J. Biochem. 107, 292-297 (1990)
Structure of Recombinant Human Interleukin 5 Produced by Chinese Hamster Ovary Cells Yoshiharu Minamitake, Shiho Kodama, Toyoko Katayama, Hideki Adachi, Shoji Tanaka, and Masafumi Tsujimoto
Received for publication, September 22, 1989
The complete peptide map of purified recombinant human interleukin 5 (rhIL-5) was determined to verify its primary structure, glycosylation sites, and disulfide bonding structure. Each peptide fragment generated by Achromobacter protease I (API) digestion was purified and characterized by amino acid analysis and amino acid sequence analysis. After digestion with API, we could identify all the peptides which were expected from human IL-5 cDNA sequence. The analyses of sulfhydryl content in rhIL-5 molecule and disulfide-containing peptide obtained from API digestion indicated that active form of rhEL-5 existed as an antiparallel dimer linked by two pairs of Cys-44 and Cys-86. In addition, we concluded that Thr-3 and Asn-28 were glycosylated. The results indicate that primary structure of rhIL-5 is highly homogeneous and observed heterogeneity is due to the difference in the content of carbohydrate.
Interleukin 5 (IL-5) (also referred to T cell replacing factor) is a lymphokine initially described as a factor that induces terminal differentiation of late-developing B cells to immunoglobulin- secreting cells (1). Recently, murine IL-5 was purified to homogeneity (2) and cDNA for murine IL-5 was cloned and sequenced (3). Subsequently, human IL-5 cDNA was also isolated using murine IL-5 cDNA as a probe (4). However, little is known about the structure of human IL-5 because adequate amount of human IL-5 protein was not available. In our previous work, we have purified and characterized recombinant human IL-5 (rhIL-5) expressed in Chinese hamster ovary (CHO) cells transfected with a plasmid comprising the genome sequence for human IL-5 (5). We and others reported that rhIL-5 existed as a dimer and dimer formation via disulfide bond(s) might be essential for the biological activity of human IL-5 (5, 6). Therefore, determination of the location of disulfide bond(s) is a prerequisite for understanding the molecular basis of the action of human IL-5. In this paper, we have characterized the structure of rhIL-5 including the location of disulfide bonds. Our data indicate that the primary structure of purified rhIL-5 is highly homogeneous and rhIL-5 forms an antiparallel dimer linked by two disulfide bonds. We also detect the glycosylation sites of the molecule. MATERIALS AND METHODS Materials—Recombinant rhIL-5 expressed in Chinese hamster ovary cells was purified by procedures previously reported (5). The preparation was shown to be biologically active in an in vitro assay for IgM secretion with a specific Abbreviations: EL, interleukin; rhIL-5, recombinant human interleukin 5; CHO, Chinese hamster ovary; 2-PDS, dithiopyridine; DTT, dithiothreitol; TFA, trifluoroacetic acid; API, Achromobacter protease I.
292
activity of 1.7 X103 units/mg. A single lot of rhIL-5 preparation was employed in all experiments. Achromobacter protease I (isolated from a strain of Achromobacter lyticus) was purchased from Wako Pure Chemical Industries (Osaka), endoglycosidase F (isolated from a strain of Flavobacterium meningosepticum) from Boehlinger Mannheim (F.R.G.), neuraminidase (isolated from a strain of Arthrobacter ureafaciens) from Nacalai Tesque (Kyoto), and endo-a-iV-acetylgalactosaminidase (isolated from a strain of Alcarigenes sp. F-1906) from Seikagaku Kogyo (Tokyo). Chemicals were reagent grade and used without further purification. Reagents and solvents used for sequence analysis were obtained from Applied Biosystems (Foster City, Calif.). Protein Concentration—The concentration of the stock solution of purified rhIL-5 used in this study was determined by amino acid analysis after 24 or 72 h hydrolysis. Otherwise mentioned, molar quantities based on dimer structure of rhIL-5 were used to present the contents of the protein. Sulfhydryl Titration—Sulfhydryl concentration was determined by the method reported by Grassetti et al using 7.06 X lOVM-cm as the molar extinction coefficient of 2-thiopyridone at 343 nm (7). In brief, to a solution of rhIL-5 (4.0nmol) in 400//I of 50 mM citrate buffer (pH 5.2) containing 5 M urea, was added dithiopyridine (2PDS) in ethanol at a protein to 2-PDS ratio of 1 : 20 (mol/ mol). The mixture was allowed to stand for 15 min at room temperature and measured its absorbance at 343 nm. As a sulfhydryl-positive control, we prepared reduced rhIL-5. In brief, rhIL-5 (4.0 nmol) was treated with 0.5 M dithiothreitol (DTT) in 0.1 M phosphate buffered saline (PBS) (pH 7.2) containing 5 M urea. The reaction was allowed to proceed for 2 h at 37*C under nitrogen atmosphere. The reaction was terminated by injection of the mixture onto a C-4 HPLC column (YMC-Pack AP-802, 5 , 4.6x150 mm) previously equilibrated with 0.1% J. Biochem.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
Suntory Institute for Biomedical Research, Shimamoto-cho, Mishima-gun, Osaka 618
293
Structure of rhIL-5
CUJCN(%)
A215
Time(min)
Fig. 1. Characterization of purified rhIL-5 employing a C-4 reversed-phase HPLC column. Protein of 15 ^g was loaded for analysis. The column was equilibrated in 0.1% TFA and eluted with linear acetonitrile gradient from 0 to 80% in 0.1% TFA. Vol. 107, No. 2, 1990
(2 nmol) was further treated with endo- a -iV-acetylgalactosaminidase (40 /x\J) at 37"C for 10 min. Both neuraminidase-treated peptide and neuraminidase and endo-a-Nacetylgalactosaminidase-treated peptide were analyzed their sensitivity to these enzymes using reversed-phase HPLC. Amino Acid Sequence—Sequence analysis of the peptides obtained from API digestion was performed using Applied Biosystem 477A gas-phase sequencer according to the manufacturer's program. The resulting phenylthiohydantoin derivatives were identified by reversed-phase HPLC (9) with an Applied Biosystem 120A on-line system. Amino Acid Analysis—The peptides were hydrolyzed in 6N HC1 containing 0.1% phenol at 110'C for 24 h in an evacuated and sealed tube, and the amino acids were analyzed using Hitachi amino acid analyzer, Model 835. The Cys residue in the disulfide-containing peptide was determined as cysteic acid after hydrolysis of the peptide previously treated with performic acid (10). RESULTS AND DISCUSSION Characterization of Purified rhIL-5—Recombinant human IL-5 was purified employing serial column chromatographies on Matrex Blue A, DEAE-Sepharose, phenyl-Sepharose and Sephacryl S200 as described previously (5). Molecular heterogeneity of purified rhIL-5 was observed by SDS-polyacrylamide gel electrophoresis analysis. While two major components of Mr around 40,000 were detected under non-reducing conditions, Mr of rhIL-5 under reducing coditions was determined to be 22,000 and 20,000. These results indicate that rhIL-5 exists in a dimeric form and that disulfide bond(s) are responsible for dimerization. When the purified rhIL-5 was subjected to reversed-phase HPLC (C-4), rhIL-5 was eluted as a single peak at 48% of acetonitrile (Fig. 1). Peptide Mapping of rhIL-5—Next, rhIL-5 was digested with API, which cleaved C-terminal side of Lys residue, to analyze its intact structure and disulfide bonding scheme. A typical peptide map of rhIL-5 is shown in Fig. 2. The peak assignments shown correspond to the cleavage sites of API for rhIL-5 in Fig. 3 and are based on amino acid analysis and sequence analysis of the purified peptides. Amino acid
CH3CN(%)
A215
S
JO
Time(min)
Fig. 2. HPLC peptide mapping of rhEL-5 on a C-18 reversedphase column. After API digestion, rhIL-5 was loaded onto a column and fractionated as described in "MATERIALS AND METHODS."
