Sulfenamides as prodrugs of NH-acidic compounds: A new prodrug option for the amide bond

Article abstract of DOI:10.1016/j.bmcl.2007.06.037

The objective of this report is to introduce the novel concept of utilizing sulfenamides as prodrugs for compounds containing an NH-acidic functionality, particularly weakly acidic amide-type functionalities (amides, ureas, carbamates, etc.). Included are the syntheses and physicochemical characterizations of some model sulfenamides to illustrate the promise of this new prodrug technology.

Full text of DOI:10.1016/j.bmcl.2007.06.037

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4910–4913
Sulfenamides as prodrugs of NH-acidic compounds: A new
prodrug option for the amide bond
Victor R. Guarino, Veranja Karunaratneà and Valentino J. Stella*
Department of Pharmaceutical Chemistry, The University of Kansas, Lawrence, KS 66047, USA
Received 3 May 2007; revised 7 June 2007; accepted 8 June 2007
Available online 13 June 2007
Abstract—The objective of this report is to introduce the novel concept of utilizing sulfenamides as prodrugs for compounds con-
taining an NH-acidic functionality, particularly weakly acidic amide-type functionalities (amides, ureas, carbamates, etc.). Included
are the syntheses and physicochemical characterizations of some model sulfenamides to illustrate the promise of this new prodrug
technology.
ꢀ 2007 Elsevier Ltd. All rights reserved.
A continual challenge in Drug Discovery is balancing
the time dedicated to lead optimization against the ulti-
mate goal of quickly identifying a suitable candidate for
development. Prodrugs1,2 can reduce this lead optimiza-
tion time, particularly when a discovery team has iden-
tified a molecule that meets most of the team’s
selection criteria, but presents a significant obstacle to
drug delivery due to limitations in properties such as sol-
ubility, dissolution rate, permeability, stability, etc. A
well-designed prodrug is a non-toxic, chemically modi-
fied derivative of the drug that is easier to deliver; but
once delivered, will reconvert, generally in a rapid and
quantitative fashion, to the drug so that it can exert its
therapeutic action. Designing a well-behaved prodrug
can be a challenging task, so the more viable options
that are available to the medicinal chemist, the greater
the chance for success. Despite the number of prodrug
approaches previously attempted for NH-acids,3 none
have provided a reliable and generic way of making pro-
drugs that are hydrolytically stable, but rapidly and
quantitatively bioreversible; more specifically, the void
in synthetic options is particularly significant for the
weakly acidic NH-acids, such as amides, ureas, carba-
mates, etc. (pKa > 12–13).3,4 The objective of this report
is to introduce a new option which utilizes sulfenamides
as prodrugs of NH-acids (Scheme 1a). And while we
have evaluated sulfenamides of NH-acids possessing a
wide range of acidities,5 this report will focus solely on
sulfenamides of weakly acidic NH-acids, since that is
the sub-group in greatest need of new prodrug ideas.
A sulfenamide6 is identified by a N–S single bond where
the sulfur atom is bivalent. This single N–S bond is typi-
cally polarized with the partial positive charge residing on
the sulfur atom. The sulfur atom is susceptible to nucleo-
philic attack resulting in a displacement reaction that
O
O
a
S
R1 NH
R1
N
R
R2
R2
NH-acid
(drug)
Sulfenamide
(prodrug)
O
O
b
c
GSH
H2O
S
S SG
R1
R1
N
R
R
R1 NH
R
R2
R2
O
O
S
S OH
N
R1 NH
R
Keywords: Prodrug; Sulfenamide; NH-acid; Amide; Bioavailability;
Solubility; Dissolution rate; Permeability; Stability.
R2
R2
Sulfenic
Acid
*
Corresponding author. Tel.: +1 785 864 3755; fax: +1 785 864
5736; e-mail: stella@ku.edu
Scheme 1. (a) Formation of sulfenamide prodrug from an NH-acid
drug; (b) bioreconverion of sulfenamide prodrug by glutathione (GSH)
to form an NH-acid drug; (c) hydrolysis of a sulfenamide prodrug to
form an NH-acid drug and a sulfenic acid.
Present address: Discovery Pharmaceutics, Bristol-Myers Squibb,
PO Box 4000, Princeton, NJ 08543, USA.
à
Present address: Department of Chemistry, University of Perad-
eniya, Peradeniya, Sri Lanka.
