Transcriptome Analysis of Shewanella oneidensis MR-1 in Response to Elevated Salt Conditions

Abstract

Whole-genomic expression patterns were examined in Shewanella oneidensis cells exposed to elevated sodium chloride. Genes involved in Na⁺ extrusion and glutamate biosynthesis were significantly up-regulated, and the majority of chemotaxis/motility-related genes were significantly down-regulated. The data also suggested an important role for metabolic adjustment in salt stress adaptation in S. oneidensis.

Shewanella species inhabit diverse environments, including spoiled food (11) and infected animals (35), deep-sea and freshwater lake sediments (8, 45, 54), and oilfield waste sites (44). Shewanella oneidensis MR-1, a facultative, gram-negative bacterium, was isolated from sediments of Lake Oneida in New York (32). The bacterium can anaerobically respire numerous organic compounds, including fumarate and dimethyl sulfoxide (28), as well as reduce metals such as Fe(III), Mn(IV), Cr(VI), and U(VI) (22, 29, 32). Because of the respiratory versatility, which may be exploited for immobilization of environmental pollutants (i.e., chromium and uranium) in soil and groundwater, the metal-reducing capabilities of Shewanella spp. have been intensively investigated (6, 14, 15, 26, 30, 33, 39).

The MR-1 genome was recently sequenced (16), and some fundamental similarities and disparities between MR-1 and other sequenced bacteria have been observed (16). To experimentally probe the genomic response of S. oneidensis to various physiologically relevant environmental stresses, a whole-genome cDNA microarray for MR-1 was constructed in this laboratory. In this study, we used this cDNA microarray to profile transcriptional responses of MR-1 to elevated sodium salt stress. The results indicated that the expression of the genes involved in osmolyte protection, cation efflux/influx, motility, and electron transport were significantly altered.

MR-1 requires a relatively high salt concentration for optimal growth.

Many Shewanella species have been isolated from marine environments, whereas some, like MR-1, have been isolated from freshwater environments (36, 39). To understand how various salt concentrations impact the growth of S. oneidensis, MR-1 cells were cultivated in triplicates in MR2A medium (12) containing different amounts of NaCl (ranging in concentration from 0 to 0.6 M) at 30°C under aerobic conditions (shake flasks, 120 rpm). Growth curves (Fig. 1) indicated that (i) the growth rate increased slightly with additional NaCl levels up to 0.4 M, (ii) cells grown in the presence of 0.4 M NaCl entered stationary-phase growth at a lower optical density (OD) than cells grown in the presence of 0.1 to 0.3 M NaCl, (iii) the growth rate decreased significantly with the addition of 0.5 M NaCl, and (iv) cell growth was drastically reduced in the presence of 0.6 M NaCl. Based on these results, MR-1 cells required NaCl levels between 0.1 to 0.3 M for optimal growth (5.8 to 17.5 g/liter) in aerobic MR2A medium. A slight decrease in overall growth was observed at 0.4 M NaCl; 0.5 M NaCl (29.2 g/liter) reduced the maximum growth rate twofold compared to the maximum growth rate observed at 0.1 to 0.3 M NaCl, and the maximum growth rate at 0.6 M NaCl was reduced over fourfold. For the present study, 0.5 M NaCl was used as a moderate stress for MR-1 cells.

FIG. 1.

Relationship between maximum growth rate of MR-1 cells grown in aerobic MR2A medium and increasing levels of sodium chloride.

Microarray analysis of salt adaptation response in MR-1.

A whole-genome cDNA microarray was constructed and described previously (7, 13, 49). Briefly, gene-specific DNA fragments (<75% homology) were selected as probes with the software PRIMEGENS (52), and the primers were designed to amplify the gene-specific DNA fragments. A total of 4,648 pairs of gene-specific primers were designed based on the known sequences (13, 16) and synthesized. Gene-specific fragments were PCR amplified in 96-well plates 8 to 16 times in 100-μl reaction mixtures, purified, pooled, quantified, and diluted to a minimum concentration of 50 ng/μl. Microarray fabrication, hybridization, and scanning were carried out as described previously (7, 13, 23, 49).

