An oxazetidine amino acid for chemical protein synthesis by rapid, serine-forming ligations

Article abstract of DOI:10.1038/nchem.2282

Amide-forming ligation reactions allow the chemical synthesis of proteins by the union of unprotected peptide segments, and enable the preparation of protein derivatives not accessible by expression or bioengineering approaches. The native chemical ligati

Full text of DOI:10.1038/nchem.2282

ARTICLES
PUBLISHED ONLINE: 22 JUNE 2015 | DOI: 10.1038/NCHEM.2282
An oxazetidine amino acid for chemical protein
synthesis by rapid, serine-forming ligations
1
1,2
*
Ivano Pusterla and Jeffrey W. Bode
Amide-forming ligation reactions allow the chemical synthesis of proteins by the union of unprotected peptide segments,
and enable the preparation of protein derivatives not accessible by expression or bioengineering approaches. The native
chemical ligation (NCL) of thioesters and N-terminal cysteines is unquestionably the most successful approach, but is not
ideal for all synthetic targets. Here we describe the synthesis of an Fmoc-protected oxazetidine amino acid for use in the
α-ketoacid–hydroxylamine (KAHA) amide ligation. When incorporated at the N-terminus of a peptide segment, this four-
membered cyclic hydroxylamine can be used for rapid serine-forming ligations with peptide α-ketoacids. This ligation
operates at low concentration (100 μM–5 mM) and mild temperatures (20–25 °C). The utility of the reaction was
demonstrated by the synthesis of S100A4, a 12 kDa calcium-binding protein not easily accessible by NCL or other amide-
forming reactions due to its primary sequence and properties.
ccess to proteins is essential for modern medicine and bio- operates at low concentrations (100 μM–5 mM) and mild tempera-
logical science. As a complement to recombinant protein tures (20–25 °C) with short reaction times. The utility of this amino
A
expression, the chemical synthesis of proteins has emerged acid was demonstrated in the synthesis of S100A4 (metastasin), a
as a viable approach to some protein targets and offers advantages 12 kDa calcium-binding protein not easily accessible by NCL due
including homogeneity, the potential to incorporate any unnatural to its primary sequence and properties.
amino acid or post-translational modification, and the ability to
prepare protein enantiomers1 . The native chemical ligation Results
–3
(NCL) of thioesters and N-terminal cysteines—reported by Kent At the outset of our studies it was unclear whether the necessary 1,2-
and co-workers over twenty years ago—has been a transformative oxazetidines could be synthesized or if they would be stable despite the
–1
10
advance in protein synthesis, but is not ideal for many synthetic high inherent ring strain (25 kcal mol , Fig. 1a) —higher even than
4
–1
targets . To identify more general and complementary protein lig- the well-studied three-membered ring oxaziridines (23 kcal mol ).
ation reactions, numerous groups have pursued the development Furthermore, no previous examples of unsubstituted (that is, N–H)
5
11
of novel methods and ligation partners .
Our own research efforts identified the union of α-ketoacids and oxazetine (N-Boc) has appeared, although a single computational
oxazetidine had been reported and only one example of a protected
1
0
hydroxylamines (KAHA ligation) as a promising new chemoselec- study suggested that such compounds might be stable .
