Crotonase catalysis enables flexible production of functionalized prolines and carbapenams

Article abstract of DOI:10.1021/ja208318d

The biocatalytic versatility of wildtype and engineered carboxymethylproline synthases (CMPSs) is demonstrated by the preparation of functionalized 5-carboxymethylproline derivatives methylated at C-2, C-3, C-4, or C-5 of the proline ring from appropriately substituted amino acid aldehydes and malonyl-coenzyme A. Notably, compounds with a quaternary center (at C-2 or C-5) were prepared in a stereoselective fashion by engineered CMPSs. The substituted-5-carboxymethyl-prolines were converted into the corresponding bicyclic β-lactams using a carbapenam synthetase. The results demonstrate the utility of the crotonase superfamily enzymes for stereoselective biocatalysis, the amenability of carbapenem biosynthesis pathways to engineering for the production of new bicyclic β-lactam derivatives, and the potential of engineered biocatalysts for the production of quaternary centers.

Full text of DOI:10.1021/ja208318d

ARTICLE
pubs.acs.org/JACS
Crotonase Catalysis Enables Flexible Production of Functionalized
Prolines and Carbapenams
Refaat B. Hamed,†,‡ Luc Henry,† J. Ruben Gomez-Castellanos,† Jasmin Mecinoviꢀc,† Christian Ducho,†,§
||
John L. Sorensen,†, Timothy D. W. Claridge,† and Christopher J. Schofield*,†
†Department of Chemistry, University of Oxford, 12 Mansfield Road, Oxford OX1 3TA, United Kingdom
‡Department of Pharmacognosy, Faculty of Pharmacy, Assiut University, Assiut 71526, Egypt
S Supporting Information
b
ABSTRACT: The biocatalytic versatility of wildtype and en-
gineered carboxymethylproline synthases (CMPSs) is demon-
strated by the preparation of functionalized 5-carboxymeth-
ylproline derivatives methylated at C-2, C-3, C-4, or C-5 of the
proline ring from appropriately substituted amino acid alde-
hydes and malonyl-coenzyme A. Notably, compounds with a
quaternary center (at C-2 or C-5) were prepared in a stereo-
selective fashion by engineered CMPSs. The substituted-
5-carboxymethyl-prolines were converted into the correspond-
ing bicyclic β-lactams using a carbapenam synthetase. The results demonstrate the utility of the crotonase superfamily enzymes for
stereoselective biocatalysis, the amenability of carbapenem biosynthesis pathways to engineering for the production of new bicyclic
β-lactam derivatives, and the potential of engineered biocatalysts for the production of quaternary centers.
’ INTRODUCTION
isomers are also resistant to dehydropeptidase hydrolysis but
their antibacterial activities are reduced17).
Proline analogues are constituents of numerous natural pro-
ducts,1 such as siderochelin A,2 bottromycin A2,3 roseotoxin B,4,5
kainic acid,6À8 and lactacystin.9,10 Proline derivatives are also
widely used as pharmaceuticals, such as captropril and saxagliptin,
and in biomedical research; for example, α-substituted prolines11
are used as templates in structureÀfunction relationship studies
directed toward elucidation of biologically active conforma-
tions.12 In organic synthesis, proline and proline analogues are
also used as catalysts for asymmetric synthesis.13
Carbapenems are an important class of β-lactam antibiotics
that are often used to treat serious infections including those
involving multidrug resistant bacteria.14 In contrast to many
β-lactam antibiotics, that is, the penicillins and cephalosporins
which are obtained by fermentation or from fermented precur-
sors, clinically useful carbapenems are produced by chemical syn-
thesis, which limits the range of compounds that can be produced
efficiently. Hence, there is interest in optimizing biocatalytic
routes to produce sufficient quantities of useful antibiotics or
intermediates for use in semisynthetic approaches to carbapenems
and other β-lactams. The substitution pattern on the bicyclic
β-lactam antibiotics affects both their antibacterial activity and
pharmacokinetic properties.15 StructureÀactivity relationship
studies on thienamycins have revealed that the introduction of
a 1β-methyl substituent results in derivatives with superior
antibacterial activity and increased resistance to inactivation by
renal dehydropeptidases.16,17 Therefore, most commercially
available carbapenems, for example, ertapenem, meropenem,
and doripenem, possess a 1β-methyl substituent (the α-methyl
Three enzymes (CarB, CarA, and CarC)18,19 are required for
biosynthesis of the simplest carbapenem (5R)-carbapen-2-em-
3-carboxylate 1 in Pectobacterium carotovorum (Scheme 1). How-
ever, more enzymes are involved in biosynthesis of the C-2,
C-6-substituted carbapenem thienamycin 2 in Streptomyces cattleya.20
In both cases, an early ring-forming step is catalyzed by carbox-
ymethylproline synthase (CMPS), a member of the crotonase
superfamily:21 CarB in P. carotovorum22À25 and ThnE in S.
cattleya.26 CarB and ThnE catalyze conversion of malonyl-CoA
3 and L-glutamate semialdehyde/5-hydroxyproline/pyrroline-
5-carboxylate (collectively L-GHP) to give (2S,5S)-carboxy-
methylproline (t-CMP) 4. CarB/ThnE catalysis proceeds via
decarboxylation of 3 to give an enolate, the anion of which is
bound to an oxyanion hole formed by residues Gly62CarB
/
Gly107ThnE and Met108CarB/Val153ThnE (Figure 1). CÀC bond
formation then proceeds via reaction of the formed enolate with
the imine form of L-GHP followed by thioester hydrolysis of the
acyl intermediate to give 4 and CoASH22,23 (Scheme 1).
