An iodine-free and directed-disulfide-bond-forming route to insulin analogues

Article abstract of DOI:10.1021/ol501252b

An iodine-free synthetic route to insulin analogues has been established via a directed disulfide bond formation strategy. This method is completely compatible with oxidation-sensitive residues. The key step is constructing the third disulfide bond via a novel procedure involving phenylacetylaminomethyl group (Phacm), immobilized Penicillin G Acylase, and Ellman's reagent. We expect that this method could be broadly utilized for synthesizing insulin-like and other cysteine-rich peptides, in particular, where oxidation-sensitive residues are present in the sequence.

Full text of DOI:10.1021/ol501252b

Letter
pubs.acs.org/OrgLett
Terms of Use
An Iodine-Free and Directed-Disulfide-Bond-Forming Route to
Insulin Analogues
Fa Liu,* Qingyuan Liu, and Adam R. Mezo
Lilly Research Laboratories, Indianapolis, Indiana 46285, United States
S
* Supporting Information
ABSTRACT: An iodine-free synthetic route to insulin analogues has been
established via a directed disulfide bond formation strategy. This method is
completely compatible with oxidation-sensitive residues. The key step is
constructing the third disulfide bond via a novel procedure involving
phenylacetylaminomethyl group (Phacm), immobilized Penicillin G Acylase, and
Ellman’s reagent. We expect that this method could be broadly utilized for
synthesizing insulin-like and other cysteine-rich peptides, in particular, where
oxidation-sensitive residues are present in the sequence.
nsulin, relaxin, and other members of the insulin superfamily
are biologically essential and structurally unique peptide
employed once or twice.4 Although this condition can
I
efficiently remove Acm (acetylaminomethyl) groups and
simultaneously form the corresponding disulfide bond, it may
also modify oxidation-sensitive residues, such as methionine,4h,5
tryptophan,6 and tyrosine.4a,7 In fact, all three of these residues
are frequently found in the native sequences of peptides from
the insulin superfamily.1,8 To avoid possible side reactions
induced by I2 oxidation, we report herein an iodine-free,
directed-disulfide-bond-forming procedure to synthesize this
class of structurally unique peptides. A porcine insulin analogue
1 with three mutations, Leu to Val at A16, Phe to Met at B1,
and Leu to Met at B17, was chosen as an illustrative target
molecule (Figure 1).
hormones.1 They consist of two individual chains, A-chain and
B-chain, which are locked together by three disulfide bonds
including one intra-A chain and two interchain (Figure 1).
Figure 1. Sequence of porcine insulin analogue 1 (red color highlights
In a recent communication, we reported that insertion of
isoacyl dipeptide segments into the A- and B-chains of human
insulin could significantly improve their solubility in aqueous
acetonitrile.4a This modification enabled their purification via
C18 reversed-phase HPLC, which provided clean A- and B-
chains suitable for ligation. Upon consideration of how to
synthesize insulin analogue 1, a similar approach could provide
A−B dimer I containing the A6−A11 and A20−B19 disulfide
bonds (Scheme 1). Both of these disulfide bonds could be
formed via a 2,2′-dithiobis(5-nitropyridine) (DTNP)-directed
thiolysis where I2-oxidation would not be involved. To
construct the last disulfide bond (A7−B7) also under I2-free
conditions, we propose a novel procedure involving both the
Phacm (phenylacetylaminomethyl) group9 and DTNB (Ell-
man’s reagent or 5,5′-dithiobis(2-nitrobenzoic acid)).10 Phacm
groups can be removed by using the enzyme Penicillin G
Acylase (PGA) or its immobilized form (iPGA), which
the mutations).
