-
Vol. 170, No. 4JOURNAL OF BACTERIOLOGY, Apr. 1988, p.
1511-15180021-9193/88/041511-08$02.00/0Copyright © 1988, American
Society for Microbiology
Molecular Cloning and Expression of theEscherichia coli Dimethyl
Sulfoxide Reductase Operon
PETER T. BILOUS AND JOEL H. WEINER*Department of Biochemistry,
University ofAlberta, Edmonton, Alberta, Canada T6G 2H7
Received 10 September 1987/Accepted 30 December 1987
The dimethyl sulfoxide (DMSO) reductase operon coding for a
membrane-bound iron-sulfur, molyb-doenzyme, which functions as a
terminal reductase in Escherichia coli, has been isolated and
cloned from anE. coli gene bank. Two clones, MV12(pLC19-36) and
MV12(pLC43-43), overexpressed both DMSO andtrimethylamine N-oxide
(TMAO) reductase activites 13- to 15-fold compared with wild-type
cells. Amplificationwas highest in cells grown anaerobically on
fumarate, while cells grown on DMSO or TMAO displayed reducedlevels
of enzyme amplification. Growth on nitrate or aerobic growth
repressed expression of the enzyme. A6.5-kilobase-pair DNA
restriction endonuclease fragment was subcloned from pLC19-36 into
the vectorpBR322, yielding a recombinant DMSO reductase plasmid,
pDMS159. Two polypeptides were amplified andidentified on sodium
dodecyl sulfate-polyacrylamide gels of proteins from E. coli HB101
harboring pDMS159:a membrane-bound protein with molecular weight
82,600 and a soluble polypeptide with molecular weight23,600. Three
plasmid-encoded polypeptides with molecular weights of 87,500,
23,300, and 22,600 weredetected by in vivo
transcription/translation studies. The smallest subunit was poorly
defined and not detectableby Coomassie blue staining. The DMSO
reductase operon was localized to the 20.0-min position on the E.
colilinkage map.
Anaerobic respiration by Escherichia coli on fumarate andnitrate
is well established due to extensive biochemical andmolecular
biological characterization of the respective ter-minal reductases
(9, 13). However, alternate forms of anaer-obic respiration are
known. Recent studies have focused ontrimethylamine N-oxide (TMAO)
reduction, due to the wide-spread distribution of this compound in
the natural environ-ment, and the demonstration of anaerobic
respiration onTMAO by various bacteria (3, 24, 28). TMAO reduction
isassociated with the marine genera, nonsulfur
photosyntheticbacteria, and certain genera of intestinal bacteria
includingE. coli. (3). Studies with E. coli have demonstrated
thepresence of one constitutive and three or four inducibleforms of
TMAO reductase in the cell (21, 23). The majorinducible form of
TMAO reductase has been purified andcharacterized (30).Although
bacterial reduction of dimethyl sulfoxide
(DMSO) has been known for some time (1), only recentlyhas its
role in anaerobic respiration been determined (19, 25,31). Like
TMAO, DMSO is associated with marine environ-ments as a by-product
of phytoplankton activity (2) and isconsidered to be an
intermediate in the global sulfur cycle(16). We recently
demonstrated that E. coli is capable ofanaerobic respiration on
DMSO (5). Anaerobic growth of E.coli on DMSO, TMAO, methionine
sulfoxide, or fumarateresults in the induction of a membrane-bound
molyb-doenzyme catalyzing the reduction of DMSO to dimethylsulfide
(4).
Studies in other bacteria have suggested that TMAO andDMSO are
reduced by the same enzyme system (20, 25). Ourinital studies
suggested some similarities and differencesbetween DMSO and TMAO
reduction in E. coli (4). Todefine the role of DMSO reductase in
the anaerobic growthof E. coli, the structural genes for the enzyme
were clonedand characterized. Cells harboring the DMSO
reductase
* Corresponding author.
plasmid displayed amplified levels of DMSO, TMAO andmethionine
sulfoxide reductase activities. Studies with thepurified enzyme
indicate that DMSO reductase has a broadsubstrate specificity,
reducing various sulfoxides and N-oxides (29). The enzyme has been
designated DMSO reduc-tase due to the high affinity for this
substrate.
MATERIALS AND METHODS
Bacterial strains and plasmids. The E. coli strains andplasmids
used in this study are listed in Table 1. E. coliMV12 carries ColEl
hybrid plasmids prepared by Clarke andCarbon (8).
