Tetrahedron Letters 58 (2017) 3421–3425

Contents lists available at ScienceDirect

Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet

Comparison of various coupling reagents in solid-phase aza-peptide

synthesis
Meeli Arujõe, Anu Ploom ⇑, Anton Mastitski, Jaak Järv
Institute of Chemistry, University of Tartu, 14A Ravila Str, Estonia

a r t i c l e i n f o a b s t r a c t

Article history: Aza-peptides are promising drug leads, however extensive study of their properties is hampered by low
Received 10 May 2017 yielding aza-peptide bond formation during conventional Fmoc SPPS. The kinetics of aza-peptide bond
Revised 6 July 2017 formation in the model peptide H-Ala-AzAla-Phe-NH2 was compared with various conventional amino
Accepted 17 July 2017
acid activators. The reaction rates and yields were dependent on the activator structure. The reaction
Available online 18 July 2017
time of aza-peptide formation using oxyma-based agents was approximately 30 times longer than in typ-
ical peptide synthesis. Therefore, new activators are required to increase the reactivity of the activated
Keywords:
amino acid to achieve effective acylation of the semicarbazide moiety during aza-peptide bond
Activators
Aza-peptide
formation.
Kinetics Ó 2017 Elsevier Ltd. All rights reserved.
SPPS

Biologically active peptides may represent ideal drug candi- influence on the binding effectiveness of aza-peptides with differ-
dates due to their biocompatibility and outstanding specificity ent binding sites in proteins.8 Thus, these peptidomimetics repre-
for various target sites. However, peptide applicability is generally sent promising drug leads.9,10
limited by their rapid degradation in living organisms.1,2 Thus, In spite of these challenging features,5 the properties of
identifying suitable chemical modifications that increase peptide aza-peptides have only been less studied.4,6,11–13 This is due to
stability is a major challenge. Several different approaches to pep- the difficulties associated with the synthesis of aza-amino acid pre-
tidomimetic design have emerged.3 One such approach is chemical cursors and building blocks.11,12 Additionally, common peptide
modification of the peptide backbone structure via nearly isosteric bond synthesis methods, wherein the peptide N-terminal amino
substitution of the a-carbon atoms in conventional amino acids 1 group is acylated by the subsequent amino acid, are unefficient
with nitrogen to yield amino acid aza-analogs 2 (Fig. 1). for aza-peptide bond synthesis. In this case, acylation of the semi-
The first study concerning biologically active aza-petides was carbazide moiety NH2–NR–C(O)– should occur instead of alky-
reported by Hess and co-workers.4 As a result, many attempts have lamine acylation, however the nucleophilicities of these two
been made to synthesize and study peptidomimetics containing nitrogen atoms are different.14,15 Unfortunately, most studies do
aza-amino acids. As expected, these studies revealed that the not consider this difference and the similarities between the che-
aza-peptide bond is significantly more stable towards degradation mistries of peptide and aza-peptide bond synthesis have been
by peptidases, such as a-chymotrypsin2,5 and subtilisin,5 than the advocated even in textbooks.16
typical peptide bond. This is in line with the general understanding The differing reactivities of amino groups in a-amino acids 1
that the carbonyl C atom within the structural fragment –NH–NR– and their aza-derivatives 2 can be theoretically justified by the
C(O)–NH– exhibits lower electrophilicity than that of the amide nucleophilicity parameters N = 13.85 (sN 0.53) and N = 11.05 (sN
group –NH–CH(R)–C(O)–NH– in the peptide backbone. 0.52) listed in Mayr’s nucleophilicity database for methylamine
The replacement of amino acids with their aza-analogs also and semicarbazide, respectively.17 The aim of this study was to
induces greater peptide backbone rigidity, since the u and w dihe-
dral angles of the aza-peptide bond are constrained.6,7
These constraints favor formation of the b-turn conformation,
as observed via X-ray and NMR studies and revealed using compu-
tational models. This conformational effect may have minimal

⇑ Corresponding author.
E-mail address: anu.ploom@ut.ee (A. Ploom). Fig. 1. Structure of amino acid (1) and aza-amino acid (2).

http://dx.doi.org/10.1016/j.tetlet.2017.07.063
0040-4039/Ó 2017 Elsevier Ltd. All rights reserved.
3422 M. Arujõe et al. / Tetrahedron Letters 58 (2017) 3421–3425

Scheme 1. Model aza-peptide H-Ala-AzAla-Phe-NH2 synthesis.