tive protein ligation . We have synthesized proteins including
Pup, CspA, UFM1 and SUMO2/3 by KAHA ligations between intramolecular N-alkylation , intramolecular C-alkylation and
C-terminal peptide α-ketoacids and segments bearing an N-term- [2+2] cycloadditions, were complicated by the unique reactivity of
inal 5-oxaproline residue (5-Opr) . This ligation has proven to be the N–O bond in our system and the well-known difficulty of
remarkably robust, but has limitations including the introduction closing four-membered rings . We reasoned that if a hydroxyl-
of a non-canonical homoserine residue at the ligation site, the for- amine-epoxide (Fig. 1b) underwent an intramolecular cyclization,
mation of esters as the primary ligation products , and a preference it could potentially give the oxazetidine by a 4-exo-tet pathway
for relatively high concentrations (10–20 mM) and temperatures rather than the disfavoured 5-endo-tet closure , as a similar strat-
50–60 °C). The success of the 5-Opr, a cyclic five-membered ring egy has proven successful for cyclobutane formation . A carba-
6,7
Our early attempts at devising a synthetic entry to 1, including
1
2
13
8
1
4
9
1
5,16
1
7
(
hydroxylamine, in the ligation appears to derive from ring strain; mate-protected hydroxylamine 2 (Fig. 1b) was synthesized and
analogous acyclic hydroxylamines are unreactive. Based on this subjected to cyclization conditions, which cleanly, but unfortu-
observation, we postulated that oxazetidine monomers—four-mem- nately, delivered 3 by 6-exo-tet ring closure of the Boc oxygen
1
8
t
bered ring hydoxylamines—would address all of the remaining atom . A similar cyclization, followed by Bu migration to give 4,
limitations of the KAHA ligation for protein synthesis by providing occurred when 2 was treated with BF ·OEt (Fig. 1b). We opted
3
2
native serine or threonine residues, operating at lower concen- instead for N-benzyl protected substrate 5, which cannot close to
trations and temperatures due to increased reactivity, and affording form a six-membered ring, and was prepared in good yield.
directly the amide products.
During its preparation, we serendipitously found that it easily
Here, we describe the first synthesis of an oxazetidine amino acid cyclized to the undesired isoxazoline 6, presumably via the interme-
by a chemoselective and stereospecific rearrangement to form the diacy of an alkyl chloride. Although we initially expected the desired
four-membered ring hydroxylamine. When incorporated at the cyclization to proceed under basic conditions, this substrate resisted
N-terminus of a peptide segment, this highly strained—but suffi- even harsh treatments (for example, potassium hexamethyldisily-
1
9
ciently stable—amino acid residue can be used for rapid serine- lamide (KHMDS), 80 °C, in dimethyl sulfoxide (DMSO)) .
forming ligations with the KAHA amide-forming reaction. It Continued studies revealed a critical role for Lewis acid activation
1
Department of Chemistry and Applied Biosciences, Laboratorium für Organische Chemie, Vladimir Prelog Weg 3, ETH Zürich, Zürich 8093, Switzerland.
Institute of Transformative Bio-Molecules (WPI-ITbM), Nagoya University, Chikusa, Nagoya 464-8602, Japan. *e-mail: bode@org.chem.ethz.ch
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2282
ARTICLES
a
b
tBu
O
O
O
O
OH
HN
HN
HN
KHMDS
HCl
N
O
O
Oxaziridine
1,2-oxazetidine
Isoxazolidine
DMSO, 80 °C
EtOH/H2O
92%
calc. 22.9 kcal mol–1
)
(calc. 25.2 kcal mol–1)
(calc. 6.7 kcal mol–1)
OH
N
(
O
90%
Bn
3
6
O
O
Pg = Boc
compound 2)
Pg
O
Pg = Bn
(compound 5)
(
N
HN
H
Cyclobutane
Azetidine
Oxetane
O
exp. 26.5 kcal mol–1
)
(exp. 25.2 kcal mol–1)
(exp. 24.7 kcal mol–1)
Bn
(
tBu
N
O
BF3•OEt2
CH2Cl2
BF •OEt
N
O
3
2
O
OH
CH2Cl2
90%
HO
57%
4
7
c
HO
O
N
H2, Pd(OH)2
EtOAc, 0 °C
OH
Ph
OH
OH
Ph
N
H
7
8
NO2
O
N
O
N
3,5-dinitrobenzoylchloride
Ph
DMAP, NEt3,
CH2Cl2
Ph
O
7
70%
9
O
NO2
Crystalline derivative
9
d
OH
N
MeO
MeO
(
1) NaBH3CN, HCl,
MeOH, 99%
(1) (2S)-glycidyl tosylate, K CO ,
2
3
OH
O
DMF, 65 °C, 90%
N
O
NH
O
(
2) Alloc-Cl, NEt3,
CH2Cl2, 97%
(2) Pd(PPh ) , morpholine,
THF, 85%
3
4
MeO
OMe
OMe
O
OMe
10
12
2,4-dimethoxybenzaldoxime
BF3•OEt2, CH2Cl2,
40 °C to 0 °C,
–
76%
MeO
MeO
(
1) DDQ, CH2Cl2 (wet),
78 °C, 8 h
O
O
N
O
–
TBSCl, imidazole
DMF, 0 °C, 73%
N
OTBS
OH
N
OTBS
Fmoc
(2) FmocCl, NEt3, 52%
OMe
OMe
14
13
15
HF (aq.)
