Correct disulfide pairing is required for the biological activity of crustacean androgenic gland hormone (AGH): Synthetic studies of AGH

Article abstract of DOI:10.1021/bi902100f

Androgenic gland hormone (AGH) of the woodlouse, Armadillidium vulgare, is a heterodimeric glycopeptide. In this study, we synthesized AGH with a homogeneous N-linked glycan using the expressed protein ligation method. Unexpectedly, disulfide bridge arran

Full text of DOI:10.1021/bi902100f

1798 Biochemistry 2010, 49, 1798–1807
DOI: 10.1021/bi902100f
Correct Disulfide Pairing Is Required for the Biological Activity of Crustacean Androgenic
Gland Hormone (AGH): Synthetic Studies of AGH†
Hidekazu Katayama,*,‡ Hironobu Hojo,*,‡ Tsuyoshi Ohira,§ Akira Ishii,‡ Takamichi Nozaki,‡ Kiyomi Goto,‡
Yuko Nakahara,‡ Tetsuo Takahashi,‡ Yuriko Hasegawa, Hiromichi Nagasawa,^ and Yoshiaki Nakahara‡
‡Department of Applied Biochemistry, Institute of Glycoscience, Tokai University, 1117 Kitakaname, Hiratsuka, Kanagawa 259-1292,
Japan, §Department of Biological Sciences, Faculty of Science, Kanagawa University, Hiratsuka, Kanagawa 259-1293, Japan,
Department of Biology, Keio University, Yokohama, Kanagawa 223-8521, Japan, and ^Department of Applied Biological
Chemistry, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113-8657, Japan
Received December 7, 2009; Revised Manuscript Received January 20, 2010
ABSTRACT: Androgenic gland hormone (AGH) of the woodlouse, Armadillidium vulgare, is a heterodimeric
glycopeptide. In this study, we synthesized AGH with a homogeneous N-linked glycan using the expressed
protein ligation method. Unexpectedly, disulfide bridge arrangement of a semisynthetic peptide differed from
that of a recombinant peptide prepared in a baculovirus expression system, and the semisynthetic peptide
showed no biological activity in vivo. To confirm that the loss of biological activity resulted from disulfide
bond isomerization, AGH with a GlcNAc moiety was chemically synthesized by the selective disulfide
formation. This synthetic AGH showed biological activity in vivo. These results indicate that the native
conformation of AGH is not the most thermodynamically stable form, and correct disulfide linkages are
important for conferring AGH activity.
Sex differentiation in crustaceans was first shown to be under
hormonal control by Charniaux-Cotton in 1954 (1). This intrinsic
factor, termed androgenic gland hormone (AGH), is synthesized
in and secreted from the male-specific organ, the androgenic
gland. In 1999, the chemical structure of AGH from a terrestrial
isopod, Armadillidiyum vulgare, was determined to be a hetero-
dimeric glycopeptide with an insulin-like amino acid sequence (2),
and a cDNA encoding the AGH precursor was cloned (3). Up to
now, cDNAs encoding AGH precursors or putative AGH
precursors have been cloned from three isopod (2, 4) and one
decapod species (5).
Similar to proinsulin, AGH is expressed as a 123-residue single
precursor peptide chain consisting of B-chain (44 aa), C-peptide
(50 aa), and A-chain (29 aa). Removing the C-peptide portion by
processing at two Lys-Arg sites transforms it into a mature
heterodimeric form which contains four disulfide bonds (Figure 1)
(2, 6). It has been shown that the mature AGH harbors an N-
linked glycan at AsnA18 in the A-chain, and this glycan moiety
was essential for conferring the biological activity (6). To study
the relationship between biological activity and glycan structure,
we sought to obtain an AGH that had a homogeneous glycan
structure.
systems usually have heterogeneous glycan structures, and it is
difficult to obtain homogeneous materials from these systems.
A chemical synthesis strategy provides a good solution for this
problem.
In 1998, Muir et al. developed the expressed protein ligation
method (7). In this method, a recombinant protein thioester
prepared from a C-terminal intein tag-fusion protein was used as
a building block and condensed with a peptide segment contain-
ing posttranslational modifications and/or unnatural amino
acids that were prepared by the ordinary solid-phase peptide
synthesis method. Up to now, several proteins such as phospho-,
glyco-, and acyl-proteins have been synthesized using this met-
hod (8). Since the glycosylation site of AGH is in the C-terminal
part of its precursor, this method may be useful for synthesizing
AGH. In this study, in order to obtain AGH with a homogeneous
glycan structure, we synthesized AGH using the expressed
protein ligation method.
EXPERIMENTAL PROCEDURES
General. MALDI-TOF mass spectra were recorded with a
Voyager-DE PRO spectrometer (Applied Biosystems, CA). Amino
acid composition was determined with a LaChrom amino acid
analyzer (Hitachi, Tokyo, Japan) after hydrolysis with 6 M HCl
at 150 ꢀC for 2 h in an evacuated sealed tube. Amino acid
sequence analysis was performed with a protein sequencer 491
(Applied Biosystems). Circular dichroism (CD) spectra were
measured with a Jasco J-820 spectropolarimeter (JASCO, Tokyo,
Japan) at room temperature with a 1 mm path length cell. Protein
concentration was determined from the absorbance at 280 nm
using the extinction constants of Tyr and Trp residues as reported
previously (9).
When large amounts of glycoprotein are required, recombi-
nant protein expression systems are used. However, recombinant
glycoproteins expressed in baculovirus, yeast, or mammalian cell
†This work was supported in part by a Grant-in Aid for Scientific
Research from the Ministry of Education, Sport, Science, and Techno-
logy of Japan (No. 20380069). We thank Tokai University for a Grant-
in-Aid for High-Technology Research and the Japan Society for the
Promotion of Science for a Grant-in-Aid for Creative Scientific Re-
search (No. 17GS0420).
*To whom correspondence should be addressed: e-mail, katay@
hotmail.co.jp (H.K.) and hojo@keyaki.cc.u-tokai.ac.jp (H.H.); phone/
fax, þ81-463-50-2075.
Expression and Purification of AGH(1-110) Thioester 1.
Isolation and characterization of a cDNA encoding AGH of
r
pubs.acs.org/Biochemistry
Published on Web 01/21/2010
2010 American Chemical Society

Article
Biochemistry, Vol. 49, No. 8, 2010 1799
1
in CHCl3); Rf 0.52 (95:5 CHCl3:MeOH); H NMR (CDCl3-
D2O) δ 4.85 (d, 1H, J = 12.0 Hz, Ph-CH2), 4.80 (d, 1H, J = 9.5 Hz,
H-1), 4.79 (d, 1H, J = 10.7 Hz, Ph-CH2), 4.62 (d, 2H, J = 12.0 Hz,
Ph-CH2), 4.57 (d, 1H, J = 10.7 Hz, Ph-CH2), 4.48 (m, 1H, Asn
R-H), 4.45 (d, 1H, J = 12.2 Hz, Ph-CH2), 4.40 (dd, 1H, J = 7.1,
10.3 Hz, Fmoc CH-CH2), 4.27 (t, 1H, J = 8.8 Hz, Fmoc CH-
CH2), 4.20 (t, 1H, J = 7.2 Hz, Fmoc CH-CH2), 3.87 (t, 1H, J =
10.0 Hz, H-2), 3.78 (t, 1H, J = 9.3 Hz, H-4), 3.45 (dd, 1H, J =
9.0, 10.3 Hz, H-3), 2.79 (dd, 1H, J = 4.9, 16.4 Hz, Asn β-H), 2.66
(dd, 1H, J = 4.0, 16.6 Hz, Asn β-H), 1.71 (s, 3H, CH3CO), 1.41
(s, 9H, t-Bu). Anal. Calcd for C52H57N3O10: C, 70.65; H, 6.50; N,
4.75; O, 18.10. Found: C, 70.67; H, 6.50; N, 4.70.
F
IGURE 1: Primary structure of AGH. (CHO), N-glycosylation site.
A. vulgare have been described previously (2). A cDNA encoding
AGH(1-110) was amplified by polymerase chain reaction (PCR)
using the following set of primers: a forward primer (50-TGGTC-
ATATGTACCAGGTACGAGGTATGAGATCCGATGTG-30)
and a reverse primer (50-GGTTGCTCTTCTGCATTTGTGC-
TCTGTCCTAATATTGC-30). The amplified PCR product was
cloned into NdeI/SapI sites of the pTWIN1 vector (New England
Biolabs, MA) to obtain an expression plasmid, pTWIN1-AGH-
(1-110).
