Controlled thioamide vs. amide formation in the thioacid-azide reaction under acidic aqueous conditions

Article abstract of DOI:10.1039/c4cc00774c

The thioacid-azide reaction and its chemoselectivity were probed with alkyl azides for a potential application to form amide bonds in aqueous solvents. Our results reveal that under acidic conditions thioamides were formed as major reaction products suggesting a competing mechanism, whereas reactions forming amides predominated at slightly higher pH values. the Partner Organisations 2014.

Full text of DOI:10.1039/c4cc00774c

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
Controlled thioamide vs. amide formation in the
thioacid–azide reaction under acidic aqueous
conditions†
Cite this: Chem. Commun., 2014,
50, 4603
Received 28th January 2014,
Accepted 12th March 2014
Michaela Muhlberg,ab Kristina D. Siebertz,bc Brigitte Schlegel,b Peter Schmiederb
¨
and Christian P. R. Hackenberger*bc
DOI: 10.1039/c4cc00774c
www.rsc.org/chemcomm
The thioacid–azide reaction and its chemoselectivity were probed
with alkyl azides for a potential application to form amide bonds in
aqueous solvents. Our results reveal that under acidic conditions
thioamides were formed as major reaction products suggesting a
competing mechanism, whereas reactions forming amides pre-
dominated at slightly higher pH values.
Scheme 1 Thioacid–azide reactions with electron-deficient azides (left)
and electron-rich azides (right).
Despite many years of extensive research, amide bond formation
still remains a challenging research field. A vast variety of synthetic Several electron-deficient azides have been applied in peptide
methods have been developed throughout the last few decades1 glycosylations,12 seleno amidations,13 the synthesis of resin-bound
ranging from standard activation and dehydration strategies N-peptidyl sulfonamides,14 short peptide ligations15 and in kinetic
between carboxylic acids and amines over radical-based to oxidative target-guided synthesis.16 In addition, researchers have shown that
methods. Other recent approaches include an umpolung strategy the amidation reaction with sulfonazides proceeds rapidly and
between a-bromonitroalkanes and amines2 or the reaction between chemoselectively.17 In contrast, considerably fewer studies have
acyltrifluoroborates and hydroxylamines.3 In addition, several chemo- addressed the performance of less electrophilic alkyl azides for
selective ligation strategies, such as the native chemical ligation the reaction with thioacids (Scheme 1, right).11a,18 Several attempts
(NCL),4 the traceless Staudinger ligation (TSL)5 and the ketoacid– have been made to improve the thioacid–azide reaction probing
hydroxylamine ligation (KAHA)6 have been employed to yield the different reaction temperatures,11 solvent systems11,15a and catalysts
desired amidation product even between polypeptides of synthetic such as RuCl3.18 Despite the apparent lower reactivity of alkyl azides,
and/or molecular biological origin.7 However, several limitations, this reaction is particularly promising for the formation of peptide
such as solubility and hydrolytic instability8 (TSL) and the necessity bonds, which could be applied as a new peptide ligation strategy
of cysteine (NCL and EPL) or homoserine (KAHA) at the ligation site with orthogonal functional groups to NCL or KAHA. In addition,
provoke a demand for further optimization or engineering of reliable this reaction could also lead to site-specific post-translational acyl-
chemoselective amidation reactions.
Another promising transformation to yield amide bonds is the nature as fatty-acylations, acetylations or ubiquitinations.19
thioacid–azide reaction,9 which is also known as the ‘‘Sulfo-Click’’
At the outset of our studies, we intended to probe the reactivity
ations on polypeptides at former azido-Lys residues, which occur in
reaction (Scheme 1, left).10 Herein, an electron-deficient sulfona- and chemoselectivity of thioacetic acid (2) in acetylation reactions
zide reacts with a thioacid through a thiatriazoline intermediate, on the unprotected and electron-rich e-azido lysine peptide 1
as postulated by Williams, which then decomposes with the (Table 1). In the ‘‘Sulfo-Click’’ reaction, best results were often
release of sulfur and nitrogen to yield an N-acyl sulfonamide.11 achieved in DMF at room temperature in the presence of lutidine
as a base.11 We thereby employed the previously reported condi-
a
tions to convert peptide 1, whose amino acid sequence is derived
from the histone protein H420 and is rich in the basic amino acids
arginine and lysine. For better conversion analysis by HPLC, a
fluorescent e-nitrobenzoxadiazole (NBD) lysine was incorporated
into the peptide sequence by standard SPPS (Table 1).
