Cr(VI) Reduction by Sulfidogenic and Nonsulfidogenic Microbial Consortia

In time course experiments, bacterial community compositionswere compared between a sulfidogenic and two nonsulfidogenicCr(VI)-reducing consortia enriched from metal-contaminated sediments.The consortia were subjected to 0 and 0.85 mM or 1.35 mM Cr(VI),and Cr(VI) reduction, growth, and denaturing gradient gel electrophoresisprofiles of PCR products of small-subunit (16S) ribosomal geneswere compared. Results showed that although Cr(VI) was completelyreduced by the three consortia, Cr(VI) inhibited cell growth,with sulfate-reducing bacteria being particularly sensitiveto Cr(VI) toxicity relative to other bacteria in the consortia.

INTRODUCTION

Top

Introduction

Chromium is a metal contaminant that, in nature, exists primarilyas the soluble, highly toxic Cr(VI) anion and the less soluble,less toxic Cr(III) species. Many facultative and strictly anaerobicbacteria commonly found in soils and marine sediments are capableof reducing Cr(VI) to Cr(III) (10, 14, 20, 34, 35). Sulfate-and iron-reducing bacteria can indirectly reduce Cr(VI) viatheir anaerobic metabolic end products, hydrogen sulfide (HS^-)and Fe(II), respectively (22, 25, 26, 29, 32). Some bacteria,such as Desulfotomaculum reducens and Pantoea (formerly Enterobacter)agglomerans strain SP-1, may use Cr(VI) as an alternate electronacceptor for anaerobic growth (10, 34), while others have cytochromes(15) (e.g., Desulfovibrio vulgaris) or specific enzymes (3,17, 21, 33) (e.g., several Pseudomonas species, Escherichiacoli, and Shewanella oneidensis strain MR-1) that reduce Cr(VI).

Despite the knowledge regarding Cr(VI) reduction mechanismsthat pure cultures have provided, we still know relatively littleabout Cr(VI) reduction by natural microbial communities. Sulfate-reducingconsortia have been studied previously and found to reduce upto 20 mM Cr(VI) (11). However, these samples were not degassedto remove hydrogen sulfide present in the medium prior to theaddition of the Cr(VI), as 15% of the Cr(VI) was removed withinthe first 30 min and the remaining Cr(VI) was removed within1 h. More recently, Marsh et al. (16) compared Cr(VI) reductionin aquifer sediment slurries amended with different electronacceptors. Results indicated that the reduction of 0.5 mM Cr(VI)in sandy sediment samples amended with sulfate or ferric ironpreceded sulfidogenesis or ferric iron reduction and appearedto be mediated by microorganisms. However, more informationregarding Cr(VI) reduction by bacterial communities is essential,since pollution levels and other environmental factors influencethe development, diversity, and activities of natural microbialcommunities in different environments.

One powerful molecular biological tool for assessing a microbialpopulation response to a toxic substance is denaturing gradientgel electrophoresis of PCR products of ribosomal small-subunit(16S) genes (PCR-DGGE). PCR-DGGE assesses the diversity andrelative dominance of community bacterial ribosomal DNA (rDNA)by usage of 16S rRNA gene banding patterns independent of culturingtechniques. Further, DNA sequencing of DGGE bands allows identificationof the dominant organisms present (8, 9, 13, 18).

In this study, the changing microbial diversities of one sulfidogenicand two nonsulfidogenic consortia over time during Cr(VI) exposurewere compared. Growth, Cr(VI) reduction, and PCR-DGGE of theconsortia in combination with DNA sequencing aided in identifyingthe predominant bacterial populations and those which mightbe important for Cr(VI) reduction in the environment.

