(S,S)-2,3-dihydroxy-2,3-dihydrobenzoic acid: Microbial access with engineered cells of Escherichia coli and application as starting material in natural-product synthesis

Article abstract of DOI:10.1002/chem.200204265

Cyclohexadiene-trans-5,6-diols such as (S,S)-2,3-dihydroxy-2,3-dihydrobenzoic acid (2,3-trans-CHD) have been shown to be of importance as chiral starting materials for the syntheses of bioactive substances, especially for the syntheses of carbasugars. By

Full text of DOI:10.1002/chem.200204265

FULL PAPER
(S,S)-2,3-Dihydroxy-2,3-dihydrobenzoic Acid: Microbial Access with
Engineered Cells of Escherichia coli and Application as Starting Material
in Natural-Product Synthesis
Dirk Franke,[a, e] Volker Lorbach,[a] Simon Esser,[a] Christian Dose,[a] Georg A. Sprenger,[b, f]
Markus Halfar,[c] Jˆrg Thˆmmes,[c, g] Rolf M¸ller,[d] Ralf Takors,[a] and Michael M¸ller*[a]
Abstract: Cyclohexadiene-trans-5,6-di-
ols such as (S,S)-2,3-dihydroxy-2,3-dihy-
drobenzoic acid (2,3-trans-CHD) have
been shown to be of importance as chiral
starting materials for the syntheses of
bioactive substances, especially for the
syntheses of carbasugars. By using meth-
ods of metabolic-pathway engineering,
the Escherichia coli genes entB and
entC, which encode isochorismatase
and isochorismate synthase, were cloned
and over-expressed in E. coli strains with
a deficiency of entA, which encodes 2,3-
dihydroxybenzoate synthase. A 30-fold
increase in the corresponding EntB/
EntC enzyme activities affects the accu-
mulation of 2,3-trans-CHD in the culti-
vation medium. Although the strains did
not contain deletions in chorismate-
utilising pathways towards aromatic
amino acids, neither chorismate nor
any other metabolic intermediates were
found as by-products. Fermentation of
these strains in a 30 L pH-controlled
stirred tank reactor showed that 2,3-
trans-CHD could be obtained in con-
centrations of up to 4.6 gLÀ1. This dem-
onstrates that post-chorismate metabo-
lites are accessible on a preparative scale
by using techniques of metabolic-path-
way engineering. Isolation and separa-
tion fromfermentation salts could be
performed economically in one step
through anion-exchange chromatogra-
phy or, alternatively, by reactive extrac-
tion. Starting from2,3- trans-CHD as an
example, we established short syntheses
towards new carbasugar derivatives.
Keywords: arene diols ¥ bioorganic
chemistry ¥ carbohydrate mimetics
¥ chorismate ¥ enzyme catalysis ¥
shikimate
the generation of a large combinatorial compound library.[4]
Introduction
c]
The neuraminidase inhibitor GS-4104[1a and its deriva-
Intermediates in the common biosynthetic pathway of aro-
matic amino acids (AAA pathway) like shikimate[1] and
tives[1d, 5] have recently been synthesised starting fromquinic
acid and shikimic acid.
Much effort has been concentrated on the molecular
characterisation of enzymes and their mechanisms that are
involved in the AAA pathway. This has finally led to
substantial progress in microbially producing many new chiral
pool substances on an industrial scale, including, for example,
c, 2]
quinate[1a
have emerged as valuable precursors in the
syntheses of natural products and pharmacologically active
substances. Starting fromquinic acid, a synthesis towards, say,
the immunosuppressant FK506[3] has been established. Addi-
tionally, shikimic acid has been used as a starting material for
[a] PD Dr. M. M¸ller, Dr. D. Franke, Dipl.-Chem. V. Lorbach,
Dipl.-Chem. S. Esser, Dipl.-Ing.(FH) C. Dose, Dr.-Ing. R. Takors
Institut f¸r Biotechnologie 2, ForschungszentrumJ¸lich GmbH
52425 J¸lich (Germany)
[e] Dr. D. Franke
Present address: Department of Chemistry and
The Skaggs Institute for Chemical Biology
The Scripps Research Institute
Fax : (‡49)2461-61-3870
10550 North Torrey Pines Road, La Jolla, CA 92037 (USA)
E-mail: mi.mueller@fz-juelich.de
[f] Prof. Dr. G. A. Sprenger
[b] Prof. Dr. G. A. Sprenger
Present address: Institut f¸r Mikrobiologie
Allmandring 31, 70569 Stuttgart (Germany)
Institut f¸r Biotechnologie 1, ForschungszentrumJ¸lich GmbH
52425 J¸lich (Germany)
[g] PD Dr. J. Thˆmmes
[c] M. Halfar, PD Dr. J. Thˆmmes
Institut f¸r Enzymtechnologie (IET)
Heinrich-Heine Universit‰t D¸sseldorf
52426 J¸lich (Germany)
Present address: IDEC Pharmaceutics Corp
11011 Torreyana Road, San Diego, CA 92121 (USA)
Supporting information for this article is available on the WWW under
http://www.chemeurj.org or from the author.
[d] PD Dr. R. M¸ller
Gesellschaft f¸r Biotechnologische Forschung mbH
Mascheroder Weg 1, 38124 Braunschweig (Germany)
4188
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200204265
Chem. Eur. J. 2003, 9, 4188 4196

4188 4196
dehydroshikimate,[6] shikimate,[7] dehydroquinate[8] and quin-
ate.[7b, 9]
in principle, be accessible by techniques of metabolic-pathway
engineering.
Nevertheless, metabolites derived from chorismate, the
metabolic branch point in the biosynthesis of ubiquinones,
menaquinones, folates and aromatic amino acids, have not
been examined in the same way. This may be due to the fact
that chorismate itself has not yet been microbially produced in
high concentrations.[10]
Recently, increasing attention has been paid to substituted
cyclohexadiene-trans-diols (trans-CHDs). Bearing in mind
that they are small and highly functionalised chiral molecules,
As has been shown by Leistner and co-workers,[14] strains of
Klebsiella pneumoniae with deficiencies in the AAA pathway
excrete two different trans-CHDs, the (S,S)-2,3-dihydroxy-
2,3-dihydrobenzoic acid (2,3-trans-CHD, 1) and the (R,R)-3,4-
dihydroxy-3,4-dihydrobenzoic acid (3,4-trans-CHD, 2) when
enzymes catalysing the reaction of chorismate (3) towards
these metabolites are overproduced. However, the applic-
ability of these microbial systems for preparative purposes has
been limited by the low final product concentrations (up to
200 mgLÀ1) and by the occurrence of chorismate as an
ubiquitous by-product, necessitating further purification
processes.[14]
In this article we describe the construction and character-
isation of recombinant Escherichia coli strains for the efficient
microbial production of 2,3-trans-CHD 1, which has been
published in a preliminary form.[15] The microbial production
could easily be scaled up, and 1 was obtained on a higher
decagramscale. Efficient isolation fromthe fermentation
broth was achieved by ion-exchange chromatography or,
alternatively, by reactive extraction. The potential of enan-
tiopure 1 as a valuable synthetic building block in the
chemistry of carbohydrate mimetics is demonstrated through
the straightforward synthesis of a carbohydrate analogue.
it is obvious that they can be regarded as valuable building
blocks for a wide variety of target molecules. Substances such
as plant metabolites of the cyclohexane epoxide class and
valienamine have already been synthesised starting from
trans-CHDs.[11]
trans-CHDs are also found in bacteria, plants and fungi, as
well as in yeast, as intermediates of the post-chorismate
pathway towards enterobactin and menaquinones[12] or are
biosynthesised fromshikimate. [13] Therefore, they should also,
Results and Discussion
Our strategic approach to strains for the production of 2,3-
trans-CHD 1 is in analogy to work with K. pneumoniae strains
(Scheme 1).[14]
Abstract in German: Cyclohexadien-trans-diole wie beispiels-
weise (5S,6S)-Dihydroxycyclohexa-1,3-diencarbons‰ure (2,3-
trans-CHD) haben sich als wichtige chirale Ausgangsverbin-
dungen zur Synthese pharmakologisch aktiver Substanzen,
insbesondere zur Synthese von Carbazuckern und Aminocar-
bazuckern erwiesen. Unter Verwendung von Techniken der
Stoffflussderegulation wurden die Escherichia coli-Gene entB
und entC, kodierend f¸r Isochorismatase und Isochorismat-
synthase, in E. coli-St‰mmen ¸berexpremiert, welche eine
Defizienz von entA, kodierend f¸r 2,3-Dihydroxybenzoatsyn-
thase, hatten. Eine 30-fache Steigerung der entsprechenden
Enzymaktivit‰ten EntB/EntC bewirkte die Bildung von 2,3-
trans-CHD im Kultivierungsmedium. Obwohl die St‰mme
keine Mutationen in den chorismatverwendenden Biosynthe-
sewegen zu den aromatischen Aminos‰uren besa˚en, wurde
weder Chorismat noch ein anderes metabolisches Intermediat
als Nebenprodukt gefunden. Fermentationen mit diesen St‰m-
men in einem pH-geregelten 30 L R¸hrkesselreaktor zeigten,
dass 2,3-trans-CHD in Konzentrationen bis zu 4.6 gLÀ1 er-
halten werden kann. Dies demonstriert zum ersten Mal, dass
Metabolite des Enterobactin-Biosyntheseweges durch Techni-
ken der Stoffflussderegulation in pr‰parativem Ma˚stab er-
halten werden kˆnnen. Aufreinigung und Separation von den
Fermentationssalzen konnte auf effiziente Weise in einem
Schritt unter Verwendung von Anionentauscherharzen oder
alternativ durch Reaktivextraktion durchgef¸hrt werden. Aus-
gehend von 2,3-trans-CHD haben wir exemplarisch eine kurze
Synthese zu ent-Streptol (ent-Valienol) durchgef¸hrt.