We harvested cells grown in the presence of 0.1 M (control condition) or 0.5 M (salt stress condition) NaCl for analysis. To evaluate biological variations, we extracted total cellular RNA from three sets of independent salt-stressed and control cultures to serve as biological replicates and that were hybridized at least twice for each replicate set by an optimized protocol (7, 13, 23, 49). The ratios of the salt-stressed samples to the control samples for an arrayed gene were normalized by a trimmed geometric mean (48). Data points that were not consistently reproducible and had a disproportionately large effect on the statistical result were removed (23). Student's t test was used to identify differentially expressed genes by comparing the means of the normalized and log-transformed control versus salt-treated data with a total of 12 replicates in each set. A significance cutoff for the t statistic (P = 0.05) of a two-tailed test was chosen, and also required genes with significant changes to show a greater than twofold average change in expression level. As a result, a balance between the number of false negatives and trends supported by concerted changes among multiple genes within the same operon or pathway is achieved. For comparison, we also used the empirical Bayesian method of Lonnstedt and Speed (24) to rank and identify genes with significant changes, and the results are consistent by both methodologies.

The quality of the microarray data was assessed based on a number of criteria. First, expression patterns for genes in the same putative operons were checked. The similarity in gene expression patterns between gene pairs predicted to be in the same operon to that of randomly chosen gene pairs was compared. Consistent with this expectation, we observed that genes within the same operon responded in a similar fashion under salt stress compared to genes randomly selected from the genome. Observed pairwise differences in log ratio expression levels were significantly smaller for the within-operon set (Kolgomorov-Smirnoff test, D = 0.3925, P = < 2.2 × 10⁻¹⁶) (37). Second, genes known to function together displayed similar changes in expression levels, as described throughout this article. One example is the consistent down-regulation of flagellar assembly genes (Table 1). Third, expression patterns of well-studied genes were verified (e.g., cation efflux transporters and Na⁺/H⁺ antiporters; Table 1S, online supplementary data [http://www.esd.ornl.gov/facilities/genomics/pubs/Table1S.xls]). Finally, we selected four predicted open reading frames (ORFs) that displayed significant changes in expression that have not been previously described as osmotic stress response genes in other organisms for real-time quantitative reverse transcription-PCR analysis (23). The expression patterns of the selected genes (pflA, aceA, acnA, and SO3874) were similar to the patterns observed with the microarrays (Table 2S, online supplementary material [http://www.esd.ornl.gov/facilities/genomics/pubs/Table2S.pdf]).

TABLE 1.

Operon organizations and expression ratios of flagellar assembly genes in regions 1, 4, 5, and 6 and chemotaxis genes in regions 2 and 3

