Hydrophobic and electronic factors in the design of dialkylglycine decarboxylase mimics

Article abstract of DOI:10.1016/j.bmc.2005.05.019

The first functional catalytic mimic of the enzyme dialkylglycine decarboxylase is described. This system utilizes a hydrophobically modified polyethylenimine polymer, a pyridoxamine cofactor, and a 2-aryl-2-alkylglycine sacrificial amine source to convert α-keto acids to α-amino acids at biologically relevant temperatures with multiple turnovers of the pyridoxamine catalyst. The effects of hydrophobic and electronic factors in the 2,2-disubstituted sacrificial amine source and the pyridoxamine catalyst on turnover frequency and turnover number are explored.

Full text of DOI:10.1016/j.bmc.2005.05.019

Bioorganic & Medicinal Chemistry 13 (2005) 5873–5883
Hydrophobic and electronic factors in the design
of dialkylglycine decarboxylase mimics
Jason J. Chruma, Lei Liu, Wenjun Zhou and Ronald Breslow*
Columbia University, Department of Chemistry, 3000 Broadway, New York, NY 10027, USA
Received 20 April 2005; accepted 10 May 2005
Available online 13 June 2005
Abstract—The first functional catalytic mimic of the enzyme dialkylglycine decarboxylase is described. This system utilizes a
hydrophobically modified polyethylenimine polymer, a pyridoxamine cofactor, and a 2-aryl-2-alkylglycine sacrificial amine source
to convert a-keto acids to a-amino acids at biologically relevant temperatures with multiple turnovers of the pyridoxamine catalyst.
The effects of hydrophobic and electronic factors in the 2,2-disubstituted sacrificial amine source and the pyridoxamine catalyst on
turnover frequency and turnover number are explored.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Dialkylglycine decarboxylase (DGD) is unique among
pyridoxamine 50-phosphate (PMP)-dependent transam-
inase and decarboxylase enzymes. Unlike typical trans-
aminase enzymes, which utilize a-monosubstituted
amino acids as sacrificial amine sources, DGD regener-
ates the PMP cofactor from the corresponding aldehyde,
pyridoxal 50-phosphate (PLP), by an oxidative decarbox-
ylation of a 2,2-dialkylglycine substrate (Scheme 1).1 In
contrast to normative decarboxylase enzymes, in which
Ca protonation is greatly favored, protonation in DGD
occurs preferentially at the benzylic C-40 position, to af-
ford a ketone and PMP after hydrolysis or transimina-
tion. Given its unique activity, DGD has experienced
significant mechanistic and crystallographic examination
in the past 10 years.2 Distinctive to DGD among other
PMP/PLP-dependent enzymes is the lack of an observa-
ble quinoid intermediate throughout the catalytic cycle.3
As part of a long-lived initiative toward the mimicry of
various pyridoxamine-dependent enzymatic systems
(e.g., transaminases,4 racemases,4e,5 tryptophan syn-
thases,6 and decarboxylases4e,7), we recently reported
the first functional mimic of DGD utilizing a laurylated
polyethylenimine polymer catalyst [PEI-C12 (8.7%)] and
a hydrophobically modified pyridoxamine cofactor (1)
in aqueous reaction conditions (Scheme 2).8 Employing
Scheme 1. PMP/PLP catalytic cycle in dialkylglycine decarboxylase.
2-phenyl-2-methylglycine (Pmg) as a sacrificial amine
source, we obtained record rate accelerations (up to
725,000-fold) and turnover numbers (up to 100 turn-
overs) for a non-enzymatic pyridoxamine-mediated
transamination.
While our initial studies with the polymer-catalyzed,
1-mediated transamination of a-keto acids to a-amino
acids demonstrated remarkably fast reaction rates in
*
Corresponding author. Tel.: +212 854 2170; fax: +212 854 2755;
e-mail: rb33@columbia.edu
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.05.019

5874
J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
Figure 1. Putative internal aldimine intermediate 2.
intermediate 2 (Fig. 1); therefore, increasing the rate of
decarboxylation should lead to higher catalyst turnover
frequencies. Electronic factors in the sacrificial amine
source and pyridoxal cofactor, as well as hydrophobic
interactions between the reagents and the polymeric
cocatalyst in the aqueous reaction mixture should have
a dramatic impact on the rate of decarboxylation.
Scheme 2. DGD mimetic system exploiting hydrophobic interactions.
PEI-C12 (8.7%) is commercial branched polyethylenimine22 in which
8.7% of the amines are laurylated and the remaining amines are
permethylated.8
We previously observed8 that electron-rich 2,2-diphenyl-
glycine is too reactive in the polymer-buffered reaction
conditions, reacting directly with the a-keto acid
substrate and obviating the hydrophobically modified
pyridoxamine catalyst.8 2-Aminoisobutyric acid (2,2-
dimethylglycine), in contrast, did not decarboxylate
appreciably under our reaction conditions. Given our
success with Pmg as a sacrificial amino source, we syn-
thesized a small set of 2-aryl-2-alkylglycines to further
investigate electronic and hydrophobic influences on
the rate of pyridoxal-mediated decarboxylation. These
a,a-disubstituted amino acids were obtained in short
order from the corresponding ketones employing a mod-
ified Bucherer–Bergs reaction,15 followed by hydrolysis
of the resulting hydantoins (Scheme 3). The amino acid
products were purified, generally by ion-exchange chro-
matography, and dissolved in slightly acidic deionized
water to afford 0.1 M stock solutions.16
comparison to pyridoxamine in water alone, the corre-
sponding reverse reaction did not proceed appreciably.
Other groups have reported similar difficulties in regen-
erating the pyridoxamine cofactor from pyridoxal using
monosubstituted a-amino acids under aqueous condi-
tions.9 This reverse reaction can be promoted, with vary-
ing degrees of success, by very high concentrations (1 M!)
of the a-monosubstituted sacrificial amine source,13a
high temperatures,10 the addition of metal10,11 or imi-
dazolium salts,12 incorporation of the pyridoxamine
cofactor into a protein environment,13 or quaternization
of the pyridoxamine/pyridoxal cofactor pyridine nitro-
gen.14,23 Presumably, all of these strategies act to reduce
the resonance stabilization of the pyridoxal aldehyde/im-
ine with the pyridine ring of the cofactor. Utilizing a,
a-disubstituted amino acid, as in our DGD-mimetic sys-
tem, represents a unique strategy to regenerate the pyri-
doxamine cofactor and obtain a truly catalytic reaction.
Attempts to synthesize either 2-(2-pyridyl)-2-methylgly-
cine or the 4-pyridyl derivative from the corresponding
hydantoins were unsuccessful, as these amino acids rap-
idly decomposed under both basic and acidic hydrolytic
conditions. 2-(3-Pyridyl)-2-methylglycine (3-Pmg)17 was
successfully obtained following the above protocol, but
it was indistinguishable from Pmg in both initial turn-
over frequency and turnover number when used as a
sacrificial amine source for our DGD-mimetic system
(data not shown).
Given our initial success, we sought to improve our DGD
mimic protocol to achieve faster turnover rates at lower,
more biologically relevant temperatures, potentially with
higher turnover numbers. In this paper, we describe the
influence of electronic and hydrophobic factors in the
2-aryl-2-alkylglycine sacrificial amine source on turnover
rate and number in our system. Furthermore, the impor-
tance of a cationic pyridoxamine cofactor is explored.
These studies indicate that enhancing hydrophobic inter-
actions between the sacrificial amine source and the poly-
mer–cofactor diad leads to greater fidelity and rate
acceleration in our DGD-mimetic protocol over strictly
electronic modifications. Additionally, quaternization of
the pyridine nitrogen of the hydrophobically modified
pyridoxal cofactor by alkylation greatly increases the rate
of cofactor-mediated decarboxylation.
2. Results and discussion
2.1. Synthesis of 2-aryl-2-alkylglycines
Scheme 3. Synthesis of 2-aryl-2-alkylglycines. Hydrolysis conditions:
(A) 2 N aqueous NaOH, D; (B) Ba(OH)2, H2O, D; (C) 6 N aqueous
HCl, D.
The rate determining step for our new catalytic system is
the decarboxylation of the putative internal aldimine

J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
5875
Table 1. Rate of 2-aryl-2-alkylglycine decarboxylation by pyridoxal
50-ethylthioether in water and 34.5% aqueous acetonitrile
2.2. Rate of decarboxylation of 2-aryl-2-alkylglycines and
influence on catalyst turnover frequency and number
The 2-aryl-2-alkylglycines described above were
designed to explore electronic and hydrophobic factors
in the rate of pyridoxal-mediated decarboxylation.