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
trifluoroacetic acid (TFA). The reduced protein was eluted with a linear gradient of acetonitrile (0 to 80% in 0.1% TFA) over 50 min at a flow rate of 1 ml/min. A peak eluted at retention time of 38.5 min was collected, freeze dried and the sample was immediately subjected to the sulfhydryl tdtration as described above. Peptide Mapping—Digestion of rhIL-5 was performed using Achromobacter Protease I (API) at a substrate to enzyme ratio of 1,000: 1 (w/w) at 3TC for 2 h. Typical procedure is as follows. To a solution of rhIL-5 (2.0 nmol) in 50 ft\ of 0.1 M PBS (pH 7.2) containing 5 M urea was added API. The reaction was allowed to proceed for 2 h at 3TC. After acidification with 1 M HC1, the mixture was subjected to a C-18 HPLC column (YMC-Pack A-302, 5 pm, 4.6x150 mm) previously equilibrated with 0.1% TFA, and the peptides were eluted with linear acetonitrile gradient from 0 to 80% in 0.1% TFA over 50 min at a flow rate of 1 ml/min. Detection of the Disulfide-Containing Peptide—Aliquots of the mixture of the API digest were incubated with DTT at 37'C for 2 h. Reduction was terminated by injection of the mixture onto a reversed-phase HPLC column. The peptides were separated under the same conditions employed for the peptide mapping. Preparation of Disulfide Homodimers—Oxidati.ve dimerization was achieved by du Vigneaud's procedure using potassium ferricyanide as an oxidizing agent (8). In brief, to a peptide solution (1 nmol) in 20 mM Tris-HCl buffer (300 ftl, pH 7.5) containing 5 M urea was added potassium ferricyanide (10 eq) in water. The mixture was allowed to stand for 1 h at room temperature. The peptide generated was analyzed by reversed-phase HPLC under the same conditions used for the peptide mapping. Detection of the Carbohydrate-Containing Peptides—1) The peptide having N-linked carbohydrates: To determine N-linked glycosylation sites, purified rhIL-5 was treated with endoglycosidase F as described previously (5). In brief, rhIL-5 (2 nmol) was treated with the enzyme (125 mU) at 37*C for 12 h. After isolation by gel permeation column chromatography, endoglycosidase F-treated rhIL-5 was digested with API and analyzed. 2) The peptide having O-linked carbohydrates: Peptide la (3 nmol) was treated with neuraminidase (3 mU) at 37*C for 1 h. One-third of the reaction mixture was withdrawn for the analysis, the rest
294
Y. Minamitake et aL A-l
1 10 20 H-Ile-Pro-Thr-Glu-Ile-Pro-Thr-Ser-Ala-Leu-Val-Lys-Glu-Thr-Leu-Ala-Leu-Leu-Ser-ThrI
la
:
1|
;
;
lb-
30 40 Hia-Arg-Thr-Leu-Leu-Ile-Ala-Asn-Glu-Thr-Leu-Arg-Ile-Pro-Val-Pro-Val-Hia-Lys-Asn—>
—>
—*•
—>
—*•
—>
— *
-->
—*•
—»•
—*•
—*•
—*•
—>•
—>
—>
—>
—>
—HI—>
2b 2cA-3 •
50 60 His-Gln-Leu-Cys-Thr-Glu-Glu-Ile-Phe-Gln-Gly-Ile-Gly-Thr-Leu-Glu-Ser-Gln-Thr-Val3
A-4 A-5 70 ' 80 Gln-Gly-Gly-Thr-Val-Glu-Arg-Leu-Phe-Lys-Asn-I^u-Ser-Leu-Ile-Lys-Lys-Tyr-Ile-Aap-
A-690 100 Gly-Gln-Lys-Lys-Lys-Cya-Gly-Glu-Glu-Arg-Arg-Arg-Val-Asn-Gln-Phe-Leu-Asp-Tyr-Leu6 Fig. 3. Primary structure of rhIL-5 produced by CHO cells. Digestive peptides obtained in this study are numbered according to their separation as shown in Figs. 2 and 4. Dashed arrows indicate the amino acids 110 115 not detected by sequence analysis. Digestive Gln-Glu-Phe-Leu-Gly-Val-Met-Asn-Thr-Glu-Trp-Ile-Ile-Glu-Ser-OH peptides (designated as A-l to A-6) expected \ from human IL-5 cDNA sequence are also shown.
sequence of rhIL-5 shown in the figure was predicted from human IL-5 cDNA (4) and N-terminal end of purified rhIL-5 was determined as described below. Digested peptides were separated and amino acid compositions and sequences were determined (Tables IS and IIS, respectively). When A-l peptide was recovered from the reversedphase HPLC (C-18) column, it was separated into two peptides, la and lb. From the data of amino acid composition analyses, both peptides were considered to be derived from the N-terminal end of the molecule. Both peptides shared the same sequence, and no other peptide derived
from the N-terminal end could be detected. We could not detect Thr-3 in both peptides by Edman degradation, suggesting majority of the 3rd threonine from the Nterminal end was O-glycosylated. In contrast, we could detect Thr-7 in the same experiments. Separation of A-l peptide into two peptides might be due to the heterogeneity of O-linked oligosaccharides. Even after treatment with endoglycosidase F, we detected at least three bands by two-dimensional gel electrophoresis, futher supporting the presence of O-linked oligosaccharides (data not shown). It should be noted that Thr-3 was the only O-glycosylated amino acid which we could detect in this study. J. Biochem.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
A-2
295
Structure of rhLL-5 CH3CN(%)
A215
30
Time(min) Time(min)
Fig. 4. Detection of peptide (3+6) as a disulfide-containlng fragments. The mixture of API digest was treated with 0.5 M DTT for 2 h at 37*C under nitrogen atmosphere. After the reaction, the mixture was loaded onto a C-18 reversed-phase column and fractionated as described in "MATERIALS AND METHODS.' Arrow indicates the retention time of untreated peptide (3 + 6).
