Native Chemical Ligation-Photodesulfurization in Flow

Article abstract of DOI:10.1021/jacs.8b03115

Native chemical ligation (NCL) combined with desulfurization chemistry has revolutionized the way in which large polypeptides and proteins are accessed by chemical synthesis. Herein, we outline the use of flow chemistry for the ligation-based assembly of polypeptides. We also describe the development of a novel photodesulfurization transformation that, when coupled with flow NCL, enables efficient access to native polypeptides on time scales up to 2 orders of magnitude faster than current batch NCL-desulfurization methods. The power of the new ligation-photodesulfurization flow platform is showcased through the rapid synthesis of the 36 residue clinically approved HIV entry inhibitor enfuvirtide and the peptide diagnostic agent somatorelin.

Full text of DOI:10.1021/jacs.8b03115

Communication
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Native Chemical Ligation−Photodesulfurization in Flow
Timothy S. Chisholm, Daniel Clayton, Luke J. Dowman, Jessica Sayers, and Richard J. Payne*
School of Chemistry, The University of Sydney, Sydney, NSW 2006, Australia
S
* Supporting Information
Scheme 1. Native Chemical Ligation−Desulfurization: (A)
ABSTRACT: Native chemical ligation (NCL) combined
with desulfurization chemistry has revolutionized the way
in which large polypeptides and proteins are accessed by
chemical synthesis. Herein, we outline the use of flow
chemistry for the ligation-based assembly of polypeptides.
We also describe the development of a novel photo-
desulfurization transformation that, when coupled with
flow NCL, enables efficient access to native polypeptides
on time scales up to 2 orders of magnitude faster than
current batch NCL−desulfurization methods. The power
of the new ligation−photodesulfurization flow platform is
showcased through the rapid synthesis of the 36 residue
clinically approved HIV entry inhibitor enfuvirtide and the
peptide diagnostic agent somatorelin.
Trans-thioesterification with a Thiol Additive; (B) Trans-
thioesterification between Two Reacting Peptide Fragments;
(C) Intramolecular S → N Acyl Shift of Thioester
Intermediate; (D) Desulfurization of Cys to Ala at the
Ligation Junction
eptides and proteins are ubiquitous molecules in living
Psystems and generally exhibit exquisite selectivity for their
targets, a property that has led to renewed interest in
polypeptides as therapeutics. These “biologics” have been
reported to have twice the approval rate of small molecule
therapeutics1 and currently make up 10% of approved drugs.2 As
a result of this renaissance in polypeptide-based therapeutics, it is
not surprising that attention has turned to the development of
methods to efficiently access these biomolecules. While solid
phase peptide synthesis (SPPS) represents the most efficient
avenue for generating peptides via chemical synthesis,3 there is a
significant limitation on the size of targets that can be produced
en bloc. This limitation of SPPS was largely addressed through the
development of native chemical ligation (NCL), a transformative
technology that enables convergent and chemoselective fusion of
unprotected peptide fragments.4 NCL is performed between a
peptide containing an N-terminal cysteine (Cys) residue and a
peptide functionalized as a C-terminal thioester, and usually
requires a suitable thiol additive to generate a reactive thioester
(SR2) that accelerates the rate-limiting trans-thioesterification
step (Scheme 1).5
An important advance to the seminal methodology was the
development of metal-based desulfurization6 and later a milder
radical-based protocol7 that facilitate the conversion of the least
abundant proteinogenic amino acid Cys to an alanine (Ala)
residue at the ligation junction (Scheme 1). The subsequent
development of thiolated amino acids (as Cys surrogates) has
served to expand the number of targets accessible via NCL
technology.8 A further innovation has been the development of
one-pot ligation−desulfurization chemistry using thiol additi-
ves,5b,9 e.g. trifluoroethanethiol (TFET).5b These additives
enhance the rate of the NCL reaction but, unlike traditionally
used aryl thiols, do not interfere with subsequent radical
desulfurization chemistry which can therefore be performed
without intermediary purification.
Given our long-standing interest in the development and
exploitation of ligation technologies for accessing peptides and
proteins, we envisaged a novel flow chemistry platform to
facilitate rapid and automated polypeptide assembly via NCL.
