Remarkably broad substrate tolerance of malonyl-CoA synthetase, an enzyme capable of intracellular synthesis of polyketide precursors [2]

Full text of DOI:10.1021/ja0028368

5822
J. Am. Chem. Soc. 2001, 123, 5822-5823
Remarkably Broad Substrate Tolerance of
Malonyl-CoA Synthetase, an Enzyme Capable of
Intracellular Synthesis of Polyketide Precursors
Scheme 1. Synthesis of Malonic Acid Derivatives
†
†
‡
Nicola L. Pohl, Marcus Hans, Hwan Young Lee,
‡
§
,|
Yu Sam Kim, David E. Cane, and Chaitan Khosla*
Departments of Chemical Engineering and
Chemistry and Biochemistry, Stanford UniVersity
Stanford, California 94305-5025
Department of Biochemistry, College of Science
Protein Network Research Center
ity of malonyl-CoA synthetase from R. trifolii. As shown below,
the enzyme has remarkable tolerance for a variety of C-2
substituted malonic acids, making it an attractive catalyst for the
in vivo or in vitro formation of building blocks for polyketide
synthesis.
Yonsei UniVersity, Seoul, Korea 120-749
Department of Chemistry, Brown UniVersity
ProVidence, Rhode Island 02912
An important consideration in the exploitation of malonyl-CoA
synthetase is the availability of suitable 1,3-dicarboxylic acid
substrates. Although malonic acid derivatives can be accessed
by alkylation of a malonic acid diester by an electrophile and
ReceiVed August 1, 2000
6
subsequent base-catalyzed saponification, this synthetic protocol
Polyketides are a structurally complex class of natural products
1
requires tedious isolation procedures to recover high yields of
dicarboxylic acids. On the basis of our previous experience with
the synthesis of N-acetyl cysteamine monothioesters of malonic
with therapeutic and agrochemical utility. Polyketide backbones
are generated by the repetitive decarboxylative condensation of
simple malonic acid derivatives by large multifunctional proteins
called polyketide synthases (PKSs). These modular enzymes can
be genetically modified to biosynthesize new “unnatural” natural
7
acid via ring-opening of Meldrum’s acid, we reasoned that in
addition to being an activating group, the isopropylidene ketal
could also be viewed as a protecting group that could be cleaved
to avoid aqueous product extractions. Ultimately, a mixture of
2
products. Notwithstanding the spectacular diversity of natural
and engineered polyketides, however, the potential structural
diversity of these molecules is seriously limited by the relatively
small number of building blocks that are naturally available for
polyketide biosynthesis within a cell. The ability to regioselec-
tively incorporate new, orthogonally reactive functional groups
into a polyketide scaffold has important implications for inves-
tigations into PKS mechanisms as well as for the medicinal
exploitation of polyketides.
8
trifluoroacetic acid and water were chosen. Deprotection took
place in minutes at ambient temperature, and the reagents and
reaction byproducts were volatilized to leave the desired diacid
9
in almost quantitative yield. Several alkylated malonic acids were
obtained in this manner (Scheme 1).
Malonyl-CoA synthetase has previously been shown to convert
malonic acid and CoASH to malonyl-CoA with hydrolysis of ATP
1
0
to AMP and diphosphate via a malonyl-AMP intermediate.
Acyltransferase domains within PKSs are responsible for
selecting malonyl-CoA or its analogues for each round of
condensation. The two most common metabolically available
substrates are malonyl-CoA and methylmalonyl-CoA, although
biosynthetic pathways for a few other R-carboxylated CoA
While malonic acid and methylmalonic acid were converted to
their corresponding monothioesters, acetate, propionate, or suc-
cinate were not. This suggested that the 1,3-diacid functionality
was crucial for enzymatic activity.
3
To probe the molecular recognition features of malonyl-CoA
synthetase, ethyl-, propyl-, allyl-, isopropyl-, dimethyl-, cyclo-
propyl-, cyclopropylmethylene-, cyclobutyl-, and benzyl-malonate
were either purchased or prepared, and assayed in the presence
thioesters are also presumed to exist. A common biosynthetic
strategy for the formation of malonyl- or methylmalonyl-CoA
entails the carboxylation of acetyl- or propionyl-CoA, respectively.
However, harnessing this pathway to extend the in vivo pool of
R-carboxylated CoA thioesters would require an enzyme that
would carboxylate a wide variety of CoA-linked acids with
substituents directly at the reactive site. Recently, an alternative
pathway for malonyl-CoA biosynthesis has been discovered in
Rhizobium trifolii, in which exogenous malonate is imported via
a membrane-bound dicarboxylate transporter protein, and is
directly activated into malonyl-CoA by an ATP-dependent
1
1,12
of the enzyme, ATP, and CoASH as described previously.