Gene ID no.a	Name	Putative function	Expression ratiob
			Avg	SD	Sig.
Region 1
SO1529(−)	motA-1	Chemotaxis protein	0.470	0.184	Yes
SO1530(−)	motB-1	Chemotaxis protein	0.421	0.109	Yes
Region 2
SO2120(−)	cheY-1	Chemotaxis protein	2.899	1.570	No
SO2121(−)	cheA-1	Chemotaxis protein	0.659	0.216	Yes
SO2122(−)	cheW-1	Purine-binding chemotaxis protein	0.407	0.222	Yes
SO2123(−)		Methyl-accepting chemotaxis protein	1.004	0.390	No
SO2124(−)	cheR-1	Chemotaxis protein methyltransferase	0.934	0.387	No
SO2125(−)	cheD-1	Chemotaxis protein	0.865	0.293	No
SO2126(−)	cheB-1	Protein-glutamate methylesterase	0.855	0.347	No
Region 3
SO2317(−)		Methyl-accepting chemotaxis protein	1.295	0.590	No
SO2318(−)	cheY-2	Chemotaxis protein	2.085	1.680	No
SO2319		Anti-anti-sigma factor, putative	1.600	1.127	No
SO2320(−)	cheA-2	Chemotaxis protein, interruption-N	1.414	1.222	No
SO2321(−)		ISSod4, transposase	No data
SO2322(−)	cheA-2	Chemotaxis protein, interruption-C	3.299	1.611	No
SO2323(−)		Methyl-accepting chemotaxis protein	1.933	1.425	No
SO2324(−)	cheW-2	Purine-binding chemotaxis protein	0.875	0.209	Yes
SO2325(−)	cheR-2	Chemotaxis protein methyltransferase	1.197	0.618	No
SO2326(−)	cheD-2	Chemotaxis protein, putative	1.373	0.666	No
SO2327(−)	cheB-2	Protein-glutamate methylesterase	0.713	0.266	Yes
Region 4
SO3202(−)	cheW-3	Purine-binding chemotaxis protein	1.247	0.495	No
SO3203(−)		CheW domain protein	0.987	0.300	No
SO3204		ParA family protein	1.103	0.368	No
SO3205		Hypothetical protein	1.233	0.398	No
SO3206(−)	cheB-3	Protein-glutamate methylesterase	1.615	0.717	No
SO3207(+)		Chemotaxis protein	1.226	0.387	No
SO3208(+)	cheZ	Chemotaxis protein	1.722	0.908	No
SO3209(+)	cheY-3	Chemotaxis protein	1.886	0.940	No
SO3210(+)	fliA	RNA polymerase sigma-28 factor	1.052	0.331	No
SO3211(+)	flhG	Flagellar biosynthetic protein	0.749	0.141	No
SO3212(+)	flhF	Flagellar biosynthetic protein	0.632	0.293	Yes
SO3213(+)	flhA	Flagellar biosynthesis protein	0.558	0.149	Yes
Region 5
SO3936(+)	motX	Sodium-type flagellar protein	0.788	0.323	No
Region 6
SO4286(+)	motB-2	Chemotaxis motB protein	0.737	0.328	No
SO4287(+)	motA-2	Chemotaxis motA protein	0.646	0.135	Yes
Region 4
SO3214		Hypothetical protein	0.369	0.154	Yes
SO3215(+)	flhB	Flagellar biosynthetic protein	0.678	0.237	Yes
SO3216(+)	fliR	Flagellar biosynthetic protein	0.439	0.159	Yes
SO3217(+)	fliQ	Flagellar biosynthetic protein	0.672	0.252	No
SO3218(+)	fliP	Flagellar biosynthetic protein	0.537	0.140	Yes
SO3219(+)	fliO	Flagellar protein	0.532	0.210	Yes
SO3220(+)	fliN	Flagellar motor switch protein	0.575	0.138	Yes
SO3221(+)	fliM	Flagellar motor switch protein	0.716	0.363	No
SO3222(+)	fliL	Flagellar protein	0.741	0.156	Yes
SO3223(+)	fliK	Flagellar hook-length control protein	0.378	0.244	Yes
SO3224(+)	fliJ	Flagellar protein	0.328	0.168	Yes
SO3225(+)	fliI	Flagellum-specific ATP synthase	0.577	0.161	Yes
SO3226(+)	fliH	Flagellar assembly protein	1.017	0.868	No
SO3227(+)	fliG	Flagellar motor switch protein	0.839	0.401	No
SO3228(+)	fliF	Flagellar M-ring protein	0.584	0.132	Yes
SO3229(+)	fliE	Flagellar hook-basal body complex	0.284	0.127	Yes
SO3230(+)	flrC	Flagellar regulatory protein C	0.352	0.137	Yes
SO3231(+)	flrB	Flagellar regulatory protein B	0.200	0.073	Yes
SO3232(+)	flrA	Flagellar regulatory protein A	0.388	0.150	Yes
SO3233(+)	fliS	Flagellar protein	0.527	0.183	Yes
SO3234		Hypothetical protein	0.497	0.273	Yes
SO3235(+)	fliD	Flagellar hook-associated protein	0.555	0.304	Yes
S03236(+)	flaG	Flagellin	0.345	0.078	Yes
SO3237(+)		Flagellin	0.317	0.097	Yes
SO3238(+)		Flagellin	0.846	0.471	No
SO3239(+)	flgL	Flagellar hook-associated protein	0.495	0.226	Yes
SO3241(+)	flgJ	Flagellar protein	0.422	0.216	Yes
SO3242(+)	flgI	Flagellar P-ring protein	0.383	0.229	Yes
SO3243(+)	flgH	Flagellar L-ring protein	0.427	0.170	Yes
SO3244(+)	flgG	Flagellar basal-body rod protein	0.389	0.175	Yes
SO3245(+)	flgF	Flagellar basal-body rod protein	0.320	0.188	Yes
SO3246		Hypothetical protein	0.434	0.413	No
SO3247(+)	flgE	Flagellar hook protein	0.269	0.136	Yes
SO3248(+)	flgD	Basal-body rod modification protein	0.276	0.066	Yes
SO3249(+)	flgC	Flagellar basal-body rod protein	0.300	0.109	Yes
SO3250(+)	flgB	Flagellar basal-body rod protein	0.142	0.053	Yes
SO3251(+)	cheR-3	Chemotaxis protein methyltransferase	0.609	0.188	Yes
SO3252(+)	cheV	Chemotaxis protein	0.556	0.113	Yes
SO3253(−)		Flagellar basal-body P-ring protein	0.257	0.071	Yes
SO3254(−)	flgM	Regulator of flagellin synthesis	0.307	0.164	Yes
SO3255(−)	flgN	Flagellar biosynthetic protein	0.190	0.086	Yes