MeCN, 0 °C
48%
(
1) (COCl)2, DMSO
DIPEA, CH2Cl2
–
78 °C to 0 °C
O
N
O
N
Stable for months
Coupled with standard SPPS methods
Stable to TFA cleavage conditions and HPLC purification
OH
OH
Fmoc
(2) NaCl2O, NaH2PO4
Fmoc
2
-methyl-2-butene
O
16
tBuOH/THF/H2O, 40%
1
Figure 1 | Synthesis of enantioenriched Fmoc-oxazetidine amino acid 1. a, Ring strain in 1,2-oxazetidine and parent compounds. Calculated or experimental
strain energies are shown in parentheses (see Supplementary Information for calculations). b, Cyclizations from the bifunctional epoxide-hydroxylamine.
c, Characterization of the oxazetidine. d, Synthetic route to Fmoc-protected amino acid 1. DMAP, 4-dimethyaminopyridine; TBS, tert-butyldimethylsilyl; DDQ,
2
,3-dichloro-5,6-dicyano-1,4-benzoquinone; Fmoc, 9-fluorenylmethyloxycarbonyl; DIPEA, diisopropylethylamine; TFA, trifluoroacetic acid.
of the epoxide, and treatment of 5 with BF ·OEt in CH Cl furn-
A literature survey of N-alkyl protecting groups that can be
3
2
2
2
ished the desired oxazetidine 7 in excellent yield. Unfortunately, removed under neutral, non-reducing conditions provided little
while the N-Bn group could be easily removed from five-membered encouragement. Despite the fact that almost no examples of amine
ring 6 without affecting the hydroxylamine, all attempts at N-Bn deprotections of oxidatively removable 2,4-dimethoxybenzyl were
2
0
removal from 7 resulted in N–O bond cleavage, reflecting the known , we selected this group in the hope of identifying conditions
much higher ring strain. This reduction confirmed the structure for its removal without destruction of the oxazetidine ring. After
by comparison to known compound 8, an assignment further ver- extensive optimization, we found that it could be removed by treat-
ified by X-ray analysis of ester derivative 9 (Fig. 1c). Unfortunately, ment with DDQ to give an N-unprotected oxazetidine that was
these results exposed 7 as a dead-end for the synthesis of the immediately protected with FmocCl. On the basis of these results,
N-Fmoc oxazetidine amino acid 1.
we developed a scalable synthesis of enantioenriched monomer 1 by
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2282
ARTICLES
N
N
NHPbf
HN
HN
OH
CONH2
NH2
O
OH
CONH2
NH2
O
HN
NH
H
N
DMSO/H2O (20 mM)
H2N
Ala-Val-Ser-Glu-Ile-Gln-Phe-Met-His-Asn-Leu-Gly-Lys-His-Leu
OH
H N Ala-Val-Ser-Glu-Ile-Gln-Phe-Met-His-Asn-Leu-Gly-Lys-His-Leu
2
Phe-Arg-Ala
O
0.1 M oxalic acid
40 °C
CO2H
S
CONH2
O
Me
HN
CO2H
S
CONH2
n
Me
HN
HO
N
NHPbf
NH
N
O
HN
Phe-Arg-Ala
HN
O
n
n = 1 (Ozt) or 2 (Opr)
1
00
90
80
70
O
O
H
N
HN
O
n = 1, Ozt
n = 2, Opr
Key features
O
Opr
Ozt
6
5
4
3
0
0
0
0
Lower operational concentration
Optimum temperature
Preferred additives
10–15 mM
40–60 °C
100 μM–5 mM
20–25 °C
None
Oxalic acid
Homoserine
Formed residue
Serine
20
10
0
O
O
H
N
O
Primary product
0
1
2
3
4
5
6
7
O
NH2
O
OH
Time (h)
Figure 2 | A comparison between KAHA ligations with oxazetidine- and oxaproline-containing peptides. The plot shows relative conversion to the product
at 20 mM, 40 °C and in the presence of oxalic acid (0.1 M), which are the optimal conditions for 5-Opr ligation.
