Previous Article | Next Article 
Applied and Environmental Microbiology, April 2005, p. 2199-2202, Vol. 71, No. 4
0099-2240/05/$08.00+0 doi:10.1128/AEM.71.4.2199-2202.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Opposite Enantioselectivities of Two Phenotypically and Genotypically Similar Strains of Pseudomonas frederiksbergensis in Bacterial Whole-Cell Sulfoxidation
Waldemar Adam,1,2
Frank Heckel,3
Chantu R. Saha-Möller,1
Marcus Taupp,3
Jean-Marie Meyer,4 and
Peter Schreier3*
Institute of Organic Chemistry,1
Institute of Food Chemistry, University of Würzburg, Würzburg, Germany,3
Department of Chemistry, Facundo University of Puerto Rico, Rio Piedras, Puerto Rico,2
Laboratoire de Microbiologie et Génétique, Université Louis Pasteur, Strasbourg, France4
Received 19 August 2004/
Accepted 12 November 2004

ABSTRACT
Soil samples were screened to select microorganisms with the
capability to oxidize organic sulfides into the corresponding
sulfoxides with differential enantioselectivities. Several bacterial
strains that preferentially produced the
S-configured sulfoxide
enantiomer were isolated. Surprisingly, one bacterial strain,
genotypically and phenotypically characterized as
Pseudomonas frederiksbergensis, selectively gave the
R enantiomer. The finding
that two apparently identical organisms displayed opposite enantioselectivities
is novel for non-genetically modified organisms.

INTRODUCTION
Sulfoxides are versatile and convenient chiral building blocks
in asymmetric synthesis. In particular, enantiomerically pure
sulfoxides play an important role either as chiral building
blocks or as stereo-directing groups (
20). The value of sulfoxide
functionality is further illustrated by their diverse biological
activities and pharmaceutical uses. In most cases, only one
enantiomer is responsible for the desired biological activity.
Enzyme-catalyzed asymmetric sulfoxidations have been developed in recent years for the preparation of chiral sulfoxides with an enantiomeric excess (ee value; for a definition, see Table 1) up to 100% (2, 3, 5, 9, 10, 17). Especially peroxidases efficiently catalyze the enantioselective sulfoxidation of a variety of sulfides (4, 11, 23). Several whole-cell systems have also been successfully employed in biocatalytic sulfoxidation; among them, some bacterial strains hold great promise for the development of environmentally benign, efficient asymmetric sulfoxidation (16, 19, 22). For example, we have shown that the topsoil bacterium Pseudomonas frederiksbergensis, recently isolated at a coal gasification site in Frederiksberg, Copenhagen, Denmark (6, 18), catalyzes the asymmetric oxidation of sulfides to enantiomerically enriched sulfoxides (1). Since the hitherto known isolated enzymes and bacterial whole-cell systems preferentially produce only one of the two enantiomers (either R or S), finding bacteria delivering both enantiomers is a promising goal. So far, only one example appears to be known, namely, a Pseudomonas putida strain, which produces one enantiomer of the sulfoxide in excellent enantioselectivity, whereas a genetically engineered mutant of this bacterium and the native strain afford the opposite enantiomers (8).
View this table:
[in this window]
[in a new window]
|
TABLE 1. Enantioselectivity in the asymmetric oxidation of phenyl methyl sulfide (Ia) to the corresponding sulfoxide IIa by isolated soil microorganisms
|
To find soil microorganisms for asymmetric sulfoxidation, a
sulfide-mediated screening was carried out. Strains were isolated
from soil samples taken in Geiselwind, Germany, by treatment
with 10 µl of phenyl methyl sulfide (Ia) per plate on
minimal agar, with glucose as the carbon source. After isolation
of the sulfur-adapted microorganisms, these pure cultures were
used for standard oxidations of phenyl methyl sulfide (Ia) as
substrate to the enantiomerically enriched phenyl methyl sulfoxide
(IIa), as displayed in Fig.