Preparation of colicin El. Colicin El was prepared from E.coli
W3110(ColE1) by the procedure of Schwartz and Hel-inski (22),
except that cells were disrupted by two passagesthrough a French
pressure cell (American Instrument Co.,Silver Spring, Md.) at 110
MPa. The crude lysate wascentrifuged at 150,000 x g for 1 h before
ammonium sulfateprecipitation of the soluble material according to
the pub-lished procedure. The final preparation was stored in 50
mMpotassium phosphate buffer (pH 7.5) containing 50% glyceroland
assayed as described previously (12).Growth of cells and
preparation of everted membrane
vesicles. For enzyme expression studies, cells were
routinelygrown in glycerol minimal medium (4) supplemented withthe
appropriate antibiotics (100 ,ug ampicillin or streptomy-cin
sulfate per ml), amino acids (0.003%), and terminalelectron
acceptor (nitrate, 100 mM; fumarate, 40 mM;TMAO, 100 mM; or DMSO,
70 mM). Cultures were grownfor 36 h at 37°C, harvested, and then
lysed by Frenchpressure cell treatment. Membranes were prepared
from thecrude lysate material as described previously (4).Enzyme
assay. Reductase activity was assayed by moni-
toring the substrate-dependent oxidation of reduced
benzylviologen at 570 nm (4). One unit of activity corresponds to
1,umol of benzyl viologen oxidized per min at 23°C. Specific
1511
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
1512 BILOUS AND WEINER
TABLE 1. Bacterial strains and plasmids
Strain or Description Sourceplasmid
E. coli strainHB101 F- hsdR hsdM pro leu lac gal thi recA rpsL
Lab collectionK38 HfrC (X) S. TaboraMV12 F+ recA Atrp thr leu thi
Lab collectionTG1 A(lac-pro) supE thi hsdDSIFl' traD36 proA+B+ la,
lacZAM15 W. ParanchychbW3110 F- X- Lab collection
PlasmidColEl ColElr, coding for colicin El Lab collectionpBR322
Apr Tcr Boehringer MannheimpDMS159 Apr dms+ This studypDMS201 Apr
dms, derivative of pDMS159 This studypDMS216 Apr Kmr dms+,
derivative of pMS159 This studypDMS219 Apr dms, derivative of
pDMSt1iS This studypDMS222 Aprdms+, pTZ18R derivative of pDMS2l6
This studypDMS229 Apr dms+, pTZ18R derivative qo pDMS216 This
studypGP1-2 Kmr clts857, coding for T7 RNA polymerase under X PL
control S. TaborpLC19-36 ColElrdms+ This studypLC43-43 ColElr dms+
This studypTZ18R Apr lacZ' Pharmacia
"Harvard Medical School, Boston, Mass.b University of Alberta,
Edmonton, Alberta, Canada.
activity is expressed as units of reductase activity per
mil-ligram of protein.
Screening of the Clarke and Carbon colony bank. Each ofthe 2,112
clones from the Clarke and Carbon colony bank (8)was growp aq
erpbically at 37°C for 36 h in screw-cap testtubes (13 by "#'mm)
containing 8.5 ml of complex medium(glucose, 0.1 Bacto-Peptone
[Difco Laboratories, Detroit,Mich.], 0.4%; yeast extract, 0.4%; 70
mM potassium phos-phate buffer,; pEH 6.8) supplemented with 40 mM
sodiumfumarate (pP 7 4hiamine (0.003%), and colicin El at 1U/ml.
Cultures" Werd mixed continuously during growth bygently rocking
horizontally on a platform shaker. Cells wereharvested at 4,400 x g
for 5 min (IEC clinical centrifuge,model CL), washed once with 5 ml
of 50 mM sodiumphosphate buffer, pH 6.8, and then suspended in 0.5
ml ofthe same buffer. A 50-,ul portion of each cell suspension
wasadded to individual wells on microtiter plates. Enzymeassays
were initiated by the rapid addition of 200 p.l of assaymixture (50
mM sodium phosphate buffer, pH 6.8, 0.5 mMdithiothreitol, 0.2 mM
benzyl viologen, 1.0 mM sodiumdithionite, 10 mM either DMSO or
TMAO). The rate ofoxidation of reduced benzyl viologen in each well
(purple tocolorless transition) was monitored visually, and the
approx-imate time for complete oxidation was recorded.
Preparation and analysis of plasmid DNA. Plasmid DNAwas isolated
from cells grown in M9CA medium by chlor-amphenicol amplification
and sodium dodecyl sulfate (SDS)lysis as described by Maniatis et
al. (17). The isolatedplasmid DNA was purified by equilibrium
centrifugation oncesium chloride-ethidium bromide gradients.
Isolation ofplasmid DNA on a smaller scale was performed by
thealkaline-SDS procedure of Birnboim and Doly (6).
Electrophoresis. SDS-polyacrylamide gel electrophoresiswas
performed on vertical slab gels of 12.5% (wt/vol)acrylamide-0.33%
bisacrylamide, with a stacking gel of 3%acrylamide-0.08%
bisacrylamide. The discontinuous SDSbuffer system of'Laemmli (14)
was used. Gels were stainedand destained as described previously
(15). Gels containing35S-labeled proteins were dried and
autoradiographed di-rectly without further treatment.
Protein determination. Protein was estimated by an
SDSmodification of the Lowry procedure (18), using
crystallinebovine serum albumin (Bio-Rad Laboratories,
Richmond,Calif.) as the protein standard.
Construction of recombinant pTZ18R plasmids. For in
vivopolypeptide expression studies, recombinant plasmidspDMS222 and
pDMS229 were constructed from pDMS159and pTZ18R (Pharmacia) as
follows. A Kmr cartridge (Gen-Block, Pharmacia) was ligated into
the unique EcoRI site ofpDMS159 to proviol an additional SalI site
for convenientisolation and subsequent ligation of the 6.5-kilobase
(kb)chromosomal insert. The resulting plasmid, pDMS2'16,
wasdigested with SalI to yield a 6.5-kb SalI-SalI fragment
ofchromosomal DNA, which was subsequently purified fromagarose gels
by electroelution (D-gel; KONTES, Vineland,N.J.). The fragment was
ligated into the SalI site of pTZ18Rand then used to transform TG1
host cells. Plasmid DNAwas isolated from transformed cells, and the
orientation ofthe insert was determined by restriction endonuclease
m'ap-ping. pDMS222 and pDMS229 contained the chromosomalinsert from
pDMS159 in opposite orientations with respectto the T7 RNA
polymerase promoter region of pTZ18R.