Nonpathogenic E. coli was selected as a host strain for
genetic modification because of the availability of potent
mutants and well-established fermentation conditions.
An increase in the metabolic production of 1 was achieved
by over-expression of entC, which encodes isochorismate
synthase, and entB, which encodes isochorismatase. In order
to prevent product loss due to metabolic processing of 1, an
entA-negative mutant was used as the production host.
Isolation and cloning of entB and entC: entB and entC were
isolated from E. coli W3110[16] (wild-type). Both fragments
were inserted into plasmid pJF119EH1,[17] either separately or
in tandem, to give plasmids pDF1, pDF2 and pDF3.
DNA sequencing of the cloned fragments proved to be in
conformity with literature data,[18] except for a point mutation
in entB at bp 525 (gcg ! gct), which should not result in any
change of the protein sequence (quiet mutation).
Cells of E. coli cloning strain DH5a,[19] E. coli wild-type
strain W3110,[16] E. coli entAÀ mutants AN193[20, 21] and
H1882[22] and E. coli entCÀ/menFÀ mutant PBB8[23] were
transformed with the plasmids pDF1, pDF2 or pDF3. Addi-
tionally, transformations were performed with the ™empty∫
plasmid pJF119EH1. Transformants were grown on LB
medium in the presence of IPTG (100 mm). SDS-PAGE
analysis of the crude cell extracts showed expression of the
inserted gene fragments in all cases.
Chem. Eur. J. 2003, 9, 4188 4196
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4189

M. M¸ller et al.
FULL PAPER
In extracts of cells containing
the empty plasmid pJF119EH1 or
cells without any plasmids, only a
very slow reaction of chorismate
(3) to isochorismate (4) and 2,3-
trans-CHD 1 was observed (Fig-
ure 1A). Nevertheless, an almost
linear conversion of 3 to other
undetectable products was ob-
served (t1/2 ˆ 15 h). Eliminating
the thermal decomposition of 3
(t1/2 ˆ 28 h; applying the same
conditions) in enzyme balancing,
a total enzymatic chorismate-deg-
radation activity of 21 mU per mg
of protein was calculated.
Crude cell extract of strains
with plasmid pDF1, harbouring
entB, showed a linear increase of
the concentration of 3,4-trans-
CHD 2 as a major product and
2,3-trans-CHD 1 as a minor prod-
Scheme 1. Biosynthesis of trans-CHD starting fromchorismate ( 3). The reaction towards 2,3-trans-CHD 1 via
isochorismate (4) is catalysed by isochorismate synthase EntC and isochorismatase EntB. trans-CHD 1 is
metabolised to 2,3-dihydroxybenzoate by 2,3-dihydroxybenzoate synthase EntA. 3,4-trans-CHD 2 is a non-
natural metabolite and is formed out of chorismate under a large excess of EntB.
Enzyme-activity test: The enzyme activities of EntC and EntB
were determined in cell-free extracts of IPTG-induced cells
by monitoring the conversion of chorismate (3) to either 3,4-
trans-CHD 2 (EntB) or to isochorismate (4) and 2,3-trans-
CHD 1 (EntC ‡ EntB). This was performed in a buffered
aqueous solution of the cell-free extracts for a period of
400 minutes. Time-course data of the analysis of the entAÀ
strain AN193 are shown in Figure 1.
uct (Figure 1B). Isochorismate (4) could not be detected at
any time. The formation of 3,4-trans-CHD 2 verifies that EntB
is overproduced in a catalytically active form. The production
of small amounts of 2,3-trans-CHD 1 indicates the weak
activity of chromosomally encoded EntC in E. coli strains
AN193, H1882, DH5a and W3110.
Cell-free extracts of strains containing plasmid pDF2,
harbouring entC, catalysed the reaction of chorismate (3) to
isochorismate (4, major product)
and 2,3-trans-CHD
1
(minor
product) (Figure 1C). The slow
and nearly linear increase of the
concentration of 1 can be attrib-
uted to the wild-type activity of
EntB.
Analysis of crude cell extract of
strains in which entB and entC
were over-expressed in parallel
(pDF3) verified the conversion of
chorismate (3) to 2,3-trans-CHD
1 via intermediate isochorismate
(4) (Figure 1D). The linearity of
increase of the concentration of 1
and the formation of 4 as an
intermediate results from the cat-
alysis by EntB as the rate-limiting
step.
Analogous experiments with
cell-free extract of strains
H1882, W3110 or DH5a showed,
within the limits of measurement
accuracy, approximately the same
enzyme activities for natural and
over-expressed EntB and EntC.
No product loss of 2,3-trans-CHD
1 or 3,4-trans-CHD 2 due to
metabolic processing in strains
Figure 1. Determination of enzyme activities in the crude cell extract of transformants of E. coli strain AN193
at 208C (chorismate 8 mm). The change of concentration of reactants and products was monitored by HPLC.
*
A) Transformants with plasmid pJF119EH1 catalyse the degradation of chorismate (3, ) slowly. The products
~
&
are 2,3-trans-CHD 1 ( ) and isochorismate (4, ) in equal parts. As a reference, time-course studies of the
nonenzymatic decomposition of chorismate in buffered solution showed slower degradation under the same
conditions (––). B) Transformants with plasmid pDF1 (EntB is over-expressed) mainly catalyse the reaction of
&
3 to 3,4-trans-CHD 2 ( ) and small amounts of 1. C) Transformants with pDF2 (EntC is over-expressed) catalyse
the reaction of 3 to 4 and small amounts of 1. D) Transformants with plasmid pDF3 containing entB and entC
completely convert 3 to 1 with 4 as a temporarily formed intermediate.
4190
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4188 4196

2,3-Dihydroxy-2,3-dihydrobenzoic acid
4188 4196
W3110 or DH5a was observed. Cell-free extract of E. coli
strain PBB8 (entCÀ/menFÀ-mutation) is characterised by the
absence of natural isochorismate synthase activity (below
quantification limit). Catalytic activities for plasmid-encoded
EntB and/or EntC were much lower in this strain; however,
this observation was examined not in detail.
Fromthe decrease in chorismate concentration ( 3) and
increase of secondary-metabolite concentrations, absolute
activities of EntB and EntC were quantified relative to total
protein concentrations (Table 1).
Figure 2. Production rates and final concentrations of 1 for cultivation of
Table 1. Enzyme activities of cell extracts in Tris-HCl buffer (100 mm, pH 8.5)
and at 208C depending on the genes over-expressed in strains AN193 and PBB8.
strain AN193/pDF3 in mineral salts medium.
Strains
Plasmid-
borne
genes
Enzyme activity
concentrations were below the detection limit (7 mm).
Even in extracts of highly concentrated cell suspensions
(DCM > 21 gLÀ1) no metabolites 1 4 were detected. Con-
centrations of 2,3-trans-CHD 1 in the cell-free extracts were
estimated to be below 45 mg per g of DCM in any case,
corresponding to an intracellular concentration of less than
86 mmolLÀ1 (cell volume).[24]
Other experiments in shake flasks showed that optimal
production conditions, with regard to production rate, growth
rate and long-termstability, were 37 8C and a pH between 6
and 8. Alternative carbon sources like galactose, fructose,
glycerol, acetate or lactate could also be used for production.
The use of glycerol or galactose allows even better cell-mass-
specific product-formation rates of up to 48 mghÀ1 per g of
DCM.
