Preformed selenoesters enable rapid native chemical ligation at intractable sites

Article abstract of DOI:10.1002/anie.201105512

Going pro: The first facile Pro-Cys ligation using a preformed prolyl selenoester is reported (see scheme; P=peptide). In a comparative study of peptide selenoesters in native chemical ligation peptide α-selenoesters are shown to be superior acyl donors and result in rate enhancements of at least two orders of magnitude when compared to the well-established peptide α-thioesters. This method permits rapid chemical ligation even at previously intractable sites, such as Pro-Cys. Copyright

Full text of DOI:10.1002/anie.201105512

Communications
DOI: 10.1002/anie.201105512
Peptides
Preformed Selenoesters Enable Rapid Native Chemical Ligation at
Intractable Sites**
Thomas Durek* and Paul F. Alewood
The discovery of native chemical ligation (NCL) by Kent and
co-workers in 1994 was a critical step, enabling researchers to
chemically access medium-sized proteins through total or
semi-synthesis.[1] Virtually hundreds of proteins have now
been prepared by the NCL method ranging in size from small
polypeptides to medium-sized proteins of more than 200
amino acids.[2] The original conceptual framework of NCL is
based on the reaction between an unprotected peptide with a
C-terminal thioester and another unprotected peptide carry-
ing an N-terminal cysteine (Scheme 1b).[2b] The sulfhydryl
Several parameters have been shown to affect the rate of
an NCL reaction, including pH, the chemical structure, and
the reactivity of exogenous thiol catalysts and the thioester
leaving group, as well as the identity of the C-terminal amino
acid of the thioester segment.[3] Systematic analysis has shown
that specific amino acid thioesters—notably Ile, Val, Thr, and
Pro—react prohibitively slowly.[3b] As long NCL reaction
times (> 24 h) may lead to significant side reactions (notably
thioester hydrolysis/direct aminolysis, tris(2-carboxyethyl)-
phosphine HCl (TCEP) mediated desulfurization) these
residues are generally unsuitable for Xaa–Cys ligation, and
traditionally have been avoided. Peptidyl prolyl thioesters are
by far the least reactive of all amino acids in a NCL
reaction.[3b,4] This lack of reactivity of the prolyl thioesters
appears to be the result of an orbital interaction of the
carbonyl oxygen atom of the adjacent amino acid residue and
the carbonyl carbon atom of the thioester functionality. Such
an interaction effectively reduces its electrophilicity.[4] This
led us to consider strategies that would increase the low
inherent reactivity of prolyl thioesters and consequently allow
them to undergo facile NCL. We reasoned that peptidyl prolyl
selenoesters would be more effective as acyl donors than the
corresponding thioesters because of the better leaving group
characteristics of a selenolate versus a thiolate. While the use
of selenocysteine (Sec) in NCL is well documented,[5]
preformed peptidyl a-selenoesters have not been described
nor has their potential in NCL reactions been investigated.
To probe their feasibility for NCL we undertook the
synthesis of the peptidyl a-selenoesters 3 and 4. The a-
thioester peptides 1 and 2 were prepared by the standard
in situ neutralization protocol for Boc (tert-butoxycarbonyl)
solid-phase peptide synthesis (SPPS; Scheme 1a),[6] and after
cleavage from the solid support and purification, substitution
of the thioester moiety with a selenoester functionality was
achieved by a two-step procedure. First, the peptide thioesters
were converted into the corresponding selenoacids 1a and 2a
by treating the unprotected peptide (5 mm) with NaHSe
(300 mm) in aqueous buffer at pH 7 for 10 hours. As the
resulting selenoacids appeared to be susceptible to hydrolysis
under conditions commonly employed for HPLC purification
of peptides (TFA/acetonitrile/water, pH 1–2; TFA = tri-
fluoroacetic acid), they were converted without purification
in one pot into the corresponding a-selenoesters 3 and 4 by
alkylation with 2-iodoacetamide at pH 4 for 2 hours, and
purified by HPLC (see the Supporting Information). This
simple procedure allows preparation of peptide a-selenoest-
ers from readily available peptide a-thioesters in near
quantitative yield and appears to be compatible with most
side-chain functionalities commonly found in complex unpro-
Scheme 1. a) One-pot synthesis of peptidyl selenoesters from the
corresponding thioesters. b) Mechanism of NCL.
group of the N-terminal cysteine residue undergoes a
transesterification with the C-terminal thioester, which is
generally believed to be the rate-determining step of the
overall reaction. The resulting thioester-linked intermediate
spontaneously and rapidly rearranges through an intra-
molecular S!N acyl shift to yield a native peptide bond.
[*] Dr. T. Durek, Prof. P. F. Alewood
Institute for Molecular Bioscience
The University of Queensland
306 Carmody Rd, St Lucia, Brisbane (Australia)
E-mail: t.durek@uq.edu.au
[**] This work was supported by The University of Queensland.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201105512.
12042
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 12042 –12045