^a+ and − represent the operon location on the positive or negative strand, respectively. Some non-motility-related genes are also included in the table to keep the entity of a complete operon.

^bThe average expression ratio of the salted sample to the control was calculated from 12 replicates together with the respective standard deviation (SD), a P = 0.05 standard t test result (significant [Sig.]).

Overall genomic expression profile of MR-1 in response to salt stress.

The overall genomic expression profiles indicated that the expression of a considerable subset of genes was affected during growth in the presence of 0.5 M NaCl. We identified a total of 518 genes (11.2% of the total gene content) as significantly upregulated and 598 genes (13%) as significantly down-regulated by a factor of 2 or more. According to the genome sequence annotations provided by The Institute for Genomic Research (http://www.tigr.org/), the majority of the up-regulated genes fell into the following functional categories: amino acid biosynthesis, protein synthesis, biosynthesis of cofactors, prosthetic groups, energy metabolism, and fatty acid/phospholipid metabolism. A large fraction of the most-highly-down-regulated genes were annotated as chemotaxis-related proteins. Similar to previous studies of microbial stress response (19, 23, 42, 53), changes in the expression of ORFs predicted to be involved with protein biosynthesis seem to play an important role in modulating cellular activities that allow adaptation to environmental stress (Table 1S).

Salt stress activated genes involved in Na⁺ efflux and K⁺ accumulation.

Na⁺ extrusion and replacement with K⁺ is the primary response of Escherichia coli to NaCl stress. To balance the large amounts of cation accumulation, E. coli will also accumulate glutamate (46). MR-1 appears to respond similarly to NaCl stress. First, genes encoding K⁺ uptake proteins were up-regulated, as well as Na⁺ efflux system components that included the Trk K⁺ uptake system, Na⁺/H⁺ antiporters, and Na⁺ efflux transporters (Table 1S). As expected, genes (SO1325 and SO4410) putatively involved in glutamate synthesis and a Na⁺/glutamate symporter gene (SO2923) were up-regulated in MR-1 by NaCl stress (Table 1S).

