Sequential α-ketoacid-hydroxylamine (KAHA) ligations: Synthesis of C-terminal variants of the modifier protein UFM1

Article abstract of DOI:10.1002/anie.201204144

3 for 3: Sequential α-ketoacid-hydroxylamine (KAHA) ligations with 5-oxaproline allow access to the modifier protein UFM1 (Ubiquitin-fold modifier 1) with a C-terminal amide, carboxylic acid, or a masked thioester. Fmoc protection of an N-terminal 5-oxaproline permits the assembly of proteins of >80 residues in good yield by a two-pot process from three readily prepared medium-sized protein segments. Copyright

Full text of DOI:10.1002/anie.201204144

Angewandte
Chemie
DOI: 10.1002/anie.201204144
Protein Synthesis
Sequential a-Ketoacid-Hydroxylamine (KAHA) Ligations: Synthesis
of C-Terminal Variants of the Modifier Protein UFM1**
Ayodele O. Ogunkoya, Vijaya R. Pattabiraman, and Jeffrey W. Bode*
The role of conjugation and deconjugation of modifier
proteins, such as ubiquitin and SUMO, to their targets has
rapidly emerged as one of the most exciting areas of
molecular biology and drug discovery.[1] The development of
new therapeutics perturbing these regulatory pathways
requires access to labeled and derivatized modifier proteins
and a means to prepare their conjugates.[2] In some cases, the
modifier protein, the target protein, and the specific ligases
can be obtained from natural sources but research is increas-
ingly turning to synthetic methods to prepare them and study
their biological roles.[2,3] In this regard, the native chemical
ligation (NCL) of C-terminal modifier protein thioesters to
a 5-thiolysine residue in a protein stands out for its promise to
impact this area.[4] The unnatural 5-thiolysine residue can be
incorporated into a target protein by total chemical syn-
thesis,[5] semi-synthesis from expressed protein segments[6] or
by ribosomal incorporation.[7] It has already been used to
attach ubiquitin to a-synuclein[6] and SUMO[7] and to prepare
di-,[4,8] tri-,[5] and tetra-ubiquitins.[9]
thiols,[15] the use of specific E1 enzymes, or the total chemical
synthesis of C-terminal thioesters. The size of these modifier
proteins (70–140 residues) generally precludes linear solid-
phase synthesis of the thioesters and many are arguably
outside the range of two-segment ligations. A three- or four-
segment ligation strategy is required.
We now report the chemical synthesis of C-terminal
variants of ubiquitin-fold modifier 1 (UFM1) by multiple
segment ligations using the a-ketoacid-hydroxylamine
(KAHA) ligation with 5-oxaproline (Opr). All of the
necessary peptide segments for ligation can be easily pre-
pared by standard Fmoc solid-phase peptide synthesis
(SPPS). The N-terminal Fmoc-protected Opr peptides are
stable to resin cleavage, purification, oxidation, and ligation
but can be rapidly and cleanly removed from the otherwise
unprotected ligation product. This work is the first alternative
to native chemical ligation for sequential chemoselective
ligations that afford native amide bonds.[16]
The key to any sequential ligation strategy (Figure 1) is
the identification of a suitable protecting group for the N or
the C terminus of a ligating segment. Protection strategies for
The execution of synthetic and semi-synthetic approaches
to preparing modifier protein–protein conjugates requires C-
terminal derivatives of modifier
proteins. Access to such molecules
is currently limited to ubiqui-
tin,[4,5,8,9,10,11a] SUMO, Nedd8, and
ISG15[11a] prepared either syntheti-
cally, semi-synthetically by NCL of
an expressed protein fragment or
with the use of their specific E1
enzymes[11a] to provide thioesters
for NCL. The extension of this
technique to the preparation and
incorporation of other modifier
proteins
including
FUB1,[12]
MUB,[13] and UFM1[14] would
require either the expression of
intein-forming fusion proteins fol-
lowed by capture with exogenous
Figure 1. Sequential KAHA ligation. PG=Protecting group, X=C-terminal variants.
