Extending the applicability of native chemical ligation

Article abstract of DOI:10.1021/ja960398s

A more general approach to native (amide-forming) chemical ligation of unprotected peptide segments is described that extends the technique beyond the previously reported X-Cys ligation site to now include X-Gly and Gly-X ligation sites. A peptide, [pepti

Full text of DOI:10.1021/ja960398s

J. Am. Chem. Soc. 1996, 118, 5891-5896
5891
Extending the Applicability of Native Chemical Ligation
Lynne E. Canne,* Steven J. Bark, and Stephen B. H. Kent
Contribution from The Scripps Research Institute, 10666 North Torrey Pines Road,
La Jolla, California 92037
X
ReceiVed February 7, 1996
Abstract: A more general approach to native (amide-forming) chemical ligation of unprotected peptide segments is
described that extends the technique beyond the previously reported X-Cys ligation site to now include X-Gly and
R
R
Gly-X ligation sites. A peptide, [peptide1] COSR, is reacted with a second peptide, HSCH2CH2(O)-N [peptide2],
under conditions promoting thioester exchange. The intermediate thioester-linked product rearranges to form a ligation
product linked by an N-substituted amide bond. If desired, the -oxyalkyl substitution on the amide bond can be
removed by facile treatment with Zn in acidic medium to give a native peptide bond at the ligation site. The techniques
described have been employed to ligate small model peptide segments to yield peptides with native or modified
backbones, proving the feasibility of this approach.
Introduction
regenerating the Cys side chain thiol. This original version of
14
native ligation uses chemistry first described by Wieland for
The application of chemical ligation to peptide synthesis is
13
reacting amino acids. As originally described, native ligation
is restricted to joining peptide segments at an X-Cys bond.
Here we report important advances which extend this
procedure, alleviating the necessity for an N-terminal Cys
residue, and thus expanding the range of sites amenable to native
chemical ligation to X-Gly and Gly-X as well as X-Cys. The
method allows for the synthesis of peptides and proteins with
either native or modified backbone structures. Scheme 1
outlines the procedure. The peptide-R-carboxythioester (1)
becoming increasingly popular.1
-8
Typically, this method
involves the chemoselective reaction of unprotected peptide
segments to give a product with an unnatural backbone structure
at the site of ligation.1
,4,5
This revolutionary concept makes
possible the synthesis of peptides of greater size than attainable
by standard methods of peptide synthesis8 as well as the
,9
2
,4,6,8,9
synthesis of peptides of unusual structure and topology.
The chemical ligation method has allowed us to harness the
power of stepwise solid phase peptide synthesis1 to routinely
make unprotected peptides of up to 60 amino acid residues in
good yield and purity,12 and to use this for the purpose of routine
total chemical synthesis of proteins.
Recently, a chemical ligation technique was reported for the
preparation of proteins with native backbone structure.13 In this
technique, an unprotected synthetic peptide bearing a C-terminal
R-thioester is reacted in a chemoselective manner with an
unprotected peptide containing an N-terminal Cys residue. Thiol
exchange yields an initial thioester-linked intermediate which
spontaneously rearranges to give a native amide bond at the
ligation site joining the two peptide segments, in the process
0,11
R
reacts, via thiol exchange, with either an N (ethanethiol) peptide
R
(
2a) or an N (oxyethanethiol) peptide (2b) to produce the
ligation product 3. This thioester-linked intermediate rearranges
through a favorable geometric arrangement involving a five-
or six-membered ring to give the amide-linked product 4,
containing a secondary amide in 4a or the analogous N-oxyalkyl
R
compound in 4b. The resulting N (substituted)amides also have
15
R
potentially favorable solubility properties; the N (oxyalkyl)-
amide ligation product has the further advantage of being stable
to HF cleavage conditions, yet is simply removed under mild
conditions. In the case of 4b, zinc dust can be added directly
to the reversed phase HPLC-purified peptide in the acidic eluant
to reduce the N-O bond of the O-alkylhydroxamate and give
the native backbone structure of ligation product 5b. Model
studies were performed to explore the scope and limitations of
this approach.
*
To whom correspondence should be addressed. Present address:
Gryphon Sciences, 250 E. Grand Ave., Suite 90, South San Francisco, CA
4080. Phone: (415) 952-8518. FAX: (415) 952-3055. E-mail: lynne@
gryphonsci.com.
9
X
Abstract published in AdVance ACS Abstracts, June 1, 1996.
1) Schn o¨ lzer, M.; Kent, S. B. H. Science 1992, 256, 221-225.
2) Dawson, P. E.; Kent, S. B. H. J. Am. Chem. Soc. 1993, 115, 7263-
(
(
Results
7
1
266.
All peptide segments were synthesized in stepwise fashion
by established solid phase methods using in situ neutralization/
(
3) Baca, M.; Kent, S. B. H. Proc. Natl. Acad. Sci. U.S.A. 1993, 90,
1638-11642.
(
(
4) Rose, K. J. Am. Chem. Soc. 1994, 116, 30-34.
2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluo-
5) Liu, C. F.; Tam, J. P. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 6584-
rophosphate (HBTU) activation protocols for tert-butoxycar-
bonyl (Boc) chemistry, purified by preparative reversed phase
6
588.
(
6) Muir, T. W.; Williams, M. J.; Ginsberg, M. H.; Kent, S. B. H.
Biochemistry 1994, 33, 7701-7708.
7) Williams, M. J.; Muir, T. W.; Ginsberg, M. H.; Kent, S. B. H. J.
