Bacterial dissimilatory metal Keywords kinetic enzyme parameters; metal reduction; outer membrane cytochromes c OmcA and OmcB; Shewanella oneidensis MR-1; terminal reductases Corresponde
Trang 1metal reduction by Shewanella oneidensis MR-1 on the
outer membrane cytochromes c OmcA and OmcB
Jimmy Borloo*, Bjorn Vergauwen*, Lina De Smet, Ann Brige´, Bart Motte, Bart Devreese
and Jozef Van Beeumen
Laboratory for Protein Biochemistry and Protein Engineering, Ghent University, Belgium
Shewanella oneidensis MR-1 is a Gram-negative
c-pro-teobacterium with an extremely versatile anaerobic
res-piratory metabolism Under anaerobic conditions, this
organism reduces a variety of organic and inorganic
substrates, including fumarate, nitrate, trimethylamine
N-oxide, dimethylsulfoxide, sulfite and thiosulfate, as well as various polyvalent metal ions and radio-nuclides, including iron(III), manganese(IV), chro-mium(VI), vanadium(V), selenium(VI), uranium(VI), and tellurium(VI) [1–7] Bacterial dissimilatory metal
Keywords
kinetic enzyme parameters; metal reduction;
outer membrane cytochromes c OmcA and
OmcB; Shewanella oneidensis MR-1;
terminal reductases
Correspondence
J Borloo, Laboratory for Protein
Biochemistry and Protein Engineering,
Ghent University, K.L Ledeganckstraat 35,
B-9000 Ghent, Belgium
Fax: +32 9 264 52 73
Tel: +32 9 264 51 26
E-mail: jimmy.borloo@ugent.be
Website: http://www.eiwitbiochemie.ugent.
be/index.html
*These authors contributed equally to this
work
(Received 28 April 2007, revised 25 May
2007, accepted 30 May 2007)
doi:10.1111/j.1742-4658.2007.05907.x
The Gram-negative bacterium Shewanella oneidensis MR-1 shows a remarkably versatile anaerobic respiratory metabolism One of its hall-marks is its ability to grow and survive through the reduction of metallic compounds Among other proteins, outer membrane decaheme cyto-chromes c OmcA and OmcB have been identified as key players in metal reduction In fact, both of these cytochromes have been proposed to be ter-minal Fe(III) and Mn(IV) reductases, although their role in the reduction
of other metals is less well understood To obtain more insight into this,
we constructed and analyzed omcA, omcB and omcA⁄ omcB insertion mutants of S oneidensis MR-1 Anaerobic growth on Fe(III), V(V), Se(VI) and U(VI) revealed a requirement for both OmcA and OmcB in Fe(III) reduction, a redundant function in V(V) reduction, and no apparent involvement in Se(VI) and U(VI) reduction Growth of the omcB– mutant
on Fe(III) was more affected than growth of the omcA– mutant, suggesting OmcB to be the principal Fe(III) reductase This result was corroborated through the examination of whole cell kinetics of OmcA- and OmcB-dependent Fe(III)-nitrilotriacetic acid reduction, showing that OmcB is
11.5 and 6.3 times faster than OmcA at saturating and low nonsaturat-ing concentrations of Fe(III)-nitrilotriacetic acid, respectively, whereas the omcA– omcB– double mutant was devoid of Fe(III)-nitrilotriacetic acid reduction activity These experiments reveal, for the first time, that OmcA and OmcB are the sole terminal Fe(III) reductases present in S oneidensis MR-1 Kinetic inhibition experiments further revealed vanadate (V2O5) to
be a competitive and mixed-type inhibitor of OmcA and OmcB, respect-ively, showing similar affinities relative to Fe(III)-nitrilotriacetic acid Nei-ther sodium selenate nor uranyl acetate were found to inhibit OmcA- and OmcB-dependent Fe(III)-nitrilotriacetic acid reduction Taken together with our growth experiments, this suggests that proteins other than OmcA and OmcB play key roles in anaerobic Se(VI) and U(VI) respiration
Abbreviation
FR, fumarate reductase.