N2-(9-Fluorenylmethoxycarbonyl)-N4-(2-acetamido-3,4,
6-tri-O-benzyl-2-deoxy-β-D-glucopyranosyl)-L-asparagine
4. Compound 3 (170 mg, 0.19 mmol) was dissolved in 90% TFA
in dichrolomethane (4 mL), and the solution was stirred for 1 h.
The solvent was removed in vacuo to obtain Fmoc-Asn-
(GlcNAcBn3)-OH 4 (120 mg, 0.14 mmol, 76%): [R]D þ10.1ꢀ
1
(c = 0.5 in DMF); Rf 0.38 (90:10:1 CHCl3:MeOH:AcOH); H
Escherichia coli ER2566 competent cells were transformed
with the expression plasmid and selected on LB plates containing
ampicillin (50 μg/mL). Bacterial cells from a single colony were
grown at 37 ꢀC overnight in LB medium containing ampicillin
and then diluted 50-fold with the same medium. The diluted
medium was incubated at 37 ꢀC for 90 min, and then isopropyl
NMR (DMSO) δ 12.63 (br s, 1H, COOH), 8.41 (d, 1H, J =
9.3 Hz, Asn γ-CONH), 7.99 (d, 1H, J = 9.3 Hz, NH-Ac), 7.48 (d,
1H, J = 8.3 Hz, Asn NH), 4.98 (t, 1H, J = 9.4 Hz, H-1), 4.70 (d,
1H, J = 11.0 Hz, Ph-CH2), 4.69 (d, 1H, J = 11.0 Hz, Ph-CH2),
4.65 (d, 1H, J = 11.0 Hz, Ph-CH2), 4.53 (d, 1H, J = 12.0 Hz, Ph-
CH2), 4.52 (d, 1H, J = 11.0 Hz, Ph-CH2), 4.46 (d, 1H, J = 12.0 Hz,
Ph-CH2), 4.40 (m, 1H, Asn R-H), 3.79 (m, 1H, H-2), 2.66 (dd, 1H,
J = 5.4, 16.1 Hz, Asn β-H), 1.77 (s, 3H, CH3CO). Anal. Calcd for
β-D-thiogalactoside (IPTG) was added to the culture at a final
concentration of 1 mM. After an additional incubation for 3 h,
bacterial cells were harvested by centrifugation. The bacterial
C48H49N3O10 1/2H2O: C, 68.89; H, 6.02; N, 5.02. Found: C,
3
1
cells were suspended in /50th culture volume of phosphate-
68.93; H, 5.96; N, 5.00.
buffered saline (PBS; 150 mM NaCl, 10 mM phosphate buffer,
pH 7.0), and the suspension was sonicated. The insoluble
material was collected by centrifugation and washed twice with
the same volume of PBS to obtain pure inclusion bodies.
Nonglycosylated AGH(111-123) 5. Starting from Fmoc-
Tyr(But)-Wang resin (0.63 mmol/g, 397 mg), the peptide chain
was elongated with an ABI 433A peptide synthesizer (Applied
Biosystems) using FastMoc protocol, and Fmoc-Arg(Pbf)-Thr-
(But)-Thr(But)-Val-Ser(But)-Leu-Tyr(But)-Cys(Trt)-Arg(Pbf)-Thr-
(But)-Tyr(But)-Wang resin was obtained. Four-fifths of the resin
was removed, and the remaining resin was used for additional
peptide chain elongation with the peptide synthesizer. After the
peptide chain elongation, H-Cys(Trt)-Asn(Trt)-Arg(Pbf)-Thr-
(But)-Thr(But)-Val-Ser(But)-Leu-Tyr(But)-Cys(Trt)-Arg(Pbf)-Thr-
(But)-Tyr(But)-Wang resin was obtained (246 mg). A part of the
resin (100 mg) was treated with reagent K (TFA/phenol/water/
thioanisole/ethanedithiol, 33:2:2:2:1, 1.5 mL) (12) for 2 h. TFA
was removed with a nitrogen stream, and the peptide was
precipitated by diethyl ether. The crude peptide was washed
twice with ether, and AGH(111-123) 5 was purified by reversed-
phase (RP) HPLC on a Mightysil RP-18 GP column (4.6 ꢀ 150 mm;
Kanto, Japan) at a flow rate of 1 mL/min. Elution was performed
with a 30 min linear gradient of 10-40% acetonitrile containing
0.1% TFA. The yield was 32.2 mg (20.4 μmol, 50%). MALDI-
TOF mass: found, m/z 1579.4; calcd, 1579.7 for (M þ H)þ.
Amino acid analysis: Asp1.00Thr2.93Ser0.95Cys0.09Val1.00Leu1Tyr
1
After solubilizing the inclusion bodies with a /100th culture
volume of 8 M urea solution, the resultant solution was diluted
using a 4-fold volume of dilution buffer to final conditions of
1.6 M urea/10% glycerol/50 mM sodium mercaptoethanesulfo-
nate (MESNa)/100 mM Hepes (pH 7.5) and stirred at room
temperature overnight to refold the intein tag part of the
recombinant protein and to digest the tag by MESNa. Proteins
in solution were precipitated by adding trichloroacetic acid
(TCA) at the concentration of 10%, and the precipitate was
washed twice with acetone. The desired product 1 was purified by
gel filtration column chromatography on a Superdex 75 10/300 GL
column (GE Healthcare UK Ltd., U.K.) with 50% acetonitrile/
0.1% trifluoroacetic acid (TFA) as a solvent at a flow rate of
0.75 mL/min.
N2-(9-Fluorenylmethoxycarbonyl)-N4-(2-acetamido-3,4,
6-tri-O-benzyl-2-deoxy-β-D-glucopyranosyl)-L-asparagine
tert-Butyl Ester 3. Fmoc-Asp-OBut (1.6 g, 3.9 mmol) was
dissolved in dichloromethane (6 mL), and N,N0-dicyclohexylcar-
bodiimide (DCC; 410 mg, 2.0 mmol) was added at 0 ꢀC. After the
mixture was stirred for 30 min at 0 ꢀC, the precipitate was filtered
off, and the solvent was removed in vacuo. The residue was
dissolved in tetrahydrofuran (6 mL) and added to a mixture of 2-
2.07Arg2.16
.
Asn(GlcNAc)112-AGH(111-123) 6. Fmoc-Arg(Pbf)-Thr-
(But)-Thr(But)-Val-Ser(But)-Leu-Tyr(But)-Cys(Trt)-Arg(Pbf)-Thr-
(But)-Tyr(But)-Wang resin (50 μmol equivalent) was treated with
20% piperidine/1-methyl-2-pyrrolidinone (NMP) and reacted
with Fmoc-Asn(GlcNAcBn3)-OBt, which was prepared from
Fmoc-Asn(GlcNAcBn3)-OH 4, 1 M DCC/NMP, and 1 M 1-
hydroxybenzotriazole (HOBt)/NMP. After the reaction at 50 ꢀC
for 1 h, the resin was washed with NMP, and the N-terminal
Fmoc group was cleaved with 20% piperidine/NMP treatment.
Fmoc-Cys(Trt)-OH was then condensed in the same manner as
acetamido-3,4,6-tri-O-benzyl-2-deoxy-β-D-glucopyranosyl azide
2 (520 mg, 1.0 mmol) and Lindlar catalyst (300 mg) (10). The
resulting solution was stirred overnight under a H2 atmo-
sphere (11). After the catalyst was removed on Celite, the solvent
was removed in vacuo. The residue was purified by silica gel
column chromatography with 95:5 CHCl3-MeOH to give
compound 3 (660 mg, 0.75 mmol, 75%): [R]D þ45.9ꢀ (c = 0.5

1800 Biochemistry, Vol. 49, No. 8, 2010
Katayama et al.
