Recombinant polypeptide production methods

Abstract

Herein is reported a method for producing a fusion-polypeptide comprising the seeps of a) cultivating a mammalian cell comprising a nucleic acid encoding a variant fusion-polypeptide wherein the amino acid sequence of the fusion-polypeptide has been modified by replacing in a pro-fusion-polypeptide the endogenous protease cleavage site between the pro-peptide and the fusion-polypeptide with an exogenous (with respect to the origins of the parts of the fusion-polypeptide) or artificial protease cleavage site, and b) recovering the fusion-polypeptide or fusion-pro-polypeptide from the cell or the cultivation median and thereby producing the (recombinant) fusion-polypeptide.

Claims

1. A method for producing a recombinant polypeptide, comprising the following steps: cultivating a eukaryotic cell comprising a nucleic acid encoding a variant polypeptide wherein the polypeptide has been modified by introducing one or more artificial glycosylation sites located in a specific glycosylation tag that is fused either directly or via a linker to the polypeptide and consists of one of the following amino acid sequences: NGTEGPNFYVPFSNATGVV (SEQ ID NO: 10), NGTEGPNFYVPFSNATGVVR (SEQ ID NO: 16), AAANGTGGA (SEQ ID NO: 17), and NATGADNGTGAS (SEQ ID NO: 19), and recovering the variant recombinant polypeptide from the cell or the cultivation medium.

2. The method according to claim 1, wherein the producing steps comprise: providing a nucleic acid encoding the polypeptide, modifying the nucleic acid to encode a variant polypeptide wherein the amino acid sequence of the polypeptide has been modified to comprise one or more artificial glycosylation sites, introducing the nucleic acid into a eukaryotic cell, cultivating the eukaryotic cell, and recovering the variant recombinant polypeptide from the cell or the cultivation medium and thereby producing the recombinant polypeptide.

3. The method according to claim 1, wherein the eukaryotic cell is a mammalian cell.

4. The method according to claim 2, wherein the eukaryotic cell is a mammalian cell.

Description

DESCRIPTION OF THE FIGURES

(1) FIG. 1 Stability of BDNF variants during cell culture. Wild-type BDNF (A) produced in E. coli and myc-tagged BDNF produced by transient expression in HEK293 cells (B) were added to a growing culture of HEK293 cells (10 μg/ml) and the BDNF concentration were determined at day 0 and day 4 in samples by reducing SDS PAGE and Western Blot analysis using an anti-BDNF antibody for staining/visualization; lane 1: reference material from Peprotech (E. coli); lane 2: molecular weight marker; the Western Blot has been cut and re-assembled for simplicity.

(2) FIG. 2 SDS PAGE/Western Blot analysis of cell culture supernatants containing expressed/secreted BDNF variant polypeptides; lane 1: molecular weight marker; lane 2: row 1 of Table 9; lane 3: row 5 of Table 9; lane 4: row 2 of Table 9; lane 5: row 4 of Table 9; lane 6: row 3 of Table 9; lane 7=mBDNF reference; the Western Blot has been cut an re-assembled for simplicity.

(3) FIG. 3 SDS PAGE/Western Blot analysis of cell culture supernatants containing expressed/secreted BDNF variant polypeptides; lane 1: molecular weight marker, lane 2: row 1 of Table 6 (glycosylated pro-BDNF); lane 3: row 2 of Table 6 (pro-BDNF); lane 4: row 3 of Table 6 (glycosylated mature BDNF); lane 5: row 4 of Table 6 (mature BDNF).

(4) FIG. 4 SDS PAGE/Western Blot analysis of cell culture supernatants containing expressed/secreted BDNF variant polypeptides as of Table 11; lanes A: row 1 of Table 11; lane B: row 2 of Table 11; lane C: row 3 of Table 11; lane D: row 4 of Table 11; lane E: row 5 of Table 11; lanes F: row 6 of Table 11; the star denotes that different expression conditions were used.

(5) The following examples, sequences and figures are provided to aid the understanding of the present invention, the true scope of which is set forth in the appended claims. It is understood that modifications can be made in the procedures set forth without departing from the spirit of the invention.

EXAMPLES

(6) Recombinant DNA Techniques

(7) Standard methods were used to manipulate DNA as described in Sambrook, J. et al., Molecular cloning: A laboratory manual; Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989. The molecular biological reagents were used according to the manufacturer's instructions.

(8) Gene Synthesis

(9) Desired genes and gene segments were prepared by chemical synthesis at Geneart GmbH (Regensburg, Germany). The synthesized genes and gene fragments were cloned into an E. coli plasmid for propagation/amplification. The DNA sequence of the subcloned genes and gene fragments were verified by DNA sequencing.

(10) Protein Determination

(11) The protein concentration of purified polypeptides was determined by determining the optical density (OD) at 280 nm, using the molar extinction coefficient calculated on the basis of the amino acid sequence of the polypeptide.

(12) Description of the Basic/Standard Mammalian Expression Plasmid

(13) Desired genes/polypeptides were expressed by transient transfection of human embryonic kidney cells (HEK293). For the expression of a desired gene/polypeptide (e.g. antibody-GFP-fusion-polypeptide, wild-type BDNF, BDNF variant polypeptides, BDNF-Fab- and BDNF-scFv-fusion-polypeptides) a transcription unit comprising the following functional elements was used: the immediate early enhancer and promoter from the human cytomegalovirus (HCMV) including intron A, a human heavy chain immunoglobulin 5′-untranslated region (5′UTR), a gene to be expressed, and the bovine growth hormone polyadenylation sequence (BGH pA).

(14) Beside the expression unit/cassette including the desired gene to be expressed the basic/standard mammalian expression plasmid contains an origin of replication from the vector pUC18 which allows replication of this plasmid in E. coli, and a beta-lactamase gene which confers ampicillin resistance in E. coli.

Example 1

(15) Generation of Antibody Expression Plasmids

(16) a) Generation of the Antibody Expression Plasmids for the Parental Human Anti-Human IGF-1R Antibody

(17) The gene segments encoding the human kappa light (Vk) and heavy chain variable regions (VH) were joined to the gene segments encoding the human kappa light chain constant region (Ck) or the human gamma-1 heavy chain constant region (CH1-Hinge-CH2-CH3), respectively. Both antibody chain genes were expressed from two separate expression plasmids including the genomic exon-intron structure of the antibody genes. The amino acid sequence of the mature (without signal sequence) heavy and light chain of anti-human IGF-1R antibody are shown in SEQ ID NO: 05 and SEQ ID NO: 06.

