.sup.81K] 141 89 11 0 0.11 85 15 0 0.15 hIL22-[CLλ-A.sup.44V/S.sup.81K] 142 48 52 0 0.52 87 13 0 0.13 hIL22-[CLλ-A.sup.44V/Q.sup.78A/S.sup.81K] 143 30 70 0 0.7 82 18 0 0.18 hIL22-[CLλ-A.sup.76-84/145] 144 83 17 0 0.34 82 18 0 0.2

[0549] A single point mutation CLλ-S.sup.81K enabled PFP conjugation to the CLλ. The LC % 1CA increased from 0% to 31% on the CLλ-S.sup.81K mutant compared against the unmutated hIL22, with more conjugation additions seemingly occurring at the newly created conjugation site than at any other single site on the antibody. Peptide mapping studies confirmed that the conjugation events were occurring on CLλ-S.sup.81K (Table 70).

TABLE-US-00072 TABLE 70 Overview of peptide mapping studies on hIL22-[CLλ-S.sup.81K]-CL. Average Δ (% conju- gation, Protein Chain Peptide n = 3) hIL22-[CLλ-S.sup.81K]-CL SEQ ID NO: 140, residues 80-85 54 ± 10 hIL22-[CLλ-S.sup.81K]-HC SEQ ID NO: 137, residues 96-103 21 ± 10 hIL22-[CLλ-S.sup.81K]-HC SEQ ID NO: 137, residues 95-103 19 ± 10

[0550] Referring back to Table 69, combining mutations at CLλ-S.sup.81K and CLλ-A.sup.44V (hIL22-[CLλ-A.sup.44V/S.sup.81K]) further increased light chain 1CA to 52%.

[0551] Interestingly, while the CLλ-Q.sup.78A/CLλ-S.sup.81K double mutation did improve conjugation of the CLλ-S.sup.81K compared to WT CLλ (1CA from 0% to 11%), the improvement was less pronounced than that seen in the single CLλ-S.sup.81K mutant. This correlates well with the model: the effect of mutating CLκ-D.sup.77 is to remove the hydrogen bond between CLκ-D.sup.77 and CLκ-H.sup.81, enabling CLκ-H.sup.81 to revert to the catalytically tautomeric Nδ form. The corresponding position to CLκ-D.sup.77 is CLλ-Q.sup.78, which both modeling and mutational analysis suggest would not have a limiting effect on CLλ-H.sup.82. Taken together, this suggests that the diminishing effect of CLλ-Q.sup.78A on the directional conjugation is likely caused by an alteration to the size and shape of the binding pocket.

[0552] Most surprising, however, was the result of the triple mutation CLλ-A.sup.44V/Q78A/S.sup.81K. Directional conjugation as measured by the 1CA % increased to 70%; reaching levels typically seen in native CLκ domains.

[0553] These results and analysis were borne out by the similar level of directional conjugation seen in the loop swap (hIL22-[CLλ-λ.sup.76-84/145]).

[0554] Overall, these data suggested that the CLλ-Q.sup.78A mutation does improve directionality of CLλ-S.sup.81K conjugation, provided the size and shape of the binding pocket is adapted for the specific Z group used. Moreover, the results of CLλ mutants suggested that directional conjugation sites can be created on immunoglobulins other than CLκ, provided that the motif KH is present in the correct 3D location. Naturally, the binding pocket must be of suitable size and shape to accommodate the specific halo-phenyl ester employed, and, as demonstrated herein, additional features, such as the presence or absence of residues corresponding to CLκ-V.sup.42, CLκ-D.sup.43, and CLκ-D.sup.77 can have significant effects on the rate and optimization of directional conjugation of the KH motif relative to the specific immunoglobulin domain and halo-phenyl ester. While PFP was used in these examples, it will be apparent that other Z groups with acceptable levels of directional conjugation may be selected, and using rational modeling techniques, a balance may be obtained between the desired Z group, size of the binding pocket and the specific mutations required to maintain an active binding pocket.

Example 45 Recreating PFP Conjugation Sites on CH Domains

[0555] The CH domains of antibodies also comprise immunoglobulin structures. Prior to modeling the domains as described in Examples 40 and 41, it was postulated that moving the conjugation motif to the EF loop portion of the EF connecting chain of other CH domains may permit directional conjugation. A sequence alignment of the CHγ1 (SEQ ID NO:147), CHγ2 (SEQ ID NO:155), CHγ3 (SEQ ID NO:158), CLκ (SEQ ID NO:6) and CLλ (SEQ ID NO:57) domains is shown in FIG. 16.