Fmoc-Asn(GlcNAcBn3)-OH. After cleavage of the Fmoc
group by 20% piperidine/NMP treatment, H-Cys(Trt)-Asn-
(GlcNAcBn3)-Arg(Pbf)-Thr(But)-Thr(But)-Val-Ser(But)-Leu-Tyr-
(But)-Cys(Trt)-Arg(Pbf)-Thr(But)-Tyr(But)-Wang resin was ob-
tained (207 mg). A part of the resin (100 mg) was treated with
reagent K (1.5 mL) for 2 h. TFA was removed with a nitrogen
stream, and the peptide was precipitated by ether. The crude
peptide was washed twice with ether and then dissolved in a
mixture of TFA/dimethyl sulfide/m-cresol (5:3:1, 0.9 mL). Triflic
acid (0.1 mL) was added to the solution, and the reaction mixture
was left at -15 ꢀC for 2 h (13-16). The crude peptide was
precipitated and washed twice with ether and separated by RP-
HPLC under the same conditions as for the nonglycosylated
AGH(111-123) to give Asn(GlcNAc)112-AGH(111-123) 6
(43.1 mg, 24.2 μmol, 37%). MALDI-TOF mass: found, m/z
unit of metalloendopeptidase (Seikagaku Kogyo, Japan) was
added. After 2 h incubation at 37 ꢀC, product 10 was purified by
RP-HPLC in the same manner as described above. The yield was
14 μg (2.0%). MALDI-TOF mass: found, m/z 9161.8; calcd,
9162.2 (average) for (M þ H)þ.
Asn(GlcNAc)112-AGH 15. The N-terminal peptide thioester
1 (1 mg) and monosaccharyl C-terminal peptide 6 (1 mg) were
used for the coupling reaction. The conditions for coupling,
folding, and processing reactions were the same as those for 10.
The isolated yield of 15 was 57 μg (8.0%). MALDI-TOF mass:
found, m/z 9366.4; calcd, 9363.4 (average) for (M þ H)þ.
Asn[(GalGlcNAcMan)2ManGlcNAc2]112-AGH 16. The
N-terminal peptide thioester 1 (1 mg) and nonasaccharyl C-
terminal peptide 7 (0.9 mg) were used for the coupling reaction.
The conditions for coupling, folding, and processing reactions
were the same as those for 10. The isolated yield of 16 was 14 μg
(1.6%). MALDI-TOF mass: found, m/z 10784.6; calcd, 10784.7
(average) for (M þ H)þ.
1782.9; calcd, 1782.8 for (M þ H)þ. Amino acid analysis: Asp1.03
-
Thr2.92Ser0.96Cys0.23Val1.28Leu1Tyr2.06Arg2.14
.
Asn[(GalGlcNAcMan)2ManGlcNAc2]112-AGH(111-123)
7. Fmoc-Asn[(GalGlcNAcMan)2ManGlcNAc2]-OH was a kind
gift from Otsuka Chemical Co. Ltd., Japan. Fmoc-Arg(Pbf)-
Thr(But)-Thr(But)-Val-Ser(But)-Leu-Tyr(But)-Cys(Trt)-Arg(Pbf)-
Thr(But)-Tyr(But)-Wang resin (40 μmol equivalent) was treated
with 20% piperidine/NMP. Fmoc-Asn[(GalGlcNAcMan)2Man-
GlcNAc2]-OH (50 μmol) was activated with 3-(diethoxyphos-
phoryloxy)-3H-benzo[d][1,2,3]-triazin-4-one (DEPBT) in NMP
and reacted with the resin (17, 18). After reaction at 50 ꢀC for 2 h,
the resin was washed with NMP, and the N-terminal Fmoc group
was cleaved by 20% piperidine/NMP treatment. Fmoc-Cys(Trt)-
OH was then condensed by DCC/HOBt activation to obtain
H-Cys(Trt)-Asn[(GalGlcNAcMan)2ManGlcNAc2]-Arg(Pbf)-
Thr(But)-Thr(But)-Val-Ser(But)-Leu-Tyr(But)-Cys(Trt)-Arg(Pbf)-
Thr(But)-Tyr(But)-Wang resin (174 mg). A part of the resin
(44 mg) was treated with reagent K (1.0 mL) for 2 h. TFA was
removed with a nitrogen stream, and the peptide was precipi-
tated with ether. The crude peptide was washed twice with
ether, and the product was purified by RP-HPLC under the
same conditions as for the nonglycosylated AGH(111-123) to
give Asn[(GalGlcNAcMan)2ManGlcNAc2]112-AGH(111-123)
7 (0.90 mg, 0.28 μmol, 2.8%). MALDI-TOF mass: found, m/z
3202.3; calcd, 3202.3 for (M þ H)þ. Amino acid analysis: Asp0.84
Bioassay. AGH activity was determined using an in vivo
bioassay as described (19, 20). In brief, samples (10 ng each of the
synthetic AGH or 0.5 AG equivalent of AG extract) were added
to a small amount of bovine serum albumin, dried under vacuum,
and dissolved in saline solution. They were injected into 10 young
A. vulgare females. After molting, the animals were examined
under a stereoscopic microscope for elongation of the endo-
podites of the first pair of abdominal legs, a criterion for
masculinization.
Trypsin Digestion. Peptides 10 and 15 (10 μg each) were
dissolved in 100 mL of 50 mM Hepes buffer (pH 7.0), and 1 μL of
TPCK-treated trypsin (Sigma, MO) aqueous solution (1 mg/mL)
was added. The solution was incubated at 37 ꢀC for 24 h, and the
tryptic digests were separated by RP-HPLC on a Mightysil RP-
18 GP column. Elution was performed with a 30 min linear
gradient of 10-40% acetonitrile containing 0.1% TFA.
Endoproteinase Glu-C Digestion. The tryptic digest was
dissolved in 100 μL of 100 mM Tris buffer (pH 6.8), and 0.5 μL of
endoproteinase Glu-C (Sigma) aqueous solution (1 mg/mL) was
added. The solution was incubated at 37 ꢀC for 24 h, and the
digests were separated by RP-HPLC in the same manner as that
for the tryptic digests.
Thr2.64Ser0.92Cys0.36Val2.00Leu1Tyr1.87Arg1.93
.
[Asn(GlcNAcBn3)112,Cys(MeOBn)102,Cys(Acm)120]-AGH
A-Chain 17. Starting from Fmoc-Tyr(But)-OCH2-C6H4-OCH2-
C6H4-resin (520 mg, 250 μmol), the peptide chain was elongated
with an ABI 433A peptide synthesizer using FastMoc proto-
col, and Arg(Pbf)-Thr(But)-Thr(But)-Val-Ser(But)-Leu-Tyr(But)-
Cys(Acm)-Arg(Pbf)-Thr(But)-Tyr(But)-OCH2-C6H4-OCH2-C6H4-
resin was obtained. A part of the resin (70 μmol) was removed and
reacted with Fmoc-Asn(GlcNAcBn3)-OBt, which was prepared
from Fmoc-Asn(GlcNAcBn3)-OH (83 mg, 100 μmol), 1 M DCC/
NMP (110 μmol), and 1 M HOBt/NMP (110 μmol). After 1.5 h,
the resin was washed with NMP and acetylated with 10% Ac2O
and 5% DIEA in NMP for 10 min. Chain elongation was
continued using the peptide synthesizer, and Glu(OBut)-Ile-
Ala-Phe-Tyr(But)-Gln(Trt)-Glu(OBut)-Cys(MeOBn)-Cys(Trt)-
Asn(Trt)-Ile-Arg(Pbf)-Thr(But)-Glu(OBut)-His(Trt)-Lys(Boc)-
Cys(Trt)-Asn(GlcNAcBn3)-Arg(Pbf)-Thr(But)-Thr(But)-Val-Ser-
(But)-Leu-Tyr(But)-Cys(Acm)-Arg(Pbf)-Thr(But)-Tyr(But)-OCH2-
C6H4-OCH2-C6H4-resin (444 mg) was obtained. A part of the
resin (108 mg) was treated with reagent K (1 mL) for 2 h. TFA
was removed with a nitrogen stream, and the peptide was
precipitated with ether. The peptide was washed twice with ether
and extracted with 50% aqueous acetonitrile (10 mL). The
Nonglycosylated AGH 10. The N-terminal peptide thioester
1 (1 mg) and nonglycosylated C-terminal peptide 5 (1 mg) were
mixed in a 1.5 mL microtube and dissolved in 100 μL of 100 mM
phosphate buffer (pH 7.0) containing 6 M guanidine hydro-
chloride. The coupling reaction was initiated by adding 4-
mercaptophenylacetic acid (MPAA) at the concentration of
2%. After reaction at room temperature for 3 days, the reaction
mixture was desalted by gel filtration column chromatography
on a Superdex 75 10/300 GL column with 50% acetonitrile/0.1%
TFA as a solvent at a flow rate of 0.75 mL/min. After
lyophilization, the resultant material was dissolved in 1 mL of
8 M urea/90 mM glutathione (reduced form, GSH). The solution
was then diluted to the final conditions of 1.3 M urea/10%
glycerol/5 mM GSH/200 mM Tris (pH 8.5), 18 mL, and cooled at
4 ꢀC. The oxidized form of glutathione (GSSG) was then added
to the solution at a concentration of 1 mM, and the solution was
stirred for 3 days at 4 ꢀC. The folded AGH 9 was purified by RP-
HPLC on a PROTEIN-RP column (YCM, Japan). Elution was
performed with a 30 min linear gradient of 20-50% acetonitrile
containing 0.1% TFA. After lyophilization, the resultant peptide
was dissolved in 100 μL of 100 mM Tris (pH 6.8), and then 0.1

Article
Biochemistry, Vol. 49, No. 8, 2010 1801
solution was added dropwise into 0.1 M ammonium acetate
containing 6 M guanidine hydrochloride (200 mL), and the
solution was gently stirred for 3 days. Acetic acid was added,
and the product was purified by RP-HPLC to obtain peptide
17 (2.7 μmol). The dibenzylated peptides were also collected
(0.61 μmol). MALDI-TOF mass: found, m/z 4207.4; calcd,
[Asn(GlcNAc)112]-AGH 22. Peptide 21 (80 nmol) was
dissolved in 20% aqueous MeOH (1.0 mL) and added to MeOH
(5.0 mL) containing 20 mM I2 in MeOH (200 μL) and 6 M HCl
(67 μL) within 7 min with mixing. After the resulting solution was
stirred for another 15 min, the reaction was terminated by adding
ascorbic acid. The product was purified by cation-exchange
column chromatography (TSKgel CM-3SW, 7.5 mm ꢀ 750 mm)
using a linear gradient from 20 mM sodium phosphate (pH 6.0)
containing 40% MeCN to the same buffer containing 0.5 M
NaCl in 20 min at a flow rate of 0.5 mL/min and then desalted by
RP-HPLC to give peptide 22 (31 nmol, 39%). MALDI-TOF
mass: found, m/z 8949.8; calcd, 8949.2 (average) for (M þ H)þ.