We hypothesized that a flow manifold may improve the efficiency
of the ligation−desulfurization transformations through auto-
mation and enhanced mixing,10 features that have been
capitalized on for SPPS but not for peptide ligation chemistry.11
Herein, we report the development of an efficient flow-based
NCL methodology. We also describe a photochemical trans-
formation that enables traceless desulfurization within minutes in
flow. We show that these two methods can be combined via “in-
line” ligation−desulfurization for the rapid and efficient assembly
of peptide targets. The power of this novel flow platform is
demonstrated through the synthesis of two clinically used
polypeptides.
Our investigations began with the development and
optimization of a model NCL reaction in flow (Scheme 2A).
We initially performed a flow ligation whereby a stream of H-
CSPGYS-NH2 1 (10 mM) in aqueous ligation buffer (6 M Gdn·
HCl, 0.1 M HEPES, pH 7.4−7.5) containing TCEP (50 mM)
and the alkyl thiol TFET5b (500 mM) was mixed through a T-
piece with a stream of peptide thioester Ac-LYRANX-
S(CH2)2CO2Et (2: X = A, 3: X = V, 12 mM) and TCEP (50
mM) in ligation buffer (pH 7.1−7.2) at equal flow rates using
Received: March 27, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.8b03115
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 2. (A) Flow NCL between Peptide 1 and Thioesters 2
and 3; (B) Flow NCL between Peptide 1 and Trifluoroethyl
Thioesters 4 and 5; (C) Schematic Diagram of the Optimized
NCL Flow System for the Reaction between Peptide 1 and
Peptide Thioester 5
Figure 1. (A) Reaction of peptide 1 and peptide trifluoroethyl thioester
5 with 2-MIM as an additive (2.5 M) to generate 7 on a 0.75 μmol scale
in batch (blue) and flow (red), on a 20 μmol scale in batch (green), and
reaction of 1 and 3 to generate 7 using TFET as an exogenous additive
on a 20 μmol scale in batch (black). (B) UPLC of the crude eluent from
the optimized flow ligation between 1 and 5. (gradient: 0 to 28%B over 5
min, λ = 214 nm). * Peak corresponds to the acyl hydrazide derivative of
5 following quenching with hydrazine; # peak corresponds to hydrolyzed
5.
28% completion after 1 h for desulfurization of 1, Figure S3,
Scheme S4). We therefore sought an alternative method to
facilitate the desulfurization of ligation products in flow.
Recently, Guo and co-workers reported the use of a ruthenium
photocatalyst, together with a phosphine and a hydrogen atom
donor, for the conversion of Cys to Ala.13 However, the kinetics
of the desulfurization process (5−16 h for reaction completion)
were deemed too slow for implementation of this method into
our flow platform. Based on early work by Hoffmann and co-
workers for the desulfurization of mercaptans14 and, more
recently, by Ge et al. for the desulfurization of deoxyglycosides,15
we instead considered the use of UV irradiation (in the absence
of an additional catalyst) to photochemically initiate the
desulfurization reaction in flow. To this end, a flow system was
assembled where a solution of peptide 1 (5 mM in ligation buffer
with 2.5 M 2-MIM and 50 mM TCEP, pH 7.3−7.4, to mimic the
NCL flow eluent) was mixed through a T-piece with 250 mM L-
glutathione (GSH) as a hydrogen atom source16 and 350 mM
TCEP in ligation buffer (pH 7.4−7.5) with or without VA-044.