Labeling of plasmid-encoded polypeptides. E. coli K38(pGP1-2),
coding for T7 RNA polymerase under c1857, APLcontrol, was
transformed with either pDMS222 orpDMS229. Cell proteins were
labeled in these strains with[35S]methionine 'as outlined by Tabor
and Richardson (26)with minor modifications. Transformed cells were
grownovernight at 30°C in LB medium (17) containing kanamycin(40
,ug/ml) and ampicillin (100 ,ug/ml) and then diluted 1:50 infresh
LB medium. Cells were grown to an A600 of 0.5, and200-,ul samples
were removed and washed twice with 1.0 mlof M9 medium (17) before
suspension in 1.0 ml of M9medium plus amino acids (0.1%, minus
methionine andcysteine), thiamine, and antibiotics. Cells were
grown for 60min at 30°C and shifted to 42°C for 15 min, and then
rifampin(200 ,ug/ml) was added followed by a further 10-min
incuba-tion at 42°C. Cells were then shifted to 30°C for 20 min,
atwhich time 6 p.Ci of L-[35S]methionine (1,330 Ci/mmol) wasadded.
At the appropriate time intervals, samples were
J. BACTERIOL.
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
CLONING OF DMSO REDUCTASE 1513
removed, added to cold trichloroacetic acid (10% final), andthen
incubated on ice for 30 min. Samples were washedtwice with 10%
trichloroacetic acid, suspended in Laemmlisolubilization buffer
(14) modified to contain 0.2 M Trizmabase (Sigma Chemical Co., St.
Louis, Mo.), electrophoresedon SDS-polyacrylamide gels, and then
autoradiographed.For pulse-chase experiments, 1.0 ml of cells was
pulsed for1 min with 24 pLCi of [35S]methionine, followed by a
chasewith 0.01% methionine.
Reagents. All chemicals used in this study were of analyt-ical
grade and obtained commercially. L-[35S]methioninewas purchased
from Amersham Canada Ltd., Oakville,Ontario.
RESULTSIsolation of DMSO reductase plasmids. To determine
the
mechanism of DMSO reduction and its relationship to
otherwell-defined anaerobic respiration pathways, cloning of
theDMSO reductase operon was carried out. Our initial ap-proach to
cloning DMSO reductase was by the mutantcomplementation procedure,
using a group of mutants wepreviously characterized to be defective
in DMSO reductaseactivity (4). The mutants were complemented with
an E. coligene bank prepared by ligating HindIlI-digested
chromo-somal DNA into plasmid vectors pBR322 and pUC13. Themutants
fell into two complementation groups, and the DNAfragments
complementing each of the two mutant classeswere cloned and
characterized. Expression studies indicatedthat neither fragment
coded for the DMSO reductase struc-tural gene. A preliminary report
of these findings has beenpresented (P. T. Bilous, and J. H.
Weiner, Abstr. Annu.Meet. Am. Soc. Microbiol. 1987, K119, p. 222),
and work isin progress to characterize the gene products.
Since the mutagenesis approach did not result in theisolation of
the structural gene for DMSO reductase, theClarke and Carbon E.
coli gene bank (8) was screened forclones which expressed amplified
levels ofDMSO or TMAOreductase activity. Elevated expression was
expected inappropriate clones, due to the multicopy nature of the
ColElvector. Each of the approximately 2,100 clones
harboringrecombinant ColEl plasmids was grown anaerobically on
aglucose-peptone medium. The cells were harvested as de-scribed in
Materials and Methods and then used directly forassay of enzyme
activity by following the DMSO- or TMAO-dependent oxidation of
reduced benzyl viologen. Previousstudies have indicated that both
substrates have readyaccess to the enzyme in whole cells or crude
lysates (data
PLASMID
pLCI9- 36
RESTRICTION MAP
Sc CPB B BPA A C SE
not shown). Two clones were identified, E. coli MV12(pLC19-36)
and MV12(pLC43-43), both of which displayed atwo- to
fourfold-faster DMSO- and TMAO-dependent oxi-dation of reduced
benzyl viologen than the average E. coliMV12 clone. Air oxidation
of the reduced benzyl viologenwas at least twofold slower than the
average substrate-dependent oxidation reaction. No clones were
found by thisscreening procedure, which amplified DMSO or
TMAOreductase individually.