Various experiments in which the inductor IPTG was added
at different cell growth phases showed that induction only
temporarily reduced growth rate for 4 to 5 hours and only
after the first addition. If IPTG was added directly at the very
beginning of incubation, no reduction of growth rate or final
cell density was observed. Formation of 1 started at the very
beginning of induction, and was found in both growing and
resting cells. Almost maximal induction was achieved with
inductor concentrations as low as 50 mm ꢀIPTG).
In order to obtain 1 on a preparative scale, strain AN193/
pDF3 was cultivated in a 30 L pH-controlled stirred tank
reactor (STR, process volume 20 L) in a mineral-salts medium
with glucose as the carbon source. Production was monitored
over a period of 50 h. Maximal cell-mass-specific product-
formation rates of 58 mghÀ1 per g of DCM were found 3 hours
after induction. A molar yield of 17% was calculated with
respect to glucose. Final product concentrations of more than
4.6 gLÀ1 (1) were achieved.
[mU per mg protein]
EntC
EntB (c)[b]
EntB (isoc)[c]
AN193
PBB8
/pJF119EH1
/pDF1
/pDF2
/pDF3
/pJF119EH1
/pDF1
/pDF2
5
7
< 0.25
77
3
entB
entC
entB, entC
n.d.[a]
224
212
< 0.25
< 0.25
34
n.d.
n.d.
< 0.25
40
< 7
442
n.d.
entB
entC
entB, entC
n.d.
< 0.25
32
< 0.25
> 74
/pDF3
74
[a] n.d. ˆ not determined. [b] chorismate. [c] isochorismate.
In summary, EntC activity in all the tested E. coli strains
was amplified at least 30-fold by plasmidic over-production
relative to strains in which EntC is not overproduced,
amounting to more than 210 mU per mg of protein (strains
DH5a, W3110, AN193 and H1882). The amplification of
EntB activity was even higher (almost 150-fold) amounting to
442 mU per mg of protein. Similar amplification factors were
observed in strains H1882, DH5a and W3110.
Fermentation experiments for the production of 1: Cultiva-
tion of the E. coli strains in shaking flasks proved that 1 was
excreted into the medium by strains containing pDF3 (over-
expressing entB/entC), whereas 1 was not detected in suspen-
sions of any other transformants, plasmid-free strains or
uninduced strains. Interestingly, neither 3,4-trans-CHD 2 nor
isochorismate (4) was detected in suspensions of any trans-
formants or in strains containing pDF1 or pDF2.
Production rates of 1 in mineral-salt media were best with
entAÀ mutants AN193 (29 mghÀ1 per g of dry cell mass
(DCM) and H1882, moderate in entCÀ/menFÀ strain PBB8
and low in wild-type strains W3110 and DH5a (Figure 2).
With regard to production rates, the excretion of 1 was
threefold higher in entAÀ strains than in strains without entA
defect.
The final product concentration of 1 after 14 h was
significantly higher in strains PBB8, AN193 and HI1882 than
in wild-type strains. Maximum concentrations of 550 mgLÀ1
(3.5 mm) were achieved in shaking-flask experiments with
strain AN193.
Several attempts to determine intracellular metabolite
concentrations by analysing crude extracts of induced cells
with RP-HPLC failed, presumably because the metabolite
Methods for isolation and purification of 1: One major
challenge for separating 1 fromfermentation salts, proteins
and other organic compounds is its instability under strongly
acidic or strongly basic conditions or if heated. On the other
hand, 1 was found to be stable in aqueous solution at
moderate pH (3 ꢀ pH ꢀ 11) and roomtemperature in the
absence of oxidising agents.
All attempts to extract 1 fromacidified fermentation
permeate (pH 3) by using alkanes, chloroalkanes, diphenyl
ether and similar solvents were characterised by low extrac-
tion coefficients (e ꢀ 7%). Nevertheless, using more polar
Chem. Eur. J. 2003, 9, 4188 4196
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4191

M. M¸ller et al.
FULL PAPER
extraction solvents like short-chain alcohols significantly
improves the extraction coefficient (e ˆ 30%, butan-1-ol);
however, at the cost of simultaneously extracting other polar
compounds as well.
Additionally, two separation methods were established that
make use of selectively binding the carboxylic acid function-
ality of 1 to a dispersed or solid-phase surface. Both methods
work on the laboratory scale as well as on a higher gramscale.
A reactive extraction process for separation: The extraction
coefficients in organic solvents are substantially improved by
adding ionic carriers that bind the product.[25] Ion-pair
formation of the anion of 1 with the cationic carrier
trioctylmethylammonium chloride (TOMAC) results in un-
charged, less hydrophilic species that can be extracted with
organic solvents with a higher extraction coefficient (e ˆ 23%,
octan-1-ol with 5% carrier). The selective re-extraction of 1
and separation of the remaining nonpolar impurities can be
done by using aqueous phases with a high anion concentra-
tion, such as brine, at neutral pH (e > 90%) (Scheme 2).
After concentration in vacuo, separation fromsalts can
easily be done in a third extraction step with butan-1-ol (e ˆ
30%) to afford 1 with 90% purity after lyophilisation (HPLC
analysis).
Although reactive extraction and re-extraction normally
work most efficiently in a multicycle, continuously operating
separation process that is integrated into the production
process, all the tested cationic carriers were characterised as
being bacteriostatic or even partially bacteriocidal. Therefore,
integration of the extraction method into the production
process (for technical application) affords further improve-
ment in preventing carrier contact with production cells.
Scheme 2. Isolation of 1 by reactive extraction with cationic carriers
(TOMAC) as extraction mediators in organic solvents like octan-1-ol.
structures in the synthesis of iso-crotepoxide and ent-senep-
oxide.[28] Nucleophilic opening results in the introduction of a
new functionality in trans-orientation to the simultaneously
formed alcohol function. Introduction of a cis-diol function-
ality is possible by the OsO4-catalysed dihydroxylation
presented here.
Compound 1 was transformed into its methyl ester 5 with
60% yield by using a solution of hydrogen chloride in
methanol (Scheme 3).
Regioselective dihydroxylation at the C3,C4 positions
followed by OsO4-catalysed oxidation with N-methylmorpho-
line N-oxide (NMO) gave two diastereomeric vicinal alcohols
6 and 7 in a ratio of 5:1, which could not be separated
chromatographically. No oxidation of the electron-poor
double bond at the C1,C2 positions was observed.
On increasing the steric demand of the alcohol function-
alities by introduction of tert-butyldimethylsilyl (TBS) pro-
Ion exchange chromatography was used advantageously to
selectively bind 1 to DOWEX 1 Â 8 (ClÀ) anion-exchange
resin[26] at pH 7 with a dynamic capacity of 15 to 16 gLÀ1 (96 to
102 mm) sedimented resin. In contrast, binding of proteins or
other by-products was not observed. Elution of the product
occurs under moderate acidic conditions at pH 2.8 by proto-
nation of the carboxylic acid functionality; this allows
isolation without any aromati-
sation. By using 2 kg of resin
27 g (0.17 mol) of 1 could be
isolated and purified in a single
operation step with 95% purity
and 75% yield after lyophilisa-
tion.
Application in organic synthe-
sis: a short synthesis of the
enantiomer of the plant-growth
inhibitor streptol:[27] The syn-
thesis of unsaturated carbasu-
gars requires stereoselective
oxidation, for example at the
C3,C4 positions. Previously, we
Scheme 3. Esterification of 1 and subsequent transformations of the methyl ester 5. Reagents and conditions:
a) HCl in MeOH, RT, 48 h; b) NMO, cat. OsO4, tBuOH/H2O (1:1), RT, 72 h; c) TBS-OTf, NEt3, CH2Cl2, RT,
18 h; d) NMO, MeSO3NH2, cat. K2OsO4, acetone/nBuOH/H2O (10:9:1), RT, 16 h; e) DIBAL-H, CH2Cl2, 08C,
3 h; f) tris(dimethylamino)sulfoniumdifluorotrismethylsilicate, CH2Cl2, À788C to RT, 16 h.
have shown regio- and stereo-
selective epoxidation of 1, and
the application of the resultant
4192
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4188 4196

2,3-Dihydroxy-2,3-dihydrobenzoic acid
4188 4196
tecting groups, the OsO4-catalysed dihydroxylation results in
9 as the main product in 56% yield. Only traces of the product
resulting fromoxidation of the C1,C2 double bond were
found.
