Method for determining site having N-linked sugar chain added thereto or proportion of said addition

Abstract

A method for detecting a site which can be modified with an N-linked glycan chain and to which an N-linked glycan chain is actually linked in a glycoprotein; and a method for determining the state of an N-linked glycan chain addition at the site are provided. A glycoprotein having an N-linked glycan chain linked thereto is subjected to an N-linked glycan chain removal treatment with a peptide N-glycanase, subsequently a site capable of being modified with an N-linked glycan chain, in which an Asn residue has been changed to an Asp residue by the action of the peptide N-glycanase, is treated with an endo-type peptidase capable of recognizing an Asp residue to thereby produce peptide fragments, and subsequently the mass of the fragments is detected. In this manner, a site which can be modified with an N-linked glycan chain and to which an N-linked glycan chain is actually linked can be detected. Furthermore, the proportion or state of of the N-linked glycan chain addition at the site can be determined from the intensity of a signal generated upon the detection.

Claims

1. A method for detecting a site where an N-linked glycan chain is linked and/or a site where an N-linked glycan chain is not linked in a glycoprotein, comprising: (A) removing an N-linked glycan chain from an N-linked glycosylated glycoprotein and changing an asparagine residue to which the N-linked glycan chain is linked to an aspartic acid residue; (B) hydrolyzing the glycoprotein obtained in step (A) specifically for the asparagine residue or aspartic acid residue to obtain peptide fragments; (C) detecting the peptide fragments obtained in step (B); and (D) by comparing with a peptide fragment obtained by hydrolyzing in the same manner as step (B) a protein having the same amino acid sequence as the original N-linked glycosylated glycoprotein but not having a glycan chain, judging that a glycan chain is linked to an N-linked glycosylation site in the case different peptide fragments are present among the peptide fragments detected in step (C) and the different peptide fragments are presumed to contain an N-linked glycosylation site or amino acid residue adjacent thereto, and/or judging that a glycan chain is not linked to an N-linked glycosylation site in the case identical peptide fragments are present among the peptide fragments detected in step (C) and those identical peptide fragments are presumed to contain an N-linked glycosylation site.

2. The method according to claim 1, wherein step (A) is carried out using a peptide N-glycanase having deamidase activity.

3. The method according to claim 2, wherein the peptide N-glycanase having deamidase activity is peptide N-glycanase F or peptide N-glycanase A.

4. The method according to claim 1, wherein step (B) is carried out using an endo-type peptidase that acts by specifically recognizing an aspartic acid residue.

5. The method according to claim 4, wherein the endo-type peptidase that acts by specifically recognizing an aspartic acid residue is Asp-N or Glu-C.

6. The method according to claim 1, wherein the detection in step (C) is carried out using a mass spectrometer.

7. A method for detecting a site in a glycoprotein where an N-linked glycan chain is linked by using the method described in claim 1 and determining the proportion of N-linked glycan chain linked from signal strength at the time of detection.

8. A method for detecting a glycoprotein for which the site where an N-linked glycan chain is linked or the proportion of that addition has changed, comprising determining the sites where an N-linked glycan chain is linked, or the proportion thereof that have been linked, for a plurality of samples using the method according to claim 1, and comparing those results.

9. A glycoprotein detected using the method according to claim 8, in which the site where an N-linked glycan chain is linked or the proportion of that addition differs in a comparison between a sample obtained from a patient with a disease and a sample obtained from a healthy individual.

10. A method for detecting a disease, comprising determining the site where an N-linked glycan chain is linked or the proportion of that addition in the glycoprotein according to claim 9 that is present in a sample.

11. A method for determining the site where an N-linked glycan chain is linked or the proportion of that addition in a glycoprotein present in a pharmaceutical by using the method according to claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a drawing showing one example of the detection method of the present invention.

(2) FIG. 2 is a drawing showing the amino acid sequence of ribonuclease 1 (to be abbreviated as RNase 1) as predicted from gene sequence information, and three N-linked glycosylation sites.

(3) FIG. 3 is a drawing showing the results of separating human recombinant RNase 1 expressed by CHO cells by SDS-PAGE and the result of PNGase F treatment thereon.

(4) FIG. 4 is a drawing showing extraction of peptide fragments derived from RNase 1 in the mass spectrometry results of Example 1.

(5) FIG. 5 is a drawing showing the presence or absence of detection of a fragment derived from an N-linked glycosylation site in the form of a peptide containing Asn34, Asn76 or Asn88 over the full-length sequence of RNase 1 in Example 1.

(6) FIG. 6 is a drawing schematically representing the relationship between the presence or absence of glycosylation at a glycosylation site and the detected peptide in Example 1.

(7) FIG. 7 is a drawing showing extraction of peptide fragments derived from RNase 1 in the mass spectrometry results of Example 2.

(8) FIG. 8 is a drawing showing the presence or absence of detection of a fragment derived from an N-linked glycosylation site in the form of a peptide containing Asn34, Asn76 or Asn88 over the full-length sequence of RNase 1 in Example 2.

(9) FIG. 9 is a drawing showing one example of an embodiment of the present invention.

(10) FIG. 10 is a drawing showing the results of separating human interferon gamma expressed in Escherichia coli by SDS-PAGE and the result of PNGase F treatment thereon in Example 3.