The crotonase superfamily is ubiquitous and is characterized
by a conserved repeated ββα-crotonase fold.21 In most cases,
their catalytic mechanisms are proposed to proceed via an
enolate intermediate, which is generated by decarboxylation of
malonyl-CoA (derivatives), as for CarB and ThnE. The alkyla-
tion of enolates is an important reaction in organic chemistry, yet
Received: September 9, 2011
Published: November 17, 2011
r
2011 American Chemical Society
471
dx.doi.org/10.1021/ja208318d J. Am. Chem. Soc. 2012, 134, 471–479
|

Journal of the American Chemical Society
ARTICLE
Scheme 1. Role of the Carboxymethylproline Synthases CarB and ThnE in the Biosynthesis of (5R)-Carbapen-2-em-3-carboxylic
Acid 1 and Thienamycin 2, Respectively
Figure 1. CarB and ThnE are crotonase superfamily enzymes. (A) View derived from a crystal structure of CarB25 with malonyl-CoA 3 and pyrroline-
5-carboxylate (L-P5C) modeled into the active site. The view highlights some of the residues proposed to be important in substrate binding/catalysis. Some of
these residues were targeted for substitution. ThnE residues analogous to those of CarB active site are shown in blue;26 (B) A list of the CarB/ThnE variants
used in this study. The figure was generated using Pymol (www.pymol.org), Clustal W,35 and Genedoc (http://www.nrbsc.org/gfx/genedoc/).
has been somewhat unexplored in biocatalysis (with one excep-
tion being the use of engineered aldolases27À30 for stereoselec-
tive CÀC bond formation), particularly with respect to the
crotonase superfamily. Recently, we have reported that wildtype
and variants of CarB/ThnE can accept analogues of L-GHP with
different chain lengths to give 6- and 7-membered carboxy-
methyl-N-heterocycles.31 Furthermore, use of appropriate var-
iants enabled the stereoselective alkylation of enolates generated
from C-2-alkylated malonyl-CoA.32 These findings stimulated us
to investigate other potential substrate analogues for CMPSs
with a view to exploring their potential as biocatalysts for the
production of 5-carboxymethylprolines substituted at any posi-
tion of the proline ring including the production of compounds
with a quaternary center at C-2 or C-5.
We now report on the use of CMPSs to prepare functional-
ized prolines substituted at C-2, C-3, C-4, and C-5 of the proline
ring. The potential therapeutic utility of the resultant products
is demonstrated by their conversion into the respective bicyclic
β-lactam derivatives by carbapenam synthetase catalysis. The
results exemplify the biocatalytic versatility of CMPSs and more
generally the applicability of crotonases in biocatalysis.
’ RESULTS AND DISCUSSION
Preparation of C-4-methylated t-CMP derivatives. In order
to test for the ability of CMPSs to accept C-4-methyl-substituted-
L-GHP derivatives as substrates, and with a view to preparing
1β-methyl-carbapenams, we synthesized protected forms 5À7 of
the C-4-methyl- and C-4-dimethyl-substituted-L-GHP deriva-
tives 8À10 starting from pyroglutamate33 (Scheme 2A); com-
pounds 5-7 were deprotected under standard acidic conditions.34
Upon incubation with CarB and 3, 4,4-dimethyl-L-GHP 10 was
converted to a single product (48% isolated yield, Table 1: entry
1, Table S1) assigned as 4,4-dimethyl-t-CMP 11 on the basis of
MS and NMR analyses (Scheme 3I, Figures 2A and S1). The (S)
stereochemistry at C-5 of the product, and all subsequently
described t-CMP derivatives, was assigned based on coupling
constant(J) values and 2D NOESY analyses (Figure S1), assuming
retention of stereochemistry at C-2. When diastereomerically pure
(>95% by 1H NMR) (4S)-4-methyl-N-Boc-L-GHP tert-butyl ester
5 and (4R)-4-methyl-N-Boc-L-GHP tert-butyl ester 6 were de-
protected and incubated with CarB and 3, they each gave a ∼1:1
mixture of the C-4 epimeric products, (4R)-4-methyl-t-CMP 12
472
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
Scheme 2. Synthesis of Potential CMPS Substratesa
a (A) 4-methyl-L-GHP derivatives. (i) Isobutylene, H2SO4, 1,4-dioxane, À78 °C to rt, 6 d, 51%; (ii) di-tert-butyl dicarbonate, 4-dimethylaminopyridine,
Et3N, CH2Cl2, 3 d, rt, 85%; (iii) lithium bis(trimethylsilyl)amide, methyl trifluoromethanesulfonate, toluene/THF 2:1, À78 °C, 6 h, 21-62%;
(iv) lithium triethylborohydride, THF, À78 °C, 84-87%; (v) 10% aq. HCOOH, 60 °C, 1 h, app. quant. An ∼1:1 mixture of both 8 and 9 was obtained
upon deprotection of either pure 5 or pure 6. (B) 3-Methyl-L-GHP derivatives. (vi) Di-tert-butyl dicarbonate, Et3N, CH2Cl2, rt, 97%; (vii) Dess-Martin
periodinane, CH2Cl2, rt, 96%; (viii) methyl(triphenylphosphoranylidene)acetate, toluene, reflux, 62%; (ix) H2 (g), palladium on carbon, ethyl acetate,
rt, 98%; (x) di-tert-butyl dicarbonate, 4-dimethylaminopyridine, Et3N, CH2Cl2, 85%; (xi) diisobutylaluminium hydride, Et2O, À78°C, 70%; (xii) 10% aq.