(18) The expression of antibody chains is controlled by a shortened intron A-deleted immediate early enhancer and promoter from the human cytomegalovirus (HCMV) including a human heavy chain immunoglobulin 5′-untranslated region (5′-UTR), a murine immunoglobulin heavy chain signal sequence, and the polyadenylation signal from bovine growth hormone (BGH pA). The expression plasmids also contain an origin of replication and a β-lactamase gene from the vector pUC18 for plasmid amplification in Escherichia coli (see Kopetzki, E., et al., Virol. J. 5 (2008) 56; Ji, C., et al., J. Biol. Chem. 284 (2009) 5175-5185).

(19) b) Generation of the Anti-Transferrin Receptor Antibody Light Chain Expression Plasmid

(20) In order to obtain a light chain for the anti-transferrin-receptor antibody, a light chain gene was chemically synthesized coding for the murine immunoglobulin heavy chain signal sequence MGWSCIILFLVATATGVHS (SEQ ID NO: 14), the VL variable domain of the rat-anti-murine transferrin receptor and the human Vkappa light chain constant region. The amino acid sequence of the VL domain of the rat antibody was obtained from Boado, R. J., et al., Biotechnol. Bioeng. 102 (2009) 1251-1258). The amino acid sequence of the chimeric rat/human anti-transferrin receptor antibody light chain is shown in SEQ ID NO: 80.

Example 2

(21) Generation of Antibody-GFP-Fusion Polypeptide Expression Plasmids

(22) a) Generation of the Expression Plasmids for the Anti-IGF-1R Antibody-GFP Fusion-Polypeptides

(23) All anti-human IGF-1R antibody heavy chain-GFP-fusion-polypeptide encoding genes were assembled by fusing a chemically synthesized DNA fragment coding for the respective GFP variant and a glycine-serine linker consisting of 2 Gly.sub.4Ser repeats and a further Gly (heavy chain . . . LSPG-gggsggggsg-GFP) to the 3′ end of the anti-IGF-1R antibody heavy chain gene coding for a slightly truncated human gamma-1 heavy chain constant region (removal of the last natural amino acid Lys). The amino acid sequence of the anti-IGF-1R antibody heavy chain eGFP, emGPF and tagGFP fusion protein is shown in SEQ ID NO: 07, SEQ ID NO: 08 and SEQ ID NO: 09, respectively.

(24) The antibody heavy and light chain genes were expressed from two separate expression plasmids including the genomic exon-intron structure of antibody genes.

(25) b) Generation of Expression Plasmids for Anti-IGF-1R Antibody Heavy Chain-eGFP-Opsin-Tag-Fusion-Polypeptides

(26) The expression plasmids for the transient expression of the anti-IGF-1R antibody heavy chain-GFP-opsin-tag-fusion-polypeptides in HEK293 cells was derived from the expression vectors described above. They differ only in the DNA segment encoding the GFP-opsin-tag where upon the 19 amino acid peptide (NGTEGPNFYVPFSNATGVV; opsin(M); SEQ ID NO: 10) was fused directly to the C-terminal end of a respective GFP. As example, the amino acid sequence of the anti-IGF-1R antibody heavy chain-eGFP-opsin(M)-tag-fusion-polypeptide is shown in SEQ ID NO: 33.

Example 3

(27) Generation of BDNF Expression Plasmids

(28) a) Generation of the Expression Plasmid for Wild-Type Pre-Pro-BDNF

(29) The DNA segment coding for human pre-pro-BDNF gene was prepared by chemical synthesis and inserted into the basic expression vector described above. For this purpose, the pre-pro-BDNF gene was ligated with the CMV-promoter at its 5′-end and with the bovine growth hormone polyadenylation sequence at its 3′-end. The amino acid sequence of the wild type pre-pro-BDNF protein is shown in SEQ ID NO: 20.

(30) b) Generation of the Expression Plasmids for BDNF Variants

(31) In order to obtain optimal production yield several BDNF variants were constructed (see Table below): in some variants the wild-type BDNF signal sequence (pre-segment) was exchanged by an signal sequence which is derived from a highly expressed murine immunoglobulin heavy chain antibody (MGWSCIILFLVATATGVHS); in some variants the codon usage of the encoding BDNF gene was exchanged to an optimized codon usage in the pro-segment and/or in the mature part of BDNF; the BDNF genes with optimized codon usages were obtained by backtranslation of the amino respective acid sequence using algorithms from Geneart (see e.g. Fath, S., et al., PLOS One 6 (2011) e17596); in some variants a T7-His6-tag (SEQ ID NO: 12) was used, since it is generally believed to enhance protein expression (see e.g. Luan, C. H., et al., Genome Res. 14 (2004) 2102-2110); in most BDNF variants a His6-tag was included in order to simplify sample preparation/purification; in some variants the last three C-terminal amino acids, the RGR motif of mature BDNF, was deleted, since it might function as cryptic protease cleavage site for proteases like furin or other PC convertases; in another variant the signal sequence and pro-segment of BDNF was exchanged by the corresponding amino acid sequence of human NGF, since it was published that this variation improves expression for another neurotrophin (Iwane et al., Appl. Microbiol. Biotechnol. 41 (1994) 225-232).