(11) FIG. 11 is a drawing showing the results of separating human interferon gamma expressed in mammalian CHO-K1 cells by SDS-PAGE and the result of PNGase F treatment thereon in Example 3.

EXAMPLES

Example 1

Detection of N-Linked Glycosylation Site by Asp-N Digestion

(12) Recombinant human pancreas-specific RNase 1 used for analysis was obtained by introducing the full length gene of human pancreas-specific RNase 1 (SEQ ID NO. 1) into cultured Chinese hamster ovary cells (CHO-K1 cells) in accordance with ordinary methods and purifying the recombinant protein secreted into the medium by affinity chromatography. More specifically, a gene sequence encoding human pancreas-specific RNase 1 (SEQ ID NO. 1) was inserted into a pcDNA3.1-mycHis vector (Life Technologies Corp.) to prepare a mammalian cell expression vector. The prepared mammalian cell expression plasmid was inserted into CHO-K1 cells using Lipofectamine 2000 (Life Technologies Corp.) and the human pancreas-specific RNase 1 secreted into the medium was purified using an affinity column immobilized with anti-RNase 1 antibody.

(13) RNase 1 is a protein composed of the amino acid sequence shown in SEQ ID NO. 2, sequences thereof enabling the addition of an N-linked glycan chain (NXS/T) are present at three locations indicated with underlines in FIG. 2, and N-linked glycan chains are able to bind to the Asn residues at positions 34, 76 and 88 (referred to as Asn34, Asn76 and Asn88, respectively). Recombinant RNase 1 expressed in the CHO cells used was confirmed to have a proportion of glycan chains linked to Asn88 of 8% of all Asn88 according to the method described in International Publication WO 2013/187371. Sites where N-linked glycan chains are actually linked and the degree of glycan chain addition at all N-linked glycosylation sites were determined in this sample.

(14) The aforementioned recombinant human pancreas-specific RNase 1 was separated into a plurality of molecular weights by electrophoresis as indicated in the left lane (PNGase F()) of FIG. 3.

(15) On the other hand, after reductively degenerating 20 g of recombinant human pancreas-specific RNase 1, since the product of removing glycan chains by treating with PNGase F (New England Biolabs Inc.) resulted in a 15 kDa band that closely coincided with the molecular weight predicted from the amino acid sequence on electrophoresis (FIG. 3, right lane, PNGase F(+)), the aforementioned plurality of molecular weights was confirmed to be the result of differences in the numbers of glycan chains linked to the N-linked glycosylation sites.

(16) Following separation by SDS-PAGE, the fragments were transferred to a PVDF membrane in accordance with ordinary methods. The transferred PVDF membrane was washed with ultrapure water followed by staining with a solution containing Coomassie Brilliant Blue R250, cutting out the portion corresponding to recombinant RNase 1 from which glycan chains had been removed, and subjecting to treatment for mass spectrometry. The cut out PVDF membrane was added to Tris buffer (pH 8.0) containing an endo-type peptidase in the form of Asp-N followed by treating for 20 hours at 37 C. The sample solution was then treated with ZipTip C18 (Millipore Corp.) and eluted into a matrix solution and spotted on a plate. After allowing to air dry, mass spectrometry was carried out according to the peptide mass fingerprint method using MALDI-TOFMS (Voyager-DE STR, Applied Biosystems, Inc.).

(17) The left side of Table 1 shows peptide fragments formed during hydrolysis of the sequence of RNase 1 predicted from the gene sequence (SEQ ID NO. 2) with Asp-N and their theoretical masses. In addition, the right side of Table 1 shows peptide fragments formed during hydrolysis with Asp-N of a sequence in which all Asn residues at N-linked glycosylation sites, namely the Asn residues at positions 34, 76 and 88, changed to Asp residues (SEQ ID NO. 3) as a result of removing N-linked glycan chains by PNGase F from RNase 1, in which glycan chains were linked to all N-linked glycosylation sites, along with their theoretical masses. Theoretical mass indicates the value obtained by calculating the mass of those fragments in which all amino acid residues were not subjected to chemical modification. In addition, those peptide fragments that were able to be assigned as peptide fragments derived from RNase 1 by mass spectrometry were indicated with a O in the detection column while those that were unable to be assigned were indicated with an X.

