Water-soluble triazabutadienes that release diazonium species upon protonation under physiologically relevant conditions

Article abstract of DOI:10.1002/anie.201411277

Triazabutadienes are an understudied structural motif that have remarkable reactivity once rendered water-soluble. It is shown that these molecules readily release diazonium species in a pH-dependent manner in a series of buffer solutions with pH ranges similar to those found in cells. Upon further development, we expect that this process will be well suited to cargo-release strategies and organelle-specific bioconjugation reactions. These compounds offer one of the mildest ways of generating diazonium species in aqueous solutions.

Full text of DOI:10.1002/anie.201411277

Angewandte
Chemie
International Edition: DOI: 10.1002/anie.201411277
German Edition: DOI: 10.1002/ange.201411277
Diazonium Salts
Hot Paper
Water-Soluble Triazabutadienes that Release Diazonium Species upon
Protonation under Physiologically Relevant Conditions**
Flora W. Kimani and John C. Jewett*
Abstract: Triazabutadienes are an understudied structural
motif that have remarkable reactivity once rendered water-
soluble. It is shown that these molecules readily release
diazonium species in a pH-dependent manner in a series of
buffer solutions with pH ranges similar to those found in cells.
Upon further development, we expect that this process will be
well suited to cargo-release strategies and organelle-specific
bioconjugation reactions. These compounds offer one of the
mildest ways of generating diazonium species in aqueous
solutions.
Whereas many biochemical events are prompted by subtle
changes in the pH value, there are relatively few small-
molecule reactions that are equally responsive. Our work on
the pH-responsive dengue virus[1] has prompted our entry into
this area. We are looking for processes that are responsive
within the relevant pH and temporal ranges of infection[2] and
ideally generate species that go on to perform other reactions.
Herein, we report a chemical system that brings us closer to
meeting this challenge, as a stable species is transformed into
a reactive species in a physiologically relevant pH range.
There are two general types of reactions that probe
changes in the pH value: Some dyes undergo a reversible
bathochromic shift upon protonation,[3] and some acid-labile
protecting groups can irreversibly uncage to reveal a reactive,
therapeutic, or diagnostic functionality. Whereas the former
are rapid indicators, it is challenging to parlay their proto-
nation into further chemical reactivity. The most prominent
uncaging reactions, namely those of Schiff bases,[4] acetals,[5]
or imidazole-based cross-linkers,[6] do not have unmasking
rates with the dramatic pH dependence that we desire. Our
search for reactivity led us to the work of Fanghänel and co-
workers and their two-decade-long study of 1,2,3-triazabuta-
dienes 1 (Figure 1b).[7] Inspired by their work, the processes
reported herein represent a clear departure from existing pH-
Figure 1. a) The transformation of a probe triggered by physiological
changes. b) Several classes of triazabutadienes that have been
reported previously. Bn=benzyl, Ts = para-toluenesulfonyl.
triggered transformations and, to the best of our knowledge,
feature unique reactivity.
Triazabutadienes have undergone a small renaissance in
recent years mainly because of the work of Bielawski and
Khramov (compound 2; Figure 1b).[8] While these molecules
are now commonly called triazenes, for the purpose of our
discussion, we will continue to call this functional group
a triazabutadiene, in part to distinguish it from the more
conventional N,N-dialkyltriazenes,[9] but also in recognition of
the role that the left half of the molecule plays in determining
its reactivity. Much of the more recent work on these
compounds has focused on their “push–pull” properties (see
compound 3, Figure 1b)[10] and Staudinger-like rearrange-
ments to expel nitrogen upon heating of the compounds.[8,11]
What fascinated us about triazabutadienes was their early
reported instability in the presence of strong acids (Fig-
ure 2a). The mechanism for their degradation, which is
summarized in Figure 2b, involves the protonation of the
N3 nitrogen atom followed by elimination of an aryl
diazonium species. This simple mechanism is complicated by
[*] F. W. Kimani, Dr. J. C. Jewett
Department of Chemistry and Biochemistry
University of Arizona
1306 E University Blvd, Tucson, AZ 85721 (USA)
E-mail: jjewett@email.arizona.edu
E/Z isomerization about the N1 N2 bond[7c] and the fact that
À
[**] We thank Jon Njardarson and Indraneel Ghosh for discussions,
Neal Jacobsen and Jixun Dai for NMR expertise, Kofi Bright for
assistance with buffer preparation, and Elisa Tomat and Craig
Aspinwall for laboratory equipment. This work was supported in
part by an ACS-PRF grant to J.C.J. and by start-up funds provided by
the Arizona Board of Regents to J.C.J. We also thank the NSF for
a departmental instrumentation grant for the NMR facility (CHE-
0840336).