Amino acid analysis: Asp7.00Thr5.42Ser1.79Glu8.92Pro4.46Gly3.16
4206.9 for (M þ H)þ. Amino acid analysis: Asp2.09Thr3.79Ser0.96
-
Glu4.25Ala11/2Cys2.46Val1.10Ile1.74Leu1.06Tyr2.86Phe0.92Lys0.99
His0.97Arg3.02
.
[Asn(GlcNAc)112,Cys(SPy)102,Cys(Acm)120]-AGHA-Chain
18. Peptide 17 (730 nmol), DPDS (6.4 mg, 29 μmol), and
thioanisole (60 μL) were dissolved in TFA (600 μL), and TfOH
(30 μL) was added at -10 ꢀC. After the solution was kept for
5 min at -10 ꢀC, the product was precipitated with ether and
purified by RP-HPLC to give peptide 18 (299 nmol, 41%).
MALDI-TOF mass: found, m/z 3925.4; calcd, 3925.7 for (M þ
H)þ. A major peak corresponding to the depyridine sulfenylated
peptide, which may be derived from decomposition during the
mass measurement, was also observed: found, m/z 3816.5; calcd,
1
Ala1 /2Cys2.66Val4.05Met0.92Ile3.75Leu4.16Tyr5.43Phe3.38Lys2.08
His1.01Arg7.20
.
[Cys(MeOBn)102,Cys(Acm)120]-AGH A-Chain 24. Start-
ing from Fmoc-Tyr(But)-OCH2-C6H4-OCH2-C6H4-resin (510 mg,
245 μmol), the peptide chain was elongated with the peptide
synthesizer using FastMoc protocol, and Glu(OBut)-Ile-Ala-
Phe-Tyr(But)-Gln(Trt)-Glu(OBut)-Cys(MeOBn)-Cys(Trt)-Asn-
(Trt)-Ile-Arg(Pbf)-Thr(But)-Glu(OBut)-His(Trt)-Lys(Boc)-Cys-
(Trt)-Asn(Trt)-Arg(Pbf)-Thr(But)-Thr(But)-Val-Ser(But)-Leu-Tyr-
(But)-Cys(Acm)-Arg(Pbf)-Thr(But)-Tyr(But)-OCH2-C6H4-OCH2-
C6H4-resin (2.1 g) was obtained. An aliquot of the resin (100 mg)
was treated with 1 mL of reagent K for 2 h at room temperature.
The peptide precipitated by ether was dissolved in 50% aqueous
acetonitrile containing 0.1% TFA (10 mL) and diluted with
0.1 M ammonium acetate (90 mL) containing 6 M guanidine
hydrochloride. The pH was adjusted to 8.0 with aqueous ammo-
nia, and the solution was stirred for 3 days. The main peak
was isolated by RP-HPLC to give peptide 24 (1.1 μmol, 4.1 mg,
9.3% yield). MALDI-TOF mass: found, m/z 3733.2; calcd,
3816.7 for [Asn(GlcNAc)112,Cys(Acm)120]-AGH A-chain (M þ H)þ.
1
Amino acid analysis: Asp2.17Thr3.83Ser0.99Glu4.64Ala1 /2Cys1.55
-
Val0.98Ile1.83Leu1.00Tyr2.88Phe0.94Lys1.05His1.00Arg3.15
.
[Cys(MeOBn)12,Cys(Acm)23]-AGH B-Chain 19. Starting
from Fmoc-Glu(OBut)-Wang resin (0.50 g, 275 μmol), the
peptide chain was elongated using the peptide synthesizer, and
Tyr(But)-Gln(Trt)-Val-Arg(Pbf)-Gly-Met-Arg(Pbf)-Ser(But)-Asp-
(OBut)-Val-Leu-Cys(MeOBn)-Gly-Asp(OBut)-Ile-Arg(Pbf)-Phe-
Thr(But)-Val-Gln(Trt)-Cys(Trt)-Ile-Cys(Acm)-Asn(Trt)-Glu(OBut)-
Leu-Gly-Tyr(But)-Phe-Pro-Thr(But)-Glu(OBut)-Arg(Pbf)-Leu-
Asp(OBut)-Lys(Boc)-Pro-Cys(Trt)-Pro-Trp(Boc)-Pro-Asn(Trt)-
Arg(Pbf)-Glu(OBut)-Wang resin (2.2 g) was obtained. An ali-
quot of the resin (50 mg) was treated with reagent K (500 μL) for
2 h at room temperature. The peptide precipitated with ether was
dissolved in 50% aqueous acetonitrile containing 0.1% TFA and
diluted with 0.1 M ammonium acetate (50 mL) containing 6 M
guanidine hydrochloride. The pH was adjusted to 8.0 with
aqueous ammonia, and the solution was stirred for 3 days. The
main peak was isolated by RP-HPLC to give peptide 19 (368
nmol, 5.9% yield based on the content of Glu residue in the
starting resin). MALDI-TOF mass: found, m/z 5397.7; calcd,
5397.2 (average) for (M þ H)þ. Amino acid analysis: As-
p5.10Thr1.96Ser0.91Glu5.10Pro3.94Gly3.001/2Cys2.50Val2.79Met0.72Ile
3733.7 for (M þ H)þ. Amino acid analysis: Asp1.96Thr3.48Ser0.90
-
Glu4.05Ala1.001/2Cys3.18Val0.90Ile1.71Leu0.99Tyr2.96Phe0.95Lys0.94
His0.94Arg2.75
.
[Cys(SPy)102,Cys(Acm)120]-AGH A-Chain 25. Peptide
24 (15 mg, 2.0 μmol), DPDS (16 mg, 72.6 μmol), and thioanisole
(60 μL) were dissolved in TFA (540 μL), and the solution was
kept at room temperature for 2.5 h. The product was precipitated
with ether and purified by RP-HPLC to give peptide 25 (5.7 mg,
0.93 μmol, 48%). MALDI-TOF mass: found, m/z 3722.9; calcd,
3722.6 for (M þ H)þ. Amino acid analysis: Asp2.20Thr3.90Ser0.99
Glu4.73Ala1.001/2Cys1.57Val0.97Ile1.86Leu1.06Tyr3.17Phe1.07Lys1.13
1.69Leu3.14Tyr1.81Phe2.13Lys1.01Arg4.74
.