Table 1: Second-order rate constants for the NCL of peptide thioesters
and selenoesters.
tected peptides (see Figures 1 and 2 in the Supporting
Information).
Standard NCL conditions[a] Selenol NCL conditions[b]
We next set out to compare the reactivity of the peptides
1–4 in NCL reactions with peptide 5 under standard NCL
conditions (6m guanidine HCl, 0.2m sodium phosphate,
50 mm 4-mercaptophenylacetic acid (MPAA), 100 mm
TCEP, pH 6.8, each peptide at a concentration of 1 mm;
Scheme 1b).[3a,7] Aliquots were removed regularly and were
quenched by adding an equal volume of a solution of 20%
acetonitrile, 5% TFA, and 75% water. The aliquots were
subsequently analyzed by HPLC, which allowed us to monitor
both product formation (Figure 1) and various reaction
intermediates (see the Supporting Information). Second-
Peptide ester
k [mꢀ1 sꢀ1
]
Factor k [mꢀ1 sꢀ1
]
Factor
LYRAP-[COSR1] 0.00057ꢁ0.00066
LYRAF-[COSR1] 0.437ꢁ0.032
LYRAP-[COSeR2] 0.019ꢁ0.002
LYRAF-[COSeR2] 0.827ꢁ0.039
1
766
33
n.d.
n.d.
n.d.
n.d.
0.198ꢁ0.007 347
1450
7.7ꢁ0.1 13500
[a] 6m guanidine HCl, 0.2m sodium phosphate, 50 mm MPAA, 100 mm
TCEP, pH 6.8; [b] 6m guanidine HCl, 0.2m sodium phosphate, 28 mm
1
ꢀ
ꢀ
ꢀ
DPDS, 100 mm TCEP, pH 6.2. n.d.=not determined; R = CH2 CH2
2
ꢀ
ꢀ
ꢀ
ꢀ
CONH CH2 COOH; R = CH2 CONH2.
3 and 4 rapidly transesterified with the thiol catalyst MPAA to
form the corresponding thio(4-carboxymethyl)phenyl ester:
transesterification of the phenylalanyl selenoester was quan-
titative in less than 1 minute, whereas the prolyl selenoester
reacted quantitatively within 10 minutes. In contrast, the
thioalkyl esters 1 and 2 were very slow to form the activated
thioaryl ester intermediate, which did not accumulate signifi-
cantly over time and was rapidly consumed (see the Support-
ing Information). It seems plausible that the rate enhance-
ment observed for a-selenoesters versus a-thioesters is
primarily due to significantly faster formation of the activated
thioaryl ester. Thus, in the case of 3 and 4, the rate-
determining step is likely to be trans-thioesterification of
the (rapidly formed) thioaryl ester with the cysteine sulf-
hydryl, whereas for 1 and 2 the overall reaction rate is limited
by formation of the thioaryl ester. Hence, similar rate
enhancements should be achievable by using preformed
MPAA thioesters.[8] Our data indicates that peptide sele-
noesters are superior acyl donors for thiol nucleophiles in
NCL, a finding that is consistent with the exceptional leaving
group ability of selenolates. The observation that transesteri-
fication of selenoesters with the MPAA catalyst leads rapidly
and quantitatively to the formation of the corresponding
thio(4-carboxymethyl)phenyl ester and selenol, suggests a
highly exergonic process that is kinetically favored. Hence,
this energetically favorable (downhill) reaction leads in fact to
deactivation of the peptide a-selenoester and loss of acylation
power, which led us seek alternative ways to preserve its
original reactivity.
In view of these findings we next turned our attention to
the NCL catalyst. As mentioned above, the thiol catalyst plays
a critical role during NCL. It serves as a mild reducing agent
helping to keep cysteine sulfhydryls in the reduced state and
during trans-thioesterification activates the less reactive
thioalkylesters as well as reverses unproductive thioesters.[3a,c]
We anticipated that a selenol catalyst such as selenophenol
(generated in situ from diphenyldiselenide (DPDS) and
TCEP) could substitute effectively for MPAA in NCL. To
this end, ligations of 1–4 with CFRANK (5) were repeated in
6m guanidine HCl, 0.2m sodium phosphate, 28 mm DPDS,
100 mm TCEP, pH 6.2 with each peptide at a concentration of
1 mm (Figure 1b, Table 1). Under these reaction conditions
both 1 and 2 reacted sluggishly, thus demonstrating both the
poor reactivity of thioalkyl esters in NCL as well as the
Figure 1. Product formation (P, as percent complete) as a function of
&
*
*
time for NCL of 1 (&), 2 ( ), 3 ( ), and 4 ( ) with CFRANK (5).
Reaction conditions: a) 6m guanidine HCl, 0.2m sodium phosphate,
50 mm 4-mercaptophenylacetic acid (MPAA), 100 mm TCEP, pH 6.8 or
b) 6m guanidine HCl, 0.2m sodium phosphate, 28 mm DPDS, 100 mm
TCEP, pH 6.2. Peptide concentration: 1 mm. Data were fitted to a
double exponential equation.
order rate constants (Table 1) were derived from the inte-
grated rate equation by following the consumption of the
starting material (CFRANK, 5) as described recently.[4]
Our data (Figure 1a, Table 1) show that the LYRAP-
thioester 1 reacts at almost three orders of magnitude slower
than the LYRAF-thioester 2, results that are in agreement
with results previously obtained by the groups of Dawson and
Kent.[3b,4] Gratifyingly, both 3 and 4 reacted significantly
faster than the corresponding a-thioesters. Detailed analysis
(see the Supporting Information) however showed, that both
Angew. Chem. Int. Ed. 2011, 50, 12042 –12045
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12043