alkylation of hydroxamic acid 10 with (2S)-(–)-glycidyl tosylate 100 times faster. Treatment of the data shown in Fig. 2 gives an
−
1 −1
followed by cyclization with BF ·OEt . Remarkably—and fortunately— approximate second-order rate constant for Ozt of ∼0.1 M s ,
3
2
−
1 −1
the cyclization proceeds via an S 2 mechanism to give the (S)-config- while for 5-Opr it is ∼0.001 M s . For comparison, the
N
−1 −1
ured oxazetidine product in 98% enantiomeric excess (e.e.). Protection second-order rate constant of NCL is ∼0.3 M s .
of the primary alcohol, oxidative 2,4-dimethoxybenzyl removal and We applied 1 to the synthesis of human S100A4 (also known as
Fmoc protection provided 15. Finally, tert-butyldimethylsilyl depro- metastasin) , a 100-residue, calcium-binding protein of the S100
tection and a careful two-step oxidation of the alcohol gave the family . S100A4 contains four cysteine residues in its primary
Fmoc-protected amino acid 1 (Fig. 1d). sequence (Cys3, Cys76, Cys81, Cys86), but these are not in strategic
2
1,22
2
3
With compound 1 in hand, we tested its stability and perform- positions for NCL and are critical to the folding and function of the
ance in KAHA ligations. We found that 1 is stable for months if protein. The required segments were prepared by SPPS according to
kept at a low temperature (–20 °C) and could be coupled under the strategy showed in Fig. 3. Three mutations were into the syn-
standard solid phase peptide synthesis (SPPS) conditions using thetic protein (Thr39Ser, Asn68Gly, Ser80Hse). The Ser80Hse
common coupling reagents. Interestingly, MALDI analysis of mutation arose from early attempts that were plagued by poor solu-
peptides containing the Fmoc-Ozt often showed the mass of a bility of the intermediates. These difficulties were overcome by
retro [2+2] cycloreversion, a fragmentation not observed by electron taking advantage of the ester-forming ligation of 5-Opr and
spray ionization analysis. We were pleased to find that the Fmoc keeping the assembled segments as the ester until folding into a
protecting group could be removed with N,N-diethylamine more soluble form. The Asn68Gly was chosen due to large
without decomposition of the oxazetidine, although care in amounts of aspartimide formation from the neighbouring Asp
handling of the unprotected oxazetidine was needed and storage during the synthesis of segment two and the fact that most other
of the unprotected Ozt is not recommended. This is true only of S100 proteins have Gly at this residue. The Thr39Ser mutation
the unprotected oxazetidine peptides; the Fmoc protected forms allowed us to both demonstrate the ability to carry the protected
are completely stable to storage and other manipulation. When Ozt-residue through multiple ligation steps and conduct the final
the deprotected oxazetidine-peptides were exposed to α-ketoacids ligation under lower (1–5 mM) concentrations.
in aqueous DMSO solutions, they underwent rapid and chemo-
The first KAHA ligation between segment 2 α-ketoacid (17)
selective ligation. Even at low concentrations (100 μM–1 mM) and bearing the Fmoc-protected Ozt and segment 3 5-oxaproline 18
at room temperature the reaction was complete within minutes. furnished the ligated ester, which was kept in this form to
2
4,25
NMR and HPLC studies of a model peptide showed that the improve handling and solubility . Fmoc-deprotection revealed
primary product of the oxazetidine ligation is the amide—no trace N-terminal oxazetidine 20, which was ligated with two equivalents
of the ester was detected—and the ligation furnished the product of segment 1 α-ketoacid 21. As hoped, the reaction gave the