Whether decarboxylation and C-40 protonation proceed
in a step-wise fashion through the intermediacy of a res-
onance-stabilized anion or represent a less likely con-
certed event in our DGD mimic, electron-withdrawing
substituents on the sacrificial amine source (e.g.,
4-NO2Ph or 4-CF3Ph) are expected to increase the rate
of decarboxylation for intermediate 2. Increasing the
electron-withdrawing potential of the amino acid a-sub-
stituents, however, could have deleterious conse-
quences. For example, the aryl methyl ketones produced
from the PLP-mediated oxidative decarboxylation of the
2,2-disubstituted amino acids will be more activated
toward aldol condensation with the pyridoxal intermedi-
ate, thus killing the catalytic cycle; we have observed
that the hydrophobically modified polyethylenimine
polymers can accelerate such aldol condensations (data
not shown). Electron-donating substituents (e.g.,
4-MeOPh), on the other hand, should retard the rate
of decarboxylation, but exhibit longer-lived catalytic
cycles due to slower ketone-induced catalyst
decomposition.
Amino k (· 10ꢀ3 minꢀ1
acid
)
krelative w/34.5% MeCN k kMeCN/k
(· 10ꢀ3 minꢀ1
)
Pmg 3.2 1.4
4-Tmg 6.2 4.3
4-Nmg 57.0 1.7
1.0
1.9
39.5 0.7
66.8 1.4
123.3 1.0
24.5 0.8
47.4 1.3
46.5 1.8
12.3
10.8
2.2
17.3
0.4
4-Mmg
Pxg
1-Aic
1.2 4.1
4.6 3.5
2.6 1.3
20.4
10.3
17.9
1.4
0.8
Reaction conditions: 1.0 · 10ꢀ4 M pyridoxal 50-ethylthioether,
1.0 · 10ꢀ3 M amino acid, 2.0 · 10ꢀ3 M EDTA, 2.5 · 10ꢀ2 M Hepes,
2.5 · 10ꢀ2 M KCl, T = 60 ꢁC, pH 7.0.
group (i.e., 4-Nmg) exhibits over a 17-fold effect. Inter-
estingly, replacing the 2-methyl substituent with an elec-
tron-withdrawing trifluoromethyl group19 resulted in
almost no observed decarboxylation, presumably due
to the significant decrease in amine nucleophilicity and
thus slower imine formation between the amino acid
and pyridoxal reagents (data not shown). In the absence
of the modified PEI polymer, increasing the hydropho-
bicity of the amino acid was not expected to greatly
influence the rate of decarboxylation. Indeed, the more
hydrophobic 2-phenyl-2-hexylglycine (Pxg) decarboxyl-
ated at a rate nearly identical to Pmg. Joining the
2-phenyl and 2-alkyl groups into a constrained five-
membered ring (i.e., 1-Aic) also displayed little effect
Hydrophobic factors are also expected to play a signifi-
cant role. Lengthening the alkyl substituents on the sac-
rificial amine source should increase hydrophobic
interactions with the modified polymer, and the result-
ing increase in local concentration should afford faster
rates of decarboxylation. A similar phenomenon was
observed in the rate of pyridoxamine-mediated trans-
aminations when the hydrophobicity of either the pyri-
doxamine cofactor or a-keto acid substrate was
increase, with the former being more influential.8 Fur-
thermore, increasing the length, and thus steric bulk,
of the alkyl substituent should reduce the rate of aldol
condensation between the resulting ketone byproduct
and the pyridoxal cofactor.
on the apparent kdecarboxylation
.
Given previous studies indicating that reducing the polar-
ity of the reaction medium can greatly enhance decarbox-
ylation reactivity,8,20 we also determined kdecarboxylation of
the 2-aryl-2-alkylglycines in 34.5% aqueous acetonitrile.
This relatively modest reduction in solvent polarity
afforded an average 12-fold increase in kdecarboxylation
.
Preliminary amino acid decarboxylation studies were
conducted in the absence of the modified PEI polymer
employing the fully water-soluble pyridoxal 50-ethyl
thioether4a (Table 1). The reactions were monitored by
UV–vis spectroscopy at 60 ꢁC,18 and pseudo-first-order
rates of decarboxylation (kdecarboxylation) were computed
following the exponential decay of the characteristic
pyridoxal/pyridoximine absorbance at 400 nm.4a To
prevent any adventitious metal ion catalysis, the decar-
boxylation reactions were performed in the presence of
2.0 · 10ꢀ3 M EDTA. As predicted, 2-aryl-2-alkylgly-
cines with electron-withdrawing groups on the aromatic
ring (i.e., 4-Tmg and 4-Nmg) showed marked increases
in kdecarboxylation over Pmg, while an electron-donating
methoxy moiety (i.e., 4-Mmg) retarded the rate of decar-
boxylation. The strength of the electron-withdrawing
aromatic substituent directly translates into increases
in kdecarboxylation; a para-trifluoromethyl group (i.e.,
By far, 2-(4-nitropheny)-2-methylglycine (4-Nmg) decar-
boxylated fastest of all the amino acids studied. In fact,
the kdecarboxylation for 4-Nmg in 34.5% acetonitrile
[(123.3 1.0) · 10ꢀ3 minꢀ1] is nearly identical to the
fastest rate of pyruvate transamination observed for our
1/PEI-C12 (8.7%) system [(3.9 0.2) · 10ꢀ2 minꢀ1].8
The above decarboxylation studies strongly support the
proposal that incorporation of electron-withdrawing
groups on the aromatic substituent of the 2-aryl-2-alkyl-
glycine sacrificial amine source should increase the turn-
over frequency of the DGD mimic catalytic cycle. To
test this hypothesis, the conversion of pyruvate
(5.0 · 10ꢀ3 M) to alanine in the presence of PEI-C12
(8.7%) (2.5 · 10ꢀ5 M), EDTA (2.0 · 10ꢀ3 M), a catalytic
amount of 1 (2.5 · 10ꢀ5 M), and either Pmg, 4-Nmg, or
4-Mmg (1.0 · 10ꢀ2 M) at pH 7–8 and 60 ꢁC was moni-
tored by previously reported HPLC techniques (Table
2).8 The initial turnover frequency (TOF), the number
4-Tmg) affords a nearly twofold increase in kdecarboxylation,
while the stronger electron-withdrawing para nitro

5876
J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
Table 2. Catalytic turnover frequency and number with pyridoxamine
1 and 2-aryl-2-alkylglycines
vert to a protonated, that is pyridinium, form. Common
arguments dictate that this pyridinium formation may
serve as an ꢀelectron-sinkꢁ to accelerate decarboxylation.
However, such a directed protonation is not feasible
under the polymer-buffered reaction conditions of the
DGD mimic described above. A similar stable pyridi-
nium species can be generated, though, by alkylation
of the pyridine nitrogen. To this end, Murakami and
coworkers reported catalytic turnover for a transam-
inase mimic composed of a quaternized pyridoxamine
in a lipid micelle at room temperature, though in the
presence of copper salts.23 Furthermore, 10-fold in-
creases in enzymatic activity have been reported by
Gong and coworkers when employing N-methyl-PLP
as a catalyst for decarboxylases.24 On the other hand,
Bach and coworkers, following high-level enzyme
active-site computational modeling, found no computa-
tional evidence for an increase in kdecarboxylation due to
pyridinium formation by protonation or methylation.25
To further explore empirically the effect of pyridinium
formation in our DGD-mimetic system, we synthesized
the pyridinium salts 8 and 15, representing quaternized
versions of 1 and the corresponding pyridoxal cofactor,
respectively.
Amino acid TOF (hꢀ1 a
)
TOFrelative TON
TONrelative
41.2 7.5 1.0
7.2 1.9 0.2
40.3 0.3 1.0
Pmg
4-Nmg
4-Mmg
0.83
P1.3
0.78
1.0
P1.6
0.9
a Initial turnover frequency (TOF) = number of turnovers in 2 h/2 h.