Trang 2reduction is known to account for the majority of the
valence transitions of Fe(III) to Fe(II) in anoxic,
non-sulfidogenic and low-temperature environments
Fur-thermore, microbial metal reduction represents a
potential strategy for the in situ immobilization and
containment of contaminant metals and radionuclides
in aqueous waste streams and subsurface
environ-ments, as some of these metals precipitate upon
reduc-tion [6,8]
Although the importance of bacterial dissimilatory
metal reduction in controlling the fate and transport
of metals and their potential for remediation purposes
are well recognized, the terminal reductases involved
are not yet identified, and nor are they sufficiently
characterized, as kinetic information on metal
reduc-tion is scarce The electron transport chain involved in
the reduction of either Fe(III) or Mn(IV) in MR-1 is
thought to be composed of cytochromes and a
qui-none, located in both the cytoplasmic membrane
(CymA and menaquinone) and the outer membrane
(OmcB, and a partial role for OmcA) [4,9–11] The
21 kDa tetraheme cytochrome c CymA (SO_4591) and
menaquinone are believed to be common central
com-ponents in the electron transport chain that branch to
several reductases downstream, as cymA– or
menaqui-none-deficient strains lose their ability to grow
anaero-bically on Fe(III), Mn(IV), V(V), nitrate, fumarate and
dimethylsulfoxide [9,10,12] OmcA (SO_1779) and
OmcB (SO_1778) are outer membrane decaheme
lipo-protein cytochromes c [13,14] that are specifically
involved in metal reduction, although distinct
func-tions have been proposed OmcB-negative MR-1
mutants are heavily affected in either Fe(III), Mn(IV)
or V(V) reduction, whereas the absence of OmcA
results in metal reduction rates that are 55% and 62%
of those of the MR-1 parent strain for Mn(IV) and
V(V), respectively [10] Purified and dithionite-reduced
preparations of both outer membrane proteins were
recently shown to directly transfer electrons to chelated
Fe(III) at comparable rates (kcat values ranging
between 1.5 and 4.1 s)1), whereas only reduced OmcB
was shown to be oxidized by uranyl acetate
(kcat< 0.01 s)1) [15] Taken together, OmcA and
OmcB function as metal reductases in MR-1, albeit
apparently behaving kinetically differently and
display-ing a rather undefined metal specificity
To address these latter issues, we constructed omcA,
omcB and omcA⁄ omcB insertion mutants of MR-1,
and analyzed them in terms of dissimilatory reduction
of a variety of metals, i.e Fe(III), V(V), U(VI), and
Se(VI) A ‘whole cell’ kinetics approach was used to
determine the kinetic parameters for OmcA- and
OmcB-dependent chelated Fe(III) reduction, which are
shown to corroborate the results of inhibition and liquid growth experiments These results identify OmcA and OmcB, for the first time to our knowledge,
as the sole terminal Fe(III) reductases, and additionally provide novel insights into the dependence of dissimila-tory metal reduction by MR-1 on OmcA and OmcB
Results
Growth analyses of anaerobically metal-respiring omcA–, omcB–and omcA–omcB–MR-1R mutants relative to their MR-1R parent
To study the substrate specificities of the outer membrane decaheme cytochromes OmcA and OmcB
in the process of dissimilatory metal reduction, omcA–, omcB– and omcA–omcB– MR-1R mutants were con-structed and evaluated in liquid broth growth experi-ments with lactate as electron donor and either Fe-nitrilotriacetic acid, Fe-citrate, V2O5, Na2SeO4 or
UO2(CH3COO)2.2H2O as the terminal electron accep-tor Complete growth curves were recorded for each experiment; those of MR-1R grown on the different metals are shown in Fig 1B, whereas the increases in density at day 3 of MR-1R and of all mutants are summarized in Fig 1A For chelated forms of Fe(III) and for V2O5, culture turbidities gradually decreased
in the order MR-1R > omcA–>> omcB–> omcA– omcB–, with the greatest effect being caused by the omcB disruption OmcA and OmcB are collectively essential for chelated Fe(III) dissimilatory reduction,
as the omcA–omcB– double mutant cannot grow on either Fe(III)-nitrilotriacetic acid or Fe(III)-citrate, whereas they appear to have an important, although redundant, function as a terminal V(V) reductase, as the omcA–omcB– double mutant still reaches 50%
of the MR-1R turbidity Knocking out either omcA or omcBturned out to have no significant growth pheno-type with either U(VI) or Se(VI) as the terminal elec-tron acceptor These results therefore provide evidence that there are differences between the electron transfer pathways towards chelated Fe(III) on the one hand and either U(VI) or Se(VI) on the other Redundancy between these pathways may explain the growth curves observed for V(V) reduction
Decaheme cytochrome c quantification of anaerobically Fe(III)-respiring omcA–, omcB–and omcA–omcB–MR-1R mutants relative to their MR-1R parent
The major impact on Fe(III) respiration by OmcB relat-ive to OmcA can be explained by one or a combination
Trang 3of the following possibilities: (a) the steady-state OmcB
concentration is greater than that of OmcA; (b) OmcB
is differentially produced (upregulated) by the omcA
insertional inactivation, but not vice versa; (c) OmcA
and OmcB show different behavior patterns in terms
of kinetics; and (d) OmcB is required to obtain
functional OmcA These possibilities are discussed
below
A heme-staining approach was used to reveal the
decaheme cytochrome c pools present in
Fe(III)-respir-ing MR-1 omcA–, omcB– and omcA–omcB– mutants
relative to their MR-1R parent Figure 2B shows the
absence of mature OmcA (83 kDa) and OmcB
(78 kDa) in an omcA– and an omcB– background,
respectively, a complete lack of both proteins in the
omcA–omcB–double mutant, and approximately equal
amounts of either decaheme cytochrome c in an
MR-1R extract Relative to MR-1R, Fig 2B does not
suggest compensatory induction of either OmcB or OmcA in an omcA– or omcB– background, respect-ively
