SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The osmolality required to activate osmosensory transporter ProP and the proportion of cardiolipin (CL) among the phospholipids of Escherichia coli rise with growth medium osmolality. Most CL synthesis has been attributed to the cls gene product. Transcription of cls increased with osmolality. The proportion of CL was low and osmolality-independent in cls bacteria. It increased more dramatically on the transition to stationary phase in cls than cls+ bacteria. Thus, Cls is responsible for osmoregulated CL synthesis and other enzymes may contribute to CL accumulation during stationary phase. The proportion of phosphatidylglycerol (PG) was elevated and it increased with medium osmolality in cls bacteria. A cls defect impaired growth of E. coli on solid and in liquid media at low and, more strongly, at high osmolality. Bacteria cultured at high osmolality without osmoprotectant were shorter and rounder than those cultured at low osmolality or with glycine betaine. Fluorescence microscopy showed that CL and ProP colocalize at the poles and near the septa of dividing E. coli cells. The polar localization of ProP was independent of its expression level but correlated with the proportion and polar localization of CL. Association with CL (and not PG) may be required for polar ProP localization.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Changes in extracellular osmotic pressure elicit transmembrane water fluxes that concentrate or dilute the cytoplasm of living cells, disrupting their structure and function. Cells respond by actively adjusting the distributions of selected solutes across the cytoplasmic membrane (Wood, 1999). Water follows, restoring cellular hydration and volume.

The osmoregulatory solutes used by bacteria include K+ and organic osmolytes that are also protein stabilizers (Wood, 1999; Bolen, 2001). In many organisms multiple, functionally redundant osmoregulatory systems adjust cellular solute content in response to osmotic stress (Wood, 2006). The ProP protein of Escherichia coli is denoted an ‘osmosensory transporter’ because it can sense increasing osmolality and respond by mediating the cytoplasmic accumulation of organic osmolytes, including proline, glycine betaine and ectoine (MacMillan et al., 1999; Racher et al., 1999). These compounds are osmoprotectants because, when provided exogenously, they stimulate bacterial growth in high (but not low) osmolality media.

For many halophilic and halotolerant bacteria, both Gram-positive and Gram-negative, the proportion of zwitterionic phospholipids [e.g. phosphatidylethanolamine (PE)] decreases and the proportion of anionic phospholipids [e.g. phosphatidylglycerol (PG), diphosphatidylglycerol, known as cardiolipin (CL)] increases with growth in media of increasing salinity (Russell, 1993; López et al., 1998; Brown et al., 2000; 2003; Trotsenko and Khmelenina, 2002; Rilfors and Lindblom, 2002; Machado et al., 2004; Danevcic et al., 2005; Vargas et al., 2005; López et al., 2006). An increase in the cyclopropane fatty acid content of Gram-negative bacteria is the most frequently observed alteration to fatty acid composition (Russell, 1989). In a few cases, analogous changes have been shown to occur when salts or non-ionic osmolytes are used to create growth media of the same osmolality (Wood, 1999). Growth of Bacillus subtilis in a saline medium is impaired by a CL deficiency (López et al., 2006). However, the physiological contributions of these phospholipid changes to bacterial osmoadaptation have not been fully elucidated.

The cytoplasmic membrane of E. coli includes approximately 75% PE, 20% PG and 5% CL (Cronan, 2003). The proportion of CL in the membrane increases as E. coli enters stationary phase or if energy metabolism is impaired by treatment with colicin K, dinitrophenol, penicillin or cyanide or by infection with bacteriophage T4 (Cronan and Vagelos, 1972; Cronan, 2003). We found that the proportion of CL in the cytoplasmic membrane and the osmolality required to activate osmosensory transporter ProP increased in parallel, as the proportion of PE decreased, when E. coli was cultivated in defined media of increasing osmolality (0.1–0.7 mol kg−1) (Tsatskis et al., 2005).