Reaction conditions: 2.5 · 10ꢀ5 M 1, 2.5 · 10ꢀ5 M PEI-C12 (8.7%),
5.0 · 10ꢀ3 M pyruvic acid, 1.0 · 10ꢀ2 M amino acid, 2.0 · 10ꢀ3 M
EDTA, T = 60 ꢁC, pH 7.5.
of turnovers after 2 h of reaction divided by 2 h, employ-
ing 4-Nmg as a sacrificial amine source, was markedly
faster than that observed with Pmg or 4-Mmg, but the
overall number of turnovers (TON) was significantly
less than that obtained from Pmg (7.2 1.9 vs
41.2 7.5). Furthermore, a large amount of unidentified
peaks rapidly accumulated on the HPLC trace, compli-
cating the ability to determine accurately the TOF value.
As expected, similar results were observed when enan-
tiopure (S)-4-Nmg21 was employed. The multiple side
products and relatively low TON observed with
4-Nmg, despite the very rapid initial turnover of the pyr-
idoxamine catalyst, indicates that, in the presence of the
polymer, the highly active 4-nitro-acetophenone pro-
duced from the oxidative decarboxylation of 4-Nmg is
reacting with the pyridoxal catalytic intermediate, thus
killing the catalytic cycle. The opposite phenomenon
was observed employing 4-Mmg as the sacrificial amine
source; the TOF was slightly less in comparison to Pmg
(0.78 vs 0.83 hꢀ1), but the TON was nearly identical and
very consistent. Importantly, for all three amino acids
tested, the corresponding phenethylamines were not
observed throughout the entirety of the reaction, sug-
gesting that oxidative decarboxylation (i.e., C-40 proton-
ation), and thus regeneration of catalyst 1, is preferred
exclusively over non-oxidative decarboxylation (i.e.,
Ca protonation) for this system.
Pyridoxamine 8 was synthesized starting from nucleo-
philic substitution of mesylate 3 by the thiolate gener-
ated from disulfide
4 and NaBH4 (Scheme 4).
Deprotection of the resulting phenol sulfate ester 5
afforded free phenol 6, which was then selectively
N-methylated with MeI. While phenolic O-methylation
may have occurred, we have observed that substituents
on the pyridoxamine/pyridoxal phenolic oxygen are
highly susceptible toward hydrolysis under standard
work-up conditions in the pyridinium species. Similarly,
phenolic O-alkylation of the pyridinium salts has proven
extremely difficult. Accordingly, Boc-protected pyridi-
nium salt 7 was obtained in good purity by simple evap-
oration of the solvent and excess iodomethane. Removal
of the Boc group with TFA followed by preparative
reverse-phase HPLC provided the fully elaborated
hydrophobic quaternized pyridoxamine 8.
It should be noted that the observed TOF and TON val-
ues are highly dependent on the particular batch of syn-
thetic PEI-C12 (8.7%) utilized, presumably due to the
inherent difficulties of producing identical batches of
modified polymers from the highly polydisperse unfunc-
tionalized commercial PEI.22 All of the results reported
in this paper were determined using the same batch of
PEI-C12 (8.7%).
2.3. Synthesis and evaluation of quaternized, hydro-
phobically modified pyridoxamine and pyridoxal cofactors
8 and 15
In enzymatic systems, the pyridine nitrogen of the PLP
cofactor is in close proximity to an appropriately placed
acidic amino acid side chain in the cofactor binding
pocket, suggesting that the cofactor may exist in or con-
Scheme 4. Synthesis of pyridoxamine 8.

J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
5877
was then N-methylated with MeI. Unlike the pyridox-
amine derivative 6, a significant amount of S-methyla-
tion accompanied formation of pyridinium 14,
necessitating purification of the protected aldehyde by
reverse-phase preparative HPLC. Conversion of the
dimethylacetal to the aldehyde with HCl in acetonitrile,
followed by another round of HPLC purification pro-
vided quaternized hydrophobic pyridoxal 15. The iden-
tity of the counteranions for both 8 and 15 was not
rigorously determined. Like their non-ionic counter-
parts, pyridinium salts 8 and 15 exhibit poor water-sol-
ubility. They were therefore dissolved in acetonitrile to
afford ca. 5.0 · 10ꢀ3 M solutions.
Pyridinium 8 converted pyruvic acid to alanine at a rate
(ktransamination) only slightly faster than the non-quatern-
ized pyridoxamine 1 [(5.2 1.4) · 10ꢀ2 minꢀ1 at 20 ꢁC
vs (3.5 0.2) · 10ꢀ2 minꢀ1]. Furthermore, competition
reactions run with 30:30:1 (pyruvic acid–phenylpyruvic
acid–8) ratios at 20 ꢁC, pH 7.5, and in the presence of
PEI-C12 (8.7%) in water revealed alanine:phenylalanine
products ratios similar to those observed with 1
[1:(11 2) vs 1:(14 2)].
Scheme 5. Synthesis of pyridoxal 15.
While the effect of pyridine quaternization on
ktransamination was slight, pyridoxal 15 exhibited marked
increases in kdecarboxylation over a non-quaternized ver-
sion, permitting the decarboxylation reactions to pro-
ceed smoothly at 30 ꢁC. This apparently contrasts with
the computational predictions of Bach and coworkers,25
and represents the first time we have observed substan-
tial cofactor-mediated amino acid decarboxylation at a
biologically relevant temperature. To insure that reagent
15 was fully solubilized, initial decarboxylation studies
were performed in 58% aqueous acetonitrile (Conditions
A, Table 3). Heterogeneous reaction mixtures were still
observed, however, when Pxg and 4-Tmg were utilized
as sacrificial amine sources, possibly accounting for
the lower than expected kdecarboxylation values observed
for these two amino acids. Regardless, kdecarboxylation
Synthesis of the quaternized pyridoxal 15 proved to be
more complicated (Scheme 5). Thioether 10 was ob-
tained from benzyl alcohol 9 and dithiane 4 employing
PBu3 in THF. The acetal protecting group was then
hydrolyzed under acidic conditions to afford benzyl
alcohol 11. After we screened a variety of oxidative con-
ditions, MnO2 proved to be the best selective reagent to
convert alcohol 11 to pyridoxal 12. Attempts to oxidize
the benzylic alcohol after pyridinium formation led to
an intractable mixture of compounds. Likewise, reacting
pyridoxal 12 with MeI to directly form the pyridinium
salt 15 afforded only complex mixtures. Therefore, a
protection–alkylation–deprotection
strategy
was
employed. Specifically, the aldehyde moiety of 12 was
masked as the corresponding dimethylacetal 13, which
Table 3. Rate of 2-aryl-2-alkylglycine decarboxylation
Amino acid
Conditions
k (· 10ꢀ3 minꢀ1
)
krelative
Relative rate to S-Et-pyridoxal; at 60 ꢁC
Pmg
Pmg
A
B
A
B
A
B
A
B
A
B
29.1 1.2
24.4 2.3
31.9 3.3
83.3 5.4
24.1 1.1
19.7 2.3
24.0 0.7
67.6 5.0
42.1 1.2
30.5 2.7
1.2
1.0
1.3
3.4
1.0
0.8
1.0
2.8
1.7
1.3
9.1
7.6
4-Tmg
4-Tmg
4-Mmg
4-Mmg
Pxg
5.1
13.4
20.1
16.4
5.2
Pxg
1-Aic
1-Aic
14.7
16.2
11.7
Reaction conditions A: 1.0 · 10ꢀ4 M 15, 1.0 · 10ꢀ3 M amino acid, 2.0 · 10ꢀ3 M EDTA, 1.3 · 10ꢀ2 M Hepes, 1.3 · 10ꢀ2 M KCl in 58% aqueous
MeCN, T = 30 ꢁC, pH 7.0. Reaction conditions B: 1.0 · 10ꢀ4 M 15, 1.0 · 10ꢀ3 M amino acid, 1.25 · 10ꢀ5 M PEI-C12 (8.7%), 2.0 · 10ꢀ3 M EDTA,
2.5 · 10ꢀ2 M Hepes, 2.5 · 10ꢀ2 M KCl, T = 30 ꢁC, pH 7.0.

5878
J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
values for the 2-aryl-2-alkylglycines with 15 at 30 ꢁC
were an average ninefold faster than those observed with
pyridoxal 50-ethyl thioether at 60 ꢁC. Pyridinium 15 rap-
idly forms the corresponding hydrate of the aldehyde in
aqueous conditions, and thus does not exhibit the char-
acteristic absorption peak at 400 nm in the UV–vis spec-
trum. Therefore, decarboxylation reactions with 15 were
monitored by HPLC following the generation of the ke-
tone corresponding to amino acid oxidative decarboxyl-
ation. As before, none of the corresponding
phenethylamines were detected throughout the entirety
of the reactions.
catalyst 8, along with the marked production of catalyst
decomposition products. These turnover studies demon-
strate the particular advantage that hydrophobic effects
can have over electronic factors.