[0556] Two mutant versions were made on the CH1 domain.

[0557] In hCHγ1-m1, the sequence LGTQT (SEQ ID NO:152), which corresponds to residues L.sup.76-T.sup.80 of SEQ ID NO:147, was removed, and replaced by EKHKV (SEQ ID NO:153), which corresponds to E.sup.79-V.sup.83 of CLκ. The resultant mutant, hCHγ1-m1, comprised SEQ ID NO:148.

[0558] In CHγ1-m2, the sequence LGTQT (SEQ ID NO:152), which corresponds to residues L.sup.76-T.sup.80 of SEQ ID NO:147, was removed, and replaced by YEKHKV (SEQ ID NO:154), which corresponds to Y.sup.78-V.sup.83 of CLκ-. The resultant mutant, hCHγ1-m2, comprised SEQ ID NO:150. The additional Y residue was incorporated to allow the hCHγ1-m2 sequence to better align with CLκ sequence.

[0559] Sequence alignment indicated that hCHγ1 lacks an Asp residue corresponding to CLκ-D.sup.43. Accordingly, two additional mutants were generated; where each of hCHγ1-m1 and hCHγ1-m2 were subjected to an additional insertional mutation of an Asp residue between CHγ1-S.sup.43 and CHγ1-G.sup.44, creating the two new mutants of hCHγ1-ml-D.sup.44 (SEQ ID NO:149) and hCHγ1-m2-D.sup.44 (SEQ ID NO:151).

[0560] A mutant version of hCHγ2 (SEQ ID NO:155) was generated, where residues N.sup.85G.sup.86 of SEQ ID NO:155 were substituted with KH to generate hCHγ2-m (SEQ ID NO:156). Sequence alignment suggested that hCHγ2 (SEQ ID NO:155) comprised an Asp residue (D.sup.50) at a location that may correspond to CLκ-D.sup.43.

[0561] A mutant version of hCHγ3 (SEQ ID NO:157) was generated, where residues Q.sup.79G.sup.80 of SEQ ID NO:157 were substituted with KH to generate hCHγ3-m (SEQ ID NO:158). Sequence alignment suggested that hCHγ3 (SEQ ID NO:155) comprised an Asn residue (N.sup.44) at a location corresponding to CLκ-D.sup.43.

[0562] Trastuzumab was used as a model Ab in the study. All the hTrast-CHγ mutants were expressed with a [CLκ-K.sup.80A] mutation, so that conjugation events would preferentially occur on the test CHγ domain. The hTrast-LC-[CLκ-K.sup.80A] (SEQ ID NO:146) mutation was generated following protocols described in QuickChange site-directed mutagenesis kit (Stratagene).

[0563] The mutations on CHγ domains were generated using overlap PCR. Mutations were introduced by oligonucleotide primers. Primer specific to the 5′ end of trastuzumab HC paired with a reverse primer carrying the desired mutation, and a forward primer carrying the desired mutation paired with the reverse primer specific to the 3′ end of trastuzumab HC were used to PCR amplify DNA fragments using trastuzumab HC as template. These two PCR products were mixed as templates; trastuzumab heavy chain forward primer and reverse primer were used in overlap PCR reaction to amplify the full length trastuzumab HC DNA with desired mutations. The PCR was then digested with restriction enzyme BgIII and NheI. The digested PCR was ligated with a modified pCEP4 plasmid (Invitrogen) cut with BgIII and NheI.

[0564] The trastuzumab antibody carrying mutations were transiently expressed in HEK 293 cells, and purified using protein A affinity column. The purified mAbs were characterized using MS.

[0565] The expressed antibody was buffer exchanged to 20 mM sodium acetate, 200 mM trehalose pH 5.5 at 20 mg/ml. The proteins were then spiked with 60 mM sodium phosphate pH 7.7. ABP was resuspended with 50% propylene glycol and mixed with the protein at a 4.3:1 molar ratio and allowed to react overnight at room temperature. All samples were diluted to 2 mg/ml and analyzed as an intact conjugated protein by size exclusion chromatography-mass spectrometry (SEC-MS) to determine the number and quantitation of conjugate forms of the protein. This technique measures the molecular weight of each protein form; multiple ABP conjugation sites are observed as distinct signals separated by the mass difference of an ABP. Relative quantitation of multiple ABP conjugation species is performed by measuring the signal magnitude.