The reagent stream was directed into a polyfluoroalkoxy (PFA)
coil inside a Rayonet RPR-100 photoreactor and irradiated at
multiple wavelengths (Scheme 3A, Schemes S5−S7, Figures S4,
S5). In the absence of VA-044, clean conversion of 1 to the
desulfurized product 8 was observed in only 15 s upon
petroleum ether as an immiscible system solvent (see Scheme S2
for flow setup). The resultant reaction stream entered a
polytetrafluoroethylene (PTFE) coil reactor at 37 °C for a
variety of residence times. Unfortunately, this led to slow and
incomplete ligation (<60% after 40 min) due to axial diffusion of
TFET into the system solvent. The use of other common thiol
additives, namely 4-mercaptophenylacetic acid (MPAA) and
MESNa, led to incomplete ligation (Figure S1). An alternative
NCL flow strategy was devised whereby preformed trifluoroethyl
thioesters Ac-LYRANX-SCH2CF3 (4: X = A, 5: X = V) were
ligated with 1 using 2-methylimidazole (2-MIM)12 as an additive
to enhance the rate of ligation (Scheme 2B). Using the same flow
setup above, these conditions led to complete ligation between
peptide 1 and thioester 4 in 2 min, and with thioester 5 in 11 min
(see Scheme 2B and 2C, Figure S2 for additional data). Having
optimized the flow method we next compared the reaction rates
to batch ligations between 1 and thioester 3 with TFET as an
additive,5b and between 1 and preactivated trifluoroethyl
thioester 5 (Figure 1A). The optimized flow procedure was
substantially faster than both of these batch reactions (4−50-
fold), the rate difference being more pronounced on a larger scale
(20 μmol) despite active mixing (Figure 1A). As expected, the
reaction end points of flow ligations did not change upon scale-
up (see Figure 1B for crude UPLC trace of the ligation between 1
and 5, after quenching with 3 vol% hydrazine) and led to
excellent isolated yields upon direct HPLC purification of the
flow eluent (18 mg, 67% for 6 and 18 mg, 65% for 7).
Scheme 3. (A) Novel Peptide Photodesulfurization; UPLC
Chromatograms for the Crude Photodesulfurization Reaction
of (B) 1 to 8 (0 to 15%B over 5 min, λ = 214 nm) and (C) 7 to
10 (0 to 28%B over 5 min, λ = 214 nm)
Having established an optimized protocol for NCL in flow, we
next investigated whether desulfurization could be carried out
under a similar manifold. Initial attempts at performing VA-044-
promoted radical desulfurization7a in flow proved unsuccessful
with inefficient and incomplete desulfurization observed (ca.
B
DOI: 10.1021/jacs.8b03115
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 4. (A) Setup for the Ligation−Photodesulfurization Procedure in Flow; Crude UPLC Trace (Gradient: 0 to 28%B over 5
min, λ = 280 nm) of the Ligation−Photodesulfurization Procedure Performed between 1 and (B) Alanine Thioester 4 to Generate
9 (Total Reaction Time 5 min) and (C) Valine Thioester 5 to Generate 10 (Total Reaction Time 18 min) (* Peak Corresponds to
the GSH Thioester of 4 and 5)
photoirradiation at λ = 254 nm (35 W, rt), which served as our
optimal conditions in the reaction screen (Scheme 3A and 3B).
We also submitted products 6 and 7 from the earlier flow NCL
reactions to these photodesulfurization conditions that cleanly
generated 9 and 10 in 60−90 s (Scheme 3C). The reaction was
also chemoselective in the presence of potentially reactive side
chains, e.g. methionine and side chain protected Cys residues
(Schemes S12−S14). Finally, we showed that peptides bearing
penicillamine16,17 and β-thiol-aspartate18 residues could be
photodesulfurized cleanly in <1 min to afford native valine and
aspartate residues, respectively (Schemes S15 and S16). It should
be noted that prolonged exposure of batch reactions to UV
irradiation led to product degradation. This highlights both the
efficiency and homogeneity of the photoirradiation under a flow
manifold.