Chemical synthesis of these peptides has been of long-standing
interest due to the needs from a SAR (structure−activity
relationship) perspective as well as the attractive nature of the
challenging chemistry itself.2 One of the main hurdles of these
syntheses is the efficient formation of three native disulfide
bonds. This has been achieved using two distinct strategies:
protein folding and directed disulfide bond formation. The
former approach relies on a thermodynamic process, which
restricts its applications to analogues with foldability similar to
or better than that of their native parents.3 Conversely, the
latter approach utilizes orthogonal protecting groups for each
pair of cysteines, which permits unambiguous stepwise
formation of each distinct disulfide bond.4
The directed-disulfide-bond-forming strategy can theoret-
ically generate analogues with any possible mutation; therefore
it is more suitable for SAR studies, where structural diversity is
highly desired. However, in all reported existing routes using
this strategy, a harsh I2 (iodine) oxidation step has been
Received: May 1, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ol501252b | Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 1. Proposed Phacm/DTNB/iPGA-Mediated
Strategy for A7−B7 Disulfide Bond Formation
Scheme 2. Synthesis of (Disulfide A6−A11, Isoacyl A8−A9,
CysA7-Phacm) A-Chain 2 (Isoacyl dipeptide segment is
indicated by a blue box; red color indicates mutation)
selectively recognizes and cleaves the phenylacetyl moiety,
followed by the release of formaldehyde and subsequent
deprotection of the sulfhydryl group.9
through the β-carboxyl of Fmoc-Asp-OtBu, which would be
converted to the native Asn of residue A21 after resin cleavage.
The resulting resin-bound peptide 3 was treated with 20% BME
(β-mercaptoethanol) to remove the STmp moiety on CysA6
and subsequently treated with DTNP and DIEA (N,N-
diisopropylethylamine) to activate CysA6 as Cys-SNPy [(4-
nitro-2-pyridinyl)thiol]. The Mmt group on CysA11 of resin 4
was then removed by 1% TFA, and the unmasked free thiol
quickly reacted with CysA6-SNPy to form the intrachain
disulfide A6−A11 to provide resin 5. Peptide cleavage from
resin 5 followed by purification by C18 RP-HPLC afforded A-
chain 2 with an overall yield of 12% based on the substitution
of the starting Rink amide resin.
When preparing A-chain 2, the StBu (tert-butylthiol) group
was also tested as a protecting group of CysA6 in resin-bound
peptide 3; however, the complete removal of this StBu was not
achieved even after overnight treatment with 25% BME in
DMF. In addition, when preparing resin 4, it was found that
DIEA was needed to drive CysA6 activation (by DTNP) to
completion. Both of these observations on CysA6 were
different from those found in a previous procedure,4a
presumably due to the bulky Phacm group attached on the
adjacent residue CysA7. The bulkiness and hydrophobicity of
the Phacm group may also explain why the yield of A-chain 2
was lower than that of a similar A-chain with CysA7 protected
by an Acm group that was reported previously.4a The Dpm
(Diphenylmethyl) group was also evaluated as a protecting
group for CysA20 in resin-bound peptide 3, since it was
reported to be a better Mmt-orthogonal group than Trt when
being treated with 1% TFA.13 However, it was found that the
purity of crude A-chain 2 that resulted from CysA20-Dpm was
similar to that of CysA20-Trt.
In the new proposed procedure, both CysA7 and CysB7 in
A−B dimer I are protected by Phacm groups (Scheme 1).
When dimer I is treated with iPGA,9 a sulfhydryl group will be
liberated upon the removal of one Phacm group (on CysA7 or
CysB7; CysA7 is shown as an example in Scheme 1). We
hypothesized that, in the presence of DTNB, this newly
generated free thiol will be immediately capped and
concomitantly activated to provide intermediate II. This
capping and activation step is expected to take place prior to
potential disulfide bond scrambling, driven by the faster
reaction kinetics of DTNB vs a Cys-Cys disulfide when
exposed to sulfhydryl groups. Subsequent removal of the
second Phacm group (on CysB7 or CysA7) results in
intermediate III, which will quickly undergo A7−B7 disulfide
bond formation via a DTNB-directed thiolysis to afford final
product IV. The possible capping of the sulfhydryl group in
intermediate III by DTNB, which would form a double-capped
side product, is not expected to be of concern since formation
of the A7−B7 disulfide bond, an intramolecular process, is
expected to be much faster than capping by DTNB, an
intermolecular process. However, one possibility which has not
been ruled out is that the faster kinetics of this intramolecular
reaction may be compromised by an unfavorable conformation
of certain analogues, such as ones with less foldability.