TABLE-US-00073 TABLE 71 Intact conjugation analysis of hTrast antibody and CHγ domain mutants. % CA Ab + ABP 0 1 2 3 Avg CA hTrast 1 16 56 24 2.112 hTrast-[CLκ-D77A] 0 1 62 34 2.338 hTrast-[CLκ-K80A] 48 41 10 2 0.665 hTrast-[CLκ-K.sup.80A]/CHγ1-m1 23 48 24 5 1.107 hTrast-[CLκ-K.sup.80A]/CHγ1-m2 24 48 24 4 1.087 hTrast-[CLκ-K.sup.80A]/CHγ1-m1-D.sup.44 33 49 17 1 0.854 hTrast-[CLκ-K.sup.80A]/CHγ1-m2-D.sup.44 41 45 12 2 0.746 hTrast-[CLκ-K.sup.80A]/CHγ2m 45 44 9 3 0.694 hTrast-[CLκ-K.sup.80A]/CHγ3m 40 44 13 3 0.78

[0566] The extent of ABP conjugation was examined separately on the light and heavy chains of trastuzumab and trastuzumab mutants. MAC product was denatured and disulfide bonds were reduced using guanidine hydrochloride and dithiothreitol. The resulting free light and heavy chains were analyzed using LCMS to determine the conjugation profile on each chain (Table 72).

TABLE-US-00074 TABLE 72 Reduced heavy, light chain conjugation analysis of hTrast antibody and CHγ domain mutants. The % of 1-HC % relative to the respective WT run is shown in the right column, as described in Table 53, although in this example, hTrast-[CLκ-K.sup.80A HC 1CA was taken as the WT FIGURE. LC % CA HC % CA Avg Avg CA- CA- 1HC 0 1 2 LC 0 1 2 HC WT % hTrast 32 66 2 0.7 83 17 0 0.17 N/A hTrast-[CLκ-D.sup.77A] 0 95 5 1.05 90 10 0 0.1 N/A hTrast-[CLκ-K.sup.80A] 98 2 0 0.02 79 21 0 0.21 hTrast-[CLκ-K.sup.80A]/CHγ1-m1 99 1 0 0.01 64 31 5 0.41 148 hTrast-[CLκ-K.sup.80A]/CHγ1-m2 98 2 0 0.02 62 32 6 0.44 152 hTrast-[CLκ-K.sup.80A]/CHγ1-m1-D.sup.44 97 3 0 0.03 74 24 2 0.28 114 hTrast-[CLκ-K.sup.80A]/CHγ1-m2-D.sup.44 97 3 0 0.03 73 24 3 0.3 114 hTrast-[CLκ-K.sup.80A]/CHγ2m 99 1 0 0.01 75 23 2 0.27 110 hTrast-[CLκ-K.sup.80A]/CHγ3m 100 0 0 0 70 27 3 0.33 129

[0567] The total conjugation on hTrast-[CLκ-K.sup.80A]/CHγ1-ml and hTrast-[CLκ-K.sup.80A]/CHγ1-m2 were increased compared to hTrast-[CLκ-K.sup.80A], from 0.66 CA to ˜1CA; and the HC 1CA of these two mutants increased from 21% to 31%, and improvement of about 150% This supports the hypothesis that directional conjugation can be introduced to immunoglobulin domains other than CLκ and CLλ by the introduction of the KH motif.

[0568] The total conjugation on hTrast-[CLκ-K.sup.80A]/CHγ2m also increased, but to a lesser extent than that of the CHγ1 mutants. A comparison of the sequences of the CHγ1 and CHγ2 sequences shows that at the residue corresponding to CLκ-D.sup.77, CHγ1 comprises Ser, and CHγ2 comprises Asp. This could suggest that a mutation similar to the CLκ-D.sup.77 mutations in CHγ2 domains could improve the extent of conjugation.

[0569] In addition, the improvement in conjugation on the CHγ3 HC, while evident, was also modest in comparison to the CHγ1 domain mutants. The CHγ3 domain sequence appeared to comprise Arg at a position corresponding to CLκ-D77, and Ser at a position corresponding to CLκ-D.sup.43.