The cls locus of E. coli encodes a well-characterized CL synthase which catalyses the condensation of two molecules of PG to yield CL and glycerol (Tropp, 1997). E. coli and B. subtilis cells with defects at their homologous cls loci are CL-deficient (Cronan, 2003; López et al., 2006). Other enzymes may contribute to CL synthesis as the genomes of both organisms include cls homologues and, at least in E. coli, cls bacteria retain traces of CL (Cronan, 2003). Despite important, well-documented roles of CL in DNA replication and cell division (Mileykovskaya and Dowhan, 2005), the previously reported impacts of CL deficiency on E. coli include only slightly reduced growth in standard media and reduced survival during stationary phase (Cronan, 2003). In this article we show that cls transcription is osmoregulated, that the proportion of PG rises in cls bacteria in response to osmotic stress, and that a pathway other than cls may contribute to stationary-phase CL synthesis. Furthermore, a cls defect reduces the osmotolerance of E. coli in the absence and presence of osmoprotectant glycine betaine.

Cardiolipin is enriched in regions of the cytoplasmic membrane near the nucleoid-free poles and septa of growing E. coli (Mileykovskaya and Dowhan, 2000) and in E. coli minicells (nucleoid-free cells resulting from polar cell division) (Koppelman et al., 2001). Many E. coli proteins are located at the cell poles, including chemotactic receptors and lactose-H+ symporter LacY (Lai et al., 2004; Nagamori et al., 2004; Matsumoto et al., 2006). Variations in protein–lipid affinity or interactions with localized cytoplasmic constituents may promote these asymmetries, but no localization mechanism has been demonstrated. Here we show that osmosensory transporter ProP colocalizes with CL at the poles of E. coli cells. Thus, although CL comprises a small proportion of total cellular lipid, ProP is embedded in a CL-enriched membrane environment. Polar localization of ProP is less pronounced in cls than in cls+ bacteria, implying that CL–ProP interactions promote the polar localization of ProP.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The CL synthase encoded by cls mediates CL accumulation in response to osmotic stress

Most CL synthesis in E. coli has been attributed to the CL synthase encoded by cls (Cls). However, residual CL is present in cls bacteria and that residue increases as cls cells enter stationary phase (Nishijima et al., 1988). The E. coli genome includes cls homologues ybhO and ymdC (Guo and Tropp, 2000). YbhO and YmdC are predicted to include a sequence motif characteristic of phospholipase D superfamily members like Cls. YbhO has CL synthase activity in vitro but neither homologue has been shown to contribute to CL synthesis in vivo (Guo and Tropp, 2000).

Expression of cls was monitored by measuring the β-galactosidase activities of bacteria harbouring cls::lacZ operon and protein fusions available in E. coli strains SOH92 and SOH93 respectively (Heber and Tropp, 1991). These fusions are present within lysogenic λ transducing phage in cis with the chromosomal cls locus. The β-galactosidase activity of bacteria with a cls::lacZ operon fusion increased nearly threefold as they were cultivated in media supplemented with NaCl or sucrose to attain osmolalities in the range 0.1–0.8 mol kg−1 (Fig. 1). A similar fold induction was observed with strain SOH93, although the β-galactosidase activities were lower as reported previously (Heber and Tropp, 1991) (data not shown). Thus, expression of cls is osmoregulated and the CL synthase encoded by cls may mediate osmoregulated CL synthesis.

image

Figure 1. Osmotic induction of cls expression. E. coli strain SOH92 [Φ(clslacZ+)] was cultivated to late exponential phase in MOPS media adjusted to the indicated osmolalities with NaCl (circles) or sucrose (squares) and the β-galactosidase activities of the harvested bacteria were measured as described in Experimental procedures.

Download figure to PowerPoint

Bacteria harbouring mutation cls::Tn10dTet3 (Heber and Tropp, 1991) were used to determine whether Cls is the only contributor to CL accumulation under osmotic stress. This Tn10dTet insertion was first isolated by screening narC+ transductants obtained from a Tn10dTet3 mutant pool for 3,4-dihydroxybutyl-1-phosphonate (DHBP) resistance, then shown to confer CL deficiency (cls had been linked to narC and other cls mutations had conferred DHBP resistance on E. coli). However, this insertion had not been shown to directly disrupt cls. The polymerase chain reaction (PCR) was used to locate this insertion near the cls initiation codon (Table 1). Sequencing of the amplicon obtained with primers cls-03 and Tn10dTet-1 (Table 1) revealed that Tn10dTet3 had inserted between the R31 and R32 codons of cls.