Having optimized the ligation and photodesulfurization
reactions in flow, we next combined these into an “in-line”
ligation−photodesulfurization flow platform. A 20 μmol flow
experiment was performed whereby a solution of 1 (10 mM in
ligation buffer with 5 M 2-MIM and 50 mM TCEP) was mixed at
a T-piece with thioester 4 or 5 (20 mM in ligation buffer with 50
mM TCEP, Scheme 4A). After passing through a PTFE coil
reactor at 37 °C (4: 3 min, 5: 11 min), the product was mixed at a
second T-piece with a stream of 250 mM GSH and 350 mM
TCEP in ligation buffer and directed to a second PTFE reactor at
37 °C (4: 1 min, 5: 6 min) to thiolyze any product thioesters
generated by trans-thioesterification between the internal Cys of
the ligation product 6 or 7 and excess thioester 4 or 5 (Figure S6,
Schemes S8−S11). It should be noted that product thioester
species accrue here due to the absence of external thiol additives
that would normally be added to facilitate transthioesterification
to regenerate a reactive thioester under typical batch NCL
conditions. Finally, the reaction mixture was photoirradiated (λ =
254 nm) for 1 min to facilitate photodesulfurization (see Scheme
4B and 4C and Figures S7 and S8 for additional data). Following
HPLC purification, the desired products were isolated in
excellent yields (9: 17 mg, 65%, 10: 17 mg, 63%). The peptides
were also generated on rapid time scales; a total reaction time of 5
min for 9 and 18 min for 10 was required, the latter representing
one of the most challenging ligation junctions (Val thioester).
To make a valid rate comparison, we next performed the novel
ligation−photodesulfurization as a one-pot procedure in batch.
Notably, we observed over an order of magnitude increase in the
reaction time when using preactivated thioesters (4: 2.6 h, 5: 3.2
h, Figure S9, Scheme S20) in batch compared to flow.
Furthermore, the flow NCL-photodesulfurization process was
2 orders of magnitude faster compared to a traditional ligation−
desulfurization reaction in batch5b (with thioesters 2 or 3 and
TFET as an additive), while affording comparable isolated yields
(see Figure S10, Schemes S18 and S19).
Having demonstrated the efficiency and rate acceleration of
the ligation−photodesulfurization flow manifold for model
peptides, we next investigated more challenging polypeptide
targets. As a target for synthesis, we first investigated the clinically
approved 36-residue peptide HIV entry inhibitor enfuvirtide that
is produced commercially by the condensation of three protected
fragments in organic solvent.19 We envisaged a flow-based
assembly whereby peptide 11 (enfuvirtide 1−18) bearing a C-
terminal lysine, derivatized as a trifluoroethyl thioester, could be
reacted with peptide 12 (enfuvirtide 19−36) bearing an N-
terminal β-thiol-asparagine residue20 (Scheme 5A). Both
peptides were made using Fmoc-SPPS (SI). Thioester 11 (2.0
equiv, 10 mM final concentration) was then ligated with β-thiol-
asparagine peptide 12 (1.0 equiv, 5 mM final concentration) in
the presence of 2.5 M 2-MIM and 50 mM TCEP at 37 °C (pH
6.8). After 32 min the reaction mixture was mixed at a T-piece
with a solution containing 250 mM GSH and 800 mM TCEP in
ligation buffer (pH 7.4−7.5) and directed into a PTFE coil
reactor at 37 °C for thiolysis (1 min). The reaction stream was
then passed through a PFA coil with UV photoirradiation (λ =
254 nm, 35 W, 3 min) to effect desulfurization and afford the
native polypeptide (see Scheme 5B). The total reaction time was
36 min, and enfuvirtide 13 was isolated in 40% yield (9.7 mg)
from a single HPLC purification (see Scheme 5C−5E). This
result demonstrates the utility of the flow methodology in
complex target assembly, as well as the compatibility with
thiolated amino acids.