Experimental Section
General methods: All reagents used were of analytical grade. Solvents
were dried by standard methods if necessary. TLC was carried out on
aluminium sheets precoated with silica gel 60F254 (Merck). Detection was
accomplished by UV light (l ˆ 254 nm). Preparative column chromatog-
raphy was carried out on silica gel 60 (Merck, 40 63 mm). 1H NMR spectra
were recorded on an AMX300 (Bruker Physik AG, Germany) with
CD3OH (d ˆ 4.87 ppm), CHCl3 (d ˆ 7.27 ppm) or HDO (d ˆ 4.81 ppm) as
internal standard, 13C NMR spectra were calibrated with 13CD3OD (d ˆ
49.15 ppm), 13CDCl3 (d ˆ 77.23 ppm) or sodium trimethylsilylpropansulfo-
nate (TSP) as internal standard. GCMS spectra were determined on an
HP6890 series GC systemfitted with an HP5973 mass-selective detector
(Hewlett Packard; column HP-5MS), 30 m  250 mm; TGC (injector) ˆ
2508C, TMS (ion source) ˆ 2008C, time program (oven): T0min ˆ 608C,
T3min ˆ 608C, T14min ˆ 2808C (heating rate 208CminÀ1), T19min ˆ 2808C.
HR-MS (EI) was performed on an A.E.I. MS50 and elemental analysis on a
While cis-dihydroxylation starting fromthe BOC-protected
2-amino analogue of 5 investigated by Steel et al. occurs in a
comparable regio- and stereoselective manner,[29] oxidation of
the corresponding cis-diastereomer of 8 by Sutherland et al.
gave a 10:1 mixture of oxidation product at the C1,C2
positions over oxidation product at the C3,C4 positions;[30]
this means that any implication from the chemical behaviour
of cis-CHD to trans-CHD should be handled with care.
The stereochemistry of 9 was proven by synthesis of ent-
streptol (ent-valienol) 11. Reduction of the ester function with
diisopropylaluminium hydride (DIBAL-H) and deprotection
with tris(dimethylamino)sulfonium difluorotrismethylsilicate
as a fluoride source gave ent-streptol 11. The spectroscopic
data of 11 are identical to those published by Sakuda et al.[27]
Streptol was isolated froma culture of a Streptomyces sp. by
Sakuda et al. and inhibited the growth of lettuce seedlings at a
concentration above 13 ppm.[27] Valienol, synonymous with
streptol, was isolated from Actinoplanes sp. in the context of
acarbose biosynthesis studies.[31] rac-11 was synthesised by
Suami et al.[32a] and Block.[32b]
¬
Vario EL (Heraeus) at the analytical department of the Kekule Institut f¸r
Organische Chemie und Biochemie (University of Bonn). Melting points
were measured on a B¸chi B-540 heating unit. UV spectra were recorded
with an Ultrospec 2000 UV/Vis spectrophotometer (Pharmacia Biotech,
Sweden). Identification and quantification of metabolites was performed
by using an HPLC (HP series 1100, Hewlett Packard), fitted with a diode-
array detector, and equipped with a LiChrospher¾ C8 column (25 cm Â
3 m m , 5mm particle size, CS Chromatographie Service GmbH, Langer-
wehe, Germany). The metabolites were quantified by integration of peaks
at a wavelength of 275 nm. The injection volume was 5 mL. The initial
mobile phase was water (0.1% trifluoroacetic acid) at a flow rate of
0.45 mLminÀ1. 5 min after injection the eluent was changed linearly to a
50% ratio of methanol/water at 40 min. Afterwards the column was
washed, maintaining this composition for 10 min, and regenerated for
15 min. Retention times are as follows: chorismate 25.0 min, isochorismate
21.2 min, 2,3-trans-CHD 4.4 min and 3,4-trans-CHD 6.7 min. The limit of
quantification was ꢀ16 mm for all metabolites.
We have shown that the dihydroxylation of trans-CHD can
be carried out in a regio- and stereoselective manner. In
combination with stereoselective epoxidation and subsequent
nucleophilic opening, this will enable straight-forward diver-
sity-oriented access to regio- and stereoisomeric (amino)-
carbasugars[33] of potential biological activity.
Bacterial strains, cloning material and growth media: The strains, plasmids
and primers used in this study are shown in Table 2. Primers were
synthesised at MWG-Biotech AG, Ebersberg, Germany. E. coli strains
were grown on Luria Bertani (LB) medium.[35] Ampicillin (100 mgLÀ1
was added when required.
)
The synthetic medium for analyses of product distribution was a slightly
modified medium of Pan and co-workers,[36] containing per L: KH2PO4
(13 g), K2HPO4 (10 g), NaH2PO4 ¥ 2H2O (6 g), (NH4)2SO4 (2 g), MgSO4 ¥
7H2O (3 g), NaCl (5 g), FeSO4 ¥ 7H2O (40 mg), CaCl2 ¥ 2H2O (40 mg),
MnSO4 ¥ 2H2O (10 mg), ZnSO4 ¥ 7H2O (2 mg), AlCl3 ¥ 6H2O (10 mg),
CoCl2 ¥ 6H2O (4 mg), Na2MoO4 ¥ 2H2O (2 mg), CuCl2 ¥ 2H2O (1 mg),
H3BO3 (0.5 mg), leucine (200 mg), proline (200 mg), adenine (200 mg),
tryptophan (200 mg), thiamine (20 mg) and 2,3-dihydroxybenzoic acid
(20 mg).
Conclusion
Production of 2,3-trans-CHD 1, an essential metabolite in
bacteria, plants and fungi, was efficiently accomplished by
methods of deregulating the metabolic flux in E. coli strains.
Advantageously, no metabolic by-products such as chorismate
were excreted even though the chorismate-utilising pathways
are not totally blocked. A deficiency of EntA activity
significantly increased the cell-mass-specific product-forma-
tion rate by a factor of 3. It was demonstrated that the
microbial production of 1 can be scaled up, thus preparing the
way for a preparative access to 1. By using a 30 L stirred tank
reactor, a final concentration of 1 of 4.6 gLÀ1 (0.17m) was
obtained within a 50 h process time. The molar yield relative
to glucose consumption was calculated to be 17%. Two
methods for separating 1 frommediumingredients were
established, the use of reactive extraction or anion exchange
chromatography. Starting from the purified product as an
example, we established short and straight-forward syntheses
of new carbasugar derivatives.
For long-termstorage, cells in the mid-exponential phase of growth were
harvested and shock-frozen in 50% glycerol suspension at À308C.
PCR amplification: EntB and EntC were amplified from chromosomal
DNA of E. coli strain W3110. PCR was performed by using a Pwo DNA
polymerase kit (Roche Diagnostics, Mannheim, Germany) and a Peltier
Thermal Cycler PTC200 (MJ Research, Waltham, MA, USA) in 100 mL
volume. The concentration of chromosomal DNA was 0.1 ngmLÀ1. EntB
was amplified with 1 pmolmLÀ1 of primers BBAM and BPST. After an
initial denaturation step of 5 min at 948C, 34 cycles of 4 min at 948C, 2 min
at 578C and 90 s at 758C were performed. A final extension step of 7 min at
758C and cooling to 48C completed the reaction. EntB was obtained as a
single product.
EntC was generated with 0.5 pmolmLÀ1 of primers CSAC and CBAM. The
reaction was subjected to an initial denaturation step of 2 min at 948C,
followed by 10 cycles of 30 s at 948C, 45 s at 408C, 1 min at 728C. Finally, 20
cycles of 30 s at 948C, 45 s at 408C and 1 min ‡ 10 s per cycle at 728C
finished the reaction before the mixture was cooled to 48C. Three products
of approximately 500, 900 and 2500 bp were obtained from PCR
amplification. The one with 900 bp (EntC) was isolated, purified on
preparative agarose gel and used for further cloning procedures.
We are presently working on an enantioselective access to
3,4-trans-CHD 2 through methods of metabolic engineering in
recombinant microorganisms.[34] We are convinced that trans-
CHD 1 and 2 and derivatives thereof can be of similar value
for preparative organic chemistry as has been described for
the well-examined and commercially available cis analogues.
PCR products were analysed by electrophoresis in 0.8% (w/v) agarose gel
in 1 Â TAE buffer (Tris (40 mm), acetic acid (20 mm), EDTA (1 mm),
Chem. Eur. J. 2003, 9, 4188 4196
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4193

M. M¸ller et al.
FULL PAPER
Table 2. Strains, plasmids and primers used in this study.
Relevant characteristic(s)
Origin and refs.