(
6) See for example: Patterson, F. L. M.; Buchanan, R. L.; Dean, F. H.
Can. J. Chem. 1965, 43, 1700-1713.
(7) Pohl, N. L.; Gokhale, R. S.; Cane, D. E.; Khosla, C. J. Am. Chem. Soc.
1
998, 120, 11206-11207.
(
(
8) Christensen, J. E.; Goodman, L. Carbohydr. Res. 1968, 7, 510-512.
9) Procedure for deprotection of isopropylidene-protected malonic acid
deriVatiVes. Isopropyl Meldrum’s acid (1b) (200 mg, 1.07 mmol) was treated
with water (0.2 mL) and trifluoroacetic acid (1.8 mL). The reaction mixture
was allowed to sit at ambient temperature for 15 min. Solvent was removed
under reduced pressure. Diethyl ether (5 mL) was added to the resulting oil
4
malonyl-CoA synthetase. Heterologous expression of these two
genes in a recombinant strain of Streptomyces coelicolor which
produces 6-deoxyerythronolide B has been shown to result in
and then removed under reduced pressure to leave a white amorphous solid
1
5
(2b) (154 mg, 1.05 mmol, 98%). 2b: (400 MHz H NMR, DMSO-d
6
) δ 2.94
dramatic improvements in macrolide productivity, indicating that
(
d, J ) 8.8 Hz, 1 H), 2.2-2.1 (m, 1 H), 0.93 (d, J ) 6.7 Hz, 6 H); (100 MHz
this precursor pathway can also utilize methylmalonate in addition
to malonate. We have therefore investigated the substrate specific-
13
1
C NMR, DMSO-d
DMSO-d ) δ 3.27 (t, J ) 7.3 Hz, 1H), 1.63 (dd, J ) 7.3 Hz, 2 H), 0.92 (m,
H), 0.38 (m, 2 H), 0.08 (m, 2 H); (100 MHz C NMR, DMSO-d
53.0, 34.3, 9.8, 5.2.
6
) δ 171.3, 60.0, 29.0, 21.4. 2c: (400 MHz H NMR,
6
1
3
2
6
) δ 171.8,
†
Department of Chemical Engineering, Stanford University.
‡
Department of Biochemistry, College of Science, Protein Network
(10) Kim, Y. S.; Kang, S. W. Biochem. J. 1994, 297, 327-333.
(11) Stock solutions of the diacids (pH 6-7) were made using aqueous
sodium hydroxide. A direct spectrophotometric assay method was used as
Research Center, Yonsei University.
§
Department of Chemistry, Brown University.
|
11
Departments of Chemistry and Biochemistry, Stanford University.
previously reported. This method is based on the measurement of the increase
(
(
1) Hopwood, D. A. Chem. ReV. 1997, 97, 2465.
in absorbance at 232 nm by the formation of thioester bond of malonyl-CoA.
The incubation mixture for this assay contains (in micromoles) potassium
2) Khosla, C.; Gokhale, R.; Jacobsen, J. R.; Cane, D. E. Annu. ReV.
Biochem. 1999, 68, 219.
2
phosphate buffer, pH 7.2, 100; sodium malonate-derivative, 10; MgCl , 2;
(
3) (a) Kim, Y. S.; Chae, H. Z. Biochem. J. 1991, 273, 511. (b) Kakavas,
ATP, 0.4; coenzyme A, 0.2; and enzyme (0.9 µg) and water in a total volume
of 1.0 mL. Control reaction mixture contains 10 µmol of malonate instead of
sodium malonate-derivatives. The rate of increase in absorbance at 232 nm is
recorded by a spectrophotometer equipped with 30 °C circulator. The molar
extinction coefficient of thioester bond of malonyl-CoA at 232 nm is 4500
S. J.; Katz, L.; Stassi, D. J. Bacteriol. 1997, 179, 7515. (c) Motamedi, H.;
Shafiee, A. Eur. J. Biochem. 1998, 256, 528. (d) Wu, K.; Chung, L.; Revill,
W. P.; Katz, L.; Reeves, C. D. Gene 2000, 251, 81.
(
(
4) An, J. H.; Kim, Y. S. Eur. J. Biochem. 1998, 257, 395-402.
-1
-1
5) Lombo, F.; Pfeifer, B.; Leaf, T.; Ou, S.; Kim, Y. S.; Cane, D. E.; Licari,
M
cm
.
P.; Khosla, C. Biotech. Prog. in press (2001).
(12) Kim, Y. S.; Bang, S. K. Anal. Biochem 1988, 170, 45-49.
1
0.1021/ja0028368 CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/26/2001

Communications to the Editor
J. Am. Chem. Soc., Vol. 123, No. 24, 2001 5823
Table 1.
