Stereocontrolled Synthesis of (2R,3S)-2-Methylisocitrate, a Central Intermediate in the Methylcitrate Cycle

Article abstract of DOI:10.1002/hlca.200390332

2-Methylisocitrate (=3-hydroxybutane-1,2,3-tricarboxylic acid) is an intermediate in the oxidation of propanoate to pyruvate (=2-oxopropanoate) via the methylcitrate cycle in both bacteria and fungi (Scheme 1). Stereocontrolled syntheses of (2R,3S)- and (2S,3R)-2-methylisocitrate (98% e.e.) were achieved starting from (R)- and (S)-lactic acid (=(2R)- and (2S)-2-hydroxypropanoic acid), respectively. The dispiroketal (6S,7S,15R)-15-methyl-1,8,13,16-tetraoxadispiro[5.0.5.4]hexadecan-14-one (2a) derived from (R)-lactic acid was deprotonated with lithium diisopropylamide to give a carbanion that was condensed with diethyl fumarate (Scheme 3). The configuration of the adduct diethyl (2S)-2-[(6S,7S,14R)-14-methyl-15-oxo-1,8,13,16-tetraoxadispiro[5.0.5.4] hexadec-14-yl]butanedioate (3a) was assigned by consideration of possible transition states for the fumarate condensation (cf. Scheme 2), and this was confirmed by a crystal-structure analysis. The adduct was subjected to acid hydrolysis to afford the lactone 4a of (2R,3S)-2-methylisocitrate and hence (2R,3S)-2-methylisocitrate. Similarly, (S)-lactic acid led to (2S,3R)-2-methylisocitrate. Comparison of 2-methylisocitrate produced enzymatically with the synthetic enantiomers established that the biologically active isomer is (2R,3S)-2-methylisocitrate.

Full text of DOI:10.1002/hlca.200390332

Helvetica Chimica Acta Vol. 86 (2003)
3991
Stereocontrolled Synthesis of (2R,3S)-2-Methylisocitrate, a Central
Intermediate in the Methylcitrate Cycle
by Dan J. Darleya), Thorsten Selmera), William Cleggb), Ross W. Harringtonb), Wolfgang Buckela), and
Bernard T. Golding*c)
a) Laboratorium f¸r Mikrobiologie, Fachbereich Biologie, Philipps-Universit‰t, D-35032 Marburg
b) Chemical Crystallography Unit, School of Natural Sciences Chemistry, Bedson Building,
University of Newcastle upon Tyne, Newcastle upon Tyne NE17RU, U.K.
c) School of Natural Sciences Chemistry, Bedson Building, University of Newcastle upon Tyne,
Newcastle upon Tyne NE17RU, U.K. (phone: ‡ 44-191-222 6647; fax: ‡44-191-222 6929;
e-mail: b.t.golding@ncl.ac.uk)
Dedicated to Duilio Arigoni on the occasion of his 75th birthday
2-Methylisocitrate (ˆ 3-hydroxybutane-1,2,3-tricarboxylic acid) is an intermediate in the oxidation of
propanoate to pyruvate (ˆ2-oxopropanoate) via the methylcitrate cycle in both bacteria and fungi (Scheme 1).
Stereocontrolled syntheses of (2R,3S)- and (2S,3R)-2-methylisocitrate (98% e.e.) were achieved starting from
(R)- and (S)-lactic acid (ˆ(2R)- and (2S)-2-hydroxypropanoic acid), respectively. The dispiroketal
(6S,7S,15R)-15-methyl-1,8,13,16-tetraoxadispiro[5.0.5.4]hexadecan-14-one (2a) derived from (R)-lactic acid
was deprotonated with lithium diisopropylamide to give a carbanion that was condensed with diethyl fumarate
(Scheme 3). The configuration of the adduct diethyl (2S)-2-[(6S,7S,14R)-14-methyl-15-oxo-1,8,13,16-tetraoxa-
dispiro[5.0.5.4]hexadec-14-yl]butanedioate (3a) was assigned by consideration of possible transition states for
the fumarate condensation (cf. Scheme 2), and this was confirmed by a crystal-structure analysis. The adduct was
subjected to acid hydrolysis to afford the lactone 4a of (2R,3S)-2-methylisocitrate and hence (2R,3S)-2-
methylisocitrate. Similarly, (S)-lactic acid led to (2S,3R)-2-methylisocitrate. Comparison of 2-methylisocitrate
produced enzymatically with the synthetic enantiomers established that the biologically active isomer is
(2R,3S)-2-methylisocitrate.