Communications
inability of the selenophenol catalyst to efficiently activate
them through transesterification.
reactions. In practical terms the requirement for an excess of 5
is acceptable, because excess peptide can be easily recovered
during the customary HPLC purification step following a
NCL.
In contrast, the a-selenoesters 3 and 4 rapidly and
quantitatively trans-selenoesterified with selenophenol (4:
< 30 s, 3: < 10 min) to form the corresponding peptidyl
selenophenyl esters. Inspection of the kinetic data revealed
Under similar reaction conditions 4 achieved near quan-
titative (> 95%) conversion within a few minutes, thus
bringing the NCL rates in line with more traditional
procedures for amide formation (e.g. anhydrides or active
esters as acyl donors in SPPS; see Figure 7 in the Supporting
Information). Despite the superior reactivity of peptidyl
selenoesters as acyl donors in NCL, the chemoselectivity is
maintained, that is, no reaction was observed between 4 and
peptide AFRANK, which lacks the essential b-mercapto
group of cysteine (see the Supporting Information). More-
over, the selenoester-mediated NCL proceeded without
epimerization at Phe or Pro, and the peptide configuration
and stereochemistry were fully retained (see the Supporting
Information).
[
]
*
pronounced biphasic product formation (Figure 1b): a fast
phase up to about 55–65% completion at t = 5 min [4, k =
(0.69 ꢁ 0.10) minꢀ1] followed by a slower phase [4, k = (4.8 ꢁ
0.8 ꢀ 10ꢀ3) minꢀ1]. This behavior coincided with the appear-
ance of a new peak in the HPLC chromatograms and was
identified as LYRAXC(LYRAX-[COS-])FRANK. Evidently,
this branched thioester originates from a side reaction, in
which the product cysteine sulfhydryl is additionally acylated
by the selenoester. Hence it seems plausible, that for the fast
phase the rate-determining step is transesterification between
the selenoester and CFRANK to form the desired ligation
product LYRAXCFRANK. As the reaction progresses, the
product will compete with CFRANK for the selenoester,
which is essentially consumed within 20 minutes. Thus, the
rate-determining step for the slow phase is likely the
dissolution of the branched thioester by transesterification
with selenophenol and re-activation of the acyl donor.
The underlying second-order rate constants (Table 1)
revealed that in the presence of selenophenol at pH 6.2 ꢁ
0.1, the a-selenoesters reacted about one order of magnitude
faster than when MPAA was used as a catalyst at pH 6.8 ꢁ 0.1
(Table 1). This directly supports our initial hypothesis that a-
selenoesters are superior acyl donors in NCL. In contrast,
selenophenol as a catalyst is less potent than MPAA in
reversing unproductive thioesters such as the branched
thioester intermediates. This is exemplified by comparing
the observed rate constants of the slow phase (Figure 1) for 4
under standard NCL conditions [k = (8.1 ꢁ 1.7 ꢀ 10ꢀ3) minꢀ1]
to those under selenol NCL conditions [k = (4.8 ꢁ 0.8 ꢀ
10ꢀ3) minꢀ1]. These findings can be explained on the basis
of the pioneering work of Hupe and Jencks on the mechanism
and structure–reactivity relationships of acyl-transfer reac-
tions.[9] Selenols are more acidic than structurally similar
thiols, for example, the pKa of the mercapto group of cysteine
is approximately 8.3, whereas that of the selenol of seleno-
cysteine has been determined to be around 5.4.[10] Conse-
quently, selenolate is a weaker nucleophile but better leaving
group than a structurally similar thiolate. Thus, in a simple
transesterification reaction between a selenoester and a
thiolate the equilibrium will favor the thioester and free
selenolate—well in agreement with our observations.
In its current form the selenoester-mediated NCL is
incompatible with unprotected nonligation site cysteines
largely because of the inefficiency of the selenol catalyst to
reverse the formation of unproductive thioesters. However,
we anticipate that selenols equipped with electron-donating
substituents leading to increased basicity/nucleophilicity of
the selenol functionality will be more efficient catalysts in
selenoester NCL than the selenophenol used herein; work in
this area is currently underway. Finally, our findings should
justify an examination of selenoacids and selenoesters in
other thioester/thioacid-based coupling techniques, such as
the direct aminolysis method.[11]
To minimize this side reaction we reasoned that a simple
solution would be to increase the ratio of 5 relative to the
selenoester given that the starting material acyl acceptor 5
will more effectively outcompete the product for the acyl
donor. This is indeed the case (Figure 2): when 3 (1.0 mm) is
reacted with increasing concentrations of 5 (1.0–5.1 mm) the
Figure 2. Product formation (P) as a function of time for NCL of 3
&
(1 mm) and increasing concentrations of CFRANK 5: 1 mm ( );
~
!
*
1.5 mm ( ); 3.4 mm ( ); 5.1 mm ( ). Data were fitted to a double
exponential equation.
plateau characterizing the end of the fast phase is shifted
considerably to higher product yields. Thus, by using pepti-
dylprolyl selenoesters such as 3 and an excess of 5 under the
conditions described above, near quantitative (> 90%) NCL
at proline is achieved in as little as 2 hours—a rate enhance-
ment of about 350 times when compared to standard NCL
[*] This biphasic product-formation behavior is also evident in NCL
reactions under standard conditions (Figure 1a), but much less
pronounced.
Received: August 4, 2011
Published online: October 13, 2011
12044 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 12042 –12045