¨
¨
Freie Universitat Berlin, Institut fur Chemie und Biochemie, Takustr. 3,
14195 Berlin, Germany
b
c
¨
Forschungsinstitut fur Molekulare Pharmakologie (FMP), Robert-Roessle-Str. 10,
13125 Berlin, Germany. E-mail: hackenbe@fmp-berlin.de; Fax: +49 (0)30 947931
¨
¨
Humboldt Universitat zu Berlin, Institut fur Organische und Bioorganische Chemie,
¨
Institut fur Chemie, Brook-Taylor-Str. 2, 12489 Berlin, Germany
Our initial attempt to perform the amidation reaction in organic
solvents under basic conditions revealed that the overall conversion
† Electronic supplementary information (ESI) available: Procedures and spectral
data. See DOI: 10.1039/c4cc00774c
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 4603--4606 | 4603

View Article Online
Communication
ChemComm
Table 1 Conditions for the reaction of thioacetic acid (2) with e-azido lysine peptide 1 (NBD = e-nitrobenzoxadiazole, Nle(6-N3) = 6-azido norleucine)
Ratiob [%]
Entry
c [mM]
TAAa [eq.]
pH
Solvent system
1
3
6
4
1
2
3
4
5
6
3.00
1.75
1.75
1.75
3.00
3.00
5
5
5
5
20
100
—
DMF, 3 mM lutidine
Phosphate buffer (100 mM)
Citrate buffer (100 mM)
Citrate buffer (100 mM)
Sodium citrate/HCl buffer (100 mM)
Sodium citrate/HCl buffer (100 mM)
—
38
60
70
8
2
6
6
9
25
6
—
15
10
5
53
15
98
38
24
16
14
79
7.0
3.0
2.0
2.0
2.0
—
a
b
TAA = thioacetic acid (2). Determined by LC–Fluorescence peak integration (ex460/em540 nm); detection by HPLC-MS.
of alkyl azides was very low at room temperature and heating to 2 showed only very little dithioacetic acid (o0.5%), which would not
40 1C was necessary. Additionally, there was hardly any formation be enough to explain thioamide formation in entry 5.
of the desired specifically acetylated Lys-peptide 3 and almost full
These results led to the assumption that, at lower pH, the
conversion to non-specifically acetylated peptides 4 by a presum- reaction might at least partially proceed via a different mechanism
ably faster acylation of amines with thioacids (Table 1, entry 1). This as the one postulated by Williams in 2006. Based on DFT calcula-
side reaction has also been described by Kent, who discovered that tions, Williams proposed the concerted formation of thiatriazolidine
Cys-[peptide]-thiocarboxylates can undergo direct intramolecular 8 starting from a CQS bond for the reaction of methyl azide 7 with
cyclization at neutral pH.21 When performing the reaction in a thioacetic acid (2) (Scheme 2A).11b At lower pH, thioacetic acid (2)
phosphate buffer at neutral pH at room temperature, we observed a exists in its neutral form to a greater extent with a CQO bond and
crucial decrease of non-specific amidation and a slight increase in contains only traces of the CQS tautomer as demonstrated theore-
product formation, though conversion of azide 1 was still incom- tically and experimentally by Liu and Gordy, respectively.23 Hence,
plete (Table 1, entry 2).