HCOOH, 60 °C, 1 h, app. quant. (C) 2-Methyl-L-GHP. (xiii) Di-tert-butyl dicarbonate, Et3N, 1,4-dioxane, rt; (xiv) 1-ethyl-3-(3-dimethylaminopro-
pyl)carbodiimide, 4-dimethylaminopyridine, 1:1 tert-butanol/CH2Cl2, rt, 78% over 2 steps; (xv) n-butyllithium, diisopropylamine, MeI, THF, À78 °C,
65%; (xvi) ruthenium tetroxide in situ from ruthenium(IV) oxide and sodium periodate, 1:4 ethyl acetate/water, rt, 85%; (xvii) lithium
triethylborohydride, THF, À78 °C, 68%; (xviii) 10% aq. HCOOH/, 60°C, 1 h, app. quant. (D) 5-Methyl-L-GHP derivatives. (xix) MeMgBr, THF,
87%; (xx) 10% aq. HCOOH, 60 °C, 1 h, app. quant. (xxi) 10% deuterated-formic acid (2HCOO2H) in 2H2O, 60 °C, 1 h, app. quant.
473
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
Table 1. Chemoenzymatic Synthesis of Functionalized 5-Carboxymethylprolines from Malonyl-CoA 3 and Methylated-L-GHP by
CMPSse
a d.r.: diastereomeric ratio as determined by 1H NMR spectroscopy, under standard incubation conditions. b The % yield (isolated) was determined over
the following steps: deprotection of amino acid aldehydes, incubation with enzyme, LC-MS purification, and lyophilization; products were quantified by
1H NMR using [2H]4-trimethylsilylpropionate as an external standard.31 c The starting materials (17, 19, and 23) were made as racemates; taking this
into account, the yields are given in parentheses. d No product was detected. e CMPSs giving the highest yield and/or diastereomeric excess are shown
(shaded) in comparison to wild-type CarB and ThnE.
and (4S)-4-methyl-t-CMP 13, as shown by LC-MS and 1H NMR
analyses (Scheme 3II, Figures 2A, S2 and S3). Evidence that C-4
epimerization occurred during the acid mediated deprotection
(likely via enol and/or enamine formation) came from deprotec-
tion of N-Boc-L-GHP tert-butyl ester in 2HCOO2H/2H2O which
led to the incorporation of two deuterium atoms at C-4 of L-
GHP.34 Given that previous studies have shown the potential of
CMPS variants to select from equilibrating mixtures of alkylma-
lonyl-CoA derivatives,32 we then investigated whether the stereo-
selective production of 4-methyl-t-CMP from a mixture of
4-methyl-L-GHP epimers (8/9) is possible. A set of active site
CarB variants (prepared as described,32 Figure 1A) were then
screened by analytical LC-MS for the formation of 12 and 13
from C-4 epimeric 8/9 and 3, under standard incubation con-
ditions. The results reveal that all tested CMPS enzymes are able
to form the two C-4 epimers of 4-methyl-t-CMP, but with varying
yields and diastereomeric ratios (4S:4R) as follows (Table S2,
Figure 2B): CarB M108A (54:46), CarB M108L (56:44), CarB
M108V (93:7), CarB M108I (90:10), CarB Q111N (70:30),
CarB W79F (61:39), CarB W79F/M108A (62:38), CarB W79A
(64:36), and CarB H229A (65:35). Likewise, screening of wild-
type ThnE and active site ThnE variants (Table S2, Figure S4)
reveals that they can catalyze the formation of the two C-4
epimers 12 and 13 with a particular bias toward the (4S)-epimer
13 (d.r. g 90 (4S):10 (4R)) with the exception of ThnE V153A
(d.r. = 74 (4S):26 (4R)). In case of both ThnE V153I and ThnE
H274A, the d.r. was g98 (4S):2 (4R) (Table S2, Figures 2B and S4).
For most of the CMPS variants tested, the reactions were scaled
up and the diastereomeric ratio (d.r.) of the two epimers con-
firmed by 1H NMR analyses (Table S2, Figure 2B). These results
demonstrate the ability of CMPS to stereoselectively produce
4-methyl-t-CMP derivatives from equilibrating mixtures of semi-
aldehyde substrates likely through a dynamic kinetic resolution
process. It is notable that all CMPSs (including wildtype ThnE)
with a β-branched residue (i.e., Val or Ile) at position 108CarB
/
153ThnE (one of the oxyanion hole forming residues) produced
13 as the major C-4-epimer, demonstrating that subtle active site
variations can have large effects on the stereochemical outcome
of CMPS-catalysis. Consistent with these results, the yield of 11
(Table S1) by CMPSs that favored the formation of 13 were
considerably lower than that of wildtype CarB (which produced
12 and 13 in near 1:1 ratio). For example, the yield of 11 by CarB
was ∼4À5 times higher than that of CarB M108 V/I (Table S1).
