DOI: 10.1002/cbic.201000168
Protein Synthesis Assisted by Native Chemical Ligation at Leucine
Ziv Harpaz, Peter Siman, K. S. Ajish Kumar, and Ashraf Brik*[a]
Total chemical synthesis of proteins offers exceptional opportu-
nities for preparing targets with exquisite control over the co-
valent structure, high purity and large quantities for functional
and structural analysis.[1] In this regard, native chemical ligation
(NCL) continues to be the method of choice for joining two
unprotected peptides through thiol capture followed by S–N
acyl transfer to form the amide bond at the ligation site.[2] The
desulfurization reaction introduced by Yan and Dawson[3] as a
post-NCL step greatly expanded the scope of ligation chemis-
try beyond Xaa-Cys (Xaa is any amino acid) by making ligation
at Xaa-Ala sites accessible in the synthesis of functional pro-
teins.[4] This seminal work has prompted several groups to
extend this strategy to other b- and g-mercapto amino acid
derivatives (e.g., Phe and Val) as part of the ongoing efforts to
extend NCL to essentially any ligation junction.[5] In addition,
by positioning a thiol handle at the Lys side chain and at the
2-acetamido group of the glycan moiety in a glycopeptide al-
lowed NCL to be used for peptide ubiquitylation[6] and glyco-
peptide ligation[7] respectively, and after a desulfurization step,
the native structures could be achieved.
Scheme 1. General strategy for NCL at Leu.
The above-described examples testify to the power of desul-
furization when combined with NCL to assist the synthesis of
naturally occurring proteins. In addition, recent reports on the
mild and highly versatile free-radical Cys reduction proto-
col,[5a,b,8] as well as the compatibility of these methods in the
presence of other thiol functionalities[9] permit the use of this
two-step approach in the synthesis of a variety of protein tar-
gets. Despite the introduction of mercaptoPhe and mercapto-
Val to assist efficient peptide ligation,[5] their utility in protein
synthesis has not yet been demonstrated. The preparation of
proteins by using ligation methods is often challenging be-
cause decreases in rate and chemoselectivity only come to
light in more complex systems. To date, only ligation at Xaa-
Cys followed by desulfurization has been demonstrated in the
synthesis of full-length proteins,[4] this diminishes the generali-
ty of the desulfurization approach in the total chemical synthe-
sis of proteins. Here we report an innovative strategy for liga-
tion at Xaa-Leu sites by using b-mercaptoleucine combined
with desulfurization (Scheme 1) and its application in the total
chemical synthesis of HIV-1 Tat protein.
commercially available threo-b-hydroxy-l-leucine. In the first
route, the b-hydroxy group could be converted to a good leav-
ing group followed by nucleophilic substitution with a thiol
nucleophile, whereas in the second one an aziridine intermedi-
ate could be employed to introduce the thiol functionality
through a regioselective ring opening. Applying the first strat-
egy by activation of the b-hydroxy group as a mesylate or to-
sylate derivative, then substitution with PMB-SH or thioacetic
acid as a source of the thiolate functionality led to formation
of the b-elimination product in a substantial amount (>60%).
Additionally, it is very likely that under these conditions the
thiol-substituted product could also be formed through a Mi-
chael addition, which would lead to a racemization at the a-
carbon, rendering this method less desirable for us. Inspired by
the recent work on the racemization-free synthesis of b-meth-
ylcysteine by starting from Thr,[10] we decided to follow a simi-
lar path to prepare the desired building block (Scheme 2).
Thus, b-hydroxy-l-leucine was converted to the methyl ester
derivative followed by treatment with p-nitrophenylsulfonyl
chloride (NsCl) to give the p-nitrosulfonamide 2 in 70% yield
(two steps). Ring closure of 2 under Mitsunobu conditions
afforded the aziridine 3 in 96% yield. Subsequently, BF3·OEt2-
mediated ring opening with PMB-SH furnished the two re-
gioisomers 4 and 5 in a quantitative yield in a 6:4 ratio respec-
tively. Apparently, the additional steric hindrance in our case
compared to Thr (isopropyl vs. methyl) decreased the desired
regioselectivity. Notwithstanding the unsatisfactory regioselec-
tivity of the aziridine opening, the regioisomers 4 and 5 were
To implement the approach outlined in Scheme 1, we first
had to design a synthetic strategy for b-mercaptoleucine, the
key residue in this strategy. We envisioned two routes to ach-
ieve the synthesis of this building block by starting from the
[a] Z. Harpaz, P. Siman, Dr. K. S. A. Kumar, Dr. A. Brik
Department of Chemistry, Ben-Gurion University of the Negev
Beer Sheva 84105 (Israel)
Fax: (+972)8-6472943
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201000168.
1232
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 1232 – 1235