protonation and alkylation of the N1 nitrogen atom is highly
preferred.[7d,12] The vast majority of the work on such
structures has been carried out with molecules featuring
a benzothiazole or benzimidazole moiety in organic solvents,
and upon making a departure from both, we were excited by
what we found.
Whereas Fanghänel and co-workers had previously shown
that strong mineral acids such as hydrochloric acid could
facilitate the decomposition of 1, they did not perform the
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201411277.
Angew. Chem. Int. Ed. 2015, 54, 4051 –4054
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4051

.
Angewandte
Communications
employing acetic acid instead of HCl, the same reaction
occurred, albeit at a slower rate.
We synthesized a water-soluble analogue by treating
methylimidazole with 1,3-propanesultone to obtain zwitter-
ionic NHC precursor 6. Upon treatment with base in the
presence of phenyl azide, triazabutadiene 7 was formed in
acceptable yield (Scheme 1b). This compound was highly
water-soluble, but upon dissolution, we were met with
a surprising level of instability. Whereas this species was
stable in methanol or dimethylsulfoxide (DMSO), it degraded
in water with a half-life of approximately 24 hours.[15] Upon
increasing the steric hindrance by replacing the methyl
substituent with a mesityl group, we obtained compound 8
from 9; only 10% of 9 degraded upon sitting in D2O for
Figure 2. a) Fanghänel and co-workers reported that 1 decomposed in
the presence of a strong acid. b) Reaction of 1 by protonation of the
N3 nitrogen atom.
reaction in water or assess its susceptibility to degradation in
physiologically more relevant pH ranges.[7e] Based on the
harsh conditions used to degrade triazabutadienes with
benzothiazole or benzimidazole moieties, we hypothesized
that these derivatives would be too stable for our purposes.
Instead, we selected an imidazole core, in no small part
because of the wealth of N-heterocyclic carbene (NHC)
literature and synthetic precedent.[13,14] To assess the reac-
tivity of the core, we synthesized triazabutadiene 4 from
dimethylimidazolium iodide (5) and phenyl azide (Sche-
me 1a). We chose this azide-based synthetic approach, which
was reported by Bielawski, over Fanghänelꢀs original coupling
of an appropriately deprotonated cyclic guanidine with an
aryl diazonium salt for reasons of synthetic ease and
versatility. No changes were observed for a methanolic
mixture of equimolar amounts of 4 and resorcinol (to trap
any diazonium species formed) until we added HCl. Upon
acidification, the solution underwent a rapid color change and
provided the known aryl azo dye Sudan Orange G (Sche-
me 1a). When we increased the pKa value of the acid by
Figure 3. a) A representative NMR experiment showing that 8 is
converted into 10 in a pH 6 phosphate/citrate buffer over time. b) The
reaction rate of the benzenediazonium salt formation from 8 in buffers
is highly pH-dependent. Error bars indicate the standard deviation
from three separate reactions. PC=phosphate/citrate buffer, P=phos-
phate buffer. c) Formation of azobenzene 11 from 8.
Scheme 1. a) Synthesis of triazabutadiene 4 and acid-promoted diazo-
nium generation. The diazonium ion is trapped to provide Sudan
Orange G. b) Synthesis of triazabutadienes 7 and 8. Mes=mesityl.