In case of CarB variants, a steric clash between the methyl-sub-
stituent on the (4R)-4-methyl-L-GHP 8 and β-branched residues
at position 108CarB provides a possible explanation for these
selectivities.
Preparation of C-3-methylated-t-CMP derivatives. Interest
in constructing a CMP-derivative substituted at C-3 arises
because clinically useful carbapenems are substituted at an equivalent
474
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
Scheme 3. Products of CMPS and CarA Catalysis (i.e., t-CMP and Bicyclic β-Lactam Derivatives, Respectively) Reported in This
Study
Figure 2. Improving the diastereoselectivity of CMPSs toward (4S)-4-methyl-t-CMP 13. (A) 1H NMR spectra of the C4-methylated-t-CMPs resulting
from incubation of malonyl-CoA 3 and C4-methylated-L-GHP derivatives (8 to 10) in the presence of the shown CMPSs. (B) Part of the 1H NMR
spectra for the two C4-epimers of 4-methyl-t-CMP (12 and 13) resulting from incubation of 3 and 4-methyl-L-GHP 8/9 in the presence of the shown
CMPSs, under standard conditions, demonstrating that stereoselective production of C4-epimers can be achieved by use of appropriate variants.
position (with a substituted cysteaminyl moiety).36 Both epimers
of 3-methyl-L-GHP were obtained, in racemic form, from com-
mercially available L-threonine-OtBu ester in seven steps (23%
overall yield, Scheme 2B). Oxidation of the protected amino acid
14 with Dess-Martin reagent37 gave ketone 15 which was reacted
under Wittig conditions to give a separable 1:1 mixture of E- and
Z-isomers of alkene 16. Hydrogenation, further protection,
reduction, and finally deprotection of Z-16 and E-16 gave,
respectively, racemic (3S)-3-methyl-L-GHP 17/18 and racemic
(3R)-3-methyl-L-GHP 19/20 (Scheme 2B). Incubation of 17/
475
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
1
Figure 3. Spectroscopic analyses of the C2-, C3-, and C5-methyl-substituted-t-CMP products. H NMR spectra of methylated-t-CMP derivatives
resulting from incubation of malonyl-CoA 3 and the shown methylated-L-GHP derivatives in the presence of the shown CMPSs, under standard
conditions.
18 and 3 in the presence of either wildtype or engineered CMPS
variants resulted in the observation of a single major chromato-
graphic peak with the anticipated mass (m/z = 188
[M+1]+) as revealed by analytical LC-MS (Figure S5). Scale-
up and LC-MS-guided purification of the product of catalysis of
wildtype CarB, which produced the highest relative yield com-
pared to the other tested CMPSs (Table 1: entry 6 and Table S3),
led to isolation of a product which was then subjected to NMR
analyses to assign its C-3 stereochemistry. Assuming the stere-
ochemistry at C-2 is (S) based on reports that D-GHP is not a
substrate for either CarB or ThnE,26,38 2D NOESY analyses
enabled assignment of the product as (3S)-3-methyl-t-CMP 21
(Scheme 3III, Figures 3 and S6). Likewise, incubation of 19/20
and 3 in the presence of wildtype CarB provided a product that
was identified as (3R)-3-methyl-t-CMP 22 (Scheme 3IV, Fig-
ures 3 and S7, Table 1: entry 8 and Table S3). These results
demonstrate that CMPSs can catalyze the formation of C-3-
substituted-t-CMP derivatives without loss of stereochemistry.
Preparation of C-2 and C-5-methylated-t-CMP deriva-
tives. Construction of quaternary stereocenters is a significant
challenge in the synthesis of heterocycles and natural products.39
We therefore investigated whether CMPSs can catalyze the pro-
duction of heterocycles with asymmetric quaternary centers, start-
ing with the C-2 position. Racemic 2-methyl-GHP 23 was syn-
thesized from L-proline in 6 steps (29% overall yield, Scheme 2C)
via C-2 alkylation of protected L-proline 24 through the generation
of a tertiary carbanion. Subsequent oxidation using ruthenium
tetroxide40 and reduction with lithium triethylborohydride41
gave the protected form of 2-methyl-GHP 25. Scale-up and
purification of the product of incubation of 23 and 3 in the
presence of the CarB H229A variant, which produced the highest
relative yield compared to other CMPS variants tested (Figure 4A,
Table 1: entry 12), followed by NMR analyses revealed a trans-
relationship between H-5 and the methyl group at C-2 of the
product (Scheme 3V, Figures 3 and S8). Therefore, the product
of CarB H229A catalysis was assigned as (2S)-2-methyl-t-CMP
26. This result demonstrates the capability of CMPSs to catalyze
the formation of t-CMP derivatives with a quaternary center at
C-2 and (at least for these substrate analogues) eliminates a
mechanism involving proton abstraction at C-2. The high yield in
the case of the CarB H229A variant may reflect (although other
factors may contribute) an increased active site volume that
enables productive binding of the quaternary C-2 stereocenter
(Figure 4B). It is notable that the CarB H229A variant provides
access to two t-CMP analogues (i.e., 26 and (2S,7S)-7-(carbo-
xymethyl)azepane-2-carboxylic acid31), both not efficiently pro-
duced by other CMPSs tested. The common theme between the
starting materials of the two products (i.e., 23 and L-aminopi-
melate semialdehyde31) is their increased steric demand com-
pared to the natural substrate L-GHP.