1. Introduction. 2-Methylisocitrate (ˆ 3-hydroxybutane-1,2,3-tricarboxylate) is a
key intermediate in the methylcitrate cycle of both bacteria [1 4] and fungi [4 6]. In
this cycle (Scheme 1), propanoate is oxidised to pyruvate (ˆ2-oxopropanoate) at the
expense of oxaloacetate (ˆoxobutanedioate), which is reduced to succinate
(ˆ butanedioate). The oxaloacetate is regenerated by oxidation of succinate, mediated
by enzymes of the Krebs cycle. The reactions of the methylcitrate cycle are closely
related to those of the glyoxylate cycle, in which acetate is oxidised to the a-oxo acid
glyoxylate (ˆoxoacetate). In the methylcitrate cycle, propanoyl-CoA condenses with
oxaloacetate to yield (2S,3S)-2-methylcitrate (ˆ(2S,3S)-2-hydroxybutane-1,2,3-tricar-
boxylate) [7], which is probably dehydrated to (Z)-2-methylaconitate (ˆ(2Z)-but-2-
ene-1,2,3-tricarboxylate) in a syn manner catalysed by methylcitrate dehydratase
(PrpD) [3][4]. Pig-heart aconitase catalyses the hydration of the double bond of (Z)-2-
methylaconitate with the opposite regioselectivity to that of methylcitrate dehydratase
[3][8]. The addition of H and OH in an anti manner would yield either (2R,3S)- or
(2S,3R)-2-methylisocitrate. By optical-rotation and circular-dichroism studies, the
enantiomer formed by aconitase has been tentatively assigned as (2R,3S)-2-methyl-

3992
Helvetica Chimica Acta Vol. 86 (2003)
Scheme 1. The Methylcitrate Cycle. The configuration of 2-methylisocitrate was determined as (2R,3S) in the
present paper.
isocitrate (ˆ(2S,3R)-3-hydroxybutane-1,2,3-tricarboxylate) [9]. Recently, it was shown
that the 2-methylisocitrate lyases from Escherichia coli and Aspergillus nidulans accept
only one of the threo-diastereoisomer pair (2R,3S)- and (2S,3R)-2-methylisocitrate as
substrate, with the erythro-diastereoisomer pair (2R,3R)- and (2S,3S)-2-methyliso-
citrate being enzymatically inactive. Amino acid sequence comparisons have revealed
that 2-methylisocitrate lyase is homologous with isocitrate lyase, which catalyses the
cleavage of (2R,3S)-isocitrate [10][11] to glyoxylate and succinate. It has been
proposed, therefore, that 2-methylisocitrate lyase stereospecifically mediates the
cleavage of (2R,3S)-2-methylisocitrate to pyruvate and succinate. A definite answer to
this stereochemical problem requires, however, an independent synthesis of bio-
logically active −monochiral× (one-handed; enantiomerically pure) [12] 2-methylisoci-
trate.