[Cys(Acm)23]-AGH B-Chain 20. Peptide 19 (790 nmol), m-
cresol (15 μL), and thioanisole (30 μL) were dissolved in TFA
(210 μL), and TfOH (26 μL) was added at 0 ꢀC. After the solution
was kept at 0 ꢀC for 20 min with occasional mixing, the product
was precipitated with ether and purified by RP-HPLC to give
peptide 20 (380 μmol, 48%). MALDI-TOF mass: found, m/z
His1.07Arg3.17.
Cys(Acm)23, 120-AGH 26. Peptide 20 (303 nmol) in 50%
acetonitrile (1.0 mL) was added dropwise to the solution of 25
(312 nmol) in 0.1 M NH4HCO3 (0.8 mL) and acetonitrile
(0.4 mL) within 1 min. The reaction was quenched by adding
acetic acid, and the product was purified by RP-HPLC to give
peptide 26 (286 nmol, 94%). MALDI-TOF mass: found, m/z
8890.5; calcd, 8890.2 (average) for (M þ H)þ. Amino acid
5273.4; calcd, 5273.5 for (M þ H)þ. Amino acid analysis: Asp5.03
-
-
Thr1.73Ser0.87Glu5.27Pro4.11Gly2.711/2Cys2.46Val2.43Met0.84Ile1.24
1
Leu2.80Tyr1.83Phe2Lys1.06Arg4.61
.
analysis: Asp6.95Thr5.24Ser1.72Glu9.42Pro4.02Gly2.66Ala1 /2Cys3.67
[Asn(GlcNAc)112,Cys(Acm)23, 120]-AGH 21. Peptide 20
(134 nmol) in 50% acetonitrile (0.8 mL) was added dropwise to a
mixture of peptide 18 (224 nmol), 0.1 M NH4HCO3 (0.8 mL),
and acetonitrile (0.4 mL) within 1 min. The reaction was
quenched by adding acetic acid, and the product was purified
by RP-HPLC to give peptide 21 (80 nmol, 60%). MALDI-TOF
mass: found, m/z 9093.3; calcd, 9093.4 (average) for (M þ H)þ.
Amino acid analysis: Asp6.82Thr5.22Ser1.67Glu8.65Pro4.41Gly3.06
Val3.35Met0.89Ile2.93Leu3.71Tyr4.72Phe2.96Lys2.08His0.96Arg7.49.
Nonglycosylated AGH 23. Peptide 26 (60 nmol) was
dissolved in distilled water (0.5 mL) and added dropwise to
MeOH (2.0 mL) containing 20 mM I2 in MeOH (60 μL) and 6 M
HCl (20 μL) within 5 min with mixing. After the resulting
solution was stirred for another 15 min, the reaction was
terminated by adding aqueous ascorbic acid solution. The
product was purified by cation-exchange column chromato-
graphy (TSKgel CM-3SW, 7.5 mm ꢀ 75 mm) using a linear gra-
dient from 20 mM sodium phosphate (pH 6.0) containing 40%
1
Ala1 /2Cys2.26Val3.99Met1.08Ile3.51Leu3.96Tyr5.22Phe3.09Lys2.04
His1.04Arg7.12
.

1802 Biochemistry, Vol. 49, No. 8, 2010
Katayama et al.
Scheme 1: Synthesis Strategy for AGH by the Expressed
Protein Ligation Method
F
IGURE 2: Preparation of the N-terminal peptide thioester 1. (a) Gel
filtration chromatogram (GFC) after the cleavage reaction of the
recombinant intein fusion protein. (b) SDS-PAGE analysis of the
peaks of GFC. Lane symbols indicate the peaks of GFC in panel a.
Scheme 2: Synthesis Route for Fmoc-Asn(GlcNAcBn3)-OH
acetonitrile to the same buffer containing 0.5 M NaCl in 20 min at
a flow rate of 0.5 mL/min and desalted by RP-HPLC to give
peptide 23 (20 nmol, 33%). MALDI-TOF mass: found, m/z
8748.6, calcd, 8746.0 (average) for (M þ H)þ. Amino acid
1
analysis: Asp6.93Thr5.24Ser1.70Glu9.55Pro4.53Gly2.64Ala1 /2Cys4.67
2 (10) was reduced and condensed using an anhydride of Fmoc-
Asp-OBut in situ by Matsuo’s method (11). The But ester was
then cleaved with trifluoroacetic acid (TFA) to obtain Fmoc-
Asn(GlcNAcBn3)-OH 4.
Val3.51Met0.96Ile3.10Leu3.77Tyr4.85Phe2.99Lys2.12His0.98Arg7.66
.
RESULTS
Preparation of the N-Terminal Peptide Thioester 1. In
order to synthesize AGH using the expressed protein ligation
method (Scheme 1), the N-terminal peptide thioester was pre-
pared using a bacterial expression system. A recombinant AGH-
(1-110) with a C-terminal intein-chitin binding protein tag was
expressed in an E. coli expression system. Furthermore, an
additional Met residue was attached at the N-terminus of the
fusion protein. Unfortunately, most of the recombinant protein
was expressed as inclusion bodies, and only a small amount was
detected in the soluble fraction after disrupting cells by sonica-
tion. No drastic change in expression level in the soluble fraction
was observed when the expression was induced under milder
Nonglycosylated C-terminal segment 5 was synthesized by
ordinary Fmoc chemistry. To synthesize monosaccharyl peptide
6, Asn derivative 4 was introduced instead of Fmoc-Asn(Trt)-OH
in the synthesis of 5 by the N,N0-dicyclohexylcarbodiimide
(DCC)/1-hydroxybenzotriazole (HOBt) method. The protected
peptide resins were treated with reagent K (12) for 2 h at room
temperature, and benzyl groups on the sugar moiety were cleaved
by low-acidity TfOH treatment (13-16). The nonasaccharyl
peptide 7 was also synthesized and purified in the same manner
as that for 5, except for Asn112 in which Fmoc-Asn[(GalGlc-
NAcMan)2ManGlcNAc2] was introduced by 3-(diethoxyphos-
phoryloxy)-3H-benzo[d][1,2,3]-triazin-4-one (DEPBT) as a con-
densation reagent (17, 18).
conditions, such as low temperature and low isopropyl β-D-
thiogalactoside (IPTG) concentration. Therefore, we decided to
obtain the desired product from the inclusion bodies by an in vitro
refolding method.
Native Chemical Ligation, Folding, and Processing. To
synthesize nonglycosylated linear AGH(1-123) 8, the N- and C-
terminal segments, 1 and 5, were condensed by the native
chemical ligation reaction (21). The reaction was monitored by
SDS-PAGE and was almost complete within 3 days (Figure 3).
Even when the reaction continued for 7 days, the N-terminal
segment was not completely consumed, suggesting that the
thioester moiety was partially hydrolyzed. Since the product
could not be separated from the hydrolyzed N-terminal segment,
AGH(1-110), by reversed-phase (RP)-HPLC, we separated it
from other additives by gel-filtration column chromatography
and used it for the next reaction without further purification.
Disulfide bonds were formed by oxidation in a redox buffer
containing reduced and oxidized forms of glutathione, and
folded propeptide 9 was purified by RP-HPLC (Figure 4).
The inclusion bodies were dissolved and denatured in an 8 M
urea solution, and the intein-tag portion was refolded by 5-fold
dilution under neutral pH. Simultaneously, the tag was cleaved
off by adding a thiol compound, sodium mercaptoethanesulfo-
nate, to obtain an AGH(1-110) thioester. The reaction mixture
was separated using gel filtration column chromatography as
shown in Figure 2. The desired product 1 (∼14 kDa) was eluted
just after the peak derived from the tag portion (∼24 kDa), and
the yield was 6.8 mg from 400 mL culture.