Table 1.  Location of cls::Tn10dTet3.
Primer pairaPredicted ampliconb
WG709 cls+WG983 cls::Tn10dTet3
  • a.

    The primer sequences (5′−3′) were cls-01 (CCCACTTCCGT TCTACTCCGC), cls-02 (GATCGAGATTGTCGGCAGCC), cls-03 (CCATATTCATGCTGCCGGTG), cls-04 (CACCGGCAGCATGAATATGG) and Tn10dTet-1 (CTTTCTAAGGCAGACCAACC).

  • b.

    Amplicons of these sizes (bp) would be expected with genomic DNA from E. coli WG709 (cls+) or WG983 (cls::Tn10dTet3) as template and the indicated primers. With the exception of those starred (*), amplicons of the expected sizes were obtained when PCR was performed as described in Experimental procedures. The starred amplicons were too long to amplify under our reaction conditions so no amplicon was observed.

cls-01/cls-0217574674*
cls-01/cls-038953812*
cls-02/cls-04882882
Tn10dTet-1/cls-02Nil1722
Tn10dTet-1/cls-03Nil860

For exponential-phase bacteria, the proportion of CL among the phospholipids of E. coli WG983 (cls::Tn10dTet3, data points and solid lines in Fig. 2) was lower than that of its cls+ parent strain and independent of growth medium osmolality. In Fig. 2, the dotted lines represent the phospholipid proportions of parent strain WG709, reported previously (Tsatskis et al., 2005). For exponential-phase bacteria, the average proportion of CL in WG983 (cls) was 0.4 mole % whereas that of WG709 (cls+) varied with osmolality from 2.3 to 8.3 mole %. Thus, the CL synthase encoded by cls mediates osmoregulated CL synthesis. The PG content of the cls bacteria was significantly higher than that of the cls+ parent strain, as previously observed (Pluschke et al., 1978; Nishijima et al., 1988), and it increased with osmolality, at the expense of PE (Fig. 2). Thus, the synthesis of PG, a CL precursor, can also be directly or indirectly modulated by growth medium osmolality.

image

Figure 2. Phospholipid headgroup composition of cls bacteria. E. coli strain WG983 (cls) was cultivated to late exponential phase in NaCl-supplemented MOPS media of the indicated osmolalities and its phospholipid composition was determined as described in Experimental procedures. The symbols represent the mole per cent PE (circles), PG (squares) or CL (triangles) determined in duplicate experiments and the solid lines were obtained by linear regression of the resulting data. The dotted lines represent the phospholipid compositions of the cls+ parent strain, WG709, determined by Tsatskis et al. (2005).

Download figure to PowerPoint

The proportions of CL and PE among the phospholipids of E. coli are known to increase during stationary phase, while the proportion of PG decreases, although the magnitude of these changes is E. coli strain-dependent (Nishijima et al., 1988). Heber and Tropp (1991) showed that cls expression was induced 2.5-fold as E. coli entered stationary phase in low-osmolality LB medium. The effects of osmotic stress and the transition to stationary phase on the proportion of CL were not additive (Table 2). For this cls+ strain, the proportion of CL rose almost twofold on entry to stationary phase in low-osmolality medium, whereas the proportion of CL in bacteria grown at high osmolality was independent of growth phase. Feedback inhibition of CL synthase may have imposed a limit on CL accumulation at high osmolality (Ragolia and Tropp, 1994). The fold increase in CL as cells entered stationary phase was greater for cls bacteria at low or high osmolality (six- and fourfold respectively) than for cls+ bacteria at low osmolality (1.3-fold), suggesting that an additional enzyme or pathway contributed to CL synthesis in stationary phase. These interpretations are based on the assumption that the total quantity of lipid per cell did not vary significantly with the imposed growth conditions, and that changes in proportion of CL resulted from changes in the synthesis of CL, not changes in the synthesis of PG or PE.