Donk, Org. Lett. 2001, 3, 1331 – 1334; c) R. J. Hondal, B. L.
Nilsson, R. T. Raines, J. Am. Chem. Soc. 2001, 123, 5140 – 5141;
d) N. Metanis, E. Keinan, P. E. Dawson, Angew. Chem. 2010,
122, 7203 – 7207; Angew. Chem. Int. Ed. 2010, 49, 7049 – 7053.
[6] M. Schnçlzer, P. Alewood, A. Jones, D. Alewood, S. B. Kent, Int.
J. Pept. Prot. Res. 1992, 40, 180 – 193.
[7] T. Durek, V. Y. Torbeev, S. B. Kent, Proc. Natl. Acad. Sci. USA
2007, 104, 4846 – 4851.
[8] D. Bang, B. L. Pentelute, Z. P. Gates, S. B. Kent, Org. Lett. 2006,
8, 1049 – 1052.
Keywords: kinetics · native chemical ligation · peptides ·
selenium · synthetic methods
.
[1] a) P. E. Dawson, T. W. Muir, I. Clark-Lewis, S. B. Kent, Science
1994, 266, 776 – 779; b) P. E. Dawson, S. B. Kent, Annu. Rev.
Biochem. 2000, 69, 923 – 960.
[2] a) S. B. Kent, Curr. Opin. Biotechnol. 2004, 15, 607 – 614; b) S. B.
Kent, Chem. Soc. Rev. 2009, 38, 338 – 351.
[3] a) E. C. Johnson, S. B. Kent, J. Am. Chem. Soc. 2006, 128, 6640 –
6646; b) T. M. Hackeng, J. H. Griffin, P. E. Dawson, Proc. Natl.
Acad. Sci. USA 1999, 96, 10068 – 10073; c) P. E. Dawson, M. J.
Churchill, M. R. Ghadiri, S. B. Kent, J. Am. Chem. Soc. 1997,
119, 4325 – 4329.
[4] S. B. Pollock, S. B. Kent, Chem. Commun. 2011, 47, 2342 – 2344.
[5] a) R. Quaderer, A. Sewing, D. Hilvert, Helv. Chim. Acta 2001,
84, 1197 – 1206; b) M. D. Gieselman, L. L. Xie, W. A. van der
[9] D. J. Hupe, W. P. Jencks, J. Am. Chem. Soc. 1977, 99, 451 – 464.
[10] a) R. E. Huber, R. S. Criddle, Arch. Biochem. Biophys. 1967,
122, 164 – 173; b) M. Muttenthaler, P. F. Alewood, J. Pept. Sci.
2008, 14, 1223 – 1239.
[11] S. Aimoto, Curr. Org. Chem. 2001, 5, 45 – 87.
Angew. Chem. Int. Ed. 2011, 50, 12042 –12045
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12045