Inr. J . Peptide Protein Reg. 39, 1992, 464-411
Efficient approach to synthesis of two-chain asymmetric cysteine
analogs of receptor-binding region of transforming growth factor- a
JAMES P TAhl and ZHI-YI S H E N
The Rockefeller Uizi\ersirJ,, New York. IVY, USA
Received 28 October 1991, accepted for publication 26 January 1992
The putative receptor-binding region of human transforming growth factor-r (TGFr) has been shown to be
contributed by two fragments: an A-chain (residue 12-18) and a 17-residue carboxyl fragment (residue 3450) that includes a disulfide-containing C-loop (residue 34-43). An approach to the synthesis of two-chain
analogs containing an intermolecular disulfide linked A-chain and the 17-residue carboxyl fragment (Cfragment) possessing receptor-binding activity is described. The synthesis was achieved by the solid-phase
method using the Boc-benzyl protecting group strategy. The single Cys of the A-chain was activated as a mixed
disulfide with 2-thiopyridine to form the intermolecular disulfide bond with Cys41 or Cys46 of the C-fragment
on the resin support. Prior to this reaction, the acetamido (Acm) protecting group of Cys41 or Cys46 was
removed by Hg(OAc)? on the resin support. The peptide and side chain protecting groups including the Smethylbenzyl moiety of the Cys34 and Cys43 \vere concomitantly cleaved by high HF. The intramolecular
disulfide with two unprotected Cys was formed in the presence of an intermolecular disulfide. This intramolecular disulfide bond formation was usually not feasible under the traditionally-held scheme at basic pH since
disulfide interchange would occur faster than intramolecular oxidation. To prevent the disulfide interchange,
a new method was devised. The intramolecular disulfide bond oxidation was mediated by dimethylsulfoxide
at an acidic pH, at which the disulfide interchange reaction was suppressed. The desired product was obtained with a 60-70", yield. In contrast, the conventional scheme of using 12 to form the intermolecular
disulfide between the Cys(Acm) of the A-chain and C-fragment with the preformed intramolecular disulfide
bond in solution phase did not result in any product. The purified two-chain analogs were found to be unstable and rearranged to the homo-diniers. This reaction was greatly accelerated in 12. which explained the
difficulty associated with the conventional scheme. When assayed against A431 and NRK clone 49F cells,
both the A-chain and the C-fragment did not exhibit any biological activity independently, but the two-chain
analogs showcd low receptor-binding activity ivith an lCso at 0.3 niM level. Unexpectedly, dimeric C-fragment,
which resulted from the rearrangement reaction, also showed receptor-binding activity. Our results demonstrate that the two-chain analogs exhibit low but distinct biological activity and provide evidence that the
putative T G F r receptor binding region ma! be discontinuous. In addition, we also provide an efficient approach to further explore the two-chain receptor-binding analogs of T G F r .
KeJ,wrd.y: disulfide formation. asymmetric: disulfidc forination. dimeth!Isulfoxidc-mediated: epidermal growth factor,
receptor-binding region: solid-phase peptide s: nthesis: transforming growth factor-z
Human transforming growth factor type r (TGFr) is a
member of the epidermal growth factor (EGF) family
(1-10). This family of growth factors, which binds to
and activates a common EGF receptor-tyrosine protein
kinase, is important for the inaintenance and growth of
both normal and malignant cells. TGFr and EGFs are
produced by a diverse range of cells and organisms
(1 1-21) and because of common occurrence and rnitogenic functions, they are important therapeutic targets for wound healing and antitumor agents.
464
TGFr contains the essential salient features of the
EGF family. It has 50 amino acids and three disulfide
bonds. The study of structure-function relationships of
TGFz and EGF has been frustrated, however, by the
lack of small peptide analogs that contain high biological activity (27-29). Part of the reason could be the
lack of definitive proof for the exact location of the
receptor-binding region of TGFa.
Two important clues contributc to the understanding
of the putative rcccptor-binding portion of TGFa. First.
Asymmetric cysteine analogs of TGFx
there is a substantial body of point-substituted synthetic analogs of TGFa or EGF with biological activity. Second, both INMR structures of EGF and TGFa
have been determined and found to be similar in the
overall backbone structures (22-26). The conclusion
based on the NMR studies suggests that the B loop,
with its dominant 8-sheet structure, acts as a scaffolding that brings A- and C-loops together, and forms the
receptor-binding region. The receptor-contact residues
(Fig. 1) are likely to comprise three discontinuous fragments: residue 12-18 of the A-chain; the tripeptide
Arg4*-Cys-Glu,of the C-loop; and the dipeptide, Asp4'Leu, of the external COOH fragment, also of the Cloop. These results indicate that the receptor-binding
region of TGFx is likely to comprise discontinuous
fragments spatially orientated by the disulfide-bonded
structure, and the 8-sheet structure of the B-loop. The
putative receptor-binding region is hypothesized to be
composed of three disconnected regions: the A-chain
(H-T-Q-F-C-F-H) consisting of residue 12-18 and the
C-fragment consisting of residue 34-50 which includes
the residue Arg", Asp", and Leu48.It would, therefore,
be useful to prepare receptor-binding analogs by combining both regions with an interchain disulfide linkage.