Am. Chem. Soc. 1994, 116, 10797-10798.
8) Canne, L. E.; Ferr e´ -D’Amar e´ , A. R.; Burley, S. K.; Kent, S. B. H. J.
Am. Chem. Soc. 1995, 117, 2998-3007.
9) Baca, M.; Muir, T. W.; Schn o¨ lzer, M.; Kent, S. B. H. J. Am. Chem.
HPLC, and characterized by electrospray mass spectrometry
(
12
(
ESMS). Peptide-R-thioesters (1) were generated from the
corresponding peptide-R-thioacids, which in turn, were synthe-
sized on a thioester resin as previously described. The peptide-
(
16
(
R-thioacids were converted to either the corresponding 3-car
Soc. 1995, 117, 1881-1887.
(10) Merrifield, R. B. J. Am. Chem. Soc. 1963, 85, 2149-2154.
(11) Kent, S. B. H. Annu. ReV. Biochem. 1988, 57, 957-989.
(12) Schn o¨ lzer, M.; Alewood, P.; Jones, A.; Alewood, D.; Kent, S. B.
(14) Wieland, T.; Bokelmann, E.; Bauer, L.; Lang, H. U.; Lau, H. Liebigs
Ann. Chem. 1953, 583, 129-149.
(15) Quibell, M.; Packman, L. C.; Johnson, T. J. Am. Chem. Soc. 1995,
117, 11656-11668.
H. Int. J. Peptide Protein Res. 1992, 40, 180-193.
13) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science
994, 266, 776-779.
(
(16) Canne, L. E.; Walker, S. M.; Kent, S. B. H. Tetrahedron Lett. 1995,
36, 1217-1220.
1
S0002-7863(96)00398-8 CCC: $12.00 © 1996 American Chemical Society

5
892 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Canne et al.
R
Scheme 1. Generalized Native Chemical Ligation of
Scheme 2. Synthesis of N (substituted) Peptide Segments
Unprotected Peptide Segments
boxy-4-nitrophenyl thioesters in 6 M guanidine‚HCl, 0.1 M
sodium acetate, pH 5.0-6.5, by reaction with 1.5 equiv of 5,5′-
dithiobis(2-nitrobenzoic acid)17 or the corresponding benzyl ester
in 6 M guanidine‚HCl, 0.1 M sodium acetate, pH 4.0, using 10
equiv of benzyl bromide.13 Both these thioesters provide
satisfactory leaving groups for the ligation reactions, with the
Scheme 3. Synthesis of Compound 8b
3
-carboxy-4-nitrophenyl thioesters demonstrating a somewhat
1
3
faster reaction rate than the corresponding benzyl thioesters.
R
R
The syntheses of N (ethanethiol) (2a) and N (oxyethanethiol)
2b) peptides are outlined in Scheme 2. The appropriate
(
R-bromocarboxylic acid, activated as the symmetric anhydride
0.5 equiv of 1,3-diisopropylcarbodiimide (DIC) in dichlo-
(
18
romethane), was coupled to the deprotected N-terminal amino
acid of peptide-resin 6 to give bromoacyl-peptide-resin 7. The
bromide was then displaced, with inversion of stereochemistry,
by the amine function of either structure 8a or 8b in DMSO to
give peptide-resin 9. Deprotection and cleavage from the resin
in anhydrous HF gave 2b directly or 2a still in the form of the
disulfide which was reduced to the free thiol.
ranging from 4 to 8 mg/mL of each peptide in either 8 M urea,
0
.1 M Na2HPO4, pH 7.0, or 6 M guanidine‚HCl, 0.1 M Na2-
HPO4, pH 7.5. Immediately after solvation of the peptide
segments, 2-5% benzyl mercaptan (model ligation no. 1) or
thiophenol (model ligation nos. 2-5), by volume of ligation
buffer, was added to keep thiol functions in the reduced form.
This had the added consequence of exchanging a significant
amount of the original peptide-R-thioester (1) to a peptide-R-
benzyl thioester or peptide-R-phenyl thioester, both of which
are still capable of reacting in the desired fashion with peptide
The aminoethanethiol derivative 8a was synthesized in one
step from the reaction of 2-aminoethanethiol and 5,5′-dithiobis-
(2-nitrobenzoic acid) in 80% acetonitrile in water. The synthesis
of the (aminooxy)ethanethiol derivative 8b was more involved
and is outlined in Scheme 3. Bromide 11 was produced from
the reaction of N-hydroxyphthalimide (10) with a large excess
of 1,2-dibromoethane.19 Bromide 11 was then converted to the
protected (aminooxy)ethanethiol derivative 12 with 4-methyl-
benzyl mercaptan in the presence of the base 1,8-diazabicyclo-
2
. The ligation reactions were followed by analytical reversed
phase HPLC and the products identified by ESMS. For clarity,
in the following text the amino acid residues involved in the
ligation are underlined.
[
5.4.0]undec-7-ene (DBU). The phthalimide group of 12 was
20
removed in a two-step process involving reduction with NaBH4
followed by treatment with acetic acid to give the desired
Model Ligation No. 1. The peptide-R-(3-carboxy-4-nitro)-
R
phenyl thioester 1 had the sequence LYRAG- COSC6H3(3-
(aminooxy)ethanethiol derivative 8b.