A-2 peptide was also separated into three peptides, 2a, 2b, and 2c. Unfortunately, we could not determine the sequences of 2b and 2c peptides entirely because of their small quantities. However, little difference of amino acid composition was detected among these peptides. Although 2b and 2c peptides had relatively high contents of Glu residues, it was likely that they contained two Glu residues in their sequence, because no more Glu residue was expected in these two peptides from human IL-5 cDNA sequence. While we could detect the presence of Asn residue in all three peptides by amino acid composition analyses, Asn-28 was detected only in 2c peptide by amino acid sequence analyses. These results suggest that while 2a and 2b are glycosylated at Asn-28, 2c was not. In fact, A-2 peptide contains possible N-linked glycosylation site at Asn-28. As for A-5 peptide, two peptides, 5a and 5b, were obtained by reversed-phase HPLC. While 5b contained two Lys residues per peptide, only one Lys residue was detected in 5a. Sequence analysis confirmed the presence of Lys-77 at the N-terminal end of 5b peptide. It is most likely that differential cleavage by API might result in occurrence of 5a and 5b peptides, because rhIL-5 has 78 Lys-"Lys sequence in the molecule (Fig. 3). As shown in Fig. 2, A-3 and A-6 peptides were eluted at the same retention time, suggesting that these peptides were linked via disulfide bond(s) (see below). In fact, both A-3 and A-6 peptides contain Cys residues at positions 44 and 86, respectively. Analysis of peptide 6, isolated after treatment of peptide (3 + 6) with DTT, indicated that the peptide was derived from C-terminal end of the molecule. It should be noted that we could detect one Ser residue, which should correspond to Ser-115 locating at the Cterminal end of rhIL-5 molecule. These results indicate that we could identify all the peptides which are expected from human IL-5 cDNA sequence. Because the recovery of each peptide from the column was high ( — 90%), the data presented here indicate that primary structure of rhIL-5 is highly homogeneous and heterogeneity observed on SDS-polyacrylamide gel electrophoresis is most likely due to the difference in the content of carbohydrate. Location of the Disulfide Bonds—It has been suggested Vol. 107, No. 2, 1990
Fig. 5. Detection of a peptide containing N-linked glycosylation sites. A: Peptide map obtained from API degestion of intact rhIL-5. B: Peptide map obtained from API digestion of endoglycosidase F treated rhIL-5.
that IL-5 exists as a dimer through intermolecular disulfide bond formation (11, 12). To determine the disulfide structure of rhIL-5, we first examined the sulfhydryl content in the molecule. The key feature to determine the correct disulfide structure is supposed to prevent the oxidation of the sulfhydryl group(s) or disulfide rearrangement in the course of the analytical procedures. Sulfhydryl titration was performed using 2-PDS which exhibited quantitative SH content even at the low pH (i.e. pH 3.4). The protein was dissolved in a de-gased and nitrogen-purged acidic buffer to prevent the possible oxidation of sulfhydryl function or rearrangement of disulfide bonds. While sulfhydryl content of reduced rhIL-5 (monomer) was calculated to be 1.98 SH per monomer unit, intact rhIL-5 (dimer) showed only 0.02 SH per dimer unit. These results indicate the absence of free sulfhydryl groups in intact rhIL-5 molecule. Next we intended to isolate the disulfide-containing fragment(s) after API digestion. To eliminate the possible disulfide bond scrambling, the digestion was performed at pH 7.2 and in a relatively short period (2 h). To survey the disulfide-containing peptides, the mixture of API digest was treated with DTT. As shown in Fig. 4, we found peptide (3+6) was the only fragment reduced by DTT yielding two peptides, 3 and 6. Sequence analysis confirmed the generation of A-3 and A-6 peptides after treatment of peptide (3 + 6) with DTT. These results indicated that peptide (3 + 6) shown in Fig. 2 was the disulfide heterodimer between A-3 and A-6 peptides. To eliminate the possibility that peak (3 + 6) was a mixture of homodimers of (3 + 3) and (6 + 6), we prepared these dimers and compared the elution profile. Both disulfide homodimers obtained by the conventional disulfide-forming reaction showed different retention times from peptide (3 + 6) (36.8 min); Le. peptide (3 + 3) (32.0 min), peptide (6 + 6) (38.2 min). Moreover, no peak corresponding to (3 + 3) or (6+6) was observed in the chromatogram of the mixture of API digest. These results indicate that rhIL-5 exists as an antiparallel dimer linked with two disulfide bonds (Cys44-Cys86' and Cys44'-Cys86). Detection of N-Linked Glycosylation Sites—Human IL-5 has two putative N-linked glycosylation sites (4). These sites are located on A-2 (Le. Asn-28) and A-4 (Le. Asn-71) peptides. Although it was reported that the N-linked carbohydrate moiety did not play an essential role in the biological activity of EL-5 (5, 11), it is important to identify the glycosylated sites.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
20
296
antiparallel dimer. Because no information on the structure of native human IL-5 is available at present, the data presented here provide an initial characterization of the molecular bases of the biological activity of human EL-5. REFERENCES 1. Takatsu, K.( Tanaka, K., Tominaga, A., Kumahara, Y., & Hamaoka, T. (1980) J. Immunol. 125, 2646-2653 2. Takatsu, K., Harada, N., Hara, Y., Takahama, Y., Yamada, G., Dobaahi, K., & Hamaoka, T. (1985) J. Immunol 134, 382-389 3. Kinaahi, T., Harada, N., Severinson, E., Tanabe, T., Sideras, P., Konishi, M., Azuma, C, Tominaga, A., Bergstedt-Lindqvist, S., Takahaahi, M., Matsuda, F., Yaoita, Y., Takatsu, K., & Honjo, T. (1986) Nature 324, 70-73 4. Aruma, C, Tanabe, T., Konishi, M., Kinashi, T., Noma, T., Matanda, F., Yaoita, Y., Takatsu, K., Hammarstrom, L., Edvard-Smith, C.I., Severinson, E., & Honjo, T. (1986) Nucleic Add. Res. 14, 9149-9158 5. Tsujimoto, M., Adachi, H., Kodama, S., Tsuruoka, N., Yamada, Y., Tanaka, S., Mita, S., & Takatsu, K. (1989) J. Biochem. 106, 23-28 6. Tavemier, J., Devos, R., Van der Heyden, J., Hauquier, G., Bauden, R., Fache, I., Kawasbima, E., Vandekerckhove, J., Contrevas, R, & Fiers, W. (1989) DNA 8, 491-501 7. Grassetti, D.R. & Murray, J.F., Jr. (1967) Arch. Biochem. Biophys. 119, 41-49 8. Hope, D.B., Murti, V.V.S., & Du Vigneaud, V. (1961) J. BioL Chem. 237, 1563-1566 9. Hewick, R.M., Hunkapiller, M.W., Hood, L.E., & Dreyer, W.J. (1981) J. Biol. Chem. 256, 7990-7997 10. More, S. (1963) J. BioL Chem. 238, 235-237 11. Takatsu, K., Tominaga, A., Harada, N., Mita, S., Matsumoto, M., Takahashi, T., Kikuchi, Y., & Yamaguchi, N. (1988) Immunol. Rev. 102, 107-135 12. Yokota, T., Arai, N., de Virues, J., Spits, H., Banchereau, J., Zlotnik, A., Rennick, D., Haward, M., Takebe, Y., Miyatake, S., Le, F., & Arai, K.-I. (1988) Immunol Rev. 108, 137-188 13. Tsuruoka, N., Funakoshi, K., Kodama, S., & Tsujimoto, M. (1990) Cell Immunol, in press
J. Biochem.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
Sequence analysis suggested that only A-2 peptide was N-glycosylated because while only a small portion of Asn-28 was detected, Asn-71 was detectable almost entirely. To confirm the N-linked glycosylation site, we compared the peptide map of rhIL-5 treated with or without endoglycosidase F. As shown in Fig. 5, only peak 2 corresponding to A-2 peptide was shifted to later retention time. No proteolytic activity in endoglycosidase F preparation was observed. In contrast, endoglycosidase F treatment had no effect on the retention time of A-4 peptide, further suggesting that the peptide was not AT-glycosylated. We also performed similar experiments employing another preparation of rhIL-5 and obtained the same conclusions. To determine whether A-l peptide contains 0glycosylated amino acid or not, effect of endo-a-Nacetylgalactosaminidase treatment on retention time of la peptide was examined. Treatment of la peptide with neuraminidase followed by endo-a-N-acetylgalactosaminidase resulted in the appearance of new peak having later retention time suggesting that la peptide contained 0linked glycosylation site(s) (presumably Thr-3) (data not shown). These results may explain our previous results that even after treatment with endoglycosidase F, molecular weight of rhIL-5 was determined to be 18,000 by SDSpolyacrylamide gel electrophoresis analysis under reducing conditions, which was higher than that of human IL-5 expected from its amino acid sequence. We are now investigating the structure and biological significance of the carbohydrate in rhIL-5. It has been suggested that the formation of intermolecular disulfide bond is essential for the biological activity of IL-5 (5, 6). In our recent data, monomeric form of IL-5, which was blocked disulfide bond formation, had no ability to bind to its cell surface receptors (23). Thus it is important to determine the location of the disulfide bonds. Our data indicate that both Cys residues in the molecule participate in the formation of disulfide bonds, forming an