Restriction mapping and subcloning. The recombinantColEl
plasmids were isolated from clones MV12(pLC19-36)and
MV12(pLC43-43), and restriction endonuclease mapswere determined
relative to a unique EcoRI site (Fig. 1). Thetwo plasmids were
found to contain a similar chromosomalDNA fragment with an
overlapping region of approximately15 kb. To identify the DNA
region coding for DMSOreductase, various restriction fragments were
generated andsubcloned into vector pBR322, yielding plasmids
pDMS159,pDMS201, and pDMS219 (Fig. 1). As shown, only E. coliHB101
cells transformed with plasmid pDMS159 expressedamplified levels of
DMSO reductase activity comparable tothe levels seen with
MV12(pLC19-36) or MV12(pLC43-43).Chromosomal mapping of DMSO
reductase operon. The
precise location of the coding region for DMSO reductasewas
determined by comparison of the restriction endonucle-ase sites on
pLC19-36 to a BamHI-EcoRI-HindIII restrictionmap of the E. coli
chromosome (10). The location, kind, andnumber of endonuclease
sites on pLC19-36 agree perfectlyover a 14-kb region with the
restriction sites determined fora region of the 210-kb NotI
chromosomal DNA fragment(Fig. 2). The NotI fragment was shown by
restriction sitecomparison (D. Daniels and F. Blattner, unpublished
data)to contain the rpsA and ompF genes which are located on
thelinkage map at 20.5 and 20.8 min, respectively. DNA se-quencing
analysis (P. T. Bilous, S. T. Cole, W. F. Andersonand J. H. Weiner,
manuscript in preparation) has demon-strated the presence of three
open reading frames associatedwith DMSO reductase, organized as an
operon. The DMSOreductase operon dms was localized to the region of
pLC19-36 shown in Fig. 2. The dms operon is therefore situated
at20.0 min on the E. coli linkage map.Growth and enzyme expression.
E. coli MV12(pLC19-36)
and HB101(pDMS159) cells were grown anaerobically
onglycerol-fumarate medium, and the enzyme activities in
themembranes and soluble fractions of the cells were deter-mined.
We have previously shown that growth on fumarateresults in the
expression of DMSO reductase at levels
DMSO REDUCTASE(Specific Activity, U/mg)
9.9Pu Sm
pLC43-43
pDMS159
pDMS201
pDMS219
Sc cPB 8 BPA A C SE
CPS B BPA AC S A... . - . . ' ' J1.
Pu Sm
c
CPS BP A A C S A cMI I --+ .-
PB B BP E P_ _ _ II_ ----- ------
6.4
10.5
.8
1.3i_ kb
FIG. 1. Partial restriction endonuclease maps of chromosomal DNA
from various plasmids containing the DMSO reductase gene
andactivities in membranes prepared from cells harboring each of
the plasmids. DMSO reductase specific activities were determined in
themembrane fractions ofE. coli MV12 (for ColEl plasmids) and E.
coli HB101 (for pBR322 plasmids) grown anaerobically on
Glycerol-fumaratemedium. A, AvaI; B, BamHI; C, ClaI; E, EcoRI; P,
PstI; Pu, PvuII; S, Sall; Sc, SacII; Sm, SmaI. Vectors: ColEl (-);
pBR322 (---).
VOL. 170, 1988
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
1514 BILOUS AND WEINER
E cOOiGENOME
20.0' 20.5S 2Q8'
N dnw rp*A ompF
iH 6lV B E H B BH H
pLC19-36H11 H A IB
NHH
A A H
kbH
N 10kbAI
IkbH
FIG. 2. Location of DMSO reductase operon on the E. coli
chromosome. The DMSO reductase operon dms was localized on the E.
colichromosome by comparison of the BamHI (B), EcoRI (E), and
HindIll (H) endonuclease restriction sites on plasmid pLC19-36 to
the kind,number, and position of these restriction sites on a
210-kb Notl (N) fragment of the E. coli chromosome. The positions
of several genes onthe Notl fragment are indicated. The restriction
map of the Notl chromosomal fragment and data on the location of
the genes shown werekindly provided by D. Daniels (personal
communication). The operon coding for DMSO reductase, dms, was
localized on the pLC19-36chromosomal fragment from DNA sequence
data. (Bilous et al., in preparation).
comparable to or better than those obtained with DMSO aselectron
acceptor in the growth medium (4; unpublishedobservations). The
results of a typical expression study areshown in Table 2.The E.
coli MV12(pLC19-36) clone isolated from the
Clarke and Carbon colony bank displayed a 13- to
15-foldamplification of the membrane-bound DMSO and TMAOreductase
activity when compared with a typical E. coliMV12 clone harboring a
random DNA insert. Approxi-mately 90% of the DMSO or TMAO activity
in these cellswas associated with the membrane fraction of the
cell, inagreement with the nitrate and fumarate reductase
activities.Interestingly, the ratio of TMAO/DMSO reductase
activitywas constant at about 4:1 both in wild type and in E.
coliMV12(pLC19-36). This suggested that one enzyme wasresponsible
for both TMAO and DMSO reductase activitiesunder these growth
conditions. A similar amplification anddistribution of activity was
observed with E. coli MV12(pLC43-43) (data not shown). Neither
clone overexpressedfumarate or nitrate reductase activity.
E. coli HB101 cells harboring the dms plasmid pDMS159expressed
elevated levels of TMAO and DMSO reductaseactivity eight- to
ninefold greater than those observed for E.coli HB101 (Table 2).
Approximately 70 to 80% of theamplified activity was membrane
associated, comparable to
that observed with E. coli MV12(pLC19-36) and in agree-ment with
previous observations with wild-type cells (4).However, the level
of enzyme expression with pDMS159 inE. coli HB101 was not as high
as observed with pLC19-36 inE. coli MV12. The results may reflect
strain differences.Previous growth studies with E. coli HB101 have
shown thatmethionine sulfoxide, an analog of DMSO, could
substitutefor DMSO in both the growth medium and the benzylviologen
assay (4). Amplification of activity (as shown byDMSO and TMAO) was
also observed with methioninesulfoxide as substrate (Table 2). The
results suggest that oneenzyme is responsible for all three
activities.