To calculate the OmcA and OmcB content in Fe(III)-respiring MR-1R and single mutants, differen-tial absorption spectra for reduced-minus-oxidized heme were recorded (Fig 2D) As these spectra are based on total heme content, it is imperative that all the other heme-containing proteins in the cells are not subjected to regulation in the respective mutants Fig-ure 2B,C shows that, apart from OmcA and OmcB, the periplasmic fumarate reductase (FR), the cytoplas-mic CymA and other, smaller (< 20 kDa), cyto-chromes are highly abundant c-type cytocyto-chromes in MR-1R, and thus contribute substantially to the
554 nm absorbance Although not fully linear and sat-urating with increasing cytochrome content, the heme staining experiments are indicative of the fact that these cytochromes are not subjected to upregulation or downregulation in the analyzed mutants We further-more monitored and compared FR activities in wild-type MR-1R and mutants The enzyme assay yielded activity values of (in lmolÆmin)1Æmg)1) 43.8 ± 0.90, 42.9 ± 0.58, 43.0 ± 0.24 and 44.3 ± 0.70 for MR-1R, omcA–, omcB– and omcA–omcB–, respect-ively, indicating no upregulation or downregulation of
FR (P¼ 0.83) On the basis of the fact that FR is not subjected to regulation under the applied conditions, and deducing from Fig 2C that all other c-type cyto-chromes are also invariantly produced in the respective mutants, we feel safe to extract OmcB and OmcA concentrations from omcA– and omcB– mutant heme values minus omcA–omcB– double mutant values, respectively The concentrations of OmcA and OmcB were subsequently calculated on the basis on the known stoichiometry of 10 heme groups per OmcA or OmcB molecule [16] This approach is valid, because
no alterations other than the expected disappearance
of either or both OmcA and OmcB in the respective mutants are apparent from the heme-staining gels The omcA– background contains 4.00 pmol of OmcB per
109cells, which, as to be expected from the heme stain
in Fig 2B, is similar to the OmcA concentration cal-culated for the omcB– background (3.43 pmol per
109cells)
By subtracting the heme concentration of the omcA–omcB– double mutant from that of MR-1R cells, we calculated a decaheme cytochrome c content (OmcA + OmcB) in MR-1R of about 6.68 pmol per
109cells This value matches the sum of both deca-heme cytochrome c concentrations in the respective single mutants, again showing that neither decaheme cytochrome c is upregulated in the absence of the
Fig 1 Anaerobic liquid growth experiments assess the role of
OmcA and OmcB in dissimilatory metal reduction Anaerobic liquid
growth of MR-1R, omcA–, omcB– and omcA–omcB– mutant
cul-tures with either Fe(III)-nitrilotriacetic acid, Fe(III)-citrate, V(V),
U(VI) or Se(VI) as terminal electron acceptor is represented as
the increase in density reached after 3 days of growth (A).
Complete curves of MR-1R grown on the different metals are
pro-vided in (B).
Trang 4other Statistical analysis (Student’s t-test) between the
MR-1R values and the sum of the values of the omcA–
and the omcB–mutants revealed that there is no
statis-tically significant difference (P¼ 0.43)
Whole cell kinetics of OmcA- and
OmcB-dependent chelated Fe(III) reduction
To establish whether differential kinetics and⁄ or
syner-gism explain the dominance of OmcB over OmcA in
dissimilatory chelated Fe(III) reduction, we determined
the kinetic parameters for each decaheme cytochrome
c using intact actively Fe(III)-respiring cells (Table 1)
Maximal activities were converted to turnover numbers
on the basis of either the OmcA or OmcB concentra-tions calculated in the above paragraph for the omcB– and omcA– single mutants, respectively As explained
in Experimental procedures, Monod-based kinetic models for whole cell kinetics simplify to Michaelis– Menten models under the conditions applied in this study
Figure 3A shows Fe(III)-nitrilotriacetic acid satura-tion curves obtained using either omcA– [OmcB-dependent Fe(III) reduction], omcB–[OmcA-dependent Fe(III) reduction] or MR-1R [OmcA + OmcB-dependent Fe(III) reduction] cells In the absence of
Table 1 Enzymatic properties of OmcA- and OmcB-dependent-Fe(III)-nitrilotriacetic acid reduction Values represent the average of triplicate experiments ± SD.
Fe(III)-nitrilotriacetic acid
kcat⁄ K m ( M )1Æs)1) 1.17· 10 6 7.33 · 10 6
V2O5
Fig 2 Heme quantifications reveal unaltered protein production profiles of both OmcA and OmcB in the respective single mutants relative to the wild-type (A) RT-PCR confirming the absence of polar effects in mutants omcA – and omcB – Specific oligonucleo-tides were used to amplify omcA (lane 2), omcB (lane 3), mtrA (lane 4) and mtrB (lane 5) in the omcA–mutant, and omcA (lane 7), omcB (lane 8), mtrA (lane 9) and mtrB (lane 10) in the omcB –
mutant MR-1R was used as a positive control to display omcA (lane 1) and omcB (lane 6) DNA standards are indicated at the left and right of the agarose gels (B) Visualization and separation of high molecular mass cytochromes c through heme staining of a Tris ⁄ glycine SDS ⁄ PAGE gel loaded with 4 · 10 7 whole cells from anaerobically grown overnight cultures of MR-1R (lane 1), mutants omcA – (lane 2), omcB – (lane 3), and omcA – omcB – (lane 4), and complemented strains omcA – ⁄ pBAD202 ⁄ D-TOPOomcA (lane 5) and omcB–⁄ pBAD202 ⁄ D-TOPOomcB (lane 6) A molecular mass standard is indicated at the right (C) Visualization of low molecular mass cytochromes c through heme staining of a Tricine ⁄ SDS ⁄ PAGE gel loaded with 4 · 10 7
whole cells from anaerobically grown overnight cultures of MR-1R (lane 1), and mutants omcA –
(lane 2), omcB – (lane 3), and omcA – omcB – (lane 4) A molecular mass standard is indicated at the left (D) Bar graph represen-tation of the cytochrome content, normalized to 10 9 CFU, and calculated from reduced-minus-oxidized heme absorption differ-ences at 554 nm (a peak) using the absorption coefficient of
21 400 M )1Æcm)1 The differences in peak height reflect the
concentrations of OmcA and OmcB in omcB – and omcA – cells, respectively.