Table 2.  Impact of cls::Tn10dTet3 on the phospholipid headgroup composition of E. coli.a
Growth phaseOsmolalityProportions of phospholipids (mole %)b
PEPGCL
cls + cls cls + cls cls + cls
  • a.

    E. coli strain WG709 (cls+) or WG983 (cls) was cultivated to exponential or stationary phase in MOPS medium adjusted to low osmolality (0.15 mol kg−1) or high osmolality (0.74 or 0.75 mol kg−1) with NaCl and the phospholipid composition was determined as described in Experimental procedures.

  • b.

    Values are cited as the mean ± range of two determinations with the exception of exponential-phase cls+ bacteria. Those values were determined only once, and they are comparable to estimates reported previously (see data reported by Tsatskis et al., 2005 and reproduced in Fig. 2).

ExponentialLow78.774.9 ± 0.617.524.8 ± 0.63.80.3 ± 0.01
StationaryLow90.5 ± 3.284.6 ± 2.1 3.1 ± 1.613.6 ± 1.46.4 ± 1.81.8 ± 0.8
ExponentialHigh73.869.8 ± 2.718.029.6 ± 2.88.30.6 ± 0.1
StationaryHigh86.4 ± 3.277.9 ± 2.7 5.0 ± 0.419.7 ± 3.08.6 ± 3.02.4 ± 0.3

Impacts of CL deficiency, osmolality and glycine betaine on bacterial growth and morphology

Growth of cls+ and clsE. coli strains in low- and high-osmolality media, with and without the osmoprotectant glycine betaine, was measured by determining colony numbers and diameters on NaCl-supplemented MOPS medium plates (Neidhardt et al., 1974) (Fig. 3) and by monitoring the optical densities (ODs) of corresponding liquid cultures (Fig. 4). The liquid media contained no salt (osmolality 0.15 mol kg−1) or 0.5 M NaCl (osmolality 1.0 mol kg−1). NaCl (50 mM) is included in standard MOPS medium to maximize the growth rate of E. coli (Neidhardt et al., 1974). However, NaCl-free MOPS was used for these experiments to maximize the osmolality range tested.

image

Figure 3. Impacts of CL deficiency, osmolality and glycine betaine on the growth of E. coli on solid medium. E. coli strains WG709 (cls+, left bar of each pair, labelled ‘+’) and WG983 (cls, right bar of each pair, labelled ‘–’) were cultivated overnight in LB medium (Miller, 1972); cells were harvested and re-suspended in saline (0.85% NaCl, w/v). Aliquots of appropriate dilutions in saline were spread on MOPS medium without or with NaCl (up to 0.6 M) and without or with glycine betaine (1 mM). The diameters of 10 colonies on each plate were recorded after 2, 3 and 4 days' incubation at 37°C. Scanned images of the plates were displayed with Paint Shop Pro (Corel) and the colony diameters in pixels, at a resolution of 600 dpi, were measured with the ruler. To convert the lengths in pixels to lengths in μm, each value was multiplied by the ratio of the diameter of the plate bottom measured with a ruler, to the diameter of the plate bottom measured with Paint Shop Pro. Representative data from one of two replicate experiments are reported as means ± standard deviations.

Download figure to PowerPoint

image

Figure 4. Impacts of CL deficiency, osmolality and glycine betaine on the growth of E. coli in batch culture. E. coli strains WG709 (cls+, solid circles) and WG983 (cls, open circles) were cultivated overnight in NaCl-free MOPS medium, then used to inoculate MOPS medium without NaCl (0.15 mol kg−1) or with NaCl (500 mM, 1 mol kg−1) and without or with glycine betaine (1 mM) to an optical density (550 nm) of 0.1. The optical density (550 nm) was then monitored as the resulting cultures were incubated at 37°C with rotary shaking at 200 r.p.m. Averages of duplicate measurements from a representative experiment are shown and each experiment was performed at least twice.