(18) TABLE-US-00001 TABLE1 Peptidefragmentsobtainedbytreating Peptidefragmentsobtainedbydeglycosylating RNase1predictedfromgeneticallyencoded RNase1havingglycanchainslinkedtoallN- sequencewithAsp-Nandtheirtheoretical linkedglycosylationsiteswithPNGaseFand masses treatingwithAsp-N Theo- Amino Theo- Amino retical Detec- Acid retical Detec- AcidNo. AminoAcidSequence Mass tion No. AminoAcidSequence Mass tion 1-13 KESRAKKFQRQHM 1673.9016 O 1-13 KESRAKKFQRQHM 1673.9016 O (SEQIDNO.4) (SEQIDNO.4) 14-15 DS 221.0768 X 14-15 DS 221.0768 X 16-52 DSSPSSSSTYCNQMMRRRN 4243.9675 X 16-33 DSSPSSSSTYCNQMMRRR 2092.8957 O MTQGRCKPVNTFVHEPLV (SEQIDNO.9) (SEQIDNO.5 34-52 DMTQGRCKPVNTFVHEPLV 2171.0736 O (SEQIDNO.10) 53-82 DVQNVCFQEKVTCKNGQGN 3362.5021 O 53-75 DVQNVCFQEKVTCKNGQGNCYKS 2592.1639 O CYKSNSSMHIT (SEQIDNO.11) (SEQIDNO.6) 76-82 DSSMHIT(SEQIDNO.12) 790.3400 X 83-120 DCRLTNGSRYPNCAYRTSP 4336.0960 X 83-87 DCRLT(SEQIDNO.13) 607.2868 O KERHIIVACEGSPYVPVHF 88-120 DGSRYPNCAYRTSPKEPHIIVAC 3748.8111 X (SEQIDNO.7) EGSPYVPVHF (SEQIDNO.14) 121-125 DASVE(SEQIDNO.8) 520.2249 X 121-125 DRSVE(SEQIDNO.8) 520.2249 X 126-128 DST 322.1245 X 126-128 DST 322.1245 X

(19) In the case of removing glycan chains with PNGase F, Asn residues where the glycan chains are linked are changed to Asp residues by the deamidase activity of PNGase F. Thus, a glycoprotein from which the glycan chains have been removed with PNGase F is subjected to hydrolysis by Asp-N on the N-terminal side of the amino acid residues where the glycan chains were originally linked. On the other hand, since Asn residues where glycan chains were not originally linked are not changed to Asp residues and remain as Asn residues, they are not subjected to hydrolysis by Asp-N. In this manner, the manner in which peptides at N-linked glycosylation sites are fragmented varies according to the presence or absence of the addition of glycan chains.

(20) Table 2 shows the masses of peptide fragments described in Table 1 that were actually detected along with the amino acid sequences predicted when considering amino acid modification.

(21) TABLE-US-00002 TABLE2 Amino ModifiedAminoAcidResidue: Acid Detected Predicted TypeofModification No. Sequence Mass(mz) Mass(mz) (AdditionalMass) 1-13 KESPAKKFQRQHM 1673.8889 1672.8816 None (SEQIDNO.4) 1689.8641 1688.8568 Methionine:Oxidation(+16) 16-33 DSSPSSSSTYCNQMMRRR 2092.8999 2091.8927 None (SEQIDNO.9) 2108.9119 2107.9049 Methionine:Oxidation(+16) 2163.9243 2162.9171 Cysteine:Acrylamddation(+71) 2179.9227 2178.9154 Methionine:Oxidation(+16) Cysteine:Acrylamddation(+71) 2196.0021 2194.9949 Methionine:Oxidation 2(+32) Cysteine:Acrylamidation(+71) 34-52 DMTQGRCKPVNTFVHEPLV 2171.0681 2170.0680 None (SEQIDNO.10) 2187.0475 2186.0402 Methionine:Oxidation(+16) 2242.1006 2241.0933 Cysteine:AcrylAmidation(+71) 2258.0760 2257.0687 Methionine:Oxidation(+16) Cysteine:Acrylamddation(+71) 53-75 DVQNVCFQEKVTCKNGQGNCYKS 2663.1499 2662.1427 Cystein:Acrylamidation(+71) (SEQIDNO.11) 53-82 DVQNVCFQEKVTCKNGQGNCYKSNSSMHIT 3433.4771 3432.4699 Cysteine:Acrylamidation(+71) (SEQIDNO.6) 83-87 DCRLT(SEGIDNO.13) 607.3008 606.2936 None

(22) Mass spectrometry spectral data is shown in FIG. 4. A plurality of peptide sequences derived from RNase 1 was detected. In the drawing, the graph shown in the small frame is an enlarged view of the m/z range from 2000 to 3500, and among those peptides derived from RNase 1, amino acid numbers were described for those peaks of the masses (m/z: 2663, 1499, 3433, 4771) of the peptide fragment adjacent to Asn76 (53-75) and the peptide fragment containing Asn76 (53-82).

(23) FIG. 5 shows those peptide fragments formed from RNase 1 that were detected and not detected in the peptide sequence of RNase 1 with respect to portions containing N-linked glycosylation sites. Solid lines indicate detected peptide fragments while broken lines indicate peptide fragments that were not detected, and their respective amino acid numbers are shown below the lines.

(24) As shown in Table 1, in the fragment derived from the sequence containing Asn34, peptides were detected having masses corresponding to fragments 16-33 (DSSPSSSSTYCNQMMRRR) (SEQ ID NO. 9) and 34-52 (DMTQGRCKPVNTFVHEPLV) (SEQ ID NO. 10), indicating that an N-linked glycan chains were linked thereto, and no peptide indicating that no N-linked glycan chain was linked thereto was detected. In the fragment derived from the sequence containing Asn76, among those fragments indicating to have glycan chains linked thereto, the fragment 53-75 (DVQNVCFQEKVTCKNGQGNCYKS) (SEQ ID NO. 11) was detected, but the fragment 76-82 (DSSMHIT) (SEQ ID NO. 12) was not detected. In addition, fragment 53-82 (DVQNVCFQEKVTCKNGQGNCYKSNSSMHIT) (SEQ NO. 6), which was indicating that a glycan chain was not linked thereto, was also detected. In the fragment derived from the sequence containing Asn88, only the fragment 83-87 (DCRLT) (SEQ ID NO. 13) that was indicating to have a glycan chain linked thereto was detected, while the fragment 88-120 (DGSRYPNCAYRTSPKERHIIVACEGSPYVPVHF) (SEQ ID NO. 14) was not detected. In addition, the fragment 83-120 (DCRLTNGSRYPNCAYRTSPKERHIIVACEGSPYVPVHF) (SEQ ID NO. 7), which was indicating not to have a glycan chain linked thereto, was not detected.