Table 1
experimentally evaluate these differences, as well as the applica- Kinetic study of aza-peptide bond formation via the reaction of Fmoc-Ala-OH with the
bility of the conventional SPPS protocol to aza-peptide bond syn- semicarbazide group of the resin-bound H-AzAla-Phe residue at 25 °C using various
thesis in general. For this purpose, a systematic investigation of coupling reagents. Eq. (1) was used to calculate the kobs and yield (1  Y1) values
using Graphpad 5 software.
the kinetics of aza-peptide bond formation in the model aza-
peptide H-Ala-AzAla-Phe-NH2 was conducted, and the applicabil- Activator Leaving group X kobs, min1 Yield pKa of HX18
ity of various amino acid activators to this reaction was analyzed. COMU 19
Oxyma 0.022 ± 0.001 0.99 ± 0.01 4.24
To the best of our knowledge, no systematic aza-peptide bond for- PyOxim20 Oxyma 0.023 ± 0.001 0.95 ± 0.01 4.24
mation kinetic study has been reported thus far. HDMC21 6-Cl-HOBt 0.016 ± 0.001 0.55 ± 0.02 4.62
HCTU22 6-Cl-HOBt 0.017 ± 0.002 0.68 ± 0.03 4.62
Scheme 1 outlines the preparation of the model aza-peptide,
HATU23 HOAt 0.017 ± 0.001 0.93 ± 0.01 4.65
H-Ala-AzAla-Phe-NH2 8 from Fmoc-protected methyl hydrazine 3 TBTU24 HOBt 0.004 ± 0.001 0.69 ± 0.05 5.65
(see ESI for details). The most critical step is formation of the PyBOP25 HOBt 0.005 ± 0.002 0.65 ± 0.14 5.65
aza-peptide bond via the coupling of activated Fmoc-Ala-OH 6 to
the nitrogen atom of the deprotected semicarbazide moiety in 5.
Therefore, a detailed kinetic study of this step was conducted using
The observed first-order rate constants (kobs) and reaction
various activating agents. Each activating agent inserts a different
yields (1  Y1) were determined at 25 °C (Table 1), and in some
leaving group X into alanine 6 to facilitate coupling of the amino
cases also at 40 °C. The reaction time was dependent on the activa-
acid to the resin-linked peptide or aza-peptide.
tor used, but product formation was monitored for 300 min in
The reaction of activated alanine 6 to the deprotected peptide
most experiments
sequence 5 was studied under conditions defined by the conven-
Initially, the time evolution of aza-tripeptide 7 formation was
tional SPPS protocol. At appropriate times, aliquots were taken
compared to synthesis of the typical peptide bond in the resin-
from the reaction mixture and analyzed via HPLC (see ESI for
bound tripeptide Fmoc-Ala-Ala-Phe. Two oxyma-based coupling
details). As a 10-fold excess of alanine relative to the number of
reagents, COMU and PyOxim, were used to activate Fmoc-pro-
resin-bound reaction sites was used, the process was described
tected alanine in the latter case. The kobs values for resin-bound
using a first-order rate equation
tripeptide Fmoc-Ala-Ala-Phe formation were 1.02 ± 0.29 and
Y ¼ ekobs t þ Y 1 ð1Þ 1.06 ± 0.24, respectively. The results of the study using aza-pep-
tides are listed in Table 1 and the kinetic curves obtained at
where kobs is the observed first-order rate constant, t is the reaction
25 °C are shown in Fig. 2.
time and Y1 is the plateau value that is reached at the end of the
COMU and PyOxim had similar effects on both reactions; nearly
acylation reaction. Parameter Y characterizes the tripeptide 7 for-
complete acylation of the resin-linked aza-peptide and the conven-
mation process and was calculated using the remaining starting
tional peptide were achieved. In the case of peptide synthesis, the
material (dipeptide) and tripeptide peak areas (S) from same chro-
reaction half-life26 (t1/2 = ln 2/kobs) is about 1 min. This is in line
matographic run:
with expectations of the fast conventional Fmoc SPPS protocol,
Sdipeptide where the amino acid coupling time is 2–5 min at room
Y¼ ð2Þ
Sdipeptide þ Stripeptide temperature.22,27
 M. Arujõe et al. / Tetrahedron Letters 58 (2017) 3421–3425 3423