we propose that the concerted cyclization could also include the
To further suppress non-selective acetylations and lower the CQO bond and therefore yield an oxatriazolidine (Scheme 2B),
nucleophilicity of side chains as well as the N-terminus, the reaction which then delivers the corresponding thioamide (Scheme 2B) as
was probed at lower pH (Table 1, entries 3–6). Initial results showed previously proposed by the group of Rademann for the reaction of
that conversion was very low (Table 1, entries 3 and 4) and the reduced azides and thioacids in the presence of Lewis acids.24 Due to the
amine peptide 5 was observed as a small side-product. We decided to larger covalent radius of sulfur versus oxygen and the less efficient
increase the concentration from 1.75 to 3.00 mM and to add more overlap of the C2p–S3p p-bond, one might expect the CQS bond to
equivalents of thioacetic acid (2). After reaction for 2 d at 40 1C and pH be more reactive than the CQO bond in the cycloaddition step.25
2 with 20 eq. of thioacetic acid (2), the reaction led to almost full
conversion of azido peptide 1 and furnished the desired amidation
product 3 in 25% conversion (Table 1, entry 5). Although reducing the
pH lowered non-selective acetylations to 14%, it could not be
completely prevented, showing that in the presence of nucleophilic
amino acids like lysine the reaction with electron-rich alkyl azides
does not proceed in a chemoselective fashion. A higher amount of
thioacetic acid (2) only led to an increase in undesired amidation of
the starting material and the formed products (Table 1, entry 6).
In addition to the desired acetamide 3, we observed formation of
the thioamide-containing peptide 6. The large amount of thioamide 6
observed at pH 2 indicated that the reaction towards thioamide 6 was
faster than the formation of acetamide 3 (Table 1, entry 5). Initially,
we assumed that thioamide formation might occur due to traces of
dithioacetic acid.22 However, 1H and 13C NMR analysis of commercial acidic conditions by (A) Williams11b and (B) our group.
Scheme 2 Proposed mechanisms for the thioacid–azide reaction under
4604 | Chem. Commun., 2014, 50, 4603--4606
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
Table 2 Conditions for thioacid–azide reactions with 9 and 10
Entry
n
c [mM]
Solvent system
Ratio 11 : 12a
Conversiona [%]
1
2
3
4
5
6
7
8
1
3
1
3
1
3
1
3
1
3
1
3
3
3
3
3
3
3
3
3
3
3
3
3
3
20
KCl/HCl buffer (100 mM, pH 2.0)
KCl/HCl buffer (100 mM, pH 2.0)
NH4OAc buffer (100 mM, pH 4.0)
NH4OAc buffer (100 mM, pH 4.0)
Phosphate buffer (100 mM, pH 7.0)
Phosphate buffer (100 mM, pH 7.0)
NH4OAc buffer (100 mM, pH 4.0), 30% DMF (pH 5)
NH4OAc buffer (100 mM, pH 4.0), 30% DMF (pH 5)
Phosphate buffer (100 mM, pH 9.0), 50% THF
Phosphate buffer (100 mM, pH 9.0), 50% THF
DMF (dry, 3 mM lutidine)
0.08 : 0.92
0.32 : 0.68
0.62 : 0.38
0.47 : 0.53
1 : 0
37
47
64
49
53
15
80
79
47
14
10
3
499
1 : 0
0.88 : 0.12
0.64 : 0.36
1 : 0
1 : 0
1 : 0
1 : 0
0.32 : 0.68
9
10
11
12
13
DMF (dry, 3 mM lutidine)
KCl/HCl buffer (100 mM, pH 2.0)
a
Determined by NMR integration.