(32) TABLE-US-00014 TABLE 14 use of signal sequence and codon usage BDNF gene/variant pre pro mature (Description of gene/construct features) (BDNF) (BDNF) (BDNF) tag-1 tag-2 pre-pro-BDNF-T7-His6 BDNF wild-type wild-type T7 His6 pre-pro-BDNF(-RGR)-T7-His6 BDNF wild-type deltaRGR; 1 T7 His6 pre-pro-BDNF(-RGR)-His6, BDNF wild-type optimized, — His6 codon-optimized deltaRGR pre-pro-BDNF(-RGR)-His6, BDNF optimized optimized, — His6 codon-optimized deltaRGR pre(Ab)-pro-BDNF, antibody; 2 optimized optimized — — pre-pro-BDNF(-RGR)-His6, BDNF optimized optimized, — His6 codon-optimized deltaRGR pre(Ab)-pro-BDNF antibody; 2 optimized wild-type, — His6 optimized pre(NGF)-pro(NGF)-BDNF NGF; 3 wild-type; wild-type, 1 — His6 NGF; 3 1: C-terminal RGR-motif of BDNF deleted (deltaRGR) 2: signal sequence derived from a highly expressed murine immunoglobulin heavy chain antibody (MGWSCIILFLVATATGVHS) as defined by the amino acid sequence shown in SEQ ID NO: 14 3: signal sequence and pro-fragment of BDNF exchanged by corresponding sequence from human NGF
c) Generation of the Expression Plasmids for Pre-Pro(IgA)-BDNF and Pre-Pro(IgA; His6) Variants, Partially Including a His6-Tag N-Terminally of Mature BDNF within the Pro-Segment

(33) The pre-pro-(IgA)-BDNF gene codes for a pro-polypeptide variant wherein the naturally occurring furin cleavage site (RVRR) was replaced by the engineered IgA protease cleavage site with the sequence GSVVAPPAP (see Table below).

(34) In addition, in some variants a R54A point mutation was introduced into the pro-polypeptide of BDNF to destroy a putative protease cleavage site (see Mowla, S. J., et al., J. Biol. Chem. 276 (2001) 12660-12666). Furthermore, in some variants also a removable His6-tag was included N-terminally of mature BDNF within the pro-fragment. This tag simplifies protein purification of the pro(IgA; His6)-BDNF variant protein. Upon final in vitro protein maturation with IgA protease the pro(IgA; His6 fragment) is removed and, thus, a potential risk of immunogenicity is avoided.

(35) The expression plasmids for the transient expression of the pre-pro(IgA)-BDNF and pre-pro(IgA; His6)-BDNF variant genes/proteins in HEK293 cells are derived from the expression vector described above which encodes for the pre-pro-BDNF(-RGR)-T7-His6 protein. They differentiate in the following characteristics:

(36) TABLE-US-00015 TABLE 15 BDNF variant R54A protease (Description of construct features) mutation cleavage site His6-tag SEQ ID NO: pre-pro-BDNF(-RGR)-T7-His6 no furin C-terminus of 35 (wild-type) BDNF(-RGR) pre-pro(IgA)-BDNF(-RGR)-His6 no IgA C-terminus of 43 BDNF(-RGR) pre-pro(His6-R54A-IgA)-BDNF(-RGR) yes IgA within pro(BDNF) 45 pre-pro(His6-R54A)-BDNF(-RGR) yes furin within pro(BDNF) 46 (wild-type) pre-pro(R54A)-BDNF(-RGR)-His6 yes furin C-terminus of 44 (wild-type) BDNF(-RGR)

(37) The numbering of the R54A mutation is based on the amino acid sequence of wildtype pre-pro-BDNF (SEQ ID NO: 20)

(38) d) Generation of the Expression Plasmids for Isoelectric Point Engineered BDNF Variants

(39) Previously some in E. coli expressed and refolded BDNF variants have been described with an engineered lowered IEP. The pre-pro-BDNF(-RGR)-T7-His6 gene was mutated accordingly and the mutant BDNF genes obtained transiently expressed in HEK293 cells (sec Table below). In addition, a myc-tagged BDNF variant was constructed, since (I) the myc-tag (EQKLISEEDL; SEQ ID NO: 90) introduces a net charge difference of about −3 and (2) the myc-tag is of human origin and thus supposed to be less immunogenic.

(40) TABLE-US-00016 TABLE 16 mutations used calculated IEP for IEP lowering source of of maturated BDNF variant within mature BDNF BDNF(deltaRGR) (description of construct features) BDNF(deltaRGR) mutations tag(s) variant SEQ ID NO: pre-pro-BDNF(-RGR)-T7-His6 none wild-type T7, 9.58 35 BDNF His6 pre-pro-BDNF(-RGD; P60E, K65D, 53915428| T7, 8.15 47 P60E, K65D, K73D, K73D, K95A gb|AAU96 His6 K95A)-T7-His6 862.1|; sequence 10 of U.S. Pat. No. 6,723,701 pre-pro-BDNF(-RGD; K65D, K73D, mBDNF: T7, 8.15 48 K65D, K73D, K95A, K95A, R97A 53915427| His6 R97A)-T7-His6 gb|AAU96 861.1| sequence 9 of U.S. Pat. No. 6,723,701 pre-pro-BDNF(-RGD)- none wild-type myc, 8.83 49 myc-His6 BDNF His6 -RGD and (deltaRGR); deletion of the last 3 C-terminal amino acids of mature BDNF; numbering of amino acid mutations starts at the first amino acid of mature BDNF; the IEP of the matured BDNF(-RGR) variants were calculated using the protein statistic program pepstats from the European Molecular Biology Open Software Suite (EMBOSS).
e) Generation of the Expression Plasmids for BDNF Variants with Additional Glycosylation Sites

(41) BDNF variants were generated which harbor (1) a C-terminal tag containing glycosylation sites or (2) one engineered glycosylation site within the matured BDNF moiety.

(42) The glycosylation tags were deduced from published sequences (e.g. Meder, D., et al., J. Cell Biol. 168 (2005) 303-313; Bulbarelli, A., et al., J. Cell Sci. 115 (2002) 1689-1702; Perlman, S., et al., J. Clin. Endocrinol. Metab. 88 (2003) 3227-3235; WO 2002/002597) and putative N-glycosylation sites were predicted with an artificial neuronal network (NetNglyc server, http://www.cbs.dtu.dk/services/NetNGlyc/).