sequential NCL include N-terminal protection of cysteine
groups as a thiazolidine and their liberation with hydroxyl-
amine and C-terminal thioesters prepared post-ligation from
rearrangement.[17] The major limitation of sequential NCL is
the need for multiple cysteine residues that must be desul-
furized to avoid interference in protein function. In contrast,
KAHA-ligation with Opr results in the formation of a-
homoserine (Hse, T§), which may serve as an innocuous
substitute for other amino acids.[18]
[*] A. O. Ogunkoya, Dr. V. R. Pattabiraman, Prof. Dr. J. W. Bode
Laboratorium fꢀr Organische Chemie, Department of Chemistry
and Applied Biosciences, ETH Zꢀrich
Wolfgang Pauli Strasse 10, 8093 Zꢀrich (Switzerland)
E-mail: bode@org.chem.ethz.ch
Homepage: http://www.bode.ethz.ch
[**] This work was supported by the Swiss National Science Foundation
(200021_131957) and an ETH Fellowship to V.R.P. (ETH FEL 06-10-
1). We are grateful to Dr. Xiangyang Zhang and Rolf Hꢁfliger of the
LOC Mass Spectrometry Service for analyses.
The primary goal of our study was to identify a protecting
group that fulfilled the following criteria: 1) orthogonal to the
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201204144.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
acid-labile protecting groups typically employed in SPPS, 2)
resistant to oxidative conditions required to transform
cyanosulfurylides (SY) to a-ketoacids (KA), and 3) stability
during the ligation and purification of unprotected peptide
segments. The use of groups cleaved by thiols, reduction, and
metal catalysts such as Ns, Troc, and Alloc, were deemed
incompatible with unprotected peptides or only partially
fulfilled the outlined criteria. A survey of base-labile protect-
ing groups suggested that they may not be compatible with
unprotected functional groups or could be difficult to
deprotect post-ligation. Despite these concerns—and lacking
a better option—we elected to examine the use of Fmoc-
based protection for sequential KAHA ligations with Opr.
We therefore prepared enantiopure (S)-N-Fmoc-Opr 1 (see
Supporting Information), using modified literature prece-
dents.[18,19]
As part of our interest in synthetic and semi-synthetic
studies of new modifier proteins, we sought to provide access
to three C-terminal variants of UFM1: 1) the C-terminal acid,
which could be conjugated to a target protein with an
appropriate E1 ligase; 2) a suitably protected C-terminal
thioester, which in combination with a target protein con-
taining a 5-thiolysine, could be conjugated by native chemical
ligation; 3) a C-terminal amide, which would serve as an
Scheme 1. Preparation of UFM1 peptide segments: a) Synthesis of UFM1 Fmoc-(30–60) 5, b) UFM1 (61–83) 7a, 7b, and 7c, c) and UFM1 (2–29)
9a and 9b. DIPEA=N,N-diisopropylethylamine, TFA=trifluoroacetic acid, TIPS=triisopropylsilane, Fmoc=9-fluorenylmethoxycarbonyl, Trt=tri-
phenylmethyl, Pbf=2,2,4,6,7-pentamethyldihydro-1-benzofurane-5-sulfonyl, HCTU is a uronium-type coupling reagent.