4052 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 4051 –4054

Angewandte
Chemie
24 hours. Presumably, the bulky mesityl group disfavors
protonation of the N3 nitrogen atom.
substituent. Triazabutadiene 13 was stable in pure water,[20]
but rapidly precipitated from buffered solutions (even at
pH 7). Fearing that an unexpected degradation pathway
might be occurring in buffered solutions, we isolated the solid,
dissolved it in deuterated methanol, and confirmed that the
starting material had remained unaltered. Upon acidification
of the methanolic solution with HCl, a color change occurred,
and the solution started bubbling upon prolonged exposure to
the acid.[21] The problems with aqueous solubility foiled our
attempts to examine 13 in physiologically relevant buffers. To
directly compare the reactivity of 13 with those of 8 and 12, we
monitored their reaction rates in a 1:1 mixture of D2O/MeOD
with a ten-fold excess of formic acid (Figure 4a). While
completely abiotic, this solvent system was unique in its
ability to keep electron-poor 13 in solution in the presence of
a high concentration of water and acid. Compounds 8 and 12
While establishing the stability of 8 in water, we noted that
the degradation reaction slowed down and eventually stopped
over the course of the 24 hours. We realized that we were
consuming a reagent in the form of hydronium ions and upon
monitoring the pH value over time, we found that it rapidly
jumped from neutral pH to pH 9.7 and eventually levelled off
at pH 9.9 after 24 hours. Consistent with this observation and
the proposed mechanism, both compounds were found to be
stable in 0.1n NaOH with no detectable degradation after
24 hours.
To test the stability at a “constant”, non-basic pH value,
we performed a series of experiments in buffered solutions.
Within the studied buffering range, we observed a sigmoidal
dependence of the rate constant on the pH value, centered at
pH 6 (Figure 3a,b).[16] Imidazole 10,
which remains in solution, offers
a convenient handle for comparing
the amounts of starting material and
product.[17] To our surprise, we
observed pseudo-zeroth-order reac-
tion kinetics as the concentration of
8 decreased linearly with time. We
rationalized this observation with
deference to the proposed overall
mechanism, which states that the
reactive N3-protonated form is much
less abundant than the non-reactive
N1-protonated form. To avoid com-
plications associated with the buffer
capacity, resorcinol was not added to
these reactions. Over time, we
observed the formation of a yellow
precipitate, which was confirmed to
be 4-phenylazophenol (11, Fig-
ure 3c). This product was formed by
the hydrolysis of one diazonium ion to
yield phenol (likely by the pH-inde-
pendent formation of the aryl
Figure 4. a) Time-dependent diazonium salt release of compounds 8, 12, 13, and 15 (15 mm) in
a water/methanol mixture (1:1) containing formic acid (10 equiv). b) Unlike 8 and 12, upon
exposure to formic acid, compound 13 exists in a different form (compound 14, observable by its
unique NMR spectrum), which slowly releases a diazonium species and 10.
cation)[18] followed by
a
reaction
with a second diazonium ion.
The relative instability of 8 at
pH 7 prompted us to synthesize sev-
eral analogues to test the effect of
electronic perturbations on the reac-
tivity. For the purposes of this discussion, we have focused on
the aryl moiety, but plan to explore other parameters that
could influence these reactions. Based on previous work[7e]
and first principles,[19] we expected that electron-donating
substituents would increase the partial charge on the N3
nitrogen atom, favoring protonation and decreasing the
stability. Indeed, triazabutadiene 12, with an electron-donat-
ing para-methoxy substituent, was found to be roughly twice
as reactive as 8 at pH 4, 5, 6, and 7 (2.4 Æ 0.1 ” 10À5 m sÀ1, 1.9 Æ
0.2 ” 10À5 m sÀ1, 1.0 Æ 0.1 ” 10À5 m sÀ1, and 0.34 Æ 0.03 ”
10À5 m sÀ1, respectively).
behaved as expected with 12 reacting at a faster rate than 8,
but unlike in the buffered systems, the rate took on a decidedly
more first-order appearance while not being strictly first
order. Comparing the initial slopes to those in water, these
reactions were considerably slower than those in buffered
water, but this finding can be attributed to the large
percentage of organic solvent present in solution, which
leads to significant alterations of the solvation and pKa values.