without detectable epimerization. A screening of the ligation desired product 22 at mild temperatures (22 °C) and at low con-
conditions showed that the reaction worked well in aqueous centrations (2–5 mM), which circumvented aggregation or solubi-
N-methyl pyrrolidine (NMP) or DMSO with no additional lity issues, and the ligation was complete within hours. The
additives. Under these conditions, the reaction proceeded to unprotected amines, carboxylic acids and thiols present in the
completion at room temperature within 30 min at 1 mM (Fig. 2). peptide segments did not interfere with the ligation or attack
Competition experiments between peptides bearing 5-Opr and Ozt the oxazetidine ring. The identity and purity of 22 were confirmed
under standard KAHA conditions (20 mM, 40 °C) exclusively gave by high-resolution mass-spectrometry and HPLC. Treatment of
the Ozt-ligated product, suggesting that the Ozt ligation is at least 22 with basic carbonate buffer resulted in Fmoc-deprotection
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2282
ARTICLES
a
Motif
α-helices
4–20
31–41
43–46
52–62
72–88
EF-hand 2 (high Ca2+ affinity)
70 80
Coil
Secondary structure
EF-hand 1 (low Ca2+ affinity)
β-strands
N-terminus 10
20
30
40
50
60
90
100
Residue number
Primary sequence and
Ca2+-binding sites
MACPL EKALD VMVST FHKYS GKEGD KFKLN KSELK ELLTR ELPSF LGKRT DEAAF QKLMS NLDSN RDNEV DFQEY CVFLS CIAMM CNEFF EGFPD KQPRK K
Segment 1 (2–38)
Segment 2 (39–79)
Segment 3 (80–101)
Synthetic strategy
b
Me
Opr ligation
1.0 equiv. 18)
23 mM, 9:1 NMP/H2O
(
O
Me
40-78)
O
Fmoc
Me
0.1 M oxalic acid,
(
N
60 °C
20 mM TCEP,
N
H
N
H
OH
O
Me
(40-78)
O
O
O
O
20 h, 35%
O
(81-101)
17
HN
O
N
N
H
N
H
H
OH
Deprotection
O
NH2
O
HNEt2 (2.0 equiv.)
DMSO
10 min, 43%
O
20
H
N
(81-101)
N
O
H
OH
18
Ozt ligation
(2.0 equiv. 21)
Me
Me
Me
5
mM, 9:1 DMSO/H2O
2 °C 6 h
Me
(2-37)
O
2
Me
O
Me
(40-78)
O
O
H
N
(81-101)
FmocHN
FmocHN
(2-37)
O
N
OH
N
H
N
H
N
H
N
H
OH
H
O
O
O
NH2
2
1
OH
22
O to N shift and deprotection
pH 10.5 carbonate buffer
2
2
0 mM TCEP
3 °C
Folding
mM CaCl2,
20 mM HEPES
pH 7.5 buffer)
0 mM NaCl,
0.2 mM TCEP
2
Me
Me
(
Me
(2-37)
O
Me
40-78)
O
O
2
H
H
H2N
N
(
N
(81-101)
N
N
N
H
N
H
H
H
OH
O
O
OH
OH
23
Figure 3 | Structure and synthesis of S100A4. a, Structural motifs, metal-binding sites and synthetic strategy. b, Assembly of S100A4 by the combination of
KAHA ligation with 5-Opr and Ozt. The final ligation was carried out at 5 mM and 22 °C. TCEP, tris(2-carboxyethyl)phosphine; HEPES, 4-(2-hydroxyethyl)-1-
piperazineethanesulfonic acid.
and O-to-N acyl shift to give 23, which was folded in the Received 1 March 2015; accepted 12 May 2015;
2
+
presence of Ca . The folding and metal binding was confirmed published online 22 June 2015
by circular dichroism (CD) spectroscopy (Fig. 4). Evaluation of
this synthetic S100A4 in an adipocyte proliferation assay con-
firmed its biological activity.
References
1. Kent, S. B. H. Bringing the science of proteins into the realm of organic
In conclusion, the synthesis of an enantioenriched oxazetidine
amino acid—formally an oxidized form of serine—enables the
first native amide-forming ligation reaction that compares to the
chemistry: total chemical synthesis of SEP (synthetic erythropoiesis protein).
Angew. Chem. Int. Ed. 52, 11988–11996 (2013).
Nilsson, B. L., Soellner, M. B. & Raines, R. T. Chemical synthesis of proteins.
Annu. Rev. Biophys. Biomol. Struct. 34, 91–118 (2005).
Hackenberger, C. P. R., Bode, J. W. & Schwarzer, D. in Amino Acids,
Peptides and Proteins in Organic Chemistry: Building Blocks, Catalysis
and Coupling Chemistry Vol. 3 (ed Hughes, A. B.) 445–493
2
.
.
26
venerable NCL in both reaction rate and chemoselectivity. By
operating at prevalent serine (6.8% abundance)2 , rather than
cysteine (1.8% abundance), this ligation offers greater flexibility in
synthetic planning. Serine residues can also easily replace many
other residues without affecting protein folding or function. We
have recently introduced solid supported linkers for Fmoc-SPPS
3
7,28
(
Wiley-VCH, 2010).
4
.
.
Dawson, P., Muir, T., Clark-Lewis, I. & Kent, S. Synthesis of proteins by native
chemical ligation. Science 266, 776–779 (1994).
Pattabiraman, V. R. & Bode, J. W. Rethinking amide bond synthesis. Nature
480, 471–479 (2011).
5
8
that afford peptide α-ketoacids directly upon resin cleavage . As
these resins and the Fmoc-(S)-Ozt monomer become widely avail- 6. Bode, J. W., Fox, R. M. & Baucom, K. D. Chemoselective amide ligations by
decarboxylative condensations of N-alkylhydroxylamines and α-ketoacids.
able, we expect that this variant of the KAHA ligation will emerge as
a leading approach to chemical protein synthesis due to its ease of
execution, traceless ligation of the peptide segments and facile
implementation of multiple segment couplings.
Angew. Chem. Int. Ed. 45, 1248–1252 (2006).
7
.
Pattabiraman, V. R., Ogunkoya, A. O. & Bode, J. W. Chemical protein synthesis
by chemoselective α-ketoacid-hydroxylamine (KAHA) ligations with
5-oxaproline. Angew. Chem. Int. Ed. 51, 5114–5118 (2012).
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
©
2015 Macmillan Publishers Limited. All rights reserved

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2282
ARTICLES
a
21
b
23
2
0
2
2
0
.5 h
1
8 19 20 21 22 23 24 25 26 27 28 29 30
12
14
16
18
Time (min)
20
22
24
c
20
CD spectrum of synthetic S100A4
2
2
1
5
0
5
0
5
3
h
1
1
8 19 20 21 22 23 24 25 26 27 28 29 30
2
2
–
6
h
–10
–15
–20
1
1
8
8
19 20 21 22 23 24 25 26 27 28 29 30
2
2
Purified
product
200
210
220
230
240
250
260
270
280
19 20 21 22 23 24 25 26 27 28 29 30
Time (min)
Wavelength (nm)
Figure 4 | Monitoring of the coupling and folding processes in the synthesis of S100A4. a, Oxezetadine ligation of 20 and 21. To facilitate monitoring, the
Fmoc group was retained on the N-terminus of 21, which displays greatly increased absorbance compared to 20. b, HPLC of final coupled and purified
product 23. c, CD spectrum of folded synthetic protein S100A4.
8
.
Wucherpfennig, T. G., Pattabiraman, V. R., Limberg, F. R. P., Ruiz-Rodríguez, J.
Bode, J. W. Traceless preparation of C-terminal α-ketoacids for chemical
protein synthesis by α-ketoacid–hydroxylamine ligation: synthesis of SUMO2/3.
Angew. Chem. Int. Ed. 53, 12248–12252 (2014).
Wucherpfennig, T. G., Rohrbacher, F., Pattabiraman, V. R. & Bode, J. W.
Formation and rearrangement of homoserine depsipeptides and depsiproteins
in the α-ketoacid–hydroxylamine ligation with 5-oxaproline. Angew. Chem Int.
Ed. 53, 12244–12247 (2014).
22. Kiss, B. et al. Crystal structure of the S100A4-nonmuscle myosin IIA tail
fragment complex reveals an asymmetric target binding mechanism. Proc. Natl
Acad. Sci. USA 109, 6048–6053 (2012).
23. Schäfer, B. W. & Heizmann, C. W. The S100 family of EF-hand calcium-binding
proteins: functions and pathology. Trends Biochem. Sci. 21, 134–140 (1996).
24. Sohma, Y., Sasaki, M., Hayashi, Y., Kimura, T. & Kiso, Y. Novel and efficient
synthesis of difficult sequence-containing peptides through O–N intramolecular
acyl migration reaction of O-acyl E isopeptides. Chem. Commun.
124–125 (2004).
25. Avital-Shmilovici, M. et al. Fully convergent chemical synthesis of ester insulin:
determination of the high resolution X-ray structure by racemic protein
crystallography. J. Am. Chem. Soc. 135, 3173–3185 (2013).
26. Kalia, J. & Raines, R. T. Advances in bioconjugation. Curr. Org. Chem. 14,
138–147 (2010).
&
9
.
1
0. Magers, D. H. & Davis, S. R. Ring strain in the oxazetidines. J. Mol. Struc.
Theochem. 487, 205–210 (1999).
1. Florio, S., Capriati, V., & Luisi, R. in Comprehensive Heterocyclic Chemistry III
Vol. 2 (eds Katritzky, A. R., Ramsden, C. A., Scriven, E. F. V. & Taylor, R. J. K.)
Ch. 14, 689–711 (Elsevier, 2008).
2. Snider, B. B. & Duvall, J. R. Synthesis of the 4-methyl-1,2-oxazetidine-4-
carboxylic acid moiety of the originally proposed halipeptin A and B structures.
Tetrahedron Lett. 44, 3067–3070 (2003).
3. Denicola, A., Einhorn, C., Einhorn, J. & Luche, J. L. Intramolecular alkylation of
carboxylic-acids—application to the synthesis of Boc-protected cyclic amino-
acids. J. Chem. Soc. Chem. Commun. 879–880 (1994).
4. Casadei, M. A., Galli, C. & Mandolini, L. Ring-closure reactions. 22. Kinetics of
cyclization of diethyl (ω-bromoalkyl)malonates in the range of 4- to 21-
membered rings. Role of ring strain. J. Am. Chem. Soc. 106, 1051–1056 (1984).
5. Baldwin, J. E. Rules for ring closure. Chem. Commun. 734–736 (1976).
6. Alabugin, I. V. & Gilmore, K. Finding the right path: Baldwin ‘rules for ring
closure’ and stereoelectronic control of cyclizations. Chem. Commun. 49,
1
1
1
1
27. Li, X., Lam, H. Y., Zhang, Y. & Chan, C. K. Salicylaldehyde ester-induced
chemoselective peptide ligations: enabling generation of natural peptidic
linkages at the serine/threonine sites. Org. Lett. 12, 1724–1727 (2010).
28. Zhang, Y., Xu, C., Lam, H. Y., Lee, C. L., & Li, X. Protein chemical synthesis by
serine and threonine ligation. Proc. Natl Acad. Sci. USA 110, 6657–6662 (2012).
Acknowledgements
This work was supported by the Swiss National Science Foundation (200020_150073) and
ETH Zürich. F. Thuaud, S. Baldauf and M. Dao are thanked for contributions to the
synthesis of compound 1, V. Pattabiraman for discussions, F. Saito for advice on kinetics
and C. Wolfrum for biological evaluation.
1
1
1
124–11250 (2013).
7. Tanino, K. et al. Total synthesis of Solanoeclepin A. Nature Chem. 3,
84–488 (2011).
1
1
Author contributions
4
J.W.B. and I.P. contributed equally to the design of the study. J.B., with contributions from
I.P., wrote the paper. I.P performed the experiments and wrote the Supplementary
Information.
8. Medjahdi, M., Gonzalez-Gomez, J. C., Foubelo, F. & Yus, M. Stereoselective
synthesis of azetidines and pyrrolidines from N-tert-butylsulfonyl(2-
aminoalkyl)oxiranes. J. Org. Chem. 74, 7859–7865 (2009).
9. Davies, S. G., Jones, S., Sanz, M. A., Teixeira, F. C. & Fox, J. F. A novel [2,3]
intramolecular rearrangement of N-benzyl-O-allylhydroxylamines. Chem.
Commun. 2235–2236 (1998).
1
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints. Correspondence and
requests for materials should be addressed to J.W.B.
2
0. Wuts, P. G. M. & Greene, T. W. in Greene’s Protective Groups in Organic
Synthesis (Wiley, 2006).
2
1. Vallely, K. M. et al. Solution structure of human Mts1 (S100A4) as determined Competing financial interests
by NMR spectroscopy. Biochemistry 41, 12670–12680 (2002).
The authors declare no competing financial interests.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
5
©
2015 Macmillan Publishers Limited. All rights reserved