[0570] Furthermore, the conjugation of the CA % HC for both hTrast-[CLκ-K.sup.80A]/CHγ1-ml-D.sup.44 and hTrast-[CLκ-K.sup.80A]/CHγ1-m2-D.sup.44 showed a very surprising result, in that the apparent increase in conjugation seen in both hTrast-[CLκ-K.sup.80A]/CHγ1-ml and hTrast-[CLκ-K.sup.80A]/CHγ1-m2 went from approximately 150% of WT to only about 114% of WT.

[0571] An explanatory hypothesis suggested that the sequence alignment of FIG. 17 may not accurately align the respective residues according to where they are found on a 3D immunoglobulin structure.

Example 46 Modeling the Immunoglobulin Fold

[0572] Consequently, crystal structure coordinates for the hCHγ1, hCHγ2 and hCHγ3 domains were obtained from the “Protein Data Bank”, maintained by Rutgers, the State University of New Jersey, Center for Integrative Proteomics Research, the San Diego Supercomputer Center (SDSC) and Skaggs School of Pharmacy and Pharmaceutical Sciences, San Diego. The structure of hCHλ1 is based on X-ray structure of 3dv, available through the “Protein Data Bank”. The structure of hCHλ2 is based on the X-ray structure of 2dts available through the “Protein Data Bank”. The structure of hCHλ3 is based on the X-ray structure of 2dts available through the “Protein Data Bank”. The structure of hCLλ is based on the X-ray structure of 4fqh available through the “Protein Data Bank”.

[0573] An homology alignment was generated, which aligned the sequences according to structure (FIG. 17). Crystal structure comparisons of CLκ and hCHγ1-ml (FIG. 19) showed that the hCHγ-CD connecting chain comprised a short α-helix. Modeling suggested that this CD α-helix presents the native S.sup.43 and the hTrast-[CLκ-K.sup.80A]/CHγ1-m1-D44 Asp insertion with side chains extending away from the binding pocket (FIGS. 20 & 21), and that the addition of CHγ1-D.sup.44 increases the size of and extends the CD α-helix. In addition, the model suggests that the CHγ1-Q.sup.79H residue is not in the optimal planar orientation, reducing its ability to participate in the reaction between PFP and the adjacent CHγ1-T.sup.78K (FIG. 21).

[0574] A comparison of the CLκ and CHγ1-m1-D44 domains of FIG. 22 suggests that the pocket shape may be optimized by removing the CD α-helix on the CD connecting chain, and replacing with a loop structure. Accordingly, a mutant CHγ1 domain, SEQ ID NO:165, was modelled and minimized. SEQ ID NO:165 comprises the CHγ1-T.sup.84K/Q.sup.79H double mutation of the CHγ1-ml mutant, as well as substituting the CHγ1 CD connecting chain from CHγ1-N.sup.42 inclusive through to CHγ1-L.sup.46 inclusive with KVDNALA (SEQ ID NO:166); SEQ ID NO:166 corresponds to residues of the CLκ domain (SEQ ID NO:6), with an additional Ala residue added. Modeling results shown in FIG. 22 suggest that CHγ1-Q.sup.79H occupies a more planar orientation, assisted by the introduced Val residue on the CD connecting chain. Mutating this residue to Leu or Ile may provide further stabilization to the imidazole ring, owing to the longer side chains being able to close the increased gap to the H residue, relative to the distance in native CLκ. Significantly, the model places the adjacent Asp residue in a suitable orientation and distance from CHγ1-Q.sup.79H to favour the δ-tautomeric form, and promote increased reactivity of CHγ1-T.sup.78K.