Fig. 2. Time evolution of the reactions between activated Fmoc-alanine and resin-bound H-AzAla-Phe during aza-peptide bond synthesis (a), and with resin-bound H-Ala-Phe
during peptide bond synthesis (b) at 25 °C in DMF. Fmoc-alanine was activated with either COMU (d) or PyOxim (s).

However, the reaction time of aza-peptide formation is approx- The aza-peptide bond formation yield was only slightly higher
imately 30 times longer, as can be seen from the kobs values listed than 50% (Fig. 3).
in Table 1. This difference is also indicated by the time-scale of the Interestingly, aza-peptide bond formation occurs more slowly
kinetic curve for this reaction (Fig. 2). This data demonstrates that with TBTU and PyBOP. The half-lives of these reactions were
synthesis of an aza-peptide bond is slower than that of a peptide 150 min. As above, these reactions were incomplete and are char-
bond. Thus, the conventional SPPS protocol cannot be applied to acterized by an extrapolated yield of 0.6. Tripeptide formation
aza-peptide synthesis without significant changes in activator was faster at 40 °C than at 25 °C. The reaction was characterized
reactivity. by kobs values of 0.021 ± 0.005, 0.014 ± 0.003, and 0.007 ± 0.002
Aza-peptide formation was faster and characterized by a kobs for HCTU, PyBOP, and TBTU, respectively. However, tripeptide for-
value of 0.040 ± 0.001 min1 at 40 °C, when COMU was used for mation was incomplete at this temperature; he yields were 0.7, 0.5,
Fmoc-alanine activation. Again, practically complete conversion and 0.8 for HCTU, PyBOP, and TBTU, respectively. Since these yields
of the dipeptide into the tripeptide was observed. Although this are similar to those obtained at 25 °C, the reaction rate is not clo-
reaction was faster, it was still not comparable to the conventional sely related to the yield.
peptide synthesis protocol, since its half-life was 17 min. Synthesis Carbodiimides are another type of amino acid activator used for
of the model peptide was then studied with various activating aza-peptide bond synthesis.6,28 Thus, the model reaction was stud-
agents. HATU, HCTU, HDMC, TBTU, and PyBOP are all triazole ied after Fmoc-alanine activation with DIC.29 Kinetic curves
derivatives and produce similar leaving groups in compound 6. obtained at 25 °C and 40 °C are shown in Fig. 4.
The results using these activating agents are shown in Fig. 3 and DIC is a less efficient aza-peptide synthesis coupling reagent
the observed rate constants and calculated acylation yields are than previously considered activators; the half-life of the reaction
listed in Table 1. with DIC-activated Fmoc-alanine was 211 min at 25 °C. However,
Fig. 3 shows that product formation reaches an asymptote after tripeptide formation is nearly complete after an extended reaction
250 min using HATU, HCTU or HDMC. The reaction rates are some- time. Similar results were obtained at 40 °C; the reaction rate
what lower than those using COMU and PyOxim. The half-lives of increases (kobs = 0.017 ± 0.002 min1), and conversion of the aza-
the first three reactions were each 40 min. Although HATU dipeptide was almost complete. However, this increase is not suf-
results in nearly complete conversion of the dipeptide into the ficient to make the reaction rate comparable to that of peptide
tripeptide, the reactions using HCTU and HDMC were incomplete. bond synthesis.