(43) For introduction of N-glycosylation sites within the maturated BDNF moiety the sequence was inspected for the presence of asparagins, serins or threonins within the matured BDNF sequence. Then, based on the three-dimensional protein structure of human BDNF (lbnd; www.rcsb.org) all non-surface localized Asn, Ser or Thr residues were excluded. For the remaining surface exposed Asn, Ser and Thr residues the adjacent amino acid residues were identified in order to engineer a putative N-glycosylation site (consensus motif: N-X-(S/T), X=any amino acid except Pro) by site-directed mutagenesis. The amino acid position of these putative engineered N-glycosylation sites were used to identify the corresponding amino acids in the structurally and functionally homologous neurotrophins NGF and NT-3 by sequence alignment and protein 3D-structure comparisons. Amino acid positions were excluded that are expected to be part of the neurotrophin::p75NTR or neutrophin::Trk(A, B) interaction interface based on homologous receptor::ligand crystal structures (e.g. 3buk, 3ij2 and 2ifg). Selected mutations for surface-exposed putative N-glycosylation sites outside the putative BDNF-TrkB/p75 interaction interfaces are mentioned in the Table below (second column).

(44) The expression plasmids for the transient expression of N-glycosylated BDNF variant proteins in HEK293 cells are derived from the expression vector coding for the BDNF variant pre-pro-BDNF(-RGR)-T7-His6 (-RGR: a truncated mature wild-type BDNF wherein the last 3 C-terminal amino acids RGR are deleted; a T7-tag and a His6-tag is attached to the truncated C-terminus of BDNF via a GSG-linker). The BDNF segments, tags, glycolinkers and mutations introduced for the generation of additional artificial N-glycosylation sites are presented in the following table.

(45) TABLE-US-00017 TABLE 17 BDNF variant mutations within mature glycosylation tag further SEQ ID NO (description of construct features) BDNF(deltaRGR) SEQ ID NO: tags (aa sequence): parental construct, none none T7; 35 pre-pro-BDNF(-RGR)- His6 T7-His6 pre-pro-BDNF(-RGR)- none 15 T7; 36 opsin(S)-T7-His6; His6 pre-pro-BDNF(-RGR)- none 15 His6 37 opsin(S)-His6 pre-pro-BDNF(-RGR)- none 16 His6 38 opsin(L)-His6 pre-pro-BDNF(-RGR)- none 18 His6 39 glyco-1 pre-pro-BDNF(-RGR)- none 17 His6 40 glyco-2 pre-pro-BDNF(-RGR)- none 19 His6 41 glyco-3 pre-pro-BDNF(-RGR; M61T none T7; 42 M61T)-T7-His6 His6 pre-pro-BDNF(-RGR; Q79S none T7; 91 Q79S)-T7-His6 His6 pre-pro-BDNF(-RGR; R81N none T7; 89 R81N)-T7-His6 His6 pre-pro-BDNF(-RGR; W19N none T7; 92 W19N)-T7-His6 His6 pre-pro-BDNF(-RGR; K25N none T7; 93 K25N)-T7-His6 His6 pre-pro-BDNF(-RGR; D30N none T7; 94 D30N)-T7-His6 His6 pre-pro-BDNF(-RGR; G33N none T7; 95 G33N)-T7-His6 His6 pre-pro-BDNF(-RGR; T35N none T7; 96 T35N)-T7-His6 His6 pre-pro-BDNF(-RGR; P43N none T7; 97 P43N)-T7-His6 His6 pre-pro-BDNF(-RGR; P43N, V42G none T7; 98 P43N, V42G)-T7-His6 His6 pre-pro-BDNF(-RGR; M61S, P60G none T7; 99 M61S, P60G)-T7-His6 His6 pre-pro-BDNF(-RGR; G62N, Y63G none T7; 100 G62N, Y63G)-T7-His6 His6 pre-pro-BDNF(-RGR; Q79T none T7; 101 Q79T)-T7-His6 His6 pre-pro-BDNF(-RGR; W76N none T7; 102 W76N)-T7-His6 His6 pre-pro-BDNF(-RGR; M92N none T7; 103 M92N)-T7-His6 His6 pre-pro-BDNF(-RGR; D106N none T7; 104 D106N)-T7-His6 His6 pre-pro-BDNF(-RGR; T112N none T7; 105 T112N)-T7-His6 His6

(46) SEQ ID NOs for amino acid mutations introduced by site-directed mutagenesis into mature BDNF(deltaRGR) are exemplarily shown for the BDNF variants pre-pro-BDNF(-RGR; M61T)-T7-His6 and pre-pro-BDNF(-RGR; R81N)-T7-His6.

(47) f) Generation of the Expression Plasmids for BDNF Variants Containing Combined Mutations for the Introduction of Multiple (Two or More) N-Glycosylation Sites and an IgA Cleavage Site

(48) In order to generate BDNF variants with multiple N-glycosylation sites some of the additional N-glycosylation sites identified in previous experiments were combined. The starting construct pre-pro(IgA)-BDNF(-RGR)-His6 variant polypeptide is characterized by an IgA protease cleavage site instead of the native furin site within the pro-segment, a C-terminally truncated mature BDNF (deletion of the last 3 amino acids RGR) and a C-terminal His6-tag. The desired mutations and tags were introduced/attached as shown in the Table below.

(49) TABLE-US-00018 TABLE 18 BDNF variant (description of additional glycosylation sites SEQ construct G62N, ID features) opsin(L)-tag K25N T35N Y63G Q79T R81N myc-tag NO: parental 43 construct, pre-pro(IgA)- BDNF(-RGR)- His6 pre-pro(IgA)- x x 50 BDNF(-RGR; R81N)-opsin(L) pre-pro(IgA)- x 53 BDNF(-RGR; R81N)-His6 pre-pro(IgA)- x x x 51 BDNF(-RGR; R81N)-opsin(L)- myc-His6 pre-pro(IgA)- x x 52 BDNF(-RGR; R81N)-myc-His6 pre-pro(IgA)- x 54 BDNF(-RGR)- opsin(L)-His6 pre-pro(IgA)- x x 55 BDNF(-RGR; K25N)-opsin(L)- His6 pre-pro(IgA)- x x 56 BDNF(-RGR; T35N)-opsin(L)- His6 pre-pro(IgA)- x x x 57 BDNF(-RGR; K25N, T35N)- opsin(L)-His6 pre-pro(IgA)- x x x 58 BDNF(-RGR; K25N, T35N, R81N)-T7-His6 pre-pro(IgA)- x x x x 59 BDNF(-RGR; K25N, T35N, Q79T, R81N)-T7- His6 pre-pro(IgA)- x x 60 BDNF(-RGR; G62N, Y63G)- opsin(L)-His6 pre-pro(IgA)- x x x x 61 BDNF(-RGR; G62N, Y63G, K25N, T35N)- opsin(L)-His6

(50) g) Generation of the Expression Plasmids for BDNF Variants Containing Combined Mutations for the Introduction of Multiple (Two or More) N-Glycosylation Sites

(51) In order to generate BDNF variants with multiple N-glycosylation sites some of the additional N-glycosylation sites identified in previous experiments were combined. For this purpose the pre-pro-BDNF(-RGR)-T7-His6 variant protein characterized by a C-terminally truncated mature BDNF (deletion of the last 3 amino acids RGR) and a C-terminal T7-His6-tag was used as starting material. The desired mutations were inserted as shown in the Table below.