[0575] Similar modeling was performed on the CHγ2 and CHγ3 domains. FIG. 24 compares key WT residues of CLκ and CHγ2. While CHγ2-D.sup.49 and CHγ2-D.sup.50 appear to be in the correct position and orientation to assist with directional conjugation, CHγ2-D.sup.82 occupies a position even closer to CHγ2-G.sup.86H in the 3D model, thereby favouring the ε-tautomeric form, and reducing the reactivity of CHγ2-N.sup.85K. This model suggests that the CHγ2-D.sup.82 residue may explain the relatively low increase in directional conjugation seen in the CHγ2m mutant. FIG. 26 illustrates a minimized model of a proposed mutant for CHγ2: comprising the mutations CHγ2-D.sup.82A/N.sup.85K/G.sup.86H (SEQ ID NO:167). An immediate difference is seen: the imidazole ring appears in a more planar orientation, and within a suitable distance to CHγ2-D.sup.50 to bias formation of the δ-tautomer, and the CHγ2-D.sup.82A mutation has opened up the binding pocket, likely leading to a reduction in steric inhibition of PFP access. FIGS. 27-29 show the corresponding comparisons of CLκ with CHγ3 WT, and CHγ3m mutant. The model suggests that despite there being no residue biasing CHγ3-G.sup.80H towards the ε-tautomer form, the availability of CHγ3-Q.sup.79K to conjugate is likely sterically hindered by the long CHγ3-R.sup.76 side-chain. FIG. 30 illustrates a minimized model of a first proposed mutant for CHγ3: hCH3γm-CD1/EF, comprising the mutations CHγ3-S.sup.43V/N.sup.44D/R.sup.76A/Q.sup.79K/G.sup.80H (SEQ ID NO:168). The modeled difference can be easily observed in FIG. 30B: the availability of the binding pocket has opened up, and reduced steric interference on the reactive Lys.

[0576] From the sidechain distances in the modeled structure of FIG. 30B, it is unlikely that CHγ3-S.sup.43V or CHγ3-N.sup.44D interact with CHγ3-G.sup.80H to the same extent as the corresponding residues in CLκ. Accordingly, a second proposed mutant for CHγ3 was also minimized and modeled: hCH3γm-CD2/EF, comprising the mutations CHγ3S.sup.43L/N.sup.44E/R.sup.76A/Q.sup.79K/G.sup.80H (SEQ ID NO:169). The introduction of suitable residues on the CD connecting chain with longer side chains (CHγ3-S.sup.43L instead of CHγ3-S.sup.43V, and CHγ3-N.sup.44E instead of CHγ3-E.sup.44D) is modeled to have a binding pocket with greater structure similarities to that of CLκ (FIG. 31), and a greater likelihood of hydrogen bonding between CHγ3-N.sup.44E and CHγ3-G.sup.86H than between CHγ3-N.sup.44D and CHγ3-G.sup.80H.

[0577] Sequences of the CHγ domains and mutants were aligned with the CLκ and CLλ (FIG. 32).

[0578] The invention thus has been disclosed broadly and illustrated in reference to representative embodiments described above. Those skilled in the art will recognize that various modifications can be made to the present invention without departing from the spirit and scope thereof. All publications, patent applications, and issued patents, are herein incorporated by reference to the same extent as if each individual publication, patent application or issued patent were specifically and individually indicated to be incorporated by reference in its entirety. Definitions that are contained in text incorporated by reference are excluded to the extent that they contradict definitions in this disclosure.

[0579] It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable sub-combination.

[0580] It is specifically contemplated that any limitation discussed with respect to one embodiment of the invention may apply to any other embodiment of the invention. Furthermore, any composition of the invention may be used in any method of the invention, and any method of the invention may be used to produce or to utilize any composition of the invention. In particular, any aspect of the invention described in the claims, alone or in combination with one or more additional claims and/or aspects of the description, is to be understood as being combinable with other aspects of the invention set out elsewhere in the claims and/or description and/or sequence listings and/or drawings

[0581] In so far as specific examples found herein do not fall within the scope of an invention, said specific example may be explicitly disclaimed.

[0582] The use of the term “or” in the claims is used to mean “and/or” unless explicitly indicated to refer to alternatives only or the alternative are mutually exclusive, although the disclosure supports a definition that refers to only alternatives and “and/or.” As used herein the specification, “a” or “an” may mean one or more, unless clearly indicated otherwise. As used herein in the claim(s), when used in conjunction with the word “comprising”, the words “a” or “an” may mean one or more than one. As used herein “another” may mean at least a second or more. Unless otherwise defined herein, scientific and technical terms used in connection with the present invention shall have the meanings that are commonly understood by those of ordinary skill in the art. Further, unless otherwise required by context, singular terms shall include pluralities and plural terms shall include the singular. The words “comprises/comprising” and the words “having/including” when used herein with reference to the present invention are used to specify the presence of stated features, integers, steps or components but does not preclude the presence or addition of one or more other features, integers, steps, components or groups thereof.