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
important negative control in biochemical conjugation stud-
ies. The sequential ligation strategy with modular protein
segments would allow the efficient preparation of C-terminal
protein variants with minimum synthetic effort. In general,
the preparation of C-terminal variants of proteins as opposed
to the N-terminal modifications is arduous by linear SPPS, as
it requires individual solid-phase synthesis of each of the C-
terminal protein variant.[20]
Our sequential strategy for synthesis of UFM1 (2–83)
variants required the preparation of three segments
(Scheme 1). We selected ligation sites at Phe29-Thr30 and
between Ala60-Gln61. This strategy would give synthetic C-
terminal modified UFM1 with two mutations to the native
sequence: T30T§ and Q61T§. It also allowed us to assess the
viability of the KAHA ligation with N-terminal Opr segments
at Ala and Phe-a-KA; our prior studies examined Leu, Tyr,
and Glu.
previously established procedure for the synthesis of peptide
a-KA (Scheme 1a). Cyanosulfurylide linker on Rink amide
resin 3 was used to prepare UFM1 (31–60) by automated
SPPS.[21] Fmoc-Opr 1 was manually coupled with HBTU.
Side-chain deprotection and cleavage of the peptide from the
resin gave 4 without any side products. Unprotected SY 4 was
oxidized with dimethyldioxirane (DMDO) and purified by
preparative HPLC to provide analytically pure a-KA 5,
leaving the N-terminal Opr30 unaffected. The Fmoc-pro-
tected Opr30 proved to be stable throughout the oxidation,
purification and handling.
UFM1 (61–83) segments with three distinct C-termini and
an unprotected N-terminal Opr61 (7a–c) (Scheme 1b) were
prepared by standard methods.[18] The C-terminal amide
segment 7a and C-terminal acid segment 7b were prepared
by Fmoc SPPS on Rink amide and HMPB resins, respectively.
Thioester precursor 7c was prepared from Rink amide resin
pre-loaded with Fmoc-N-Me-Cys(o-NO2-Bn)-OH.[9] In each
case, automated SPPS was used to extend the peptides to
Preparation of Fmoc-Opr-(31–60)-a-KA 5, the founda-
tion of this sequential strategy, was accomplished by using our
Scheme 2. Synthesis of UFM1 proteins by sequential KAHA ligations: a) HPLC monitoring of the ligation between UFM1 Fmoc-(30–60)-KA 5 and
(61–83)-N-Me-Cys(o-NO2Bn) 7c, b) HPLC monitoring of ligation between UFM1 (2–29)-KA 9a and (30–83)-N-Me-Cys(o-NO2Bn) 11c.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
residue 62 (6a–c) and (S)-N-Boc-Opr61 2 was introduced by
manual coupling with HCTU. Acid-promoted cleavage from
the resin followed by purification gave unprotected peptides
7a–c.
water in DMSO with 0.1m oxalic acid at 508C. It was
essentially complete within 24 h and provided 12a in 45%
isolated yield with no observable isomeric product (Sche-
me 2b). Qualitatively, this ligation was more efficient than the
first and proceeded at lower temperature, despite the fact that
a bulkier a-KA (Phe vs. Ala) was used for the ligation. We
attribute the improved reactivity at lower temperature to the
enhanced solubility of the peptides and increased proportion
of water in the ligation mixture.
The sequential ligation procedures were readily extended
to the synthesis of C-terminal UFM1 carboxylic acid 12b and
masked thioester 12c. In all the cases, the first and second
ligation, as well as the in situ Fmoc-deprotection proceeded
under similar conditions and yields as described for the
synthesis of UFM1 variant 12a. Only minor variations in the
reaction conditions due to solubility and reaction scale were
made. For example, the second ligation of 9a with 11b to give
12b was performed at 9 mm in 25% water in DMSO but was
essentially finished after 24 h at 508C.
The purity, sequence identity and exact mass of the final
proteins were confirmed by analytical HPLC, high-resolution
mass spectrometry (FTMS), MS/MS analysis and SDS-PAGE
(Figure 2b). Also of importance for the synthetic UFM1
variants is their tertiary structure, which has a key role in its
biological function. NMR studies of UFM1 protein have
shown that it contains multiple b-strands and a-helical
regions.[23] CD spectra recorded at 238C for the synthetic
UFM1 variants 12a, 12b and 12c in 10 mm sodium phosphate
buffer with 100 mm NaCl at pH 6.0 indicated the presence of
folded proteins with b-strand and a-helical motifs (Fig-
ure 2a).