A number of synthetic methods have been developed for 2-methylisocitrate, which
lead either to a mixture of all possible isomers [13] or to distinct diastereoisomer pairs
((2R,3R)/(2S,3S) or (2R,3S)/(2S,3R)) [14]. Although these methods provided a
substrate that was useful for assaying 2-methylisocitrate lyase [4], the configurationally
nonhomogeneous nature of the materials limited their usefulness. Furthermore, none
of the procedures can be adapted for generating single enantiomers of 2-methyliso-
citrate. −Monochiral× 2-methyliso-citrate has been synthesised in small amounts
enzymatically [4][15], but this did not permit the unambiguous assignment of

Helvetica Chimica Acta Vol. 86 (2003)
3993
configuration. In this paper, we describe highly diastereoselective routes to (2R,3S)-
and (2S,3R)-2-methylisocitrate from (R)- and (S)-lactic acid (ˆ(2R)- and (2S)-2-
hydroxypropanoic acid), respectively. Comparison of 2-methylisocitrate produced
enzymatically [4][15] with these synthetic enantiomers established that the biologically
active isomer is (2R,3S)-2-methylisocitrate.
2. Results and Discussion. Ley and co-workers devised a method for coupling one
enantiomer of lactic acid at the C(a) atom with an electrophile, which in the case of an
aldehyde gave rise to a product that contains two adjacent chiral centres in a
configurationally defined relationship. In this method, the enolate derived by
deprotonation of a dispiro ketal of lactic acid (e.g., 2b from (S)-lactic acid) was
condensed with an electrophile (e.g., an aldehyde) to give, after acidic deprotection, a
chiral product of high enantiomer purity (see A in Scheme 2) [16]. The lactate Me
group occupied a pseudo-equatorial position and dictated the dispiro ketal config-
urations as (6R,7R) from (S)-lactic acid and (6S,7S) from (R)-lactic acid. Upon enolate
formation, the dispiro ketal configurations were preserved, and this determined the
mode of attack by the electrophile. We found that Ley×s dispiro-ketal-protected lactates
[16] can be used to enable efficient syntheses of −monochiral× 2-methylisocitrates.
Scheme 2. Proposed transition states for the addition of prochiral electrophiles (E ˆ CO2Et) to the dispiro ketal
lithium enolate. A: Addition of benzaldehyde to the least sterically hindered orientation with the Ph group
adopting a pseudo-equatorial position (adapted from Ley and co-workers [16]). B: Addition of diethyl
fumarate with both ester groups adopting pseudo-equatorial positions in the transition state. C: Addition of
diethyl fumarate with both ester groups occupying pseudo-axial positions in the transition state resulting in a
1,3-diaxial interaction. D: One of the two possible modes of addition of diethyl maleate.
We envisaged that coupling the enolate of the dispiro ketal 2a from (R)-lactic acid
and 1 with a suitable succinate (ˆ butanedioate) or fumarate (ˆ(2E)-but-2-enedioate
derivative would lead, with control of configuration at both newly formed stereogenic
centres, to a precursor 3a or 3a' of 2-methylisocitrate. In practice, we obtained only
precursor 3a of (2R,3S)-2-methylisocitrate starting from (R)-lactic acid, and solely

3994
Helvetica Chimica Acta Vol. 86 (2003)
precursor 3b of (2S,3R)-2-methylisocitrate from (S)-lactic acid (Scheme 3). Thus, when
the enolate of the spiro ketal 2b from (S)-lactic acid was reacted with diethyl fumarate
(or rac-diethyl bromosuccinate; see below), a single adduct 3b was obtained, which,
after deprotection by treatment with hydrochloric acid, afforded lactone 4b. The
configuration of 4b was assigned as either (2R,3S) or (2S,3R) by comparison of its 1H-
and 13C-NMR spectra with data for the authentic diastereoisomer pairs of 2-
methylisocitrates [4][14]. This conclusion was supported by reversed-phase-HPLC
analysis, which showed that lactone 4b eluted with the same retention time as the
(2R,3S)/(2S,3R) diastereoisomer pair (data not shown). The lactone 4b was also
analysed by chiral HPLC, with a system known to separate the four stereoisomers of 2-
methylisocitrate [4]. A single peak (Fig. 1,c) was observed that corresponded to the
earlier-eluting peak of the synthetic (2R,3S)/(2S,3R)-diastereoisomer pair (cf. Fig. 1,b),
but did not match either peak of the synthetic (2R,3R)/(2S,3S) pair (cf. Fig. 1,a).