CO2H-4-NO2) (observed mass 760 Da, calcd 760 Da). The
Model Ligations. The results of the model ligations are
summarized in Table 1. All ligations were run at concentrations
R
N (ethanethiol) peptide 2a, protected as the 5-thio-2-nitroben-
zoic acid disulfide, had the sequence [(3-CO2H-4-NO2)-C6H3-
R
(17) Muir, T. W.; Dawson, P. E.; Kent, S. B. H. Tetrahedron Lett.,
S-SCH2CH2]-GAGPAGD- CONH2 (observed mass 800 Da,
submitted for publication.
calcd 800 Da) which was reduced to [HSCH2CH2]-GAGPAGD-
(
(
(
18) Robey, F. A.; Fields, R. L. Anal. Biochem. 1989, 177, 373-377.
19) Bauer, L.; Suresh, K. S. J. Org. Chem. 1963, 28, 1604-1608.
20) Osby, J. O.; Martin, M. G.; Ganem, B. Tetrahedron Lett. 1984, 25,
R
CONH2 (observed mass 603 Da, calcd 603 Da) in the ligation
mixture. After 1 h, the product of the ligation, 4a, [LYRAG-
R
R
2
093-2096.
G((N -CH2CH2SH)AGPAGD- CONH2 (observed mass 1163 (

Extending the Applicability of NatiVe Chemical Ligation
Table 1. Summary of Model Ligations
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5893
c
approximate yield (%)
temp reaction
model ligation
peptide 1a
peptide 2b
pH (°C) time (h) unrearranged product (3) rearranged product (4)
R
1
2
3
LYRAG-SNB N (etsh)-GAGPAGD
7.0
25
25
37
37
37
37
37
37
4
16
11.5
10
17.5
6.5
19
not detected
not detected
90
75
(35)]
64
(52)]
69
not detected
not detected
R
LYRAG-SNB N (oetsh)-GRNTATIMMQRGNFR 7.5
R
LYRAF-Sbzl N (oetsh)-GRNTATIMMQRGNFR [7.5
30
0
39
20
58
52
d
d
d
4
.5
[7.5
.5
7.5
.5
R
4
5
LYRAG-Sbzl N (oetsh)-AARHTVHQRHLHG
4
R
LYRAF-Sbzl N (oetsh)-AARHTVHQRHLHG
4
22
a
b
R
Thioester peptides where SNB ) 3 carboxy-4-nitrophenyl thioester and Sbzl ) benzyl thioester. N (substituted) peptides where etsh )
R
R
c
N (ethanethiol) and oetsh ) N (oxyethanethiol). Based on peptide 2, estimated from analytical reversed phase HPLC (peak areas) and ESMS.
d
Adjusted to pH 4.5 after the indicated time at pH 7.5.
R
R
G(N -OCH2CH2SH)RNTATIMMQRGNFR- CONH2 (4b; ob-
served mass 2388 ( 1 Da, calcd 2389 Da) had formed to a
R
significant degree, coeluting with unreacted N (oxyethanethiol)
peptide 2b (peak e). After further reaction overnight at room
temperature, the ligation product was purified by HPLC, and
zinc dust was added directly to the collected peptide in HPLC
eluant and stirred overnight at room temperature. Under these
conditions, the zinc effectively reduced the N-O bond. Reduc-
tions of this type are possible through a variety of reagents.21
The resulting peptide (peak f) gave a mass consistent with
reduction of the N-O bond to produce LYRAGGRNTATIM-
R
MQRGNFR- CONH2 (5b; observed mass 2313 ( 1 Da, calcd
2
313 Da). This confirmed the rearrangement of 3b to 4b as
shown in Scheme 1, since cleavage of the N-O bond in
unrearranged 3b would have resulted in the formation of two
distinct peptides of significantly lower masses.
Model Ligation No. 3. Figure 2 shows the course of model
ligation no. 3. The peptide-R-thioester segment 1 (peak d)
consisted of a peptide with a C-terminal phenylalanine thioester
R
R
[LYRAF- COSCH2C6H5 (observed mass 775 Da, calcd 775
2 2
Figure 1. Model ligation no. 2 to form amide -Gly(N -OCH CH -
Da)], thus providing a more sterically hindered model than the
previous cases (model ligations nos. 1 and 2) which involved
C-terminal glycine thioesters. The N (oxyethanethiol) peptide
2b (peak a) consisted of the sequence [HSCH2CH2O]-GRN-
TATIMMQRGNFR- CONH2 (observed mass 1827 ( 1 Da,
calcd 1828 Da). The presence of the more sterically hindered
thioester slowed the reaction relative to the unhindered models.
However, heating at 37 °C was found to accelerate the rate of
initial ligation. Figure 2A shows the ligation reaction after 11.5
h at 37 °C. Surprisingly, the rate of rearrrangement of 3b to
SH)Gly- compound 4b. Carried out in 6 M guanidine‚HCl, 0.1 M
Na
2
HPO
4
, pH 7.5. (A) Analytical HPLC (10-50% B over 30 min) at
R
R
t ) 0; peak a, thioester peptide 1, LYRAG- COSC
peak b, N (oxyethanethiol) peptide 2b, [HSCH
MQRGNFR- CONH
6
H
CH
3
(3-CO
O-GRNTATIM-
. (B) Analytical HPLC (10-50% B over 30 min)
2
H-4-NO
2
;
R
2
2
R
R
2
R
at t ) 1 h; peak c, non-peptide impurity; peak d, LYRAG- COSC
resulting from transthioesterification of 1 (peak a) with thiophenol; peak
e, intermediate ligation product 4b, LYRAGG(N -OCH
6
H
5
,
R
2
CH
2
SHR-
R
NTATIMMQRGNFR- CONH
2
and a minor amount of unreacted
R
[
HSCH
2
CH
2
O-GRNTATIMMQRGNFR- CONH
2
(2b), as determined
by electrospray ionization MS. (C) Analytical HPLC (15-40% B over
0 min) of peak e after HPLC purification and treatment with Zn in
4
b was slowed enough to observe the unrearranged product 3b,
3
R
R
LYRAF-[ COSCH2CH2O]-GRNTATIMMQRGNFR- C-
acidic HPLC solvent overnight; peak f, final ligation product 5b,
R
LYRAGGRNTATIMMQRGNFR- CONH
purity.