The N-terminal segments, UFM1 (2–29)-a-KA 9a and 9b
(Scheme 1c), were prepared by SPPS on 3 followed by acidic
resin cleavage, oxidation with DMDO and purification by
HPLC. This procedure afforded the Phe a-KA with either
a free N-terminus (9a) or an Fmoc-labeled N-terminus (9b).
Our studies on the assembly of the segments began with
the ligation of a-KA 5 and Opr 7a, using a small excess of the
a-KA segment (1:1.3–2.0 equiv, respectively). In our previous
report of the Opr KAHA ligation, we found that higher
proportions of water accelerate the ligation.[18] Unfortunately,
due to peptide solubility, it was difficult to perform ligations
with acceptable concentrations of peptides (10–20 mm) con-
taining higher than 10% water in DMSO (v/v). Nonetheless,
with 10% water in DMSO and 0.1m oxalic acid at 608C, the
ligation (14–15 mm in peptides) proceeded smoothly within
20 h to give 10a (Scheme 2a). Importantly, we observed no
loss of Fmoc from the protected Opr on segment 5.
Purification of the ligation mixture by HPLC afforded the
ligated product in 46% yield. Unlike our previous work and
the second ligation (see below) we observed a small amount
of an isomeric side-product (ca. 10%) corresponding to the
same mass as the desired product.[22] This isomeric product
was easily separated from the desired ligation product by
preparative HPLC.
We were pleased that the Fmoc-protected Opr30 was
stable to the ligation and purification conditions; the remain-
ing milestone was its effective and high-yielding deprotection.
Fmoc removal is typically conducted on fully protected
peptides and a protocol for the rapid and high-yielding
deprotection of otherwise unprotected ligation products
required some development. The choice of base and reaction
temperature proved essential to avoid a side product that was
detected under some conditions. After careful optimization,
we determined that treating peptide 10a at 10–158C with 5%
Et2NH in DMSO for 5–10 min cleanly gave the desired
product 11a. For development studies, we opted to purify the
deprotected peptide by HPLC prior to examining the
conditions for the final ligation. In practice, and in our
subsequent syntheses of C-terminally modified UFM1 pro-
teins, we performed the Fmoc-deprotection immediately after
the ligation in the same reaction vessel. For example, the
ligation mixture of 5 and 7b to afford 10b was cooled to 10–
158C and treated with a 10% solution of Et2NH in DMSO
(final concentration ca. 5% Et2NH) for 10 min. The unpro-
tected ligation product 11b was purified and isolated by
preparative HPLC in 45% overall yield. This procedure
reduces the number of manipulations needed for sequential
ligations and consequently improves the overall yields. We
have adopted this in situ deprotection in all of our subsequent
studies on the synthesis of UFM1 and other protein targets.
The final ligations required for the synthesis of UFM1
variants was initially examined with ligation product 11a and
Fmoc-labeled peptide 9b, as this allowed for facile monitoring
due to the Fmoc-chromophore. The ligation was performed
by using 1.0 equiv 11a and 1.4 equiv 9b at 20 mm in 30%
Figure 2. a) CD spectrum of UFM1 (2–83)-N-Me-Cys(o-NO2-Bn) 12c at
238C; inset: solution NMR structure of UFM1.[23] b) SDS-PAGE of
protein ladder (14.3–220 kDa, lane 1) and pure 12c (9.0 kDa, lane 2).
c) Deconvoluted ESI-FTMS of 12c.
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!

Angewandte
Chemie
[10] L. A. Erlich, K. S. A. Kumar, M. Haj-Yahya, P. E. Dawson, A.