Y. Minamitake et al.
Structure of rhIL-5
297
Supplemental Materials Table IS. Amioo acid nalytit of API genented pepodet from rhIL-5 Numtwrt In oarerthtus are the ttworatical valua (or th» amlno adds. Celculatsd values are based- on mol** ol wudrw (1a. 1b. 2a, 2b, 2c 3+6. 3,6, and 4) or gtydn* (5a and 5b). . A-1-
am] no add
5a
1b 1.00 (1) 1.90 1.19 1.12 2.03
(2) (1) (1) (2)
1.89 1.10 1.12 2.13
(2) (1) (1) (2)
0.90 (1) 0.85 11) 1.16 (2) 1.00 (1)
0.94 0.98 1.89 1.00
(1) (1) (2) (1)
1.00 (1) 2.00 (2)
0.93 (1)
1.13 (1)
0.92 (1)
1.02
(1)
0.99
1.01
(1)
0.92 (1)
1.00 (1)
1.00
12
Total Position recovery (»)*
5b
1-12
12 1-12 3.1 nmol (87 %)
0.90
(1)
(1)
0.99 (1) 0.96 (1)
1.00 (1)
0.84 (1) 1.00 (1) 1.84 (2)
6 71-76 3.1 nmol (87 %)
78-83 77-63 3.2 nmol (89 %)
- A-2 2b
2a
CyaO3H Aap Thr Ser Glu Pro Gly Ma v«i Htt lie L*u Tyr Phe Ly» Bis Trp Ar9
(1) (4) (1) (2) (2)
1.53 4.11 1.44 2.81 1.81
(1) (4) (1) (2) (2)
1.48 4.27 1.20 2.48
(1) (4) (1) (2)
2.40
(2)
2.00 1.96
(2) (2)
2.10 2.03
(2) (2)
2.38 2.37
(2) (2)
2.11 6.00
(2) (6)
2.27 6.00
(2) (6)
(2)
1.07 1.91
(1) (2)
1.20 1.96
(1) (2)
1.08 (1) 2.36 (2)
1.99
(2)
1.93
(2)
1.89 (2)
27
(2)
13-39
3 . 1 nmol ( 8 7 %)
Recovery
(2) (4) (5) (2) (15)
N.D. 1.04 3.81 0.98 8.11
(1) (1) (4)
6.42
(6)
4.07
3.93 0.72 3.30 6.00 1.04 3.85 2.70 1.00 N.D. 3.86
(4) (1) (4) (6) (1) (4) (3) (1) (1) (4)
1.98 4.08 4.66 2.34 14.92
27
27
13-39
13-39
-A-6 —
—A-3 — 3
3+6
1.05 3.85 0.94 2.12 2.01
Total
Poaition
2c
(1) (8)
N.D. 3.07 1.00 1.07 7.17
(1) (3) (1) (1) (7)
(4)
2.18
(2)
1.98 (2)
1.98 0.92 1.27 3.00 1.03 2.03 1.80
(2) (1) (2) (3) (1) (2) (2)
N.D. 2.93
(1) (3)
1.99 (2) 3.00 (31 1.94 (2) 0.91 (1) 0.97 (1) 1.02
(1)
31
63 (40-70) (84-115)
40-70
3 . 1 nmol ( 8 7 %)
1 . 2 nmol (86 %)•*
32 84-115 1.1 (79
nmol %)•"
Recovery (%) la calculated relative to tht dole* of the starting u t e r i a l d . 8 nmol of diner rhIL-5) . calculated relative to the moles of the DTT-treated API digesta (0.7 nmol of dimer rhIL-5)
Table US.(continued) Table US. Sequence imlyiU of API genented peptidet from mIL-5.
Papudea
Peptides Cycle # l 2 3 4 5 6 7 8 9 10 11 12 Initial yield(l)
1a I P T E I P T 3 A L V K
480pmol 300 ND 151 268 245 66' 29 204 204 153 125 66.2
1b I P T E I P T S
115 poolN 133 L MD S 4 L 55 I 56 K 12 7 A 30 L 18 V 14 It 8 62.7
4 253 pool Y 384 I 36 D 213 G 121 Q 90 K
5a
5b
336paolK 224 Y 168 I 239 D 251 G 107 Q
K
31.6
86.6
761pmol 697 704 308 595 596 456
52.1
»•
2a
Cycle* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
E T L A L L S T H R T L L I A H E T L R I P V P V H K
IB 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Initial yield(%)
332D»O1
90 343 406 299 344 42 44 30 49 30 112 166 99 13 ND 67 19
NO
E T L A L L 3 T H R T L L I A N E T L
16paol 5 6 16 14 15 6 2 HQ NQ
E T L A L L 3 T H
NO
T L L
4 7 3 3 HD 2
NO 5
36 46 40 39 45 43 7 15
46.9
2C#
51.2
R
I A N E
3
15 psolN 5 H 7 Q 10 L 6 C 9 T NQ E E NO I HQ NQ F NO Q G NO 5 I NQ G NQ T 2 L E NO 3 0 T V Q G G T V E R L F K
63.2
6
U5pmol 14 67 69 ND 16 41 52 25 26 36 30 25 34 6 33 17
HO 16 5 6 14 11 17 3 4 6
NO 7 2
NO
36 .5
NO: Not quantitatively determined ND: Not detected •: Analysis did not r«sch to C-ttrainal end due to their small quantities.