3. Conclusions
An excellent mimic of the enzyme dialkylglycine decar-
boxylase has been created using a pyridoxal derivative
and a modified polyethylenimine. While standard trans-
aminations are used in both directions by transaminase
enzymes, our work and that of others indicates that sim-
ple transaminations in one direction—converting a pyr-
idoxal and an amino acid to a pyridoxamine and a keto
acid—are not efficient in model systems. The decarbox-
ylation process with disubstituted glycines substitutes
nicely for this inefficient process, and its conversion of
a pyridoxal to a pyridoxamine then couples with our
previous studies of highly efficient reactions of a
pyridoxamine derivative with an a-keto acid to form
the a-amino acid and the pyridoxal derivative.
Even more remarkable results were obtained when the
decarboxylation reactions were carried out in the pres-
ence of PEI-C12 (8.7%) in water (Conditions B, Table
3). Unlike the results in 58% aqueous acetonitrile, all
of the reaction mixtures were apparently homogeneous
in the presence of the polymer. Furthermore, the ob-
served kdecarboxylation values at 30 ꢁC were an average
11-fold greater than those observed with pyridoxal
50-ethyl thioether at 60 ꢁC. 4-Tmg, which possesses an
electron-withdrawing trifluoromethyl group on the aro-
matic substituent, decarboxylated fastest under these
reaction conditions [(83.3 5.4) · 10ꢀ3 minꢀ1]. How-
ever, in the presence of the hydrophobically modified
polymer, the hydrophobic amino acid Pxg also decar-
boxylated at a remarkably fast rate [(67.6 5.0) ·
10ꢀ3 minꢀ1], supporting the previous hypothesis that
hydrophobic interactions between the polymer catalyst,
the sacrificial amine source, and the pyridoxal interme-
diate would increase the local concentration of the re-
agents and thus increase the reaction rate.
In our present study, the other product of the decarboxyl-
ation process, besides the pyridoxamine, is a ketone. This
is the same ketone that is used to synthesize the disubsti-
tuted glycine by reaction with NaCN and ammonium
carbonate, followed by hydrolysis of the resulting
hydantoin. Thus, the ketone can be considered a catalyst
that is regenerated, and the overall reductive amination
of the pyridoxal is formally accomplished with the con-
sumption of ammonium cyanide and aqueous base; in-
deed the ammonia is even regenerated in the
hydrolysis. The overall result is the catalytic conversion
of a keto acid to an amino acid with the consumption of
these simple inorganic materials. The pyridoxamine
derivative and the polymer are simply catalysts, but
for a very effective process. In addition, both parts of
the overall cycle mimic well-established biochemical
reactions.
It is important to note that the kdecarboxylation values ob-
served for 4-Tmg and Pxg are close to, if not greater
than the ktransamination value observed for pyruvic acid
with quaternized pyridoxamine 8, implying that decar-
boxylation may no longer be a rate limiting step, and
that faster catalytic turnover should be observed. In-
deed, the average initial turnover frequencies observed
with pyridinium 8 at 30 ꢁC is faster than with non-quat-
ernized cofactor 1 at 60 ꢁC (Table 4). Quaternization of
the pyridoxamine cofactor also appears to accelerate
catalyst decomposition, as evidenced by the lower
TON values observed for 8 versus 1. The highest TOF
and TON values were observed utilizing the hydropho-
bic sacrificial amine source Pxg. As with 4-Nmg and cat-
alyst 1, 4-Tmg displayed the lowest TON value with
We recently reported that covalent linkage of a pyridox-
amine to peptide-derived homochiral oligoamines affords
an effective reagent for the conversion of various a-keto
acids to the corresponding a-amino acids in moderate to
good enantiomeric excesses even though the reactions
were performed under aqueous conditions.26 However,
chiral oligoamine-cofactor studies have not yet been re-
ported. Since peptide-derived chiral polyamines can be
Table 4. Catalytic turnover frequency and number with pyridoxamine 8 and 2-aryl-2-alkylglycines
Amino acid
TOF (hꢀ1 a
)
TOFrelative
TON
TONrelative
Pxg
4-Tmg
4-Mmg
1.14
0.97
0.81
1.00
0.85
0.71
14.6 7.9
9.4 0.1
12.3 0.2
1.0
0.6
0.8
a Initial turnover frequency (TOF) = number of turnovers in 2 h/2 h. Reaction conditions: 2.5 · 10ꢀ5 M 8, 2.5 · 10ꢀ5 M PEI-C12 (8.7%), 5.0 · 10ꢀ3 M
pyruvic acid, 2.0 · 10ꢀ3 M EDTA, 1.0 · 10ꢀ2 M amino acid, T = 60 ꢁC, pH 7.5.

J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
5879
synthesized in a defined and homochiral fashion,26 they
could serve as replacements for PEI-C12 (8.7%) in our
DGD mimetic system, potentially alleviating the current
polymer batch-dependence issues and allowing for the
catalytic, biomimetic, and enantioselective synthesis of
a-amino acids from a-keto acids, 2-aryl-2-alkylglycine
sacrificial amines, and a hydrophobically modified pyri-
doxamine cofactor. These studies are currently underway,
and the results will be reported in due course.
4.1.4. 5-Phenyl-5-hexylhydantoin.28 White solid; MS
(CI+) m/z 261.03 (M+H)+; 1H NMR (d6-DMSO,
300 MHz) d: 10.76 (s, 1H), 8.66 (s, 1H), 7.50–7.29 (m,
5H), 2.05–1.84 (m, 2H), 1.22 (br s, 6H), 0.84 (t,
J = 6.5 Hz, 3H).
4.1.5. Spiro[imidazolidine-4,10-indan]-2,5-dione.29 White
solid; MS (CI+) m/z 203.14 (M+H)+; 1H NMR (d6-
DMSO, 300 MHz) d: 9.39 (br s, 1H), 8.41 (s, 1H),
7.30–7.02 (m, 4H), 3.06–2.96 (m, 2H), 2.56–2.46 (m,
1H), 2.13 (dt, J = 13.2 and 8.3 Hz, 1H).
4. Experimental
4.1. Synthesis
4.1.6. 2-(4-Nitrophenyl)-2-methylglycine (4-Nmg).23 5-
(4-Nitrophenyl)-5-methylhydantoin (517 mg, 2.2 mmol)
and 6 N HCl (22 mL) were combined in a sealed high-
pressure tube and the resulting slurry was stirred at
130 ꢁC for 67 h. After cooling to ambient temperature,
the mixture was filter through glass wool, concentrated,
and dried azeotropically from PhCH3. The resulting yel-
low solid was purified by reverse-phase HPLC (80 mg,
0.3 mmol, 15%): 1H NMR (d4-MeOH, 300 MHz) d:
8.32 (d, J = 8.9 Hz), 7.75 (d, J = 8.9 Hz), 1.97 (s, 3H).
All reagents were purchased from commercial sources
and utilized as received unless otherwise noted. DMF,
THF, CH2Cl2, and CH3CN were dried through individ-
ual alumina-based solvent purification columns. 2-Ami-
no-2-phenylpropionic acid (Pmg), purchased from
Aldrich Co., was purified by preparative RP-HPLC with
a Vydac Protein & Peptide C18 column. All reactions
were carried out under atmospheric conditions, unless
otherwise noted. Flash chromatography was performed
on 230–400 mesh silica (Silica Gel 60) from EM Science.
NMR spectra were obtained on a Bruker DPX
300 MHz spectrometer. CI MS spectra were taken on
a Nermag R-10-10 instrument. FAB MS spectra were
taken on a JEOL JMS-DX-303 HF instrument using
either glycerol or p-nitrobenzyl alcohol as matrices.