To test for the production of C-5-methylated-t-CMP,
5-methyl-L-GHP 27 was synthesized by Grignard addition to the
protected L-pyroglutamate 28 followed by deprotection (Scheme 2D).
Screening of the CMPS variants led to the identification of ThnE
V153A as the variant with the highest relative yield (Figure 4C,
Table 1: entry 15). Notably, with wildtype CMPSs, yields for the
formation of 5-methyl-t-CMP were very low (ThnE) or non-
existent (CarB) (Figure 4C, Table 1: entries 13 and 14) possibly
476
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
Figure 4. Improving the yield of quaternary center-containing t-CMP derivatives by engineered CMPSs. (A) Ion-extracted LC-MS chromatograms
(positive electrospray ionization) displaying the relative yields for the formation of (2S)-2-methyl-t-CMP 26 (red peak) from racemic 2-methyl-GHP 23
and malonyl-CoA 3 catalyzed by the shown CMPSs. ThnE V153L, ThnE V153M, ThnE H274A, and ThnE V153A also produce 26, and the relative
yields are 17%, 13%, 12%, and 7%, respectively. (B) A view derived from a CarB crystal structure25 with the acetyl-CoA enolate and the imine form of
2-methyl-L-GHP 23 modeled into the active site. His229 is predicted to be one of the closest residues to the imine form of 23 (∼3 Å from τ-nitrogen of
His229 to the methyl group at C-2 of 23). (C) Ion extracted LC-MS chromatograms (positive electrospray ionization) showing the relative yields for the
production of (5S)-5-methyl-t-CMP 29 (red peak) resulting from the incubation of 5-methyl-L-GHP 27, 3 and the shown CMPSs. ThnE H274A
and ThnE W124F also produce 29 and the relative yields are 7% and 3%, respectively. p-Aminosalicylic acid was used as an internal standard
(IS, blue peak).
2
reflecting the fact that the nucleophilic attack of the intermediate
enolate occurs at C-5 of the imine form (L-P5C)22 (the electron
donating effect of the methyl group can, in part, affect the CÀC
bond forming reaction). The yield of the ThnE V153A variant
(Figure 4C, Table 1: entry 15) was ∼10 times higher than that of
wildtype ThnE revealing that substitution can alter yield sub-
stantially. Scale-up and LC-MS-guided purification of ThnE
V153A product of catalysis led to isolation of 5-methyl-t-CMP,
which is characterized by the presence of an AB quartet for the
diastereotopic H-6 protons (δH 2.47, 2.53, J = 16 Hz) in its 1H
NMR spectrum (Figure 3). 2D NOESY analyses (Figures S9 and
S10) reveal a trans-relationship between H-2 and the methyl
group at C-5, identifying the product as (5S)-5-methyl-t-CMP 29
(Scheme 3VI). It is notable that CMPS variants can accommo-
date such an increase in steric bulk adjacent to the site of CÀC
bond formation.
10% deuterated-formic acid (2HCOO2H) in H2O; substantial
incorporation of five deuterium atoms (>90% 2H) at C-4/C-6 of
5-methyl-L-GHP was detected by LC-MS analyses (Scheme 2D,
Figure S12B). Incubation of C-2 [2H]-methylmalonyl-CoA and
31 with ThnE/ThnE variants revealed the formation of a product
observed as a single chromatographic peak with mass spectral
data implying the formation of a product with six deuterium
atoms as the major product (m/z = 208 [M+H]+, Figure S12E).
This result eliminates mechanisms involving attack of the enolate
form, derived by decarboxylation of methylmalonyl-CoA, on the
ring open form of 27 (5-methyl-L-GSA), followed by desatura-
tion and intramolecular addition to form 5-methyl-t-CMP-CoA,
and supports a mechanism involving attack of the enolate on the
imine form of 27 (5-methyl-L-P5C) (Figure S12III), analogous
to that of L-GHP.22,23
Conversion of the Prepared t-CMP Derivatives into Bicy-
clic β-Lactams. β-Lactams are finding new therapeutic applica-
tions in addition to their established use as antibiotics via inhi-
bition of penicillin binding proteins (PBPs). β-Lactams are
clinically used as β-lactamase inhibitors.42 They have also been
developed as protease inhibitors43 (e.g., elastase), as cholesterol
We also investigated whether the 5-methyl-L-GHP 27 sub-
strate analogue follows the same mechanism as L-GHP,22,23
employing labeled [2H]-methylmalonyl-CoA (Figure S11) and
4,4,6,6,6-[2H]5-5-methyl-L-GHP 31. The latter was prepared by
deprotection of N-Boc-5-methyl-L-GHP tert-butyl ester 30 in
477
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
absorption inhibitors44 (e.g., ezetimibe, a monobactam45), as a
delivery vehicle for anticancer drugs,43 and to regulate levels of
the excitatory neurotransmitter glutamate46 (via interaction
with the glutamate transporter GLT1). The production of bicyclic
β-lactams with quaternary centers on the ring core structure is of
particular interest because such compounds can show improved
stability to hydrolysis.47,48 There is precedent for the biocatalytic
production of bicyclic β-lactams with quaternary centers in work
on isopenicillin N synthase (IPNS). IPNS is a promiscuous oxidase
catalyzing production of the penicillin nucleus from a tripeptide
precursor δ-(L-α-aminoadipoyl)-L-cysteinyl-D-valine (ACV).49,50
Substitution of L-cysteine for (3R)-methylcysteine in ACV and