In conclusion, we have shown that Fmoc-protected 5-
oxaproline is suitable for sequential KAHA ligation for
protein synthesis. The development of this strategy made
possible the facile synthesis of three C-terminal variants of
UFM1, an important but under-studied modifier protein of
contemporary interest. These are the largest proteins with
native amide bonds prepared to date by segment ligations
using a method other than native chemical ligation. These
studies will both improve access to synthetic modifier proteins
for biochemical studies and provide a new approach to
iterative segment assembly for chemical protein synthesis.
Brik, Org. Biomol. Chem. 2010, 8, 2392 – 2396.
[11] a) P. P. Geurink, F. El Qualid, A. Jonker, D. S. Hameed, H.
Ovaa, ChemBioChem 2012, 13, 293 – 297; b) O. N. Burchak, M.
Jaquinod, C. Cottin, L. Mugherli, K. Iwai, F. Chatelain, M. Y.
Balakirev, ChemBioChem 2006, 7, 1667 – 1669.
[12] K. Kas, E. Schoenmakers, W. Van de Ven, G. Weber, M.
Nordenskjold, L. Michiels, J. Merregaert, C. Larsson, Genomics
1993, 17, 387 – 392.
[13] B. P. Downes, S. A. Saracco, S. S. Lee, D. N. Crowell, R. D.
Vierstra, J. Biol. Chem. 2006, 281, 27145 – 27157.
[14] M. Komatsu, T. Chiba, K. Tatsumi, S. Lemura, I. Tanida, N.
Okazaki, T. Ueno, E. Kominami, T. Natsume, K. Tanaka, EMBO
J. 2004, 23, 1977 – 1986.
[15] a) T. W. Muir, Annu. Rev. Biochem. 2003, 72, 297 – 303; b) R. M.
Hofmann, Curr. Opin. Biotechnol. 2002, 13, 297 – 303; c) C.
Chatterjee, R. K. McGinty, J. P. Pellois, T. W. Muir, Angew.
Chem. 2007, 119, 2872 – 2876; Angew. Chem. Int. Ed. 2007, 46,
2814 – 2818.
[16] a) I. E. Valverde, F. Lecaille, G. Lalmanach, V. Aucagne, A. F.
Delmas, Angew. Chem. 2012, 124, 742 – 746; Angew. Chem. Int.
Ed. 2012, 51, 718 – 722; b) Z. Miao, J. P. Tan, J. Am. Chem. Soc.
2000, 122, 4253 – 4260.
Received: May 28, 2012
Published online: && &&, &&&&
Keywords: amides · chemoselectivity · ligation reactions ·
.
protein synthesis · proteins
[1] a) M. Hochstrasser, Nature 2009, 458, 422 – 429; b) E. R. Strieter,
D. A. Korasick, Chem. Biol. 2012, 7, 52 – 63.
[17] a) M. Villain, J. Vizzavona, K. Rose, Chem. Biol. 2001, 8, 673 –
679; b) S. Ueda, M. Fujita, H. Tamamura, N. Fujii, A. Otaka,
ChemBioChem 2005, 6, 1983 – 1986; c) G. Fang, Y. Li, F. Shen, Y.
Huang, J. Li, Y. Lin, H. Cui, L. Lui, Angew. Chem. 2011, 123,
7787 – 7791; Angew. Chem. Int. Ed. 2011, 50, 7645 – 7649; d) N.
Ollivier, J. Vicogne, A. Vallin, H. Drobecq, R. Desmet, O.
El Mahdi, B. Leclercq, G. Goormachtigh, V. Fafeur, O. Melnyk,
Angew. Chem. 2012, 124, 213 – 217; Angew. Chem. Int. Ed. 2012,
51, 209 – 213.
[18] V. R. Pattabiraman, A. O. Ogunkoya, J. W. Bode, Angew. Chem.
2012, 124, 5204 – 5208; Angew. Chem. Int. Ed. 2012, 51, 5114 –
5118.
[19] A. Vasella, R. Voeffray, J. Chem. Soc. Chem. Commun. 1981,
97 – 98.
[20] J. Alsina, F. Albericio, Biopolymers 2003, 71, 454 – 477.