However, the low abundance of the CQS bond at lower pH could
Regarding reaction conversions, the results obtained from the
lead to thioamide formation, if the azide itself is reactive enough to NMR spectroscopy confirm previous observations that modestly
undergo a cycloaddition with the CQO bond. electron-poor and electron-rich azides tend to react very slowly in
Since it was shown that thioamides are effective quenchers for DMF with 3 mM 2,6-lutidine (Table 2, entries 11 and 12). In addition,
several fluorophores26 and partial peak overlapping in the fluores- we observed a large increase in conversion around pH 5 (Table 2,
cence chromatograms made qualitative analysis in part difficult, we entries 7 and 8), which drops again at lower pH values of 2–4 (Table 2,
decided to further investigate thioamide formation for two electro- entries 1–4). These findings indicate that careful pH handling can
nically different azido peptides – namely azido glycine peptide 9 enhance conversion rates with alkyl azides up to 80%. In addition, we
and g-azido butanoic acid peptide 10 – and the influence of the could show that by increasing the concentration of the reaction
solvent system in more detail by conducting 1D and 2D NMR mixture from 3 to 20 mM we could drive the reaction from 47% to
experiments (Table 2). In order to focus on the thioamide vs. amide almost full conversion without any additives, such as RuCl3. Further-
formation, peptides 9 and 10 did not possess any nitrogen- more, the latter result shows that the thioamide/amide ratio depends
containing side chains that could lead to non-selective acetylations. solely on the electronic nature of the employed azide and the pH
For better comparison of conversion rates and product ratios, during the thioacid–azide reaction (Table 2, entries 2 and 13).
both peptides were reacted with 20 eq. of thioacetic acid (2) in
In the final experiment, we wanted to probe thioacid–azide
different solvent systems with varying pH values for 2 d at 40 1C reactions employing small water-soluble alkyl azido molecules 13
(Table 2). After removal of excess thioacetic acid (2) and all volatiles, and 14 to determine isolated yields (Table 3). The employed azides
each sample was redissolved in D2O and checked by NMR spectro- 13 and 14 should bear similar electronic properties as azido peptides
scopy to determine the various reaction products (for spectra see the 9 and 10, respectively. Thioacid–azide reactions were performed
ESI†). Notably, formation of thioamide 12 decreased from pH 2 to under the reaction conditions used for the previously reported
pH 7 with amide 11 as the sole product at pH 7 and higher (Table 2, highest azide conversion (Table 2, entry 7) and the highest thioamide
entries 5, 6 and 9–12). These results support the previously observed formation (Table 2, entry 1). In addition, the concentration was set to
increase in thioamide formation under highly acidic conditions 20 mM for all subsequent reactions to achieve higher conversion
(Table 2, entries 1 and 2). At pH 2, this effect seemed to be much rates. As a result, all reactions showed higher azide conversion
stronger for the modestly electron-poor azido glycine peptide 9 than (480%) compared to the previous results with one exception of
for the electron-rich g-azido butanoic acid peptide 10. Due to the azide 14 at pH 5 (Table 3, entry 4), which might be due to its slightly
presumably higher reactivity of azido glycine peptide 9, the azide different electronic properties than the previously applied azides 9
seems to be reactive enough to undergo oxatriazolidine formation and 10. The preference for thioamide or amide formation does not
with the CQO bond of thioacetic acid (2) yielding a higher amount change for azide 13 and 14 though the ratio is not completely
of thioamide 12. In contrast, the low reactivity of the electron-rich transferable from azido glycine peptide 9 to azido glycine derivative
g-azido butanoic acid peptide 10 promotes the reaction with the less 13 under acidic conditions (Tables 2 and 3, entry 1) underlining the
abundant but more reactive CQS bond. With increasing pH, strong influence of the electronic nature of the azide on the reaction
thioacetic acid (2) exists predominantly in its deprotonated form outcome. We could thereby isolate the desired thioamides at pH 2 in
and the reaction seems to increasingly follow the mechanism for about 30% yield (Table 3, entries 1 and 2) and the corresponding
thiocarboxylates proposed by Williams and co-workers (Scheme 2A). amides at higher pH in moderate to good yields (20–50%)
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun., 2014, 50, 4603--4606 | 4605

View Article Online
Communication
ChemComm
Table 3 Thioacid–azide reactions with 13 and 14 (20 mM, 10 eq. thioacetic acid (2))
Entry
Azide
Solvent system
Ratio (amide : thioamide)a
Conversiona [%]
Yieldb [%]
1
2
3
4
13
14
13
14
KCl/HCl buffer (100 mM, pH 2.0)
KCl/HCl buffer (100 mM, pH 2.0)
NH4OAc buffer (100 mM, pH 4.0), 30% DMF
NH4OAc buffer (100 mM, pH 4.0), 30% DMF
0.29 : 0.71
0.40 : 0.60
0.93 : 0.07
0.61 : 0.39
80
87
100
61
30 (16)
32 (18)
50 (15)
18 (17)
a
b
Determined by NMR integration. Main product.