1. Few strains possessed acceptable
efficiency for asymmetric sulfoxidation (Table
1) and were therefore
characterized by partial sequencing of the 16S rRNA gene. The
preparation of intact bacterial cells for PCR was performed
as described in the literature (
24). Single-stranded DNA was
sequenced with internal primers by using the
Taq-cycle DyeDeoxy
Terminator method. Finally, a homology search was performed
(
15) with the acquired sequence data against the European Molecular
Biology Laboratory (EMBL) database (
14). The results presented
in Table
1 reveal that only one strain, namely, strain 33 (Table
1, entry 3), produced the
R enantiomer of phenyl methyl sulfoxide
(IIa). All other strains produced mainly the
S enantiomer; indeed,
in the case of
Arthrobacter aurescens (Table
1, entry 5), an
enantiomeric excess (
ee) of 92% (
S) was achieved. The identified
Arthrobacter chlorophenolicus strain (Table
1, entry 1), as
well as the
P. putida,
Pseudomonas fluorescens, and
Pseudomonas aureofaciens strains (Table
1, entries 2, 4, and 6), converted
the substrate phenyl methyl sulfide (Ia) to the corresponding
sulfoxide IIa in
ee values of 50 to 60%.
Strain 33 was characterized by standard phenotypical methods.
Therefore, its appearance (by microscopic study, which included
Gram stain) and behavior (optimal growth temperature and pH
value) were assessed. The motility and the oxidase and catalase
activities were positive. The classical API 20 NE test (bioMerieux,
Nürtingen, Germany) was performed to compare strain 33
with the commercially available strain
P. frederiksbergensis DSM 13022, which was purchased from the Deutsche Sammlung für
Mikroorganismen und Zellkulturen (DSMZ; Braunschweig, Germany).
This method combines 8 conventional tests and 12 assimilation
tests for the identification of nonfastidious, gram-negative
rods that belong to the
Enterobacteriaceae, e.g.,
Pseudomonas.
All observed color reactions and assimilation tests were identical
for strain 33 and
P. frederiksbergensis DSM 13022; for example,
potassium nitrate was reduced to nitrite by both strains. The
color reactions for detecting indole production with tryptophan
as substrate and acidification with glucose were negative. The
various tests for enzyme activity were negative, including arginine
dihydrolase, urease, ß-glucosidase and protease, and
ß-galactosidase. The assimilation tests with standard
substrates as carbon sources (glucose, arabinose, mannose, mannitol,
N-acetyl-glucosamine, gluconate, caprate, malate, citrate, and
phenyl acetate) were positive; only maltose and adipate were
negative. Finally, tetramethyl-
p-phenylenediamine yielded a
positive cytochrome oxidase reaction. To confirm the phenotypic
identity between strain 33 and
P. frederiksbergensis DSM 13022,
the cellular fatty acid profile of strain 33 was determined
by the identification service of the DSMZ and shown to be typical
for
Pseudomonas organisms. Altogether, the phenotypical characterization
identified strain 33 as
P. frederiksbergensis.
A direct comparison of the 16S rRNA partial sequences indicated that both bacteria belong to the species P. frederiksbergensis. The sequences of P. frederiksbergensis DSM 13022 and strain 33 are identical for 441 base pairs, which strongly suggests that the 16S rRNA genes of the two strains are phylogenetically the same.
For further verification, siderophore typing, a powerful tool for the identification of even nonfluorescent pseudomonads, was used. This newly developed technique (21) measures the iron uptake of the siderophores located in the P. frederiksbergensis strains by means of a cross-incorporation comparison. Siderophore typing experiments were carried out according to the previously described procedure (21) in duplicate with strain 33 by employing the commercially available P. frederiksbergensis DSM 13022 and P. frederiksbergensis JAJ 28 from J.-M. Meyer's collection (21). The siderophore system of P. frederiksbergensis DSM 13022 accomplished an iron uptake of 95% efficiency compared to that of strain 33. Conversely, strain 33 has a siderophore-mediated iron uptake of 99% efficiency with regard to P. frederiksbergensis DSM 13022. Identical results were obtained in comparison with P. frederiksbergensis JAJ 28. Thus, our strain 33 possesses the same siderotype as P. frederiksbergensis DSM 13022 and JAJ 28; accordingly, strain 33 belongs to the same siderovar, namely that of P. frederiksbergensis.