A
B
C
D
Trang 5synergism, the OmcA- and OmcB-dependent substrate
saturation curves should add up to form the MR-1R
(OmcA + OmcB) curve; this is a valid assumption, as
we could not identify differential protein production
profiles as mentioned in the previous paragraph At
full Fe(III)-nitrilotriacetic acid saturation, the modeled
summation function corresponds well with the MR-1R
curve, whereas it shows slightly lower than
experiment-ally determined activities at nonsaturating Fe(III)-nitrilotriacetic acid concentrations This suggests that OmcA might synergistically enhance, albeit slightly, the affinity of OmcB for its metal substrate However, the curves totally refute the reverse possibility, i.e that OmcB is needed to get functional OmcA
On the other hand, the derived kinetic parameters for OmcA- and OmcB-dependent chelated Fe(III) reduction summarized in Table 1 do rationalize the dominance of OmcB in dissimilatory Fe(III) reduction: under physiologically relevant low micromolar concen-trations of Fe(III), OmcA should outnumber OmcB six-fold to catalyze electron transfer at a similar rate Complementation of the omcA– and omcB– mutants restored Fe(III)-nitrilotriacetic acid reduction activity
to MR-1R levels (Fig 3B)
Inhibition assays of OmcA- and OmcB-dependent chelated Fe(III) reduction as a measure of enzyme specificity
To determine whether the lack of phenotype of omcA–omcB– strains observed during anaerobic growth on either of the electron acceptors U(VI) and Se(VI) is due to the decaheme cytochromes c not recognizing either of these electron acceptors, we probed the relative affinities via competition assays Figure 4 shows the IC50plots of the inhibition data of whole cell OmcA- and OmcB-dependent Fe(III)-nitrilo-triacetic acid reduction by either V(V), U(VI), or Se(VI) Only V(V) appears to significantly inhibit Fe(III) reduction, as characterized by IC50s of 10.7 lm and 81.4 lm for inhibition of OmcA and OmcB, respectively
Modes of inhibition of either OmcA or OmcB
by V(V) The modes of inhibition of either OmcA- or OmcB-dependent Fe(III)-nitrilotriacetic acid reduction by V(V) were investigated for the two following reasons: (a) to derive the relevant inhibition constants; and (b)
to establish whether both decaheme cytochromes c may differ mechanistically Fe(III)-nitrilotriacetic acid saturation curves in the absence and in the presence of two different concentrations of V(V) were plotted and modeled to obtain the apparent Vmax and Km values (Fig 5A,B) These parameters were subsequently used
to generate double-reciprocal Lineweaver–Burk plots
to easily determine inhibitor modality (Fig 5C,D; Table 1)
OmcA inhibition by V(V) is characterized by an increase in apparent Km and no change in apparent
Fig 3 Kinetic characterization of OmcA- and OmcB-dependent
Fe(III)-nitrilotriacetic acid reduction rationalizes the dominance of
OmcB in anaerobic ferric iron respiration (A) Monod-based kinetic
model curves [34] for Fe(III)-nitrilotriacetic acid reduction by MR-1R
cells (inverted triangles), omcA – cells (squares), and omcB – cells
(triangles) As explained in Experimental procedures, the two latter
curves simplify to the Michaelis–Menten formulation under the
con-ditions applied Adding up these curves generates the dotted-line
curve, which, as explained in Experimental procedures, should
resemble the MR-1R curve Because this assumption is only valid
at saturating Fe(III)-nitrilotriacetic acid concentrations, slight synergy
may modulate activity when both OmcA and OmcB are present in
the outer membrane (B) In trans complementation of omcA – and
omcB–cells restores Fe(III) reductase activity to MR-1R levels See
Experimental procedures for details.
Trang 6Vmax, generating Lineweaver–Burk lines with
intersect-ing y-axis intercepts, which is the characteristic
signa-ture of competitive inhibition We calculated a Ki
value of 22.5 lm, suggesting that the kinetics of V2O5 binding to OmcA are similar to those for binding of Fe(III)-nitrilotriacetic acid
Fig 4 Competition assays of OmcA- (left panel) and OmcB-dependent (right panel) Fe(III)-nitrilotriacetic acid reduction with other metals show that only V(V) may represent an alternative substrate for both cytochromes Fe(III)-nitrilotriacetic acid reductase activity in the absence
of a competing metal substrate is set to 100% Relative activities are plotted as a function of increasing concentrations of either V(V) (as vanadate; red), U(VI) (as uranyl acetate; green), or Se(VI) (sodium selenate; purple) Inhibition curves were fitted to the standard hyperbolic inhibition equation (see Experimental procedures).
Fig 5 Analysis of the modes of inhibition of OmcA- and OmcB-dependent Fe(III) reduction by V(V) reveals mechanistic differences between the two cytochromes (A, B) Direct plots of the steady-state velocities of OmcA-dependent (A) and OmcB-dependent (B) Fe(III)-nitrilotriacetic acid reduction in the absence and the presence of two increasing V(V) concentrations (C, D) Theoretical double reciprocal plots using the kinetic parameters obtained by fitting the data from the direct plots.