This will require the synthesis of a two-chain asymmetric cystine TGFa analog (Fig. 1) such as the structures
of compound 1 and 2.
There are several approaches to form the interchain
disulfide chain. A convenient approach is to use nonspecific oxidation in which all sulfhydryls are concomitantly subjected to oxidation to arrive at the desired
product. In the second approach, Cys of both chains
are selectively protected by Acm or trityl (30-32) and
the interchain disulfide is formed by an electrophilic
agent such as 12. Such an approach may lead to the
formation of both hetero- and homo-disulfide products.
To provide specific disulfide formation, the third approach (33) uses thiol disulfide interchange reaction in
which one chain is activated and the other chain contains a free sulfhydryl. In this paper, we describe the
synthesis of the two-chain TGFa receptor-binding region analogs by the latter two approaches. Moreover,
we use dimethylsulfoxide (DMSO) as a new disulfide
and oxidation reagent which overcomes several limitations of the conventional approaches.
FIGURE I
Occurrence of A-chain and C-fragment that forms the putative
receptor-binding rcgion of TGFr.
EXPERIMENTAL PROCEDURES
Muterial and methods
Methylbenzhydrylamine hydrochloride resin (MBHA
resin) was purchased from Advanced Chemtech. PAM
resin and Boc amino acids were purchased from Bachem
Inc. USA; CHzClz from Fisher Scientific, D M F from
Baxter, trifluoroacetic acid from Halocarbon, and N,N'dicyclohexylcarbodiimide was from Fluka Chemical
1-hydroxybenzotrizole, p-cresol, N,N-diisopropylethylamine, dimethylsulfide, and 2,2' -thiopyridine were obtained from Aldrich Chemical.
Amino acid hydrolysis of the free peptide was performed in 5.7 N HCI heating at 110" for 24 h. The
hydrolysates were analyzed on a Beckman 6300 high
performance amino acid analyzer. Analytical and preparative HPLC were performed on Shimadzu and Water's HPLC. Mass spectrometric analysis was determined by electrospray ionization.
Synthesis of C,frugrnent analogs of TGFx
Two C-fragments (residue 34-50): Cys(MeBz1)His ( D n p)- S er (B z I)- G 1y -T yr ( B r Z )-Val - G 1y - A 1aArg ( T o s ) - C y s ( M e B z I ) - G l u ( 0 B z 1)- H i s ( D n p)Cys(Acm)-Asp(OBzl)-Leu-Leu-Ala-OCH2-Pam
resin
and
Cys(MeBz1)-His(Dnp)-Ser(Bz1)-Gly-Tyr(BrZ)Val-Gly-Cy s(Acm)-Arg(Tos)-Cys(MeBz1)-Glu(OBz1)His(Dnp)-Ala- Asp(OBzl)-Leu-Leu-Ala-OCH2-Pam
resin, were synthesized separately with an automatic
peptide synthesizer (ABI 430A) on Boc-Ala-OCH2Pam resin (0.43 g, substitution 0.8 mmol/g) (34-38).
Standard protocols of double coupling with performed
symmetric anhydrides (39) in D M F were used except
for Arg(Tos), Asn, and Gln which were coupled as
hydroxybenzotriazole esters (40). The yield of 1.5 g was
98 ofo based on theoretical yield.
Synthesis ?f A chain corresponding to the residue 12-18
of A-loop (Ac-His-Thr-Gln-Phe-Cys-Phe-His-NHr
)
The A-chain (residue 12- 18) was synthesized manually
on p-methylbenzhydrylamine resin (2.55 g, substitution 0.6 mmol/g) (41). Double couplings were performed with symmetric anhydrides for Cys and Phe in
CH2C12. His(Dnp), Gln and Thr(Bz1) were coupled as
hydroxybenzotriazole esters. At the completion of the
synthesis, the resin was acylated with 45% acetic anhydride in D M F for 10min. The resin was washed
thoroughly by D M F and CH2CL. 3.56 g (theory: 4.35 g)
of peptide resin was obtained (theory: 4.35 g ) to give
82% yield.
The resin (0.36 g) was treated four times with 10%
thiophenol in D M F for 8 h to remove the N'"-Dnp
protecting group of His42 and then with 50"; (v/v)
CF3COOH/CH2Clr for 20 min to remove the N"-Boc
group. The dried peptide resin was treated with the
low/high H F method (43, 44). After HF, the crude
peptide was washed with cold ether three times, extracted into 100mL of 50% acetic acid, diluted to
465
J.P. Tam and Z.-Y. Shen
500 mL. and lyophilized. The lyophilized powder was
purified by preparative C- 18 reverse-phase chromatography.
To remove the Cys(Acm), Hg(0Ac)Z was used (45).