Download figure to PowerPoint

The cls mutation reduced colony size on solid medium and slowed bacterial growth in liquid medium, its impact being greatest at high osmolality (Figs 3 and 4). A similar effect was observed by others using only low-osmolality, liquid media (Pluschke et al., 1978; Nishijima et al., 1988). The cultures prepared in high-osmolality liquid medium without glycine betaine reached a higher OD than those prepared in the other media (Fig. 4), implying that the application of osmotic stress in the absence of osmoprotectants had enhanced growth yield. In fact, viable counts performed using LB plates showed that the cell population densities of the high-osmolality cultures without glycine betaine were lower than or similar to those of the other cultures. For example, the final cell population densities for the cultures represented in Fig. 4 were as follows (means of duplicates). For strain WG709 (cls+), the cultures in high-osmolality medium without and with glycine betaine attained mean population densities of 1.5 × 109 (corresponding to a mean OD of 4.8) and 2.1 × 109 (corresponding to a mean OD of 3.1) respectively. For strain WG983 (cls), the corresponding values were 1.2 × 109 (corresponding to a mean OD of 4.8) and 1.4 × 109 (corresponding to a mean OD of 3.1) respectively. It is possible that the presence of dead cells reduced the titres of the cultures of the cls strain relative to those of the cls+ bacteria, despite their similar ODs.

Bacteria in samples drawn from representative liquid cultures were examined to determine whether differences in cell morphology might account for the different ODs. Neither light microscopy nor fluorescence microscopy of cells treated with membrane stain FM4-64 [N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl) hexatrienyl)pyridinium dibromide] revealed morphological differences between cls+ and cls bacteria. However, bacteria from the high-osmolality cultures lacking glycine betaine were shorter and rounder than those from the low-osmolality cultures and from all cultures containing glycine betaine (Fig. 5, top). Measurements of their lengths and widths showed that bacteria subjected to osmotic stress without osmoprotectant were shorter and wider than those grown at low osmolality, or at high osmolality with betaine (Fig. 5, bottom panels). These shorter, rounder cells may have elevated the ODs of cultures prepared in high-osmolality, betaine-free medium by scattering light more powerfully than their more rod-shaped counterparts.

image

Figure 5. Impacts of CL deficiency, osmolality and glycine betaine on the morphologies of E. coli cells in batch culture. Samples were withdrawn from cultures analogous to those illustrated in Fig. 4 when the optical density (550 nm) had reached 1. Bacteria in each sample were stained with fluorescent dye FM4-64 and visualized with a fluorescence microscope as described in Experimental procedures. Top panel: Representative images of E. coli WG709 cells (cls+). These cells were indistinguishable in morphology from E. coli WG983 cells (cls) cultivated under the same conditions. Bottom panels: Mean lengths (black bars) and widths (grey bars) for 100 bacteria of each strain cultivated under each of the indicated conditions (error bars represent standard deviations).

Download figure to PowerPoint

Subcellular distributions of CL and ProP

The proportion of cardiolipin and its distribution in cls+ and cls bacteria were visualized by staining cells with DAPI (4′,6-diamidino-2-phenylindole) and NAO (10-N-nonyl acridine orange) (Fig. 6, left). DAPI stains nucleoids, yielding blue fluorescence. NAO is a membrane stain that binds anionic but not zwitterionic phopsholipids, binding with a higher stoichiometry and affinity to CL than to other anionic phospholipids (Petit et al., 1992). It yields green fluorescence when CL is present at low levels and a characteristic red, excimer fluorescence when associated with CL oligomers in CL-rich membranes (Mileykovskaya et al., 2001). Red fluorescence was detected when NAO was applied to cls+ bacteria (all culture conditions), or to cls bacteria cultivated to stationary phase in high-osmolality medium (Fig. 6, left). These cells contained 2.4–8.6 mol % CL (Table 2). Only green fluorescence was observed when NAO was applied to cls bacteria that had been cultivated to exponential phase or to stationary phase at low osmolality (Fig. 6, left). These cells contained less than 2 mol % CL. In all cells except those with only traces of CL (exponential phase, cls bacteria), NAO fluorescence was concentrated at the cell poles and septa.