(25) Among those fragments that were not detected, although the fragment containing Asn88 having amino acid Nos. 88-120 was present since fragment 83-87, to which a pair of glycan chains were linked, was detected, since it is a molecule having a comparatively large molecular size, it is thought to not have been detected due to limitations on the measuring instrument. In the case of this fragment, since the theoretical mass when not taking into consideration amino acid modification is 3748.8111, a fragment having a mass larger than this is thought to not be detected due to limitations on the measuring instrument. Accordingly, although peptide fragments corresponding to amino acid Nos. 16-52 (theoretical mass: 4243.9675) and amino acid nos. 83-120 (theoretical mass: 4336.0960) were not detected, since these are molecules having larger theoretical masses, whether or not they were initially not present or were unable to be detected despite being present due to limitations on the measuring instrument was unable to be confirmed. Furthermore, amino acid Nos. 76-82 are located in a region of high background noise and were not detected.

(26) In this manner, in the case of Asn34, Asn76 and Asn88, molecules were clearly present to which glycan chains were respectively linked. In addition, with respect to those molecules to which glycan chains were not linked, although a molecule to which a glycan chain was not linked was clearly present in Asn76, this was unable to be confirmed with respect to Asn34 and Asn88.

(27) In addition, as shown in FIG. 3, although the recombinant RNase 1 expressed in CHO cells was suggested to have a plurality of types of molecular species having different states of glycan chain addition, according to the aforementioned analysis method, a portion of those glycan chain addition states was clearly determined. Namely, since two types of peptide fragments having different masses were detected that were derived from the sequence of RNase 1 containing Asn76, as shown in FIG. 6, in the case of Asn76 at N-linked glycosylation sites, a mixed state was able to be confirmed by mass spectrometry that consisted of a peptide fragment having a glycan chain linked thereto and a glycan chain not having a glycan chain linked thereto.

Example 2

Determination of N-Linked Glycosylated Sites of RNase 1 Treated with Asp-N and Lys-C

(28) RNase 1 expressed in CHO cells used in Example 1 was carbamidomethylated in accordance with ordinary methods to obtain an analysis sample (carbamidomethylated RNase 1), and the analysis sample was fragmented into peptides with endo-type peptidases Asp-N and Lys-C, and then analyzed with a mass spectrometer. The following provides a description of the analysis procedure. After deglycosylating the carbamidomethylated RNase 1 with PNGase F in the same manner as Example 1, the fragments were separated by SDS-PAGE and transferred to a PVDF membrane in accordance with ordinary methods. After washing the transferred PVDF membrane with ultrapure water, the membrane was stained with a solution containing Coomassie Brilliant Blue R250, and the portion corresponding to the recombinant RNase 1 that had been removed of glycan chains was cut out and subjected to treatment for mass spectrometry. The cut out PVDF membrane was added to Tris buffer (pH 8.0) containing the endo-type peptidase Lys-C, and after treating for 20 hours at 37 C., the endo-type peptidase Asp-N was added followed by further treating for 20 hours at 37 C. The sample solution was treated with ZipTip C18 (Millipore Corp.) and eluted into a matrix solution and spotted on a plate. After allowing to air dry, mass spectrometry was carried out according to the peptide mass fingerprint method using MALDI-TOFMS (AXIMA-Confidence, Shimadzu Corp., measuring range: m/z 800-4000).

(29) The left side of Table 3 shows peptide fragments formed during hydrolysis of the sequence of RNase 1 predicted from the gene sequence with Asp-N and Lys-C and their theoretical masses, while the right side shows peptide fragments formed during hydrolysis with Asp-N and Lys-C after removing glycan chains with PNGase F from RNase 1 in the case of N-linked glycan chains linked to all N-linked glycosylation sites, along with their theoretical masses. Theoretical mass indicates the value obtained by calculating the mass of those fragments in which all amino acid residues were not subjected to chemical modification. Those peptide fragments that were able to be assigned as peptide fragments derived from RNase 1 by mass spectrometry were indicated with a O in the detection column while those that were unable to be assigned were indicated with an X.