Sediment samples from two sites served as inocula for the consortia.One sediment sample was taken from Green Sands Beach at MareIsland Naval Shipyard (MINSY), San Francisco, Calif., in September1996, where total Cr concentrations of $~$ 4.8 mM have been reported(1). By using Widdel's marine medium (39) with lactate (20 mM)as the carbon source, two enrichment consortia, MI-S and MI-A,were obtained. In spite of the presence of sulfate (28 mM) inthe medium, only consortium MI-S produced hydrogen sulfide,indicating the presence of sulfate-reducing bacteria (SRB);MI-A did not. Another sample was taken from surficial sedimentsin June 1999 from Point Mugu Naval Air Weapons Station, LosAngeles, Calif., where total Cr concentrations of 380 mM havebeen reported (37). The consortium PACC was enriched from thissample by using brackish lactate medium (39) but without sulfate.All cultures were grown in serum bottles with nitrogen gas inthe headspace; reductant was excluded to avoid the chemicalreduction of Cr(VI). To ensure stable consortia with reproduciblegrowth rates and Cr(VI)-reducing capacities, all three consortiawere grown with Cr(VI) concentrations at the upper end of theirCr(VI) tolerance range and transferred often (approximately10 times).

Time course measurements of Cr(VI) reduction and microbial communitycomposition via PCR-DGGE with each enrichment condition wereperformed in duplicate experiments to assess reproducibility.To prevent the abiotic reduction and precipitation of chromium,the sulfidogenic enrichment consortium, MI-S, was asepticallydegassed with N₂ for 25 min prior to inoculation to remove HS^-.HS^- was not detectable in the experimental bottles immediatelyafter inoculation. For the experiments amended with Cr(VI),0.85 mM Cr(VI) was added to MI-S and MI-A and 1.35 mM Cr(VI)was added to PACC just prior to inoculation. Consortia withoutCr(VI) and uninoculated Cr(VI)-containing media served as controls.

Hydrogen sulfide concentrations were measured colorimetricallyat 480 nm with copper sulfate (6). Cr(VI) was measured colorimetricallyat 540 nm with diphenylcarbazide (36). The amount of proteinwas determined as a proxy for microbial biomass with a bicinchoninicacid kit (Pierce) after cells were digested in 1 N NaOH for5 min at 100°C. Water blanks with 0.85 or 1.35 mM Cr(VI),corresponding to the growth conditions when Cr(VI) was present,were used because there was a slight interference with the bicinchoninicacid assay.

At each time point, high-molecular-weight DNA was isolated.To ensure cell disruption, 1/3 volume of 0.22-µm acid-washedbeads was added during the lysis stage and cells were homogenized(Mini Biospec Beadbeater) for 2.5 min. After bead beating andcentrifugation, DNA was extracted and purified with a QIAamptissue kit protocol (Qiagen Inc., Chatsworth, Calif.). DNA wasamplified with bacterial primer 1055 and the universal small-subunitrRNA primer 1392, which incorporates a 5' GC clamp (9, 28).Sequence data were obtained by using these primers sans GC clamp.Standard PCRs were performed by using a single annealing temperatureof 50°C. We avoided touchdown PCR protocols to prevent selectiveamplification of certain templates (7) at the expense of othersthat may be relevant to Cr(VI) reduction. Three different concentrationsof DNA template (approximately 25, 50, and 100 ng/µl)were used to obtain triplicate PCR samples. The PCR productswere pooled and purified with a QIAquick PCR purification kit(Qiagen Inc.).

DGGE was carried out with a CBS Scientific Co. (Del Mar, Calif.)system (DGGE-2000). Equal amounts ( $~$ 300 ng) of pooled triplicatePCR products were loaded (except where otherwise noted) on 35to 70% denaturant gels and run at 60°C for either 6 h at200 V or 16 h at 80 V. Gels were stained with Sybr Gold (MolecularProbes, Eugene, Oreg.), visualized under UV light, and analyzedwith a NucleoVision gel documentation system and Gel Expertsoftware (NucleoTech Corp., San Mateo, Calif.). DNA extractedfrom gel bands by recommended protocols (30, 31) served as templatesfor PCR reamplification following the same protocol as aboveand utilizing primers 1055F and 1392R, without GC clamps. PCRproducts were sequenced according to the manufacturer's instructions(Big Dye v. 2.0) by using an Applied Biosystems 373A automatedsequencer. Sequence comparisons to known sequences were performedby using the National Center for Biotechnology Information BLASTwebsite (2). PCR products that did not produce coherent sequenceswere cloned into the TA cloning vector pCR2.1-TOPO (Invitrogen,San Diego, Calif.), and the clones were sequenced.