2000, 2, 2141; (c) R. Kleineweischede and C. P. R. Hackenberger,
Angew. Chem., Int. Ed., 2008, 47, 5984.
6 For a review: J. W. Bode, Top. Organomet. Chem., 2013, 44, 13.
7 For a review: C. P. R. Hackenberger and D. Schwarzer, Angew. Chem.,
Int. Ed., 2008, 47, 10030.
8 (a) M. Mu¨hlberg, D. M. M. Jaradat, R. Kleineweischede, I. Papp,
D. Dechtrirat, S. Muth, M. Broncel and C. P. R. Hackenberger,
Bioorg. Med. Chem., 2010, 18, 3679; (b) S. Sowa, M. Mu¨hlberg,
K. M. Pietrusiewicz and C. P. R. Hackenberger, Bioorg. Med. Chem.,
2013, 21, 3465.
9 (a) G. H. Hakimelahi and G. Just, Tetrahedron Lett., 1980, 21, 2119;
(b) T. Rosen, I. M. Lico and D. T. W. Chu, J. Org. Chem., 1988,
53, 1580.
(Table 3, entries 3 and 4). One should note that the conversions in
entries 1–3 are excellent and that the drop in the isolated yield was
partially due to difficult separation of the two products by HPLC.
In summary, we have shown that at lower pH values the thioacid–
azide reaction with electron-rich and modestly electron-poor azides
proceeds with high conversion rates and without any additives. With
the exception of very basic amino acid side chains such as lysine, the
reaction is highly selective in the presence of other functional groups.
In addition, we observed an increased formation of thioamides at a
pH o 7. We could show that the thioamide/amide ratio can be
controlled by varying pH with an increase of up to 92% thioamide
conversion for an azido glycine peptide and 68% thioamide conver-
sion for a g-azido butanoic acid peptide at pH 2. As this effect seems
to be stronger for modestly electron-poor azides, such as azido
glycine peptide 9, it would be interesting to see if the employment
of highly electron-deficient azides, e.g., sulfonyl azides, might even
lead to complete thioamide formation under strong acidic condi-
tions. During the last few decades, thioamides have gained much
importance, e.g., as a new class of drugs,27 as amide isosteres in
peptides,28 and as quenching units in fluorescent proteins to study
conformational changes.29 A new strategy for their selective synthesis
might enrich the field of thioamide containing probes.
10 D. T. S. Rijkers, R. Merkx, C.-B. Yim, A. J. Brouwer and R. M. J.
Liskamp, J. Pept. Sci., 2010, 16, 1.
11 (a) N. Shangguan, S. Katukojvala, R. Greenberg and L. J. Williams,
J. Am. Chem. Soc., 2003, 125, 7754; (b) R. V. Kolakowski, N. Shangguan,
R. R. Sauers and L. J. Williams, J. Am. Chem. Soc., 2006, 128, 5695.
12 K. Rohmer, O. Keiper, J. Mannuthodikayil and V. Wittmann, Peptides
2010: tales of peptides; proceedings of the thirty-first European Peptide
Symposium, San Diego, Prompt Scientific Publ, 2010, p. 218.