To examine if protein foldability could potentially impact the
utility of the proposed strategy, a folding-prohibiting mutation,
Leu to Val at residue A16, was incorporated into porcine insulin
analogue 1. The placement of Val at A16 was previously
reported to block both insulin chain combination and the in
vitro refolding of proinsulin.11 Moreover, to confirm its
compatibility with oxidation-sensitive residues, two methio-
nines were included in the sequence by mutating Phe to Met at
B1 and Leu to Met at B17 (Figure 1).
(Disulfide A6-A11, Isoacyl A8-A9, CysA7-Phacm) A-chain 2
was prepared via a similar scheme as described previously.4a
Each cysteine was protected with orthogonal protecting groups:
CysA6-STmp (2,4,6-trimethoxyphenylthio),12 CysA7-Phacm,
CysA11-Mmt (4-methoxytrityl), and Cys20-Trt (Scheme 2).
The synthesis began by loading Asp to Rink amide resin
(Isoacyl B26−B27, CysB7-Phacm, CysB19-SNPy) B-chain 6
was assembled on ChemMatrix resin14 with cysteines protected
as CysB7-Phacm and CysB19-Trt. After a simultaneous resin
cleavage and CysB19 activation by DTNP,4a,b,15 a purification
step using a C18 RP-HPLC column provided B-chain 6 with a
yield of 22%.
The chain ligation of A-chain 2 and B-chain 6 was completed
within 60 min in an acetate buffer (pH 4.5) and afforded
[Disulfide (A6−A11, A20−B19), Isoacyl (A8−A9, B26−B27)]
A−B dimer 7 in 60% yield (Scheme 3). To form the A7−B7
B
dx.doi.org/10.1021/ol501252b | Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Finally, in a moderate scale-up reaction, 24 mg of isoacyl
dimer 7 afforded 8 mg of porcine insulin analogue 1 with a
yield of 35% without detection of any oxidation of methionine
or tyrosine (Scheme 3 and Figure 2). In contrast, direct
Scheme 3. Synthesis of ValA16, MetB1, MetB17-Porcine
Insulin 1 via an I2-Free Route (Isoacyl dipeptide segment is
indicated by a blue box; red color indicates mutation)
disulfide bond, Tris buffer (pH 8.0) was chosen since it
provided sufficient solubility (>2 mg/mL) for isoacyl dimer 7.
Upon dissolution in Tris buffer (pH 8.0), the O-to-N shift of
isoacyl dimer 7 was completed within 10 min and generated all-
amide-backbone [Disulfide (A6−A11, A20−B19)] A−B dimer
8. Subsequently, iPGA and DTNB were added to facilitate
removal of the Phacm groups and the formation of the A7−B7
disulfide bond. 5.5 U of iPGA was used for every μmol of
Phacm group,9 and this ratio completely removed Phacm
groups of A−B dimer 8 within 3 h.
To identify the optimal amount of DTNB needed to achieve
the best yield of insulin analogue 1, a variety of molar ratios of
DTNB to isoacyl dimer 7 were evaluated, including 0, 0.3, 0.7,
1, 2, 4, and 8. It was found that 2 to 4 equiv of DTNB offered
the best crude product profile as judged by analytical HPLC
(Figure S4). Both desired product 1 and disulfide-bond-
scrambled side products were observed when 0.7 or 1 equiv of
DTNB was used. No product was identified when 0 or 0.3
equiv of DTNB was used. The hypothesized intermediate II
(Scheme 1) was observed when more than 1.0 equiv of DTNB
was used; comparatively, the DTNB-double-capped side
product was only detectable when more than 4 equiv of
DTNB was used. DMSO (10%, v/v) was also tested as a
DTNB surrogate for forming disulfide bonds; however, only
scrambled side products were observed (Figure S4). We
speculate that this is due to the slower reaction kinetics for the
DMSO oxidation as compared to competing disulfide bond
scrambling at pH 8.0. 2,2′-Dithiodipyridine (DTDP) and
DTNP were also attempted as thiol capping and activating
agents; however, neither reagent possessed sufficient solubility
in Tris buffer (pH 8.0) to enable their evaluation.