Trang 7OmcB inhibition by V(V) is characterized by a
decrease in apparent Km and Vmax By plugging the
values of the modeled apparent kinetic parameters into
the double-reciprocal Lineweaver–Burk equation and
plotting the resulting linear functions, we obtained the
graph in Fig 5D The lines intersect at negative values
of 1⁄ [S] and 1 ⁄ v, which is a characteristic signature
of noncompetitive inhibition Thus, V(V) apparently
binds both the free OmcB enzyme and the binary
OmcB–Fe(III)-nitrilotriacetic acid complex, and the
binding is kinetically favored upon
Fe(III)-nitrilotriace-tic acid binding We calculated Kic and Kiu values of
65.9 lm and 11.5 lm, respectively, which again appears
to have physiologic significance Hence, besides having
significantly different turnover rates, OmcA and OmcB
may also behave differently in terms of binding their
metallic substrates
Discussion
In the present study, we could not detect
an-aerobic Fe(III)-nitrilotriacetic acid respiration for
omcA–omcB–double mutant cells Virtually no biomass
was generated in minimal medium containing lactate
and Fe(III)-nitrilotriacetic acid as the electron donor
and acceptor, respectively (Fig 1), and baseline
reduc-tion of Fe(III)-nitrilotriacetic acid was seen in the
ferro-zine-based whole cell kinetic approach (data not
shown) The collective action of both decaheme
cyto-chromes c, OmcA and OmcB, appears to be crucial for
anaerobic soluble Fe(III) respiration, and, because
of their outer membrane localization, one or both
cytochromes probably function as terminal Fe(III)
reductases Both these outer membrane-localized
cytochromes are reduced through an as yet incompletely
identified electron transport chain, which at an early
point receives electrons from the NADH pool, in our
study obtained by lactate supplementation In a recent
study, Marshall et al [15] established almost equally
fast direct electron transfer from either
dithionite-reduced MR-1 OmcA or OmcB to chelated Fe(III),
pro-viding the first biochemical evidence that both decaheme
cytochromes c are in fact functional Fe(III) reductases
As an OmcA⁄ OmcB double mutant strain does not
show any Fe(III) reduction activity, our study not only
strengthens, but also exceeds, this evidence, in that
OmcA and OmcB are found to be the sole Fe(III)
reduc-tases present in MR-1 Furthermore, the outer
mem-brane localization and partial extracellular exposure of
both cytochromes c, combined with the fact that the
result of adding up the OmcA and OmcB
Fe(III)-nitrilo-triacetic acid reduction curves conforms to the MR-1R
curve, allow us to deduce that the electron transport
chain does not bifurcate any further, but ends at this point before transferring electrons to the subject metal species, indicating that OmcA and OmcB are the ter-minal Fe(III) reductases in MR-1 Other MR-1 cyto-chromes c, previously shown to be ferric iron reductases
in vitro, such as MtrA [17] and Ifc3 in S frigidimarina [18], appear to be not directly involved in the process of anaerobic chelated Fe(III) respiration
Notably, the apparent maximal rate reported for Fe(III)-nitrilotriacetic acid-dependent OmcB oxidation
is approximately 50 times slower than the kcat for OmcB-dependent Fe(III)-nitrilotriacetic acid reduction (205 s)1), determined here using a whole cell kinetics approach, which has the advantages of: (a) maintain-ing the complete electron transport chain used durmaintain-ing metal respiration; and (b) keeping the terminal reduc-tases in their native cellular compartment For OmcA, the in vitro Kobs values determined by Marshall et al [15] and the in vivo kcatvalues determined in our study also differ, although to a lesser extent (six-fold) This dis-crepancy can most likely be accounted for by the fact that the purified cytochromes used in the in vitro approach lack some factor(s), such as one or more pro-tein partners or lipids that generate maximal activity Reduced activity due to detergent-based solubilization
of the outer membrane cytochromes is an alternative explanation
Growth experiments as well as the whole cell Fe(III) reduction kinetics presented here agree with previous findings that OmcB is more important than OmcA in anaerobic Fe(III) respiration [19] Using a heme-quanti-fication approach, we have presented evidence showing that this relative difference is not based on differential protein production profiles of either the omcA or omcB gene in the presence or absence of the other Shi et al [19] provided evidence for synergistic complex forma-tion between both decaheme cytochromes, which may explain the dominance of OmcB over OmcA in dissimi-latory Fe(III) reduction Our whole cell-based kinetic analysis, however, refutes the possibility that OmcB is necessary to reconstitute fully functional OmcA, as the Fe(III)-reducing activities of omcA–and omcB–cells add