Condensation of the dispiro ketal 2a derived from (R)-lactic acid with diethyl
fumarate (or rac-diethyl bromosuccinate; see below) gave a single adduct 3a that by
acid-catalysed deprotection gave lactone 4a, analysis of which by chiral HPLC showed
a single peak corresponding to the later-eluting peak of the (2R,3S)/(2S,3R) pair
(Fig. 1,d). These results are consistent with those reported for reactions of the dispiro
ketal enolate with aldehydes, whereby predominantly one diastereoisomeric product
arises [16][17]. For these types of reaction, Ley and co-workers proposed a 6-
‡
membered, chair-like transition state, in which the Li ion coordinated to both the
carbonyl and enolate O-atom. The orientation of the aldehyde with respect to Re or Si-
Scheme 3. Synthesis of the Lactones 4a and 4b of (2R,3S)- and (2S,3R)-2-Methylisocitrate, Respectively
a) 1m HCl, Et2O, r.t., 25 h (cf. [15]). b) LDA, rac-diethyl bromobutanedioate, THF, À788, 1 h; 76%. c) 6m HCl,
reflux 3 h, r.t., 12 h, 91%

Helvetica Chimica Acta Vol. 86 (2003)
3995
Fig. 1. Elution profiles of 2-methylisocitrate samples: a) (2R,3R)/(2S,3S), b) (2R,3S)/(2S,3R), c) (2S,3R), and
d) (2R,3S) from a Chirex-d-penicillamine column. The sharp, early eluting peaks are injection artefacts
stemming from acid-induced leaching of Cu2‡ ions from the column.
addition to the enolate is dictated by the steric influence of the group attached to the
aldehyde, with this group adopting a pseudo-equatorial orientation in the transition
state. This model explains why steric control operates at both stereogenic centres.
Indeed, of the four possible stereoisomers that could arise from the addition of the
dispiro ketal 2b enolate to benzaldehyde, only the (S,S)-enantiomer was observed. For
the addition of diethyl fumarate to a dispiro ketal enolate, an analogous chair-like 6-
membered-ring transition state can be formulated. Assuming that the C2h-symmetric
fumarate attacks the enolate from the less-hindered face, thus avoiding the 1,3-diaxial
interaction with the CÀO of the pyran ring, then there are two possible transition
states, as shown by B and C in Scheme 2. In B, the fumarate approaches the enolate in
an orientation that results in both ester groups adopting pseudo-equatorial positions as
the chair-like transition state develops. Models indicate very little steric crowding of the
ester groups in this transition state. By contrast, when the fumarate approaches the
enolate in the alternative orientation C, both ester groups adopt pseudo-axial positions
in the transition state, resulting in an unfavourable 1,3-diaxial interaction between the
distal fumarate ester group and the Me group of the dispiro ketal. Such an interaction
would raise the energy of this transition state C relative to the alternative form B (cf.
Scheme 2). The transition state B would lead to either (2R,3S)- or (2S,3R)-2-

3996
Helvetica Chimica Acta Vol. 86 (2003)
methylisocitrate depending on the chirality of the precursor lactate, whereas the
alternative C would yield either the (2R,3R)- or (2S,3S)-enantiomer. This analysis leads
to the conclusion that the dispiro ketal 2a of (R)-lactic acid gave the lactone 4a of
(2R,3S)-2-methylisocitrate, whereas the dispiro ketal 2b of (S)-lactic acid gave the
lactone 4b of (2S,3R)-2-methylisocitrate. Gratifyingly, these conclusions were con-
firmed by crystal structure analyses of dispiro ketals 3a and 3b.