Interestingly, des-MetB1 porcine insulin 1 was detected by
mass spectrometry as a side product in the amount of 10−20%
of the desired product 1 (Figure S3). This may be due to the
PGA enzyme, known as a hydrophobic-group-preferred N-
terminal hydrolase,16 possibly recognizing methionine at B1 as
a weak substrate. To investigate if pH had an impact on the
observed aminopeptidase-like activity, 100 mM phosphate
buffer (pH 7.4) was also tested as the solvent; 10% DMSO was
added to facilitate full dissolution of isoacyl dimer 7 at a
concentration of ∼2 mg/mL. It was found that there was no
significant difference in the ratio of Des-MetB1-1 to 1;
however, the reaction kinetics in phosphate buffer (pH 7.4)
was significantly slower than in Tris buffer (pH 8.0) (Figure
S4).
Figure 2. Synthesis of ValA16, MetB1, MetB17-porcine insulin 1 via
an I2-free route (HPLC traces at γ = 220 nm).
treatment of isoacyl dimer 7 or dimer 8 with I2 only resulted in
methionine-oxidized products, a portion of which was also
mono- or bis-iodinated (Figures S5 and S6). This confirmed
that the reported method is fully compatible with oxidation-
sensitive residues. The overall yield of chain ligation and
Phacm/DTNB/iPGA-mediated A7−B7 disulfide bond forma-
tion was 21%.
The native disulfide bonding pattern of insulin analogue 1
was unambiguously confirmed by Glu-C (Staphylococcus aureus
Protease V8) digestion of 1, with comparisons to all three
possible disulfide isomers of the A(5−17)−B(1−13) fragment
(Table S2 and Figures S7−S9). These three disulfide bond
isomers were all generated by independent synthetic routes
(Figures S10−S12). The protein structure of synthetic analogue
1 was also confirmed as nearly identical to human insulin by
comparing their far UV-CD spectra (Figure S13). Its biological
activity was found to be similar to human insulin as determined
by their binding affinities to the insulin receptor (Table S3).
Both the structural and biological data of synthetic insulin
analogue 1 agreed well with previously reported data for the
ValA16-insulin analogues.11
In conclusion, we have developed a concise I2-free and
directed-disulfide-bond-forming approach to synthesize insulin,
and potentially relaxin and other insulin-like peptides. The key
reaction involving the formation of the last interchain disulfide
bond was accomplished by treating a precursor peptide
containing two Cys-Phacm groups with the enzyme iPGA in
the presence of Ellman’s reagent. In this conversion, the first
newly generated free thiol upon removal of one Phacm group
by iPGA is immediately capped and activated by Ellman’s
reagent. Subsequently, unmasking of the second Cys-Phacm
moiety concurrently facilitates formation of the last disulfide
bond. The successful synthesis of an unfoldable porcine insulin
analogue 1 suggests that this Phacm/DTNB/iPGA-mediated
disulfide bond formation method is fully directed by chemical
reactivity and independent of the protein foldability. This novel
strategy adds an additional orthogonal approach to the existing
collection of peptide disulfide bond formation methods. Due to
the frequent occurrence of oxidation-sensitive residues in the
native sequences of cysteine-rich peptides, we envision that this
strategy could be broadly utilized for the preparation of these
synthetically challenging peptides.
C
dx.doi.org/10.1021/ol501252b | Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Rosengren, K. J.; Tregear, G. W.; Wade, J. D. ChemBioChem 2008, 9,
1816−1822.
ASSOCIATED CONTENT
* Supporting Information
The solid-phase peptide synthesis; the screening of optimal
molar ratio of DTNB to isoacyl A−B dimer 7; the disulfide
bond mapping of insulin 1. This material is available free of
charge via the Internet at http://pubs.acs.org.
■
S
(5) (a) Chipiso, K.; Mbiya, W.; Morakinyo, M. K.; Simoyi, R. H. Aust.
J. Chem. 2014, 67, 626−635. (b) Young, P. R.; Hsieh, L. S. J. Org.
Chem. 1982, 47, 1419−1423.
(6) (a) Sieber, P.; Kamber, B.; Riniker, B.; Rittel, W. Helv. Chim. Acta
1980, 63, 2358−2363. (b) Hartdegen, F. J.; Rupley, J. A. J. Am. Chem.
Soc. 1967, 89, 1743−1745.