4.1.7. 2-(4-Trifluoromethylphenyl)-2-methylglycine (4-
Tmg). 5-(4-Trifluoromethylphenyl)-5-methylhydantoin
(536 mg, 2.1 mmol) and 2 N NaOH (8.2 mL) were com-
bined in a sealed high-pressure tube, and the resulting
solution was stirred at 115 ꢁC overnight. After cooling
to 0 ꢁC, the resulting solution was acidified with concen-
trated HCl, and the resulting white precipitate was re-
moved by filtration. The resulting filtrate was passed
through a column containing Dowex 50x-8 ion exchange
resin, (eluted with 1.7 M aqueous NH4OH) to afford,
after lyophilization, a white solid (46 mg, 0.2 mmol,
9%): 1H NMR (d6-DMSO, 300 MHz) d: 7.74 (d,
J = 6.7 Hz, 2H), 7.56 (d, J = 6.9 Hz, 2H), 1.50 (s, 3H);
19F NMR (d6-DMSO, 282.4 MHz) d: ꢀ59.63 (s, 3F).
4.1.1. Representative procedure for 5-aryl-5-alkylhydan-
toin synthesis;15 5-(4-nitrophenyl)-5-methylhydantoin.27
A 100 mL flask was charged with 4-nitroacetophenone
(0.8 g, 5.0 mmol), (NH4)2CO3 (3.6 g, 37.9 mmol), EtOH
(7.5 mL), H2O (7.5 mL), and a 32.6% (w/v) aqueous
solution of NaCN (0.8 mL, 6.5 mmol). The resulting
mixture was ultrasonicated in a 45 ꢁC water bath for
5 h, then cooled in an ice-bath and brought to neutral
pH with 1 N HCl. The resulting white precipitate was
obtained by filtration and washed with copious amounts
of 1:1 EtOH/H2O to afford the desired hydantoin as a
white powdery solid (1.1 g, 4.6 mmol, 92%) in spectro-
4.1.8. 2-(4-Methoxyphenyl)-2-methylglycine (4-Mmg).30
5-(4-Methoxyphenyl)-5-methylhydantoin
(495 mg,
2.3 mmol) was dissolved in 2 N NaOH (6 mL) and stir-
red at reflux (125 ꢁC) for 24 h. After cooling to ambient
temperature and filtration through glass wool, the
resulting solution was cooled to 0 ꢁC and brought to
neutral pH with 6 N HCl. The resulting precipitate
was removed by filtration, and the filtrate was concen-
trated in vacuo to a white solid. This solid was extracted
with MeOH, and the resulting extract was concentrated
to a white solid which was recrystallized from EtOH/
1
scopically pure form: H NMR (d6-DMSO, 300 MHz)
d: 10.92 (br s, 1H), 8.77 (s, 1H), 8.25 (d, J = 9.0 Hz,
2H), 7.76 (d, J = 9.0 Hz, 2H), 1.70 (s, 3H).
4.1.2.
5-(4-Trifluoromethylphenyl)-5-methylhydatoin.
White solid; mp 165.0–169.0 ꢁC; HRMS (FAB+) m/z
259.0681 [(M+H)+; calcd for C11H10O2N2F3
:
H2O (144 mg, 0.7 mmol, 33%): H NMR (d6-DMSO,
300 MHz) d: 7.84 (br s, 3H), 7.44 (d, J = 6.7 Hz, 2H),
6.87 (d, J = 6.7 Hz, 2H), 3.73 (s, 3H), 1.62 (s, 3H).
+
1
1
259.0694]; H NMR (d6-DMSO, 300 MHz) d: 10.89 (s,
1H), 8.74 (s, 1H), 7.76 (d, J = 8.5 Hz, 2H), 7.69 (d,
J = 8.4 Hz, 2H), 1.68 (s, 3H); 19F NMR (d6-DMSO,
282.4 MHz) d: –60.15 (s, 3F); 13C NMR (d6-DMSO,
100 MHz) d: 176.37, 156.19, 144.48, 128.52 (q,
JCꢀCꢀF = 31.9 Hz), 126.39, 124.16 (q, JCꢀF = 272.9 Hz),
125.46, 63.90, 25.14.
4.1.9. 2-Phenyl-2-hexylglycine (Pxg). 5-Phenyl-5-hexyl-
hydantoin (260 mg, 1.0 mmol) was dissolved in 2 N
NaOH (4.0 mL) and stirred at reflux (120 ꢁC) for 24 h,
then cooled to 0 ꢁC, and acidified with 6 N HCl. The
resulting precipitate was removed by filtration and the
filtrate was passed through a column containing Dowex
50x-8 ion exchange resin (eluted with 1.7 M aqueous
NH4OH) to afford, after lyophilization, a white solid
(191 mg, 0.8 mmol, 81%): 1H NMR (d6-DMSO,
300 MHz) d: 7.47 (d, J = 7.6 Hz, 2H), 7.30–7.17 (m,
4.1.3. 5-(4-Methoxyphenyl)-5-methylhydantoin.15. White
solid; MS (CI+) m/z 221.15 (M+H)+; 1H NMR (d6-
DMSO, 300 MHz) d: 10.64 (br s, 1H), 8.52 (s, 1H),
7.35 (d, J = 8.9 Hz, 2H), 6.93 (d, J = 8.9 Hz, 2H), 3.74
(s, 3H), 1.62 (s, 3H)

5880
J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
3H), 5.56 (br s, 3H), 2.06–1.93 (m, 2H), 1.21 (br s, 6H),
0.84 (t, J = 6.6 Hz, 3H).
3H), 3.30 (m, 4H), 2.95–2.76 (m, 7H), 1.85–1.32 (m,
41H), 1.00 (t, J = 6.5 Hz, 6H).
4.1.10. 1-Aminoindane-1-carboxylic acid (1-Aic).31
4.1.13. {5-[2-(Bisdecylcarbamoyl)ethylsulfanylmethyl]-3-
hydroxy-2-methylpyridin-4-ylmethyl} carbamicacid tert-
butyl ester (6). To a solution of 5 (1.0 g, 1.7 mmol) in
EtOH (10 mL) was added a 21% solution of EtONa in
EtOH (10.0 mL), and the reaction mixture was stirred
for 12 h at ambient temperature. The solvent was then
removed by rotary evaporation, and the mixture was
neutralized with saturated aqueous NH4Cl. The product
was extracted from the resulting aqueous solution with
CHCl3 (150 mL), and the resulting organic extract was
concentrated in vacuo to afford a waxy solid which
was purified by flash chromatography (740 mg,
1.4 mmol, 85%): Rf (20:1 CH2Cl2:MeOH) = 0.30; MS
(CI+) m/z 634.8 (M+H)+; 1H NMR (CDCl3,
300 MHz) d: 11.49 (s, 1H), 8.25 (s, 1H), 4.69 (d,
J = 6.3 Hz, 2H), 4.09 (s, 2H), 3.34 (m, 4H), 2.95 (m,
7H), 1.89–1.39 (m, 41H), 1.02 (t, J = 6.4 Hz, 6H).
Spiro[imidazolidine-4,10-indan]-2,5-dione
(407 mg,
2.0 mmol), Ba(OH)2Æ8H2O (1.0 g, 3.2 mmol), and H2O
(8.0 mL) were combined in a sealed high-pressure tube,
and the solution was stirred at 150 ꢁC overnight, then
cooled to ambient temperature and filtered. To the
resulting filtrate was added (NH4)2CO3 (514 mg,
2.9 mmol), and the resulting precipitate was removed
by centrifugation following by decanting the superna-
tant, which was evaporated to a white solid. This solid
was dissolved in 0.3 N HCl and passed through a col-
umn containing Dowex 50x-8 ion exchange resin (eluted
with 1.7 M aqueous NH4OH), to afford, after lyophili-
zation, a white solid (296 mg, 1.7 mmol, 84%): 1H
NMR (d6-DMSO, 300 MHz) d: 7.92 (br s, 3H), 7.36
(d, J = 7.3 Hz, 1H), 7.23–7.13 (m, 3H), 3.07–2.89 (m,
2H), 2.68–2.59 (m, 1H), 2.04–1.95 (m, 1H).
4.1.11.
3-[2-(Bisdecylcarbamoyl)ethyldisulfanyl]-N,N-
4.1.14.
5-[2-(Bisdecylcarbamoyl)ethylsulfanylmethyl]-4-
bisdecylpropionamide (4). To a solution of 3-(2-car-
boxy-ethyldisulfanyl)propionic acid (2.1 g, 10.0 mmol)
in benzene (20 mL) was added SOCl2 (10.0 mL,
137.1 mmol). The reaction was stirred at reflux for 2 h,
after which the solution became clear. The solvent then
was removed by distillation and the resulting acid chlo-
ride was dissolved in CH2Cl2 (20 mL). This CH2Cl2
solution was added dropwise to a flask containing
NH(C10H21)2 (6.0 g, 20.1 mmol), and Et3N (3.0 mL,
21.5 mmol) in CH2Cl2 (20 mL), at 0 ꢁC, and the result-
ing reaction mixture was stirred overnight. The reaction
mixture was washed sequentially with 1 N HCl and 1 N
NaOH, then concentrated in vacuo to afford the title
compound. The resulting product was carried on with-
out further purification: MS (CI+) m/z 770 (M+H)+;
1H NMR (CDCl3, 300 MHz) d: 3.34 (m, 8H) 3.05 (t,
J = 7.1 Hz, 4H), 2.81 (t, 4H), 1.65 (m, 8H), 1.40 (m,
56H) 0.98 (t, 12H).