We next investigated a second target, the 44-residue peptide
somatorelin, used as a diagnostic agent for growth hormone
deficiency (Scheme 6A).21 Peptide 14 (somatorelin 1−18),
bearing a C-terminal thioester, and peptide 15 (somatorelin 19−
44), containing an N-terminal Cys residue, were prepared using
Fmoc-SPPS (see SI). Peptide 15 (115 mg, 30 μmol, 5 mM final
concentration) was then ligated with peptide 14 (149 mg, 60
μmol, 10 mM final concentration) in flow (11 min). Following
thiolysis (1 min) the stream was directed into a PFA coil and
C
DOI: 10.1021/jacs.8b03115
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 5. (A) Synthesis of Enfuvirtide via NCL−
Scheme 6. (A) Synthesis of Somatorelin via NCL−
Photodesulfurization in Flow; (B) UPLC Trace of the Crude
Reaction Mixture after in-Line Flow NCL−
Photodesulfurization in Flow; (B) ESI-MS of Synthetic
Somatorelin 16 (MS Data were Collected over the Entire UV
Peak in the UPLC-MS Chromatogram); (C) MALDI-TOF
Photodesulfurization (Gradient: 0 to 70%B over 5 min, λ =
214 nm; * Peak Corresponds to GSH Thioester of 11); (C)
Analytical UPLC Trace of Purified Enfuvirtide 13 (Gradient:
0 to 70%B over 5 min, λ = 214 nm); (D) ESI-MS of Synthetic
Enfuvirtide 13 (MS Data were Collected over the Entire UV
Peak in the UPLC-MS Chromatogram); (E) MALDI-TOF of
a
MS of Synthetic Somatorelin 16
a
Enfuvirtide 13
a
See Figure S12 and Schemes S24−S25 for additional details.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.8b03115.
■
S
Experimental, reaction, rate and NMR data, Figures S1−
S58 and Schemes S1−S25 (PDF)
a
AUTHOR INFORMATION
Corresponding Author
*richard.payne@sydney.edu.au
See Figure S11 and Schemes S22−S23 for additional details.
■
ORCID
photoirradiated (λ = 254 nm, 35 W, 3 min). After a total reaction
time of 15 min, HPLC purification provided somatorelin 16 in
excellent yield (57%, 103 mg, see Scheme 6B, 6C and Scheme
S25 for analytical data).
Richard J. Payne: 0000-0002-3618-9226
Notes
The authors declare no competing financial interest.
In summary, we describe here the first flow platform for
polypeptide assembly using NCL. We demonstrate an improved
rate of ligation using our flow protocol compared to traditional
batch chemistry, particularly at sterically hindered junctions. We
also describe a novel photodesulfurization transformation for the
rapid and efficient conversion of Cys to Ala in peptides. We show
that combining these two reactions into an in-line flow-based
ligation−photodesulfurization process enables access to poly-
peptide targets much faster than traditional batch chemistry and
can now be extended to larger scale production of polypeptides
under a continuous flow regime. The power of the flow ligation−
photodesulfurization platform was illustrated by the synthesis of
the HIV drug enfuvirtide and the diagnostic agent somatorelin.
Given the enabling nature of this technology, we anticipate its
application to the scalable synthesis of proteins and biologics as
well as extension to other ligation reactions, e.g. diselenide-
selenoester ligation chemistry22 in the future.
ACKNOWLEDGMENTS
■
This work was supported by the ARC Future Fellowship Scheme
(FT130100150), a Northcote Scholarship (J.S.), and John
Lamberton Research Scholarships (L.J.D. and J.S.). We would
like to thank Dr. Ian Luck from the University of Sydney for
support with NMR spectroscopy.
REFERENCES
■
(1) Hay, M.; Thomas, D. W.; Craighead, J. L.; Economides, C.;
Rosenthal, J. Nat. Biotechnol. 2014, 32, 40.
(2) (a) Fosgerau, K.; Hoffmann, T. Drug Discovery Today 2015, 20,
122. (b) Usmani, S. S.; Bedi, G.; Samuel, J. S.; Singh, S.; Kalra, S.; Kumar,
P.; Ahuja, A. A.; Sharma, M.; Gautam, A.; Raghava, G. P. PLoS One 2017,
12, e0181748.
(3) (a) Merrifield, R. B. J. Am. Chem. Soc. 1963, 85, 2149. (b) Kent, S.
B. Annu. Rev. Biochem. 1988, 57, 957.
(4) (a) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. Science
1994, 266, 776. (b) Kent, S. B. H. Chem. Soc. Rev. 2009, 38, 338.
D
DOI: 10.1021/jacs.8b03115
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(5) (a) Johnson, E. C. B.; Kent, S. B. H. J. Am. Chem. Soc. 2006, 128,
6640. (b) Thompson, R. E.; Liu, X.; Alonso-García, N.; Pereira, P. J. B.;
Jolliffe, K. A.; Payne, R. J. J. Am. Chem. Soc. 2014, 136, 8161.