Fig. 3. Time evolution of the reaction between activated Fmoc-alanine and resin-
bound H-AzAla-Phe during aza-peptide bond synthesis at 25 °C in DMF. Fmoc- Fig. 4. Time evolution of the reaction between DIC-activated Fmoc-alanine and
alanine was activated with TBTU (h), PyBOP (s), HDMC (r), HCTU (N), and HATU resin-bound H-AzAla-Phe for aza-peptide bond synthesis at 25 °C (d) and 40 °C (s)
(.). in DMF.
3424 M. Arujõe et al. / Tetrahedron Letters 58 (2017) 3421–3425

Fig. 5. Structures of leaving groups in compound 6: Oxyma (9), HOAt (10), 6-Cl-HOBt (11), HOBt (12).

Synthesis of the aza-peptide bond was much slower than that of

the typical peptide bond, including when conventional coupling
agents were used for amino acid activation. Moreover, both the
reaction rates and the yields were dependent on the activator
structure. Oxyma-based activators COMU and PyOxim lead to
nearly complete aza-peptide bond formation. Triazole based HATU
is as efficient in achieving complete coupling, but it requires a
longer reaction time. These results demonstrate that the conven-
tional SPPS protocol cannot be directly applied to aza-peptide syn-
thesis, and that new coupling agents are needed to increase the
reactivity of the activated amino acid in order to achieve effective
acylation of the semicarbazide moiety during aza-peptide bond
formation.

Acknowledgment

Fig. 6. Dependence of log kobs values of the aza-peptide bond synthesis reaction on This research was supported by institutional research funding
the pKa values of acids that correspond to the leaving group X of activated Fmoc- IUT (IUT20-15) of the Estonian Ministry of Education and Research.
alanine 6 in Scheme 1 (9 oxyma, 10 HOAt, 11 6-Cl-HOBt, 12 HOBt). The log kobs
values were taken from Table 1, and the pKa values were compiled from the
literature.18 A. Supplementary data