up to the counterpart activities of MR-1R cells A per-fect fit, however, only becomes possible after slightly increasing the affinity of OmcB for its chelated Fe(III) substrate (Fig 3A) Complex formation may thus cause some synergism only at low micromolar and therefore physiologically relevant substrate concentrations The kinetics for OmcA- and OmcB-dependent Fe(III)-nitrilotriacetic acid reduction (Table 1) do rationalize the different roles of these proteins in Fe(III) respiration Both cytochromes have similar low micro-molar affinities for their Fe(III) substrate; however,
Trang 8completion of the electron transfer pathway takes
11.5 times longer for OmcA than for OmcB Taking
into account the specificity constants, OmcA should
out-number OmcB about six-fold if it is to substitute for the
latter in anaerobic Fe(III) respiration at physiologic
fer-ric iron concentrations, a hypothesis that will be pursued
further in our laboratory Note that the division of labor
established here for OmcA and OmcB cytochromes
should not necessarily apply to homologs from different
backgrounds; the OmcA homolog from S frigidimarina,
for example, has been found to be as fast (206 s)1) as
the S oneidensis MR-1 OmcB reductase [20]
It has previously been recognized that both
cyto-chromes, OmcA and OmcB, appear to have some
sub-strate specificity, as purified reduced batches lack
activity towards nitrite, nitrate and, in the case of
OmcA, uranyl acetate [15] OmcB was shown to have
some activity towards U(VI); however, the turnover
number (Kobs1¼ 0.039 s)1) is more than 100 times
lower than that for Fe(III)-nitrilotriacetic acid
(Kobs1¼ 4.1 s)1) [15] Our anaerobic growth
experi-ments show that neither decaheme cytochrome c is
necessary for dissimilatory uranyl acetate reduction
(Fig 1) OmcA, as expected, but also OmcB does not
bind U(VI) in the competition assay shown in Fig 4
The 100-fold lower Kobs1 for U(VI) reduction
com-pared to Fe(III)-nitrilotriacetic acid reduction reported
by Marshall et al [15] thus appears to result not from
disturbed catalysis, but rather from hampered
sub-strate binding Of the other metals tested in this study
[V(V) and Se(VI)], only vanadate was shown to be a
substrate for either OmcA or OmcB Inhibition
experi-ments suggest that Fe(III) and V(V) bind both
cyto-chromes with similar efficiencies (Table 1) However,
whereas omcA–omcB– double mutant cells did not
grow on chelated Fe(III), they do grow on V(V) to
about 50% of the MR-1R stationary-phase density
(Fig 1) In the case of V(V), the electron transport
chain may thus bifurcate to one or several other, as
yet unrecognized, terminal reductases Redundancy in
terminal metal reductases has been clearly shown here,
as MR-1 does not suffer from the omcA–omcB–
dou-ble mutants in anaerobic growth on the terminal
elec-tron acceptors Se(VI) and U(VI), and as none of these
metals inhibits OmcA- and OmcB-dependent whole
cell Fe(III)-nitrilotriacetic acid reduction In summary,
metal reduction appears to be a selective process in
which the reduction potential and the topology and
accessibility of the presented metal play crucial roles in
terms of binding efficiencies and subsequent reduction
by the appropriate enzyme The identification and
characterization of alternative terminal metal
reductas-es will be the subject of future rreductas-esearch
Experimental procedures
Bacterial strains
S oneidensis MR-1 was originally isolated from Oneida Lake sediments (Oneida Lake, NY, USA) [21], and was obtained from the LMG culture collection (LMG 19005; Ghent, Belgium) S oneidensis MR-1R is a spontaneous rif-ampicin-resistant mutant of strain MR-1 that was isolated in-house Escherichia coli strain TAM1pir+and E coli S17-1kpir cells were used for cloning purposes and conjugation experiments, respectively
Growth conditions MR-1R, omcA–, omcB–and omcA–
omcB–S oneidensis cul-tures were routinely grown overnight at 28C in LB broth and subsequently inoculated in M1 defined medium [22] sup-plemented with l-serine (1 lgÆmL)1), l-arginine (1 lgÆmL)1),
l-glutamate (1 lgÆmL)1), lactate (15 mm), and fumarate (20 mm) For growth experiments, fumarate was replaced
by either Fe(III)-citrate (2 mm), Fe(III)-nitrilotriacetic acid (0.5 mm), Na2SeO4 (1 mm), or UO2(CH3COO)2.2H2O (0.5 mm) (all products: Sigma-Aldrich, Bornem, Belgium) Growth on V(V) was studied using VM medium [23] Anae-robicity was achieved using a Coy anaerobic chamber (Coy Laboratories, Grass Lake, MI) containing 90% N2, 8% CO2, and 2% H2 The presence of H2in the anaerobic chamber did not affect metal reduction (data not shown) Growth curves were recorded by measuring the attenuance (D655) of the cultures at regular time intervals for 3 days The average rise in density after 3 days ± SEM for triplicate readings are summarized in Fig 1A, whereas the growth curves for MR-1R grown on the different metals are shown in Fig 1B
Construction of the omcA–and omcB–single mutants and of the omcA–omcB–double mutant strains of MR-1