Vol. 107, No. 2, 1990
K
474D001
r. 456 HD c G E E R R R V N Q
r L D Y L Q E F L G V H N T E H I I E S
37. 9
363 260 294 53 72 86 329 283 291 270 259 109 141 226 206 170 174 166 129 145 146 121 24 73 NQ 60 77 30 »Q
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
Thr Ser Glu Pro Gly Ala val lie Leu Tyr Lya
amlno acid
. A-5
Asp
Table IS. (continued)
Structure of Recombinant Human Interleukin 5 Produced by Chinese Hamster Ovary Cells Yoshiharu Minamitake, Shiho Kodama, Toyoko Katayama, Hideki Adachi, Shoji Tanaka, and Masafumi Tsujimoto
Received for publication, September 22, 1989
The complete peptide map of purified recombinant human interleukin 5 (rhIL-5) was determined to verify its primary structure, glycosylation sites, and disulfide bonding structure. Each peptide fragment generated by Achromobacter protease I (API) digestion was purified and characterized by amino acid analysis and amino acid sequence analysis. After digestion with API, we could identify all the peptides which were expected from human IL-5 cDNA sequence. The analyses of sulfhydryl content in rhIL-5 molecule and disulfide-containing peptide obtained from API digestion indicated that active form of rhEL-5 existed as an antiparallel dimer linked by two pairs of Cys-44 and Cys-86. In addition, we concluded that Thr-3 and Asn-28 were glycosylated. The results indicate that primary structure of rhIL-5 is highly homogeneous and observed heterogeneity is due to the difference in the content of carbohydrate.
Interleukin 5 (IL-5) (also referred to T cell replacing factor) is a lymphokine initially described as a factor that induces terminal differentiation of late-developing B cells to immunoglobulin- secreting cells (1). Recently, murine IL-5 was purified to homogeneity (2) and cDNA for murine IL-5 was cloned and sequenced (3). Subsequently, human IL-5 cDNA was also isolated using murine IL-5 cDNA as a probe (4). However, little is known about the structure of human IL-5 because adequate amount of human IL-5 protein was not available. In our previous work, we have purified and characterized recombinant human IL-5 (rhIL-5) expressed in Chinese hamster ovary (CHO) cells transfected with a plasmid comprising the genome sequence for human IL-5 (5). We and others reported that rhIL-5 existed as a dimer and dimer formation via disulfide bond(s) might be essential for the biological activity of human IL-5 (5, 6). Therefore, determination of the location of disulfide bond(s) is a prerequisite for understanding the molecular basis of the action of human IL-5. In this paper, we have characterized the structure of rhIL-5 including the location of disulfide bonds. Our data indicate that the primary structure of purified rhIL-5 is highly homogeneous and rhIL-5 forms an antiparallel dimer linked by two disulfide bonds. We also detect the glycosylation sites of the molecule. MATERIALS AND METHODS Materials—Recombinant rhIL-5 expressed in Chinese hamster ovary cells was purified by procedures previously reported (5). The preparation was shown to be biologically active in an in vitro assay for IgM secretion with a specific Abbreviations: EL, interleukin; rhIL-5, recombinant human interleukin 5; CHO, Chinese hamster ovary; 2-PDS, dithiopyridine; DTT, dithiothreitol; TFA, trifluoroacetic acid; API, Achromobacter protease I.
292
activity of 1.7 X103 units/mg. A single lot of rhIL-5 preparation was employed in all experiments. Achromobacter protease I (isolated from a strain of Achromobacter lyticus) was purchased from Wako Pure Chemical Industries (Osaka), endoglycosidase F (isolated from a strain of Flavobacterium meningosepticum) from Boehlinger Mannheim (F.R.G.), neuraminidase (isolated from a strain of Arthrobacter ureafaciens) from Nacalai Tesque (Kyoto), and endo-a-iV-acetylgalactosaminidase (isolated from a strain of Alcarigenes sp. F-1906) from Seikagaku Kogyo (Tokyo). Chemicals were reagent grade and used without further purification. Reagents and solvents used for sequence analysis were obtained from Applied Biosystems (Foster City, Calif.). Protein Concentration—The concentration of the stock solution of purified rhIL-5 used in this study was determined by amino acid analysis after 24 or 72 h hydrolysis. Otherwise mentioned, molar quantities based on dimer structure of rhIL-5 were used to present the contents of the protein. Sulfhydryl Titration—Sulfhydryl concentration was determined by the method reported by Grassetti et al using 7.06 X lOVM-cm as the molar extinction coefficient of 2-thiopyridone at 343 nm (7). In brief, to a solution of rhIL-5 (4.0nmol) in 400//I of 50 mM citrate buffer (pH 5.2) containing 5 M urea, was added dithiopyridine (2PDS) in ethanol at a protein to 2-PDS ratio of 1 : 20 (mol/ mol). The mixture was allowed to stand for 15 min at room temperature and measured its absorbance at 343 nm. As a sulfhydryl-positive control, we prepared reduced rhIL-5. In brief, rhIL-5 (4.0 nmol) was treated with 0.5 M dithiothreitol (DTT) in 0.1 M phosphate buffered saline (PBS) (pH 7.2) containing 5 M urea. The reaction was allowed to proceed for 2 h at 37*C under nitrogen atmosphere. The reaction was terminated by injection of the mixture onto a C-4 HPLC column (YMC-Pack AP-802, 5 , 4.6x150 mm) previously equilibrated with 0.1% J. Biochem.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
Suntory Institute for Biomedical Research, Shimamoto-cho, Mishima-gun, Osaka 618
293
Structure of rhIL-5
CUJCN(%)
A215
Time(min)
Fig. 1. Characterization of purified rhIL-5 employing a C-4 reversed-phase HPLC column. Protein of 15 ^g was loaded for analysis. The column was equilibrated in 0.1% TFA and eluted with linear acetonitrile gradient from 0 to 80% in 0.1% TFA. Vol. 107, No. 2, 1990
(2 nmol) was further treated with endo- a -iV-acetylgalactosaminidase (40 /x\J) at 37"C for 10 min. Both neuraminidase-treated peptide and neuraminidase and endo-a-Nacetylgalactosaminidase-treated peptide were analyzed their sensitivity to these enzymes using reversed-phase HPLC. Amino Acid Sequence—Sequence analysis of the peptides obtained from API digestion was performed using Applied Biosystem 477A gas-phase sequencer according to the manufacturer's program. The resulting phenylthiohydantoin derivatives were identified by reversed-phase HPLC (9) with an Applied Biosystem 120A on-line system. Amino Acid Analysis—The peptides were hydrolyzed in 6N HC1 containing 0.1% phenol at 110'C for 24 h in an evacuated and sealed tube, and the amino acids were analyzed using Hitachi amino acid analyzer, Model 835. The Cys residue in the disulfide-containing peptide was determined as cysteic acid after hydrolysis of the peptide previously treated with performic acid (10). RESULTS AND DISCUSSION Characterization of Purified rhIL-5—Recombinant human IL-5 was purified employing serial column chromatographies on Matrex Blue A, DEAE-Sepharose, phenyl-Sepharose and Sephacryl S200 as described previously (5). Molecular heterogeneity of purified rhIL-5 was observed by SDS-polyacrylamide gel electrophoresis analysis. While two major components of Mr around 40,000 were detected under non-reducing conditions, Mr of rhIL-5 under reducing coditions was determined to be 22,000 and 20,000. These results indicate that rhIL-5 exists in a dimeric form and that disulfide bond(s) are responsible for dimerization. When the purified rhIL-5 was subjected to reversed-phase HPLC (C-4), rhIL-5 was eluted as a single peak at 48% of acetonitrile (Fig. 1). Peptide Mapping of rhIL-5—Next, rhIL-5 was digested with API, which cleaved C-terminal side of Lys residue, to analyze its intact structure and disulfide bonding scheme. A typical peptide map of rhIL-5 is shown in Fig. 2. The peak assignments shown correspond to the cleavage sites of API for rhIL-5 in Fig. 3 and are based on amino acid analysis and sequence analysis of the purified peptides. Amino acid
CH3CN(%)
A215
S
JO
Time(min)
Fig. 2. HPLC peptide mapping of rhEL-5 on a C-18 reversedphase column. After API digestion, rhIL-5 was loaded onto a column and fractionated as described in "MATERIALS AND METHODS."