In the absence of significant anomalous scattering effects (no elements heavier than
O), the crystallographic data do not permit the determination of the absolute
configuration, but the relative configuration at all the chiral centres was established and
found to be the same for 3a and 3b, confirming that they are enantiomers. Indeed, all
the crystallographic results for both compounds are essentially identical, the two
compounds crystallising with opposite chirality in the same space group with the same
cell parameters. However, as the configuration of the lactate defines the dispiro ketal
configurations (e.g., as in 2a) and these are not affected in the condensation with diethyl
fumarate, the configurations of the newly formed stereogenic centres in 3a are defined
by reference to those at dispiroketal centres. The molecular structure of 3a, with the
absolute configuration derived from that of the starting material, is shown in Fig. 2. All
conformational parameters are as expected. It was also confirmed by X-ray analysis
that the relative configuration of the starting material was retained in the conversion of
dispiro ketal 3 into lactone 4 (results not shown).
Fig. 2. Molecular structure of 3a determined by X-ray crystallography. Arbitrary numbering.
When diethyl fumarate was substituted by diethyl maleate (ˆdiethyl (2Z)-but-2-
enedioate), no product was observed from the reaction with the lithium enolate of the
dispiro ketal 2b. According to the model proposed above, both possible diethyl maleate
orientations with a chair transition state would result in one of the ester groups
adopting a pseudo-axial orientation (see, e.g., D in Scheme 2), thus raising the
transition-state energy.
An alternative approach to a −monochiral× 2-methylisocitrate was attempted
through reaction of diethyl bromosuccinate with the lithium enolate of dispiro ketal 2b.
If this reaction proceeded through inversion at the chiral centre of the bromosuccinate,

Helvetica Chimica Acta Vol. 86 (2003)
3997
condensing (for example) diethyl (R)-bromosuccinate with dispiro ketal 2b from (S)-
lactic acid would lead to the lactone of (2S,3S)-2-methylisocitrate. However, regardless
of which diethyl bromosuccinate was employed ((R), (S), or rac), lactone 4b of
(2S,3R)-2-methylisocitrate was the exclusive product in every case. These results
suggested that, under the basic conditions of the reaction, diethyl fumarate was being
generated in situ by dehydrobromination of the diethyl bromosuccinate. TLC and
1H-NMR Analysis of the crude reaction mixture showed that diethyl fumarate was
formed during the reaction (data not shown). However, the isolated yield (76%) of the
dispiro ketal 3b was significantly higher than that obtained from the direct addition of
fumarate (44%). Possibly, in situ formation of diethyl fumarate limits the extent of its
polymerisation during the reaction.
Hydrolysis of the lactone 4a as described for the diastereomeric lactones ((2R,3S)/
(2S,3R) and (2R,3R)/(2S,3S)) [4] gave (2R,3S)-2-methylisocitrate. This isomer, but not
(2S,3R)-2-methylisocitrate derived from hydrolysis of 4b, was a substrate for both 2-
methylisocitrate lyase and 2-methylisocitrate dehydratase. By enzymatic determination
with 2-methylisocitrate lyase and lactate dehydrogenase, the synthesised (2S,3R)-2-
methylisocitrate was found to contain 1.0% of (2R,3S)-enantiomer (i.e., 98% e.e.).
3. Conclusions. We have shown that lactone 4a of (2R,3S)-2-methylisocitrate can
be obtained with very high diastereoselectivity by addition of diethyl fumarate to the
enolate of dispiro ketal 2a of (R)-lactic acid. Similarly, its enantiomeric lactone 4b of
(2S,3R)-2-methylcitrate can be derived from (S)-lactic acid. This selectivity can be
explained by a six-membered, chair-like transition state. The failure of diethyl maleate
to react in an analogous way precludes the possible extension of the method to the
synthesis of the lactones of either (2R,3R)- or (2S,3S)-2-methylisocitrate. The synthetic
(2R,3S)-2-methylisocitrate obtained by hydrolysis of lactone 4a was a substrate for 2-
methylisocitrate lyase and was chromatographically identical to 2-methylisocitrate
enzymatically produced from succinate and pyruvate via the methylcitrate cycle [4].