(52) TABLE-US-00019 TABLE 19 BDNF variant additional glycosylation sites (description of G62N, SEQ ID NO construct features) K25N T35N Y63G Q79T R81N (aa sequence): parental construct, 35 pre-pro-BDNF(-RGR)- T7-His6 pre-pro-BDNF(-RGR; x x 62 G62N, Y63G, Q79T)- T7-His6 pre-pro-BDNF(-RGR; x x 63 T35N, Q79T)-T7-His6 pre-pro-BDNF(-RGR; x x 64 K25N, Q79T)-T7-His6 pre-pro-BDNF(-RGR; x x x 65 T35N, G62N, Y63G, Q79T)-T7-His6 pre-pro-BDNF(-RGR; x x 106 Q79T, R81N)-T7-His6 pre-pro-BDNF(-RGR; x x 66 T35N; G62N; Y63G)- T7-His6 pre-pro-BDNF(-RGR; x x 67 K25N; G62N; Y63G)- T7-His6 pre-pro-BDNF(-RGR; x x 68 K25N; T35N)-T7-His6 pre-pro-BDNF(-RGR; x x 69 G62N, Y63G, R81N)- T7-His6 pre-pro-BDNF(-RGR; x x x 70 K25N; T35N; Q79T)- T7- His6

Example 4

(53) Generation of BDNF Antibody Fragment Fusion-Polypeptides Expression Plasmids

(54) a) Generation of the Expression Plasmids for the BDNF-Fab Antibody Heavy Chain Fusion-Polypeptides

(55) In order to obtain BDNF-(Gly.sub.4Ser).sub.a-Fab(anti-IGF-1R antibody heavy chain) fusion-polypeptides, plasmids for transient expression in HEK293 cells were constructed which harbored a chemically synthesized DNA fragment of a CDS (coding DNS sequence) coding for polypeptides with the following characteristics: the wild-type pre-pro-BDNF moiety deleted for the C-terminal RGR motif is fused at the C-terminus with a glycine-rich linker followed by a Fab heavy chain portion (VH-CH1) of the human anti-IGF-1R antibody and a C-terminal His6-tag; the glycine-rich linker consists of a (G4S)2-GG or a (G4S)4-GG or a (G4S)6-GG motif (sec Table below).

(56) b) Generation of the Expression Plasmids for the BDNF-Fab Antibody Light Chain Fusion-Polypeptides

(57) In order to obtain BDNF-(Gly.sub.4Ser).sub.n-Fab fusion-polypeptides, plasmids for transient expression in HEK293 cells were constructed which harbored a chemically synthesized DNA fragment of a CDS coding for polypeptides with the following characteristics: the wild-type pre-pro-BDNF moiety deleted for the C-terminal RGR motif is fused at the C-terminus with a glycine-rich linker followed by the Fab VL-Ckappa light chain domains of the human anti-IGF-1R antibody and a C-terminal His6 tag; the glycine-rich linker consists of a (G4S)2-GG or a (G4S)4-GG or a (G4S)6-GG motif (see Table above).

(58) TABLE-US-00020 TABLE 20 linker between BDNF-antibody variant, BDNF and Fab Fab fragment SEQ ID NO (description of construct features) fragment (anti-IGF-1R mAB) Tag (aa sequence): pre-pro-BDNF(-RGD)_(G4S)2GG_VL<IGF-1R>-Ck-His6 (G.sub.4S).sub.2-GG VL-Ckappa His6 71 pre-pro-BDNF(-RGD)_(G4S)4GG_VL<IGF-1R>-Ck-His6 (G.sub.4S).sub.4-GG VL-Ckappa His6 72 pre-pro-BDNF(-RGD)_(G4S)6GG_VL<IGF-1R>-Ck-His6 (G.sub.4S).sub.6-GG VL-Ckappa His6 73 pre-pro-BDNF(-RGD)_(G4S)2GG_VH<IGF-1R>-CH1-His6 (G.sub.4S).sub.2-GG VH-CH1 His6 74 pre-pro-BDNF(-RGD)_(G4S)4GG_VH<IGF-1R>-CH1-His6 (G.sub.4S).sub.4-GG VH-CH1 His6 75 pre-pro-BDNF(-RGD)_(G4S)6GG_VH<IGF-1R>-CH1-His6 (G.sub.4S).sub.6-GG VH-CH1 His6 76

(59) c) Generation of the Expression Plasmid for the Anti-IGF-1R Antibody Light Chain

(60) The native anti-IGF-1R antibody light chain was used for the generation anti-IGF-1R based BDNF-Fab complexes. The generation of the anti-IGF-1R antibody light chain expression plasmid is described in example 1.

(61) d) Generation of the Expression Plasmid for the BDNF-Fab(Anti-IGF-1R Antibody Heavy Chain) Fusion-Polypeptides Containing a Negatively Charged Gly-Asp Linker

(62) In order to obtain BDNF-(G3D)4-Fab(anti-IGF-1R antibody heavy chain fusion-polypeptide, a plasmid for transient expression in HEK293 cells was constructed which harbored a chemically synthesized DNA fragment of a CDS coding for the polypeptide with the following characteristics: the wild-type pre-pro-BDNF moiety deleted for the C-terminal RGR motif is fused at the C-terminus with a glycine-rich negatively charged linker followed by the Fab VH-CH1 heavy chain domains of the human anti-IGF-1R antibody and a C-terminal His6-tag; the glycine-rich negatively charged linker consists of the (G3D)4-GGGS motif.