4.82 mg Hg(0Ac)z (15.16 pmol) was added to 7.82 mg
of purified heptapeptide in 0.41 mL HzO (pH was adjusted to 4) while stirring at room temperature. Using
C-18 reverse-phase HPLC to monitor the deprotection
process, the reaction was completed within 1 h. The
solution was diluted to 7 mL, and mercaptoethanol was
added while stirring for 2 h. The mixture was purified
by C18 reverse-phase HPLC by aqueous CH3CN containing 0.0445% TFA with a linear gradient from lo",
CHlCN to 227" in 1 h at 2.5 mLImin. The major peak
was pooled under a stream of Nr and lyophilized to
obtain 4.2mg of product in 58", yield. Amino acid
analysis gave the expected molar ratio of the title compound. Mass spectrometric analysis gave (M + H)'
1030.9 and (M + 2H)2+ 1031.2 (calc. 1032.1).
Prepamtion of
Ac-His- Thr-Gln-Phe-Ci(SPxr)-Phe-His-NH>
Mixed disulfide ,formation oil the peptide. Hg(0Ac)l
(44.36 mg, 0.139 mmol) in 0.44 mL of 3y0 HOAc was
added to the purified heptapeptide (71.98 mg,
0.07 mmol) in 7 mL H.0 and 1 mL of 30,; HOAc for
4 h at room temperature. The mixture was diluted with
water, H2S was bubbled into the reaction solution to
precipitate Hg(I1) salt. After 10 min, the black solution
was filtered through a layer of celite and sodium acetate (0.6 g) was added to the filtrate, followed by 2pyridyl disulfide (154 mg, 0.7 mmol) in 3.5 mL propanol
and 2 mL of 0 . 1 sodium
~
acetate buffer. The solution
turned turbid and another 7 mL propanol and 20 mL
H 2 0 was added to clarify the solution. The reaction
was completed within 15 min based on C18 reversephase HPLC monitoring. After 1 h. the solution was
adjusted to pH 3, filtered and purified by C-18 reverse
phase HPLC by aqueous CH&N containing 0.045",
TFA with a linear gradient from lo"/, CH3CN to 44",
in 45 min. The major peak appeared in the 33", of B
buffer and was pooled. After lyophilization, 72.6 mg of
the desired product (70"; of peptide content) was obtained to give a 71% yield. The amino acid analysis
gave the expected molar ratio and the mass spectrometric analysis gave (M t H) + , 1068.8 and (M + 2H)'+
1069.2 (calc. 1069.3).
On the resin. A-chain peptide resin (130 mg, 42.8 pmol)
with N1"'-Dnp deprotected from His was shaken with
0.1 M Hg(OAc):! in D M F for 4 h, drained, washed with
D M F for four times. and then washed with lo", mercaptoethanol in D M F solution overnight. After the resin
was thoroughly washed by D M F to remove mercaptoethanol completely, it was reacted with 2-pyridyl disulfide (18.83 mg, 85.6 pmol) in D M F solution containing 30°, propanol and 5 drops of 1 M NaHC03. The
466
reaction was monitored by the change of UV absorption at 343 nm. After 22 h, the reaction was completed.
The resin was washed with D M F 3 times and dried in
a desiccator under high vacuum overnight. The resin
was treated with 9 mL H F and 1 mL anisole at 0" for
1 h (44). No thiol or sulfide scavenger should be added
to the H F cleavage step. After the H F cleavage, the
anisole was removed with CH3CN, and the peptide
was extracted into 107, deaerated HOAc, diluted, and
purified by C18 reverse-phase HPLC. The major peak
was pooled and was identical with the title compound
obtained from the method for the peptide (above).
Preparation of the two-chain analogs 1 and 2
C loop (residue 34-50) resin (60 mg, 16.1 pmol) was
treated with 10:; thiophenol in D M F solution for 8 h
four times to remove the Dnp protecting group of His.
Aftcr the resin was washed thoroughly with DMF, the
Cys(Acm) was removed by adding 0.1 M Hg(0Ac)z in
D M F solution and stirring for 6 h. The resin was
drained, washed with D M F again and shaken in 10%
mercaptoethanol D M F solution overnight to remove
Hg(I1) salt. The resin was drained again, washed with
deaerated D M F to remove mercaptoethanol.
Cys(SPyr)-heptapeptide (40.62 mg, 38 pmol) was
added to the resin in 2.43 mL of DMF-propanol and
40 pL DIEA. After the reaction, 24.74 mg of the starting material was recovered, therefore the actual amount
used was 15.85 mg. This reaction was monitored by
HPLC and the change of UV absorption at 343 mn due
to the release of 2-thiopyridone. After 1.5 h, A343 was
0.0749, and 22 h later, A343 increased to 0.2576 and
became steady. Using HPLC, Cys( SPyr)-heptapeptide
was 98"" and 2-thiopyridone was 27,. 1.5 h after the
reaction started. 22 h later, Cys( SPyr)-heptapeptide
was 840; and the thiopyridone became 16% and steady.