Identification of DMSO reductase gene product. The pro-tein
electrophoretic pattern of membranes and soluble frac-tions from E.
coli HB101 harboring pDMS159 is shown inFig. 3. An amplified
protein band with molecular weight of82,600 + 2,000 (mean standard
deviation, eight determi-nations) was evident in the membrane
fraction of E. coliHB1O1(pDMS159) (lane 2) when compared with the
wild-type control (lane 1). A polypeptide with identical
molecularweight was present in the soluble fraction of the cell
(lane 5),but the intensity of this band varied from preparation
topreparation. It is probably identical to the
membrane-boundpolypeptide of the same size. A membrane-bound
polypep-tide with molecular weight of approximately 72,000 is
evi-
TABLE 2. Expression and localization of reductase activities in
E. coli MV12 and HB101 harboring DMSO reductase plasmidsa
Sp act (U/mg)E. coli strain Cell fraction
Nitrate Fumarate TMAO DMS0 MetSO
MV12(pLC22-12)b Membranes 1.3 (>%)c 6.3 (98) 2.6 (90) 0.75
(93) dSoluble
-
CLONING OF DMSO REDUCTASE 1515
1 2 3 4 5
-.
AN* ~ *s
_~ ~
FIG. 3. SDS-polyacrylamide gel electrophoresis of membrane-bound
and soluble proteins from E. ccli HB101 and HB101(pDMS159) grown on
Glycerol-fumarate minimal medium. Cellswere grown for 36 h on
glycerol-fumarate minimal medium, and
membranes and soluble fractions were prepared as outlined in
Materials and Methods. Samples (75 p.g of protein per lane) were
run
on a 12.5% SDS-polyacrylamide gel and then stained for
proteinwith Coomassie brilliant blue. Lane 1, E. coli HB101
membranefraction; lane 2, E. coli HB101(pDMS159) membrane fraction;
lane3, protein molecular weight standards; lane 4, E. ccli HB101
solublefraction; lane 5, E. ccli HB101(pDMS159) soluble fraction.
Themolecular weights of protein standards shown in lane 3 are
as
follows: phosphorylase b, 97,400; bovine albumin, 66,000;
eggalbumin, 45.000; carbonic anhydrase, 29,000; trypsin
inhibitor,
20,100. Arrowheads mark amplified protein bands in cells
harboringthe DM80 reductase plasmid pDMS159.
dent in lane 2 and is believed to be a proteolytic fragment
ofthe 82,600-molecular-weight polypeptide as its intensity
in-creased during storage of the samples. An additional ampli-fied
protein band with molecular weight of 23,600 ± 800
(mean ± standard deviation, eight determinations) was
present in the soluble fraction of E. coli HB101(pDMS159)(lane
5). DNA sequence analysis of pDM.159 has indicated
the presence of three open reading frames organized in an
operon, suggesting a three-subunit structure for DM80
reductase with calculated molecular weights of 87,350,
23,070, and 30,789 (Bilous et al., in preparation). The
23,600-
molecular-weight polypeptide present in the soluble fraction
of the cell during isolation is probably equivalent to the
23,070-molecular-weight polypeptide coded by the dms op-eron and
is loosely associated with the 82,600-molecular-
weight membrane-bound subunit. The two subunits have
been shown to copurify during purification of DM80 reduc-
tase (29). It would appear that the 30,789-molecular-weight
polypeptide, which has a very hydrophobic amino acid
composition, is not detected by Coomassie blue staining of
crude cellular preparations.Effect of growth conditions on DM80
reductase activity. It
was previously shown that optimal levels of DM80 reduc-
tase were induced by anaerobic growth on fumarate (4).
Theglycerol-fumarate medium was the preferred medium forgrowth
studies because the products of both DMSO andTMAO reduction are
volatile and have unpleasant odors.However, it was of interest to
determine the activity levelsof DMSO reductase in E. coli
HB1O1(pDMS159) whengrown with various terminal electron acceptors.
Cells weregrown for 48 h (early stationary phase) on nitrate,
fumarate,TMAO, and DMSO minimal media. Crude membranes wereprepared
from E. coli HB1O1(pDMS159) and then assayedfor reductase activity
levels. The results are shown in Table3. The presence of the DMSO
reductase plasmid resulted inmaximal DMSO and TMAO reductase
activities only whengrown anaerobically on fumarate. Surprisingly,
unelevatedlevels of activity were obtained when cells were grown
onDMSO. Growth on TMAO resulted in a similar low level ofexpression
when compared with growth on fumarate, but afive- to
sevenfold-higher level of expression when comparedwith E. coli
HB101 grown on glycerol-TMAO. Anaerobicgrowth on nitrate or aerobic
growth (data not shown) com-pletely repressed the synthesis of
DMSO, TMAO, andfumarate reductase activities. A similar pattern of
enzymeexpression was demonstrated with E. coli MV12(pLC19-36)grown
under the conditions described above (unpublishedobservations). In
agreement with these results, examinationof the protein
electrophoretic pattern on SDS-polyacryl-amide gels revealed an
amplified protein band with molecu-lar weight 82,600 present only
in the membranes of glycerol-fumarate-grown cells (data not
shown).