(tert-butoxycarbonylaminomethyl)-3-hydroxy-1,2-di-methyl-
pyridinium iodide (7). To a solution of 6 (0.3 g, 0.6 mmol)
in CH3CN (10 mL) CH3I (2.0 mL, 32.1 mmol) was
added. The reaction was heated to reflux (80 ꢁC) and stir-
red for 30 min. The solvent was removed by rotary evap-
oration affording an yellow waxy solid (247 mg,
0.4 mmol, 65%). The resulting product was obtained
with sufficient purity for further reaction: MS (CI+)
m/z 650.5 (MꢀI)+; 1H NMR (CD3OD, 300 MHz) d:
8.53 (s, 1H), 4.61 (s, 2H), 4.37 (s, 3H), 4.11 (s, 2H),
3.00–2.80 (m, 7H), 1.90–1.41 (m, 41H), 1.03 (t,
J = 6.5 Hz, 6H).
4.1.15.
sulfanylmethyl]-3-hydroxy-1,2-dimethylpyridinium
4-Aminomethyl-5-[2-(bisdecylcarbamoyl)ethyl-
ha-
lide (8). To a solution of 7 (0.2 g, 0.4 mmol) in CH2Cl2
(10 mL), TFA (5 mL) was added, and the resulting mix-
ture was stirred at ambient temperature overnight. The
solvent was removed by rotary evaporation and the
resulting product was purified using preparative re-
verse-phase HPLC [50% to 100% aqueous MeOH
(0.1% TFA buffer) over 50 min, rt = 26–45 min] to af-
ford a yellow waxy solid (164 mg): MS (CI+) m/z
4.1.12. Methanesulfonic acid 5-[2-(bisdecyl-carbam-
oyl)ethylsulfanylmethyl]-4-(tert-butoxycarbonyl-aminom-
ethyl)-2-methylpyridin-3-yl ester (5). To a solution of
3-hydroxy-5-hydroxymethyl-2-methylpyridin-4-ylmethyl)
carbamic acid tert-butyl ester8 (1.5 g, 3.5 mmol) in
CH2Cl2 (20 mL) at 0 ꢁC, was added Et3N (1.6 mL,
11.5 mmol) followed by MsCl (0.87 mL, 11.2 mmol).
The reaction mixture was stirred for 15 min at 0 ꢁC,
and then washed with saturated aqueous NH4Cl. Evap-
oration of the resulting organic layer afforded a solid
product. This product (i.e., 3) was then dissolved in
EtOH (10 mL). Meanwhile, in a separate flask contain-
ing 4 (2.1 g, 2.7 mmol) in EtOH (20 mL), NaBH4 (1.0 g,
26.4 mmol) was added. After the mixture was stirred for
20 min, the previous ethanolic solution of 3 was added,
and the mixture was stirred for 2 h. The solvent was
removed by rotary evaporation and the resulting
product was purified by flash chromatography to afford
1
550.6 (MꢀX)+; H NMR (CD3OD, 400 MHz) d: 8.45
(s, 1H), 4.43 (s, 2H), 4.31 (s, 3H), 4.02 (s, 2H), 2.72
(m, 7H), 1.65 (m, 4H), 1.39 (br s, 28H), 0.98 (t,
J = 6.6 Hz, 6H).
4.1.16.
N,N-Bisdecyl-3-(2,2,8-trimethyl-4H-[1,3]-diox-
ino[4,5-c]pyridin-5-ylmethylsulfanyl)propionamide (10).
To a solution of (2,2,8-Trimethyl-4H-[1,3]dioxino-[4,
5-c]pyridin-5-yl)methanol8 (9, 1.1 g, 5.3 mmol) and 4
(1.9 g, 2.5 mmol) in THF (20 mL), Pbu3 (2.5 g,
10.0 mmol) was added, and the mixture was stirred under
argon at ambient temperature for 10 h. After evapora-
tion of the solvent, the resulting product was purified
using flash chromatography to afford a waxy solid
1
a
waxy solid (1.6 g, 2.6 mmol, 75%): Rf (20:1
(2.0 g, 3.4 mmol, 67%): MS (CI+) m/z 579 (M+H)+; H
CH2Cl2:MeOH) = 0.31; MS (CI+) m/z 712.8 (M+H)+;
1H NMR (CDCl3, 300 MHz) d: 8.49 (s, 1H), 5.72 (br
s, 1H), 4.71 (d, J = 6.2 Hz, 2H), 4.03 (s, 2H), 3.60 (s,
NMR (CDCl3, 300 MHz) d: 7.96 (s, 1H), 5.03 (s, 2H),
3.71 (s, 2H), 3.35 (m, 4H), 2.89–2.51 (m, 7H), 1.79 (s,
6H), 1.75–1.23 (m, 32H), 1.01 (t, J = 6.6 Hz, 6H).

J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
5881
4.1.17. N,N-Bisdecyl-3-(5-hydroxy-4-hydroxymethyl-6-
methylpyridin-3-ylmethylsulfanyl)propionamide (11). To
a solution of 10 (1.5 g, 2.6 mmol) in MeOH (20 mL)
was added concentrated aqueous HCl (20 mL), and
the resulting mixture was stirred overnight. The solvent
was evaporated, and the resulting residue was dissolved
in CH2Cl2 and washed with saturated aqueous NaH-
CO3. The organic layer was dried (Na2SO4), filtered,
and concentrated to afford a spectroscopically pure
waxy solid (1.4 g, 2.5 mmol, 98%): MS (CI+) m/z 538
After evaporation of the solvent, the resulting product
was purified using preparative reverse-phase HPLC [50
to 100% aqueous CH3CN (0.1% TFA buffer) over
45 min, rt = 24–32 min] to afford a deep red waxy solid
(145 mg): Rf (CH3OH) = 0.76; MS (CI+) m/z 548.0
(MꢀX)+; 1H NMR (CD3CN, 400 MHz) d: 10.58 (s,
1H), 8.39 (s, 1H), 4.22 (s, 3H), 4.18 (s, 2H), 3.30 (m,
4H), 2.75–2.62 (m, 7H), 1.50 (m, 4H), 1.43 (br s, 28H),
0.96 (t, J = 6.8 Hz, 6H).
1
(M+H)+; H NMR (CDCl3, 300 MHz) d: 7.99 (s, 1H),
4.2. Decarboxylation and Catalyst Turnover Studies
5.16 (s, 2H), 3.87 (s, 2H), 3.40 (m, 4H), 2.90–2.59 (m,
7H), 1.95–1.01 (m, 38H).
The concentrations of PEI-C12 (8.7%), pyridoxal 50-eth-
ylthioether, 8, and 15 solutions were determined by H
1
4.1.18.
N,N-Bisdecyl-3-(4-formyl-5-hydroxy-6-methyl-
NMR analysis using terephthalic acid disodium salt as
an internal standard. UV–vis spectra and kinetics were
taken on a Varian Cary IE UV–vis spectrometer with
accompanying kinetics-calculating software. Analytical
HPLC was run on an HP1090 liquid chromatography
(series II) equipped with a DR5 pumping system, a tem-
perature-controlled autosampler, and a diode-array
UV–vis detector. Vydac Protein & Peptide C18 reverse
phase analytical columns (4.6 · 150 mm) were used
exclusively. Derivatization Reagent: A solution of
N-Boc-L-cysteine (219 mg, 1.0 mmol) and o-phthalyldi-
aldehyde (140 mg, 1.0 mol) in MeOH (5 mL). Derivati-
zation Buffer: 1.0 M KH2PO4, pH 8.0, buffer solution.
pyridin-3-ylmethylsulfanyl) propionamide (12). To a solu-
tion of 10 (1.5 g, 2.5 mmol) in CH2Cl2 (50 mL), MnO2
(5.0 g, 57.5 mmol) was added, and the reaction was stir-
red at ambient temperature for 3 h. The solution was
then filtered through Celite and concentrated to a spec-
troscopically pure yellow waxy solid. (1.0 g, 1.8 mmol,
70%). MS (CI+) m/z 536 (M+H)+; H NMR (CD3CN,
300 MHz) d: 10.56 (s, 1H), 8.09 (s, 1H), 4.13 (s, 2H),
3.30 (m, 4H), 2.82 (t, J = 6.9 Hz, 2H), 2.61 (t, 2H), 2.56
(s, 3H), 1.70–1.25 (m, 32H), 1.01 (t, J = 6.9 Hz, 6H).