The resin was drained and Cys(SPyr)-heptapeptide
(24.74 mg) was recovered. TFA was then used to remove the Boc groups. The resin was washed with D M F
and dried in a dessicator under high vacuum for 1 h.
The resin was treated with H F (9 ml) in the presence
of anisole (1 mL) at 0 a for 50 min. After the removal
of HF, the crude peptide was extracted into 50 mL of
lo", acetic acid. It was diluted into 5% acetic acid
solution, and then 11 mL DMSO was added before
adjusting the pH to 4 with 1.6 mL of NHdOH for disulfide oxidation. The oxidation process was monitored
by HPLC. After 8 h, preparative C18-reverse phase
HPLC was used for purification using aqueous acetonitrile containing 0.0450, TFA with a linear gradient
from 5 a , CN,CN to 25 7; CH3CN in 50 min. The major
product was 43.4 mg and was purified again by a preparative column with a linear gradient from 15% B
to 35", in 35 min and the desired product elated at
B conc. of 32.627, to give 567, yield. Amino acid
analysis revcalcd: Asp( 1):1.1, Thr( 1):0.8, Ser( 1):0.9,
Glu(2):1.8, Gly(2):2.1, Ala(2):2.0, Val(l):l.O, Leu(2):2.2,
Tyr(l):0.8, Phe(2):1.7, His(4):3.5, Arg(1):l.l. Mass
Asymmetric cysteine analogs of TGFa
Spectrometry analysis: theory, 2791.1, found
(M+4H)"+, 2789.6, ( M + 5H)5+,2791.0 (M+ 3H)3+,
2789.4. The major side product was the homo-dimer.
Amino acid analysis showed that Asp( 1):1, Ser( 1):1,
Glu( 1):1, Gly(2):2, Ala(2):2, Val( 1):1, Leu(2):2,
Tyr( 1):0.9, His(2):2, Arg(1): 1. Mass spectrometric analysis theory: 3662.1 found: (M + 3H)3
3661.5,
(M + 4H)4+ 3661.0, (M + 5)5+ 3661.5, (M + 6H)6+
3662.4.
1 a,b
H-T-Q-F-C-F-H-NH,
2 a,b
Ac-H-T-Q-F-C-F-H-NH,
+
Biological ussa,vs
The mitogen assay of synthetic analogs was performed
on normal rat kidney cells in a modified Eagle's medium with 10Yo heat-inactivated calf serum for 24 h and
maintained in a medium containing 0.204 calf serum for
3 days before the assay as described (1). Incorporation
of [ 3H]thymidine was counted after 24 h exposure.
Dose-response curves of the radioreceptor were obtained and performed as described (1). Inhibition of
[ 1'51]-EGF binding to the EGF-receptor was determined by subconfluent monolayers of formalin-fixed
A-43 1 cells after 1 h incubation at 22 with either synthetic TGFa or natural mouse EGF.
O
3 a,b
4 a,b
b= p - A c
FIGURE 3
Two-chain TGFm analogs (1 and 2) which consistcd of A-chain
(residue 14-20) and C-fragmcnt (residue 34-50) of T G F r (the numbering is in accordance with T G F r ) and the two homo-dimers of
C-fragment (3 and 4). The a series contain free r-amino group on the
C-fragments while the b series contains an acetylatcd r-amino group
on the C-fragments. Note that in both series the A-chain is always
acetylated.
RESULTS AND DISCUSSION
Background and rationale f o r a new scheiiie
Our original scheme for the synthesis of the two-chain,
two-disulfide analogs 1 and 2 is shown in Fig. 2. The
single cysteinyl moiety of the A-chain was protected
with the Acm protecting group (46, 47). The three cysteines of the C-fragment were blocked selectively by
two types of protecting groups. The Cys at position 41
(1) or Cys 46 (2) of the C-fragment was protected with
Acm while cysteines at position 34 and 43 were protected with the p-methylbenzyl group. In such a scheme,
the intramolecular disulfide bond at positions 34 and 43
after H F treatment was formed without difficulty by air
oxidation while the remaining Cys(Acm) was then used
to form the intermolecular disulfide with the A-chain by
1 2 oxidation (Fig. 3). However, no intermolecular disulfide product was obtained despite many attempts and
changes of protocol. The major product was the dimeric
C-fragments (Fig. 7).
The difficulty associated with the intermolecular disulfide could be attributed to the instability of the intermolecular disulfide which rearranged to the dimeric
C-H-S-G-Y-V-G-A-R-C-E-H-C-D-L-L~~
12
+
No product
FIGURE 2
Conventional scheme used for selective disulfide formation of the
two-chain T G F r analogs. The desired product was not isolated from
the reaction mixture.
products under the reaction condition (see below for
further explanation). In view of the difficulty, a new
scheme was devised to form the intermolecular disulfide bond prior to the formation of the intramolecular
disulfide bond using the same scheme of the protecting
groups. To achieve this scheme, the thiol group of the
A-chain had to be activated and we chose to use the
mixed disulfide of 2-thiopyridine. However, the key to
success was the formation of the intramolecular disulfide mediated by dimethylsulfoxide at low pH which
minimized the disulfide interchange reaction.