TABLE-US-00075 TABLE 73 Toxin # IUPAC name STRUCTURE  #54 2-Methylalanyl-N-[(3R,4S,5S)-3- methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-2-methyl-3-oxo-3-{[(1S)- 2-phenyl-1-(1,3-thiazol-2- yl)ethyl]amino}propyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide (#54 [00111] #226 N,2-dimethylalanyl-N-[(3R,4S,5S)- 3-methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-3-{[(2S)-1-methoxy-1- oxo-3-phenylpropan-2-yl]amino}-2- methyl-3-oxopropyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide (#226) [00112] #194 N,2-{[(3S)-3-fluoropyrrolidin-3- yl]carbonyl}-N-[(3R,4S,5S)-3- methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-2-methyl-3-oxo-3-{[(1S)- 2-phenyl-1-(1,3-thiazol-2- yl)ethyl]amino}propyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide trifluoroacetic acid salt (#194) [00113] #192 N,2-{[(3R)-3-fluoropyrrolidin-3- yl]carbonyl}-N-[(3R,45,5S)-3- methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-2-methyl-3-oxo-3-{[(1S)- 2-phenyl-1-(1,3-thiazol-2- yl)ethyl]amino}propyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide trifluoroacetic acid salt (#192) [00114] #201 1,2-dimethyl-D-prolyl-N- [(3R,4S,5S)-1-{(2S)-2-[(1R,2R)-3- {[(2S)-3-(4-aminophenyl)-1- methoxy-1-oxopropan-2-yl]amino}- 1-methoxy-2-methyl-3- oxopropyl]pyrrolidin-1-yl}-3- methoxy-5-methyl-1-oxoheptan-4- yl]-N-methyl-L-valinamide, formate salt (#201) [00115] #158 N-{(2R,3R)-3-methoxy-3-[(2S)-1- {(3R,4S,5S)-3-methoxy-5-methyl-4- [methyl(N-{[(2S)-2- methylpiperidin-2-yl]carbonyl}-L- valyl)amino]heptanoyl}pyrrolidin- 2-yl]-2-methylpropanoyl}-L- phenylalaninate, trifluoroacetic acid salt (#158) [00116]  #70 2-methylalanyl-N-[(3R,4S,5S)-3- methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-3-{[(2S)-1-methoxy-1- oxo-3-phenylpropan-2-yl]amino}- 2-methyl-3-oxopropyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide (#70) [00117]  #47 N.sup.2-[(1-Aminocyclopentyl)carbonyl]- N-[(3R,4S,5S)-3-methoxy-1-{(2S)- 2-[(1R,2R)-1-methoxy-2-methyl-3- oxo-3-{[(1S)-2-phenyl-1-(1,3- thiazol-2- yl)ethyl]amino}propyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide (#47) [00118] #130 N,2-dimethylalanyl-N-[(3R,4S,5S)- 1-{(2S)-2-[(1R,2R)-3-{[(2S)-1-tert- butoxy-1-oxo-3-phenylpropan-2- yl]amino}-1-methoxy-2-methyl-3- oxopropyl]pyrrolidin-1-yl}-3- methoxy-5-methyl-1-oxoheptan-4- yl]-N-methyl-L-valinamide (#130) [00119] #159 methyl N-{(2R,3R)-3-methoxy-3- [(2S)-1-{(3R,4S,5S)-3-methoxy-5- methyl-4-[methyl(N-{[(2R)-2- methylpiperidin-2-yl]carbonyl}-L- valyl)amino]heptanoyl}pyrrolidin- 2-yl]-2-methylpropanoyl}-L- phenylalaninate, trifluoroacetic acid salt (#159) [00120] #209 1,2- dimethyl-L-prolyl-N-[(3R,4S,5S)- 3-methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-3-{[(2S)-1-methoxy-1- oxo-3-phenylpropan-2-yl]amino}-2- methyl-3-oxopropyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide, trifluoroacetic acid salt (#209) [00121] #131 2-methyl-D-prolyl-N-[(3R,4S,5S)-3- methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-3-{[(2S)-1-methoxy-1- oxo-3-phenylpropan-2-yl]amino}-2- methyl-3-oxopropyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide, trifluoroacetic