The sulfidogenic consortium, MI-S.

Top

The sulfidogenic consortium, mi...

In the absence of Cr(VI), the protein concentration reachedits maximum (1.8 mg/ml) by 24 h (Fig. 1A) and the greatest amountof HS^- ( $~$ 10 mM) was formed by $~$ 48 h (Fig. 1B). When MI-S was grownwith Cr(VI), biomass was depressed until after the completereduction of Cr(VI), which occurred by 24 h (Fig. 1A and C).HS^- was below the detection limit, indicating the absence ofsulfate reduction until 100 h (Fig. 1B). The highest levelsof protein (1.3 mg/ml) and hydrogen sulfide (6.3 mM) were reachedat the end of the experiment (220 h). These findings suggestthat Cr(VI) strongly inhibited the activity and growth of theSRB.

View larger version (12K):
[in this window]
[in a new window]
FIG. 1. Effect of 0.85 mM Cr(VI) on protein concentration, HS^- production, and Cr(VI) reduction by the sulfidogenic enrichment consortium, MI-S, from MINSY. (A) Protein concentration of the control enrichment without Cr(VI) ( ${blacklozenge}$ ) and the experimental enrichment subjected to 0.85 mM Cr(VI) ( ${square}$ ). (B) HS^- production by the control enrichment ( ${blacklozenge}$ ) and the experimental enrichment subjected to 0.85 mM Cr(VI) ( ${square}$ ). (C) Reduction of 0.85 mM Cr(VI) by medium alone ( ${blacklozenge}$ ) and by MI-S ( ${square}$ ). Error bars are standard deviations for duplicate experiments.

Temporal changes in the microbial composition of MI-S in thepresence and absence of Cr(VI) were visualized by PCR-DGGE (Fig.2), and the relative band intensities were analyzed. In theabsence of Cr(VI), MI-S had similar profiles at 24 and 48 h(Fig. 2, lanes 2 and 4). DNA sequencing of the bands allowedtentative identification of the dominant members of the consortium(Table 1). Bacteria most closely related to SRB comprised 40to 50% of the bacterial community, based on band intensity.By 168 h, a Pantoea sp. had become more dominant in the no-Cr(VI)bacterial community, and the SRB had decreased to $~$ 10% (Fig.2, lane 6). In the presence of Cr(VI), at 24 h there were considerablyfewer cells than at other time points. Cr(VI) had a strikingeffect on the microbial composition of the culture at 24 and48 h, as most of the bands were absent or faint (Fig. 2, lanes3 and 5). In these samples, the band corresponding to the Pantoeasp. was dominant and remained abundant from the beginning tothe end of the experiment ( $~$ 20 to 80%). By 48 h, the bands correspondingto the Clostridium sp., Desulfovibrio sp., and Desulfomicrobiumsp. were barely detectable, indicating that these organismswere greatly reduced in abundance in the Cr(VI) consortium andwere inhibited until long after Cr(VI) was completely reduced(Fig. 2). By $~$ 168 h, however, the PCR-DGGE profile of the consortiumonce again resembled the original inoculum (Fig. 2, lane 1).These data corroborate the absence of measurable HS^-, indicatingthat SRB were not active at 24 or 48 h, even though Cr(VI) hadbeen completely reduced by this time (Fig. 1B and C).

View larger version (40K):
[in this window]
[in a new window]
FIG. 2. Time course PCR-DGGE analysis of the sulfidogenic enrichment consortium MI-S exposed to 0.85 mM Cr(VI). 16S rDNA fragments were analyzed on a denaturing gel (35 to 70% denaturant; Sybr Gold stained). Each lane represents the MI-S community DNA at a specific time point in the absence or presence of 0.85 mM Cr(VI). Because there were considerably fewer cells in the presence of Cr(VI) at 24 h than at other time points, for comparative purposes, 150 ng of total rDNA was loaded into lanes 2 and 3. All other lanes were loaded with 300 ng of total rDNA.