Upon addition of the acid to 13, we observed a bathochromic
shift of the solution and a new species by NMR spectroscopy.
After twelve hours, only 38% of 13 had been consumed to
yield the diazonium species and the cyclic guanidine. Based
on these data, we hypothesized that the new species is
To test the corollary and study a derivative with increased
stability, we synthesized triazabutadiene 13 with a para-nitro
Angew. Chem. Int. Ed. 2015, 54, 4051 –4054
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4053

.
Angewandte
Communications
[7] a) E. Fanghänel, R. Hänsel, W. Ortmann, J. Hohlfeld, J. Prakt.
Chem. 1975, 317, 631 – 640; b) H.-T. Dorsch, H. Hoffmann, R.
Hansel, G. Rasch, E. Fanghänel, J. Prakt. Chem. 1976, 318, 671 –
680; c) E. Fanghänel, R. Hänsel, J. Hohlfeld, J. Prakt. Chem.
1977, 319, 485 – 493; d) E. Fanghänel, H. Poleschner, R. Rade-
glia, R. Hänsel, J. Prakt. Chem. 1977, 319, 813 – 826; e) E.
Fanghänel, J. Hohlfeld, J. Prakt. Chem. 1981, 323, 253 – 261; f) R.
Radeglia, R. Wolff, T. Steiger, S. Simova, E. Fanghänel, J. Prakt.
Chem. 1984, 326, 511 – 514; g) E. Fanghänel, W. Ortmann, A.
Hennig, J. Prakt. Chem. 1988, 330, 27 – 34; h) E. Fanghänel, W.
Ortmann, J. Prakt. Chem. 1989, 331, 721 – 725; i) E. Fanghänel,
J. U. Bauroth, H. Hentschel, F. Gußmann, H. Alzyadi, W.
Ortmann, J. Prakt. Chem. 1992, 334, 241 – 247.
[8] D. M. Khramov, C. W. Bielawski, Chem. Commun. 2005, 4958 –
4960.
[9] a) S. Dahmen, S. Bräse, Org. Lett. 2000, 2, 3563 – 3565; b) S.
Bräse, Acc. Chem. Res. 2004, 37, 805 – 816.
[10] D. Jishkariani, C. D. Hall, A. Demircan, B. J. Tomlin, P. J. Steel,
A. R. Katritzky, J. Org. Chem. 2013, 78, 3349 – 3354.
[11] D. M. Khramov, C. W. Bielawski, J. Org. Chem. 2007, 72, 9407 –
9417.
[12] A. G. Tennyson, E. J. Moorhead, B. L. Madison, J. A. V. Er,
V. M. Lynch, C. W. Bielawski, Eur. J. Org. Chem. 2010, 6277 –
6282.
[13] W. Herrmann, C. Kçcher, Angew. Chem. Int. Ed. Engl. 1997, 36,
2162 – 2187; Angew. Chem. 1997, 109, 2256 – 2282.
[14] N. Marion, S. Díez-Gonzµlez, S. P. Nolan, Angew. Chem. Int. Ed.
2007, 46, 2988 – 3000; Angew. Chem. 2007, 119, 3046 – 3058.
[15] Based on integration of the mesityl region.
[16] In pH 8 PBS buffer (the edge of the buffering window) we
observed a decrease in rate with time, and the pH value was
found to be 9.2 upon completion of the reaction.
[17] We never observed the reverse reaction to reform the triazabu-
tadiene under our conditions.
[18] A. F. Hegarty in The Chemistry of Diazonium and Diazo
Groups, Vol. 2 (Ed.: S. Patai), Wiley, New York, 1978, pp. 511 –
591.
[19] L. P. Hammett, J. Am. Chem. Soc. 1937, 59, 96 – 103.
[20] No degradation was observed after 24 hours, and the pH value of
the solution remained neutral over this time period.
[21] This precipitate could be resuspended by neutralizing the
solution with NaOH prior to bubbling.