Sjazthesis of the thiol-activated A-chain
The synthesis of Cys( SPyr)-containing and thiolactivated A-chain was accomplished in two ways
(Figs. 4 and 5). In the first approach (Fig. 4) the Achain was synthesized by the solid-phase method and
purified. The Acm protecting group of Cys was removed to allow the sulfur-sulfur bond formation of
Cys(SPyr). In the second approach (Fig. 5 ) all the manipulations including the sulfur-sulfur bond formation
of Cys(SPyr) were performed on the solid support to
allow the attainment of the final product after the H F
cleavage from the resin. The advantage of the first approach was that since some of the byproducts were
removed during the intermediate purification step, the
desired product appearing as a clear major peak that
facilitated both purification and identification (Fig. 7).
The advantage of the second approach was that the
inherent advantage of the solid-phase scheme was exploited to the fullest without manipulation of the intermediates. However, the final product was obtained as
461
J.P. Tam and Z.-Y. Shen
The deblocking of Cys(Acm) with Hg(0Ac)z was
accomplished in 4 h and the Hg2+ salt was removed
with lo", mercaptoethanol in D M F in 18 h. Although
(1) Hg(OAc),
the model study with resin anchored dipeptide containing Cys(Acm) showed that the reaction was facile and
complete removal of Cys(Acm) could be achieved in
0.5 h (45), the long reaction time for the deblocking of
(2) H2S
A-chain served only a prudent measure. 12 or other
electrophilic reagents could also be used to effect the
same deblocking reaction. The second step ofthe mixed
disulfide reaction with pyridyl disulfide was carried out
in a mixed solvent containing 30% propanol. The reAc-His-Thr-Glu-Phe-Cys(S
Pyr)-Phe-His-NH, action was slow as monitored by the UV absorption at
343 nm due to the release of thiopyridone and required
FIGURE 4
22 h for completion. H F cleavage of the peptide resin
Preparation of activated A-chain in solution phase.
in the absence of any thiol or sulfide scavengers produced the desired product in 50% yield (43,44). A side
which accounted for 45% of the major prodAc-His-Thr-Glu-Phe-Cys(Acm)-Phe-His- product
uct was identified as the homo-dimer which was formed
during the activation step.
I
Ac-His-Thr-Glu-Phe-Cy
s(Acm)-Phe-His-NH,
I
1
@
Hg(OAc),
Ititertriolecular disul@de bond formation on resin
The C-fragment was prepared by the solid-phase
method. The protection scheme for the three Cys involved p-methylbenzyl for Cys34 and Cys43 but Acm
for Cys4 1. The two key steps of the disulfide formation
involved the interdisulfide formation between Cys41
and the A-chain, and the intradisulfide formation beAc-His-Thr-Glu-Phe-Cys(S
Pyr)-Phe-Histween Cys34 and Cys43 (Fig. 6). The Acm group of
Cys41 was removed on the resin support by Hg(0Ac)z
and the thiol-activated A-chain was coupled to the resulting free thiol on the resin support in a mixed solvent
of D M F and propanol. The reaction was monitored
Ac-His-Thr-Glu-Phe-Cys(S mr)-Phe-His-NH, by UV absorption at 343 nm due to the release of 2thiopyridone. After 22 h, the reaction was judged to be
FIGURE 5
complete as the level of UV-absorption became steady.
Preparation of activated A-chain in solid-phase.
Ac-His-Thr-Glu-Phe-Cys(
SH)-Phe-His-@
@
a mixture with two other side products which were
separated by C-18 reverse-phase HPLC. The overall
yield by either approach was similar and in the range
of 50%.
In the solid-phase, the deblocking of the Cys(Acm)
and the mixed disulfide formation with thiolpyridine of
the A-chain (Fig. 5) were performed on the resin support (45). There are advantages to performing these
reactions on the resin support. The reagents could be
used in large excess to force the reaction to completion.
Separation of the product and the reagents could be
performed by simple filtration, an inherent advantage of
the solid-phase scheme. However, there are also limitations. The reaction rates in solid-phase are usually
slower than in the solution phase and are often dependent on the rate of diffusion of the reagents. Thus, it is
difficult to predict the time required for the completion
of thc reaction unless a monitoring step is instituted.
The long reaction time may also be accompanied by
side reactions that might escape detection.