View this table:
[in this window]
[in a new window]
TABLE 1. Bacterial sequences from PCR-DGGE of the Cr(VI)-reducing consortia

The initial hypothesis for rapid Cr(VI) reduction was that hydrogensulfide produced by the enrichment consortium would chemicallyreduce Cr(VI) to Cr(III) (Fig. 1C). However, there is a 100-to 150-h lag before the onset of HS^- production and a lack ofan inverse relationship between Cr(VI) reduction and HS^- production.In addition, there is an absence of SRB in the PCR-DGGE analysesuntil Cr(VI) is completely reduced. All of these findings seemto exclude the possibility of HS^- reducing Cr(VI). Instead,either the SRB were lysed or growth was inhibited by Cr(VI);only after several days, after complete Cr(VI) reduction, weremeasurable HS^- produced and SRB-like bacteria detected in theconsortium.

The bacterial population implicated by PCR-DGGE as being responsiblefor Cr(VI) reduction in MI-S is a member of the genus Pantoea:a Pantoea sp. sequence was the only abundant bacterial sequencepresent in the Cr(VI) enrichment at 24 h, and the Pantoea sp.was the dominant organism at 48 h (Fig. 2), a time during whichCr(VI) was being reduced (Fig. 1C). Recently, Francis et al.(10) demonstrated that P. agglomerans strain SP1 had the capacityto couple the oxidation of lactate, acetate, and H₂ to the reductionof 0.1 mM Cr(VI), although SP1 was much less tolerant of Cr(VI)than the MI-S consortium. Whether the tolerance of MI-S to high(>0.8 mM) Cr(VI) concentrations is due to the tolerance andCr(VI) reduction capacity of the Pantoea sp. or whether otherorganisms in the consortium moderated the toxicity of Cr(VI)requires further investigation.

We cannot exclude the potential role of SRB in the initial rapidreduction of Cr(VI) in this consortium, even though all HS^-was removed from the inoculum. Desulfovibrio spp. are well knownto have a c₃ cytochrome that functions as a Cr(VI) reductase(15, 23). Thus, although SRB and most other bacteria are stronglyaffected by Cr(VI) and disappear early on in this culture, itis possible that c₃ cytochromes released from lysing SRB couldcontribute to Cr(VI) reduction.

The nonsulfidogenic consortia, MI-A and PACC.

Top

The nonsulfidogenic consortia,...

Both nonsulfidogenic consortia grew well in the absence of Cr(VI)as the protein concentration increased from time zero and reachedmaximum protein levels by 20 to 24 h (Fig. 3A and C). In Cr(VI)-stressedenrichments, growth of both MI-A and PACC was slightly inhibitedand biomass did not equal that of the control sample until 48h ( $~$ 0.4 and 0.35 mg of protein/ml, respectively), when approximately40% of the initial Cr(VI) concentration had been reduced (Fig.3A and C). However, both consortia were able to completely reduceCr(VI) within approximately 200 h (Fig. 3B and D). Even thoughthe time for reduction was similar, the PACC consortium couldwithstand and reduce a greater concentration of Cr(VI): 1.35versus 0.85 mM.

View larger version (19K):
[in this window]
[in a new window]
FIG. 3. Effect of 0.85 or 1.35 mM Cr(VI) on protein concentration and Cr(VI) reduction by the nonsulfidogenic enrichment consortia MI-A (A and B) and PACC (C and D). Protein concentration of the control enrichment ( ${blacklozenge}$ ) and the experimental enrichment subjected to Cr(VI) ( ${square}$ ) is shown for MI-A [0.85 mM Cr(VI)] (A) and PACC [1.35 mM Cr(VI)] (C). (B) Reduction of 0.85 mM Cr(VI) by medium alone ( ${blacklozenge}$ ) and by enrichment consortium MI-A ( ${square}$ ). (D) Reduction of 1.35 mM Cr(VI) by medium alone ( ${blacklozenge}$ ) and by enrichment consortium PACC ( ${blacksquare}$ ). Error bars are standard deviations for duplicate experiments.