Preparation of the C-Terminal Segment. The Asn unit
bearing a GlcNAc moiety was prepared as shown in Scheme 2. 2-
Acetamido-3,4,6-tri-O-benzyl-2-deoxy-β-D-glucopyranosyl azide

Article
Biochemistry, Vol. 49, No. 8, 2010 1803
F
IGURE 3: SDS-PAGE analysis of the ligation products: lane 1,
nonglycosylated AGH; lane 2, monosaccharyl AGH; lane 3, non-
asaccharylAGH. A, the positionof12; B, the positions of8 and 11; C,
the positions of 1 and hydrolyzed material of 1.
The C-peptide portion of 9 was removed by metalloendopepti-
dase treatment at 37 ꢀC for 2 h, and mature AGH 10 was
obtained (Figure 5). The isolated yield based on the amount of
the N-terminal segment used for the peptide coupling reaction
was 2.0%.
Monosaccharyl (11) and nonasaccharyl (12) propeptides were
also synthesized by the native chemical ligation and were folded
in the same manner as for the nonglycosylated peptide 9, which
provided folded propeptides 13 and 14. The C-peptide was
cleaved off in the same manner as for 10, resulting in hetero-
dimeric monosaccharyl (15) and nonasaccharyl (16) AGHs.
Their isolated yields were 8.0% and 1.6%, respectively.
F
IGURE 4: RP-HPLC elution profiles after the refolding reaction:
(a) nonglycosylated AGH 9; (b) monosaccharyl AGH 13; (c) non-
asaccharyl AGH 14.
Circular dichroism (CD) spectra of the synthetic AGHs were
similar regardless of their glycan structures (Figure 6). In
addition, the spectral patterns showed that these AGHs pos-
sessed R-helical structures, which were similar to those of other
insulin-family peptides (22, 23).
Biological Activity and Confirmation of Folding. Biolo-
gical activities of semisynthetic AGHs were assessed by in vivo
bioassay for AGH. Unexpectedly, no sex reversal was observed
in the female animals after an injection (200 ng/individual) of 10
or 15, although injection of an AG extract (0.6 AG equivalent)
induced masculinization (nine of nine individuals were
masculinized). Since it has been reported that the native AGH
isolated from the androgenic gland showed the activity at an
injected dose of 0.1 ng/individual (6), it was likely that the
synthetic AGHs did not have proper conformations.
In order to confirm folding, 10 and 15 were digested with
endoproteases, and disulfide linkages of the digests were deter-
mined by mass spectral and N-terminal amino acid sequence
analyses. First, AGHs were digested with TPCK-treated trypsin,
and the digests were separated by RP-HPLC. On each HPLC
chromatogram, two major peaks were observed. The analytical
data for the fragments are summarized in Figure 7. These results
indicated that each fragment consisted of three peptide chains
connected by two disulfide bonds. When trypsin-2 fragment, a
tryptic digest of nonglycosylated AGH 10, was directly analyzed
with a protein sequencer, di-PTH-Cys2 was detected at cycles 5
and 7, establishing the arrangements of disulfide bonds between
Cys21 and Cys38 and between Cys23 and Cys120. The sequence
analysis for trypsin-20, one of the tryptic digests of monosaccha-
ryl AGH 15, gave the same result as trypsin-2. To determine the
other two disulfide linkages, trypsin-1 and -10 fragments were
further digested with endoproteinase Glu-C. After digestion,
fragments containing disulfide bonds were purified by RP-HPLC
and analyzed with a protein sequencer. In the analysis of GluC-1,
a digest of trypsin-1, di-PTH-Cys2 was detected at cycles 1 and 5,
F
IGURE 5: RP-HPLC elution profiles after metalloendopeptidase
treatment: (a) nonglycosylated AGH 10; (b) monosaccharyl AGH
15; (c) nonasaccharyl AGH 16.
but not at cycle 2, establishing the arrangements of disulfide
bonds between Cys12 and Cys103 and between Cys102 and Cys111
.
GluC-10, a digest of trypsin-10, gave the same results as GluC-1.
These disulfide bond arrangements were not identical to that of
recombinant AGH prepared in a baculovirus expression system,
which had disulfide bonds between Cys12-Cys102, Cys21-Cys38,
Cys23-Cys120, and Cys103-Cys111 (6).
To obtain an AGH having correct disulfide bonds, we carried
out refolding reactions under different conditions (reduced/

1804 Biochemistry, Vol. 49, No. 8, 2010
Katayama et al.
F
IGURE 7: Analytical data and structures of protease digests. Solid
lines indicate the disulfide bonds. Dotted lines indicate the disulfide
bonds observed in the recombinant AGH expressed in a baculovirus
expression system.
Scheme 3: Synthesis Procedure for Asn(GlcNAc)112-AGH 22
F
IGURE 6: Circular dichroism spectra: (a) semisynthetic nonglycosyl-
ated AGH 10; (b) semisynthetic monosaccharyl AGH 15; (c) semi-
synthetic nonasaccharyl AGH 16.
oxidized glutathione concentrations, pH, and additives) and also
done by the stepwise dialysis method. However, these reactions
gave the same results, and we could not obtain an AGH having
native conformation.
Total Chemical Synthesis of AGH. To confirm that
biological inactivity of the semisynthetic AGH was due to the
disulfide isomerization, total chemical synthesis with selective
disulfide bond formation was performed. The synthetic proce-
dure is shown in Scheme 3. The A-chain with a GlcNAc moiety
was synthesized using a peptide synthesizer. First, to form the
intrachain disulfide bond, Cys103 and Cys111 were protected by a
trityl (Trt) group. Cys102 was protected by a methoxybenzyl
(MeOBn) group for the conversion to a pyridylsulfenyl (SPy)
group by dipyridyl disulfide (DPDS). To form the final disulfide
bond by iodine oxidation, Cys120 was protected by an acetami-
domethyl (Acm) group. The introduction of Asn112 was accom-
plished using Fmoc-Asn(GlcNAcBn3) 4 by the DCC/HOBt
method. The protected peptide resin was treated with reagent
K for 2 h at room temperature. This treatment removed a part of
the benzyl groups on the GlcNAc moiety.
Before purification, the crude peptide was air-oxidized in
acetate buffer (pH 8) containing 6 M guanidine hydrochloride
for 3 days to form an intrachain disulfide bond between Cys103
and Cys111. The solution was loaded on an HPLC column, and
A-chain congeners with tri- and dibenzyl GlcNAc were collected
separately. The combined yield of these peptides was 19% based
on the amount of Tyr on the initial resin. Peptides 17 were further
treated with TfOH-TFA in the presence of DPDS for 5 min at
-10 ꢀC (24), in order to deprotect the MeOBn group of Cys102
and benzyl groups of carbohydrate moiety and also to introduce
the SPy group into Cys102. After purification, peptide 18 was
obtained in 41% yield.
peptide was air-oxidized and purified by RP-HPLC to obtain
peptide 19 in 5.9% yield based on the amount of Glu on the initial
resin. The peptide was then treated with TfOH-TFA at 0 ꢀC for
20 min to remove the MeOBn group of Cys12 and was purified by
RP-HPLC, obtaining peptide 20 in 48% yield.
The first interchain disulfide bond between Cys12 and Cys102
was constructed as follows. Peptide 20 dissolved in 50% aqueous
acetonitrile was added dropwise into a solution of peptide 18 in
67 mM NH4HCO3 and 33% acetonitrile within 1 min, and the
The B-chain was also prepared with a peptide synthesizer.
After cleavage from the protected resin with reagent K, the crude

Article
Biochemistry, Vol. 49, No. 8, 2010 1805
solution was kept at room temperature. The reaction was almost
completed within 2 min and was stopped by adding acetic acid.
Peptide 21 was obtained by RP-HPLC in 60% yield.
The final disulfide bond was formed by iodine oxidation.
Peptide 21 was dissolved in 20% aqueous MeOH and added
dropwise into the I2 solution in MeOH containing HCl. After
25 min, the reaction was quenched with ascorbic acid, and the
product was purified by ion-exchange chromatography, followed
by RP-HPLC to give the final product 22 in 39% yield. The
MALDI-TOF mass spectrum and amino acid composition of
peptide 22 corresponded well to the theoretical values. Nongly-
cosylated AGH 23 was also synthesized using the same method
as 22.
FIGURE 8: Circular dichroism spectra: circle, synthetic mono-
saccharyl AGH 22; triangle, synthetic nonglycosylated AGH 23.
CD Spectra and Biological Activities of Synthetic AGHs.