2
; peak g, non-peptide im-
ONH2 (peak b; observed mass 2478 ( 1 Da, calcd 2478 Da),
eluting slightly before the rearranged product 4b, LYRAFG(N -
OCH2CH2SH)RNTATIMMQRGNFR- CONH2 (peak c; ob-
served mass 2478 ( 1 Da, calcd 2478 Da). Having identical
masses, the unrearranged and rearranged products 3b and 4b
were identified by their zinc reduction products. Intermediate
R
R
1
Da, calcd 1163 Da)] had already formed to a significant
degree. After purification of the ligation product by reversed
phase HPLC, it was treated with 5,5′-dithiobis(2-nitrobenzoic
acid) for 1 h in 6 M urea, 0.1 M phosphate, pH 6.0, to yield the
expected S-nitrobenzoic acid disulfide product LYRAGG[N -
CH2CH2S-SC6H3(3-CO2H-4-NO2)]AGPAGD- CONH2 (ob-
served mass 1360 ( 1 Da, calcd 1360 Da). This confirmed
the rearrangement of 3a to form 4a, since 3a would have been
unreactive toward 5,5′-dithiobis(2-nitrobenzoic acid).
Model Ligation No. 2. Figure 1 shows the course of model
ligation no. 2. The peptide-R-thioester segment 1 (peak a) had
the sequence LYRAG- COSC6H3(3-CO2H,4-NO2) (observed
mass 760 Da, calcd 760 Da). The N (oxyethanethiol) peptide
b (peak b) consisted of the sequence [HSCH2CH2O]-GRN-
TATIMMQRGNFR- CONH2 (observed mass 1827 ( 1 Da,
calcd 1828 Da). After 1 h of reaction at room temperature in
the presence of added thiophenol, the ligation product LYRAG-
3
b, peak b, gave two peptides upon zinc reduction with masses
R
of 669 and 1753 Da corresponding to the peptide sequences
R
R
LYRAF- COOH (thioester hydrolysis) and GRNTATIMM-
R
QRGNFR- CONH2, respectively. It was subsequently deter-
mined that lowering the pH to 4.5 after initial ligation by diluting
the crude reaction mixture (to 5 times the volume) with 6 M
guanidine‚HCl, 0.1 M sodium acetate, pH 4.0, accelerated the
rate of rearrangement. Figure 2B shows that rearrangement of
the initial ligation product 3b to 4b (peak c) was complete after
R
R
1
0 h at pH 4.5 at 37 °C. HPLC purification of this peak and
subsequent zinc reduction gave a peptide (peak f) of the expected
2
R
(
21) Keck, G. E.; McHardy, S. F.; Wager, T. T. Tetrahedron Lett. 1995,
36, 7419-7922.

5
894 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Canne et al.
Da) was ligated to [HSCH2CH2O]-AARHTVHQRHLHG-
R
COOH 2b (observed mass 1595 ( 1 Da, calcd 1596 Da) at
pH 7.5. Though the initial (i.e., unrearranged) ligation product
3
R
b was observed, LYRAF- [COSCH2CH2O]-AARHTVHQRHL-
R
HG- COOH (observed mass 2246 ( 1 Da, calcd 2247 Da),
there was no evidence of rearrangement over time, even at lower
pH. The presence of side chains on both sides of the ligation
site apparently provided too much steric hindrance for the
rearrangement to occur via a six-membered ring intermediate
under the conditions used.
Discussion
The observations reported here serve to emphasize the
extraordinary facility of our original native ligation chemistry,13
which recently was independently repeated in essentially
2
2
identical form in model ligation studies. We have used this
original native peptide bond-forming ligation reaction in the
chemical synthesis of a number of proteins with full biological
1
3
activity, including the chemokine IL-8, the enzymes HIV-1
protease and barnase, the serine proteinase inhibitors turkey
ovomucoid third domain and eglin C, and a b/HLH transcription
factor. This ligation reaction was based on principles enunciated
R
23
Figure 2. Model ligation no. 3 to form a -Phe(N -OCH
2
CH
2
SH)Gly-
by Max Brenner and made use of chemistry first described
1
4
linked product 4b. Ligation and rearrangement carried out at 37 °C.
A) Analytical HPLC (20-40% B over 30 min) after 11.5 h in 6 M
, pH 7.5; peak a, N (oxyethanethiol)
O]-GRNTATIMMQRGNFR- CONH
unrearranged intermediate ligation product 3b, LYRAF-[ COSCH
by Wieland.
