Hydrophobic effects on rates and substrate selectivities in polymeric transaminase mimics

Article abstract of DOI:10.1021/ja028151k

The amination of ketoacids to amino acids by pyridoxamine is greatly accelerated when the pyridoxamine is covalently linked to polyethylenimine carrying N-methyl and N-lauryl groups. Michaelis-Menten kinetics is seen with all substrates, from which the effect of the lauryl groups and the methyl groups can be determined with respect to the strength of binding of the substrate and the rate constant k₂ within the complex. The polyamine catalyzes the reaction using acid and base groups, the lauryl groups increase k₂ by producing a nonpolar medium in which the reaction occurs, and the lauryl groups promote binding of hydrophobic substrates. The result is that the amination of indolepyruvic acid to produce tryptophan is accelerated by 240000-fold. Copyright

Full text of DOI:10.1021/ja028151k

Published on Web 10/05/2002
Hydrophobic Effects on Rates and Substrate Selectivities in Polymeric
Transaminase Mimics
Lei Liu, Mary Rozenman, and Ronald Breslow*
Department of Chemistry, Columbia UniVersity, New York, New York 10027
Received August 15, 2002
Recently we reported a great increase (ca. 2300-fold) of the
transamination rate for the pyridoxamine-pyruvic acid system when
we covalently attached pyridoxamine to polyethylenimine (PEI)
carrying some attached lauryl groups.1 We suggested that the
excellent general acid-base catalysis exerted by the polymeric
partially protonated amines is one of the reasons for the large rate
enhancement.1,2 We also found that the rate enhancement of the
polymer over that of simple pyridoxamine 1 was a steep function
of the length of the alkyl chains added in polymer with roughly
the same percentage of alkylation and pyridoxamine attachment.1
We suggested that the hydrophobic chains create regions in which
the transamination can take place in a less than fully aqueous
environment.
We have now explored this system further and discovered a
number of important features. First of all, we find that the amination
reaction of pyruvic acid with simple pyridoxamine is buffer-
catalyzed but the reaction with our polymer-linked pyridoxamine
shows no catalysis by external buffers. This supports the argument
that the acid and base groups of the polymer are performing the
catalytic proton transfers in the transamination process. Extrapolat-
ing [buffer] to zero, the PEI-pyridoxamine system is actually
10000-fold faster than is simple pyridoxamine with pyruvic acid.
Furthermore, we find that the polymer system shows saturation
effects when titrated with substrates, Michaelis-Menten kinetics.3
We report the k2 and KM values of different substrate-polymer
reaction pairs in Table 1 (see footnote b). As the entries show,
there are two striking effects of the added C-12 chains in the
pyridoxamine-PEI hybrid. The k2 constants are modified, and the
substrate binding constants are modified as well. This latter effect
shows up strongly with substrates that are themselves hydrophobic,
such as the keto acids corresponding to phenylalanine and tryp-
tophan.
acid (8 vs 9), which overcomes the weaker binding of relatively
non-hydrophobic pyruvic acid by 9.
Apparently k2 is increased by the lowered local dielectric constant
which results from the presence of nonpolar lauryl groups.7 In the
more polar unmethylated PEI laurylation has a lesser effect on the
rate for pyruvic acid transamination (6 vs 7).
Strikingly, for the phenylpyruvic acid case we see an impressive
increase of rate, laurylation causing the k2/KM value to jump 190-
fold for PEI-Methylated (8 vs 9), much more than the 12-fold