Identification of plasmid-encoded polypeptides. To identifyall
gene products which were plasmid encoded, an in
vivotranscription/translation expression study was
performed.Recombinant plasmids pDMS222 and pDMS229 were
con-structed as described in Materials and Methods and asshown in
Fig. 4. Plasmids pDMS222 and pDMS229 containthe chromosomal insert
in opposite orientations with respectto the T7 promoter site of
pTZ18R. E. coli K38, containingplasmid pGP1-2 which expresses a T7
RNA polymeraseunder temperature control, was transformed with
pDMS222or pDMS229. Cells under the appropriate expression
condi-tions were pulsed for 3 min with [35S]methionine.
Totalcellular protein was precipitated with trichloroacetic
acid,separated by SDS-polyacrylamide gel electrophoresis, andthen
visualized by autoradiography. The resulting autoradio-gram is
shown in Fig. SA.
TABLE 3. Reductase activity levels in membranes of E. coliHB101
and HB101(pDMS159) grown on glycerol minimal medium
with various terminal electron acceptorsa
Growth Sp act (U/mg)Snmedium Nitrate Fumarate TMAO DMSO
HB101 GLY-nitrate 26.3 NDb ND NDGLY-FUM 4.5 3.7 3.2 1.4GLY-TMAO
4.1 1.0 1.0 0.46GLY-DMSO 5.0 3.1 6.8 1.1
HB101(pDMS159) GLY-nitrate 14.6 ND ND NDGLY-FUM 0.6 3.2 29.0
7.8GLY-TMAO 0.9 1.3 7.0 2.5GLY-DMSO 0.7 1.4 5.6 1.9
Cells were grown anaerobically for 48 h at 37°C in 250 ml of
glycerol(GLY) minimal medium with nitrate, fumarate (FUM), TMAO, or
DMSO aselectron acceptor. Membranes were prepared and reductase
activities wereassayed as described in Materials and Methods.
b ND, Not detected.
VOL. 170, 1988
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
1516 BILOUS AND WEINER
preparation of DMSO reductase on SDS-polyacrylamidegels
(29).
DISCUSSION
In this paper we report on the isolation and cloning of
ananaerobically induced, membrane-bound, terminal reduc-tase of E.
coli (4, 5). The isolation of the operon wasfacilitated by the use
of a whole-cell enzyme assay to screena gene bank of potential
clones expressing higher levels ofDMSO or TMAO reductase activity.
Two clones wereidentified which amplified both DMSO and TMAO
reductaseactivity. No clone was discovered which amplified only
oneof the two substrates tested. The results suggest that oneenzyme
is responsible for the reduction of both substrates.We have
designated the cloned enzyme as DMSO reductasefor the following
reasons. (i) The enzyme is geneticallydistinct from the genetic
loci reported for TMAO reductase.(ii) The purified enzyme displays
a higher affinity for DMSOthan methionine sulfoxide, TMAO, or other
N-oxide com-pounds (29). (iii) Whole cells challenged with both
substratesreduced DMSO at a faster rate than TMAO (P. T. Bilous,B.
D. Sykes, and J. H. Weiner, unpublished observations).
Analysis of protein electrophoretic patterns on
SDS-poly-acrylamide gel electrophoresis suggested that a
membrane-bound polypeptide with molecular weight of 82,600
waslikely associated with DMSO reductase activity. In
vivopolypeptide expression studies identified three
polypeptideswith molecular weights of 87,500, 23,300, and 22,600
asso-
AB12 3 4 5
FIG. 4. Construction of recombinant pTZ18R plasmids carryingthe
DMSO reductase gene. Recombinant plasmids pDMS222 andpDMS229
containing the chromosomal insert from pDMS159 in bothorientations
with respect to the T7 promoter region were con-structed from DMSO
reductase plasmid pDMS159 and pTZ18R asdescribed in Materials and
Methods. The direction of transcriptionof the dms operon and its
location on the chromosomal DNA insertwere determined from DNA
sequencing data (Bilous et al., inpreparation). Apr, Ampicillin
resistance; dms, DMSO reductaseoperon; E, EcoRI; H, HindIII; Kmr,
kanamycin resistance; T7, T7RNA polymerase promoter; S, SaIl.
Two polypeptides with molecular weights of 87,500 ± 900(mean
standard deviation, three determinations) and23,300 300 (three
determinations) were clearly expressedby pDMS222 (lane 3), but not
by pDMS229 (lane 4) or in theappropriate controls (lanes 1 and 2).
An additional polypep-tide with molecular weight of 22,600 ± 400
(three determi-nations) was evident as a fuzzy band in lane 3.
These resultsestablish the direction of transcription of the DMSO
reduc-tase operon. To determine whether proteolytic activity
dur-ing the 3-min pulse experiment was perhaps responsible forthe
ill-defined 22,600-molecular-weight protein band, apulse-chase
experiment was performed with a 1-min pulsefollowed by 2-, 5-, and
10-min chases with cold methionine.The results shown in Fig. SB
display an identical pattern andrelative intensity to that seen
with the pulse experiment (Fig.5A, lane 3). These results
demonstrate the absence of anyfurther modification of the
polypeptides during the course ofthese experiments and suggest a
three-subunit structure forDMSO reductase. A polypeptide pattern
identical to thatshown in Fig. 5A, lane 3, has been observed with a
purified
FIG. 5. In vivo expression of plasmid-encoded polypeptides byT7
RNA polymerase-promoter expression system. For pulse-la-beling
experiments (A), E. coli K38(pGP1-2) cells alone (lane 1)
orcontaining pTZ18R (lane 2), pDMS222 (lane 3), or pDMS229 (lane
4)were pulse-labeled for 30 min with [35S]methionine as described
inMaterials and Methods. Trichloroacetic acid-precipitated
proteinswere electrophoresed on 12.5% SDS-polyacrylamide gels,
fixed,dried, and then autoradiographed. For pulse-chase studies
(B), E.coli K38(pGP1-2) cells transformed with pDMS222 were pulsed
for1 min with [35S]methionine (lane 2) and then chased with
unlabeledmethionine for 2 (lane 3), 5 (lane 4), and 10 (lane 5)
min. Molecularweight standards shown in lanes 1 (B) and 5 (A) are
identical to thosedescribed in the legend to Fig. 3, with the
addition of ot-lactalbumin,14,200 molecular weight.