(63) e) Generation of the Expression Plasmids for the BDNF-Fab(Anti-IGF-1R Antibody Heavy Chain) Fusion Proteins Harboring an IgA Cleavage Site within the Pro-BDNF Segment, a Negatively Charted GlyAsp Linker and Multiple N-Glycosylation Sites

(64) In order to obtain pro(IgA)-BDNF-(G3D)4-Fab(anti-IGF-1R antibody heavy chain fusion-polypeptides with multiple N-glycosylation sites, plasmids for transient expression in HEK293 cells were constructed which harbored a chemically synthesized DNA fragment of a CDS which code for polypeptides with the following characteristics: the pre-pro(IgA)-BDNF moiety deleted for the C-terminal RGR motif of mature BDNF is fused at the C-terminus with the extended opsin-tag (NGTEGPNFYVPFSNATGVVR; opsin(L); SEQ ID NO: 16) followed by a negatively charged glycine-aspartic-acid-rich linker and a Fab heavy chain portion (VH-CH1, partially extended with the hinge-derived peptide EPKSC) of the human monoclonal antibody directed against human insulin-like growth factor 1 (IGF-1R) and a C-terminal His6-tag; the negatively charged glycine-aspartic-acid-rich linker consists either the (G3D)4-GGGS motif or the (G2D)5-G2SG motif.

(65) The details of the constructs are summarized in the Table below.

(66) TABLE-US-00021 TABLE 21 introduced additional linker N-glycosylation sites between origin of BDNF-antibody variant, protease within BDNF Fab heavy hinge (description of construct His- cleavage opsin- BDNF and Fab chain EPKSC SEQ features) tag site tag (deltaRGR) fragment fragment peptide ID NO: pre-pro(IgA)-BDNF(- His6 IgA none (G3D)4- <IGF-1R> yes 85 RGR)_(G3D)4- GGGS G3S_VH<IGF-1R>-CH1- EPKSC-His6 pre-pro(IgA)-BDNF(-RGR; His6 IgA opsin(L) +R81N GSG-- <IGF-1R> no 86 R81N)_opsin(L)-(G3D)4- opsin(L)- G3S_VH<IGF-1R>-CH1- (G3D)4- His6 GGGS pre-pro(IgA)-BDNF(- His6 IgA none (G3D)4- <IGF-1R> no 84 RGR)_(G3D)4- GGGS G3S_VH<IGF-1R>-CH1- His6 pre-pro(IgA)-BDNF(-RGR; His6 IgA opsin(L) +R81N GSG- <IGF-1R> no 87 R81N)_opsin(L)-(G2D)5- opsin(L)- G2SG_VH<IGF-1R>- (G2D)5- CH1-His6 G2SG

(67) f) Generation of the Expression Plasmids for the BDNF-Fab(Anti-Transferrin Receptor Antibody Heavy Chain) Fusion-Polypeptides Harboring an IgA Cleavage Site within the Pro-BDNF Segment, a Negatively Charted GlyAsp Linker and Multiple N-Glycosylation Sites

(68) In order to obtain pro(IgA)-BDNF-(G3D)4-Fab(anti-TfR) antibody heavy chain fusion-polypeptides with a negatively charged GlyAsp linker and multiple N-glycosylation sites, plasmids for transient expression in HEK293 cells were constructed which harbored a chemically synthesized DNA fragment of a CDS which code for polypeptides with the following characteristics: the pre-pro(IgA)-BDNF moiety deleted for the C-terminal RGR motif is fused at the C-terminus with the opsin(L)-tag followed by a negatively charged glycine-aspartic-acid-rich linker and a chimeric rat/human Fab heavy chain portion wherein the VH variable domain is derived from the rat 8D3 monoclonal antibody which is directed against the mouse transferrin receptor (mTfR) and wherein the CH1 domain is derived from human IgG and a C-terminal His6-tag; the amino acid sequence of VH domain of the antibody was obtained from Boado, R J., et al., Biotechnol. Bioeng. 102 (2009) 1251-1258); the negatively charged glycine-aspartic-acid-rich linker consists of the (G3D)4-GGGS motif; in most variants the endogenous furin/PC convertase cleavage site between the pro-segment and the mature part of BDNF was exchanged by an IgA protease cleavage site (GSVVAPPAP); in some variants the truncated wild-type mature BDNF moiety (BDNF; deltaRGR) was exchanged by a truncated mature BDNF variant harboring an artificial R209N glycosylation site (BDNF(deltaRGR; R209N)).

(69) The details of the constructs are summarized in the Table below.

(70) TABLE-US-00022 TABLE 22 introduced additional linker N-glycosylation sites between origin of BDNF-antibody variant, protease BDNF BDNF Fab heavy SEQ ID (description of construct His6- cleavage opsin- (ΔRGR) and Fab chain hinge NO (aa features) tag site tag moiety fragment fragment EPKSC sequence): pre-pro(IgA)-BDNF(- His6 IgA none none (G3D)4- <TfR>8D3 yes 107 RGR)_(G3D)4- GGGS G3S_VH<TfR>8D3-CH1- EPKSC-His6 pre-pro(IgA)-BDNF(-RGR; His6 IgA opsin R81N GSG- <TfR>8D3 no 81 R81N)_opsin(L)-(G3D)4- (L) opsin(L)- G3S_VH<TfR>8D3-CH1-His6 (G3D)4- GGGS pre-pro(IgA)-BDNF(- His6 IgA none none (G3D)4- <TfR>8D3 no 108 RGR)_(G3D)4- GGGS G3S_VH<TfR>8D3-CH1-His6 pre-pro(IgA)-BDNF(- none IgA none None (G3D)4- <TfR>8D3 No 109 RGR)_(G3D)4- GGGS G3S_VH<TfR>8D3-CH1 pre-pro(IgA)-BDNF(- His6 IgA opsin none GSG- <TfR>8D3 No 79 RGR)_opsin(L)-(G3D)4- (L) opsin(L)- G3S_VH<TfR>8D3-CH1-His6 (G3D)4- GGGS

Example 6

(71) Generation of the Expression Plasmid for the BDNF-scFv Fusion-Polypeptides

(72) In order to obtain a BDNF(G4S)3-scFv-anti-IGF-1R antibody heavy chain-fusion-polypeptide, a plasmid for transient expression in HEK293 cells was constructed which harbored a chemically synthesized DNA fragment of a CDS which codes for a polypeptide with the following characteristics: the wild-type pre-pro-BDNF moiety deleted for the C-terminal RGR motif is fused at the C-terminus with a (G4S)3 linker followed by a scFv moiety of the human anti-human IGF-1R antibody; the scFv moiety of the anti-human IGF-JR antibody is build up of a VL domain, followed by a (G4S)4-GG linker, a VH region and a His6-tag; the amino acid sequence of the pre-pro-BDNF(-RGR)_(G4S)3_scFv-His6<IGF-1R>fusion protein is shown in SEQ ID NO: 78.