(
Compared with the original native chemical ligation at
R
guanidine‚HCl, 0.1 M Na
peptide 2b, [HSCH
2
HPO
4
13
X-Cys sites, the most notable feature of the ligation chemistry
R
2
CH
2
2
; peak b,
CH O]-
reported here is the slower rearrangement of the initial thioester
ligation product. Clearly, intramolecular attack by the R-NH-
(oxyalkyl) group via a six-membered ring intermediate in the
R
2
2
R
GRNTATIMMQRGNFR- CONH
2
; peak c, rearranged intermediate
CH SH)RNTATIMMQRGNFR-
R
ligation product 4b, LYRAFG(N -OCH
2
2
2
4
R
R
current instance is considerably less favored than the facile
CONH
Analytical HPLC (20-40% B over 30 min) after 10 h in 6 M
guanidine‚HCl, 0.1 M NaCH CO , pH 4.5. (C) Analytical HPLC (0-
7% B over 30 min) of HPLC-purified peak c after treatment with Zn
2 2 6 5
; peak d, thioester peptide 1, LYRAF- COSCH C H . (B)
five-membered ring mediated attack of the unsubstituted R-NH2
1
3
3
2
group in the X-Cys ligation.
Except for the Gly-Gly
6
ligation, in the present study we were able to isolate the
thioester-linked initial ligation product (Table 1), which rear-
in acidic HPLC solvent overnight; peak f, final ligation product 5b,
R
LYRAFGRNTATIMMQRGNFR- CONH
2
.
R
ranged only slowly. Remarkably, for C -substituted amino acids
on both sides of the ligation site, we were unable to observe
subsequent rearrangement to an amide bond: the thioester-linked
intermediate was indefinitely stable under the conditions of
reaction. These slow rearrangements are similar to those
observed in an (N-substituted)amide-forming ligation chemistry
R
mass, LYRAFGRNTATIMMQRGNFR- CONH2 (5b; observed
mass 2403 ( 1 Da, calcd 2404 Da).
R
Model Ligation No. 4. LYRAG- COSCH2C6H5 1 (observed
mass 685 Da, calcd 685 Da) was ligated to [HSCH2CH2O]-
AARHTVHQRHLHG- COOH 2b (observed mass 1595 ( 1
R
5
previously described.
Da, calcd 1596 Da) at pH 7.5. This model provided steric
hindrance, in the form of an Ala residue, on the opposite side
of the ligation site from model ligation no. 3. The rate of
reaction was similar to that of the previous model, the rate being
significantly slower than those of the unhindered models (model
ligation nos. 1 and 2), but enhanced with heating at 37 °C. Both
The use in the current study of a temporary auxiliary
functional group in amide-forming ligation is reminiscent of
the “thiol capture” strategy proposed by Kemp. Both native
25
chemical ligation and the thiol capture method have as their
stated goal the use of unprotected peptides in a segment
condensation strategy to achieve the synthesis of long polypep-
tide chains. The key aspect of the native chemical ligation
approach is reaction under conditions that promote exchange
of the thiol moiety of the initial thioester-linked intermediate
products, to give regioselective ligation at the N-terminal Cys
R
unrearranged (3b, LYRAG- [COSCH2CH2O]-AARHTVHQRHL-
R
R
HG- COOH) and rearranged (4b, LYRAGA(N -OCH2C-
R
H2SH)ARHTVHQRHLHG- COOH) ligation products were
formed, observed mass (both unrearranged and rearranged) 2156
(
1 Da, calcd 2157 Da. However, unlike model ligation no.
, the rate of rearrangement was not enhanced by lowering the
13
even in the presence of other Cys residues in both segments.
By contrast, in the thiol capture strategy and related methods,
3
26
pH to 4.5 after initial ligation, though rearrangement was nearly
complete after 2 days at 37 °C regardless of pH as shown by
successful reduction to the amide. It should also be noted that
the presence of a significant number of histidines in the final
regiospecific protection of Cys side chain functionalities prior
27
to the ligation reaction is a necessity, thus frustrating the
original intent.
R
ligation product 5b (LYRAGAARHTVHQRHLHG- COOH;
(22) Tam, J. P.; Lu, Y. A.; Liu, C. F.; Shao, J. Proc. Natl. Acad. Sci.
U.S.A. 1995, 92, 12485-12489.
observed mass 2080 ( 1 Da, calcd 2080 Da) resulted in binding
of the peptide to the zinc. EDTA had to be added to the HPLC
buffer/Zn mixture to free the peptide from the zinc after
reduction of the N-O bond.
(23) Brenner, M. In Peptides. Proceedings of the Eighth European
Peptide Symposium; Beyerman, H. C., Ed., North-Holland: Amsterdam,
1
967; pp 1-7.
(24) Mandolini, L. J. Am. Chem. Soc. 1978, 100, 550-554.
(25) Fotouhi, N.; Galackatos, N. G.; Kemp, D. S. J. Org. Chem. 1989,
Model Ligation No. 5. A final model ligation was performed
5
4, 2806-2817.
26) Liu, C. F.; Rao, C.; Tam, J. P. Tetrahedron Lett. 1996, 37, 933-
936.
which provided steric bulk on both sides of the ligation site.
(
R
LYRAF- COSCH2C6H5 1 (observed mass 775 Da, calcd 775

Extending the Applicability of NatiVe Chemical Ligation
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5895
Similarly, the use in the current study of an activated
molecular species, using the program MacSpec (Sciex). Calculated
masses were based on average isotope composition and were derived
using the program MacProMass (Terry Lee and Sunil Vemuri, Beckman
Research Institute, Duarte, CA). All other mass spectrometry was
performed at The Scripps Research Institute Mass Spectrometry Facility.