(30) TABLE-US-00003 TABLE3 Peptidefragments Peptidefragmentsobtainedby obtainedbytreatingRNase1predicted deglycosylatingRNase1havingglycanchains fromgeneticallyencodedsequence linkedtoallN-linkedglycosylationsites withAsp-N/Lys-Candtheirtheoreticalmasses withPNGaseFandtreatingwithAsp-N/Lys-C Amino Amino Acid Amino Theoretical Detec- Acid Theoretical Detec- No. AcidSequence Mass tion No. AminoAcidSequence Mass tion 1 K 147.1128 X 1 K 147.1128 X 2-6 ESRAK 590.3256 X 2-6 ESRAK(SEQIDNO.15) 590.3256 X (SEQIDNO.15) 7 K 147.1128 X 7 K 147.1128 X 8-13 FQRQHM 846.4039 O 3-13 FQRQHM(SEQIDNO.16) 846.4039 O (SEQIDNO.16) 14-15 DS 221.0768 X 14-15 DS 221.0768 X 16-52 DSSPSSSSTYCNQMMRPRNM 4243.9675 X 16-33 DSSPSSSSTYCNQMNRRP 2092.8957 O TQGRCKPVNTFYEEPLV (SEQIDNO.9) (SEQIDNO.5) 34-52 DMTQGRCKPVNTFVHEPLV 2171.0736 O (SEQIDNO.10) 53-62 DVQNVCFQEK 1209.5568 O 53-62 DVQNVCFQEK 1209.5568 O (SEQIDNO.17) (SEQIDNO.17) 63-66 VTCK(SEQIDNO.18) 450.2381 X 63-66 VTCK(SEQIDNO.18) 450.2381 X 67-74 NGQGNCYK 883.3727 X 67-74 NGQGNCYK 883.372 X (SEQIDNO.19) (SEQIDNO.19) 75-82 SNSSMHIT 876.3880 X 75 S X (SEQIDNO.20) 76-82 DSSMHIT 790.3400 X (SEQIDNO.12) 83-102 DCRLTNGSRYPNCAYRTSPK 2302.0815 O 83-87 DCRLT(SEQIDNO.13) 607.2868 O (SEQIDNO.21) 88-102 DGSRYPNCAYRTSPK 1714.7966 O (SEQIDNO.23) 103-120 ERHIIVACEGSPYVPVHF 2110.0538 O 103-120 ERHIIVACEGSPYVPVHF 2110.0538 O (SEQIDNO.22) (SEQIDNO.22) 121-125 DASVE(SEQIDNO.8) 520.2249 X 121-125 DASVE(SEQIDNO.8) 520.2249 X 126-128 DST 322.1245 X 126-128 DST 322.1245 X

(31) Table 4 shows the masses of peptide fragments that were actually detected along with the amino acid sequences predicted when considering amino acid modification.

(32) TABLE-US-00004 TABLE4 Amino Detected Predicted Acid Mass Mass ModifiedAminoAcidResidue:Typeof No. Sequence (m/z) (m/z) Modification(AdditionalMass) 2-15 ESRAKKFQRQHMDS 1747.8600 1746.8527 None (SEQIDNO.27) 7-13 KFQRQHM(SEQIDNO.24) 974.4300 973.4227 None 7-15 KFQRQHMDS(SEQIDNO.25) 1192.5500 1191.5427 Methionine:Oxidation(+16) 8-13 FQRQHM(SEQIDNO.16) 846.4100 845.4027 None 16-33 DSSPSSSSTYCNQMMRRR 2093.0800 2092.0727 None (SEQIDNO.9) 2109.0600 2108.0527 Methiondne:Oxidation(+16) 34-52 DMTQGRCKPVNTFVHEPLV 2171.1600 2170.1527 None (SEQIDNO.10) 2187.2000 2186.1927 Methionine:Oxidation(+16) 53-62 DVQNVCFQEK 1209.5300 1208.5227 None (SEQIDNO.17) 1266.6800 1265.6727 Cysteine:Carbamidomethylation(+57) 67-82 NGQGNCYKSNSSMHIT 1813.8700 1812.8627 Cysteine:Carbamidomethylatdon(+57) (SEQIDNO.26) Methionine:Oxidation(+16) 83-102 DCRLTNGSRYPNCAYRTSPK 2301.2800 2302.0815 None (SEQIDNO.21) 2416.3600 2415.3527 Cysteine:Carbamidomethylation 2 (+114) 88-102 DGSRYPNCAYRTSPK 1714.8600 1713.8527 None (SEQIDNO.23) 1771.8900 1770.8827 Cysteine:Carbamidomethylation:(+57) 103-120 ERHIIVACEGSPYVPVHF 2053.1400 2052.0251 None (SEQIDNO.22) 2110.1900 2109.1827 Cysteine:Carbamidomethylation(+57)

(33) Furthermore, although amino acid nos. 2-15 (SEQ ID NO. 27), 7-13 (SEQ ID NO. 24), 7-15 (SEQ ID NO. 25) and 67-82 (SEQ ID NO. 26) are peptide fragments not described in Table 3, these are presumed to have been formed due to hydrolysis by Asp-N and Lys-C not proceeding completely.

(34) As described in Example 1, in the case of having removed glycan chains with PNGase F, Asn residues having a glycan chain linked thereto changed to Asp residues due to the deamidase activity of PNGase F. Thus, in a glycoprotein in which glycan chains have been removed with PNGase F, Asn residues originally having glycan chains linked thereto are changed to Asp residues, and the N-terminal sides thereof are subjected to decomposition by Asp-N. On the other hand, since Asn residues originally not having a glycan chain linked thereto do not change to Asp residues, they are not subjected to hydrolysis by Asp-N. In addition, since Lys-C is an enzyme that hydrolyzes peptide bonds on the C-terminal side of a lysine group, with the exception of the case in which a proline residue is linked to the C-terminal side, they are not involved in peptide fragmentation around glycosylation sites. Accordingly, in the case of using Lys-C in addition to Asp-N, the manner in which fragments containing N-linked glycosylation sites are fragmented varies according to the presence or absence of the addition of glycan chains. Moreover, since treatment with both Asp-N and Lys-C is expected to result in shorter detected peptide fragments than treatment with Asp-N alone, fragments that were unable to be detected due to limitations on the performance of the measuring instrument as in Example 1 are thought to be able to be determined.