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.06.037

V. R. Guarino et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4910–4913
4911
breaks the N–S single bond and forms an NH-acid prod-
O
uct. Sulfenamides are known to rapidly react with free thi-
ols, explaining their effectiveness as sulfenylating agents
in the synthesis of asymmetrical disulfides.7 Since free thi-
ols are ubiquitous in vivo (cysteine, glutathione (GSH),
etc.), a sulfenamide prodrug should conceivably regener-
ate the parent NH-acid upon introduction into the blood,
as shown in Scheme 1b. Because the R-group in the pro-
moiety is not explicitly defined, this approach would al-
low introduction of either hydrophilicity or lipophilicity
to the parent drug, depending on the properties of the
chosen R-group, allowing great flexibility in prodrug de-
sign. Furthermore, the attachment of the promoiety tem-
porarily removes a hydrogen bond donor and adds steric
bulk via the promoiety, both of which could increase the
solubility through disruption of crystal packing interac-
tions in the solid state.8 Therefore with the appropriate
promoiety, it is possible to increase the lipophilicity and
the aqueous solubility simultaneously, which could have
a dramatic impact on oral bioavailability.
R1 NH
R2
O
R
S
S
SO2Cl2
a
Cl
R1
N
R
S
R
S
R
R2
b
Scheme 2. Generic synthetic pathway for sulfenamides. Reagents and
conditions: (a) THF, 0 ꢁC for 10 min; (b) TEA, THF, 0 ꢁC for 20 min,
rt for 60 min. (Note. TEA not needed if salt of NH-acid is used.)
The immediate hydrolysis products of a sulfenamide are
expected to be the corresponding NH-acid and sulfenic
acid components (Scheme 1c). Due to the general insta-
bility of sulfenic acids,11 the N–S hydrolysis was fol-
lowed by monitoring the loss of sulfenamide and the
formation of the corresponding NH-acid, using
HPLC/UV. Due to good hydrolytic stability, the pH-
rate profile for degradation of 1 to form benzamide
was determined at 50, 60, and 70 ꢁC.5 At a given pH
and temperature, the hydrolysis of 1 followed pseudo-
first-order kinetics. The pH-rate profile for 1 at 25 ꢁC
(Fig. 1) was projected by extrapolation from the ele-
vated temperature results using the Eyring equation,
and is shown as a U-shaped profile with a maximum
hydrolytic stability at ꢀpH 6. Assuming the zwitterionic
form of 1 was the predominant species degrading over
the pH range studied, the projected points in Fig. 1 were
fit12 to the following kinetic model (Eq. 1), which is a
summation of acid-catalyzed (kH[H+]), base-catalyzed
(kOHKw/[H+]), and uncatalyzed (ko) hydrolytic mecha-
nisms (Kw = 1.0 · 10ꢁ14 at 25 ꢁC). From this mathemat-
ical fit, values were obtained for kH (5.37( 0.12) ·
10ꢁ2 minꢁ1 Mꢁ1), ko (3.38( 1.99) · 10ꢁ8 minꢁ1) and
To evaluate whether both hydrophilic and lipophilic
sulfenamides of various weakly acidic NH-acids can be-
have as prodrugs, 1, 2, and 3–4 were synthesized as
model sulfenamide prodrugs of benzamide (amide; mod-
el compound), carbamazepine (urea; drug example), and
linezolid (amide; drug example), respectively, and were
evaluated for aqueous chemical stability and reconver-
sion kinetics.5,9 To our knowledge, this work5 represents
the first formal aqueous pH-stability characterization
for sulfenamides. Procedures for making sulfenamides
have been established in the synthetic literature.6 The
general synthetic pathway utilized for 1–4 is shown in
Scheme 2. In this approach, the appropriate disulfides
were converted to sulfenyl chlorides in situ through reac-
tion with sulfuryl chloride, and these sulfenyl chlorides
were further reacted, without purification, with either
the free NH-acid (in the presence of triethylamine), or
the salt of the NH-acid, to form the sulfenamide prod-
uct. Although this is a standard synthetic procedure
for making sulfenamides, there have been procedural
modifications using thiophthalimides instead of sulfenyl
halides that, while not used in this work, have resulted in
cleaner reactions and higher yields.10
-3
10
-4
10
-5
10
O
O
O
O
O
O
S
N
N
NH3
HN S
NH2
H
-6
10
1
2
O
F
F
O
N
N
O
O
O
-7
10
N
N
O
O
O
N
3.5
4.5
5.5
pH
6.5
7.5
N
S
S
Figure 1. pH-rate profiles of 1 (d) and 2 (s) at 25 ꢁC in 35 mM
buffered solutions (ionic strength of 0.15 using NaCl). The dashed and
solid lines represent mathematical fits for 1 (Eq. 1) and 2 (Eq. 5),
respectively.