[22] Initial attempts to avoid this complication with other electron-
deficient aryl species (such as para-chloro and para-trifluoro-
methyl-substituted systems) were thwarted by a lack of water
solubility preventing a comparison of the stabilities at neutral
pH values.
compound 14 (Figure 4b), which adds an extra layer of
complication to the reaction analysis.[22]
We thought to circumvent the complication imposed by
the formation of protonated species such as 14 by synthesizing
meta-nitro-substituted derivative 15, where a similar process
is unlikely. As for the other electron-deficient compounds
that we had synthesized and evaluated, 15 was soluble and
stable in D2O but insoluble in buffers. The yellow precipitate
that formed upon acidification was collected and dissolved in
[D6]DMSO. This compound was identified as a new species
and assigned as the stable N1-protonated form. The addition
of triethylamine to the NMR tube returned 15 unaltered, thus
bolstering our hypothesis (Supporting Information, Fig-
ure S1). Again, because of its insufficient solubility in water,
we treated 15 with formic acid in a methanol/water mixture
(as before), and this species was indeed found to be reactive in
the presence of an acid. Moreover, 15 formed the diazonium
species at a much faster rate than 13 (Figure 4a). After eight
hours, 15 had been nearly completely consumed whereas only
approximately 50% of 13 was consumed over 24 hours.
In conclusion, we have reported a water-soluble triaza-
butadiene system that reacts in a pH-dependent manner in
buffered solutions. We expect that its bond-cleaving reactivity
in combination with the liberation of a reactive species will
make this process suitable for the interrogation of complex
systems. For example, diazonium species have previously
been shown to be useful in biochemical conjugations to
tyrosine.[23] Moreover, exploiting the diazonium ion into
phenol conversion (Figure 3c) opens the door to pH-depen-
dent fluorogenic probes and cargo-release strategies. The
current pH reactivity profile likely precludes the level of
precision that we require for diazonium salt release, but
additional design elements, such as internal acids and bases or
enzymatically triggered variations, could enable the develop-
ment of a more broadly applicable reactivity profile.[24] These
studies are currently ongoing and will be reported in due
course.
Keywords: diazonium salts · kinetics · protonation ·
triazabutadienes
How to cite: Angew. Chem. Int. Ed. 2015, 54, 4051–4054
Angew. Chem. 2015, 127, 4123–4126
[23] a) B. M. Tracey, D. E. G. Shuker, Chem. Res. Toxicol. 1997, 10,
1378 – 1386; b) J. M. Hooker, E. W. Kovacs, M. B. Francis, J. Am.
Chem. Soc. 2004, 126, 3718 – 3719; c) J. Gavrilyuk, H. Ban, M.
Nagano, W. Hakamata, C. F. Barbas III, Bioconjugate Chem.
2012, 23, 2321 – 2328; d) L. Wang, V. Gruzdys, N. Pang, F. Meng,
X.-L. Sun, RSC Adv. 2014, 4, 39446.
[24] Preliminary experiments to confirm the reactivity of the tyrosine
analogue and the lack of thiol background reactivity have been
promising; see the Supporting Information.
[1] Y. Modis, S. Ogata, D. Clements, S. C. Harrison, Nature 2004,
427, 313 – 319.
[2] C. D. Blanchette, Y. H. Woo, C. Thomas, N. Shen, T. A. Sulchek,
A. L. Hiddessen, PLoS One 2009, 4, e6056.
[3] J. Han, K. Burgess, Chem. Rev. 2010, 110, 2709 – 2728.
[4] a) J. Kalia, R. T. Raines, Angew. Chem. Int. Ed. 2008, 47, 7523 –
7526; Angew. Chem. 2008, 120, 7633 – 7636; b) J. Z. Du, X. J. Du,
C. Q. Mao, J. Wang, J. Am. Chem. Soc. 2011, 133, 17560 – 17563.
[5] E. H. Cordes, H. G. Bull, Chem. Rev. 1974, 74, 581 – 603.
[6] A. Luong, T. Issarapanichkit, S. D. Kong, R. Fonga, J. Yang, Org.
Biomol. Chem. 2010, 8, 5105 – 5109.
Received: November 20, 2014
Published online: February 6, 2015
4054 www.angewandte.org
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 4051 –4054