Hence, it is (2R,3S)-2-methylisocitrate that participates in the methylcitrate cycle,
which can now be completed with respect to stereochemistry as shown in Scheme 1.
The authors thank the Deutsche Forschungsgemeinschaft, EPSRC (UK), European Commission, and
Fonds der Chemischen Industrie for funding this research. Dr. Matthias Brock (Universit‰t Hannover) is
thanked for performing the enzymatic assays.
Experimental Part
1. General. Tetrahydrofuran (THF) was distilled from sodium benzophenone ketyl (ˆdiphenylmethanone
radical ion(1À)). TLC: aluminium sheets pre-coated with silica gel 60 F254 (0.2 mm). Column chromatography
(CC): silica gel 60 (−flash×). Reversed-phase and chiral HPLC: as described in [4]. M.p.: Gallenkamp apparatus;
1
uncorrected. Optical rotations: Perkin-Elmer 241 polarimeter. H- and 13C-NMR Spectra: chemical shifts d in
ppm with residual protons of the deuterated solvents as internal standard, coupling constants J in Hz.
2. Enzyme Assays and Enantiomer Purity. The enzyme assays were performed spectrophotometrically at
340 nm (e ˆ 6.3 mmÀ1 cmÀ1) with 2-methylisocitrate lyase and lactate dehydrogenase as described [4]. To
determine the amount of (2R,3S)-2-methylisocitrate in the synthesised sample of the (2S,3R)-enantiomer, the
lactone was hydrolysed with NaOH and neutralised. The cuvette (d ˆ 1 cm) contained, in a total volume of
1.0 ml, 50 mm potassium phosphate pH 7.0, 2 mm MgCl2, 2 mm dithiothreitol, 0.2 mm NADH, 1.5 U of lactate
dehydrogenase, and 1.0 mm of the sample. The measured DE340 ˆ 0.06 indicated that the (2S,3R)-2-
methylisocitrate sample contained 1.0% of the (2R,3S)-enantiomer. Unfortunately this method does not allow

3998
Helvetica Chimica Acta Vol. 86 (2003)
the determination of the enantiomer purity of the synthesised (2R,3S)-2-methylisocitrate with the same
precision. In a kinetic assay under saturating substrate conditions with 1 mm (2R,3S)-2-methylisocitrate, the
specific activity of 2-methylisocitrate lyase was 28 U/mg, 32% higher than that determined with the equal
mixture of (2R,3S)- and (2S,3R)-2-methylisocitrate [4].
3. Diethyl Bromosuccinates (ˆ Diethyl Bromobutanedioates). A soln. of rac-bromosuccinic acid (2.0 g,
10.2 mmol) in 3% (v/v) HBr in EtOH was stirred at r.t. overnight. The solvent and HBr were removed, and the
residual oil was taken up in Et2O (50 ml). The Et2O soln. was washed with 5% NaHCO3 soln. (2 Â 25 ml), dried
(MgSO4), and evaporated to give rac-diethyl bromosuccinate (2.1 g, 81%) as a pale yellow oil that was used
directly for condensations.
(R)- and (S)-Bromosuccinic acid were prepared as described [18] and were converted to (R)- and (S)-
diethyl bromosuccinate, respectively, as described above for rac-bromosuccinic acid.