increase seen for pyruvic acid. The 190-fold increase is due to both
an increase of k2 by 28 times, slightly smaller than the k2 effect
with pyruvic acid, and an increased binding strength reflected in a
decrease of KM by 7 times. We believe that the increased binding
strength reflects hydrophobic interaction of the lauryl chains with
the phenyl group of the substrate. For PEI-Unmethylated (6 vs 7)
laurylation increases the transamination rate by 9 times for
phenylpyruvic acid and 2 times for pyruvic acid, a less impressive
demonstration of the hydrophobic effect. We then determined the
effect of laurylation on the transamination rates for other hydro-
phobic substrates. (Table 2). The results show that the activity (k2/
Km) of glyoxylic acid in transamination is similar to that of pyruvic
acid. 4-Methyl-2-oxopentanoic acid, the precursor of leucine, shows
only slightly higher reactivity as the lower KM from some
hydrophobic binding is compensated by a lower k2. R-Ketoglutaric
acid shows a 5-fold increase in k2/Km compared to glyoxylic acid
and pyruvic acid, due primarily to stronger binding indicated in a
smaller KM. This likely reflects a carboxylate-ammonium salt bridge
interaction. Haring and Distefano9 have described a protein-based
transaminase mimic with a k2 of 4.8 × 10-3 min-1 for the
R-ketoglutaric acid substrate and a KM of 1.8 mM. The resulting
k2/KM of 2.7 min-1 M-1 can be compared with the 16 min-1 M-1
k2/KM value for our polymer system.
We have also prepared a transaminase mimic 5 by linking
pyridoxamine to polyallylamine (PAA) and leaving the nitrogens
unmethylated.4 As the data in Table 1 show, 5 shows a slightly
smaller rate of transamination of pyruvic acid (k2 and k2/KM) than
is seen with the analogous 6, derived from PEI. PAA is linear,
while PEI is highly branched.4 Complete methylation5 decreased
the k2 of PEI-supported pyridoxamine with pyruvic acid and
phenylpyruvic acid substrates (6 vs 8), while having only small
effects on KM. Bulky phenylpyruvic acid shows slightly lesser
reactivity than pyruvic acid with 5, 6, and 8.
Another mode of modifying PEI is alkylation with long
hydrocarbon chains.6 We have reported that partial (ca. 10%)
laurylation of PEI-Methylated increases the transamination rate for
pyruvic acid substrate by ca. 14-fold.1 We now see that the effect
of laurylation is due to a considerable increase in k2 for pyruvic
Phenylpyruvic acid shows a 10-fold increase in reactivity (k2/
KM) over pyruvic acid, primarily due to a 5-fold decrease in KM as
a result of hydrophobic interactions between the lauryl groups on
the PEI and the phenyl group on the ketoacid substrate. An even
greater hydrophobic effect is seen in the case of indole-3-pyruvic
acid, for which the hydrophobic interactions between the indole
group and lauryl groups causes a 22-fold decrease in KM and 24-
fold increase in k2/KM when compared with the pyruvic acid case.
Since amination of pyruvic acid by compound 9 showed a 10000-
fold acceleration relative to pyridoxamine at [buffer] ) 0 and since
we had earlier found that the amination of pyruvic acid and of
indolepyruvic acid by pyridoxamine had the same rate,10 this means
that our hybrid 9 accelerates the amination of indolepyruvic acid
to tryptophan by ca. 240000-fold relative to pyridoxamine.
These substantial contributions of hydrophobic effect to the
substrate selectivity of the polymeric transaminase mimic were
confirmed by competition reactions with HPLC product analyses.
* To whom correspondence should be addressed. E-mail: rb33@columbia.edu.
9
12660
J. AM. CHEM. SOC. 2002, 124, 12660-12661
10.1021/ja028151k CCC: $22.00 © 2002 American Chemical Society