incubation with IPNS allowed the production of 5α-methyl
penicillin.48 The product showed no antibacterial activity but
was noticeably more stable than penicillin G to β-lactamase 1
from Bacillus cereus.48 The seven carboxymethyl-functionalized
N-heterocycles, generated by CMPSs catalysis in this study (i.e.,
11, 12, 13, 21, 22, 26, and 29), were tested as substrates for the
β-lactam ring forming enzyme carbapenam synthetase (CarA)
from P. carotovorum.41,51 CarA catalyzed the conversion of all
tested t-CMP derivatives to the respective bicyclic β-lactam
derivatives (32 to 38, respectively) as demonstrated by LC-MS
analyses (Scheme 3, Figure S13). In the case of the CarA
catalyzed reaction of a ∼1:1 mixture of the C-4 epimers 12
and 13, produced by wildtype CarB catalysis, CarA exhibited
higher diastereoselective bias toward the (4S)-epimer 13 as a
substrate (d.r. of products =24 (33):76 (34), as determined by
LC-MS under standard assay conditions, Figure S14). Notably,
the major product ((1S)-1-methyl-carbapenam 34) has the same
stereochemistry at C-4 as many commercially available carbape-
nems with improved stability toward dehydropeptidases.52 We
also found that 6,6-dimethyl-t-CMP 39,21 another CarB/CarB
variant catalytic product, is a substrate for CarA producing the
6,6-disubstituted carbapenam 40 (Scheme 3VII, Figure S13). In
the cases of 11 and 39, the CarA reaction was scaled-up and
the structures of the LC-MS isolated products (32 and 40,
respectively) were confirmed by NMR analyses (Figures S15
to S17). The C-5 stereochemistry of the product of CMPS
catalysis has to be inverted during carbapenem biosynthesis53,54
to confer antibacterial activity to those compounds. Thus, as
anticipated, the substituted carbapenams were not active anti-
biotics as found by holed-plate assays versus cephamycin C (data
not shown). Although the yields obtained for the methylated
bicyclic carbapenams varied, these results demonstrate the capa-
bility of the carbapenam synthetase CarA for converting sub-
strate analogues of t-CMP, and it is likely that yields and/or
diastereoselectivities can be improved by active site residue
substitutions. These results further expand the scope of substrates
for CarA which has been shown to accept at least three of the four
possible isomers of CMP41 and the 6- and 7-membered ring
analogues of t-CMP.31
stable to hydrolysis (e.g., by renal dehydropeptidase I) compared
to the analogous 5-unsubstituted carbapenems.47 Thus, CMPSs/
CarA may be used to prepare bicyclic β-lactams with improved
stabilities.
’ CONCLUSIONS
The results presented here demonstrate the capacity of CMPSs
for the preparation of t-CMP derivatives functionalized at all four
carbon positions of the proline ring, including the sterically
demanding C-2 and C-5 positions. The use of active site variants
enabled the generation of CMPS-catalysts operating with much
higher yields than wildtype enzymes (e.g., in case of C-2 and C-5-
methylated-t-CMP, Figure 4, Table 1: entries 10À15). Certain
variants are also capable of the stereoselective product formation
including via preferential conversion of one diastereomer from
an equilibrating mixture of C-4-methyl-L-GHP epimers 8/9
(Figure 2B, Table 1, entries 3À5), likely via a dynamic kinetic
resolution process. It is notable that single residue substitutions
in the wildtype CarB/ThnE active sites can lead to such a sub-
stantial improvement in yield and/or diastereoselectivity, and it
is plausible that further improvements can be achieved if CMPSs
are the subject of commercially focused work. Such studies would
also likely involve the use of modified enzyme in cell cultures, due
to the costs of coenzyme A derivatives. Despite these potential
limitations, CMPS catalysis in vitro is useful for the preparation
of functionalized prolines which are suitable for further mod-
ification as exemplified by the CarA-catalyzed production of
functionalized bicyclic β-lactams from the t-CMP products. Such
compounds are of interest with respect to inhibition of nucleo-
philic enzymes and are not readily accessible via synthesis, a
limitation that has hindered studies on their biological activities
and chemical properties.
The crotonase superfamily is large and present in most, if not
all, life forms. Somewhat surprisingly, to date, there have been
only limited studies on their utility as biocatalysts.31,32,55,56 The
combined work on CMPSs coupled to knowledge that catalysis
by most crotonase superfamily enzymes proceeds via conserved
mechanisms involving the generation of stabilized enolates,
suggests that the crotonase superfamily may be especially well
adapted for protein engineering studies aiming at the production
of useful small-molecules. Borrowing from a medicinal chemistry
term,57 we propose that certain enzyme families, including the
crotonase superfamily, may be “privileged” from a biocatalytic
perspective.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental procedures and
b
compound characterization data. This material is available free of
charge via the Internet at http://pubs.acs.org.