The PCR-DGGE results (Fig. 4) indicate that Cr(VI) did not affectthe diversity of the bacterial rDNA of the MI-A consortium asgreatly as it affected the sulfidogenic consortium. The Pantoeasp. appears to be the dominant organism. The Pantoea sp. islikely the same organism that was implicated in reducing Cr(VI)in MI-S (Table 1), as the sequences (in base pairs analyzed)were identical. A Spirochaeta sp. is another dominant organismin this consortium. Spirochaeta spp. are known to produce acetateduring fermentation (5), which could conceivably be utilizedby the Pantoea sp. for the reduction of Cr(VI), similar to P.agglomerans strain SP-1, which reduced Cr(VI) in the presenceof acetate (10).

View larger version (56K):
[in this window]
[in a new window]
FIG. 4. Time course PCR-DGGE analysis of the nonsulfidogenic enrichment consortium MI-A exposed to 0.85 mM Cr(VI). 16S rDNA fragments were analyzed on a denaturing gel (35 to 70% denaturant; Sybr Gold stained). Each lane represents the community MI-A DNA at a specific time point in the absence and presence of 0.85 mM Cr(VI); 300 ng was loaded onto each lane. The PCR product used for preparing the inoculum was run on a different DGGE gel and presented alongside the time course experiment. It is apparent that bands in the inoculum correlate reasonably well with the time course experimental profile.

There were two main differences in the PCR-DGGE profiles betweenCr(VI)-containing and non-Cr(VI)-containing consortia (Fig.4). One was that the unidentified top two bands and the Clostridiumsp. band were completely absent when Cr(VI) was present anddid not reappear until $~$ 168 h (Fig. 4, lane 7), when Cr(VI) wasalmost completely reduced (Fig. 3B). Another notable differencein the PCR-DGGE banding profile was the response of Shewanellaalga, a metal-reducing bacterium. Without Cr(VI), the S. algaband is a minor member of the consortium at each time point( $~$ 4% of the total community DNA). In the presence of Cr(VI),this band is more abundant, comprising $~$ 20% of the microbialcommunity. The Shewanella sp. may have contributed to the reductionof Cr(VI) in our consortium, as Shewanella species are knownto reduce Cr(VI) enzymatically (19). Alternatively, S. algamay have indirectly reduced Cr(VI) via a catalytic Fe(III)/Fe(II)cycle in which the respiration of Fe(III) produces Fe(II), whichin turn reduces Cr(VI) and produces Fe(III) (24, 38, 40). Feis present as a trace nutrient (7 µM) in Widdel's mediumand is a common contaminant in laboratory medium components.

Figure 5 presents microbial community profiles of the PACC consortium.PCR-DGGE analysis indicated that Cr(VI) did not affect the microbialcommunity as greatly as it affected both MINSY sediment consortia.The diversity of organisms in this consortium was similar withand without Cr(VI) at 15 and 42 h (Fig. 5, lanes 3 and 5).

View larger version (55K):
[in this window]
[in a new window]
FIG. 5. Time course PCR-DGGE analysis of the nonsulfidogenic enrichment consortium PACC exposed to 1.35 mM Cr(VI). 16S rDNA fragments were analyzed on a denaturing gel (35 to 70% denaturant; Sybr Gold stained). Each lane represents the PACC community DNA at a specific time point in the absence or presence of 1.35 mM Cr(VI); 300 ng was loaded onto each lane.