acid salt (#131) [00122] #117 2-methyl-L-prolyl-N-[(3R,4S,5S)-3- methoxy-1-{(2S)-2-[(1R,2R)-1- methoxy-3-{[(2S)-1-methoxy-1- oxo-3-phenylpropan-2-yl]amino}-2- methyl-3 oxopropyl]pyrrolidin-1- yl}-5-methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide, trifluoroacetic acid salt (#117) [00123] #115 N,2-dimethylalanyl-N-{(1S,2R)-4- {(2S)-2-[(1R,2R)-3-{[(1S)-1- carboxy-2-phenylethyl]amino}-1- methoxy-2-methyl-3-oxopropyl] pyrrolidin-1-yl}-2-methoxy-1-[(1S)- 1-methylpropyI]-4-oxobutyl}-N- methyl-L-valinamide, trifluoroacetic acid salt (115) [00124]  #69 2-Methylalanyl-N-[(3R,4S,5S)-1- {(2S)-2-[(1R,2R)-3-{[(1S)-1- carboxy-2-phenylethyl]amino}-1- methoxy-2-methyl-3-oxopropyl] pyrrolidin-1-yl}-3-methoxy-5- methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide (#69) [00125] #151 1,2-dimethyl-L-prolyl-N-{(1S,2R)- 4-{(2S)-2-[(1R,2R)-3-{[(1S)-1- carboxy-2-phenylethyl]amino}-1- methoxy-2-methyl-3-oxopropyl] pyrrolidin-1-yl}-2-methoxy-1- [(1S)-1-methylpropyI]-4-oxobutyl}- N-methyl-L-valinamide (#151) [00126] #162 N-{(2R,3R)-3-methoxy-3-[(2S)-1- {(3R,4S,5S)-3-methoxy-5-methyl-4- [methyl(N-{[(2S)-2- methylpiperidin-2-yl]carbonyl}-L- valyl)amino]heptanoyl}pyrrolidin- 2-yl]-2-methylpropanoyl}-L- phenylalanine, trifluoroacetic acid salt (#162) [00127] #153 1,2-dimethyl-D-prolyl-N-{(1S,2R)- 4-{(2S)-2-[(1R,2R)-3-{[(1S)-1- carboxy-2-phenylethyl]amino}-1- methoxy-2-methyl-3-oxopropyl] pyrrolidin-1-yl}-2-methoxy-1- [(1S)-1-methylpropyl]-4-oxobutyl}- N-methyl-L-valinamide (#153) [00128] #118 2-methyl-L-prolyl-N-[(3R,4S,5S)-1- {(2S)-2-[(1R,2R)-3-{[(1S)-1- carboxy-2-phenylethyl]amino}-1- methoxy-2-methyl-3-oxopropyl] pyrrolidin-1-yl}-3-methoxy-5- methyl-1-oxoheptan-4-yl]-N- methyl-L-valinamide, trifluoroacetic acid salt (#118) [00129] #163 N-{(2R,3R)-3-methoxy-3-[(2S)-1- {(3R,4S,5S)-3-methoxy-5-methyl- 4-[methyl(N-{[(2R)-2- methylpiperidin-2-yl]carbonyl}-L- valyl)amino]heptanoyl}pyrrolidin- 2-yl]-2-methylpropanoyl}-L- phenylalanine, trifluoroacetic acid salt. (#163) [00130] #217 N-{(2R,3R)-3-[(2S)-1-{(3R,4S,5S)- 4-[(N-{[(3R)-3-fluoropyrrolidin-3- yl]carbonyl}-L- valyl)(methyl)amino]-3-methoxy-5- methylheptanoyl}pyrrolidin-2-yl]-3- methoxy-2-methylpropanoyl}-L- phenylalanine, trifluoroacetic acid salt (#217) [00131] #112 2-methylalanyl-N-[(3R,4S,5S)-1- {(2S)-2-[(1R,2R)-3-{[(1S,2R)-1- hydroxy-1-phenylpropan-2- yl]amino}-1-methoxy-2-methyl-3- oxopropyl]pyrrolidin-1-yl}-3- methoxy-5-methyl-1-oxoheptan-4- yl]-N-methyl-L-valinamide (#112) [00132] MMAD The wavy line indicates a typical location for linker attachment. In non-conjugated form, the wavy line is typically conneted to a H atom. [00133] MMAE The wavy line indicates a typical location for linker attachment. In non-conjugated form, the wavy line is typically conneted to a H atom. [00134] MMAF The wavy line indicates a typical location for linker attachment. In non-conjugated form, the wavy line is typically conneted to a H atom. [00135] [00136]