H NMR spectra were recorded on a Bruker 250-MHz spectrophotom-
eter, and chemical shifts are reported in parts per million downfield
4
from Me Si. Microanalyses were performed at The Scripps Research
Institute X-ray Crystallographic Facility and agreed with calculated
values (0.4%. Boc-L-amino acids and HBTU were purchased from
Novabiochem (La Jolla, CA). [[4-(Hydroxymethyl)phenyl]acetamido]-
methyl (PAM) resins and diisopropylethylamine (DIEA) were obtained
from Applied Biosystems (Foster City, CA), and methylbenzhydry-
lamine (MBHA) resin was obtained from Peninsula Laboratories, Inc.
R-thiocarboxyl function (i.e., an R-COSR) is reminiscent of
28-30
chemistry previously used in peptide segment condensation.
However, these syntheses were based on conventional non-
chemoselective attack by the R-amine nucleophile of the second
1
+
R
segment on a Ag -activated peptide- COSR, and thus neces-
sitated regiospecific (re)protection of all other R- and ꢀ-amine
functional groups in both segments.2
8-30
By contrast, thioester-mediated amide-forming ligation chem-
istry is compatible with the use of completely unprotected
peptide segments with the full range of side chain functionalities
found in nature, including thiols.9
,13
For this reason, native
chemical ligation is simple and practical, and is a general
approach to the total chemical synthesis of proteins provided
they contain appropriate ligation sites.
(San Carlos, CA). Synthesis grade dimethylformamide (DMF) was
obtained from Baker, and AR grade CH
2
Cl and HPLC grade CH CN
2
3
were obtained from Fisher. Trifluoroacetic acid (TFA) was obtained
from Halocarbon (New Jersey). HF was purchased from Matheson
Gas. 4-Methylbenzyl mercaptan was obtained from Lancaster. All
other reagents were AR grade or better and were obtained from Aldrich
Chemical or from Fisher.
The work reported here increases the utility of the native
1
3
chemical ligation method reported by Dawson et al. by
extending the number of dipeptide sequences that can be used
as ligation sites. In addition to the X-Cys ligation site of that
study, we can now make use of X-Gly and Gly-X ligation
Peptide Segment Synthesis. Chain Assembly. Peptides were
synthesized in stepwise fashion by established machine-assisted or
manual solid phase methods using in situ neutralization/HBTU activa-
tion protocols for Boc chemistry.12 Side chain protection was as
follows: Boc-Arg(p-toluenesulfonyl)-OH, Boc-Asn(xanthyl)-OH, Boc-
Asp(O-cyclohexyl)-OH, Boc-His(dinitrophenyl)-OH, Boc-Thr(benzyl)-
OH, and Boc-Tyr(2-bromobenzyloxycarbonyl)-OH. Boc-Gln-OH and
Boc-Met-OH were used without side chain protection. Coupling
sites. Our previous studies have shown that X can be any amino
acid, including â-branched amino acids such as Val.1
3,31
Our
results potentially extend the number of suitable sites for native
chemical ligation by a factor of 3 to more than 50 of the 400
dipeptide sequences found in proteins. Because there is
considerable latitude in choosing a ligation site in a target
sequence, this extended applicability will render most polypep-
tides accessible by native chemical ligation. Continuing studies
will establish the feasibility of a further extension of the
applicability of ligation chemistry based on the same principles
32
reactions were monitored by quantitative ninhydrin assay and were
typically >99%. After chain assembly was complete, peptides were
deprotected and simultaneously cleaved from the resin by treatment
with HF containing 5% p-cresol for 1 h at 0 °C to give the peptide-
R
R
R
to an even wider range of dipeptide sequences.
Native chemical ligation in its original form13 or in the form
described here provides direct synthetic access to polypeptide
2 2
COSH, - CONH , or - CO H. After removal of the HF under reduced
2
pressure, the crude peptide was precipitated in anhydrous Et O,
dissolved in HPLC buffer (40-50% B), and lyophilized.
Peptide-r-thioesters (1). Thioacid peptides were synthesized on
the appropriate Boc-aminoacyl-S-resins, made by coupling [4-[R-(N-
t-Boc-aminoacyl-S)benzyl]phenoxy]acetic acid, DCHA salt16 (2.0 equiv),
and aminomethyl-resin (1 equiv, washed with 10% DIEA in DMF)
chains the size of typical protein domains. Other ligation
chemistries1
,4,5
can be used to join synthetic domains in a
modular fashion to produce large (i.e., >20 kDa), fully
8
,9
functional synthetic proteins.
In its most general form,
with HBTU (1.6 equiv) added as an activating agent and DIEA (1
incorporating all suitable chemistries,1
,4,5,13
the chemical ligation
R
equiv), in DMF. Peptide- COSC
H
6 3
(3-CO
2
H-4-NO
2
) thioesters were
approach represents the next stage in the evolution of methods
for the chemical synthesis of polypeptides and, for the first time,
provides for reproducible, practical total chemical synthesis of
proteins.
R
generated by dissolving the crude peptide- COSH (15-20 mg) in 6 M
guanidine‚HCl, 0.1 M sodium Acetate, pH 5.0-6.5, to which was added
1.5 equiv of 5,5′-dithiobis(2-nitrobenzoic acid).1
3,17
The mixture was
vortexed briefly and purified after 10 min. The identity of the peptide-
R
6 3 2 2
COSC H (3-CO H-4-NO ) ester was unambiguously confirmed by
2
6
precise electrospray mass measurements, in contrast with Liu.