13 X. Wu and L. Hu, Tetrahedron Lett., 2005, 46, 8401.
14 R. Merkx, M. J. van Haren, D. T. S. Rijkers and R. M. J. Liskamp,
J. Org. Chem., 2007, 72, 4574.
15 (a) R. Merkx, A. J. Brouwer, D. T. S. Rijkers and R. M. J. Liskamp,
Org. Lett., 2005, 7, 1125; (b) R. Raz and J. Rademann, Org. Lett., 2012,
14, 5038.
16 S. S. Kulkarni, X. Hu, K. Doi, H.-G. Wang and R. Manetsch,
ACS Chem. Biol., 2011, 6, 724.
17 X. Zhang, F. Li, X.-W. Lu and C.-F. Liu, Bioconjugate Chem., 2009,
20, 197.
18 F. Fazio and C.-H. Wong, Tetrahedron Lett., 2003, 44, 9083.
19 For a review: C. T. Walsh, S. Garneau-Tsodikova and G. J. Gatto Jr.,
Angew. Chem., Int. Ed., 2005, 44, 7342.
20 D. L. Hartl, D. Freifelder and L. A. Snyder, Basic Genetics, Jones and
Bartlett Publishers, Boston, 1988.
The authors acknowledge financial support from the DFG (SFB
765 and SPP 1623), the Fonds der Chemischen Industrie (FCI),
the Einstein Foundation, the Boehringer-Ingelheim Foundation
(Plus 3 award) and the Studienstiftung des deutschen Volkes. The
¨
authors thank Prof. Jorg Rademann for helpful discussions and
¨
21 M. Schnolzer and S. B. H. Kent, Science, 1992, 256, 221.
Maria Glanz for experimental contributions.
22 R. V. Kolakowski, N. Shangguan and L. J. Williams, Tetrahedron Lett.,
2006, 47, 1163.
23 (a) A. Lu, M. Ke, W. Ding and R. Liu, Chin. J. Chem., 2009, 27, 227;
(b) W. Gordy, J. Chem. Phys., 1946, 14, 560.
Notes and references
¨
1 For reviews: (a) E. Valeur and M. Bradley, Chem. Soc. Rev., 2009, 38, 606; 24 R. Raz, PhD thesis, Freie Universitat Berlin, 2011.
(b) V. R. Pattabiraman and J. W. Bode, Nature, 2011, 480, 471.
2 B. Shen, D. M. Makley and J. N. Johnston, Nature, 2010, 465, 1027.
25 E. Schaumann, Supplement A, in The Chemistry of double-bonded
functional groups, ed. S. Patai, Wiley, 1989.
3 A. Dumas, G. A. Molander and J. W. Bode, Angew. Chem., Int. Ed., 26 J. M. Goldberg, S. Batjargal and E. J. Petersson, J. Am. Chem. Soc.,
2012, 51, 5683. 2010, 132, 14718.
4 P. E. Dawson, T. W. Muir, I. Clark-Lewis and S. Kent, Science, 1994, 27 (a) D. S. Copper, N. Engl. J. Med., 2005, 352, 905; (b) B. Srinivas and
266, 776. For a review: L. Raibaut, N. Ollivier and O. Melnyk, Chem.
Soc. Rev., 2012, 41, 7001.
R. D. Su¨ssmuth, ChemBioChem, 2010, 11, 1335.
28 A. Choudhary and R. T. Raines, ChemBioChem, 2011, 12, 1801.
5 (a) B. L. Nilsson, L. L. Kiessling and R. T. Raines, Org. Lett., 2000, 29 S. Batjargal, Y. J. Wang, J. M. Goldberg, R. F. Wissner and E. J.
2, 1939; (b) E. Saxon, J. I. Armstrong and C. R. Bertozzi, Org. Lett., Petersson, J. Am. Chem. Soc., 2012, 134, 9172.
4606 | Chem. Commun., 2014, 50, 4603--4606
This journal is ©The Royal Society of Chemistry 2014