Example 7

(73) Transient Expression, Purification and Analytical Characterization of Polypeptides

(74) Transient Expression

(75) The polypeptides were generated by transient transfection of HEK293 cells (human embryonic kidney cell line 293-derived) cultivated in F17 Medium (Invitrogen Corp.). For transfection “293-Free” Transfection Reagent (Novagen) was used. The antibody and antibody fragment (Fab and scFv) comprising fusion-polypeptides were expressed from one, two or three different plasmids using an equimolar plasmid ratio upon transfection. Transfections were performed as specified in the manufacturer's instructions. The recombinant polypeptide-containing cell culture supernatants were harvested four to seven days after transfection. Supernatants were stored at reduced temperature until purification.

(76) General information regarding the recombinant expression of human immunoglobulins in e.g. HEK293 cells is given e.g. in Meissner, P., et al., Biotechnol. Bioeng. 75 (2001) 197-203.

(77) Purification

(78) a) GFP Fusion-Polypeptides were Purified Using a Two-Step Procedure Including a Protein a Chromatography and a Size Exclusion Chromatography on a Superdex 200™ Column

(79) GFP fusion-polypeptide containing culture supernatants were filtered. Thereafter the GFP fusion-polypeptides were captured by affinity chromatography using HiTrap MabSelectSuRe (GE Healthcare) equilibrated with PBS (1 mM KH.sub.2PO.sub.4, 10 mM NaHPO.sub.4, 137 mM NaCl, 2.7 mM KCl, pH 7.4). Unbound polypeptides were removed by washing with equilibration buffer, and the fusion-polypeptide was recovered with 0.1 M citrate buffer, pH 2.8. Immediately after elution the fractions were neutralized to pH 6.0 with 1 M Tris-base, pH 9.0.

(80) Size exclusion chromatography on Superdex 200™ (GE Healthcare, Uppsala, Sweden) was used as second purification step. The size exclusion chromatography was performed in 50 mM histidine buffer, 0.15 M NaCl, pH 6.8. The recovered GFP fusion-polypeptides were stored at −80° C.

(81) b) Histidine-Tagged Proteins were Purified Using a Two-Step Protocol Starting with an Immobilized Metal Ion Affinity Chromatography (IMAC) and Followed by a Size Exclusion Chromatography on a Superdex 75™ Column

(82) The histidine-tagged polypeptide-containing culture supernatants were adjusted with NaCl to a final NaCl concentration of 500 mM. The filtered culture supernatant was loaded onto a Ni-Sepharose™ 6 Fast Flow column pre-equilibrated with a NiA-buffer (50 mM TRIS, 300 mM NaCl, 5 mM imidazole containing an EDTA-free protease inhibitor cocktail tablet as specified in the manufacturer's instructions; EDTA-free Complete Mini Tablets; Roche Applied Science) at a flow of 1 ml/min using an ÄKTA explorer 100 system (GE Healthcare, Uppsala, Sweden). The column was washed with NiA-buffer until the UV reading reached back close to baseline. The histidine-tagged polypeptide was eluted with a 5 mM to 300 mM linear imidazole gradient in 50 mM TRIS and 500 mM NaCl, pH 8.0 in 10 column volumes.

(83) Size exclusion chromatography on Superdex 75™ (GE Healthcare, Uppsala, Sweden) was used as second purification step. The size exclusion chromatography was performed in 50 mM histidine buffer, 0.15 M NaCl, pH 6.8. The eluted histidine-tagged proteins were stored at −80° C.

(84) Analytical Characterization

(85) The protein concentrations of the purified polypeptides were determined by measuring the optical density (OD) at 280 nm, using the molar extinction coefficient calculated on the basis of the amino acid sequence. Purity and proper dimer formation of polypeptides were analyzed by SDS-PAGE in the presence and absence of a reducing agent (5 mM 1,4-dithiothreitol) and staining with Coomassie brilliant blue. Aggregate content of the Fc-fusion-polypeptide preparations was determined by high-performance SEC using a Superdex 200™ analytical size-exclusion column (GE Healthcare, Uppsala, Sweden). The integrity of the amino acid backbone of reduced polypeptides was verified by Nano Electrospray QTOF mass spectrometry after removal of N-glycans by enzymatic treatment with a combination of neuraminidase, O-glycanase and peptide-N-glycosidase F (Roche Applied Science, Mannheim, Germany).

(86) Determination of the BDNF Concentration in Culture Supernatant

(87) The concentration of wild-type BDNF, BDNF variants and BDNF containing fusion-polypeptides in culture supernatants was determined by semi-quantitative Western Blot analysis using recombinant human BDNF from Peprotech (catalog number: 450-02) as reference standard. A rabbit anti-BDNF antibody (Santa Cruz; catalog number: sc-20981) (first antibody) and a horseradish peroxidase conjugated sheep anti-rabbit antisera (diluted 1:5000, Roche Diagnostics GmbH, Germany) (secondary antibody) and enhanced chemiluminescence substrate (LUMI-Light plus Western Blotting substrate, Roche Diagnostics GmbH, Germany) was used for staining.

(88) The concentration of BDNF fusion-polypeptides was also determined with a BDNF ELISA using the BDNF Emax® ImmunoAssay Kit from Promega (catalog number: G7610) according to the instructions of the supplier.

Example 8

(89) In Vitro Functional Characterization

(90) Determination of the Biological Activity of Wild-Type GFP and GFP-Containing Fusion-Polypeptides

(91) The biological activity of purified wild-type GFP and GFP-containing fusion-polypeptides was monitored by its bioluminescent properties (GFP-specific fluorescence).