Single omcA–and omcB–mutants and a double omcA–omcB– mutant strain of MR-1 were generated by insertional inacti-vation using the pKNOCK-based system [24] The primers used in this study are summarized in Table 2 Briefly, internal PCR-amplified fragments of the omcA and omcB genes were 5¢-phosphorylated and cloned into EcoRV-digested and calf intestinal phosphatase-treated pKNOCK-Km and pKNOCK-Cm, respectively, using T4 DNA Ligase (all enzymes: New England Biolabs, Ipswich, MA), yielding pKNOCK-Km-omcA and pKNOCK-Cm-omcB These con-structs were transformed into E coli S17-1kpir cells Equal amounts of overnight-grown transformed E coli S17-1kpir cells and rifampicin-resistant S oneidensis cells were mixed and spotted on LB⁄ Rif plates (10 lgÆmL)1) After a 6 h incu-bation period (necessary for the conjugation to take place), the cells were resuspended in 500 lL of LB broth [25] and
Trang 9plated on LB⁄ Rif plates containing either kanamycin
(25 lgÆmL)1) or chloramphenicol (25 lgÆmL)1) (Duchefa,
Haarlem, The Netherlands) After overnight incubation
at 28C, colonies were analyzed via PCR using the
oligo-nucleotides OMCA-F⁄ OMCA-R and OMCB-F ⁄ OMCB-R
(Table 2), designed to amplify the entire omcA gene and
omcBgene, respectively Homology-based insertional
integ-ration of the pKNOCK constructs enlarged the omcA
(2207 bp) and omcB (2015 bp) gene amplicons by 2700 and
2500 bp, respectively (data not shown) The omcA–omcB–
double mutant was constructed by applying a similar
proce-dure to that described above, using the omcA–mutant as the
recipient strain in conjugation As omcA and omcB are part
of the gene cluster mtrDEF–omcA–mtrCAB (omcB is also
known as mtrC), and the genes mtrCAB form a single
operon, we expected polar effects to occur when disrupting
omcB RT-PCR experiments proved the absence of such
polar effects (Fig 2A) and confirmed that we had obtained
the omcA–and the omcB–mutants
Complementation of the MR-1 omcA–and omcB–
mutant strains
Oligonucleotides OMCA-PBAD-F⁄ OMCA-PBAD-R and
OMCB-PBAD-F⁄ OMCB-PBAD-R (Table 2) were used to
amplify the omcA and omcB genes from MR-1 genomic
DNA, respectively These genes were subsequently cloned
into vector pBAD202⁄ D-TOPO (Invitrogen, Carlsbad,
CA), and the constructs were transformed into the
appro-priate omcA– or omcB– mutants of MR-1 by
electropora-tion, generating the in trans complemented strains As
pBAD202⁄ D-TOPO carries a kanamycin resistance region,
the ability to complement the omcA– mutant was shown
using a pKNOCK-Cm-based omcA– mutant, instead of
the pKNOCK-Km-based mutant that was applied in all
other experiments Full complementation of either the omcA
or omcB insertional mutation by the wild-type genes,
controlled by an arabinose promoter [26], was achieved as visualized by heme staining of SDS⁄ PAGE gels (Fig 2B),
as well as at the level of activity (see further)
Visualization of c-type cytochromes using heme staining
High and low molecular mass c-type cytochromes were resolved by SDS⁄ PAGE according to Laemmli [27] and Schaegger & von Jagow [28] (tricine gels), respectively In either case, 4· 107
whole cells of anaerobically grown over-night cultures were applied to the gels, which were then heme stained according to Thomas et al [29] The outer membrane cytochromes c OmcA and OmcB, the periplas-mic FR, and the cytoplasperiplas-mic tetraheme cytochrome c CymA were unambiguously identified via MS from heme-stained Tris⁄ glycine gels and tricine gels, respectively
Spectral quantification of the outer membrane decaheme cytochromes c OmcA and OmcB The heme content of whole cells was determined using the difference absorption coefficient of 21 400 m)1Æcm)1 [16] at
554 nm for the pyridine ferrohemochrome minus pyridine ferrihemochrome spectrum In that study, the difference absorption coefficient was determined at pH 8.0, whereas all our experiments were carried out at pH 7.5 We observed
no differences between spectra measured at pH 8.0 and 7.5 (data not shown) Sodium dithionite was used as the redu-cing chemical Overnight anaerobically grown cells (with
20 mm fumarate as the electron acceptor) were washed with and suspended in an equal volume of air-saturated NaCl⁄ Pi
(pH 7.5), and incubated at room temperature for 1 h to ensure oxidation of the outer membrane cytochromes Absorption spectra of 1 mL fractions were recorded at
554 nm using a double-beam spectrophotometer (Uvikon, Kontron, Herts, UK) in the absence and the presence of a few crystals of sodium dithionite (Sigma-Aldrich) The decaheme cytochrome c concentration was calculated as explained in Results, taking into account 10 heme groups per molecule of either OmcA or OmcB and our experiment-ally derived correlation between D655and cell concentration (a 1 mL MR-1 culture with a D655 of 1.0 contains 1.44· 109
cells) The values presented are means of tripli-cate experiments ± SEM To quantify FR, lysed MR-1R omcA–, omcB–and omcA–omcB–cells were assayed for this specific enzyme activity according to Maklashina et al [30]
Whole cell kinetics of ferric iron reduction The Fe(III) reductase activity of whole cells was measured using the ferrozine-based method [31] The chromophore formed by ferrous iron and ferrozine was measured at
562 nm [32] Whole cells for the Fe(III) reductase assays were
Table 2 Synthetic oligonucleotides used in this study.