(35) Mass spectrometry spectral data is shown in FIG. 7. A plurality of peptides derived from RNase 1 were detected. In the drawing, the graph shown in the small frame is an enlarged view of the m/z range from. 1500 to 2500, and among those peptides derived from RNase 1, amino acid numbers were described for those peaks of the mass of peptide fragment 88-102 containing Asn88 (m/z: 1714.8600, 1771.8900) and the mass of peptide fragment 83-102 (m/z: 2301.2800, 2416.3600).

(36) As is clear from Table 3, Asn34 and Asn88 were clearly determined to each have molecules where glycan chains are linked. In addition, with respect to Asn76, a fragment shown to have a glycan chain linked thereto in the form of amino acid Nos. 76-82 (SEQ ID NO. 12) was in an area of high background noise and was unable to be detected. On the other hand, with respect to molecules to which glycan chains are not linked, since the fragment containing Asn34 (amino acid Nos. 16-52) has a large theoretical mass (4243.9675) that is beyond the determination limit of the mass spectrometer (4000), whether it was unable to be detected despite being present due to limitations on the measuring instrument or was initially not present was unable to be confirmed. With respect to Asn76, since a fragment corresponding to amino acid Nos. 67-82, which was predicted to have been formed as a result of hydrolysis by endo-type peptidase Lys-C not proceeding completely, was detected as shown in Table 4, a molecule to which a glycan chain is not linked was clearly determined to be present. In addition, since amino acid Nos. 83-102 were detected as a fragment containing Asn88, a molecule to which a glycan chain is not linked was clearly determined to be present in Asn88.

(37) In addition, FIG. 8 shows those peptide fragments formed from RNase 1 that were detected and not detected with respect to portions containing N-linked glycosylation sites in the peptide sequence of RNase 1. Solid lines indicate detected peptide fragments while broken lines indicate peptide fragments that were not detected, and their respective amino acid numbers are shown below the lines.

(38) In the case of Example 1, the peptide fragment of amino acid Nos. 83-120 formed in the case a glycan chain is not linked to Asn88 of RNase 1 (DCRLTNGSRYPNCAYRTSPKERHIIVACEGSPYVPVHF) (SEQ ID NO. 7) and a peptide fragment from amino acid Nos. 88-120 among peptide fragments in the case a glycan chain is linked to Asn88 (DGSRYPNCAYRTSPKERHIIVACEGSPYVPVHF) (SEQ ID NO. 14) were unable to be detected. In contrast, since the molecular weights of peptide fragments obtained by digesting with Asn-N and Lys-C were lower, a peptide fragment of amino acid Nos. 83-102 formed in the case a glycan chain is not linked to Asn88 of RNase 1 (DCRLTNGSRYPNCAYRTSPK) (SEQ ID NO. 21) and a peptide fragment of amino acid Nos. 88-102 formed in the case a glycan chain is linked to Asn88 (DGSRYPNCAYRTSPK) (SEQ ID NO. 23) were able to be detected in Example 2. In Example 2, since two types of peptide fragments consisting of peptide fragments containing an asparagine residue at position 88, or an Asp residue to which it had changed, were detected at amino acid Nos. 83-102 and amino acid residues 88-102, the recombinant RNase 1 expressed in CHO cells used as a sample was confirmed to consist of a mixture of RNase 1 in a state in which a glycan chain is linked to Asn88 and a state in which a glycan chain is not linked.

(39) In this manner, the use of not only an enzyme that recognizes Asp residues, but also a different endo-type peptidase that does not recognize Asp residues or Asn residues, makes it possible to detect peptide fragments more precisely.

(40) Based on the results of Examples 1 and 2, recombinant RNase 1 expressed in CHO cells was clearly determined to consist of mixture of N-linked glycan chain addition states at Asn76 and Asn88.

Example 3

Analysis of Glycan Chain Addition State of Human Interferon Gamma

(41) Human interferon gamma (SEQ ID NO. 28), which is used in pharmaceuticals in the form of antivirus drugs or anticancer agents, has two N-linked glycosylation sites as predicted from its amino acid sequence (Asn25, Asn97).

(42) Commercially available research reagents in the form of recombinant human interferon gamma expressed in an Escherichia coli expression system (Peprotech Inc., Cat. No. 300-02) and recombinant human interferon gamma expressed in a. CHO cell expression system (Sino Biological Inc., Cat. No. 11725-HNAS) were respectively acquired and analyzed for the presence or absence of glycosylation at the N-linked glycosylation sites as unglycosylated recombinant human interferon gamma expressed in Escherichia coli and glycosylated recombinant human interferon gamma expressed in CHO cells.