(92) Determination of the BDNF Binding Affinity Via Surface Plasmon Resonance (SPR, BIAcore)

(93) Amine coupling of around 750 resonance units (RU) of a capturing system (capturing mAb specific for human IgG, Jackson Immunoresearch) was performed on a CM5 chip at pH 4.5 using an amine coupling kit according to the manufacturer's manual (supplied by GE Healthcare, Uppsala, Sweden). Human Fc-tagged TrkB (R&D Systems, catalog number: 688-TK-100) was captured at a concentration of 5 μg/ml. Excess binding sites were blocked by injecting a human Fe mixture at a concentration of 1.25 μM (Biodesign, catalog number: 50175). Different concentrations of BDNF containing fusion-polypeptides ranging from 0.1 nM to 50 nM were passed with a flow rate of 10 μL/min through the flow cells at 298 K for 120 to 240 sec. The dissociation phase was monitored for up to 600 sec and triggered by switching from the sample solution to running buffer. The surface was regenerated by 1 min washing with a 100 mM phosphoric acid solution at a flow rate of 30 L/min. For all experiments HBS-P+ buffer supplied by GE Healthcare was chosen (10 mM HEPES ((4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid)), pH 7.4, 150 mM NaCl, 0.05% (v/v) Surfactant P20).

(94) Bulk refractive index differences were corrected for by subtracting the response obtained from a blank-coupled surface. Blank injections are also subtracted (double referencing).

(95) The equilibrium dissociation constant (Kd), defined as ka/kd, was determined by analyzing the sensorgram curves obtained with several different concentrations, using BIAevaluation 4.1 software package. The fitting of the data followed a suitable binding model.

(96) Determination of the BDNF Binding Affinity Via ELISA

(97) The binding properties of BDNF-containing fusion proteins were determined with a TrkB ELISA. Maxisorb plates were coated with 1 μg/mL of a TrkB-Fc fusion (R&D Systems) in PBS overnight at 4° C. After blocking the plate with PBSTC (PBS with 0.05% Tween-20 and 2% Chicken serum (Gibco)) for 1 h at RT and three washes with PBST, BDNF-containing fusion proteins or BDNF alone were added to the wells at concentrations of 15 to 500 ng/mL in PBSTC and incubated for 2 h at RT. After six washes, the wells were incubated with mouse-anti-BDNF antibody (clone 4F11.1A1, 1 μg/mL in PBSTC) and, after further washes, with anti-mouse-HRP antibody (1:10000 in PBSTC), both for 1 h at RT. After three washes with PBST, HRP activity was detected using ABTS substrate and photometric quantification at a wavelength of 492 nm.

(98) Determination of the Biological Activity of BDNF Containing Fusion Proteins Via a TrkB Reporter Gene Assay

(99) The biological activity of BDNF variants and BDNF-containing fusion proteins was determined with a TrkB-transfected CHO cell line containing a stably transfected luciferase reporter gene under the control of a SRE (serum-response element)-containing-promoter (CHO-K1-hTrkB/pSRE-Luc). The day before the experiments cell medium was changed from growth medium (Ham's F12 containing 10% FCS, 2 mM L-glutamine, 300 μg/mL G418 and 3 μg/mL puromycin) to the same medium without FCS for starvation. The next day, 10.sup.5 cells were seeded per well of a 96-well plate in 50 μL medium, then BDNF fusion proteins were added at concentrations between 0.02 nM and 115 nM, in 50 μL medium. After incubation for 4 h at 37° C., 7.5% CO.sub.2, cells were equilibrated for 30 min at RT and 100 μL of BrightGlo Luciferase Assay reagent (Promega) was added per well. Luminescence was read out after 5 minutes incubation using a Tecan plate reader (integration time 100 ms).

(100) Determination of the Biological Activity of BDNF Containing Fusion Proteins in a SH-SY5Y Neurite Outgrowth Assay

(101) The biological activity of BDNF variants and BDNF-containing fusion proteins was determined with a neurite outgrowth assay using human SH-SY5Y neuroblastoma cells. Briefly, SH-SY5Y cells were plated in a 96-well plate at 4000 cells per well in normal growth medium (Ham's F12, 1×non-essential amino acids (PAN), 10% FCS, 2 mM L-glutamine, 1×sodium pyruvate (PAN)) under addition of 10 μM retinoic acid (Sigma) to induce neuronal differentiation. After three days, medium was replaced with growth medium containing different concentrations of BDNF fusion proteins. After three additional days, cells were fixed using 4% paraformaldehyde in PBS for 10 min. at RT, washed, briefly permeabilized (0.1% Triton-X-100), blocked with 1% BSA in PBS and stained for anti-beta-tubulin immunoreactivity using the TuJ1 antibody (Covance) at a dilution of 1:1000 in PBS/i % BSA, followed by three washes and incubation with a Alexa-488-labeled anti-goat antibody (Invitrogen). Numbers of neurites were determined by fluorescence microscopy, evaluating one visual field per well.

(102) Determination of the Binding Activity of an Antibody/Antibody Fragment Containing Fusion Proteins Via FACS

(103) The binding activity of BDNF fusion proteins to the respective target receptors (transferrin receptor, IGF-1R) was determined by FACS. Cells expressing the respective receptor (mouse transferrin receptor: MEF-1 mouse embryonal fibroblasts; IGF-1R: 3T3 fibroblasts stably transfected with human IGF-1R) were harvested from their growth media, washed with PBS, and resuspended in FACS buffer (PBS+5% FCS; 100 μL containing 3×10.sup.5 cells per well of 96-well round-bottom plate). Primary antibody (depending on the BDNF fusion protein used, e.g., anti-human-Fab (Jackson ImmunoRescarch) or anti-His6 antibody (Roche)) was added at 1-10 μg/mL and cells incubated for two hours on ice. After three washes with FACS buffer, bound antibody was detected using PE-labeled secondary antibody (Jackson ImmunoResearch, 1:5000-1:10000) for one hour on ice. Cells were washed again and mean fluorescence measured on a FACS Canto cytometer (Becton-Dickinson).