1
4.1.19. N,N-Bisdecyl-3-(4-dimethoxymethyl-5-hydroxy-6-
methylpyridin-3-ylmethylsulfanyl)propionamide (13). To
a solution of 12 (2.0 g, 3.7 mmol) in HC(OCH3)3
(20.0 mL) and CH3OH (10 mL), TsOH (1.0 g,
5.8 mmol) was added, and the reaction was heated to
110 ꢁC and stirred at that temperature for 3 h. After
evaporation of the solvent, the product was purified
using flash chromatography to obtain a waxy solid
(1.6 g, 2.7 mmol, 72%): Rf (20:1 CH2Cl2:MeOH) = 0.09;
MS (CI+) m/z 580.3 (M+H)+; 1H NMR (CDCl3,
300 MHz) d: 8.92 (s, 1H), 7.95 (s, 1H), 6.18 (s, 1H),
3.88 (s, 2H), 3.56 (s, 6H), 3.41–3.26 (m, 4H), 2.95 (t,
J = 7.2 Hz, 2H), 2.62 (t, J = 7.2 Hz, 2H), 2.53 (s, 3H),
1.73 (br s, 4H), 1.40 (br s, 28H), 0.99 (t, J = 6.1 Hz, 6H).
4.2.1. General procedure for determining kdecarboxylation of
2-aryl-2-alkylglycines with pyridoxal 50-ethylthioether.
Three UV cuvettes charged with pyridoxal 50-ethylthioe-
ther (1.0 · 10ꢀ4 M), amino acid (1.0 · 10ꢀ3 M), EDTA
(2.0 · 10ꢀ3 M), pH 7.0, Hepes buffer (2.5 · 10ꢀ2 M),
and KCl (2.5 · 10ꢀ2 M) in either pure deionized H2O
or 34.5% aqueous acetonitrile were heated to 60 ꢁC
and monitored by UV–vis spectroscopy at 400 nm every
5 s for 4–24 h, depending on the reaction rate. The
resulting exponential curves were converted into pseu-
do-first-order rate constants using the spectrometer soft-
ware and the average value standard deviation was
reported as kdecarboxylation in Table 1.
4.1.20.
5-[2-(Bisdecylcarbamoyl)ethylsulfanylmethyl]-
4-dimethoxymethyl-3-hydroxy-1,2-dimethylpyridinium
iodide (14). To a solution of 13 (0.5 g, 0.9 mmol) in
CH3CN (20 mL), CH3I (2.0 mL, 32.1 mmol) was added,
and the reaction was stirred at ambient temperature for
5 h, monitoring by TLC [MeOH, Rf (starting mate-
rial) = 0.85, Rf (product) = 0.75]. After evaporation of
the solvent, the resulting product was purified using pre-
parative reverse-phase HPLC [50% to 100% aqueous
MeOH (0.1% TFA buffer) over 40 min, then 100%
MeOH (0.1% TFA buffer) for 20 min, rt = 41–49 min]
to afford a yellow waxy solid: MS (CI+) m/z 594.4
4.2.2. General procedure for determining kdecarboxylation of
2-aryl-2-alkylglycines with 15.
4.2.2.1. Conditions A. A capped HPLC autosampler
vial was charged with 15 (1.0 · 10ꢀ4 M), amino acid
(1.0 · 10ꢀ3 M), EDTA (2.0 · 10ꢀ3 M), pH 7.0, Hepes
buffer (2.5 · 10ꢀ2 M), and KCl (2.5 · 10ꢀ2 M) in 58%
aqueous acetonitrile. The reaction mixture was heated
to 30 ꢁC and resulting ketone production (see below)
was monitored by analytical HPLC/UV–vis (250 nm).
Pseudo-first-order rate constants were calculated from
the resulting exponential area/time curves. A minimum
ofthreerunswereperformedandtheaveragevalue stan-
dard deviation are reported as kdecarboxylation in Table 3.
1
(MꢀI)+; H NMR (CD3OD, 300 MHz) d: 7.82 (s, 1H),
6.08 (s, 1H), 4.19 (s, 3H), 4.11 (s, 2H), 3.40 (m, 4H),
2.80 (m, 4H), 2.67 (s, 3H), 1.70–1.35 (m, 32H), 1.01 (t,
J = 6.8 Hz, 6H).
4.2.2.2. Conditions B. A capped HPLC autosampler
vial was charged with 15 (1.0 · 10ꢀ4 M), amino acid
(1.0 · 10ꢀ3 M), EDTA (2.0 · 10ꢀ3 M), PEI-C12 (8.7%)
(1.25 · 10ꢀ5), pH 7.0, Hepes buffer (2.5 · 10ꢀ2 M), and
KCl (2.5 · 10ꢀ2 M). The reaction mixture was heated
to 30 ꢁC. Analysis and kdecarboxylation determination were
performed as per Conditions A.
4.1.21. 5-[2-(Bisdecylcarbamoyl)ethylsulfanylmethyl]-4-
formyl-3- hydroxy-1,2-dimethylpyridinium halide (15).
To a solution of 14 (0.2 g, 0.3 mmol) in CH3CN
(10 mL) was added 1 N aqueous HCl (10 mL), and the
reaction was stirred at ambient temperature for 12 h.

5882
J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
4. (a) Some recent publications: Liu, L.; Breslow, R. Bioorg.
4.2.2.3. Sample intervals, ketone elution gradient
Med. Chem. 2004, 12, 3277; (b) Liu, L.; Breslow, R. J. Am.
Chem. Soc. 2003, 125, 12110; (b) Zhou, W.; Liu, L.;
Breslow, R. Helv. Chim. Acta 2003, 86, 3560; (d) Liu, L.;
Rozenman, M.; Breslow, R. J. Am. Chem. Soc. 2002, 124,
12660; (e) Liu, L.; Rozenman, M.; Breslow, R. Bioorg.
Med. Chem. 2002, 10, 3973.
conditions, and retention times. Acetophenone (from
Pmg): 10 min intervals, 45 to 55% aqueous MeOH
(0.1% TFA buffer) over 8 min at 1.0 mL/min,
rt = 5.92 min; 4-trifluoro-methylacetophenone (from
4-Tmg): 15 min intervals, 50 to 65% aqueous MeOH
(0.1% TFA buffer) over 12 min at 1.0 mL/min,
rt = 9.38 min; 4-methoxyacetophenone (from 4-Mmg):
10 min intervals, 50% to 58.8% aqueous MeOH (0.1%
TFA buffer) over 7 min at 1.0 mL/min, rt = 5.57 min;
heptanophenone (from Pxg): 10 min intervals, 75.8 to
84.5% aqueous MeOH (0.1% TFA buffer) over 7 min
at 1.0 mL/min, rt = 6.44 min; 1-indanone (from 1-Aic):
10 min intervals, 45% to 55% aqueous MeOH (0.1%
TFA buffer) over 8 min at 1.0 mL/min, rt = 5.88 min.
5. (a) Liu, L.; Breslow, R. Tetrahedron Lett. 2001, 42, 2775;
(b) Chmielewski, J.; Breslow, R. Heterocycles 1987, 25,
533.
6. Weiner, W.; Winkler, J.; Zimmerman, S. C.; Czarnik, A.
W.; Breslow, R. J. Am. Chem. Soc. 1985, 107, 4093.
7. Koh, J. PhD thesis, Department of Chemistry. Columbia
University, 1994.
8. Liu, L.; Zhou, W.; Chruma, J.; Breslow, R. J. Am. Chem.
Soc. 2004, 126, 8136.
9. Toth, K.; Amyes, T. L.; Richard, J. P.; Malthouse, J. P.;
´
´
NıBeilliu, M. E. J. Am. Chem. Soc. 2004, 126, 10538.
10. Metzler, D. E.; Snell, E. E. J. Am. Chem. Soc. 1952, 74,
979.