Besides the primary response, a secondary response (i.e., the accumulation of compatible osmolytes) may occur when a cell is subjected to salt concentrations of 0.5 M or higher, as observed in E. coli (46). Genes that encode the enzymes for trehalose and estoine biosynthesis, however, have not been identified in MR-1, and the corresponding compounds have not been reported. Sequence annotation of the MR-1 genome revealed two operons that contain proABC genes encoding enzymes for proline synthesis (SO1121, SO1122, and SO3354), but the expression of these genes was not significantly changed under the salt stress conditions examined (Table 1S). Interestingly, the accumulation of glycine betaine was observed in S. oneidensis cells grown in the presence of salted and smoked salmon (20). The authors stated that exogenous choline in the fish was transported and converted to glycine betaine (20). Therefore, MR-1 appears to have the ability to synthesize glycine betaine from choline. Generally, choline is first oxidized to glycine betaine aldehyde by the enzyme choline dehydrognase (BetA) in E. coli or by a type III alcohol dehydrogenase (GbsB) in Bacillus subtilis. The intermediate glycine betaine aldehyde is then further oxidized to glycine betaine by glycine betaine aldehyde dehydrogenase BetB in E coli or GbsA in B. subtilis (46). We identified two candidates (SO3496 and SO4480) for aldehyde dehydrogenase, one gene for type II (SO1490) and one gene for type III alcohol dehydrogenase (SO2054), but no candidates for choline dehydrogenase. These candidates, however, may function together to convert choline into glycine betaine in MR-1. The two putative alcohol dehydrogenase genes (SO3498 and SO4480) were slightly but not significantly up-regulated, and the other two aldehyde dehydrogenase genes (SO1490 and SO2054) were significantly down-regulated. It is therefore unlikely that glycine betaine biosynthesis was enhanced under the growth conditions tested.

Up-regulation of respiration-related genes.

Microarray analyses indicated that genes involved in both aerobic and anaerobic respiration were significantly up-regulated in salt-stressed MR-1 cells (Fig. 2 and Table 1S). The up-regulated genes involved in aerobic respiration included tricarboxylic acid (TCA) cycle enzymes and ATP synthase (SO4746 to SO4753), and ORFs predicted to play a role in anaerobic respiration included components of fumarate, nitrate, and nitrite reductases. Consistent with the activation of these enzymes, key genes involved in the biosynthesis of such cofactors as molybdopterin (2, 17, 50), heme (55), and menaquinone (31, 41) were also up-regulated (Table 1S). Up-regulated genes reported to be involved in fermentation were also observed, including formate dehydrogenase, quinone-reactive Ni/Fe hydrogenase, and acetate kinase.

FIG. 2.

TCA cycle and associated energy metabolic pathways. On the right side of gene symbols, the blue vertical bars denote no change in expression, whereas red upward arrows and black downward arrows denote significant up- or down-regulation in expression, respectively. Acetyl-CoA, acetyl coenzyme A.

Pyruvate can be respired either aerobically through the TCA cycle or anaerobically by formate dehydrogenation and fermentation. The pyruvate formate-lyase encoded by pflAB is the key enzyme that catalyzes pyruvate to formate (1), leading to the final products H₂ and CO₂. Significant up-regulation of the pflAB genes was observed, suggesting a possible redirection of pyruvate. At the same time, an operon that contained aconitase, methylcitrate synthase, methylisocitrate lyase, and a conserved hypothetical protein was up-regulated 6.7- to 9.1-fold. The apparent up-regulation of both aerobic and anaerobic respiration genes has also been reported for E. coli cells exposed to seawater for 20 h (40).

The glyoxylate bypass can reduce NADH production as well as allow a partial TCA to function to generate intermediates for anabolic reactions (e.g., amino acid biosynthesis) without the decarboxylation steps that result in loss of carbon (CO₂). The methylcitric acid pathway can provide additional energy from fatty acid and acetate catabolism. Apparently, the cell needs energy to survive the stress, but the aerobic respiration that can produce more energy may simultaneously generate extra reactive oxygen species as by-products (43), thus resulting in oxidative stress. This effect was observed in the moderately halophilic Shewanella sp. strain CN32, which requires 5 to 6% NaCl for optimal growth (4). Up-regulation of anaerobic respiration could help reduce oxidative stress to the cell. In addition, the cells may undergo clumping as a protective response to osmotic stress, as observed in Azospirillum brasilense (18) and Vibrio cholerae (51), and therefore may experience microaerophilic or anoxic conditions. However, the aggregation of MR-1 cells during salt stress was investigated as previously described (18), and significant aggregation was not detected for either the control or salt-stressed cells (data not shown). Observation of the cells by light microscopy also supported this conclusion. However, the cells were shaken during incubation, and significant clumping might have been prevented. Further work is needed to discern the possible connection between clumping and anaerobic metabolism.

Down-regulation of flagellar assembly genes impacted cell motility.