The CD spectra of 22 and 23 were measured in phosphate buffer
at room temperature (Figure 8). The spectrum of 22 was quite
similar to that of 15, demonstrating that the disulfide swapping
did not significantly affect the overall conformation. In contrast,
the spectral pattern of 23 differed slightly from that of 22,
suggesting that the N-glycosylation at Asn112 affected the con-
formation.
The biological activities of 22 and 23 were assessed by in vivo
bioassay. Injection of monosaccharyl AGH 22 caused the
masculinization of female animals at a dose of 200 ng/individual
(three of ten individuals were masculinized), although nonglyco-
sylated peptide 23 showed no biological activity, even at a dose of
1 μg/individual.
activity; however, the semisynthetic AGHs did not, demonstrat-
ing that the native-type disulfide bond arrangement is important
for AGH activity.
Different from other insulin-family peptides, AGH has four
disulfide bonds; it has an additional intrachain disulfide bond
between Cys21-Cys38 in the B-chain. In addition, the arrangement
of the other three disulfide bonds is not identical to that of the
other insulin-family peptides; AGH has an intrachain disulfide
bond in the A-chain with 2-3 pattern rather than the traditional
insulin-like 1-3 pattern. In a pioneering study, two disulfide
isomers of human insulin were prepared by direct chemical
synthesis (26). One of the isomers, which contained non-native
2-3 pattern pairings (CysA7-CysA11, CysA6-CysB7, and CysA20
-
CysB19), showed biological activity comparable to the native
peptide (CysA6-CysA11, CysA7-CysB7, and CysA20-CysB19), indi-
cating that a disulfide swap in insulin is not critical for its
biological activity. In another previous study of bombyxin, an
insulin-family peptide isolated from the silkmoth, disulfide iso-
mers also showed biological activities comparable to the native
peptide (27). These results differ from the present results for
AGH. It has been reported that the N-linked glycan of AGH is
essential for conferring AGH activity, and the results obtained in
the present study support this. These observations suggest that
the glycan moiety may be recognized, in part, by AGH receptor.
It is likely that the disulfide swap of AGH may disrupt the
formation of the folding motif which is critical for binding to the
AGH receptor.
Interestingly, CD spectra of semisynthetic and synthetic
AGHs were similar to those of other insulin-family peptides (22,
23). The CD spectrum of a disulfide isomer of human insulin
showed a lesser R-helical pattern than that of the native pep-
tide (23). CD spectra of native bombyxin and its disulfide isomer
showed that a disulfide swap partially disrupted the helical
structure (24). On the other hand, the spectra of semisynthetic
and synthetic monosaccharyl AGHs were similar, indicating that
the disulfide bond isomerization in AGH did not significantly
affect the CD spectral pattern. These results suggested that the
disulfide swap in AGH did not disrupt the helical conformation
typical of insulin-family peptides. Furthermore, these results also
support that the AGH receptor partially recognizes the glycan
moiety.
DISCUSSION
It is widely believed that a refolding reaction will transform a
protein into its most thermodynamically stable form (25). In our
AGH folding experiments, the semisynthetic AGH folded into a
non-native form with wrong disulfide linkages regardless of its N-
linked glycan structure. These results indicated that the native
AGH conformation is not the most stable at least in vitro.
Our previous report showed that the recombinant AGH
expressed in an E. coli expression system and refolded in a redox
buffer had the same disulfide bond arrangement as that of the
baculovirus recombinant AGH (6). However, in vitro refolding
reaction of the semisynthetic AGH under the conditions identical
to those for the recombinant AGH gave only the disulfide isomer.
These results were contradictory and required reexamination of
the disulfide bond pairing in the recombinant AGH.
In order to cleave the C-peptides from the semisynthetic
AGHs, we examined several commercially available proteases.
A metalloendopeptidase from Grifola frondosa, which specifically
cleaves the peptide bonds at the N-terminal side of a Lys residue,
gave the most favorable results, although the two residues Lys-
Arg were still retained at the N-terminus of the A-chain. An
additional Met residue was also attached at the N-terminus of the
B-chain due to the initial residue of protein translation. When
recombinant AGH was expressed in the baculovirus expression
system, the C-peptide part was cleaved with lysylendopeptidase,
which specifically cleaves the peptide bonds at the C-terminal side
of Lys residues (6). Although this recombinant peptide had a Lys
residue at the C-terminus of the B-chain and a part of the C-
peptide (11 residues) at the N-terminus of the A-chain, it showed
biological activity in vivo at an injected dose of 1 ng/individual (6).
Therefore, it is unlikely that the three additional residues in the
semisynthetic peptide completely abolished the biological acti-
vity. Our synthetic AGH with a glycan moiety showed biological
Our synthetic AGH with a GlcNAc moiety at Asn112 certainly
possessed a biological activity in vivo, although it was much
weaker than that of the native peptide, which showed activity at
an injected dose of 38 pg/individual (6, 19). It is likely that the
GlcNAc moiety may not be insufficient for generating complete
activity, and a larger glycan moiety may be required for full
activity of AGH.

1806 Biochemistry, Vol. 49, No. 8, 2010
Katayama et al.
In a eukaryotic cell, polypeptide chains are synthesized on
ribosomes and then transferred to the endoplasmic reticulum
(ER). For glycoprotein synthesis, Glc3Man9GlcNAc2 is first
attached at an N-glycosylation site by an oligosaccharyltransferase
in the ER and then rapidly trimmed by exoglycosidase to generate
a Glc1Man9GlcNAc2 polypeptide (28). The polypeptide portion is
then folded in the ER. Glycoprotein-specific molecular chaper-
ones, calnexin and calreticulin, are present in the ER and aid in the
glycoprotein folding by binding to the glycan portion of the
Glc1Man9GlcNAc2 polypeptide (29). The folding pathway of
AGH in vivo may require such molecular chaperone(s) in order
to generate a thermodynamically less stable conformation.
It has been reported that human insulin-like growth factor
(IGF) I is unable to form its native disulfide linkages quantitatively
by in vitro refolding reactions, which produces not only the native
conformation but also a disulfide isomer (30). IGF-I is coexpressed
with IGF binding protein (IGFBP) I in the liver (31). It has been
shown that IGF-I is folded quantitatively in the presence of 30 μM
IGFBP-I (32), which suggests that IGFBP-I assists in IGF-I native
disulfide formation. At present, there is no evidence that an AGH
binding protein is present in the androgenic glands to direct the
formation of correct (non-insulin type) disulfides in AGH.
Forming thermodynamically less stable structure is generally
unfavorable for life, because more energy is required for synthe-
sizing and maintaining the unstable structure. Why is AGH
folded into the less stable form? It is likely that the degradation of
the unstable protein structure is faster than that of the more
stable form; i.e., the function of the less stable protein is
controlled by the proteolytic pathway more easily than that of
the stable form. The woodlouse may chose the such system for the
strict control of sex determination.
In conclusion, we synthesized the AGH of A. vulgare with a
homogeneous N-linked glycan moiety by the expressed protein
ligation and total chemical synthesis methods. Semisynthetic
AGHs had disulfide bond arrangements that differed from those
of recombinant AGH prepared in the baculovirus expression
system and showed no biological activities regardless of their
glycan structures. In contrast, the synthetic AGH with a GlcNAc
moiety showed weak biological activity. These results strongly
suggest that the native conformation of AGH is not the most
thermodynamically stable form and that AGH requires mole-
cular chaperone(s) or other factor(s) for forming the correct (A-
chain isomeric form of insulin) disulfide bridge arrangement.
Furthermore, it was suggested that the correct disulfide linkages
are necessary for conferring biological activity.
3. Martin, G., Sorokine, O., Moniatte, M., Bulet, P., Hetru, C., and
Dorsselaer, A. V. (1999) The structure of a glycosylated protein
hormone responsible for sex determination in the isopod, Armadilli-
dium vulgare. Eur. J. Biochem. 262, 727–736
4. Ohira, T., Hasegawa, Y., Tominaga, S., Okuno, A., and Nagasawa,
H. (2003) Molecular cloning and expression analysis of cDNAs
encoding androgenic gland hormone precursors from two porcel-
lionidae species, Porcellio scaber and P. dilatatus. Zoolog. Sci. 20,
75–81
5. Manor, R., Weil, S., Oren, S., Glazer, L., Aflalo, E. D., Ventura, T.,
Chalifa-Caspi, V., Lapidot, M., and Sagi, A. (2007) Insulin and
gender: An insulin-like gene expressed exclusively in the androgenic
gland of the male crayfish. Gen. Comp. Endrocrinol. 150, 326–336
6. Okuno, A., Hasegawa, Y., Nishiyama, M., Ohira, T., Ko, R.,
Kurihara, M., Matsumoto, S., and Nagasawa, H. (2002) Preparation
of an active recombinant peptide of crustacean androgenic gland
hormone. Peptides 23, 567–572
7. Muir, T. W., Sondhi, D., and Cole, P. A. (1998) Expressed protein
ligation: A general method for protein engineering. Proc. Natl. Acad.