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
trifluoroacetic acid (TFA). The reduced protein was eluted with a linear gradient of acetonitrile (0 to 80% in 0.1% TFA) over 50 min at a flow rate of 1 ml/min. A peak eluted at retention time of 38.5 min was collected, freeze dried and the sample was immediately subjected to the sulfhydryl tdtration as described above. Peptide Mapping—Digestion of rhIL-5 was performed using Achromobacter Protease I (API) at a substrate to enzyme ratio of 1,000: 1 (w/w) at 3TC for 2 h. Typical procedure is as follows. To a solution of rhIL-5 (2.0 nmol) in 50 ft\ of 0.1 M PBS (pH 7.2) containing 5 M urea was added API. The reaction was allowed to proceed for 2 h at 3TC. After acidification with 1 M HC1, the mixture was subjected to a C-18 HPLC column (YMC-Pack A-302, 5 pm, 4.6x150 mm) previously equilibrated with 0.1% TFA, and the peptides were eluted with linear acetonitrile gradient from 0 to 80% in 0.1% TFA over 50 min at a flow rate of 1 ml/min. Detection of the Disulfide-Containing Peptide—Aliquots of the mixture of the API digest were incubated with DTT at 37'C for 2 h. Reduction was terminated by injection of the mixture onto a reversed-phase HPLC column. The peptides were separated under the same conditions employed for the peptide mapping. Preparation of Disulfide Homodimers—Oxidati.ve dimerization was achieved by du Vigneaud's procedure using potassium ferricyanide as an oxidizing agent (8). In brief, to a peptide solution (1 nmol) in 20 mM Tris-HCl buffer (300 ftl, pH 7.5) containing 5 M urea was added potassium ferricyanide (10 eq) in water. The mixture was allowed to stand for 1 h at room temperature. The peptide generated was analyzed by reversed-phase HPLC under the same conditions used for the peptide mapping. Detection of the Carbohydrate-Containing Peptides—1) The peptide having N-linked carbohydrates: To determine N-linked glycosylation sites, purified rhIL-5 was treated with endoglycosidase F as described previously (5). In brief, rhIL-5 (2 nmol) was treated with the enzyme (125 mU) at 37*C for 12 h. After isolation by gel permeation column chromatography, endoglycosidase F-treated rhIL-5 was digested with API and analyzed. 2) The peptide having O-linked carbohydrates: Peptide la (3 nmol) was treated with neuraminidase (3 mU) at 37*C for 1 h. One-third of the reaction mixture was withdrawn for the analysis, the rest
294
Y. Minamitake et aL A-l
1 10 20 H-Ile-Pro-Thr-Glu-Ile-Pro-Thr-Ser-Ala-Leu-Val-Lys-Glu-Thr-Leu-Ala-Leu-Leu-Ser-ThrI
la
:
1|
;
;
lb-
30 40 Hia-Arg-Thr-Leu-Leu-Ile-Ala-Asn-Glu-Thr-Leu-Arg-Ile-Pro-Val-Pro-Val-Hia-Lys-Asn—>
—>
—*•
—>
—*•
—>
— *
-->
—*•
—»•
—*•
—*•
—*•
—>•
—>
—>
—>
—>
—HI—>
2b 2cA-3 •
50 60 His-Gln-Leu-Cys-Thr-Glu-Glu-Ile-Phe-Gln-Gly-Ile-Gly-Thr-Leu-Glu-Ser-Gln-Thr-Val3
A-4 A-5 70 ' 80 Gln-Gly-Gly-Thr-Val-Glu-Arg-Leu-Phe-Lys-Asn-I^u-Ser-Leu-Ile-Lys-Lys-Tyr-Ile-Aap-
A-690 100 Gly-Gln-Lys-Lys-Lys-Cya-Gly-Glu-Glu-Arg-Arg-Arg-Val-Asn-Gln-Phe-Leu-Asp-Tyr-Leu6 Fig. 3. Primary structure of rhIL-5 produced by CHO cells. Digestive peptides obtained in this study are numbered according to their separation as shown in Figs. 2 and 4. Dashed arrows indicate the amino acids 110 115 not detected by sequence analysis. Digestive Gln-Glu-Phe-Leu-Gly-Val-Met-Asn-Thr-Glu-Trp-Ile-Ile-Glu-Ser-OH peptides (designated as A-l to A-6) expected \ from human IL-5 cDNA sequence are also shown.
sequence of rhIL-5 shown in the figure was predicted from human IL-5 cDNA (4) and N-terminal end of purified rhIL-5 was determined as described below. Digested peptides were separated and amino acid compositions and sequences were determined (Tables IS and IIS, respectively). When A-l peptide was recovered from the reversedphase HPLC (C-18) column, it was separated into two peptides, la and lb. From the data of amino acid composition analyses, both peptides were considered to be derived from the N-terminal end of the molecule. Both peptides shared the same sequence, and no other peptide derived
from the N-terminal end could be detected. We could not detect Thr-3 in both peptides by Edman degradation, suggesting majority of the 3rd threonine from the Nterminal end was O-glycosylated. In contrast, we could detect Thr-7 in the same experiments. Separation of A-l peptide into two peptides might be due to the heterogeneity of O-linked oligosaccharides. Even after treatment with endoglycosidase F, we detected at least three bands by two-dimensional gel electrophoresis, futher supporting the presence of O-linked oligosaccharides (data not shown). It should be noted that Thr-3 was the only O-glycosylated amino acid which we could detect in this study. J. Biochem.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
A-2
295
Structure of rhLL-5 CH3CN(%)
A215
30
Time(min) Time(min)
Fig. 4. Detection of peptide (3+6) as a disulfide-containlng fragments. The mixture of API digest was treated with 0.5 M DTT for 2 h at 37*C under nitrogen atmosphere. After the reaction, the mixture was loaded onto a C-18 reversed-phase column and fractionated as described in "MATERIALS AND METHODS.' Arrow indicates the retention time of untreated peptide (3 + 6).