In order to determine substrate selectivity and the effects of variation in the sulfur substituent on the sense (R versus S) and extent (enantiomeric excess) of the selectivity of the sulfoxidation reactions, a series of sulfides (Ia to Ih) were tested in whole-cell cultures. The bacterial sulfoxidation experiments were carried out in 300-ml Erlenmeyer flasks, which were charged with 100 ml of Dworkin mineral nutriment solution (13) and 375 µl of Pseudomonas trace element solution (7). This medium was sterilized at 121°C for 16 min prior to the addition of 750 µl of 50% sterile glucose solution. After inoculation with bacterial material, the culture was pregrown for 24 h and subsequently shaken in the presence of 100 µmol of the appropriate sulfide (analytical grade; obtained from Fluka [Seelze, Germany], Sigma-Aldrich [Taufkirchen, Germany], or Lancaster [Morecambe, United Kingdom]) for 18 h at 30°C and 120 rpm. Subsequently, the bacterial cells were treated by ultrasound to release any sulfoxides retained in the cells. After centrifugation (9,000 rpm for 15 min in a Beckman J2-21 centrifuge equipped with a JA 10 rotor), the bacterial suspension was successively extracted twice with 100 ml of ethyl ether and 100 ml of dichloromethane (the solvents were freshly distilled before use). Finally, the products were analyzed by gas chromatography coupled with mass spectrometry (GC-MS). The chiral analysis of the products was performed by a multidimensional GC-MS combination on a 2,3-diethyl-6-t-butyldimethylsilyl-ß-cyclodextrin column. The absolute configurations of the sulfoxides were determined by comparison with the authentic samples, prepared by enzymatic sulfoxidation with chloroperoxidase according to the previously reported procedure (11), and with data from the literature (12).
The results for P. frederiksbergensis strain 33 and for the commercially available P. frederiksbergensis DSM 13022 are compared in Table 2. The yields ranged from 4 to 48% (data not shown) and were not optimized, since the emphasis of this work lies on the opposite enantioselectivities of the P. frederiksbergensis strains. As for the specific results, phenyl methyl sulfide (Ia) was sulfoxidized in an enantiomeric excess (ee) of 91% in favor of the S enantiomer by DSM 13022, whereas strain 33 produced the R enantiomer with an ee value of 62% (Table 2, entry 1). Introduction of a methyl group in the para position (Table 2, entry 2, Ib) afforded the enantiopure S-sulfoxide for DSM 13022, and again, mainly the R enantiomer was produced by strain 33. To assess the effect of the alkyl side chain on enantioselectivity, the n-propyl derivative Ic was employed as a substrate. A significantly lower enantioselectivity (81% ee in favor of S) for DSM 13022 and a poor ee value (that is, <5%) for strain 33 compared to the methyl derivative Ia were observed (Table 2, entry 3). Thus, the chain length has a pronounced effect on the enantioselectivity of sulfoxidation by both P. frederiksbergensis strains. Moreover, a branched alkyl chain dramatically lowers the substrate acceptance; for example, the isopropyl derivative Id was barely consumed by both bacterial strains (Table 2, entry 4). The enantioselectivity of strain 33, in contrast to that of DSM 13022, depends on the electronic properties of the para substituents. The oxidation of the electron-donating methoxy derivative Ie by strain 33 yielded the R sulfoxide with the highest ee value of 89% (Table 2, entry 5), while the electron-withdrawing chloro derivative If afforded the same sulfoxide enantiomer in 56% ee (Table 2, entry 6). In contrast, DSM 13022 oxidized both derivatives in ee values of 96 to 99% with a preference for the S enantiomer. Not only the alkyl aryl sulfides Ia to If but also dialkyl sulfides, e.g., cyclohexyl methyl sulfide (Ig), were enantioselectively sulfoxidized by DSM 13022 (70% S) and strain 33 (23% R), as shown in Table 2, entry 7. In contrast, the benzyl methyl sulfide (Ih) was not accepted by either strain (Table 2, entry 8).
View this table:
[in this window]
[in a new window]
|
TABLE 2. Enantioselective oxidation of sulfides I to sulfoxides II by P. frederiksbergensis DSM 13022 and P. frederiksbergensis strain 33
|
The fact that the standard characterization method (partial
sequencing of the 16S rRNA) exposed that strain 33 exclusively
biosynthesizes the
R-configured enantiomer in the sulfoxidation
of aryl alkyl sulfides, in contrast to the commercially available
strain DSM 13022, which preferentially expresses the
S-configured
sulfoxide, obliged extensive additional experimental probing
into this novel phenomenon. This extensive examination included
color reactions, assimilation tests, fatty acid profiles, and
the powerful identification tool known as siderophore typing,
which all confirmed a strong similarity between the two strains
described. Herein, by all of the tested methods,
P. frederiksbergensis strain 33 and
P. frederiksbergensis DSM 13022 are similar in
nature. Thus, for the first time two natural, genetically unmanipulated
P. frederiksbergensis strains show inverse senses of enantioselectivity
in the sulfoxidation of a variety of sulfide substrates. To
date, only one other example of such a phenomenon in enantioselectivity
has been reported (
8), namely, the sulfoxidation by a wild-type
Pseudomonas putida strain that contains naphthalene dioxygenase.