Oligonucleotide name Sequence (5¢- to 3¢)
AAACGGTTCAATTTC
AACGCACAAAAATCA
Trang 10prepared as follows Anaerobically grown cells (with
fuma-rate as the terminal electron acceptor) were collected by
centrifugation at 10 000 g (Beckman Coulter Avanti J-301
centrifuge, JA-30.50 rotor), washed twice with NaCl⁄ Pi
sup-plemented with 1 mm lactate (unless otherwise mentioned),
and placed on ice These preparations retained full activity
for at least 4 h Comparison of the reduced-minus-oxidized
spectra of anaerobically grown MR-1R cells washed with
NaCl⁄ Pi(pH 7.5) on the one hand, or water on the other,
revealed no differences in heme content, indicating that the
salt treatment did not lead to unwanted release of outer
membrane cytochromes Assays were conducted in microtiter
plates at 25C in a final volume of 200 lL of NaCl ⁄ Pi
(pH 7.4), and were monitored using a Bio-Rad model 680
microplate reader (Bio-Rad, Hercules, CA) A standard
reac-tion mixture contained 1 mm
3-(2-pyridyl)-5,6-bis(4-phenyl-sulfonic acid)-1,2,4-triazine monosodium salt (ferrozine;
Sigma-Aldrich), 1 mm lactate (unless otherwise mentioned),
a 1 : 100 dilution of the washed cell preparation, and
Fe(III)-nitrilotriacetic acid at concentrations ranging from 0.5 lm to
1.5 mm Phosphate did not interfere with the reduction assay
(data not shown), which is in accordance with the results
reported by Ruebush [33] For inhibition studies, the
stand-ard reaction mixture containing 100 lm
Fe(III)-nitrilotriace-tic acid (unless otherwise mentioned) was supplemented with
either V(V) (as V2O5), Se(VI) (as Na2SeO4) or U(VI) [as
UO2(CH3COO)2.2H2O], ranging in concentration from
0.5 lm to 1 mm Inhibition curves were fitted using a least
squares algorithm (graphpad prism Version 4.00; GraphPad
Software, Inc., San Diego, CA) to the equation:
mr¼ 100 ðImax½Me=ðIC50þ ½MeÞÞ
where vr is the relative activity, Imax is the maximal
response amplitude, [Me] is the supplemented initial
con-centration of inhibiting metallic substrate, and IC50 is the
half-maximal concentration of inhibiting metallic substrate
To analyze kinetic data, we used Monod-based kinetic
models [34] that actually simplify to a Michaelis–Menten
for-mulation under the applied conditions The kinetic rate is
determined solely by the electron acceptor, as the electron
donor used (lactate, 1 mm) is supplied in excess The effect of
bacterial growth on Fe(III)-nitrilotriacetic acid reduction can
be neglected, as the initial cell concentration used was high,
and growth-supporting nutrients were excluded We also
assumed that cell decay can be neglected, because the activity
proceeded linearly during our 1 h analyses Therefore, the
Monod model takes a form similar to the Michaelis–Menten
expression v¼ VmS⁄ (Ks+ S), where Vmequals the maximal
activity for the initial bacterial concentration, S is the initial
Fe(III)-nitrilotriacetic acid concentration, and Ksis the
half-velocity constant As we have determined the OmcA and
OmcB concentrations present in omcB– and omcA– cells,
respectively, and because omcA–
omcB–double mutant cells completely lack Fe(III) reductase activity, we can, using the
single mutants, convert Vmvalues to kcat values, and safely
assume Ksto be Km, the familiar Michaelis–Menten constant for enzyme-catalyzed reactions Activity data were fitted to the regular Michaelis–Menten equation using graphpad prism Version 4.00 For MR-1R- and OmcA-dependent kinetics, the Michaelis–Menten equation was adjusted for substrate inhibition
Acknowledgements
This work was supported by a personal grant to
J Borloo from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen) J Van Beeumen and B Devreese are indebted to the Fund for Scientific Research (FWO-Vlaanderen) for granting research project G.0190.04, as well as to the Bijzonder Onderzoeksfonds of Ghent Uni-versity for Concerted Research Action GOA 120154
References
1 Krause B & Nealson KH (1997) Physiology and enzy-mology involved in denitrification by Shewanella putre-faciens Appl Environ Microbiol 63, 2613–2618
2 Moser DP & Nealson KH (1996) Growth of the faculta-tive anaerobe Shewanella putrefaciens by elemental sul-fur reduction Appl Environ Microbiol 62, 2100–2105
3 Myers CR, Carstens BP, Antholine WE & Myers JM (2000) Chromium(VI) reductase activity is associated with the cytoplasmic membrane of anaerobically grown Shewanella putrefaciens MR-1 J Appl Microbiol 88, 98–106
4 Myers JM & Myers CR (2001) Role for outer mem-brane cytochromes OmcA and OmcB of Shewanella putrefaciens MR-1 in reduction of manganese dioxide Appl Environ Microbiol 67, 260–269
5 Myers JM & Myers CR (2003) Overlapping role of the outer membrane cytochromes of Shewanella oneidensis MR-1 in the reduction of manganese(IV) oxide Lett Appl Microbiol 37, 21–25
6 Carpentier W, Sandra K, De Smet I, Brige A, De Smet
L & Van Beeumen J (2003) Microbial reduction and precipitation of vanadium by Shewanella oneidensis Appl Environ Microbiol 69, 3636–3639
7 Klonowska A, Heulin T & Vermeglio A (2005) Selenite and tellurite reduction by Shewanella oneidensis Appl Environ Microbiol 71, 5607–5609
8 Davis JA, Kent DV, Rea BA, Maest AS & Garabedial
SP (1993) Influence of redox environment and aqueous speciation on metal transport in groundwater: prelimin-ary results of tracer injection studies In Metals in ground water(Allen HE, Perdue EM & Brown DS, eds),
pp 223–273 Lewis Publishers, Chelsea, MI
9 Schwalb C, Chapman SK & Reid GA (2003) The tetraheme cytochrome CymA is required for anaerobic