468
DitriethylsulJbxideoxide(DMSO) mediated formation of
ititratnoleculur disul@de bond ut low p H
The key to the new approach is the use of DMSOmediated oxidation of the intramolecular disulfide bond
(48) in the presence of the intermolecular disulfide bond
(Fig. 6). DMSO is a mild oxidant selective for disulfide
bond formation in peptides. Facile disulfide bond formation by DMSO in aqueous buffered solutions is
found to proceed in a wide range of pH. Since the
disulfide bond interchange reaction occurs at neutral
and basic pH, the oxidation of the intramolecular disulfide bond can be performed at the acidic pH to minimize the disulfide bond interchange reaction. Furthermore, the rate of the disulfide formation is dependent
on the concentration of DMSO and the rate of the
intramolecular disulfide bond formation can be conveniently adjusted with an increased concentration of
DMSO. We have found that a 20% DMSO at about
3 M concentration would render most disulfidc oxidation complete within 2 h (48).
The side-chain protecting groups and the peptide
Asymmetric cysteine analogs of TGFa
Ac-His-Thr-Glu-Phe-Cys(S
Pyr)-Phe-His-NH,
4-
Sp
34
+
Ac-His-lhr-Glu-Phe-Cys-Phe-His-NH~
I
'i
d-0-CH2@
31
c
HF
Time (min)
1
Sh
SH
1
COOH
DMSO
Ac-His-ThrGlu-Phe- s Phe His MI,
&&
O
iOH
FIGURE 6
Scheme used for preparation of the two-chain analogs.
were cleaved by the high H F procedure (43,44) without the use of any thiol or sulfide scavengers to preserve
the integrity of the intermolecular disulfide bond. The
intermolecular disulfide bond between Cys34 and Cys4 1
was oxidized by the presence of 20% DMSO in the
reaction mixture. The low pH oxidation greatly minimized the disulfide exchange reaction and allowed the
selective intramolecular reaction to occur. After HF,
the crude C-fragment was extracted into deaerated
aqueous acetic acid under argon, and the solution was
adjusted to pH 4 containing 207, DMSO. The oxidation was completed in 8 h as monitored by HPLC. This
reaction could not be achieved by electrophilic agent
such as 12 at acidic pH (vide infra) or air oxidation at
basic pH because of the disulfide interchange reaction.
The two-chain analogs were purified by C 18-reverse
phase HPLC (Fig. 7) without difficulty, and the purified
product gave the expected molar ratio of amino acids
and molecular mass. This scheme of reaction was performed on four two-chain disulfide analogs (Fig. 3) and
the yield, based on the C-fragment, ranged from 6070%. The major side products of the reaction were the
homomeric dimers.
Instability of inter-disul@e bond and formation of
homo-dinieric C-fragment
We were intrigued by the difficulties associated with the
conventional scheme of forming the two-chain analogs
FIGURE 7
Analytical C I 8reverse-phase HPLC. (A) reactants for Fig. 2, purified
A-chain containing Acm (peak 1) co-injected with the purified Cfragment also containing Acm (peak 2). Crude reaction products of
activated A-chain with 3-thiopyridine prepared by the solution
method (B, see Fig. 4) and by the solid-phase method (C, see Fig. 5).
The desired product (peak 3 and 6 ) is found in the mixture with thc
homo-dimers (4, 7, 8). The 2-thiopyridone is often the most prominent peak in these spectra (peaks 5. 9, and 12). Crude (D) and purified (E) product of the disulfide-linked A-chain and C-fragment.
The desired product is indicated by peaks 10 and 13. Peak 11 is the
homo-dimer.
(see Fig. 2). Several laboratories have shown successfully that such a scheme works well for heteromeric
disulfide formation (30-33). One probable explanation
for our difficulty may be the presence of four histidines,
two on each chain. The imidazole side chains could be
correctly positioned to catalyze the intermolecular disulfide bond disproportionation and subsequent rearrangement into homo-dimers. Purified two-chain analogs l and 2 were tested for their stability at pH 3-4.
30% of the homo-dimeric product of C-fragment from
1 was observed in 5 days. The rate of rearrangement of
2 was 3-fold faster than 1. However, the rate of rearrangement was accelerated nearly 20-fold in the presence of I 2 and iodide ion. These results show that the
two-chain analogs are unstable under the scheme of I2
oxidation and would rearrange rapidly under the reaction condition to the more stable homo-dimeric products.
Biological uctivity
The primary in vitro assays for TGFa have been the
mitogenic and radioreceptor assays using the normal
rat kidney fibroblasts (NRK) and the human epidermal
carcinoma A43 1 cells which are rich in EGF receptors,
respectively. For all synthetic analogs, the receptorbinding assay in concentrations as high as 1 mM was
used for screening biological activity and those analogs
469
J.P. Tam and Z.-Y. Shen
TABLE 1
Biological activirj of rnv-chuiri uridogs 1-4
Analog'
Biological activit) (nlM)
Receptor-binding
IC50
A-chain
C-fragment
la o r b
?a or b
3
4
'
Mitogenicity'
EDw
>I
>I
>I
0.1
>I
>I
0.3
1
(I 1
0.03
0. I
0.3
1 - Receptor-binding was performed on ,4431 cells.
2 - Mitogenicity was performed on NRK clone 49 cells
3 - See Fig. 3 for structures.