In this case, the
S-configured sulfoxide is preferably produced,
whereas the recombinant
Pseudomonas putida strain with toluene
dioxygenase selects the
R-configured enantiomer. It should also
be emphasized that these different enantioselectivities between
a native strain and a genetically modified
Pseudomonas strain
were not observed for all substrates used in the study (
8).
Of course, both
P. frederiksbergensis strains investigated here
must differ at the molecular level in order to express opposite
sulfoxide enantiomers; thus, presumably the amino acid sequence
for the enzyme responsible for the asymmetric sulfoxidation
with opposite enantioselectivities must be distinct. Consequently,
it should be a priority to determine the structure of the sulfoxidizing
enzymes and their amino acid sequences. We speculate that the
naphthalene dioxygenase and toluene dioxygenase enzymes described
for the
P. putida bacteria (
8) might be involved in the sulfoxidation
reactions of the
P. frederiksbergensis strains described herein.
Tentatively, the present results may be explained either in
terms of a mutation within the enzyme responsible for enantioselective
sulfoxidation or by the existence of two (or more) competing
enzyme activities. We anticipate that our novel findings will
initiate an intensive search for other such natural microorganisms.

Nucleotide sequence accession numbers.
The 16S rRNA partial sequence has been submitted to GenBank
with the accession number
AY817058.
P. frederiksbergensis strain
33 was deposited at DSMZ under the DSMZ number 16916.

ACKNOWLEDGMENTS
We are grateful to U. Vogel of the Institute of Hygiene and
Microbiology, University of Würzburg, Würzburg, Germany,
for the genotypical characterization (16S rRNA analysis) of
the isolated soil bacteria.
The Deutsche Forschungsgemeinschaft (SFB 347) and the Fonds der Chemischen Industrie are thanked for generous financial support.

FOOTNOTES
* Corresponding author. Mailing address: Institute of Food Chemistry, University of Würzburg, Am Hubland, D-97074 Würzburg, Germany. Phone: 49-(0)-931-8885480. Fax: 49-(0)-931-8885484. E-mail:
schreier{at}pzlc.uni-wuerzburg.de.


REFERENCES
1 - Adam, W., F. Heckel, C. R. Saha-Möller, M. Taupp, and P. Schreier. 2004. A highly enantioselective biocatalytic sulfoxidation by the topsoil bacterium Pseudomonas frederiksbergensis. Tetrahedron Asymmetry 15:983-985.[CrossRef]
2 - Adam, W., F. Heckel, C. R. Saha-Möller, and P. Schreier. 2002. Biocatalytic synthesis of optically active oxyfunctionalized building blocks with enzymes, chemoenzymes and microorganisms. J. Organomet. Chem. 661:17-29.[CrossRef]
3 - Adam, W., M. Lazarus, C. R. Saha-Möller, O. Weichold, U. Hoch, D. Häring, and P. Schreier. 1999. Biotransformations with peroxidases. Adv. Biochem. Eng. Biotech. 63:73-108.[Medline]
4 - Adam, W., C. Mock-Knoblauch, and C. R. Saha-Möller. 1999. Asymmetric synthesis with the enzyme Coprinus peroxidase: kinetic resolution of chiral hydroperoxides and enantioselective sulfoxidation. J. Org. Chem. 64:4834-4839.[CrossRef][Medline]
5 - Adam, W., M. N. Korb, K. J. Roschmann, and C. R. Saha-Möller. 1998. Titanium-catalyzed asymmetric sulfoxidation of alkyl aryl sulfides with optically active hydroperoxides. J. Org. Chem. 63:3423-3428.[CrossRef]
6 - Andersen, S. M., K. Johnsen, J. Sørensen, P. Nielsen, and C. S. Jacobsen. 2000. Pseudomonas frederiksbergensis sp. nov., isolated from soil at a coal gasification site. Int. J. Syst. Evol. Microbiol. 50:1957-1964.[Abstract]
7 - Atlas, R. M. 1993. Handbook of microbiological media, 2nd ed. CRC Press, Boca Raton, Fla.