Escherichia coli strains
FÀ endA1 hsdR17(rkÀmk ) recA1 supE44 thi-1 D(lacZYA-argF)U169 F80lacZDM15
[19]
[16]
[20, 21][a]
[22][a]
[23][b]
‡
DH5a
W3110
AN193
H1882
PBB8
FÀ l IN(rrnD-rrnE) prototroph
trpE38 leuB6 proC14 lacY1 fhuA23 rpsL109(strR) lÀ entA403
araD139 D(argF-lac)169 lÀ flhD5301 rpsL150(strR) D(fepA-ent)
‡
FÀ lacIq D(Ion) hflA150:Tn10 D(argF-lac)169 proA araD139 rpsL lÀ entCÀ menFÀ
Plasmids
pJF119EH1
pDF1
pDF2
pDF3
ampR, cloning vector
[17]
ampR (entB obtained with primers BBAM and BPST cloned into pJF119EH1 cut with BamHI and PstI)
ampR (entC obtained with primers CSAC and CBAM cloned into pJF119EH1 cut with SacI and BamHI)
ampR (entC obtained with primers CSAC and CBAM cloned into pDF1 cut with SacI and BamHI)
this study
this study
this study
Primers
BBAM
BPST
CSAC
CBAM
(BamHI) 5'-TATGGATCCACGCGCATCAGCCTGAA-3'
(PstI) 5'-GGGCTGCAGACATTTTTACCGCTG-3'
(SacI) 5'-GGCGAGCTCATTATTAAAGCCTTT-3'
(BamHI) 5'-TGCGGATCCTCGCTCCTTAATGC-3'
this study
this study
this study
this study
[a] Obtained fromProfessor Dr. Klaus Hantke, University of T¸bingen, Germany. [b] Obtained fromProfessor Dr. Eckhard Leistner, University of Bonn
,
Germany.
TM
pH 8), stained with ethidiumbromide. GeneRuler
1 kb DNA Ladder
above. At an optical density of 3.0 (l ˆ 600 nm), cells were harvested
(5000 g, 58C, 5 min) and resuspended in 100 mL synthetic medium (pH 7
with 1m NaOH) that contained IPTG (100 mm) and glucose (15 gLÀ1).
Incubation was done in 1 L shake flasks at 378C and 150 rpm. Aliquots
were taken at 1 h intervals. Maximal excretion rates (1) were found 2 h
after incubation.
(MBI Fermentas, Vilnius, Lithuania) was used as molecular size marker.
Construction of plasmids/transformation of strains: Plasmids were con-
structed by using standard recombinant techniques.[35] Plasmid pJF119EH1
and the PCR product of EntB were cut with BamHI and PstI and
afterwards ligated with T4 DNA ligase. Transformation of E. coli DH5a
with the resulting pDF1 gave the recombinant strain which shows
ampicillin resistance. Correct ligation and transformation were verified
by restriction site analyses and sequence analysis of the polylinker region
and the EntB insert.
For the determination of optimal fermentation conditions the following
parameters were varied. The pH was regulated (1n HCl or 1n NaOH) from
pH 3 to pH 9 (steps of 1 pH unit). The cultivation temperature was tested at
208C, 30 8C and 378C. Production of 1 was maximal under physiological
conditions at 378C and pH 7. Glucose was substituted by mass-equivalent
quantities of the following other C sources: galactose, glycerol, fructose,
yeast extract, acetic acid and lactic acid. For the induction experiments
IPTG was added to the preculture and production media to the following
final concentrations: 0, 5, 10, 20, 40, 60, 80, 100, 200, 500 mm.
The PCR product of EntC was restricted with PstI and SacI and
analogously cloned into the vectors pJF119EH1 and pDF1 to give vectors
pDF2 and pDF3, respectively. Transformation of E. coli DH5a with these
constructs, followed by restriction analysis and sequencing of the isolated
plasmid, verified the correct insertions of EntC.
Metabolite quantification: For the quantification of metabolites, aliquots of
cell suspension were centrifuged (5000 g, 48C, 10 min), and afterwards the
supernatant was diluted with water until maximum concentrations of
500 mgLÀ1 of 2,3-trans-CHD were achieved. The diluted samples were
analysed by HPLC.
Strains W3110, AN193, H1882 and PBB8 were also transformed with
plasmids pJF119EH1, pDF1, pDF2 and pDF3.
Enzyme expression analyses: For verification of enzyme expression, cell
extracts of all E. coli strains (host DH5a W3110, AN193, H1882, PBB8 with
plasmids pJF119EH1, pDF1, pDF2 or pDF3) were analysed with sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). E. coli
precultures (100 mL) from LB medium (5 mL, 150 rpm, 378C, overnight)
were used as inoculum for 100 mL volume of LB medium containing IPTG
(100 mm) as inductor. Incubation was done in 1 L shake flasks (150 rpm,
378C, 10 h). Cells from30 mL suspension were harvested by centrifugation
(5000 g, 58C, 10 min) and resuspended in distilled water (1 mL). Cell
disruption was done with an Ultrasonic Cell Disruptor ™Sonifier 250∫
(Branson Inc., Danbury, USA) at 20% duty cycle and output 2 for 4 min on
ice. The cell homogenate was centrifuged (19000 g, 58C, 10 min) to remove
cell debris, and the supernatant was analysed. Total protein concentrations
Fed-batch fermentation in a 30 L stirred tank bioreactor: Fermentations in
a 30 L stirred tank bioreactor (Chemap, Switzerland, 20 L working volume)
were done at pH 6.8 (controlled) and 378C with a mineral-salts medium.
1 L preculture from LB medium (150 rpm, 378C, overnight) served as
inoculum. The initial concentration of glucose was 30 gLÀ1 (0.17m).
Glucose was added in portions so that the concentration was maintained
in the range of 5 to 10 gLÀ1 (28 to 56 mm). Induction by addition of IPTG
(100 mm final concentration) was done at a cell concentration of 5 gLÀ1 (dry
cell mass). The aeration rate was regulated to 25 LminÀ1 (air). Antifoam
AF298 (Sigma Co., St. Louis, USA) was added in a sterile manner as
required. Fermentation was stopped after 50 h. Cells and insoluble
components were removed by centrifugation (4000 rpm, 48C, 15 min) to
give a brown solution.
were determined with protein assay reagent (Bio-Rad Laboratories,
[37]
Hercules, CA, USA) by using bovine serumalbumin as standard.
SDS-
PAGE was performed with a XcellII Mini-Cell electrophoresis apparatus
(NOVEX Experimental Technology, San Diego, CA, USA). Proteins were
separated in NuPAGE 4 12% (w/v) Bis-Tris polyacrylamide gels and
stained with Coomassie brilliant blue R-250 (Serva Electrophoresis GmbH,
Heidelberg, Germany). Sigmamarker Low Range (Sigma Chemical Co., St.
Louis, USA) was used as standard.
Preparative ion-exchange chromatography: Preparative ion-exchange
chromatography was done with a IndEX Tm Column 100, inner diameter
100 mm, 3.5 L maximal bed volume, from Amersham Pharmacia Biotech
(Uppsala, Sweden), by using DOWEX¾ 1 Â 8 (2 kg, 100 200 mesh, ClÀ
form) from Merck (Darmstadt, Germany). The resin was packed to a bed
volume of 2.6 L with the aid of a hydraulic stamp. Elution of the product
was monitored at a wavelength of 280 nm by using a UV/Vis detector
K-2001 fromKnauer (Berlin, Germany). The flow rate during the whole
Enzyme activity test: Enzyme activities were determined by monitoring the
degradation of chorismate (3). Solution A contained chorismic acid
(8 mm), MgCl2 ¥ 6H2O (10 mm) and Tris-HCl (200 mm) in aqueous solution
at pH 8.5. Solution B contained crude cell extract at a total protein
process was set to 70 mLminÀ1
.
À1
concentration of 85 mg m L . For analyses 50 mL of each solution A and B
For one separation cycle the resin was equilibrated with dipotassium
hydrogen phosphate (50 mm) at pH 8 and 9.5 mScmÀ1 (phosphoric acid,
potassiumhydroxide). Cell-free fermentation supernatant (8 L, pH 7)
containing product 1 (4.6 gLÀ1) was then passed through the column,
allowing the adsorption of the product. Elution was performed with formic
were gently mixed, and the degradation of 3 towards isochorismate (4), 2,3-
trans-CHD 1 or 3,4-trans-CHD 2 was monitored at 208C by HPLC.
Examination of excretion of 1: For the examination of product excretion,
100 mL cell suspensions from LB medium were prepared as described
4194
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4188 4196

2,3-Dihydroxy-2,3-dihydrobenzoic acid
4188 4196
acid (50 mm, pH 2.3, 1.2 mScmÀ1). Compound 1 was generally eluted at
pH 2.8 and 0.6 mScmÀ1. Regeneration of the DOWEX resin was done with
brine (2m).
were identical to the data obtained by Trost et al.[11a] for the (À)-
enantiomer. [a]D20 ˆ ‡322.1 (c ˆ 1.0 in CHCl3).