(43) Results of LC-MS/MS Analysis of Recombinant Human interferon Gamma Expressed in Escherichia coli

(44) The results of separating 2 g of recombinant human interferon gamma expressed in Escherichia coli by SDS-PAGE electrophoresis in accordance with ordinary methods followed by CBB staining are shown in FIG. 10. Since the recombinant protein is not subjected to glycosylation in the Escherichia coli expression system, a single band was detected at about 16 Da (Lane 1). Even if this sample was treated with PNGase F, a band was obtained at the same location as that of the protein prior to treatment (Lane 2).

(45) On the other hand, 20 g of recombinant human interferon gamma expressed in Escherichia coli were separated by SOS-PAGE and then transferred to a PVDF membrane in accordance with ordinary methods. The transferred PVDF membrane was washed with ultrapure water and then stained with a solution containing Coomassie Brilliant Blue R250, after which the 16 kDa band corresponding to the recombinant human interferon gamma expressed in Escherichia coli was cut out and used as a sample for mass spectrometry. The cut out PVDF membrane was added to Tris buffer (pH 8.0) containing the endo-type peptidase Asp-N and treated for 20 hours at 37 C. After separating the sample solution with an HPLC system (Advance UHPLC System, Michrom Bioresources, Inc.), analysis was carried out using a mass spectrometer connected thereto (Thermo Scientific LTQ Orbitrap XL Mass Spectrometer, Thermo Fisher Scientific Inc.).

(46) Peptides having masses of 1963.01 Da, 1618.80 Da, 1503.78 Da and 1430.66 Da were detected in LC-MS/MS analysis (Table 5). Moreover, when the internal sequences were determined for each of these peaks by tandem MS, peptides exhibiting each of these masses were confirmed to be peptides corresponding to amino acid nos. 24-40, 62-75, 63-75 and 90-101 of the sequence shown in SEQ ID NO. 28.

(47) TABLE-US-00005 TABLE 5 Results of LC-MS/MS Analysis of Recombinant Human Interferon Gamma Expressed in Escherichia coli (Absence of PNGase Treatment) Detected Peptide Mass (Da) m/z Ionic Valence Sequence Range 1963.01 982.51 2 24-40 655.34 3 24-40 1618.80 810.41 2 62-75 540.61 3 62-75 1503.78 752.90 2 63-73 502.27 3 63-73 1430.66 716.34 2 90-101

(48) Next, a sample obtained by treating 20 g of recombinant human interferon gamma expressed in Escherichia coli with PNGase F was treated with the endo-type peptidase Asp-N using the same method as explained above and then analyzed by LC-MS/MS (Table 6).

(49) TABLE-US-00006 TABLE 6 Results of LC-MS/MS Analysis of Recombinant Human Interferon Gamma Expressed in Escherichia coli (Presence of PNGase Treatment) Detected Ionic Sequence Peptide Mass (Da) m/z Valence Range 2209.08 1105.55 2 2-20 737.37 3 2-20 1963.01 982.51 2 24-40 655.34 3 24-40 1618.80 810.41 2 62-75 540.61 3 62-75 1503.78 752.90 2 63-75 502.27 3 63-75 1770.92 886.47 2 76-89 591.31 3 76-89 1430.66 716.34 2 90-101 1315.63 658.82 2 91-101

(50) As is clear from Tables 5 and 6, peptide fragments detected in Table 5 (Results of LC-MS/MS Analysis of Recombinant Human Interferon Gamma Expressed in Escherichia coli (Absence of PNGase Treatment)) were also detected in Table 6. In particular, the results of Tables 5 and 6 completely coincided with respect to peptides corresponding to amino acid Nos. 24-40 and 90-101 of SEQ ID NO. 28, which are peptide fragments containing N-linked glycosylation sites. Accordingly, the recombinant human interferon gamma expressed in Escherichia coli can be judged to not have a glycan chain linked to the N-linked glycosylation site. This result agrees with the fact that glycan chains are not linked to recombinants expressed in Escherichia coli. In addition, this result indicates that Asn not having a glycan chain linked thereto is not converted to Asp during the course of treatment with PNGase F, and even in the case of having treated a glycoprotein that is partially subjected to N-linked glycosylation with PNGase F, Asn residues at N-linked glycosylation sites where a glycan chain is not linked are not converted to Asp residues, thereby indicating that Asn having an N-linked glycan chain linked thereto can be distinguished from Asn not having an N-linked glycan chain linked thereto by using the series of methods presented in the present invention.

(51) Results of LC/MS-MS Analysis of Recombinant Human Interferon Gamma Expressed in CHO Cells

(52) Since recombinant human interferon gamma obtained in a CHO cell expression system is subjected to N-linked glycosylation, in the case of not subjecting to deglycosylation treatment with PNGase F, a plurality of bands of about 14 kDa, about 17 kDa and about 20 kDa were detected in which molecular weights had shifted to higher molecular weights by an amount equal to the mass of the N-linked glycan chains on SDS-PAGE (FIG. 11). When this sample was treated with PNGase F, since the bands at about 17 kDa and about 20 kDa disappeared and converged into the band at about 14 kDa, the band at about 17 kDa was thought to represent two glycosylation sites, with a glycan chain linked to one of the sites, and the band at about 20 kDa was thought to represent two glycosylation sites, with a glycan chain linked to two of the sites. Namely, the analyzed sample was confirmed to be in a mixed state in which glycan chains were not linked to a portion of the glycosylation sites.