4. Diethyl (2S)-2-[(6S,7S,14R)- and (2R)-2-(6R,7R,14S,14-Methyl-15-oxo-1,8,13,16-tetraoxadispir-
i
o[5.0.5.4]hexadec-14-yl]butanedioate (3a and 3b, resp.). To Pr2NH (0.4 g, 0.6 ml, 4.0 mmol) in THF (10 ml)
at À788, 2.2m BuLi in THF (1.8 ml, 4.5 mmol) was added dropwise, and the soln. was allowed to warm to r.t. over
20 min. The resulting mixture was cooled to À788 and the (6S,7S,15R)-15-methyl-1,8,13,16-tetraoxadispi-
ro[5.0.5.4]hexadecan-14-one (2a; from (R)-lactic acid) or (6R,7R,15S)-15-methyl-1,8,13,16-tetraoxadispi-
ro[5.0.5.4]hexadecan-14-one (2b; from (S)-lactic acid) (1.0 g, 3.9 mmol) in THF (10 ml) at À788 was added
via cannula. The mixture was stirred for 1 h at À788. Diethyl bromosuccinate (0.75 ml, 1.0 g, 4.0 mmol) or
diethyl fumarate (0.65 ml, 0.68 g, 4.0 mmol) was added, and the soln. was stirred for 1 h at À788. The mixture
was quenched by the addition of H2O (10 ml). The resulting mixture was extracted with Et2O (3 Â 25 ml), the
combined org. layer dried (MgSO4) and evaporated, and the crude yellow oil purified by medium-pressure CC
(silica gel, petroleum ether/Et2O 3 :2): 3a or 3b (1.5 g, 88%) as a colourless oil that was crystallised from
petroleum ether (40 608): colourless crystals (1.3 g, 76%). 1H-NMR (CDCl3, 300 MHz): 4.11 (q, J ˆ 7.2, 4 H);
4.00 3.65 (m, 4 H); 3.43 (t, J ˆ 7.6, 1 H); 2.80 (d, J ˆ 7.6, 2 H); 2.04 1.50 (m, incl. s at 1.63 (3 H), 15 H); 1.25 (t,
J ˆ 7.1, 3 H); 1.23 (t, J ˆ 7.2, 3 H). 13C-NMR (CDCl3 300 MHz): 172.2; 171.4; 170.6; 104.5; 96.3; 76.3; 62.7; 62.3;
61.01; 60.94; 49.6; 32.5; 29.4; 29.0; 25.2; 24.6; 18.5; 17.8; 14.49; 14.48. EI-MS: 228, 199, 168 (100, 154, 111, 98.
Data of 3a: M.p. 74 768. [a]2D5 ˆ ‡77.4 (c ˆ 1.0, CHCl3). Anal. calc. for C21H32O9: C 58.87, H 7.53; found:
C 59.00, H 7.78.
Data of 3b: M.p. 74 768. [a]2D5 ˆ À77.8 (c ˆ 1.0, CHCl3). Anal. calc. for C21H32O9: C 58.87, H 7.53; found:
C 58.79, H 7.80.
5. Lactones of (2R,3S)- and (2S,3R)-2-Methylisocitrate (ˆ2R,3S)- and (2S,3R)-Tetrahydro-2-methyl-5-
oxofuran-2,3-dicarboxylate; 4a and 4b, resp.). A soln. of 3a or 3b (100 mg, 0.25 mmol) in 6m HCl (20 ml) was
refluxed for 3 h, cooled, and allowed to stand at r.t. overnight. The H2O and HCl were evaporated. The residual
oil was taken up in H2O and evaporated 3 times to remove traces of HCl. The obtained yellow oil was
crystallised from acetone/benzene: 4a or 4b (41 mg, 91%). Colourless crystals. 1H-NMR ((D6) acetone,
300 MHz): 3.63 (dd, 3J ˆ 10.6, 9.7, 1 H); 3.06 (dd, 2J ˆ 17.6, 3J ˆ 10.6, 1 H); 2.82 (dd, 2J ˆ 17.6, 3J ˆ 9.7, 1 H); 1.79
‡
(s, 3 H). 13C-NMR ((D6) acetone, 300 MHz): 174.9; 171.6; 171.3; 85.4; 51.1; 32.8. MS-EI: 189 ([M ‡ 1] ), 143,
115, 98, 55, 43 (100).