image

Figure 6. Colocalization of cardiolipin and ProP at the bacterial poles and septa. E. coli strains WG994 (cls+) and WG1008 (cls), both expressing MVCCPGCC-ProP, were cultivated to exponential phase [an optical density (600 nm) of 1] or stationary phase (overnight) in low- or high-osmolality MOPS medium (0.15 or 0.7 mol kg−1) without arabinose. Cells were stained with DAPI (nucleoids, blue fluorescence) and NAO (CL, left panel, green or red fluorescence) or FlAsH-EDT2 (MVCCPGCC-ProP, right panel, green fluorescence) and fluorescence was recorded as described in Experimental procedures. Red NAO fluorescence, indicating high levels of CL, is shown for those preparations in which it was observed (cls+ bacteria cultivated under all conditions and cls bacteria cultivated in high-osmolality medium to stationary phase). Green NAO fluorescence is shown where no red NAO fluorescence was observed (cls bacteria cultivated at low osmolality or to exponential phase in high-osmolality medium).

Download figure to PowerPoint

FlAsH-EDT2 (Fluorescein Arsenical Helix binder, bis-EDT adduct) is a biarsenical fluorescein derivative that emits green fluorescence only after reacting with the helical CCXXCC motif in target proteins (Adams et al., 2002). A plasmid encoding a ProP variant with the FlAsH tag VCCPGCC inserted after the initiating methionine was created to support visualization of ProP using FlAsH-EDT2 (see Experimental procedures). To avoid regulation of their expression by osmolality and growth phase (characteristic of the native proP promoters), the genes encoding wild-type and FlAsH-tagged ProP were expressed from the araBAD promoter in plasmid pBAD24. Without induction, FlAsH-tagged ProP was expressed at a lower level than wild-type ProP. However, the levels of the two ProP variants could be made equal by inducing expression of FlAsH-tagged ProP with arabinose (Fig. 7, inset). The osmotic activation profile of FlAsH-tagged ProP matched that of the wild-type transporter (Fig. 7, note that relative uptake rates are plotted). As expected, the cellular levels of wild-type and FlAsH-tagged ProP were independent of growth medium osmolality and growth phase in both cls+ and cls bacteria (Fig. 8).

image

Figure 7. Activities and expression levels of MVCCPGCC-ProP and wild-type ProP. Bacteria were cultivated in low-osmolality MOPS medium (0.15 mol kg−1), proline uptake assays were conducted and Western blots were performed as described in Experimental procedures. Relative uptake activities (a0/Amax) are shown and the regression lines were obtained by fitting each data set to the equation a0/Amax = {1 + exp[–(Π − Π1/2)/(RTB)]}−1 by non-linear regression. The Amax and Π1/2/RT values were: 63.9 ± 0.7 nmoles min−1 (mg protein)−1 and 0.216 ± 0.002 mmol kg−1 for wild-type ProP in cells without arabinose induction (circles), 7.4 ± 0.3 nmoles min−1 (mg protein)−1 and 0.197 ± 0.007 mmol kg−1 for MVCCPGCC-ProP in cells without arabinose induction (triangles), and 39.5 ± 0.4 nmoles min−1(mg protein)−1 and 0.203 ± 0.002 mmol kg−1 for MVCCPGCC-ProP in cells induced with 33 μM arabinose (squares).

Download figure to PowerPoint

image

Figure 8. Expression of ProP and MVCCPGCC-ProP in cls+ and cls bacteria as a function of growth medium osmolality and growth phase. Proteins expressed by E. coli were detected by Western blotting after cultivation as described in Experimental procedures and the legend for Fig. 6. The E. coli strains employed for these experiments were WG709 (cls+) and WG983 (cls), each expressing wild-type ProP, as well as WG994 (cls+) and WG1008 (cls), each expressing MVCCPGCC-ProP.