O
O
3
4

4912
V. R. Guarino et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4910–4913
kOH (11.6( 0.7) minꢁ1 Mꢁ1). Compound 1 is a very sta-
ble sulfenamide with regard to hydrolysis, as shown by
its maximum projected half-life of ꢀ6.3 years at pH
6.0 and 25 ꢁC (Table 1). However, a dramatic accelera-
tion in N–S degradation was seen when cysteine was
added into a 23.3 lM, pH 5.5, solution of 1, where the
final cysteine (Cys) concentration was 233 lM, a 10-
times excess13 relative to 1. Because the observed pseu-
do-first order degradation rate constant in the presence
of Cys ðkꢂobs ¼ 1:39 ꢃ 10ꢁ2 minꢁ1Þ was 4–5 orders of
magnitude greater than the projected hydrolysis rate
constant in the absence of Cys (kobs = 2.34 ·
10ꢁ7 minꢁ1), the Cys-based degradation profile could
be mathematically fit according to Eq. (2), allowing
for a calculation of kcys (59.7 minꢁ1 Mꢁ1) at pH 5.5
and 25 ꢁC.
4.0–5.5, followed by subsequent projection to 25 ꢁC
using the Eyring equation. For most of the pH range
studied, the predominant hydrolysis reaction for 2 was
actually ethyl ester hydrolysis in the promoiety (deter-
mined by LC/MS), not N–S bond hydrolysis. Only at
the very low end of the pH range did the ethyl ester be-
come relatively stable enough such that N–S bond
hydrolysis became the predominant reaction. Over the
pH range studied, it was assumed that the protonated
form of 2 was the predominant species undergoing deg-
radation, and a kinetic model (Eq. 3) was developed
describing two basic reactions: a base-catalyzed ethyl es-
ter hydrolysis (kOEtKw/[H+]) and an uncatalyzed sulfena-
mide hydrolysis (kNS). The fractional term (f+),
describing the protonated amount of 2 (Eq. 4), can be
substituted into Eq. (3) to get Eq. (5), which is a more
explicit form of Eq. (3) containing the acid dissociation
constant for the amine conjugate acid (Ka). Using Eq.
(5) to fit12 the pH-rate profile for 2, values for kOEt
(6.85( 0.49) · 103 Mꢁ1minꢁ1), kNS (2.15( 0.67) · 10ꢁ6
minꢁ1), and Ka (1.92( 0.27) · 10ꢁ7) were calculated.
kOHKw
½Hþꢄ
kobs ¼ kH½Hþꢄ þ ko þ
ð1Þ
kꢂobs ¼ kcys½Cysꢄ
ð2Þ
ꢀ
ꢁ
kOEtKw
kobs ¼ f ꢃ kNS
þ
ð3Þ
ð4Þ
ð5Þ
Although not tested, it is probable that the assumed
Cys-based thiolysis of 1 is driven mainly by the thiolate
fraction present in solution, and therefore would be ex-
pected to occur even faster at pH 7.4 since the concen-
tration of thiolate in solution would be ꢀ1–2 orders of
magnitude greater than at pH 5.5. Based on these re-
sults, 1 certainly possesses the desired duality of good
chemical stability for formulation purposes plus fast
reconversion kinetics to form the parent in the presence
of a thiol (cysteine in this case), and therefore supports
the notion that sulfenamides can behave as successful
prodrugs of weakly acidic NH-acids.
þ
½Hþꢄ
½Hþꢄ
f ¼
þ
½Hþꢄ þ Ka
kNS½Hþꢄ þ kOEtKw
½Hþꢄ þ Ka
kobs
¼
Given the ethyl ester side reaction that was predominant
over the majority of the pH range, 2 is certainly not the
ideal model compound to directly evaluate N–S hydro-
lytic stability. But it can be indirectly gathered that over
the majority of the pH range studied, the N–S bond was
at least more stable than the ethyl ester functionality in
the promoiety; furthermore, at pH 4.0, where the degra-
dation was predominantly driven by N–S hydrolysis, the
half-life of 2 was projected (from elevated temperature
studies) to be ꢀ180.7 days at 25 ꢁC (Table 1), which is
very good aqueous stability for most formulation pur-
poses, thereby providing a second example that sulfena-
mides designed from weakly acidic NH-acids can exhibit
sufficient hydrolytic stability for formulation purposes.