Methyl (3R,4R,5S,6S)-5,6-bis(tert-butyldimethylsilanyloxy)-3,4-dihydroxy-
cyclohex-1-enecarboxylate (9): A solution of K2OsO4 (7 mg, 19 mmol)
and LiOH (3 mg, 0.1 mmol) in water (1 mL) was added to a solution of 8
(775 mg, 1.9 mmol), N-methylmorpholine N-oxide (395 mg, 2.9 mmol) and
methanesulfonamide (278 mg, 2.9 mmol) in acetone (10 mL) and butan-1-
ol (9 mL). The resulting solution was stirred vigorously at RT for 16 h,
during which time it turned black. The reaction was quenched with
saturated aqueous NaHSO3 (5 mL). All volatiles were removed in vacuo,
and the crude product was purified by flash chromatography (isohexane/
EtOAc 10:1) to obtain 9 (467 mg, 56%) as a colourless oil. Rf ˆ 0.42
(isohexane/EtOAc 6:1); [a]2D0 ˆ À9.9 (c ˆ 0.9 in CHCl3); 1H NMR
(300 MHz, CDCl3, 238C): d ˆ 0.09 (s, 3H; CH3), 0.10 (s, 3H; CH3), 0.11
(s, 3H; CH3), 0.24 (s, 3H; CH3), 0.84 (s, 9H; tBu), 0.87 (s, 9H; tBu), 2.94 (d,
J ˆ 10 Hz, 1H), 3.78 (s, 3H; OCH3), 3.93 (m, 1H), 4.04 (d, J ˆ 9.5 Hz, 1H),
4.21 (dd, J ˆ 4.2, 2.9 Hz, 1H), 4.37 (m, 1H), 4.46 (m, 1H), 6.89 (dd, J ˆ 2.2,
1.6 Hz, 1H); 13C NMR (75 MHz, CDCl3, 238C): d ˆ À4.85, À4.83, À4.81,
À4.7 (SiCH3), 18.0, 18.2 (Cq), 25.7, 25.8 (CH3), 52.1 (OCH3), 66.8, 68.1, 70.1,
70.9 (CH), 129.3 (C-1), 141.6 (C-2), 166.5 (CO); MS (70 eV, EI): m/z (%):
Without neutralisation the eluate (7.8 L) was concentrated in vacuo to give
a yellow oil (500 mL). The syrup was lyophilised at 5 Â 10À2 mbar to give a
yellow-white solid (27.5 g). This corresponds to 75% yield for the
separation step. The purity was determined with HPLC analysis and
1H NMR to be 95%.
(S,S)-5,6-Dihydroxycyclohexa-1,3-dienecarboxylic acid [(S,S)-5,6-Dihy-
droxy-5,6-dihydrobenzoic acid] (1): [a]2D0 ˆ ‡3.8 (c ˆ 0.6 in ethanol);
1H NMR (300 MHz, [D4]MeOH, 238C): d ˆ 4.10 (d, J ˆ 2.5 Hz, 1H;
H-5), 4.50 (d, J ˆ 2.5 Hz, 1H; H-6), 6.20 (m, 2H; H-3,4), 7.06 (dd, J ˆ 3.3,
3.2 Hz, 1H; H-2); 13C NMR (75 MHz, [D4]MeOH, 238C): d ˆ 68.8, 70.4,
125.1 (CH), 130.8 (C-1), 134.3, 134.4 (CH), 170.2 (CO); IR(KBr): nÄ ˆ 1699,
1644, 1586, 1258, 1075, 1008 cmÀ1; UV/Vis (H2O): lmax(e) ˆ 279 nm(4900);
‡
‡
MS (70 eV, EI): m/z (%): 156 (6) [M ], 138 (100) [M À H2O], 110 (63), 93
(10), 82 (29), 65 (13); HR-MS: calcd for C7H8O4: 156.0423; found:
156.0424.
Methyl (S,S)-5,6-dihydroxycyclohexa-1,3-dienecarboxylate [Methyl (S,S)-
5,6-dihydroxy-5,6-dihydrobenzoate] (5): A solution of 1 (3.11 g, 96%
purity, 19 mmol) in anhydrous methanol (200 mL) was treated with a
solution of hydrogen chloride in methanol (ꢁ1.25m) and stirred for 48 h at
RT. The solution was neutralised by addition of sodiumhydrogen
carbonate (10.5 g) and concentrated in vacuo. The crude product was
purified by flash chromatography (EtOAc) to give 5 as yellow crystals
(2.02 g, 60%). M.p. 79 808C; Rf ˆ 0.35 (EtOAc); [a]2D0 ˆ ‡4.6 (c ˆ 0.45 in
ethanol); 1H NMR (300 MHz, CDCl3, 238C): d ˆ 2.50 (s, 1H; OH), 3.82 (s,
3H; OCH3), 4.05 (s, 1H; OH), 4.57 (ddd, J ˆ 9.3, 3.3, 1.9 Hz, 1H; H-5), 4.76
(dd, J ˆ 9.3, 1.3 Hz, 1H; H-6), 6.08 (ddd, J ˆ 9.6, 5.4 Hz, 1.9 Hz, 1H; H-3),
6.29 (ddd, J ˆ 9.6, 3.3, 0.9 Hz, 1H; H-4), 6.99 (d, J ˆ 5.4 Hz, 1H; H-2);
13C NMR (75 MHz, CDCl3, 238C): d ˆ 52.3 (OCH3), 72.5, 72.7, 122.9 (CH),
129.4 (C-1), 133.4, 136.4 (CH), 167.9 (CO); IR(KBr): nÄ ˆ 2971, 2931, 1725,
1638, 1454, 1372, 1276, 1109 cmÀ1; UV/Vis (H2O): lmax ˆ 282 nm; MS
‡
‡
‡
432 (0.5) [M ], 417 (2) [M À CH3], 399 (2), 375 (80) [M À C(CH3)3], 357
‡
(100) [M À C(CH3)3,H2O], 271 (20), 243 (20), 225 (15), 215 (16), 197 (14),
147 (15); IR(KBr): nÄ ˆ 3457, 2954, 2931, 2897, 2858, 1721, 1472, 1389, 1361,
1256, 1111.
(1R,2R,5S,6S)-5,6-Bis-(tert-butyl-dimethylsilanyloxy)-4-hydroxymethylcy-
clohex-3-ene-1,2-diol (10): An aliquot (1.5 mL) of a solution of DIBAL-H
in methylene chloride (1m) was added dropwise to a solution of 9 (200 mg,
0.45 mmol) in dry methylene chloride (5 mL) at 08C. The resulting pale
yellow solution was stirred for 3 h. The reaction was quenched by addition
of MeOH (2 mL) and hydrochloric acid (2 mL, 2m). The mixture was
dissolved in hydrochloric acid (10 mL, 2m), then brine (10 mL) and diethyl
ether (30 mL) were added. The aqueous layer was extracted with EtOAc
(3 Â 10 mL). The combined organic layers were dried (MgSO4) and
concentrated in vacuo. The crude product was recrystallised (isohexane/
EtOAc 20:1) to obtain 10 as white needles (173 mg, 92%). M.p. ˆ 110
1128C; Rf ˆ 0.38 (isohexane/EtOAc 4:1); [a]2D0 ˆ ‡2.4 (c ˆ 0.5 in CHCl3);
1H NMR (300 MHz, CDCl3, 238C): d ˆ 0.11 (s, 6H; 2  CH3), 0.17 (s, 3H;
CH3), 0.21 (s, 3H; CH3), 0.85 (s, 9H; tBu), 0.90 (s, 9H; tBu), 2.80 (d, J ˆ
11 Hz, 1H), 3.75 (dd, J ˆ 9.9, 4.7 Hz, 1H), 3.92 (m, 1H), 4.05 (m, 1H), 4.15
(m, 3H), 4.29 (d, J ˆ 11 Hz, 1H), 5.73 (d, J ˆ 1.4 Hz, 1H); 13C NMR
(75 MHz, CDCl3, 238C): d ˆ À5.0, À4.64, À4.62, À4.3 (SiCH3), 18.0, 18.1
(Cq), 25.8, 25.9 (CH3), 64.4 (CH2), 66.4, 68.9, 70.7, 71.0 (CH), 127.5 (C-3),
136.3 (C-4); IR(KBr): nÄ ˆ 3516, 3334, 2953, 2930, 2896, 2859, 1472, 1463,
1385, 1361, 1254, 1108, 1075, 1056 cmÀ1; elemental analysis calcd (%) for
C19H40O5Si2: C 56.39, H 9.96, found C 56.10, H 9.84.