Data of 4a: M.p. 142 1448. [a]2D5 ˆ À25.0 (c ˆ 1.0, acetone).
Data of 4b: M.p. 140 1428. [a]2D5 ˆ ‡25.3 (c ˆ 1.0, acetone). Anal. calc. for C7H8O6: C 44.69, H 4.29; found:
C 45.09, H 4.68.
6. X-Ray Crystallography. Crystals of 3a and 3b were examined on a Bruker SMART-CCD diffractometer
at 150 K, with Mo-Ka radiation (l 0.71073 ä). Results for 3a are presented here. Crystal data: C21H32O9, M
428.5, monoclinic, space group C2, a ˆ 25.7317(14), b ˆ 8.3941(4), c ˆ 10.3023(5) ä, b ˆ 93.712(2)8, V ˆ
2220.57(19) ä3, Z ˆ 4; R ˆ 0.037 for refinement of 275 parameters from 2810 unique data. In the absence of
significant anomalous scattering effects, Friedel pairs were merged. The crystallographic results for 3b are
essentially identical within experimental error, since the absolute configuration is undetermined1).
1
)
CCDC-220257 and -220256 contain the supplementary crystallographic data for this paper. These data can
be obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html (or from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ; fax ‡441223336033; e-mail
deposit@ccdc.cam.ac.uk).

Helvetica Chimica Acta Vol. 86 (2003)
3999
REFERENCES
[1] S. Textor, V. F. Wendisch, A. A. De Graaf, U. M¸ller, M. I. Linder, D. Linder, W. Buckel, Arch. Microbiol.
1997, 168, 428.
[2] A. R. Horswill, J. C. Escalante-Semerena, J. Bacteriol. 1999, 181, 5615.
[3] A. R. Horswill, J. C. Escalante-Semerena, Biochemistry 2001, 40, 4703.
[4] M. Brock, D. Darley, S. Textor, W. Buckel, Eur. J. Biochem. 2001, 268, 3577.
[5] H. Uchiyama, M. Ando, Y. Toyonaka, T. Tabuchi, Eur. J. Biochem. 1982, 125, 523.
[6] M. Brock, R. Fischer, D. Linder, W. Buckel, Mol. Microbiol. 2000, 35, 961.
[7] S. Brand‰nge, O. Dahlmann, L. Mˆrch, J. Chem. Soc., Chem. Commun. 1980, 555.
[8] M. Brock, C. Maerker, A. Sch¸tz, U. Vˆlker, W. Buckel, Eur. J. Biochem. 2002, 269, 6184.
[9] R. L. Beach, T. Aogaichi, G. W. Plaut, J. Biol. Chem. 1977, 252, 2702.
[10] H. Katsura, J. Chem. Soc. Japan 1961, 82, 92.
[11] A. L. Patterson, C. K. Johnson, D. van der Helm, J. A. Minkin, J. Am. Chem. Soc. 1962, 84, 309.
[12] J. W. Cornforth, Aust. J. Chem. 1993, 46, 157.
[13] C. Rach, Ann. Chim. (Paris) 1886, 234, 35.
[14] G. W. Plaut, R. L. Beach, T. Aogaichi, J. Biol. Chem. 1975, 250, 6351.
[15] O. Gawron, K. P. Mahajan, Biochemistry 1966, 5, 2335.
[16] G. J. Boons, R. Downham, K. S. Kim, S. V. Ley, M. Woods, Tetrahedron 1994, 50, 7157.
[17] D. P. Zhang, C. Bleasdale, B. T. Golding, W. P. Watson, Chem. Commun. 2000, 1141.
[18] J. A. Frick, J. B. Klassen, A. Bathe, J. M. Abramson, H. Rapoport, Synthesis 1992, 7, 621.
Received September 11, 2003