Furthermore, due to the flexibility inherent in the sulfe-
namide approach, it should be fairly straightforward to
design a more hydrolytically stable carbamazepine-
based sulfenamide by choosing a promoiety that does
not contain labile groups such as the ethyl ester in 2.
Compound 2 was also evaluated for aqueous stability
over the same pH range as 1 (Fig. 1).5 It is important
to note that the promoiety in 2 is the ethyl ester version
of the cysteinyl promoiety in 1. For a given non-ioniz-
able NH-acid, it is conceivable that this esterified cys-
teinyl promoiety would lend improved dissolution rate
and solubility properties to the prodrug, relative to the
non-esterified cysteinyl promoiety found in 1, since the
latter would be predominantly in the zwitterionic state
across the great majority of the physiologically relevant
pH range. The degradation of 2 was followed at 25 ꢁC
for pH 6.0–7.0 and at elevated temperatures for pH
Table 1. Projected and actual observed hydrolysis rate constants and
half-lives for compounds 1 and 2 at 25 ꢁC
pH
Compound 1
kobs (minꢁ1 · 10ꢁ6
Compound 2
t1/2 (yr) kobs (minꢁ1 · 10ꢁ5
)
t1/2 (day)
Finally, the two linezolid-based sulfenamide prodrugs,
3–4, were evaluated for reconversion kinetics in dog
whole blood.5 While a formal pH-stability study was
not conducted on these sulfenamides, they were cer-
tainly stable enough to conduct, without significant
hydrolytic interference, both whole blood reconversion
experiments and MDCK (Madin–Darby canine kidney)
cell transport permeability studies,5 the latter of which is
not described in this communication. When spiked into
37 ꢁC pre-heated Beagle dog whole blood to form an
)
4.0 5.32
4.5 1.85
5.0 0.559
5.5 0.234
6.0 0.208
0.25
0.71
2.36
5.64
6.33
—
0.266
0.437
0.870
2.58
5.40a
14.1a
23.4a
180.7
110.2
55.4
18.7
8.9a
6.5 —
7.0 1.20
3.4a
1.10
2.1a
a Actual 25 ꢁC data. All other data in table is projected from elevated
temperatures.

V. R. Guarino et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4910–4913
4913
7. Harpp, D. N.; Back, T. G. J. Org. Chem. 1971, 36, 3828.
8. (a) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.;
Feeney, P. J. Adv. Drug Del. Rev. 1997, 23, 3; (b) Lipinski,
C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv.
Drug Del. Rev. 2001, 46, 3.
initial concentration of ꢀ40 lM, compounds 3 and 4
yielded equimolar amounts of linezolid in essentially
an instantaneous and quantitative fashion. In fact, the
conversion to linezolid was complete for both 3 and 4
before the first spiked blood sample could be sufficiently
mixed and subsequently quenched (<2 min).14 This ra-
pid and quantitative reconversion is highly desired for
the vast majority of prodrug uses, and provides strong
support for these compounds performing as prodrugs
in vivo, particularly since the in vivo reconversion for
sulfenamide prodrugs is expected to be chemically dri-
ven by reaction with thiols, not enzymatic in nature.
1
9. Compound 1: H NMR (DMSO-d6) d 3.00–3.12 (m, 1H),
3.20–3.28 (m, 1H), 3.40–4.20 (br), 4.07–4.20 (br m, 1H),
7.46 (t, 2H), 7.53 (t, 1H), 7.80 (d, 2H), 8.30–8.50 (br), 9.65
(s, 1H); HRMS (FAB+, PEG300) m/z calcd for
C10H13N2O3S (M+H)+ 241.0647, found 241.0643.Com-
pound 2: 1H NMR (DMSO-d6) d 1.15 (t, 3H), 2.83 (m,
1H), 3.05 (m, 1H), 4.02 (br m, 1H), 4.15 (q, 2H), 6.95 (s,
2H), 7.08 (s, 1H), 7.33 (m, 2H), 7.41 (br m, 6H), 8.40 (br
s); HRMS (FAB+, PEG300) m/z calcd for C20H22N3O3S
(M+H)+ 384.1382, found 384.1369.Compound 3: 1H
NMR (CDCl3) d 1.21 (t, 3H) 2.35 (s, 3H), 2.62 (t, 2H),
3.00 (m, 4H), 3.13 (m, 2H), 3.63 (dd, 1H), 3.67 (dd, 1H),
3.82 (m, 4H), 3.98 (t, 1H), 4.07 (m, 1H), 4.14 (q, 2H) 4.86
(m, 1H), 6.88 (t, 1H), 7.05 (dd, 1H), 7.40 (dd, 1H); HRMS
(FAB+, PEG600) m/z calcd for C21H28F1N3O6S (M)+
469.1683, found 469.1684. Compound 4: 1H NMR
(CDCl3) d 2.35 (s, 3H), 3.00 (m, 4H), 3.70 (dd, 1H), 3.83
(m, 4H), 3.88 (dd, 1H), 3.98 (t, 1H), 4.07 (dd, 1H), 4.88 (m,
1H), 6.88 (t, 1H), 7.07 (t, 3H), 7.22 (t, 1H), 7.35 (m, 3H);
HRMS (ES+, 99% methanol, 5mM ammonium acetate)
m/z calcd for C22H25F1N3O4S (M+H)+ 446.1550, found
446.1545.