Peptide- COSCH
crude peptide- COSH (15-20 mg) in 6 M guanidine‚HCl, 0.1 M
sodium acetate, pH 4.0, to which was added 10 equiv of benzyl
bromide. The mixture was vortexed briefly and purified after 1 h.
LYRAG- COSC H (3-CO H-4-NO ) (observed mass 760 Da, calcd 760
Experimental Section
R
2
C
6
H
5
thioesters were generated by dissolving the
R
Materials and Methods. Machine-assisted solid phase peptide
syntheses were carried out on a custom-modified Applied Biosystems
1
2
13
4
30A peptide synthesizer. Reversed phase HPLC was performed on
R
a Rainin HPLC system with 214-nm UV detection, using Vydac C-18
6
3
2
2
analytical (5 µm, 0.46 × 15 cm) and semiprepatative (10 µm, 1.0 ×
Da) was purified by semipreparative HPLC (20-60% B over 40 min)
R
2
5 cm) columns. Chromatographic separations were achieved using
to give 20-30% yield. LYRAG- COSCH C H (observed mass 685
2
6
5
linear gradients of buffer B in A (A ) 0.1% TFA in water, B ) 90%
CH CN/10% water containing 0.09% TFA) over 30-60 min at 1 mL/
Da, calcd 685 Da) was purified by semipreparative HPLC (15-45%
R
3
B over 60 min) to give 25-30% yield. LYRAF- COSCH C H
2
6
5
min (analytical) or 3 mL/min (semipreparative). Mass spectra of all
peptide segments were obtained with a Sciex API-III electrospray
quadrupole mass spectrometer; observed masses were derived from the
experimental m/z values for all observed protonation states of a
(observed mass 775 Da, calcd 775 Da) was purified by semipreparative
HPLC (30-60% B over 60 min) to give 25-30% yield. Most of the
losses in yield arose simply from HPLC recoveries.
r
r
N (ethanethiol) and N (oxyethanethiol) Peptides (2). These
peptides were synthesized on either MBHA or the appropriate Boc-
(
27) Kemp, D. S.; Carey, R. I. In Peptides: Proceedings of the 11th
aminoacyl-OCH
2
-PAM-resins. After chain assembly was complete and
American Peptide Symposium; Rivier, J. E., Marshall, G. R., Eds.;
ESCOM: Lieden, 1990; pp 920-922.
R
the N Boc group removed with neat TFA (two 1-min treatments) and
neutralized with 10% DIEA in DMF (two 1-min treatments), the peptide
was bromoacetylated by the method of Robey.18 Bromoacetic acid (2.0
(
28) Blake, J.; Li, C. H. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4055-
4
058.
(
34.
29) Yamashiro, D.; Li, C. H. Int. J. Peptide Protein Res. 1988, 31, 322-
2 2
mmol) was dissolved in CH Cl (2 mL), to which was added DIC (1
mmol). After activation for 10-15 min to form the symmetric
anhydride, the mixture was diluted with DMF (2 mL), added to the
3
(30) Hojo, H.; Yoshimura, S.; Aimoto, S. In Peptide Chemistry; Okada,
Y., Ed.; Protein Research Foundation: Osaka, 1994; pp 9-12.
(
31) Simon, R. J. Personal communication. Ligations involving â-branched
amino acids at the ligation site proceed at slower rates than ligations
involving other amino acids.
(32) Sarin, V. K.; Kent, S. B. H.; Tam, J. P.; Merrifield, R. B. Anal.
Biochem. 1981, 117, 147-157.

5
896 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Canne et al.
peptide-resin, and coupled for 30 min. The resin was then washed
with DMSO, and 8a (2 M in DMSO) or 8b (1 M in DMSO) was added
and allowed to react with the bromoacetylated peptide-resin for 8-23
h. The peptides were purified without further modification after
N-[2-[S-(4-Methylbenzyl)]thio]ethoxy]phthalimide (12). Bromide
11 (16.6, 62 mmol), 4-methylbenzyl mercaptan (8.5 mL, 63 mmol)
and DBU (9.5 mL, 64 mmol) were combined in benzene (150 mL)
and stirred at room temperature for 8 h. Solids were filtered and washed
with benzene, and the filtrate was washed with 1 N HCl (2 × 35 mL),
water (1 × 35 mL), and saturated NaCl (1 × 35 mL) and dried over
cleavage from the resin. [(3-CO
2
H-4-NO
2
)-C
6
H
3
-S-SCH
2
CH
2
]-AG-
R
PAGD- CONH
2
(2a) (observed mass 800 Da, calcd 800 Da) was
purified by semipreparative HPLC (0-67% B over 60 min) to give
4
MgSO . Volatiles were removed in Vacuo and the resulting solid
R
∼
24% yield. [HSCH
2
CH
2
O]-GRNTATIMMQRGNFR- CONH
2
(2b)
recrystallized from EtOAc/hexane to yield 12 as a white solid (14.8 g,
1
(
observed mass 1827 ( 1 Da, calcd 1828 Da) was purified by
3
74%). H NMR (CDCl ): δ 7.80 (m, 4H), 7.18 (d, 2H, J ) 8.0 Hz),
semipreparative HPLC (15-40% B over 60 min) to give ∼25% yield.