With regard to the effects of substitution on the properties of
carbapenams, during the NMR studies on the bicyclic products
of CarA catalysis, we observed that functionalization with two
methyl groups at C-6 (or to a lesser extent C-4) has a significant
stabilizing effect on the carbapenams with respect to the rate of
β-lactam hydrolysis: the t1/2 for 40 was ∼2 weeks; whereas that of
32 was ∼2 days (as determined by 1H NMR analyses at 4 °C and
pH ∼ 7). Although we did not carry out studies on the stability
of 5-methylcarbapenam prepared by CMPS/CarA catalysis,
5-methylcarbapenems have been prepared by total synthesis
and while possessing relatively low antibiotic activity are more
’ AUTHOR INFORMATION
Corresponding Author
christopher.schofield@chem.ox.ac.uk
Present Addresses
§Department of Chemistry, University of Paderborn, Warburger
Str. 100, 33 098 Paderborn, Germany.
334 Parker Building, Department of Chemistry, University of
Manitoba, Winnipeg, Manitoba, R3T 2N2, Canada.
478
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479

Journal of the American Chemical Society
ARTICLE
’ ACKNOWLEDGMENT
(31) Hamed, R. B.; Mecinovic, J.; Ducho, C.; Claridge, T. D. W.;
Schofield, C. J. Chem. Commun. 2010, 46, 1413–1415.
(32) Hamed, R. B.; Gomez-Castellanos, J. R.; Thalhammer, A.;
Harding, D.; Ducho, C.; Claridge, T. D. W.; Schofield, C. J. Nat. Chem.
2011, 3, 365–371.
(33) Charrier, J.-D.; Duffy, J. E. S.; Hitchcock, P. B.; Young, D. W.
J. Chem. Soc., Perkin Trans. 1 2001, 2367–2371.
(34) Ducho, C.; Hamed, R. B.; Batchelar, E. T.; Sorensen, J. L.;
Odell, B.; Schofield, C. J. Org. Biomol. Chem. 2009, 7, 2770–2779.
(35) Thompson, J. D.; Higgins, D. G.; Gibson, T. J. Nucleic Acids Res.
1994, 22, 4673–4680.
(36) Sasaki, A.; Sunagawa, M. Chem. Heterocycl. Compd. 1998, 34,
1249–1265.
(37) Kelly, G. T.; Sharma, V.; Watanabe, C. M. H. Bioorg. Chem.
2008, 36, 4–15.
(38) Sorensen, J. L.; Sleeman, M. C.; Schofield, C. J. Chem. Commun.
2005, 1155–1157.
(39) Christoffers, J.; Baro, A. Quaternary stereocenters: challenges and
solutions for organic synthesis; Wiley-VCH Verlag GmbH& Co. kgAa:
Weinheim, 2005.
We would like to thank Dr. Edward Batchelar and Mr. Daniel
Harding for their help during the course of this study. Biotech-
nology and Biological Sciences Research Council (RBH), the
Ministry of Higher Education (Egypt, RBH during D. Phil), Berrow
Foundation (LH), CONACyT and FIDERH (Mexico, RGC),
and the Deutsche Akademie der Naturforscher Leopoldina
(Germany, CD).
’ REFERENCES
(1) Mauger, A. B. J. Nat. Prod. 1996, 59, 1205–1211.
(2) Liu, W. C.; Fisher, S. M.; Wells, J. S., Jr. J. Antibiot. 1981, 34,
791–799.
(3) Waisvisz, J. M. J. Am. Chem. Soc. 1957, 79, 4520–4521.
(4) Springer, J. P.; Cole, R. J.; Dorner, J. W.; Cox, R. H.; Richard, J. L.;
Barnes, C. L.; Van Der Helm, D. J. Am. Chem. Soc. 1984, 106, 2388–2392.
(5) Engstrom, G. W.; DeLance, J. V.; Richard, J. L.; Baetz, A. L.
J. Agric. Food Chem. 1975, 23, 244–253.
(6) Murakami, S.; Takemoto, T.; Shimizu, Z.; Daigo, K. Jpn.
J. Pharm. Chem. 1953, 25, 571–574.
(7) Johnston, G. A. R.; Curtis, D. R.; Davies, J.; McCulloch, R. M.
Nature 1974, 248, 804–805.
(40) Yoshifuji, S.; Tanaka, K.-I.; Kawai, T.; Nitta, Y. Chem. Pharm.
Bull. 1986, 34, 3873–3878.
(41) Gerratana, B.; Stapon, A.; Townsend, C. A. Biochemistry 2003,
42, 7836–7847.
(42) Page, M. G. P. Drug Resist. Updates 2000, 3, 109–125.
(43) Hamilton-Miller, J. J. Antimicrob. Chemother. 1999, 44, 729–734.
(44) Vaccaro, W. D.; Sher, R.; Davis, H. R. Biorg. Med. Chem. Lett.
1998, 8, 319–322.
(45) Earl, J.; Kirkpatrick, P. Nat. Rev. Drug Discovery 2003, 2, 97–98.
(46) Rothstein, J.; Dykes-Hoberg, M.; Pardo, C.; Bristol, L.; Jin, L.;
Kuncl, R.; Kanai, Y.; Hediger, M.; Wang, Y.; Schielke, J. Neuron 1996,
16, 675–686.
(47) Onoue, H.; Narukawa, Y. J. Antibiot. 1989, 42, 1100–1113.
(48) Baldwin, J. E.; Adlington, R. M.; Moss, N.; Robinson, N. G.
J. Chem. Soc., Chem. Commun. 1987, 1664–1667.
(49) Baldwin, J. E.; Bradley, M. Chem. Rev. 1990, 90, 1079–1088.