While there remains the need to evaluate the pharmaco-
kinetics and toxicology of some selected sulfenamide
prodrugs, some of which is being explored, the results
presented in this communication provide strong support
for sufenamides successfully behaving as prodrugs of
weakly acidic NH-acids. As mentioned previously, there
is a significant void in the synthetic literature for design-
ing prodrugs of weakly acidic NH-acids, and it appears
as if sulfenamides could fill this current need in the
medicinal chemist’s portfolio of prodrug technologies.
10. (a) Harpp, D. N.; Back, T. G. Tetrahedron Lett. 1971, 52,
4953; (b) Woulfe, S. R.; Iwagami, H.; Miller, M. J.
Tetrahedron Lett. 1985, 26, 3891; (c) Liu, L.; Tanke, R. S.;
Miller, M. J. J. Org. Chem. 1986, 51, 5332.
11. Behforouz, M.; Kerwood, J. E. J. Org. Chem. 1969, 34, 51.
12. SigmaPlot 2004 version 9.01. Fit kobs vs pH with a
References and notes
1. Prodrugs: Challenges and Rewards; Stella, V. J., Borc-
hardt, R. T., Hageman, M. J., Oliyai, R., Maag, H., Tilley,
J. W., Eds.; Springer: New York, 2007; Parts 1 & 2.
2. Bundgaard, H. Design of Prodrugs; Elsevier: Amsterdam,
1985, pp 1–360.
weighting factor of 1/kobs
13. A 10 times excess of cysteine was used to create pseudo-
.
first order degradation conditions for 1 allowing a
straightforward calculation of kcys
3. Guarino, V. R.; Stella, V. J. In Prodrugs: Challenges and
Rewards; Stella, V. J., Borchardt, R. T., Hageman, M. J.,
Oliyai, R., Maag, H., Tilley, J. W., Eds.; Springer: New
York, 2007; Part 2, pp 133–187.
4. (a) Johansen, M.; Bundgaard, H. Arch. Pharm. Chem.,
Sci. Ed. 1979, 7, 175; (b) Bundgaard, H.; Johansen, M. Int.
J. Pharm. 1980, 5, 67; (c) Bundgaard, H.; Johansen, M.
Int. J. Pharm. 1984, 22, 45.
5. Guarino, V. R. Ph.D. Dissertation; University of Kansas:
May 2004.
6. (a) Craine, L.; Raban, M. Chem. Rev. 1989, 89, 689; (b)
Koval, I. V. Russian Chem. Rev. 1990, 59, 836; (c) Davis,
F. A. Int. J. Sulfur Chem. 1973, 8, 71; (d) Koval, I. V.
Russian J. Org. Chem. 1996, 32, 1239.
.
14. Beagle dog whole-blood experiment. For a given com-
pound (3, 4 or linezolid), 15 lL of an 8 mM DMSO
stock solution was added to 3 mL blood (37 ꢁC) and
vortexed for 5 s. A 400 lL aliquot of spiked blood was
mixed with 400 lL chilled acetonitrile and 100 lL
chilled saturated aqueous (NH4)2SO4 solution. The
mixture was vortexed for 5 s and centrifuged for
30 s. The supernatant was collected and immediately
analyzed by HPLC. Control experiments showed that
3, 4, and linezolid were chemically stable in the
supernatant fraction over the time course of the
sample preparation and analysis.