ES Km E
-.. 1.5kb
J. BACTERIOL.
!!TNWWWW
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
CLONING OF DMSO REDUCTASE 1517
ciated with the cloned DMSO reductase, which agrees withthe
subunits associated with purified DMSO reductase (29).A consistent
difference in molecular weight of the largersubunit was noted
between growth and expression studiesand the values obtained from
in vivo labeling experiments.The possibility of posttranslational
modification of this sub-unit is under investigation.
It was perhaps fortuitous that the Clarke and Carboncolony bank
was screened in a complex medium supple-mented with fumarate.
Growth studies with E. coliHB101(pDMS159) or MV12(pLC19-36) have
shown thatgrowth in the presence of DMSO or TMAO result in
nearwild-type levels of DMSO reductase activity. In agreementwith
the activity results, the 82,600-molecular-weight sub-unit is
barely visible in the membrane fraction of DMSO- andTMAO-grown
cells. Due to the different generation timesassociated with growth
on the various terminal electronacceptors, cultures were grown to
stationary phase. Allcultures were harvested at the same time and
treated in anidentical fashion. It is therefore unlikely that
proteolyticdigestion could be responsible for the observed results.
It ispossible that the end products of reduction are toxic-to
thecells and a repression mechanism exists when cells aregrown on
Glycerol-DMSO or Glycerol-TMAO.
Studies of TMAO and DMSO reduction in Rhodobactercapsulatus and
Proteus vulgaris have concluded that onlyone enzyme is responsible
for both activities (20, 25).However, in E. coli multiple forms
ofTMAO reductase havebeen reported (23). The major inducible form
has beenpurified and characterized (30). Genetic studies have
local-ized two inducible E. coli TMAO reductase genes to the
28.3(21)- and 77- to 84 (27)-min region of the chromosome. It isnot
clear if the major inducible form of the enzyme which hasbeen
purified by Yamamoto et al. (30) is coded by either ofthese
genes.
In the present study, we have cloned a membrane-boundterminal
reductase from E. coli which is situated at 20.0 minon the linkage
map. The enzyme is induced by anaerobiosis,but does not require the
presence of any added sulfoxide orN-oxide substrates for
expression. The enzyme is thusanaerobically constitutive. The
enzyme can use DMSO,TMAO, and methionine sulfoxide as substrates,
but is ge-netically distinct from the reported TMAO reductases.
Twomethionine sulfoxide reductases have been identified in E.coli,
one reducing free methionine sulfoxide (11) and theother reducing
protein-bound residues (7). Both have beenpurified and have
molecular weights of 21,000 and 18,000 to20,000, respectively.
Although the cloned DMSO reductaseis able to reduce methionine
sulfoxide, it appears to bedistinct from the reported methionine
sulfoxide reductasesbased on physical properties.
ACKNOWLEDGMENTS
We thank Stan Tabor for stain K38 and plasmid pGP1-2 andNancy
Chung for construction of pDMS201. We are especiallyindebted to
Donna Daniels and Frederick Blattner for providing theE. coli
restriction map data.
This work was supported by a grant (MT5838) from the
MedicalResearch Council of Canada. P.T.B. is a postdoctoral fellow
of theAlberta Heritage Foundation for Medical Research.
LITERATURE CITED
1. Ando, H., M. Kumagai, T. Karashimada, and H. lida.
1957.Diagnostic use of dimethylsulfoxide reduction test within
Ente-robacterioaceae. Jpn. J. Microbiol. 1:335-338.
2. Andreae, M. 0. 1980. Dimethylsulfoxide in marine and
fresh-
waters, Limnol. Oceanogr. 25:1054-1063.3. Barrett, E. L., and H.
S. Kwan. 1985. Bacterial reduction of
trimethylamine oxide. Annu. Rev. Microbiol. 39:131-149.4.
Bilous, P. T., and J. H. Weiner. 1985. Dimethyl sulfoxide
reductase activity by anaerobically grown Escherichia coliHB101.
J. Bacteriol. 162:1151-1155.
5. Bilous, P. T., and J. H. Weiner. 1985. Proton
translocationcoupled to dimethyl sulfoxide reduction in
anaerobically grownEscherichia coli HB101. J. Bacteriol.
163:369-375.
6. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline
extractionprocedure for screening recombinant plasmid DNA.
NucleicAcids Res. 7:1513-1523.
7. Brot, N., L. Weissbach, J. Werth, and H. Weissbach.
1981.Enzymatic reduction of protein-bound methionine
sulfoxide.Proc. Natl. Acad. Sci. USA 78:2155-2158.
8. Clarke, L., and J. Carbon. 1976. A colony bank
containingsynthetic Col El hybrid plasmids representative of the
entire E.coli genome. Cell 9:91-99.