A-2 peptide was also separated into three peptides, 2a, 2b, and 2c. Unfortunately, we could not determine the sequences of 2b and 2c peptides entirely because of their small quantities. However, little difference of amino acid composition was detected among these peptides. Although 2b and 2c peptides had relatively high contents of Glu residues, it was likely that they contained two Glu residues in their sequence, because no more Glu residue was expected in these two peptides from human IL-5 cDNA sequence. While we could detect the presence of Asn residue in all three peptides by amino acid composition analyses, Asn-28 was detected only in 2c peptide by amino acid sequence analyses. These results suggest that while 2a and 2b are glycosylated at Asn-28, 2c was not. In fact, A-2 peptide contains possible N-linked glycosylation site at Asn-28. As for A-5 peptide, two peptides, 5a and 5b, were obtained by reversed-phase HPLC. While 5b contained two Lys residues per peptide, only one Lys residue was detected in 5a. Sequence analysis confirmed the presence of Lys-77 at the N-terminal end of 5b peptide. It is most likely that differential cleavage by API might result in occurrence of 5a and 5b peptides, because rhIL-5 has 78 Lys-"Lys sequence in the molecule (Fig. 3). As shown in Fig. 2, A-3 and A-6 peptides were eluted at the same retention time, suggesting that these peptides were linked via disulfide bond(s) (see below). In fact, both A-3 and A-6 peptides contain Cys residues at positions 44 and 86, respectively. Analysis of peptide 6, isolated after treatment of peptide (3 + 6) with DTT, indicated that the peptide was derived from C-terminal end of the molecule. It should be noted that we could detect one Ser residue, which should correspond to Ser-115 locating at the Cterminal end of rhIL-5 molecule. These results indicate that we could identify all the peptides which are expected from human IL-5 cDNA sequence. Because the recovery of each peptide from the column was high ( — 90%), the data presented here indicate that primary structure of rhIL-5 is highly homogeneous and heterogeneity observed on SDS-polyacrylamide gel electrophoresis is most likely due to the difference in the content of carbohydrate. Location of the Disulfide Bonds—It has been suggested Vol. 107, No. 2, 1990
Fig. 5. Detection of a peptide containing N-linked glycosylation sites. A: Peptide map obtained from API degestion of intact rhIL-5. B: Peptide map obtained from API digestion of endoglycosidase F treated rhIL-5.
that IL-5 exists as a dimer through intermolecular disulfide bond formation (11, 12). To determine the disulfide structure of rhIL-5, we first examined the sulfhydryl content in the molecule. The key feature to determine the correct disulfide structure is supposed to prevent the oxidation of the sulfhydryl group(s) or disulfide rearrangement in the course of the analytical procedures. Sulfhydryl titration was performed using 2-PDS which exhibited quantitative SH content even at the low pH (i.e. pH 3.4). The protein was dissolved in a de-gased and nitrogen-purged acidic buffer to prevent the possible oxidation of sulfhydryl function or rearrangement of disulfide bonds. While sulfhydryl content of reduced rhIL-5 (monomer) was calculated to be 1.98 SH per monomer unit, intact rhIL-5 (dimer) showed only 0.02 SH per dimer unit. These results indicate the absence of free sulfhydryl groups in intact rhIL-5 molecule. Next we intended to isolate the disulfide-containing fragment(s) after API digestion. To eliminate the possible disulfide bond scrambling, the digestion was performed at pH 7.2 and in a relatively short period (2 h). To survey the disulfide-containing peptides, the mixture of API digest was treated with DTT. As shown in Fig. 4, we found peptide (3+6) was the only fragment reduced by DTT yielding two peptides, 3 and 6. Sequence analysis confirmed the generation of A-3 and A-6 peptides after treatment of peptide (3 + 6) with DTT. These results indicated that peptide (3 + 6) shown in Fig. 2 was the disulfide heterodimer between A-3 and A-6 peptides. To eliminate the possibility that peak (3 + 6) was a mixture of homodimers of (3 + 3) and (6 + 6), we prepared these dimers and compared the elution profile. Both disulfide homodimers obtained by the conventional disulfide-forming reaction showed different retention times from peptide (3 + 6) (36.8 min); Le. peptide (3 + 3) (32.0 min), peptide (6 + 6) (38.2 min). Moreover, no peak corresponding to (3 + 3) or (6+6) was observed in the chromatogram of the mixture of API digest. These results indicate that rhIL-5 exists as an antiparallel dimer linked with two disulfide bonds (Cys44-Cys86' and Cys44'-Cys86). Detection of N-Linked Glycosylation Sites—Human IL-5 has two putative N-linked glycosylation sites (4). These sites are located on A-2 (Le. Asn-28) and A-4 (Le. Asn-71) peptides. Although it was reported that the N-linked carbohydrate moiety did not play an essential role in the biological activity of EL-5 (5, 11), it is important to identify the glycosylated sites.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
20
296
antiparallel dimer. Because no information on the structure of native human IL-5 is available at present, the data presented here provide an initial characterization of the molecular bases of the biological activity of human EL-5. REFERENCES 1. Takatsu, K.( Tanaka, K., Tominaga, A., Kumahara, Y., & Hamaoka, T. (1980) J. Immunol. 125, 2646-2653 2. Takatsu, K., Harada, N., Hara, Y., Takahama, Y., Yamada, G., Dobaahi, K., & Hamaoka, T. (1985) J. Immunol 134, 382-389 3. Kinaahi, T., Harada, N., Severinson, E., Tanabe, T., Sideras, P., Konishi, M., Azuma, C, Tominaga, A., Bergstedt-Lindqvist, S., Takahaahi, M., Matsuda, F., Yaoita, Y., Takatsu, K., & Honjo, T. (1986) Nature 324, 70-73 4. Aruma, C, Tanabe, T., Konishi, M., Kinashi, T., Noma, T., Matanda, F., Yaoita, Y., Takatsu, K., Hammarstrom, L., Edvard-Smith, C.I., Severinson, E., & Honjo, T. (1986) Nucleic Add. Res. 14, 9149-9158 5. Tsujimoto, M., Adachi, H., Kodama, S., Tsuruoka, N., Yamada, Y., Tanaka, S., Mita, S., & Takatsu, K. (1989) J. Biochem. 106, 23-28 6. Tavemier, J., Devos, R., Van der Heyden, J., Hauquier, G., Bauden, R., Fache, I., Kawasbima, E., Vandekerckhove, J., Contrevas, R, & Fiers, W. (1989) DNA 8, 491-501 7. Grassetti, D.R. & Murray, J.F., Jr. (1967) Arch. Biochem. Biophys. 119, 41-49 8. Hope, D.B., Murti, V.V.S., & Du Vigneaud, V. (1961) J. BioL Chem. 237, 1563-1566 9. Hewick, R.M., Hunkapiller, M.W., Hood, L.E., & Dreyer, W.J. (1981) J. Biol. Chem. 256, 7990-7997 10. More, S. (1963) J. BioL Chem. 238, 235-237 11. Takatsu, K., Tominaga, A., Harada, N., Mita, S., Matsumoto, M., Takahashi, T., Kikuchi, Y., & Yamaguchi, N. (1988) Immunol. Rev. 102, 107-135 12. Yokota, T., Arai, N., de Virues, J., Spits, H., Banchereau, J., Zlotnik, A., Rennick, D., Haward, M., Takebe, Y., Miyatake, S., Le, F., & Arai, K.-I. (1988) Immunol Rev. 108, 137-188 13. Tsuruoka, N., Funakoshi, K., Kodama, S., & Tsujimoto, M. (1990) Cell Immunol, in press