Surprisingly, the enzyme converted all these diacids into their
corresponding monothioesters (Table 1). Moreover, except for
benzylmalonate, in all cases the bimolecular rate constants for
product formation were greater than 1% that for the natural
(0.5 mM for methylmalonyl-CoA, 5 mM for other thioesters),
the relative product yields could be estimated at the end of the
incubation period. The molar ratios of products derived from
methylmalonyl-CoA, ethylmalonyl-CoA, propylmalonyl-CoA and
malonyl-CoA were 100:8:6:<0.1, respectively. Thus, although
this methymalonyl-CoA specific module can discriminate against
both smaller and bulkier substituents at the R-carbon, it appears
to have greater tolerance toward unnatural precursors that are
ordinarily not available in a cell.
substrate, malonate. As a point of reference, the kcat and K
M
for
-1
malonate are 19 ( 0.3 s and 208 ( 6 µM, respectively, whereas
-
1
the kcat for a typical PKS is in the range of 1 min . Therefore,
rate constants equivalent to 1% that for the natural substrate should
be adequate to ensure nonlimiting generation of a desired substrate
in vivo.
In summary, the R. trifolii malonyl-CoA synthetase provides
an attractive route to expand the lexicon of precursors available
for polyketide biosynthesis. Parenthetically, we note that several
orthologs of this enzyme have recently been reported in the
The above results suggest that malonyl-CoA synthetase has
broad tolerance toward unnatural diacid substrates, especially
those with hydrocarbon side chains. Moreover, the enzyme does
not appear to have a fixed angle requirement for the two
carboxylate functionalities, as both cyclobutyl dicarboxylic acid
and cyclopropyl dicarboxylic acid were turned over. Most
interestingly, the enzyme can handle substrates such as allylma-
lonate and cyclopropylmethylenemalonate which, if incorporated
into polyketide backbones, could provide functional handles for
regioselective modifications.
As mentioned above, acyltransferase domains in PKS modules
are the principal determinants of extender unit selectivity. These
domains can strongly discriminate against alternative extender
units available in a cell.13 To test whether PKS modules might
be able to process these unusual extender units, methylmalonyl-
CoA, ethylmalonyl-CoA, and propylmalonyl-CoA were synthe-
sized using malonyl-CoA synthetase and assayed against the
terminal module of the 6-deoxyerythronolide B synthase using
16
genomic literature with approximately 30% sequence identity;
these enzymes may have complementary substrate preferences
to the enzyme described here. Since PKS modules appear to have
modest tolerance toward some of these unnatural precursors,
coexpression of PKSs with slightly altered acyltransferase selec-
tivity and malonyl-CoA synthetase could yield polyketides with
unnatural functional groups.
Acknowledgment. This work was funded by NIH Grants to C.K.
(CA66736) and to D.E.C. (GM22172) and Grants from the Korea Science
and Engineering Foundation (97-05-01-06-01-5) to Y.S.K. N.L.P. is the
recipient of a National Institute of General Medical Studies Postdoctoral
Fellowship (1 F32 GM19540-02). We thank Allis Chien at the Stanford
Mass Spectrometry Laboratory for carrying out the mass spectrometric
measurements.
(
2S,3R)-2-methyl-3-hydroxy-pentanoyl-S-N-acetylcysteamine as
14
JA0028368
the source of primer units. Mass spectrometric analysis of
overnight incubations of each reaction mixture revealed the
expected triketide lactone product. This assay could not be used
for accurate kinetic analysis. However, under saturation conditions
(
15) Observed masses (ESI-MS): 172.0, 185.1, and 199.0 for the methyl-
15
malonyl-, ethylmalonyl- and propylmalonyl-derived â-ketolactones, respec-
tively. In addition, tandem MS analysis of the molecular ions yielded the
2
expected M-H O species. (described in Weissman, K. J.; Kearney, G. C.;
(
13) For example, the methylmalonyl-CoA-dependent acyltransferase do-
main from the terminal module of the 6-deoxyerythronolide B synthase is
60-fold more selective toward its natural substrate than toward malonyl-
CoA (Lau, J. Ph.D. Thesis, Stanford University, 2000).
14) Assay conditions are described in Wu, N.; Kudo, F.; Cane, D. E.;
Khosla, C. J. Am. Chem. Soc. 2000, 122, 4847.
Leadlay, P. F. and Staunton, J. Rapid Commun. Mass Spectrom. 1999, 13,
2103-2108).
3
(16) For example, see: (a) Fuldam, M.; Heinz, E.; Wolter, F. P. Mol. Gen
Genet. 1994, 242, 241-249. (b) Blobel, F.; Erdmann, R. Eur. J. Biochem.
1996, 240, 468-476. (c) Koo, H. M.; Kim, Y. S. Arch. Biochem. Biophys.
2000, 378, 167.
(