We have also shown that the two-chain analogs posses low biological activity. The low biological activity of
analogs 1 and 2 could result from the conformation and
orientation of the A-chain posed by the interiiiolecular
disulfide bond not being optimal. However, this approach will allow further exploration of two-chain analogs to define the receptor-binding region of T G F K
ACKNOWLEDGMENT
This vork was supported by PHS Grant number CA36544. awarded
b> the National Canccr Institute. and DHHS. We thank Dr. B. Chait
of the Rockefeller University Mass Spectrometry Center for the mass
spectrometric analyses, Dr. Paul Haser for his comments. and Ms.
Jennq Park for the preparation of the manuscript.
REFERENCES
that displayed activity higher than 1 mM were considered inactive (Table 1). Both the A-chain and the Cfragment were found to be inactive based on the above
criterion. The two-chain analogs 1 and 2 were found to
contain low receptor-binding activity with an ICS,, in
the range 0.1-0.3 mM which was about 10000-fold less
active than TGFx. There was essentially no differences
between the acetylated or non-acetylated analogs of 1
and 2. Unexpectedly, the homo-dimeric analogs of the
C-fragment 3 and 4 were found to also contain low
receptor-binding activity with an ICSOranging from
0.3-0.4 mM. More interestingly, their mitogenicity with
an E G Oat 0.03-0.1 m M was fold more active than the
two-chain analogs.
CONCLUSION
Our results describe a new approach for the formation
of two-chain, two-disulfide peptides for structurefunction relationship studies. The salient features of our
approach are as follow:
1
This approach fully exploits the attendant advantages of the solid-phase scheme. The activation of
the A-chain to form the sulfur-sulfur bond of mixed
disulfide with thiopyridine, the deprotection of the
Cys(Acm) groups, and the formation of the intermolecular disulfide bond can be performed on the
solid-phase scheme.
Contrary to conventional thinking, the disulfide bond
is relatively stable in the high H F condition in the
absence of thiol or sulfide scavcngers. This advantage can be exploited to allow the formation of the
intermolecular disulfide bond between two chains
on the resin support to increase yield and rccycling
of the excess reagent.
The formation of an intramolecular disulfide bond
from two unprotected Cys in the presence of the
intermolecular disulfide bond can bc accomplished
selectively and efficiently by DMSO-mediated oxidation at an acidic pH.
470
1. DeLarco. J.E. & Todaro. G.J. (1978) Proc. N d .Acrid. Sci. USA
75. 400 1-4005
7.
_
Ozanne, B., Fulton. J . 8: Kaplan, P.L. (1980) J . Cell, Physiol.,
163- 180
3. hfoses, H.L. & Robinson, R.A. (1982) Fed. Proc., 3008-3011
1.Marquardt, H., Hunkapiller. M.W.. Hood. L.E., Twardzik. D.R.,
DeLarco. J . & Todaro. G.J. (1983) Science, 1079-1082
_
5 . Massague. J. (1983) J . Biol. Chem., 13606-13613
6. Lee, D.C.. Rose. T.M., Webb, N.R. & Todaro, G.J. (1985)
.~t7r!ll'e/f1JIldOll~ ??. 489-491
7. Derynck. R . Roberts, A.B.. Winkler, M.E., Chen. E.Y. & Goeddel. D.V. (1983) Cell 41. 287-297
8. Savage, C.R.. Jr.. Inagami, J. & Cohen. S . (1972) J. Biol. Chem.,
76 12-7621
9. Gregory. H. ( 1 975) Norlire (Lorrdoir), 325-327
10. Stroobant. P., Rice, A P . . Gullick, W.J.. Chcng, D.J., Kcrr, I.M.
& Waterfield. M.D. (1985) Cell 42, 383-393
1 1 . Chang, W.. Upton, C., Hu, S.-L., Purchio, A.F. & hlcFadden,
G. (1987) Mol. Cell. Biol. 7, 535-540
12 Upton, C.. Macen, J.L. & McFaddcn, G . (1987) J . Vird 61,
1271-1275
13. Gra), A.. Jull. T.J. & Ullrich, A. (1983) Nature (London) 303,
722-725
11 Scott, J.. Urdea. M.. Quiroga. M.. Sanchez-Pescador, R.. Fong,
N.. Selb). M., Rutter. W.J. 61 Bell, G.I. (1983) Scierice 211,
236-240
15. Doolittle. R.F., Feng, D.F. & Johnson, M . S . (1984) Nnrure 307,
558-560
16. Blornquist. M.C., Hunt. L.T. & Barker, W.C. (1984)Proc. Narl.
Actrd. Sci. U S A 81. 7363-7367
17. Saski. hl.. Kato, S., Kohno, K., Martin, G.R. & Yarnada, Y.
(1987) Proc. ,2'arl. Acad. Sci. U S A 84. 935-939
18. Jackman. R.W.. Beeler, D.L.. Fritze, L., Soff, G . & Roseberg,
R.D. ( 1987) Proe Nrrtl. Acad. Sci. USA 84, 6425-6429
19. "harton. K.A., Johansen. K.M., Xu, T. & Artavanis-Tsakonas,
S. (1985) Cell 43, 567-58 I
20. Greenwald. I . (1985) Cell 43. 583-590
2 1 Hursh, D.A.. Andrew. M.E. & Raff, R.A. (1987) Scieirce 237,
1187-1190
??