Download figure to PowerPoint

When cells expressing FlAsH-tagged ProP were labelled with FlAsH-EDT2, green fluorescence was found at the cell poles and septa. No fluorescence was observed when cells of E. coli WG709, expressing untagged, wild-type ProP, were examined in the same way. The same pattern of ProP localization was observed in cells without arabinose induction (Fig. 6, right) or after full arabinose induction (cultures supplemented with 13.3 mM arabinose). Interestingly, the polar/septal localization of ProP was strongest when the proportion of CL was high and CL was also concentrated at the poles and septa, as indicated by phospholipid headgroup analysis (Table 2) and CL visualization (Fig. 6, left) respectively. Thus, the localization of FlAsH-tagged ProP, and not its expression, was impaired in CL-deficient bacteria.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The following evidence indicates that at least two pathways contribute to CL synthesis by E. coli. Transcription of the cardiolipin synthase encoded by cls is enhanced threefold as bacteria are cultivated in media of increasing osmolality (0.1–0.7 mol kg−1) (Fig. 1) and insertion cls::Tn10dTet3, which interrupts the cls open reading frame after the 31st codon, renders bacterial CL levels osmolality independent during both exponential and stationary-phase growth (Fig. 2, Table 2). Despite the impact of the cls lesion on CL synthesis, exponential-phase bacteria retained a trace of CL (approximately 0.4 mol % lipid). Further, a more pronounced increase in proportion of CL occurred on the transition to stationary phase for cls bacteria (four- to sixfold) than for cls+ bacteria (up to 1.3-fold) (Table 2). Nishijima et al. (1988) suggested that the trace of CL found in exponential-phase cls cells may be produced by phosphatidyl group transfer from CDP-diacylglycerol to PG catalysed by phosphatidylserine synthase. That enzyme or the CL synthase homologues YbhO (which has CL synthase activity in vitro; Guo and Tropp, 2000) or YmdC could contribute to stationary-phase CL synthesis.

The changes in proportion of CL discussed above were accompanied by quantitatively larger, complementary changes in proportion of PG. PG levels were not osmoregulated in cls+ bacteria (Tsatskis et al., 2005; Fig. 2, Table 2), but the PG/PE ratio was higher under all conditions in cls than in cls+ bacteria and the proportion of PG in the membranes of cls bacteria rose with growth medium osmolality (Fig. 2, Tables 2 and 3). As CL is made from PG, these observations suggest that the increased proportion of CL in cls+ bacteria under osmotic stress results from increases in both PG synthesis and PG conversion to CL. Each cardiolipin molecule contributes two phosphate moieties to the membrane surface whereas PE and PG contribute only one. Thus, the net effect of these changes was to minimize the difference in the proportion of the headgroup phosphate moieties present in anionic lipid (and hence the membrane surface charge) between cls+ and cls bacteria (Table 3). Impacts of the cls defect on cell physiology must be interpreted in the context of these compensatory changes.

Table 3.  Maintenance of membrane surface charge in E. coli.a
Growth phaseOsmolalityPG/PE mole ratioAnionic headgroups (%PG + %CL) (mol %)Phosphate in anionic headgroupsb (mol %)
cls + cls cls + cls cls + cls
  • a.

    Values are calculated from the data in Table 2.

  • b.

    The per cent of phosphate in anionic headgroups is %PG + 2 %CL/%PE + %PG + 2 %CL.

ExponentialLow0.220.3321252425
StationaryLow0.030.1610151517
ExponentialHigh0.240.4226303231
StationaryHigh0.060.2514222024

The growth of E. coli in liquid medium is inhibited by osmotic up-shocks and recovery is contingent on the availability of organic osmoprotectants. In their absence, K+ is taken up and glutamate may accumulate as a K+ counterion. When present, osmoprotectants are taken up to accumulate or be converted to compatible solutes (solutes that can accumulate to high cytoplasmic levels without disrupting cellular functions). K+ accumulation does not restore cellular hydration, turgor pressure or bacterial growth to pre-stress levels. Osmoprotectant uptake and organic osmolyte accumulation result in more complete rehydration and fully restore growth (Wood, 1999; Cayley et al., 2000; Cayley and Record, 2003).

The experiments reported here employed E. coli strains in which osmosensory transporter ProP, alone, mediates uptake of the osmoprotectant glycine betaine. The cls defect impaired their growth under all conditions tested and it had the strongest effects during growth on high-osmolality media (Figs 3–5). A cls defect also impaired the growth of B. subtilis in high-osmolality media (López et al., 2006). Thus, CL is important for the growth of E. coli, particularly under osmotic stress, and PG cannot fully replace it. The cls defect affected growth with and without glycine betaine similarly. Thus, ProP retained the ability to mediate osmoprotection in cls cells but other membrane enzymes or transporters were adversely affected by the CL deficiency and compensatory PG accumulation.