(7) (a) Morikawa, S.; Kanatani, A.; Kobayashi, R.; Yoshimoto, T.;
Tsuru, D. Agric. Biol. Chem. 1991, 55, 2751−2756. (b) Richards, H. R.;
Speakman, J. B. Nature 1953, 171, 751.
AUTHOR INFORMATION
Corresponding Author
■
(8) Conlon, J. M. Peptides 2001, 22, 1183−1193.
*E-mail: liufx@lilly.com.
Notes
(9) (a) Gongora-Benitez, M.; Basso, A.; Bruckdorfer, T.; Royo, M.;
Tulla-Puche, J.; Albericio, F. Chem.Eur. J. 2012, 18, 16166−16176.
(b) Royo, M.; Alsina, J.; Giralt, E.; Slomcyznska, U.; Albericio, F. J.
Chem. Soc., Perkin Trans. 1 1995, 1095−1102.
The authors declare no competing financial interest.
(10) Ellman, G. L. Arch. Biochem. Biophys. 1959, 82, 70−77.
(11) Liu, M.; Wan, Z.-L.; Chu, Y.-C.; Aladdin, H.; Klaproth, B.;
Choquette, M.; Hua, Q.-X.; Mackin, R. B.; Rao, J. S.; De Meyts, P.;
Katsoyannis, P. G.; Arvan, P.; Weiss, M. A. J. Biol. Chem. 2009, 284,
35259−35272.
ACKNOWLEDGMENTS
■
We thank Mr. Steven D. Kahl (Eli Lilly) for conducting the
insulin receptor binding assay and Ms. Caitlyn M. Krukau (Eli
Lilly) for performing CD analysis.
(12) Postma, T. M.; Giraud, M.; Albericio, F. Org. Lett. 2012, 14,
5468−5471.
REFERENCES
■
(13) Gongora-Benitez, M.; Mendive-Tapia, L.; Ramos-Tomillero, I.;
Breman, A. C.; Tulla-Puche, J.; Albericio, F. Org. Lett. 2012, 14, 5472−
5475.
(1) (a) Litwack, G.,; Ed. Vitamins and Hormones: Insulin and IGFs, 1st
ed.; Elsevier: 2009. (b) Chan, L. J.; Hossain, M. A.; Samuel, C. S.;
Separovic, F.; Wade, J. D. Protein Pept. Lett. 2011, 18, 220−229.
(c) Schwabe, C.; Bullesbach, E. E. FASEB J. 1994, 8, 1152−1160.
(2) (a) Belgi, A.; Hossain, M. A.; Tregear, G. W.; Wade, J. D.
Immunol., Endocr. Metab. Agents Med. Chem. 2011, 11, 40−47.
(b) Wade, J. D.; Lin, F.; Hossain, M. A.; Shabanpoor, F.; Zhang, S.;
Tregear, G. W. Ann. N.Y. Acad. Sci. 2009, 1160, 11−15. (c) Mayer, J.
P.; Zhang, F.; DiMarchi, R. D. Biopolymers 2007, 88, 687−713.
(3) (a) Avital-Shmilovici, M.; Mandal, K.; Gates, Z. P.; Phillips, N. B.;
Weiss, M. A.; Kent, S. B. J. Am. Chem. Soc. 2013, 135, 3173−3185.
(b) Sohma, Y.; Hua, Q.-X.; Whittaker, J.; Weiss, M. A.; Kent, S. B. H.
Angew. Chem., Int. Ed. 2010, 49, 5489−5493. (c) Sohma, Y.; Kent, S. B.
H. J. Am. Chem. Soc. 2009, 131, 16313−16318. (d) Tofteng, A. P.;
Jensen, K. J.; Schaffer, L.; Hoeg-Jensen, T. ChemBioChem 2008, 9,
2989−2996. (e) Marglin, A.; Merrifield, R. B. J. Am. Chem. Soc. 1966,
88, 5051−5052. (f) Katsoyannis, P. G.; Tometsko, A.; Zalut, C. J. Am.
Chem. Soc. 1966, 88, 166−167. (g) Kung, Y. T.; Du, Y. C.; Huang, W.
T.; Chen, C. C.; Ke, L. T. Sci. Sin. (Engl. Ed.) 1965, 14, 1710−1716.