(53) 10 g of recombinant human interferon gamma expressed in CHO cells and treated with PNGase F were separated by SDS-PAGE and then transferred to a PVDF membrane in accordance with ordinary methods. The transferred PVDF membrane was washed in ultrapure water and then stained with a solution containing Coomassie Brilliant Blue R250, after which a 14 kDa band corresponding to recombinant human interferon gamma from which glycan chains had been removed was cut out and subjected to treatment for mass spectrometry. The cut out PVDF membrane was added to Tris buffer (pH 8.0) containing the endo-type peptidase Asp-N and treated for 20 hours at 37 C., after which the resulting sample was analyzed by LC-MS/MS. The sample solution was separated with an HPLC system (Advance UHPLC System, Michrom Bioresources, Inc.) connected to a mass spectrometer (Thermo Scientific LTQ Orbitrap XL Mass Spectrometer, Thermo Fisher Scientific Inc.).

(54) Peptides having masses of 1963.99 Da, 1618.81 Da, 1503.78 Da, 1770.92 Da, 866.40 Da, 1430.66 Da and 1315.63 Da were detected in LC-MS/MS analysis (Table 7). Moreover, when the internal amino acid sequences were determined for each of the detected peptide fragments having respective masses by tandem MS, peptides indicating each mass were confirmed to be peptides corresponding to amino acid Nos. 24-40, 62-75, 63-75, 76-89, 90-96, 90-101 and 91-101 shown in SEQ ID NO. 28. Since peptide fragments corresponding to amino acid Nos. 62-75, 63-75 and 76-89 in SEQ ID NO. 28 do not have internal N-linked glycosylation sites, these fragments are not subjected to changes in the amino acid sequence as a result of going through step (A). Since the amino acid sequence of this fragment coincides with the previously reported amino acid sequence of human interferon gamma, the analyzed sample was confirmed to be human interferon gamma. A fragment of 1963.99 Da was detected for a peptide containing an N-linked glycosylation site (Asn25). This peptide fragment was confirmed to consist of DDGTLFLGILKNWKEES (SEQ ID NO. 29) by tandem mass spectrometry. The sequence on the side of the N-terminal of this peptide was the sequence AspAsp, and the second Asp is an amino acid derived from the N-linked glycosylation site (Asn25). This peptide fragment (SEQ ID NO. 29) differs from a peptide fragment obtained by similarly hydrolyzing human interferon gamma not having a glycan chain, and since it contains an N-linked glycosylation (position 25), a glycan chain can be judged to be linked to Asp25 serving as the N-linked glycosylation site.

(55) Furthermore, in the case a glycan chain is linked to Asn of a sequence consisting of AsnAsn as in the present example, although a phenomenon is observed in which the peptide cannot be completely digested in step (B) even if using Asp-N, it is still possible to judge the presence or absence of the addition of a glycan chain. In the case this phenomenon is observed, the peptide can be completely digested by using endoproteinase Glu-C, which is known to recognize an Asp moiety and decompose the peptide on the C-terminal side thereof in phosphate buffer (pH 7.8).

(56) On the other hand, three fragments of 866.40 Da, 1430.66 Da and 1315.63 Da were detected for the peptide containing an N-linked glycosylation site (Asn97) or peptide adjacent thereto. Although an 866.40 Da peptide fragment was detected corresponding to amino acid nos. 90-96 of SEQ ID NO. 28, since this peptide fragment differed from the peptide fragment obtained by similarly hydrolyzing human interferon gamma not having a glycan chain with Asp-N, and did not contain an amino acid residue adjacent to the N-linked glycosylation site (position 97), an N-linked glycan chain can be judged to be linked to the Asn residue at position 97 of SEQ ID NO. 28. On the other hand, although peptide fragments having masses of 1430.66 Da and 1315.63 Da corresponding to amino acid nos. 90-101 and 91-101 of SEQ ID NO. 28 were detected, since these were the same peptides as the peptide fragment obtained by similarly hydrolyzing human interferon gamma not having a glycan chain, and contained an N-linked glycosylation site (position 97), an N-linked glycan chain can be judged to not be linked to the Asn residue at position 97 of SEQ ID NO. 28. On the basis of these two results, both an Asn residue where a glycan chain is linked and an Asn residue where a glycan chain is not linked can be confirmed to exist for the Asn residue at position 97 of recombinant human interferon gamma expressed in CHO cells. This coincided with the sample not subjected to deglycosylation treatment by PNGase F (Lane 2) that indicated bands corresponding to a plurality of molecular weights in FIG. 11.

(57) TABLE-US-00007 TABLE 7 Results of LC-MS/MS Analysis of Recombinant Human Interferon Gamma Expressed in CHO Cells Detected Peptide Ionic Mass (Pa) m/z Valence Sequence Range 1963.99 983.00 2 24-40 (Asn25.fwdarw.Asp25) 655.67 3 24-40 (Asn25.fwdarw.Asp25) 1618.81 810.41 2 62-75 540.61 3 62-75 1503.78 752.90 2 63-75 502.27 3 63-75 1770.92 591.31 3 76-89 (Met77, oxidation) 866.40 434.21 2 90-96 1430.66 716.8381 2 90-101 1315.63 658.82 2 91-101