‡
‡
(70 eV, EI): m/z (%): 170 (27) [M] , 152 (54) [M À H2O], 138 (91), 128 (2),
121 (89), 110 (100), 93 (38), 82 (69), 65 (50), 53 (38); HR-MS: calcd for
C8H10O4: 170.0579; found: 170.0580.
Methyl (3R,4R,5R,6S)-3,4,5,6-tetrahydroxycyclohex-1-enecarboxylate (6)
and methyl (3S,4S,5R,6S)-3,4,5,6-tetrahydroxycyclohex-1-enecarboxylate
(7):
a) OsO4 (2 mg, 8 mmol) was added to a solution of 5 (400 mg, 2.3 mmol) and
N-methylmorpholine N-oxide (653 mg, 4.8 mmol) in tert-butanol (5 mL)
and water (5 mL). The solution was stirred at RT for 72 h, during which
time it turned black. The reaction was quenched by addition of aqueous
NaHSO3 (1 mL). All volatiles were removed in vacuo, and the crude
product was purified by flash chromatography (EtOAc/MeOH 4:1) to give
a chromatographically inseparable 5:1 mixture of 6 and 7 as a white solid
(139 mg, 29%).
(1R,2R,3R,4S)-5-hydroxymethylcyclohex-5-ene-1,2,3,4-tetraol (ent-strep-
tol) (11): Compound 10 (110 mg, 0.27 mmol) was dissolved in dry
methylene chloride (5 mL) and tris(dimethylamino)sulfonium difluoro-
trismethylsilicate (280 mg, 1.01 mmol) was added at À788C. The solution
was stirred for 16 h and allowed to warmto RT. All volatiles were removed
in vacuo, and the crude product was purified by flash chromatography
(MeOH/EtOAc 1:2) to obtain 11 as a white solid (46 mg, 96%). Rf ˆ 0.45
(EtOAc, MeOH 2:1); [a]D20 ˆ À92.5 (c ˆ 0.2 in H2O); 1H NMR (300 MHz,
D2O, 23 8C): d ˆ 3.60 (dd, J ˆ 11, 4 Hz, 1H; H-2), 3.69 (dd, J ˆ 11, 8 Hz, 1H;
H-3), 4.13 (d, 14 Hz, 1H; CH2), 4.16 (m, 1H; H-4), 4.20 (d, 14 Hz, 1H;
CH2), 4.27 (d, 5 Hz, 1H; H-1), 5.83 (d, 5 Hz, 1H; H-6); 13C NMR (75 MHz,
D2O, 23 8C, TSP): d ˆ 64.1 (CH2), 68.9, 73.5, 75.0, 75.3 (CH), 124.9 (C-5),
144.9 (C-6); IR(KBr): nÄ ˆ 3331, 2913, 1420, 1099, 1064, 1019, 994.
b) A solution of 9 (50 mg, 0.1 mmol) in CH3CN (2 mL) was treated with
aqueous HF (50 mL, 40%) and the mixture was stirred at 08C for 3 h. Water
(5 mL) was added, and the solution was extracted with EtOAc (3 Â 15 mL).
The combined organic layers were dried over MgSO4, filtrated and
concentrated in vacuo to give 6 as a white wax without further purification
(7 mg, 30%). Rf ˆ 0.18 (CH3CN); 6: 1H NMR (300 MHz, [D4]MeOH,
238C): d ˆ 3.76 (dd, J ˆ 7.5, 4 Hz, 1H; H-3), 3.82 (s, 3H; OCH3), 4.05 (dd,
J ˆ 7.5, 5 Hz, 1H; H-4), 4.32 (d, J ˆ 5 Hz, 1H; H-6), 4.40 (pt, J ˆ 5 Hz, 1H;
H-5), 6.78 (d, J ˆ 4 Hz, 1H; H-2); 13C NMR (75 MHz, [D4]MeOH, 238C):
d ˆ 52.5 (OCH3), 67.3, 70.4, 71.4, 73.0 (CH), 134.1 (C-1), 139.3 (C-2), 168.4
(CO). 7: 1H NMR (300 MHz, [D4]MeOH, 238C): d ˆ 6.73 (dt, J ˆ 3, 0.9 Hz,
1H; H-2), all other signals are superposed; 13C NMR (75 MHz, [D4]MeOH,
238C): d ˆ 52.5 (OCH3), 68.6, 69.9, 72.2, 75.7 (CH), 133.2 (C-1), 141.1 (C-2),
168.3 (CO).
Acknowledgement
Methyl (S,S)-5,6-bis(tert-butyldimethylsilanyloxy)cyclohexa-1,3-dienecar-
boxylate (8): A solution of methyl ester
5
(18 mg, 0.11 mmol) and
D.F. was financially supported by the ™Graduiertenfˆrderung des Landes
Nordrhein Westfalen∫ and by the ™Studienstiftung des deutschen
Volkes∫. We thank the following individuals: Professor Dr. K. Hantke,
University of T¸bingen, and Professor Dr. E. Leistner, University of Bonn,
for kindly providing cultures of E. coli mutants H1882, AN193 and BN117;
U. Degner and R. Halbach for help with genetic work; Dipl.-Ing. H.-J.
Brandt and S. Stevens for assistance with fermentations; P. Geilenkirchen
and S. Bode for technical assistance. We gratefully thank Professor Dr.
triethylamine (20 mg, 0.2 mmol) in dry methylene chloride (10 mL) was
treated with tert-butyldimethylsilyltriflate (34 mg, 1.3 mmol). After being
stirred for 18 h at RT, the remaining reactant was hydrolysed by quenching
with saturated aqueous Na2CO3 (4 mL). The solution was diluted with
diethyl ether and dried (MgSO4). Volatiles were removed in vacuo, and the
crude product was purified by flash chromatography (isohexane/EtOAc
20:1 to 9:1) to give 8 as colourless oil (42 mg, 95%). All analytical data
Chem. Eur. J. 2003, 9, 4188 4196
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4195

M. M¸ller et al.
FULL PAPER
Christian Wandrey for the generous support always granted. We gratefully
acknowledge the reviewers for helpful comments and suggestions.
[15] D. Franke, G. A. Sprenger, M. M¸ller, Angew. Chem. 2001, 113, 578
581; Angew. Chem. Int. Ed. 2001, 40, 555 557.
[16] a) C. W. Hill, B. W. Harnish, Proc. Natl. Acad. Sci. USA 1981, 78,
7069 7072; b) K. F. Jensen, J. Bacteriol. 1993, 175, 3401 3407.
[17] J. P. F¸rste, W. Pansegrau, R. Frank, H. Blˆcker, P. Scholz, M.
Bagdasarian, E. Lanka, Gene 1986, 48, 119 131.
[18] a) M. S. Nahlik, T. J. Brickman, B. A. Ozenberger, M. A. McIntosh, J.
Bacteriol. 1989, 171, 784 790; b) B. A. Ozenberger, T. J. Brickman,
M. A. McIntosh, J. Bacteriol. 1989, 171, 775 783.
[19] D. Hanahan, J. Mol. Biol. 1983, 166, 557 580.
[20] G. B. Cox, F. Gibson, R. K. J. Luke, N. A. Newton, I. G. O×Brien, H.
Rosenberg, J. Bacteriol. 1970, 104, 219 226.
[21] Characterisation of E. coli strain AN193 cf. a) R. Wayne, K. Frick, J. B.
Neilands, J. Bacteriol. 1976, 126, 7 12; b) T. P. Fleming, M. S. Nahlik,
M. A. McIntosh, J. Bacteriol. 1983, 156, 1171 1177; c) M. S. Nahlik,
T. P. Fleming, M. A. McIntosh, J. Bacteriol. 1987, 169, 4163 4170;
d) Ref. [18b].
[1] For some recent examples see: a) J. C. Rohloff, K. M. Kent, M. J.
Postich, M. W. Becker, H. H. Chapman, D. E. Kelly, W. Lew, M. S.
Louie, L. R. McGee, E. J. Prisbe, L. M. Schultze, R. H. Yu, L. Zhang,
J. Org. Chem. 1998, 63, 4545 4550; b) M. Federspiel, R. Fischer, M.
Hennig, H.-J. Mair, T. Oberhauser, G. Rimmler, T. Albiez, J. Bruhin,
H. Estermann, C. Gandert, V. Gˆckel, S. Gˆtzˆ, U. Hoffmann, G.
Huber, G. Janatsch, S. Lauper, O. Rˆckel-St‰bler, R. Trussardi, A. G.
Zwahlen, Org. Process Res. Rev. 1999, 3, 266 274; c) M. Karpf, R.