(h) Meienhofer, J.; Schnabel, E.; Bremer, H.; Brinkhoff, O.; Zabel, R.;
Sroka, W.; Klostermayer, H.; Brandenburg, D.; Okuda, T.; Zahn, H. Z.
Naturforsch. B 1963, 18, 1120−1121. (i) Zaykov, A. N.; Mayer, J. P.;
Gelfanov, V. M.; DiMarchi, R. D. ACS Chem. Biol. 2014, 9, 683−691.
(j) Kaur, Z. P.; Ochman, A. R.; Mayer, J. P.; Gelfanov, V. M.;
DiMarchi, R. D. ACS Chem. Biol. 2013, 8, 1822−1829. (k) Tang, J.-G.;
Wang, Z.-H.; Tregear, G. W.; Wade, J. D. Biochemistry 2003, 42,
2731−2739. (l) Barlos, K. K.; Gatos, D.; Vasileiou, Z.; Barlos, K. J.
Pept. Sci. 2010, 16, 200−211.
(14) (a) de la Torre, B. G.; Jakab, A.; Andreu, D. Int. J. Pept. Res.
Ther. 2007, 13, 265−270. (b) Garcia-Martin, F.; Quintanar-Audelo,
M.; Garcia-Ramos, Y.; Cruz, L. J.; Gravel, C.; Furic, R.; Cote, S.; Tulla-
Puche, J.; Albericio, F. J. Comb. Chem. 2006, 8, 213−220.
(15) Ghosh, A. K.; Fan, E. Tetrahedron Lett. 2000, 41, 165−168.
(16) (a) Oh, B.; Kim, K.; Park, J.; Yoon, J.; Han, D.; Kim, Y. Biochem.
Biophys. Res. Commun. 2004, 319, 486−492. (b) van der Mey, M.; de
Vroom, E. Bioorg. Med. Chem. Lett. 1994, 4, 345−348.
(4) (a) Liu, F.; Luo, E. Y.; Flora, D. B.; Mezo, A. R. Angew. Chem., Int.
Ed. 2014, 53, 3983−3987. (b) Liu, F.; Luo, E. Y.; Flora, D. B.; Mayer,
J. P. Org. Lett. 2013, 15, 960−963. (c) Barlos, K. K.; Gatos, D.;
Vasileiou, Z.; Barlos, K. J. Pept. Sci. 2010, 16, 200−211. (d) Hossain,
M. A.; Belgi, A.; Lin, F.; Zhang, S.; Shabanpoor, F.; Chan, L.; Belyea,
C.; Truong, H.-T.; Blair, A. R.; Andrikopoulos, S.; Tregear, G. W.;
Wade, J. D. Bioconjugate Chem. 2009, 20, 1390−1396. (e) Samuel, C.
S.; Lin, F.; Hossain, M. A.; Zhao, C.; Ferraro, T.; Bathgate, R. A. D.;
Tregear, G. W.; Wade, J. D. Biochemistry 2007, 46, 5374−5381.
(f) Bathgate, R. A. D.; Lin, F.; Hanson, N. F.; Otvos, L., Jr.; Guidolin,
A.; Giannakis, C.; Bastiras, S.; Layfield, S. L.; Ferraro, T.; Ma, S.; Zhao,
C.; Gundlach, A. L.; Samuel, C. S.; Tregear, G. W.; Wade, J. D.
Biochemistry 2006, 45, 1043−1053. (g) Akaji, K.; Fujino, K.; Tatsumi,
T.; Kiso, Y. J. Am. Chem. Soc. 1993, 115, 11384−11392. (h) Bullesbach,
E. E.; Schwabe, C. J. Biol. Chem. 1991, 266, 10754−10761. (i) Sieber,
P.; Kamber, B.; Hartmann, A.; Joehl, A.; Riniker, B.; Rittel, W. Helv.
Chim. Acta 1974, 57, 2617−2621. (j) Akhter, H. M.; Bathgate, R. A.
D.; Kong, C. K.; Shabanpoor, F.; Zhang, S.; Haugaard-Jonsson, L. M.;
D
dx.doi.org/10.1021/ol501252b | Org. Lett. XXXX, XXX, XXX−XXX