J. Biochem.
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
Sequence analysis suggested that only A-2 peptide was N-glycosylated because while only a small portion of Asn-28 was detected, Asn-71 was detectable almost entirely. To confirm the N-linked glycosylation site, we compared the peptide map of rhIL-5 treated with or without endoglycosidase F. As shown in Fig. 5, only peak 2 corresponding to A-2 peptide was shifted to later retention time. No proteolytic activity in endoglycosidase F preparation was observed. In contrast, endoglycosidase F treatment had no effect on the retention time of A-4 peptide, further suggesting that the peptide was not AT-glycosylated. We also performed similar experiments employing another preparation of rhIL-5 and obtained the same conclusions. To determine whether A-l peptide contains 0glycosylated amino acid or not, effect of endo-a-Nacetylgalactosaminidase treatment on retention time of la peptide was examined. Treatment of la peptide with neuraminidase followed by endo-a-N-acetylgalactosaminidase resulted in the appearance of new peak having later retention time suggesting that la peptide contained 0linked glycosylation site(s) (presumably Thr-3) (data not shown). These results may explain our previous results that even after treatment with endoglycosidase F, molecular weight of rhIL-5 was determined to be 18,000 by SDSpolyacrylamide gel electrophoresis analysis under reducing conditions, which was higher than that of human IL-5 expected from its amino acid sequence. We are now investigating the structure and biological significance of the carbohydrate in rhIL-5. It has been suggested that the formation of intermolecular disulfide bond is essential for the biological activity of IL-5 (5, 6). In our recent data, monomeric form of IL-5, which was blocked disulfide bond formation, had no ability to bind to its cell surface receptors (23). Thus it is important to determine the location of the disulfide bonds. Our data indicate that both Cys residues in the molecule participate in the formation of disulfide bonds, forming an
Y. Minamitake et al.
Structure of rhIL-5
297
Supplemental Materials Table IS. Amioo acid nalytit of API genented pepodet from rhIL-5 Numtwrt In oarerthtus are the ttworatical valua (or th» amlno adds. Celculatsd values are based- on mol** ol wudrw (1a. 1b. 2a, 2b, 2c 3+6. 3,6, and 4) or gtydn* (5a and 5b). . A-1-
am] no add
5a
1b 1.00 (1) 1.90 1.19 1.12 2.03
(2) (1) (1) (2)
1.89 1.10 1.12 2.13
(2) (1) (1) (2)
0.90 (1) 0.85 11) 1.16 (2) 1.00 (1)
0.94 0.98 1.89 1.00
(1) (1) (2) (1)
1.00 (1) 2.00 (2)
0.93 (1)
1.13 (1)
0.92 (1)
1.02
(1)
0.99
1.01
(1)
0.92 (1)
1.00 (1)
1.00
12
Total Position recovery (»)*
5b
1-12
12 1-12 3.1 nmol (87 %)
0.90
(1)
(1)
0.99 (1) 0.96 (1)
1.00 (1)
0.84 (1) 1.00 (1) 1.84 (2)
6 71-76 3.1 nmol (87 %)
78-83 77-63 3.2 nmol (89 %)
- A-2 2b
2a
CyaO3H Aap Thr Ser Glu Pro Gly Ma v«i Htt lie L*u Tyr Phe Ly» Bis Trp Ar9
(1) (4) (1) (2) (2)
1.53 4.11 1.44 2.81 1.81
(1) (4) (1) (2) (2)
1.48 4.27 1.20 2.48
(1) (4) (1) (2)
2.40
(2)
2.00 1.96
(2) (2)
2.10 2.03
(2) (2)
2.38 2.37
(2) (2)
2.11 6.00
(2) (6)
2.27 6.00
(2) (6)
(2)
1.07 1.91
(1) (2)
1.20 1.96
(1) (2)
1.08 (1) 2.36 (2)
1.99
(2)
1.93
(2)
1.89 (2)
27
(2)
13-39
3 . 1 nmol ( 8 7 %)
Recovery
(2) (4) (5) (2) (15)
N.D. 1.04 3.81 0.98 8.11
(1) (1) (4)
6.42
(6)
4.07
3.93 0.72 3.30 6.00 1.04 3.85 2.70 1.00 N.D. 3.86
(4) (1) (4) (6) (1) (4) (3) (1) (1) (4)
1.98 4.08 4.66 2.34 14.92
27
27
13-39
13-39
-A-6 —
—A-3 — 3
3+6
1.05 3.85 0.94 2.12 2.01
Total
Poaition
2c
(1) (8)
N.D. 3.07 1.00 1.07 7.17
(1) (3) (1) (1) (7)
(4)
2.18
(2)
1.98 (2)
1.98 0.92 1.27 3.00 1.03 2.03 1.80
(2) (1) (2) (3) (1) (2) (2)
N.D. 2.93
(1) (3)
1.99 (2) 3.00 (31 1.94 (2) 0.91 (1) 0.97 (1) 1.02
(1)
31
63 (40-70) (84-115)
40-70
3 . 1 nmol ( 8 7 %)
1 . 2 nmol (86 %)•*
32 84-115 1.1 (79
nmol %)•"
Recovery (%) la calculated relative to tht dole* of the starting u t e r i a l d . 8 nmol of diner rhIL-5) . calculated relative to the moles of the DTT-treated API digesta (0.7 nmol of dimer rhIL-5)
Table US.(continued) Table US. Sequence imlyiU of API genented peptidet from mIL-5.
Papudea
Peptides Cycle # l 2 3 4 5 6 7 8 9 10 11 12 Initial yield(l)
1a I P T E I P T 3 A L V K
480pmol 300 ND 151 268 245 66' 29 204 204 153 125 66.2
1b I P T E I P T S
115 poolN 133 L MD S 4 L 55 I 56 K 12 7 A 30 L 18 V 14 It 8 62.7
4 253 pool Y 384 I 36 D 213 G 121 Q 90 K
5a
5b
336paolK 224 Y 168 I 239 D 251 G 107 Q
K
31.6
86.6
761pmol 697 704 308 595 596 456
52.1
»•
2a
Cycle* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
E T L A L L S T H R T L L I A H E T L R I P V P V H K
IB 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Initial yield(%)
332D»O1
90 343 406 299 344 42 44 30 49 30 112 166 99 13 ND 67 19
NO
E T L A L L 3 T H R T L L I A N E T L
16paol 5 6 16 14 15 6 2 HQ NQ
E T L A L L 3 T H
NO
T L L
4 7 3 3 HD 2
NO 5
36 46 40 39 45 43 7 15
46.9
2C#
51.2
R
I A N E
3
15 psolN 5 H 7 Q 10 L 6 C 9 T NQ E E NO I HQ NQ F NO Q G NO 5 I NQ G NQ T 2 L E NO 3 0 T V Q G G T V E R L F K
63.2
6
U5pmol 14 67 69 ND 16 41 52 25 26 36 30 25 34 6 33 17
HO 16 5 6 14 11 17 3 4 6
NO 7 2
NO
36 .5
NO: Not quantitatively determined ND: Not detected •: Analysis did not r«sch to C-ttrainal end due to their small quantities.
Vol. 107, No. 2, 1990
K
474D001
r. 456 HD c G E E R R R V N Q
r L D Y L Q E F L G V H N T E H I I E S
37. 9
363 260 294 53 72 86 329 283 291 270 259 109 141 226 206 170 174 166 129 145 146 121 24 73 NQ 60 77 30 »Q
Downloaded from https://academic.oup.com/jb/article-abstract/107/2/292/868100 by Stockholm University Library user on 19 January 2019
Thr Ser Glu Pro Gly Ala val lie Leu Tyr Lya
amlno acid
. A-5
Asp
Table IS. (continued)