9. Cole, S. T., C. Condon, B. D. Lemire, and J. H. Weiner.
1985.Molecular biology, biochemistry and bioenergetics of
fumaratereductase, a complex membrane-bound iron-sulfur
flavoenzymeof Escherichia coli. Biochim. Biophys. Acta
811:381-403.
10. Daniels, D. L., and F. R. Blattner. 1987. Mapping using
geneencyclopaedias. Nature (London) 325:831-832.
11. Ejiri, S.-I., H. Weissbach, and N. Brot. 1980. The
purification ofmethionine sulfoxide reductase from Escherichia
coli. Anal.Biochem. 102:393-398.
12. Herschman, H. R., and D. R. Helinski. 1967. Purification
andcharacterization of colicin E2 and colicin E3. J. Biol. Chem.
242:5360-5368.
13. Ingledew, W. J., and R. K. Poole. 1984. The respiratory
chainsof Escherichia coli. Microbiol. Rev. 48:222-271.
14. Laemmli, U. K. 1970. Cleavage of structural proteins during
theassembly of the head of bacteriophage T4. Nature
(London)227:680-685.
15. Lemire, B. D., J. J. Robinson, and J. H. Weiner. 1982.
Identi-fication of membrane anchor polypeptides of Escherichia
colifumarate reductase. J. Bacteriol. 152:1126-1131.
16. Lovelock, J. E., R. J. Maggs, and R. A. Rasmussen.
1972.Atmospheric dimethyl sulphide and the natural sulphur
cycle.Nature (London) 237:452-453.
17. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982.
Molecularcloning: a laboratory manual. Cold Spring Harbor
Laboratory,Cold Spring Harbor, N.Y.
18. Markwell, M. A. K., S. M. Haas, L. L. Bieber, and N.
E.Tolbert. 1978. A modification of the Lowry procedure to sim-plify
protein determination in membrane and lipoprotein sam-ples. Anal.
Biochem. 87:206-210.
19. McEwan, A. G., S. J. Ferguson, and J. B. Jackson.
1983.Electron flow to dimethylsulphoxide or
timethylamine-N-oxidegenerates a membrane potential in
Rhodopseudomonas capsu-lata. Arch. Microbiol. 136:300-305.
20. McEwan, A. G., H. G. Wetzstein, 0. Meyer, J. B. Jackson,
andS. J. Ferguson. 1987. The periplasmic nitrate reductase
ofRhodobacter capsulatus; purification, characterisation and
dis-tinction from a single reductase for
trimethylamine-N-oxide,dimethylsulphoxide and chlorate. Arch.
Microbiol. 147:340-345.
21. Pascal, M.-C., J.-F. Burini, and M. Chippaux. 1984.
Regulationof the trimethylamine N-oxide (TMAO) reductase in
Esche-richia coli: analysis of tor::Mudl operon fusion. Mol.
Gen.Genet. 195:351-355.
22. Schwartz, S. A., and D. R. Helinski. 1971. Purification
andcharacterization of colicin E1. J. Biol. Chem.
246:6318-6327.
23. Shimokawa, O., and M. Ishimoto. 1979. Purification and
someproperties of inducible tertiary amine N-oxide reductase
fromEscherichia coli. J. Biochem. 86:1709-1717.
24. Str0m, A. R., J. A. Olafsen, and H. Larsen. 1979.
Trimethyla-mine oxide: a terminal electron acceptor in anaerobic
respira-tion of bacteria. J. Gen. Microbiol. 112:315-320.
25. Styrvold, 0. B., and A. R. Str0m. 1984. Dimethylsulphoxide
andtrimethylamine oxide respiration of Proteus vulgaris.
Evidencefor a common terminal reductase system. Arch.
Microbiol.140:74-78.
VOL. 170, 1988
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/
-
1518 BILOUS AND WEINER
26. Tabor, S., and C. C. Richardson. 1985. A bacteriophage T7RNA
polymerase/promoter system for controlled exclusiveexpression of
specific genes. Proc. Natl. Acad. Sci. USA 82:1074-1078.
27. Takagi, M., and M. Ishimoto. 1983. Escherichia coli
mutantsdefective in trimethylamine N-oxide reductase. FEMS
Micro-biol. Lett. 17:247-250.
28. Takagi, M., and T. Tsuchiya, and M. Ishimoto. 1981.
Protontranslocation coupled to trimethylamine N-oxide reduction
inanaerobically grown Escherichia coli. J. Bacteriol.
148:762-768.
29. Weiner, J. H., D. P. MacIsaac, R. E. Bishop, and P. T.
Bilous.1988. Purification and properties of Escherichia coli
dimethylsulfoxide, an iron-sulfur molybdoenzyme with broad
substratespecificity. J. Bacteriol. 170:1505-1510.
30. Yamamoto, I., N. Okubo, and M. Ishimoto. 1986.
Furthercharacterization of trimethylamine N-oxide reductase
fromEscherichia coli, a molybdoprotein. J. Biochem.
99:1773-1779.
31. Zinder, S. H., and T. D. Brock. 1978. Dimethyl sulfoxide as
anelectron acceptor for anaerobic growth. Arch. Microbiol.
116:35-40.
J. BACTERIOL.
on April 1, 2021 by guest
http://jb.asm.org/
Dow
nloaded from
http://jb.asm.org/