(50) Robinson, J. A.; Gani, D. Nat. Prod. Rep. 1985, 2, 293–319.
(51) Miller, M. T.; Gerratana, B.; Stapon, A.; Townsend, C. A.;
Rosenzweig, A. C. J. Biol. Chem. 2003, 278, 40996–41002.
(52) Nussbaum, F. v.; Brands, M.; Hinzen, B.; Weigand, S.; H€abich,
D. Angew. Chem., Int. Ed. 2006, 45, 5072–5129.
(8) McGeer, E. G.; Olney, J. W.; McGeer, P. L. Kainic acid as a tool in
neurobiology; Raven Press: New York, 1978.
(9) Omura, S.; Matsuzaki, K.; Fujimoto, T.; Kosuge, K.; Furuya, T.;
Fujita, S.; Nakagawa, A. J. Antibiot. 1991, 44, 117–118.
(10) Omura, S.; Fujimoto, T.;Otoguro, K.;Matsuzaki, K.;Moriguchi,
R.; Tanaka, H.; Sasaki, Y. J. Antibiot. 1991, 44, 113–116.
(11) Calaza, M. I.; Cativiela, C. Eur. J. Org. Chem. 2008, 3427–3448.
(12) Karoyan, P.; Sagan, S.; Lequin, O.; Quancard, J.; Lavielle, S.;
Chassaing, G. Targets in Heterocycl. Syst. Chem. Prop. 2005, 8, 216–273.
(13) List, B. Tetrahedron 2002, 58, 5573–5590.
(14) Frase, H.; Shi, Q.; Testero, S. A.; Mobashery, S.; Vakulenko,
S. B. J. Biol. Chem. 2009, 284, 29509–29513.
(15) D€urckheimer, W.; Blumbach, J.; Lattrell, R.; Scheunemann,
K. H. Angew. Chem., Int. Ed. 1985, 24, 180–202.
(16) Shih, D. H.; Baker, F.; Cama, L.; Christensen, B. G. Heterocycles
1984, 21, 29–40.
(17) Shih, D. H.; Cama, L.; Christensen, B. G. Tetrahedron Lett.
1985, 26, 587–590.
(18) Li, R. F.; Stapon, A.; Blanchfield, J. T.; Townsend, C. A. J. Am.
Chem. Soc. 2000, 122, 9296–9297.
(19) McGowan, S. J.; Sebaihia, M.; O’Leary, S.; Hardie, K. R.;
Williams, P.; Stewart, G. S. A. B.; Bycroft, B. W.; Salmond, G. P. C.
Mol. Microbiol. 1997, 26, 545–556.
(53) Sleeman, M.; Smith, P.; Kellam, B.; Chhabra, S.; Bycroft, B.;
Schofield, C. ChemBioChem 2004, 5, 879–882.
(54) Stapon, A.; Li, R.; Townsend, C. A. J. Am. Chem. Soc. 2003,
125, 8486–8493.
(55) Grogan, G.; Graf, J.; Jones, A.; Parsons, S.; Turner, N. J.; Flitsch,
S. L. Angew. Chem., Int. Ed. 2001, 113, 1111–1114.
(56) Siirola, E.; Grischek, B.; Clay, D.; Frank, A.; Grogan, G.; Kroutil,
W. Biotechnol. Bioeng. 2011, 108, 2815–2822.
(20) Nunez, L. E.; Mendez, C.; Brana, A. F.; Blanco, G.; Salas, J. A.
Chem. Biol. 2003, 10, 301–311.
(21) Hamed, R. B.; Batchelar, E. T.; Clifton, I. J.; Schofield, C. J. Cell.
Mol. Life Sci. 2008, 65, 2507–2527.
(57) DeSimone, R. W.; Currie, K. S.; Mitchell, S. A.; Darrow, J. W.;
Pippin, D. A. Comb. Chem. High Throughput Screening 2004, 7, 473–94.
(22) Batchelar, E. T.; Hamed, R. B.; Ducho, C.; Claridge, T. D. W.;
Edelmann, M. J.; Kessler, B.; Schofield, C. J. Angew. Chem., Int. Ed. 2008,
47, 9322–9325.
(23) Gerratana, B.; Arnett, S. O.; Stapon, A.; Townsend, C. A.
Biochemistry 2004, 43, 15936–15945.
(24) Sleeman, M. C.; Schofield, C. J. J. Biol. Chem. 2004, 279,
6730–6736.
(25) Sleeman, M. C.; Sorensen, J. L.; Batchelar, E. T.; McDonough,
M. A.; Schofield, C. J. J. Biol. Chem. 2005, 280, 34956–34965.
(26) Hamed, R. B.; Batchelar, E. T.; Mecinoviꢀc, J.; Claridge,
T. D. W.; Schofield, C. J. ChemBioChem 2009, 10, 246–250.
(27) Schultz, P.; Lerner, R. Science 1995, 269, 1835–1842.
(28) Reymond, J. J. Mol. Catal. B: Enzym. 1998, 5, 331–337.
(29) Dean, S.; Greenberg, W.; Wong, C. Adv. Synth. Catal. 2007,
349, 1308–1320.
(30) Bolt, A.; Berry, A.; Nelson, A. Arch. Biochem. Biophys. 2008,
474, 318–330.
479
dx.doi.org/10.1021/ja208318d |J. Am. Chem. Soc. 2012, 134, 471–479