Phylogenetic analysis suggested that S. oneidensis flagellar motor proteins were more closely related to the sodium-driven motors in Vibrio species than to proton-driven motors. In addition, homologs of the MotAB and MotXY proteins, which are thought to be associated with sodium-driven motors, were present in the MR-1 genome (5). Notably, 47 of 49 flagellar assembly genes were repressed by the NaCl stress (Table 1). All flagellar assembly genes are located in region 4 except for the motor-encoding genes (Table 1). Apart from a few methyl-accepting chemotaxis protein (MCP) genes (less than 5% of the total MCP) that are dispersed throughout the genome, almost all chemotaxis-related genes were either significantly down-regulated or unaffected (Table 1 and Table 1S).

To test whether the observed down-regulation of chemotaxis-related protein genes indeed impacted cell motility, cell motility was qualitatively tested with soft agar inoculations. We prepared both solid (1% agar) and semisolid (0.3% agar) MR2A plates in combination with different salt concentrations for motility assessments. Cells (5 μl; OD₆₀₀ = 0.45) were applied to the center of the plate, the plates were cultivated at 30°C for 20 h, and the swarming behavior of the cells was observed. As expected, the cell motility was adversely affected under salt stress even at decreased NaCl concentrations (Fig. 3). These results indicated that down-regulation of flagellar assembly genes caused a decrease in motility, which agrees with previously reported observations for E. coli (21), B. subtilis (47), and Salmonella enterica serovar Typhimurium (34).

FIG. 3.

NaCl at a concentration of 0.3 M or higher completely halts MR-1 cell motility. The solid medium plates (1% agar) are designed to control possible different cell growth rates over varied NaCl concentrations as indicated, whereas the semisolid medium plates (0.3% agar) show the cell ability for motility.

Transcriptional regulation of flagellar and chemotaxis genes has been well studied (3, 4, 25) and has been documented in detail for bacteria of the Enterobacterales (9), Bucillaceae (47), and Vibrionaceae (27). Except for 28 MCP genes that are located in different operons, MR-1 has more than 60 flagellar assembly and other chemotaxis genes organized in at least 17 probable operons (Table 1). The operon organization of MR-1 flagellar ORFs most closely resembles that of V. cholerae. Maintenance of these large flagellar systems would seem to be a sizable investment with respect to cellular economy. In V. cholerae, the operons constitute a large, coordinately regulated flagelar regulon that is divided into three temporally regulated, hierarchical transcriptional levels: early, middle, and late (27). In V. cholerae, FlrA, acting as a σ⁵⁴-dependent transcription factor, activates transcription of flrBC, a two-component signal transduction system. The phosphorylation of FlrC by FlrB is required to activate middle-level flagellar genes (38), which includes most flagellar assembly genes, and fliA, which encodes a specialized sigma factor, σ²⁸. σ²⁸ activity controls transcription of the late-level genes like the flagellin, motor, and anti-sigma factor genes (27). Salt stress repressed the expression of flrA and flrC, the master transcriptional regulator genes in MR-1, leading to a complete shutdown of middle- and late-level flagellar assembly genes (Table 1). MR-1 may be similar to E. coli in terms of flagellar gene expression regulation, in which a promoter or promoters of the master operon flhDC receive a number of global regulatory signals, including the concentration of inorganic salt (9). The simultaneous detection of the whole-genomic expression patterns in response to a specific environmental stress can provide details about the possible connections between components in regulatory networks.

Concluding remarks.

The up-regulation of energy metabolism, including electron transport, and down-regulation of flagellar biosythesis in response to elevated salt conditions suggested that MR-1 needs more ATP to pump sodium out of the cell. In addition, an increase in electron transport may directly contribute to the efflux of sodium via the sodium-translocating electron transport complex I. Under high-salt conditions, MR-1 may repress the expression of flagellar genes to conserve energy necessary for sodium transport. The genomic expression profile of MR-1 in response to the sodium salt stress together with comparative genomics analyses indicated that MR-1 resembled responses observed in V. cholerae. As with Vibrio (10), a majority of Shewanella species reside in oceans, costal waters, and estuaries and were therefore more tolerant to sodium salt stress. More genomic similarities of MR-1 to V. cholerae clearly outline the connections between environments where the microorganisms naturally reside.