[21] a) L. Ju, J. W. Bode, Org. Biomol. Chem. 2009, 7, 2259 – 2264;
b) L. Ju, A. R. Lippert, J. W. Bode, J. Am. Chem. Soc. 2008, 130,
4253 – 4255.
[22] J. Wu, J. Ruiz-Rodriguez, J. M. Cornstock, J. Z. Dong, J. W.
Bode, Chem. Sci. 2011, 2, 1976 – 1979.
[23] H. Sasakawa, E. Sakata, Y. Yamaguchi, M. Komatsu, K. Tatsumi,
E. Kominami, K. Tanaka, K. Kato, Biochem. Biophys. Res.
Commun. 2006, 343, 21 – 26.
[2] T. Fekner, X. Li, M. K. Chan, ChemBioChem 2011, 12, 21 – 33.
[3] P. Siman, A. Brik, Org. Biomol. Chem. 2012, DOI: 10.1039/
C2OB25149C.
[4] a) K. S. Ajish Kumar, M. Haj-Yahta, D. Olschewski, H. A.
Lashuel, A. Brik, Angew. Chem. 2009, 121, 8234 – 8238;
Angew. Chem. Int. Ed. 2009, 48, 8090 – 8094; b) R. Yang, K. K.
Pasunooti, F. Li, X. W. Liu, C. F. Liu, J. Am. Chem. Soc. 2009,
131, 13592 – 13593; c) F. El Qualid, R. Merkx, R. Ekkebus, D. S.
Hameed, J. J. Smit, A. de Jong, H. Hilkmann, T. K. Sixma, H.
Ovaa, Angew. Chem. 2010, 122, 10347 – 10351; Angew. Chem.
Int. Ed. 2010, 49, 10149 – 10153.
[5] S. N. Bavikar, L. Spasser, M. Haj-Yahya, S. V. Karthikeyan, T.
Moyal, K. S. A. Kumar, A. Brik, Angew. Chem. 2012, 124, 782 –
787; Angew. Chem. Int. Ed. 2012, 51, 758 – 763.
[6] M. Hejjaoui, M. Haj-Yahya, K. S. A. Kumar, A. Brik, H. A.
Lashuel, Angew. Chem. 2011, 123, 425 – 429; Angew. Chem. Int.
Ed. 2011, 50, 405 – 409.
[7] S. Virdee, P. B. Kapadnis, T. Elliott, K. Lang, J. Madrzak, D. P.
Nguyen, L. Riechmann, J. W. Chin, J. Am. Chem. Soc. 2011, 133,
10708 – 10711.
[8] K. S. A. Kumar, L. Spasser, L. A. Erlich, S. N. Bavikar, A. Brik,
Angew. Chem. 2010, 122, 9312 – 9317; Angew. Chem. Int. Ed.
2010, 49, 9126 – 9131.
[9] K. S. A. Kumar, S. N. Bavikar, L. Spasser, T Moyal, S. Ohayon,
A. Brik, Angew. Chem. 2011, 123, 6261 – 6265; Angew. Chem.
Int. Ed. 2011, 50, 6137 – 6141.
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Protein Synthesis
A. O. Ogunkoya, V. R. Pattabiraman,
J. W. Bode*
&&&& — &&&&
Sequential a-Ketoacid-Hydroxylamine
(KAHA) Ligations: Synthesis of C-
Terminal Variants of the Modifier Protein
UFM1
3 for 3: Sequential a-ketoacid–hydroxyl-
amine (KAHA) ligations with 5-oxaproline
allow access to the modifier protein
UFM1 (Ubiquitin-fold modifier 1) with
a C-terminal amide, carboxylic acid, or
a masked thioester. Fmoc protection of
an N-terminal 5-oxaproline permits the
assembly of proteins >80 residues in
good yield by a two-pot process from
three readily prepared medium-sized
protein segments.
6
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 6
These are not the final page numbers!