7.04 (d, 2H, J ) 8.0 Hz), 4.22 (t, 2H, J ) 7.4 Hz), 3.75 (s, 2H), 2.79
R
[
HSCH
2
CH
2
O]-AARHTVHQRHLHG- COOH (2b) was synthesized by
(t, 2H, J ) 7.4 Hz), 2.27 (s, 3H). FAB MS: calcd for [C18
H
17NO
17NO
3
S,
3
S: C,
+
the above method using the racemate of 2-bromopropionic acid instead
of bromoacetic acid. The resulting crude lyophilized product was a
mixture of the desired peptide and the peptide that results from
H ] 328.1007, found 328.1016. Anal. Calcd for C18
H
66.03; H, 5.23; N, 4.28; S, 9.79. Found: C, 66.04; H, 4.95; N, 4.30;
S, 9.58.
elimination of HBr from the bromopropionylated peptide (CH
ARHTVHQRHLHG- COOH; observed mass 1502 ( 1 Da, calcd 1502
Da). The mixture was purified by semipreparative HPLC (10-40%
2
dCHCO-
S-(4-Methylbenzyl)-2-(aminooxy)ethanethiol (8b). S-(4-methyl-
benzyl)-2-(aminooxy)ethanethiol was synthesized by the method of
Osby.20 The N-substituted phthalimide 12 (7.4 g, 23 mmol) was
suspended in 2-propanol (203 mL) and water (35 mL), to which was
R
B over 60 min) to give [HSCH
observed mass 1595 ( 1 Da, calcd 1596 Da) in ∼10% yield.
S-[(3-Carboxy-4-nitrophenyl)thio]-2-aminoethanethiol (8a). 2-Ami-
noethanethiol (1.0 g, 13.0 mmol) and 5,5′-dithiobis(2-nitrobenzoic acid)
2 2
CH O]-AARHTVHQRHLHG-RCOOH
(
4
added NaBH (3.5 g, 92 mmol). The mixture was stirred at room
temperature overnight. Acetic acid (25 mL) was slowly added till
bubbling ceased and the flask stoppered and heated to 50 °C for 2-3
h. Volatiles were removed in Vacuo, and the resulting solution was
diluted with 1 N NaOH and extracted with EtOAc (4 × 50 mL). The
combined EtOAc extractions were washed with saturated NaCl (1 ×
(
(
1.75 g, 4.4 mmol) were combined in acetonitrile (100 mL) and water
25 mL) and stirred at room temperature for 14 h. The reaction mixture
was diluted with water (450 mL) and purified by preparative reversed
5
0 mL) and dried over MgSO
4
. Volatiles were removed in Vacuo,
phase HPLC on a Waters Delta Prep 4000 with a Vydac 5.0 × 2.5 cm
1
and the resulting oil was purified by flash chromatography (1:1 hexane/
EtOAc) to yield 8b as a clear, colorless oil (3.2 g, 72%). H NMR
preparative C-18 column to give 8a (1.0 g, 40%). H NMR (D
2
O): δ
1
7
.95 (d, 1H, J ) 8.7 Hz), 7.61 (dd, 1H, J ) 8.7, 2.2 Hz), 7.58 (d, 1H,
(
(
(
CDCl
br s, 2H, D
3
): δ 7.21 (d, 2H, J ) 8.0 Hz), 7.12 (d, 2H, J ) 8.0 Hz), 5.40
O exchanged), 3.77 (t, 2H, J ) 6.5 Hz), 3.71 (s, 2H), 2.64
J ) 2.2 Hz), 3.18 (t, 2H, J ) 6.5 Hz), 2.92 (t, 2H, J ) 6.7 Hz).
2
N-(2-Bromoethoxy)phthalimide (11). N-(2-Bromoethoxy)phthal-
t, 2H, J ) 6.5 Hz), 2.33 (s, 3H). FAB MS: calcd for [C10
H15NOS,
imide was synthesized by a modification of the procedure of Bauer
+
H ] 198.0953, found 198.0958. Anal. Calcd for C10
H15NOS: C, 60.88;
19
and Suresh. N-Hydroxyphthalimide (16.0 g, 98 mmol), triethylamine
30 mL, 215 mmol), and 1,2-dibromoethane (35 mL, 406 mmol) were
H, 7.66; N, 7.10; S, 16.25. Found: C, 60.79; H, 7.88; N, 7.03; S,
6.11.
(
1
combined in DMF (115 mL) and stirred at room temperature overnight.
Solids were filtered and washed with DMF. The filtrate was diluted
with water (800 mL) and the resulting precipitate filtered, dissolved in
EtOAc (200 mL), and washed with 1 N HCl (2 × 50 mL), water (1 ×
Acknowledgment. This work was supported by funds from
the NIH (Grants PO1 HL31950 and RO1GM43397, S.B.H.K.).
We gratefully acknowledge the assistance of Dr. Gary Suizdak
and Dr. Raj K. Chandra, The Scripps Research Institute, for
FAB MS measurements and for elemental analyses, respectively,
and our other colleagues at The Scripps Research Institute for
numerous helpful discussions.
5
0 mL), and saturated NaCl (1 × 50 mL), and dried over MgSO
4
.
Volatiles were removed in Vacuo, and the resulting solid was recrystal-
1
lized from 95% EtOH to give 11 as a white solid (16.6 g, 63%).
NMR (CDCl ): δ 7.82 (m, 4H), 4.49 (t, 2H, J ) 6.9 Hz), 3.65 (t, 2H,
J ) 6.9 Hz). FAB MS (sodium ion): calcd for [C10
91.9585, found 291.9579.
H
3
+
8 3
H BrNO , H ]
2
JA960398S