Sci. U.S.A. 95, 6705–6710
8. Muir, T. W. (2003) Semisynthesis of proteins by expressed protein
ligation. Annu. Rev. Biochem. 72, 249–289. Muralidharan, V., and
Muir, T. W. (2006) Protein ligation: An enabling technology for the
biophysical analysis of proteins. Nat. Methods 3, 429–438
9. Magne, A, Gerard, D., Hirth, L., and Laustriat, G. (1977) Fluore-
scence study of tobacco mosaic virus protein. Biochim. Biophys. Acta
495, 189–194
10. Liu, J., and Gin, D. Y. (2002) C2-amidoglycosylation. Scope and
mechanism of nitrogen transfer. J. Am. Chem. Soc. 124, 9789–9797
11. Matsuo, I., Nakahara, Y., Ito, Y., Nukada, T., Nakahara, Y., and
Ogawa, T. (1995) Synthesis of a glycopeptide carrying a N-linked core
pentasaccharide. Bioorg. Med. Chem. 3, 1455–1463
12. King, D. S., Fields, C. G., and Fields, G. B. (1990) A cleavage method
which minimizes side reactions following Fmoc sold phase peptide
synthesis. Int. J. Pept. Protein Res. 36, 255–266
13. Tam, J. P., Heath, W. F., and Merrifield, R. B. (1986) Mechanisms for
the removal of benzyl protecting groups in synthetic peptides by
trifluoromethanesulfonic acid-trifluoroacetic acid-dimethyl sulfide.
J. Am. Chem. Soc. 108, 5242–5251
14. Hojo, H., Haginoya, E., Matsumoto, Y., Nakahara, Y., Nabeshima,
K., Toole, B. P., and Watanabe, Y. (2003) The first synthesis of
peptide thioester carrying N-linked core pentasaccharide through
modified Fmoc thioester preparation: Synthesis of an N-glycosylated
Ig domain of emmprin. Tetrahedron Lett. 44, 2961–2964
15. Takano, Y., Hojo, H., Kojima, N., and Nakahara, Y. (2004) Synthesis
of a mimic for the heterogenous surface of core 2 sialoglycan-linked
glycoprotein. Org. Lett. 6, 3135–3138
16. Hojo, H., Matsumoto, Y., Nakahara, Y., Ito, E., Suzuki, Y., Suzuki,
M., Suzuki, A., and Nakahara, Y. (2005) Chemical synthesis of 23
kDa glycoprotein by repetitive segment condensation: A synthesis of
MUC2 basal motif carrying multiple O-GalNAc moieties. J. Am.
Chem. Soc. 127, 13720–13725
17. Ye, Y. H., Li, H., and Jiang, X. (2005) DEPBT as an efficient coupling
reagent for amide bond formation with remarkable resistance to
racemization. Biopolymers 80, 172–178.
18. Yamamoto, N., Takayanagi, A., Sakakibara, T., Dawson, P. E., and
Kajihara, Y. (2006) Highly efficient synthesis of sialylglycopeptides
overcoming unexpected aspartimide formation during activation of
Fmoc-Asn(undecadisialyloligosaccharide)-OH. Tetrahedron Lett. 47,
1341–1346.
ACKNOWLEDGMENT
19. Okuno, A., Hasegawa, Y., and Nagasawa, H. (1997) Purification and
properties of androgenic gland hormone from the terrestrial isopod,
Armadillidium vulgare. Zoolog. Sci. 14, 837–842.
20. Hasegawa, Y., Haino-Fukushima, K., and Katakura, Y. (1987)
Isolation and properties of androgenic gland hormone from the
terrestrial isopod, Armadillidium vulgare. Gen. Comp. Endrocrinol.
67, 101–110.
21. Dawson, P. E., Muir, T. W., Clark-Lewis, I., and Kent, S. B. (1994)
Synthesis of proteins by native chemical ligation. Science 266, 776–
779.
22. Hua, Q.-X., Jia, W., Frank, B. H., Phillips, N. F. B., and Weiss, M. A.
(2002) A protein caught in a kinetic trap: Structures and stabilities of
insulin disulfide isomers. Biochemistry 41, 14700–14715.
23. Hua, Q.-X., Mayer, J. P., Jia, W., Zhang, J., and Weiss, M. A. (2006)
The folding nucleus of the insulin superfamily: A flexible peptide
model foreshadows the native state. J. Biol. Chem. 281, 28131–28142.
24. Maruyama, K., Nagata, K., Tanaka, M., Nagasawa, H., Isogai, A.,
Ishizaki, H., and Suzuki, A. (1992) Synthesis of bombyxin-IV, and
insulin superfamily peptide from the silkworm, Bombyx mori, by
We are grateful to Mr. Hirotaka Haibara of the Graduate
School of Agricultural and Life Sciences, The University of
Tokyo, for providing the optimized conditions for the C-peptide
processing reaction. We also thank Mr. Hiroaki Asai of Otsuka
Chemical Co. Ltd. for providing us Fmoc-Asn[(GalGlcNAcMan)2-
ManGlcNAc2].
REFERENCES
1. Charniaux-Cotton, H. (1954) Discovery in an amphipod crustacean
(Orchestia gammarella) of an endocrine gland responsible for the
differentiation of primary and secondary male sex characteristics.
C. R. Acad. Sci., Paris 239, 780–782.
2. Okuno, A., Hasegawa, Y., Ohira, T., Katakura, Y., and Nagasawa,
H. (1999) Characterization and cDNA cloning of androgenic gland
hormone of the terrestrial isopod Armadillidium vulgare. Biochem.
Biophys. Res. Commun. 264, 419–423.

Article
Biochemistry, Vol. 49, No. 8, 2010 1807
stepwise and selective formation of three disulfide bridges. J. Protein
Chem. 11, 1–12.
25. Anfinsen, C. B. (1973) Principles that govern the folding of protein
chains. Science 181, 223–230.
29. Williams, D. B. (2006) Beyond lectins: The calnexin/calreticulin
chaperone system of the endoplasmic reticulum. J. Cell Sci. 119,
615–623. Ellgaard, L., and Frickel, E. M. (2003) Calnexin, calreticu-
lin, and ERp57: Teammates in glycoprotein folding. Cell Biochem.
Biophys. 39, 223–247.
30. Hober, S., Forsberg, G., Palm, G., Hartmanis, M., and Nilsson, B.
(1992) Disulfide exchange folding of insulin-like growth factor I.
Biochemistry 31, 1749–1756.
31. Humbel, R. E. (1990) Insulin-like growth factors I and II. Eur. J.
Biochem. 190, 445–462.
32. Hober, S., Hansson, A., Uhlen, M., and Nilsson, B. (1994) Folding of
insulin-like growth factor I is thermodynamically controlled by
insulin-like growth factor binding protein. Biochemistry 33, 6758–
6761.
26. Sieber, P. S., Eisler, K., Kamber, B., Riniker, B., Rittel, W., Marki, F.,
and deGasparo, M. (1978) Synthesis and biological activity of two
disulphide bond isomer of human insulin [A7-A11,A6-B7-cystine]-
and [A6-A7,A11-B7-cystine]insulin (human). Hoppe-Seyler’s Z. Phy-
siol. Chem. 359, 113–123.
27. Nagata, K., Maruyama, K., Nagasawa, H., Urushibata, I., Isogai, A.,
Ishizaki, H., and Suzuki, A. (1992) Bombyxin-II and its disulfide bond
isomers: Synthesis and activity. Peptides 13, 653–662.
28. Kornfeld, R., and Kornfeld, S. (1985) Assembly of asparagines-linked
oligosaccharides. Annu. Rev. Biochem. 54, 631–664.