Because the Shewanella sp. was implicated in Cr(VI) reductionin the MI-A consortium, we expected it to be active in the PACCconsortium as well. However, the Shewanella species is actuallyless dominant with Cr(VI) at 42 h (Fig. 5, lanes 2 and 5). Itis difficult to assess which organisms may be involved in Cr(VI)reduction in the PACC consortium. At this time, the capacityfor reduction of Cr(VI) by Spirochaeta spp., Propionibacteriumspp., Ralstonia spp., and Acetobacterium spp. has not been established.The Pseudomonas sp. may be contributing to Cr(VI) reductionin this consortium, as reductases capable of reducing Cr(VI)have been found in Pseudomonas ambigua (33) and in several otherPseudomonas spp. (4, 12, 17). Generally, pseudomonads are resistantto extreme Cr(VI) concentrations (10 mM) but do not reduce Cr(VI)(17).

Phylogenetic analysis.

Top

Phylogenetic analysis.

Sequence analyses of the PCR-DGGE bands were used to identifythe dominant organisms in the consortia, and in a few cases,bacterial isolates were obtained from the consortia (Table 1).In most cases the sequence similarity to known bacteria washigh (>97%) based on the $~$ 300 bases examined. One SRB-likebacterial sequence was only 88% similar to a Desulfomicrobiumsp. sequence. The other SRB was 99% related to the Desulfovibriostrain LZK1, found in naturally occurring radioactive material(27). Some bacterial species were present in two consortia (aPantoea sp., a Shewanella sp., and the uncultured bacteriumDCE-25, a Spirochaeta sp.). These bacterial sequences were alignedwith each other and determined to be identical in the sequencesexamined. Of the isolates obtained in pure culture, none wereable to reduce 0.85 mM Cr(VI) in the presence of either acetateor lactate as an electron donor (Table 1).

Conclusions.

Top

Conclusions.

This report establishes that the reduction of Cr(VI) at concentrationsup to 1.35 mM can occur with microbial consortia enriched fromcontaminated sites. Although this study examined bacterial consortia,these data also facilitate the understanding of bacterial interactionsin natural communities exposed to hexavalent Cr. Since manybacteria are inhibited by Cr(VI), it is inferred that the bacteriaresistant to Cr(VI) and/or responsible for its reduction willcontinue to grow, while other strains in the environment willremain inactive or dormant until Cr(VI) is reduced to noninhibitorylevels. Knowing which bacteria are active or thrive during Cr(VI)reduction is useful for site assessment and for designing bioremediationstrategies for chromium-contaminated sites.

That SRB are strongly inhibited by Cr(VI) is important becauseit is often assumed that if SRB are present in Cr(VI)-contaminatedenvironments, then the SRB will be metabolically active andproduce HS^-, which would abiotically reduce Cr(VI). While thismay be true at low Cr(VI) levels, in areas where Cr(VI) levelsare high, SRB are probably inhibited relative to other membersof the microbial community and, therefore, not the key playersin Cr(VI) reduction. This interpretation is consistent withdata of Marsh et al. (16), who showed that sulfate reductionand HS^- production did not occur when sandy sediments were exposedto 0.5 mM Cr(VI). Similar results have been repeatedly demonstratedin our laboratory in other pure and enrichment cultures of SRB.Our results suggest that facultative anaerobes such as speciesof Pantoea and Shewanella, at least some of which are capableof Fe(III) reduction, are more likely than SRB to be the keyplayers in Cr(VI) reduction in Cr-contaminated anaerobic environments.

ACKNOWLEDGMENTS

We thank Shelly Anghera for providing sediment samples fromPoint Mugu and Mike Johnson for collecting samples for us atMINSY. We also are grateful to Anna Obraztsova for advice, includingcritical comments on the manuscript, and to Chris Francis, currentTebo laboratory members, and the reviewers for their helpfulsuggestions on the manuscript.

Y.M.A. was supported by a MARC Graduate Fellowship through theNIH. This research was supported in part by Office of NavalResearch grant N00014-99-1-0107 and the University of CaliforniaToxic Substances Research and Training Program.

FOOTNOTES

* Corresponding author. Mailing address: Marine Biology Research Division and Center for Marine Biotechnology and Biomedicine, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0202. Phone: (858) 534-5470. Fax: (858) 822-5130. E-mail: btebo{at}ucsd.edu.

REFERENCES

Top