Supplementary data associated with this article can be found, in

All model peptide synthesis reactions were conducted under the online version, at http://dx.doi.org/10.1016/j.tetlet.2017.07.
similar conditions, including reagent concentration and sample 063.
processing conditions. Therefore, the kinetic data obtained during
experiments with different activators should be comparable. References
Specifically, the rate constant kobs should characterize the reactiv-
ities differences of the activated alanines 6 during acylation. This 1. Pollaro L, Heinis C. Med Chem Commun. 2010;1:319–324.
2. Tal-Gan Y, Freeman NS, Klein S, Levitzki A, Gilon C. Chem Biol Drug Des.
hypothesis was confirmed by the linear log kobs vs pKa plot
2011;78:887–892.
(Fig. 6), where the pKa values measured in a 95% acetonitrile-water 3. Stawikowski M, Fields GB. In Board John E, editor, Curr. Protoc. Protein Sci.
mixture at 25 °C18 quantify the acidities of compounds that Coligan Al; 2002 [Chapter, Unit-18.1].
4. Hess HJ, Moreland WT, Laubach GD. JACS. 1963;85:4040–4041.
correspond to the leaving group X in the activated Fmoc-alanine
5. Gupton BF, Carroll DL, Tuhy PM, Kam CM, Powers JC. J Biol Chem.
(compound 6 in Scheme 1) with following structures (Fig. 5). 1984;259:4279–4287.
However, in parallel to the influence of leaving group acidity 6. Boeglin D, Lubell WD. J Comb Chem. 2005;7:864–878.
that seems to quantify polar effects, it is possible that steric effects 7. Thormann M, Hofmann HJJ. Mol Struct Theochem. 1999;469:63–76.
8. Boeglin D, Xiang Z, Sorenson NB, Wood MS, Haskell-Luevano C, Lubell WD.
may also play some role in this acylation reaction, in analogy with Chem Biol Drug Des. 2006;67:275–283.
ester hydrolysis.30 Therefore, before final conclusions about the 9. Proulx C, Sabatino D, Hopewell R, Spiegel J, García Ramos Y, Lubell WD. Future
role of the leaving group in the activated amino acid molecule, Med Chem. 2011;3:1139–1164.
10. Zega A. Curr Med Chem. 2005;12:589–597.
the reactivity of more bulky substrates should be analyzed. 11. Freeman NS, Tal-Gan Y, Klein S, Levitzki A, Gilon C. JOC. 2011;76:3078–3085.
To understand why the reaction yield varies by activator, it is 12. Freeman NS, Hurevich M, Gilon C. Tetrahedron. 2009;65:1737–1745.
important to emphasize that the process was monitored via direct 13. Chingle R, Ratni S, Claing A, Lubell WD. Pept Sci. 2016;106:235–244.
14. Nigst TA, Antipova A, Mayr H. J Org Chem. 2012;77:8142–8155.
measurement of resin-bound H-AzAla-Phe conversion to the aza- 15. Garcia-Ramos Y, Proulx C, Lubell WD. Can J Chem. 2012;90:985–994.
tripeptide Fmoc-Ala-AzAla-Phe. This ensures that none of the 16. Trabocchi A, Guarna A. Peptidomimetics in Organic and Medicinal Chemistry. The
reagents were consumed in parallel reactions. Therefore, the low Art of Transforming Peptides in Drugs. John Wiley & Sons; 2014.
17. Mayr’s Database of Reactivity Parameters. <http://www.cup.lmu.de/oc/mayr/
reaction yields observed with certain activators should be related
reaktionsdatenbank/> Accessed: April 2017.
to side-reactions of the activated Fmoc-alanine. As the yield is 18. Fathallah MF, Khattab SN. J Chem Soc Pak. 2011;33:324–332.
not dependent on the reaction temperature, the aza-peptide syn- 19. El-Faham A, Subirós Funosas R, Prohens R, Albericio F. Chem Weinh Bergstr Ger.
2009;15:9404–9416.
thesis and putative side-reaction should have similar temperature
20. Subirós-Funosas R, El-Faham A, Albericio F. Org Biomol Chem.
dependencies. Incomplete acylation has also been observed in typ- 2010;8:3665–3673.
ical peptide bond formation. In this case, repeating the process was 21. El-Faham A, Albericio F. J Org Chem. 2008;73:2731–2737.
recommended.27 In our case, these side-reactions may play even 22. Hood CA, Fuentes G, Patel H, Page K, Menakuru M, Park JH. J Pept Sci.
2008;14:97–101.
more important role, as the aza-peptide bond formation reaction 23. Alewood P, Alewood D, Miranda L, Love S, Meutermans W, Wilson D. Academic
is very slow. Press. 1997;289:14–29.
 M. Arujõe et al. / Tetrahedron Letters 58 (2017) 3421–3425 3425

24. Knorr R, Trzeciak A, Bannwarth W, Gillessen D. Tetrahedron Lett. 28. Doan ND, Zhang J, Traoré M, Kamdem W, Lubell WD. J Pept Sci.
1989;30:1927–1930. 2015;21:387–391.
25. Coste J, Le-Nguyen D, Castro B. Tetrahedron Lett. 1990;31:205–208. 29. Sheehan J, Cruickshank P. J Org Chem. 1961;26:2525–2528.
26. Atkins PW. Physical Chemistry. 1st ed. San Francisco: W.H. Freeman and 30. Williams A. Free Energy Relationships in Organic and Bio-organic
Company; 1978. Chemistry. Cornwall: Royal Society of Chemistry; 2003.
27. Chantell CA, Onaiyekan MA, Menakuru M. J Pept Sci. 2012;18:88–91.