(6) Yan, L. Z.; Dawson, P. E. J. Am. Chem. Soc. 2001, 123, 526.
(7) (a) Wan, Q.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2007, 46,
9248. (b) Jin, K.; Li, T.; Chow, H. Y.; Liu, H.; Li, X. Angew. Chem., Int.
Ed. 2017, 56, 14607.
(8) (a) Dawson, P. E. Isr. J. Chem. 2011, 51, 862. (b) Kulkarni, S. S.;
Sayers, J.; Premdjee, B.; Payne, R. J. Nat. Rev. Chem. 2018, 2, 122.
(c) Malins, L. R.; Payne, R. J. Aust. J. Chem. 2015, 68, 521. (d) Malins, L.
R.; Payne, R. J. Curr. Opin. Chem. Biol. 2014, 22, 70. (e) Malins, L. R.;
Payne, R. J. Top. Curr. Chem. 2014, 362, 27. (f) Bondalapati, S.; Jbara,
M.; Brik, A. Nat. Chem. 2016, 8, 407.
(9) Huang, Y. C.; Chen, C. C.; Gao, S.; Wang, Y. H.; Xiao, H.; Wang,
F.; Tian, C. L.; Li, Y. M. Chem. - Eur. J. 2016, 22, 7623.
(10) (a) Pastre, J. C.; Browne, D. L.; Ley, S. V. Chem. Soc. Rev. 2013, 42,
8849. (b) Plutschack, M. B.; Pieber, B.; Gilmore, K.; Seeberger, P. H.
Chem. Rev. 2017, 117, 11796.
(11) Mijalis, A. J.; Thomas, D. A.; Simon, M. D.; Adamo, A.; Beaumont,
R.; Jensen, K. F.; Pentelute, B. L. Nat. Chem. Biol. 2017, 13, 464.
(12) Sakamoto, K.; Tsuda, S.; Mochizuki, M.; Nohara, Y.; Nishio, H.;
Yoshiya, T. Chem. - Eur. J. 2016, 22, 17940.
(13) Gao, X.-F.; Du, J.-J.; Liu, Z.; Guo, J. Org. Lett. 2016, 18, 1166.
(14) Hoffmann, F. W.; Ess, R. J.; Simmons, T. C.; Hanzel, R. S. J. Am.
Chem. Soc. 1956, 78, 6414.
(15) Ge, J.-T.; Li, Y.-Y.; Tian, J.; Liao, R.-Z.; Dong, H. J. Org. Chem.
2017, 82, 7008.
(16) Haase, C.; Rohde, H.; Seitz, O. Angew. Chem., Int. Ed. 2008, 47,
6807.
(17) Chen, J.; Wan, Q.; Yuan, Y.; Zhu, J.; Danishefsky, S. J. Angew.
Chem., Int. Ed. 2008, 47, 8521.
(18) Thompson, R. E.; Chan, B.; Radom, L.; Jolliffe, K. A.; Payne, R. J.
Angew. Chem., Int. Ed. 2013, 52, 9723.
(19) Bray, B. L. Nat. Rev. Drug Discovery 2003, 2, 587.
(20) Sayers, J.; Thompson, R. E.; Perry, K. J.; Malins, L. R.; Payne, R. J.
Org. Lett. 2015, 17, 4902.
(21) Vlieghe, P.; Lisowski, V.; Martinez, J.; Khrestchatisky, M. Drug
Discovery Today 2010, 15, 40.
(22) (a) Mitchell, N. J.; Malins, L. R.; Liu, X.; Thompson, R. E.; Chan,
B.; Radom, L.; Payne, R. J. J. Am. Chem. Soc. 2015, 137, 14011.
(b) Mitchell, N. J.; Sayers, J.; Kulkarni, S. S.; Clayton, D.; Goldys, A. M.;
Ripoll-Rozada, J.; Pereira, P. J. B.; Chan, B.; Radom, L.; Payne, R. J.
Chem. 2017, 2, 703.
E
DOI: 10.1021/jacs.8b03115
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX