Product Selectivity in KAHA Ligations: Ester vsAmide Formation with Cyclic Hydroxylamines

Article abstract of DOI:10.1055/s-0036-1588480

Cyclic hydroxylamines form esters instead of the expected amides as major product upon reaction with α-ketoacids. In this report, we document a systematic investigation into the effect of the hydroxylamine structure and the solvent mixture on the product

Full text of DOI:10.1055/s-0036-1588480

SYNLETT
0
9
3
6
-
5
2
1
4
1
4
3
7
-
2
0
9
6
© Georg Thieme Verlag Stuttgart · New York
2017, 28, A–E
cluster
A
en
F. Rohrbacher et al.
Cluster
Synlett
Product Selectivity in KAHA Ligations: Ester vs. Amide Formation
with Cyclic Hydroxylamines
Florian Rohrbachera
Simon Baldaufa
Thomas G. Wucherpfenniga
Jeffrey W. Bode*a,b
a Laboratorium für Organische Chemie, Department of
Chemistry and Applied Biosciences, ETH Zürich, 8093
Zürich, Switzerland
bode@org.chem.ethz.ch
b Institute of Transformative Bio-Molecules (WPI–ITbM),
Nagoya University, Chikusa, Nagoya 464-8602, Japan
Published as part of the Cluster Recent Advances in Protein
and Peptide Synthesis
Received: 30.04.2017
Accepted after revision: 04.06.2017
Published online: 19.07.2017
O
O
H
N
R2
OH
N
H
R1
O
DOI: 10.1055/s-0036-1588480; Art ID: st-2017-w0309-c
O
1
2
Abstract Cyclic hydroxylamines form esters instead of the expected
amides as major product upon reaction with α-ketoacids. In this report,
we document a systematic investigation into the effect of the hydroxyl-
amine structure and the solvent mixture on the product ratio of amides
vs. ester in the KAHA ligation. We show that the ratio is almost exclu-
sively determined by the structure of the hydroxylamine, with only mi-
nor contributions from the reaction solvent or the structure of the α-
ketoacid.
O
HO
CO2H
R1
R1
O
O
– H2O
O
R2
N
R2
N
N
O
N
H
O
H
3
4
– CO2 + H+
Key words ligation, mechanism, hydroxylamines, α-ketoacids, esters,
peptides
O
O
H2
N
HO
H
R1
R2
R2
R1
N
O
R
C
+ H2O
R2
N
N
N
H
H
In 2006, we reported the α-ketoacid-hydroxylamine
(KAHA) amide-forming ligation,1 which complements the
well-established native chemical ligation2 (NCL) for the
chemical synthesis of proteins by chemoselective couplings
of unprotected peptide segments.3
N
H
O
O
HO
7
6
5
+ H2O – H+
O
The most widely used implementation of the KAHA li-
gations uses a cyclic hydroxylamine – 5-oxaproline – as sta-
ble and chemoselective ligation partner that forms a homo-
serine residue upon ligation with α-ketoacids.4 The primary
product of this reaction is ester 8 (Scheme 1), which is rear-
ranged to amide 9 under basic conditions.5 In contrast, the
oxazetidine analogue of 5-oxaproline (four- vs. five-mem-
bered ring) exclusively yields the amide product.6
R2
O
H3N
H
N
N
R2
R1
H
basic
aqueous buffer
R1
O
N
H
OH
O
O
amide-9
ester-8
Scheme 1 Current understanding of the mechanism
We have previously demonstrated that the structure of
the α-ketoacid only has a minor influence on the ester to
amide product ratio (Table 1).5 We have not, however, re-
ported a systematic investigation into the effect of the hy-
droxylamine structure on the formation of amides vs. esters
in the KAHA ligation. In this report, we document compre-
hensive studies on the effect of hydroxylamine structure on
the product distribution and consider the implications of
these results for the understanding of the mechanism of the
KAHA ligation.
In a preliminary study we could isolate iminoether 6 as
an intermediate of the reaction leading to the ester product
and hypothesized that nitrilium 5 could be a reactive inter-
mediate leading to the formation of iminoether 6.7
© Georg Thieme Verlag Stuttgart · New York — Synlett 2017, 28, A–E

B
F. Rohrbacher et al.
Cluster
Synlett
As show in Scheme 2, we expected the main influences
on the product selectivity to be: 1) solvents and 2) the
structure of the hydroxylamine (variation of n and R2). We
therefore carried out ligations with small model peptides.
Table 2 Influence of the Solvent on Product Selectivity
O
OH
R1
O
OH
O
O
O
10
0.1 M oxalic acid
2
O
NHR2
R1
O
NHR2
2
R1
N
OH
solvent, 60 °C, 12 h
O
R1
O
O
H
NH2
ester-12
R2
O
KAHA ligation
solvent
R2
O
NHR2
R1
N
H
OH
R1
O
n = 0–N2 H2
amide-12
n = 0–2
R2
R1 = Trp-Thr-Leu
R2 = 4-F-Phe
NH
O
O
amide product
ester product
HN
11
n = 0–2
Scheme 2 Evaluation of various hydroxylamines to assess product se-
lectivity of ester vs. amide product
Entry
Solvent
Ester (%)a
Amide (%)a
1
2
3
4
5
6
7:3 DMSO/H2O
7:3 NMP/H2O
99:1 DMSO/H2O
99:1 NMP/H2O
99:1 CH3CN/H2O
H2O
89
87
97
90
95
73
11
13
3
We assessed the influence of solvents by varying the ra-
tio of water/organic solvent and changing the organic sol-
vent itself. Restricting the organic solvents to those capable
of solubilizing proteins we chose dimethylsulfoxide (DMSO),
N-methylpyrrolidone (NMP), and MeCN. The product ratio
of the reaction between tripeptide α-ketoacid 10 and 5-ox-
aproline-containing dipeptide 11 was determined by ana-
lytical HPLC. As shown in Table 2 most amide product am-
ide-12 was formed in pure water (Table 2, entry 6). Under
our standard conditions (7:3 organic solvent/H2O; Table 2,
entries 1 and 2) a ratio of approximately 9:1 is observed in-
dependent on the organic solvent. Using almost pure or-
ganic solvents (99:1; Table 2, entries 3, 4, and 5) the ratio of
the ester increased up to 20:1 in the case of DMSO and
MeCN and remained at 9:1 in the case of NMP.
10
5
27
a Determined by analytical HPLC of the reaction mixture at 220 nm.
Various types of hydroxylamines undergo KAHA liga-
tion with α-ketoacids. O-Unsubstituted hydroxylamines –
which we have previously classified as type I hydroxyl-
amines – appear to react by a different mechanism and are
not part of this study.8 Substituted hydroxylamines – which
we have previously termed type II hydroxylamines – pro-
ceed via a different mechanism. We have reported two ma-
jor types of substituted hydroxylamines: 1) O-acyl and O-
carbamoyl hydroxylamines and 2) O-alkyl hydroxylamines.
Linear O-alkyl hydroxylamines react with α-ketoacids only
very slowly and under forcing conditions. In contrast,
strained, cyclic hydroxylamines are excellent reaction part-
ners in the KAHA ligation, although in the case of 5-oxapro-
line the ester is observed as the major product.
We synthesized cyclic hydroxylamines (Table 3) to
probe their influence on the product selectivity. As previ-
ously reported, four-membered 1,2-oxazetidines exclusive-
ly yield the amide product at a much faster reaction rate
(Table 3, entry 1).6 5-Oxaproline (Table 3, entry 2) affords a
mixture of ester and amide product in a ratio of 9:1, while
the regioisomer of 5-oxaproline (Table 3, entry 3) exclusive-
ly yields the ester product. 5-Oxaprolinol (Table 3, entry 4)
showed essentially the same product distribution of 9:1 as
5-oxaproline. Unsubstituted isoxazolidine (Table 3, entry 5)
and 1,2-oxazinane (Table 3, entry 6) showed a slight in-
crease in amide product (8:2 and 6:4) with increasing ring
size, along with a decrease in the reaction rate. The seven-
membered 1,2,5-oxadiazepanes (Table 3, entries 7 and 8)
afforded a 7:3 mixture of the ester and amide products, re-
gardless whether the transannular amine is protected or
not. The linear analogue of isoxazolidine (Table 3, entry 9)
exclusively yields the amide product.
Table 1 Influence of the α-Ketoacid on Product Selectivity
O
OH
R1
OH
O
O
O
2
0.1 M oxalic acid
8:2 DMSO/H2O
O
O
NHR2
R1
O
NHR2
R1
2
N
H
NH2
ester
60 °C, overnight
O
amide
R2 = 4-F-Phe
NHR2
NH
O
Entry R1
Ratio ester/amidea Isolated yield of ester (%)
Me
Me
1
93:7
61
Fmoc
N
H
HO2C
2
3
4
93:7
59
59
55
86:14
87:13
Me
Ph
a Determined by analytical HPLC of the reaction mixture at 220 nm.
© Georg Thieme Verlag Stuttgart · New York — Synlett 2017, 28, A–E

C
F. Rohrbacher et al.
Cluster
Synlett
Table 3 Influence of the Hydroxylamine on Product Selectivity
R2
O
O
O
R2
0.1 M oxalic acid
R2
OH
R1
R1
O
NH2
R1
N
H
OH
NH
DMSO/H2O
60 °C
O
O
ester
amide
Entry
1b
R1
Hydroxylamine
Ratio ester/amidea
<1:99c
Isolated yield of ester (%)
Me
O
Me
ester: 0
amide: 69
NHR
R
O
N
H
NH
R = Gly-Phe-Gly
R = Fmoc-Ala-
O
Me
Me
NHR
2d
3e
4g
93:7
>99:1f
92:8
61
40
43
NH
O
Fmoc
N
R = 4-F-Phe
H
O
OH
HO
O
HN
Me
Me
OH
NH
O
Fmoc
N
H
5h
6h
79:21i
58:42i
27
14
NH
O
NH
O
H
N
O2N
7j
8j
71:29
73:27
25
27
O
NH
Cbz
O2N
N
O
NH
Me
Me
9k
<1:99c
79
HO
R
N
H
NH
OBz
R = Trp-Thr-
a Determined by analytical HPLC of the reaction mixture at 220 nm.
b Previously reported (see Supporting Information of ref. 6), in 10:1 DMSO/H2O at r.t. for 30 min.
c Only the amide product was observed.
d Previously reported (see Supporting Information of ref. 5), in 8:2 DMSO/H2O overnight.
e In 8:2 DMSO/H2O for 5 h, racemic hydroxylamine was used.
f Only the ester product was observed.
g In 8:2 DMSO/H2O for 2 h.
h In 8:2 DMSO/H2O for 14 h.
i Based on isolated yields after purification by preparative HPLC.
j In 9:1 DMSO/H2O for 12 h.
k In 7:3 DMSO/H2O for 16 h.
© Georg Thieme Verlag Stuttgart · New York — Synlett 2017, 28, A–E

D
F. Rohrbacher et al.
Cluster
Synlett
The remarkable difference in product distribution of li-
gations with isoxazolidine and its linear analogue is illus-
trated in Scheme 3 (A and B). Even though 18 contains a hy-
droxyl group that would allow a similar 6-endo-dig attack
as observed for KAHA ligations with isoxazolidine (14), it
exclusively yields the amide product. When 18 is used in
KAHA ligations under anhydrous conditions (Scheme 3, C)
we observed both amide and ester products in a ratio of
8:2.
contributions from the reaction solvent or the structure of
the α-ketoacid. These experiments provide further data
supporting a nitrilium intermediate in type II KAHA liga-
tions and suggest that ester formation is preferred in cyclic
hydroxylamines due to the formation of an ion pair with fa-
vorable geometry for intramolecular cyclization. We believe
that these insights will support the further improvement of
the KAHA ligation and its application in chemical protein
synthesis.
These results could be most easily be explained by the
formation of the ion pair 15 upon decarboxylation in the
case of cyclic hydroxylamines. This ion pair could readily
undergo 6-endo-dig cyclization to form the iminoether,
leading ultimately to the ester product. DFT calculations
implicate the formation of similar ion pairs for Ugi reac-
tions.9 Since this ion pair is unlikely to be formed with the
(protonated) hydroxyl group of linear hydroxylamine 18, 6-
endo-dig cyclization is slower and is only observed under
strictly anhydrous conditions.
Funding Information
This work was supported by the Swiss National Science Foundation
(150073, 169451)
Swsi
N
oaitn
a
l
Secince
F
o
u
n
d
oaitn
1(
5
0
0
7
3
S)wsi
N
oaitn
a
l
Secince
F
o
u
n
d
oaitn
1(
6
9
4
5
1)
Acknowledgment
We thank the MS and NMR service of the Laboratorium für Organ-
ische Chemie at ETH Zürich for analyses.
In summary, we have shown that cyclic hydroxylamines
– with the notable exception of four-membered 1,2-oxaz-
etidines – afford the ester product in significant amounts.10
The ester-to-amide ratio is almost exclusively determined
by the structure of the hydroxylamine, with only minor
Supporting Information
Supporting information for this article is available online at
https://doi.org/10.1055/s-0036-1588480.
S
u
p
p
ortiInfogrmoaitn
S
u
p
p
o
nrtogI
f
rmoaitn
(A)
O
N
O
N
0.1 M oxalic acid
Ph
6-endo-dig
Ph
OH
Ph
Ph
O
NH
O
8:2 DMSO/H2O
60 °C
O
NH2
O
O
13
14
15
nitrilium ion pair
16
ester-17
79%a
O
Ph
nitrilium trapped by water
N
H
OH
amide-17
21%a
(B)
H
H
O
O
O
N
H
nitrilium trapped by water
0.1 M oxalic acid
OH
N
OH
R1
R
R
N
H
OH
BzO
O
7:3 DMSO/H2O
60 °C, 12 h
O
10
1 = Trp-Thr-Leu-
H
amide-20
18
19
exclusive product
R
H
H
(C)
O
O
O
N
N
6-endo-dig
0.1 M oxalic acid
H
R
O
NH2
OH
R
N
OH
R
R
O
O
BzO
DMSO (anhydrous)
ester-24
17%a
O
21
60 °C, 12 h
H
23
18
22
R
1 = 4-O2NC6H4(CH2)2-
O
nitrilium trapped by water
R
N
OH
H
amide-24
83%a
Scheme 3 KAHA ligations with (A) isoxazolidine, (B) linear hydroxylamine 18, and (C) linear hydroxylamine 18 under anhydrous conditions. a Ester-to-
amide ratio.
© Georg Thieme Verlag Stuttgart · New York — Synlett 2017, 28, A–E

E
F. Rohrbacher et al.
Cluster
Synlett
References and Notes
reaction was directly purified by preparative HPLC (Shiseido
Capcell Pak MGII C18 column, 20 × 250 mm, r.t., 40–90% MeCN
with 0.1% TFA in 20 min, flow rate 10 mL/min).
Typical Analytical Data
N-(3-Hydroxypropyl)-3-(4-nitrophenyl)propanamide
(Amide-24)
(1) Bode, J. W.; Fox, R. M.; Baucom, K. D. Angew. Chem. Int. Ed. 2006,
45, 1248.
(2) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. Science
1994, 266, 776.
1H NMR (600 MHz, CDCl3): δ = 8.14–8.10 (m, 2 H, CH), 7.38–
7.34 (m, 2 H, CH), 6.07 (br s, 1 H, NH), 3.54 (t, J = 5.6 Hz, 2H,
CH2), 3.37 (q, J = 6.1 Hz, 2 H, CH2), 3.13 (br s, 1 H, OH), 3.07 (t,
J = 7.5 Hz, 2 H, CH2), 2.53 (t, J = 7.5 Hz, 2 H, CH2), 1.66–1.59 (m, 2
H, CH2). 13C NMR (150 MHz, CDCl3): δ = 172.4 (CO), 148.8 (C),
146.7 (CNO2), 129.4 (CH), 123.9 (CH), 59.6 (CH2), 37.5 (CH2),
36.7 (CH2), 32.1 (CH2), 31.4 (CH2). ESI-HRMS: m/z calcd for
(3) Other amide-forming ligations include the serine/threonine
ligation (a) and the traceless Staudinger ligation (b): (a) Zhang,
Y.; Xu, C.; Lam, H. Y.; Lee, C. L.; Li, X. Proc. Natl. Acad. Sci. U.S.A.
2013, 110, 6657. (b) Kleineweischede, R.; Hackenberger, C. P. R.
Angew. Chem. Int. Ed. 2008, 47, 5984.
(4) Pattabiraman, V. R.; Ogunkoya, A. O.; Bode, J. W. Angew. Chem.
Int. Ed. 2012, 51, 5114.
(5) Wucherpfennig, T. G.; Rohrbacher, F.; Pattabiraman, V. R.; Bode,
J. W. Angew. Chem. Int. Ed. 2014, 53, 12244.
(6) Pusterla, I.; Bode, J. W. Nat. Chem. 2015, 7, 668.
(7) Rohrbacher, F.; Deniau, G.; Luther, A.; Bode, J. W. Chem. Sci.
2015, 6, 4889.
(8) Pusterla, I.; Bode, J. W. Angew. Chem. Int. Ed. 2012, 51, 513.
(9) Chéron, N.; Ramozzi, R.; Kaïm, L. E.; Grimaud, L.; Fleurat-Les-
sard, P. J. Org. Chem. 2012, 77, 1361.
C
12H17N2O4 [M + H]+: 253.1183; found: 253.1183.
3-{[3-(4-Nitrophenyl)propanoyl]oxy}propan-1-aminium
2,2,2-trifluoroacetate (Ester-24)
1H NMR (600 MHz, (D3C)2SO): δ = 8.19–8.14 (m, 2 H, CH), 7.73
(br s, 3 H, NH3+), 7.56–7.51 (m, 2 H, CH), 4.06 (t, J = 6.3 Hz, 2 H,
CH2), 3.00 (t, J = 7.5 Hz, 2 H, CH2), 2.82 (t, J = 7.4 Hz, 2 H, CH2),
2.73 (t, J = 7.5 Hz, 2 H, CH2), 1.88–1.80 (m, 2 H, CH2). 13C NMR
(150 MHz, (D3C)2SO): δ = 171.9 (CO), 157.9 (q, J = 32.0 Hz,
CF3COO–), 148.9 (C), 146.1 (CNO2), 129.7 (CH), 123.5 (CH),
117.36 (q, J = 301 Hz, CF3), 61.3 (CH2), 36.2 (CH2), 34.1 (CH2),
29.9 (CH2), 26.3 (CH2). ESI-HRMS: m/z calcd for C12H17N2O4 [M +
H]+: 253.1183; found: 253.1186.
(10) General Procedure for KAHA Ligations
Ketoacid (1.00–2.00 equiv) and hydroxylamine (1.00–2.00
equiv) were dissolved in 8:2 DMSO/H2O (20–100 mM) and
heated to 60 °C for 12 h. Reaction progress was monitored by
analytical HPLC (Shiseido Capcell Pak UG80 C18 column (4.6 ×
250 mm), heated to 60 °C, 30–70% MeCN in 20 min). The crude
© Georg Thieme Verlag Stuttgart · New York — Synlett 2017, 28, A–E