Trussardi, J. Org. Chem. 2001, 66, 2044 2051; d) C. U. Kim, W. Lew,
M. A. Williams, H. Liu, L. Zhang, S. Swaminathan, N. Bischofberger,
M. S. Chen, D. B. Mendel, C. Y. Tai, W. G. Laver, R. C. Stevens, J. Am.
Chem. Soc. 1997, 119, 681 690.
[2] A. Barco, S. Benetti, C. D. Risi, P. Marchetti, G. P. Pollini, V. Zanirato,
Tetrahedron: Asymmetry 1997, 8, 3515 3545, and references therein.
[3] J. D. White, S. G. Toske, T. Yakura, Synlett 1994, 591 593.
[4] a) D. S. Tan, M. A. Foley, M. D. Shair, S. L. Schreiber, J. Am. Chem.
Soc. 1998, 120, 8565 8566; b) D. S. Tan, M. A. Foley, B. R. Stockwell,
M. D. Shair, S. L. Schreiber, J. Am. Chem. Soc. 1999, 121, 9073 9087;
c) B. R. Stockwell, D. S. Tan, M. A. Foley, M. D. Matthew, S. L.
Schreiber, S. M. Scott, WO 00/06525, 2000 [Chem. Abstr. 2000, 132,
151600r].
[5] a) W. Lew, M. A. Williams, D. B. Mendel, P. A. Escarpe, C. U. Kim,
Bioorg. Med. Chem. Lett. 1997, 7, 1843 1846; b) L. Zhang, M. A.
Williams, D. B. Mendel, P. A. Escarpe, C. U. Kim, Bioorg. Med. Chem.
Lett. 1997, 7, 1846 1850.
[6] a) I. I. Salamon, B. D. Davis, J. Am. Chem. Soc. 1953, 75, 5567 5571;
b) K. Li, M. R. Mikola, K. M. Draths, R. M. Worden, J. W. Frost,
Biotechnol. Bioeng. 1999, 64, 61 73; c) K. Li, J. W. Frost, Biotechnol.
Prog. 1999, 15, 876 883.
[7] a) B. D. Davis, J. Biol. Chem. 1951, 191, 315 325; b) K. M. Draths,
D. R. Knop, J. W. Frost, J. Am. Chem. Soc. 1999, 121, 1603 1604;
c) J. M. Gibson, P. S. Thomas, J. D. Thomas, J. L. Barker, S. S.
Chandran, M. K. Harrup, K. M. Draths, J. W. Frost, Angew. Chem.
2001, 113, 1999 2002; Angew. Chem. Int. Ed. 2001, 40, 1945 1948;
d) S. S. Chandran, J. Yi, K. M. Draths, R. von Daeniken, W. Weber,
J. W. Frost, Biotechnol. Prog. 2003, 19, 808 814.
[8] a) U. Weiss, B. D. Davis, E. S. Mingioli, J. Am. Chem. Soc. 1953, 75,
5572 5576; b) K. Li, J. W. Frost, Biotechnol. Prog. 1999, 15, 876 883.
[9] a) K. M. Draths, T. L. Ward, J. W. Frost, J. Am. Chem. Soc. 1992, 114,
9725 9726; b) J. W. Frost, K. M. Draths, WO 94/08015, 1994 [Chem.
Abstr. 1994, 121, 7466g].
[10] a) F. Gibson, Biochem. Prep. 1968, 12, 94 97; b) C. E. Rieger, J. L.
Turnbull, Prep. Biochem. Biotechnol. 1996, 26, 67 76; c) C. Grisos-
tomi, P. Kast, R. Pulido, J. Huynh, D. Hilvert, Bioorg. Chem. 1997, 25,
297 305.
[11] a) B. M. Trost, L. S. Chupak, T. L¸bbers, J. Am. Chem. Soc. 1998, 120,
1732 1740; b) T. K. M. Shing, E. K. W. Tam, J. Org. Chem. 1998, 63,
1547 1554; c) T. K. M. Shing, E. K. W. Tam, Tetrahedron: Asymmetry
1996, 7, 353 356; d) M. R. Demuth, P. E. Garrett, J. D. White, J. Am.
Chem. Soc. 1976, 98, 634 635; e) S. Ogawa, T. Takagaki, J. Org.
Chem. 1985, 50, 2356 2359; f) S. Ogawa, T. Takagari, Bull. Chem. Soc.
Jpn. 1987, 60, 800 802; g) R. H. Schlessinger, A. Lopes, J. Org. Chem.
1981, 46, 5253 5254; h) T. Hudlicky, G. Seoane, T. Pettus, J. Org.
Chem. 1989, 54, 4239 4243.
[12] C. F. Earhart in Escherichia coli and Salmonella, Vol. 1, 2nd ed., (Ed.:
F. C. Neidhardt) ASM, Washington, DC, 1996, pp. 1075 1102.
[13] T. A. Cropp, D. J. Wilson, K. A. Reynolds, Nat. Biotechnol. 2000, 18,
980 983.
[22] a) M. J. Casadaban, J. Mol. Biol. 1976, 104, 541 555; b) M. J.
Casadaban, S. N. Cohen, Proc. Natl. Acad. Sci. USA 1979, 76, 4530
4533.
[23] a) R. M¸ller, C. Dahm, G. Schulte, E. Leistner, FEBS Letters 1996,
378, 131 134; b) C. Dahm, R. M¸ller, G. Schulte, K. Schmidt, E.
Leistner, Biochim. Biophys. Acta 1998, 1425, 377 386.
[24] This was calculated by using the following data: a) Total dry weight
per cell is 2.8 Â 10À13 g, cf. F. C. Neidhardt, H. Umbarber in Escher-
ichia coli and Salmonella, Vol. 1, 2nd ed., (Ed.: F. C. Neidhardt) ASM,
Washington, DC, 1996, pp. 13 16; b) Cell volume is 1.6 Â 10À18 m3, cf.
J. D. Watson, Molecular Biology of the Gene, Benjamin, New York,
1965.
[25] a) R. Haensel, W. Halwachs, K. Sch¸gerl, Chem. Eng. Sci. 1986, 41,
1811 1815; b) M. M. Bora, N. N. Dutta, K. G. Bhattacharya, Chem.
Eng. Commun. 2000, 179, 15 34; c) P. Dzygiel, P. Wieczorek, K.
Mathiasson, J. ä. Jˆnsson, Anal. Lett. 1998, 31, 1261 1274; d) J. A.
Adarkar, S. B. Sawant, J. B. Joshi, V. G. Pangarkar, Biotechnol. Prog.
1997, 13, 493 496.
[26] By using methods of ion-exchange chromatography, trans-CHDs have
already been isolated fromcultivation mediumfor analytical pur-
poses: a) I. G. Young, F. Gibson, C. G. MacDonald, Biochim. Biophys.
Acta 1969, 192, 62 72; b) Ref. [14].
[27] A. Isogai, S. Sakuda, J. Nakayama, S. Watanabe, A. Suzuki, Agric.
Biol. Chem. 1987, 51, 2277 2279.
[28] V. Lorbach, D. Franke, M. Nieger, M. M¸ller, Chem. Commun. 2002,
494 495.
[29] I. B. Masesane, P. G. Steel, Synlett 2003, 735 737.
[30] A. J. Blacker, R. J. Booth, G. M. Davies, J. K. Sutherland, J. Chem.
Soc. Perkin Trans. 1 1995, 22, 2861 2870.
[31] T. Mahmud, I. Tornus, E. Egelkrout, E. Wolf, C. Uy, H. G. Floss, S.
Lee, J. Am. Chem. Soc. 1999, 121, 6973 6983.
[32] a) T. Toyokuni, Y. Abe, S. Ogawa, T. Suami, Bull. Chem. Soc. Jpn.
1983, 56, 505 513; b) O. Block, Dissertation, University of Wuppertal
(Germany), 2000.
[33] a) D. Franke, C. Dose, S. Esser, V. Lorbach, M. M¸ller, unpublished
results; b) D. Franke, Dissertation, University of Bonn (Germany),
2001.
[34] D. Franke, G. A. Sprenger, M. M¸ller, ChemBioChem 2003, 4, 775
777.
[35] J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning:
A
Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory, N.Y.
(USA) 1989.
[36] J. G. Pan, J. S. Rhee, J. M. Lebeault, Biotechnol. Lett. 1987, 9, 89 94.
[37] M. M. Bradford, Anal. Biochem. 1976, 72, 248 254.
[14] R. M¸ller, M. Breuer, A. Wagener, K. Schmidt, E. Leistner, Micro-
biology 1996, 142, 1005 1012.
Received: July 22, 2002
Revised: May 19, 2003 [F4265]
4196
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 4188 4196