__
Maya, K.H (1985) Biochemistr~24, 3783-3794
23 Marco, -4.D , Mayo, K.H.. Bartolotti. F., Scalia, S., Menegatti,
E. & Kaplcin. R. (1986)J. Bid. Chein. 261, 13510-13516
34 hlontclionc. G.T.. Wuthrich. K., Nice, E.C., Burgess. A.W. &
Scheraga. H.A. (1986) /'roc. N u / / .Actrd. Sci. U S A 83, 8594-8598
Asymmetric cysteine analogs of TGFa
25. Cooke. R.M., Wilkinson. A.J., Baron, M., Pastore, A., Tappin,
M.J., Campbell, I.D., Gregory, H . & Sheard, B. (1987) Nature
327, 339-341
26. Montellione, G.T., Wutrich, K., Nice, E.C., Burgess, A.W. &
Scheraga, H.A. (1987) Prnc. Nutl. Acad. Sci. USA 84,5226-5230
27. Tam, J.P., Lin, Y.-Z., Liu. W., Wang, D.-Z.,Ke, X.-H.&Zhang,
J.-W. (1991) Int. J. Peptide Protein Res. 38, 204-21 1
28. Konioriya, A,, Hortsch, M., Meyers. C., Smith, M., Kanety, H.
& Schlessinger, J. (1984) Prnc. Natl. Acud. Sci. USA 81. 13511355
29. Nestor, J.J. Jr., Newman, S.R.. DeLustro, B., Todaro, G.J. &
Schreiber, A.B. (1985) Einchem. Binphys. Res. Cnmmurt. 129,
226-232
30. Kamber, B., Hartniann, A., Eisler, K., Riniker, B., Rink, H.,
Siebcr & Rittel, W. (1980) Heh. ChZn7. Actu 63, 899-901
31. Roman], S., Moroder, L.. Gohring, W., Scharf, R., Wunsch, E.,
Baede, Y.A. & Thocnen, H. (1987) Int. J. Peptide Protein Res..
107-117
32. Wunsch, E., Moroder, L.. Gohring-Romani, S., Musiol. H.-J.,
Gohring, W. & Bovermann, G . (1988) Int. J . Peptzde Protein Res.
32, 368-383
33. Kullmann, W. & Gutte, B. (1978) Int. J . Peptide Protein Res. 12,
17-26
34. Merrifield, R.B. (1963) J . Am. Chem. Snc. 85, 2149-2154
35. Merrifield, R.B. (1986) Science 232, 341-347
36. Mitchell, A.R.. Erickson, B.W., Ryabtsev, M.N., Hodges, R.S.
& Merrifield, R.B. (1976) J. A m . Chenz. Soc. 98, 7357-7362
37. Tam, J.P., Kent, S.B.H., Wong. T.W. & Merrifield, R.B. (1979)
Sjnthesis, 955-957
38. Tam, J.P. (1985) J . Org. Chenz. 50, 5291-5298
39. Hagenmaier, H. & Frank, H. (1972) Hnppe-Sevlers 2. Ph.vsiol.
Chenz. 353, 1973-1976
40. Konig, W. & Geiger, R. (1970) Cheni. Eer. 103, 788-798
41. Matsueda, G.R. & Stewart, J.M. (1981) Peplides 2, 45-50
42. Lin, M.C., Guttle, B., Caldi, D.G., Moore, S. & Merrifield. R.B.
(1972) J . Bid. Chem. 247. 4768-4772
43. Tam, J.P., Heath, W.F. & Merrifield. R.B. (1983) J . Am. Cheni.
Snc. 105, 6442-6455
44. Sakakibara, S. (1971) in Chemistry und Eiochernist~rqf Amino
Acids, Peptides, und Pvntein.r (Weinstein, B.. ed.), Vol. 1, pp. 5185, Dekker, New York
45. Liu, W., Shiue, G.H. & Tam, J.P. (1990) Peprides: Chefnistrj..
Structure and Einlngy (Rivier, J.E. and Marshall. G.R.. eds.),
pp. 271-273. Escom Science Publishers, Leiden
46. Veber, S.,Milkowski, J.D., Varga, S.L.. Denkewalter, R.G. &
Hirschmann, R. (1972) J . Am. Chem. Snc. 94, 5456-5461
47. Kamber, B., Hartmann, A,, Eisler, K., Rmiker, B., Rink, H . ,
Sieber, P. & Rittel, W. (1980) Hels. Chiwi. Acru 63, 899-915
48. Tam, J.P., Wu, C.W., Liu, W. & Zhang, J.W. (1991) J . A m .
Chem. Snc. 113, 6657-6662
Address:
James P. Tam
The Rockefeller University
1230 York Ave.
New York
NY 10021
USA
Tel (212)-570-8433
Fax (212)-570-7779
411