8 - Boyd, R. B., N. D. Sharma, S. A. Haughey, M. A. Kennedy, B. T. McMurray, G. N. Sheldrake, C. C. R. Allen, H. Dalton, and K. Sproule. 1998. Toluene and naphthalene dioxygenase-calalysed sulfoxidation of alkyl aryl sulfides. J. Chem. Soc. Perkin Trans. I 12:1929-1933.
9 - Colonna, S., S. del Sordo, N. Gaggero, G. Carres, and P. Pasta. 2002. Enzyme-mediated catalytic asymmetric oxidations. Heteroatom Chem. 13:467-473.[CrossRef]
10 - Colonna, S., N. Gaggero, G. Carrea, and P. Pasta. 2001. Oxidation of sulfides, p. 227-235. In T. Katsuki (ed.), Asymmetric oxidation reactions. Oxford University Press, Oxford, United Kingdom.
11 - Colonna, S., N. Gaggero, L. Casella, G. Carrea, and P. Pasta. 1992. Chloroperoxidase and hydrogen peroxide: an efficient system for enzymic enantioselective sulfoxidations. Tetrahedron Asymmetry 3:95-106.
12 - Colonna, S., N. Gaggero, A. Manfredi, L. Casella, M. Gullotti, G. Carrea, and P. Pasta. 1990. Enantioselective oxidations of sulfides catalyzed by chloroperoxidase. Biochemistry 29:10465-10468.[CrossRef][Medline]
13 - Dworkin, M., and J. W. Forster. 1958. Experiments with some microorganisms which utilize ethane and hydrogen. J. Bacteriol. 75:592-603.[Free Full Text]
14 - Harmsen, D., and H. Karch. 2004. 16S rDNA for diagnosing pathogens: a living tree. ASM News 70:19-24.
15 - Harmsen, D., J. Rothgänger, C. Singer, J. Albert, and M. Frosch. 1999. Intuitive hypertext-based molecular identification of microorganisms. Lancet 353:291-298.[Medline]
16 - Holland, H. L. 1988. Chiral sulfoxidation by biotransformation of organic sulfides. Chem. Rev. 88:473-485.[CrossRef]
17 - Hudlick
, M. 1990. Oxidation in organic chemistry. American Chemical Society, Washington, D.C.
18 - Johnsen, K., S. Andersen, and C. S. Jacobsen. 1996. Phenotypic and genotypic characterization of phenanthrene-degrading fluorescent Pseudomonas biovars. Appl. Environ. Microbiol. 62:3818-3825.[Abstract]
19 - Kelly, D., C. Knowles, J. Mahdi, G. Jassem, I. Taylor, and M. Wright. 1996. The enantioselective oxidation of sulfides to sulfoxides with Acinetobacter sp. NCIMB 9871, Pseudomonas sp. NCIMB 9872, Xanthobacter autotrophicus DSM 431 (NCIMB 10811) and the black yeast NV-2. Tetrahedron Asymmetry 7:365-368.[CrossRef]
20 - Mata, E. G. 1996. Recent advances in the synthesis of sulfoxides from sulfides. Phosphorus 117:231-286.
21 - Meyer, J. M., V. A. Geoffroy, N. Baida, L. Gardan, D. Izard, P. Lemanceau, W. Achouak, and N. J. Palleroni. 2002. Siderophore typing, a powerful tool for the identification of fluorescent and nonfluorescent pseudomonads. Appl. Environ. Microbiol. 68:2745-2753.[Abstract/Free Full Text]
22 - Ohta, H., Y. Okamoto, and G. Tsuchihashi. 1985. Studies on enzymic oxidation of sulfides. Part II. Microbial oxidation of alkyl aryl sulfides to the corresponding optically active sulfoxides. Agric. Biol. Chem. 49:671-676.
23 - van Deurzen, M. P. J., I. J. Remkes, F. van Rantwijk, and R. A. Sheldon. 1997. Chloroperoxidase-catalyzed oxidations in t-butyl alcohol/water mixtures. J. Mol. Catal. A Chem. 117:329-337.[CrossRef]
24 - Woods, C. R., J. Versalovic, T. Koeuth, and J. R. Lupski. 1993. Whole-cell repetitive element sequence-based polymerase chain reaction allows rapid assessment of clonal relationships of bacterial isolates. J. Clin. Microbiol. 31:1927-1931.[Abstract/Free Full Text]
Applied and Environmental Microbiology, April 2005, p. 2199-2202, Vol. 71, No. 4
0099-2240/05/$08.00+0 doi:10.1128/AEM.71.4.2199-2202.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.