C O MMU N I C A T I O N S
Table 1: Michaelis Constants (k2 × 1000 (min-1), KM (mM)) of the Polymeric Pyridoxamine Reagentsa,b
pyruvic acid
phenylpyruvic acid
reagent
k2
K
M
k2/K
k2
K
M
k2/K
M
M
PAA-Unmethylated, 5
PEI-Unmethylated, 6
PEI-lauryl-Unmethylated, 7
PEI-Methylated, 8
20 ( 1
55 ( 1
15 ( 1
17 ( 1
11 ( 0
12 ( 2
42 ( 6
1.3
3.2
6.4
0.60
6.9
21 ( 4
49 ( 5
43 ( 8
21 ( 5
0.49
2.3
21
0.48
92
70 ( 3
50 ( 5
2.4 ( 0.6
27 ( 3
7.2 ( 0.7
290 ( 60
13 ( 1
370 ( 30
PEI-lauryl-Methylated, 9
4.0 ( 0.6
a All constants reported are averaged measurements from two trials and data analysis with Eadie-Hofstree and Hanes-Wolff8 kinetic approximations.
Five or six distinct substrate concentrations were used to define the kinetic curve for each polymer/substrate pair. Pyridoxamine concentrations in the
polymer were determined by UV measurements (pyridoxamine, λ ) 324 nm), ca. 1.0 × 10-4 M, T ) 40 °C, pH ) 7.5. Find details in Supporting Information.
b k2 is the rate constant for the reaction of an intermediate going to product. KM has its usual meaning: (k2 + k-1)/k1, where k1 and k-1 refer to the reversible
formation of the reaction intermediate.
Scheme 1
The result is high rate acceleration for all substrates, and particularly
high acceleration for hydrophobic substrates.
Acknowledgment. M.M.R. acknowledges funding from the
Morris K. Udall Foundation. We thank the NIH and NSF for
financial support of this work.
Supporting Information Available: Procedures for the synthesis
of the polymeric transaminase mimics and model compounds 2-4;
detailed kinetic data for the buffer effect and acetonitrile effect. (PDF)
This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Liu, L.; Breslow, R. J. Am. Chem. Soc. 2002, 124, 4978.
(2) We have made pyridoxamine model compounds 2-4 containing one, two,
and four amino groups in the side chain. The transamination rates of these
compounds were found to be 12, 36, and 68 times higher than the free
simple pyridoxamine 1 at pH 7.0 in aqueous HEPES buffer. (See the
Supporting Information.) These are much smaller accelerations than we
have shown with the polymer-based reagents.
Table 2. Michaelis Constants (k2 × 1000(min-1), KM (mM)) of the
PEI-lauryl-Methylated Transaminase Mimic, 9a
substrate
k2
K
M
k2/K
M
glyoxylic acid
pyruvic acid
4-methyl-2-oxo-pentanoic acid
R-ketoglutaric acid
phenylpyruvic acid
indole-3-pyruvic acid
92 ( 6
100 ( 0
55 ( 4
120 ( 0
200 ( 1
110 ( 2
30 ( 5
3.1
3.6
4.2
16
36
85
28 ( 1
13 ( 3
7.4 ( 0.9
5.5 ( 0.6
1.3 ( 0.2
(3) Experiments were carried out under conditions of excess substrate, that
is S0 . C0. C0 represents one reaction site (i.e., pyridoxamine unit) on
the polymer and S0 the ketoacid substrate. Under these circumstances initial
observed rate constants kobs are measured which can fit into equation:
kobs ) k2S0/(KM + S0). For detailed kinetics see Supporting Information.
For another example of Michaelis-Menten kinetics in a simple polymeric
system see: Suh, J.; Scarpa, I. S.; Klotz, I. N. J. Am. Chem. Soc. 1976,
98, 7060.
(4) Poly(allylamine) [-CH2CH(CH2NH2)-]n is a linear polymer, average Mw
ca. 17000; polyethylenimine is a branched polymer which has a molecular
weight of ca. 60000, containing ca. 1400 monomeric ethylamine residues.
About 25% of the amino groups are primary, about 50% are secondary,
and about 25% are tertiary. For more details see: Suh, J. Synlett 2001,
1343.
(5) Polymers were completely methylated using reductive amination with
HCHO and NaBH3CN.
(6) Klotz, I. M.; Royer, G. P.; Scarpa, I. S. Proc. Natl. Acad. Sci. U.S.A.
1971, 68, 263.
(7) Addition of acetonitrile to aqueous pyridoxamine/pyruvic acid system was
found to increase the transamination rate significantly. See Supporting
Information. For the medium effects of the polymers, also see: Hollfelder,
F.; Kirby, A. J.; Tawfik, D. S. J. Org. Chem. 2001, 66, 5866.
(8) Tinoco, I., Jr.; Sauer, K.; Wang, J. C. Physical Chemistry, 3rd ed.; Prentice
Hall Press: New Jersey, 1995.
(9) Haring, D.; Distefano, M. D. Bioorg. Med. Chem. 2001, 9, 2461.
(10) Breslow, R.; Hammond, M.; Lauer, M. J. Am. Chem. Soc. 1980, 102,
421-422.
a Reported constants have been determined from UV measurement, T )
30 °C, pH ) 7.5. All constants reported are averaged measurements from
two trials and data analysis with Eadie-Hofstree and Hanes-Wolff8 kinetic
approximations. Five or six distinct substrate concentrations were used to
define the kinetic curve for each reagent/substrate pair. See footnote b of
Table 1.
Competition reactions were run with 20:20:1 (pyruvic acid:
phenylpyruvic acid:pyridoxamine unit, or pyruvic acid:indole-3-
pyruvic acid:pyridoxamine unit) ratios at 28 °C, pH 7.5, for 1 h.
Amino acid product ratios determined from HPLC analysis were
1:8.5, for the alanine-phenylalanine case and 1:20 for the alanine-
tryptophan case, consistent with the results from the kinetic studies
in Table 2 under slightly different conditions.
Thus, we have shown that these polyamine-supported pyridox-
amine transaminase mimics demonstrate Michaelis-Menten kinet-
ics and substrate selectivity. Laurylation of the polymers both
increases k2 by producing a less aqueous reaction environment and
selectively binds the more hydrophobic substrates, decreasing KM.
JA028151K
9
J. AM. CHEM. SOC. VOL. 124, NO. 43, 2002 12661