4.2.3. General procedure for determining TOF and TON
for 2-aryl-2-alkylglycines with PEI-C12 (8.7%) and 1 or 8.
A tightly capped vial was charged with 1 or 8
(2.5 · 10ꢀ5 M), sodium pyruvate (5.0 · 10ꢀ3 M), amino
acid (1.0 · 10ꢀ2 M), EDTA (2.0 · 10ꢀ3 M), and PEI-
C12 (8.7%) (2.5 · 10ꢀ5 M). The reaction mixture was
brought to pH 7.5 with 1 N HCl or 1 N NaOH and
heated to 60 ꢁC for 1 or 30 ꢁC for 8. Two separate con-
trols reactions, one in which the pyridoxamine catalyst 1
or 8 was excluded (no turnover, control #1) and another
in which the amino acid was excluded (one turnover,
control #2) were run simultaneously. At regular inter-
vals, 10 lL aliquots from the reaction mixture and the
two controls were removed and individually combined
with 80 lL Derivatization Reagent and 10 lL Derivati-
zation Buffer. The resulting three solutions of 1-thioiso-
indoles were analyzed by HPLC with UV–vis detection
at 344.5 nm. Retention time: DL-alanine = 17–18 min.
Turnover was calculated by dividing the area under
the DL-alanine peak for the reaction mixture by that of
control #2. A minimum of three runs were performed
and the average TOF and TON values standard devi-
ation are reported in Tables 2 and 4.
11. Metzler, D. E.; Ikawa, M.; Snell, E. E. J. Am. Chem. Soc.
1954, 76, 648.
12. Bruice, T. C.; Topping, R. M. J. Am. Chem. Soc. 1963, 85,
1493, and references therein.
13. (a) Imperiali, B.; Roy, R. S. J. Org. Chem. 1995, 60,
1891; (b) Shogren-Knaak, M. A.; Imperiali, B. Bioorg.
Med. Chem. 1999, 7, 1993; (c) Imperiali, B.; Roy, R. S.
J. Am. Chem. Soc. 1994, 116, 12083; (d) Roy, R. S.;
Imperiali, B. Tetrahedron Lett. 1996, 37, 2129; (e) Roy,
R. S.; Imperiali, B. Protein Eng. 1997, 10, 691; (f)
Kuang, H.; Brown, M. L.; Davies, R. R.; Young, E. C.;
Distefano, M. D. J. Am. Chem. Soc. 1996, 118, 10702;
(g) Qi, D. F.; Kuang, H.; Distefano, M. D. Bioorg.
Med. Chem. Lett. 1998, 8, 875; (h) Haring, D.;
Distefano, M. D. Bioorg. Med. Chem. 2001, 9, 2461;
(i) Kuang, H.; Distefano, M. D. J. Am. Chem. Soc.
1998, 120, 1072.
14. (a) Murakami, Y.; Kikuchi J-i; Hisaeda, Y.; Nakam-
ura, K.; Kitazaki, T.; Kaya, H. Bull. Chem. Soc. Jpn.
1990, 63, 2339; (b) Murakami, Y.; Hisaeda, Y.;
Nakamura, K.; Kikuchi, J.-i. Chem. Lett. 1990, 1765;
(c) Maley, J. R.; Bruice, T. C. J. Am. Chem. Soc. 1968,
90, 2843.
15. Li, J.; Li, L.; Wang, J. Ind. J. Chem. Sec. B 1998, 37,
298.
Acknowledgments
16. For less water soluble amino acids, 1 drop of 1 N HCl was
added to afford completely homogeneous solutions.
17. 3-Pmg was obtained by Ba(OH)2-mediated hydrolysis of
the corresponding known hydantoin: Teague, P. C.;
Ballentine, A. R.; Rushton, G. L. J. Am. Chem. Soc.
1953, 75, 3429.
We thank the NIH and NSF for financial support of this
work. J.J.C. has an NIH postdoctoral fellowship, and
L.L. was partially supported by a Guthikonda graduate
fellowship.
18. Our previous studies (Ref. 8) determined 60 ꢁC to be the
ideal reaction temperature for pyridoxal-mediated
amino acid decarboxylations, affording the best
ratio of kdecarboxylayion to byproduct formation/pyridoxal
decomposition.
19. 2-Phenyl-2-(trifluoromethyl)glycine was synthesized fol-
lowing published procedures: Amii, H.; Kishikawa, Y.;
Kageyama, K.; Uneyama, K. J. Org. Chem. 2000, 65,
3404.
20. Askley, J. A.; Lo, C.-H. L.; McElhaney, G. P.; Wirsching,
P.; Janda, K. D. J. Am. Chem. Soc. 1993, 115, 2515.
21. Davis, F. A.; Lee, S.; Zhang, H.; Fanelli, D. L. J. Org.
Chem. 2000, 65, 8704.
22. The parent polyethylenimine (PEI) used in this study is a
branched polymer purchased from Aldrich Co. and has a
molecular weight of ca. 60,000, containing ca. 1400
monomeric ethylamine residues and a polydispersity index
of 12.5.
References and notes
1. (a) Aaslestad, H. G.; Larson, A. D. J. Bacteriol. 1964, 88,
1296; (b) Bailey, G. B.; Dempsey, W. B. Biochemistry
1967, 6, 1526.
2. (a) Toney, M. D. Arch. Biochem. Biophys. 2005, 433, 279;
(b) Zhou, X.; Jin, X.; Medhekar, R.; Chen, X.; Dieck-
mann, T.; Toney, M. D. Biochemistry 2001, 40, 1367; (c)
Toney, M. D.; Hohenester, E.; Keller, J. W.; Jansonius, J.
N. J. Mol. Biol. 1995, 245, 151; (d) Toney, M. D.
Biochemistry 2001, 40, 1378; (d) Zabrinski, R. F.; Toney,
M. D. J. Am. Chem. Soc. 2001, 123, 193; (f) Li, J.; Brill, T.
B. J. Phys. Chem. A 2003, 107, 5993.
3. Quinoid intermediates were detected when Ala was used as
a substrate but not when Aib the standard enzymatic
substrate was used: see Ref. 2b.

J. J. Chruma et al. / Bioorg. Med. Chem. 13 (2005) 5873–5883
5883
23. Kikuchi, J.-i.; Zhang, Z.-Y.; Murakami, Y. J. Am. Chem.
Soc. 1995, 117, 5383.
24. Gong, J.; Hunter, G. A.; Ferreira, G. C. Biochemistry
1998, 37, 3509.
25. Bach, R. D.; Canepa, C.; Glukhovtsev, M. N. J. Am.
Chem. Soc. 1999, 121, 6542.
26. Zhou, W.; Yerkes, N.; Chruma, J. J.; Liu, L.; Breslow, R.
Bioorg. Med. Chem. Lett. 2005, 15, 1351.
27. Stilz, H. U.; Guba, W.; Jablonka, B.; Just, M.; Klinger,
O.; Koenig, W.; Wehner, V.; Zoller, G. J. Med. Chem.
2001, 44, 1158.
28. Brown, M. L.; Zha, C. C.; Van Dyke, C. C.; Brown, G. B.;
Brouillette, W. J. J. Med. Chem. 1999, 42, 1537.
29. Matveeva, E. D.; Podrugina, T. A.; Morozkina, N.; Yu
Zefirova, O. N.; Seregin, I. V.; Bachurin, S. O.; Pellicciari,
R.; Zefirov, N. S. Russ. J. Org. Chem. 2002, 38, 1769.
30. Washburn, W. N.; Sun, C.-Q.; Bisachhi, G.; Wa, G.;
Cheng, P. T.; Sher, P. M.; Ryono, D.; Gavai, A. V.; Poss,
K.; Girotra, R. N.; McCann, P. J.; Mikkilineni, A. B.;
Dejneka, T. C.; Wang, T. C.; Merchant, Z.; Morella, M.;
Arbeeny, C. M.; Harper, T. W.; Slusarchyk, D. A.;
Skwish, S.; Russell, A. D.; Allen, G. T.; Tesfamariam, B.;
Frohlich, B. H.; Abboa-Offei, B. E.; Cap, M.; Waldron, T.
L.; George, R. J.; Young, D.; Dickinson, K. E.; Seymour,
A. A. Bioorg. Med. Chem. Lett. 2004, 14, 3525.
31. Connors, T. A.; Ross, W. C. J. J. Chem. Soc. 1960, 2119.