Bacteria (cls+ or cls) cultivated in high-osmolality liquid medium without osmoprotectant were shorter and rounder than those cultured at low osmolality or in the presence of glycine betaine (Fig. 5). The shorter, rounder bacteria appeared to scatter light more powerfully than their longer counterparts (Figs 4 and 5). Light scattering increases transiently after E. coli cells are subjected to an osmotic upshift (Meury, 1994) and osmoprotectant transporter-deficient E. coli scatter light more powerfully than those able to transport the osmoprotectants available during growth in high-osmolality human urine (Culham et al., 1998). The enhanced ability to scatter light may arise as the cells' refractive index increases and their shape changes upon dehydration. These observations all reflect the cells' failure to regain normal morphology and hydration if they cannot accumulate compatible solutes. They show that care must be taken when using OD measurements to estimate the growth of bacteria under osmotic stress.

In this study, FlAsH labelling of ProP and NAO labelling of CL were used to demonstrate that they colocalize at the poles of E. coli and near the septa of dividing cells (Fig. 6). FlAsH-tagged ProP is likely to report the localization of the wild-type transporter as the FlAsH-tagged variant was fully functional (Fig. 7) and its polar localization was independent of its expression level. The same polar localization was seen when cells were programmed to express FlAsH-tagged ProP at levels below that of wild-type ProP (expression from vector pBAD24 without arabinose induction, Fig. 6) or at levels greatly above that of wild-type ProP (fully arabinose-induced expression from the pBAD24 vector, data not shown).

Further, polar localization of ProP and polar localization of CL were correlated (Fig. 6). Polar localization of CL was particularly evident in bacteria for which CL constituted 2.4% or more of membrane lipid (Table 2 and Fig. 6). Polar localization of FlAsH-tagged ProP was also particularly evident in bacteria with a high proportion of CL (Fig. 6), even though the overall, cellular levels of untagged or FlAsH-tagged ProP did not vary with the proportion of CL, medium osmolality or growth phase (Fig. 8). Thus, association with CL may be required for polar ProP localization and elevation of PG is not sufficient to substitute for CL in this regard. To our knowledge, this is the first evidence that polar localization of an E. coli membrane protein is CL-dependent.

Other energy-transducing enzymes are concentrated at the poles of E. coli cells, including H+-lactose symporter LacY (a ProP homologue) (Nagamori et al., 2004) and ATPase subunits A and B (Lai et al., 2004). In yeast, CL associates with and is required by respiratory chain complexes and the ADP/ATP antiporter (Mileykovskaya et al., 2005). Haines and Dencher (2002) proposed that CL mediates proton transfer among energy-conserving complexes. CL may also play a special role in the energy-transducing cytoplasmic membrane of E. coli. Osmotic stress impairs respiration in E. coli (Houssin et al., 1991), and this work adds osmotic stress to the list of treatments, all impairing energy metabolism, that increase the proportion of CL in the membrane (Cronan and Vagelos, 1972). Perhaps impaired energy metabolism signals a need for elevated CL, and elevating the proportion of CL can improve the efficiency of energy metabolism.

This study was inspired by our observation that the osmolality required to activate ProP correlates with the CL content for bacteria cultivated in media with diverse osmolalities (Tsatskis et al., 2005). Our characterization of cls bacteria revealed that PG synthesis is also osmoregulated in E. coli, so cls bacteria cannot be used to independently assess the impacts of CL content and PG content on ProP. Future studies will be designed to resolve those effects, thereby testing the hypothesis that CL is a key structural and functional partner for ProP.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and growth conditions

The following E. coli strains were used for this study. E. coli SOH92 harbours a λ phage-encoded clslacZ operon fusion [Φ(clslacZ+)] and E. coli SOH93 harbours a λ phage-encoded clslacZ protein fusion [Φ(cls–‘lacZ)hyb]. Both strains are λ lysogens of E. coli P90C [araΔ(lac–proAB) thi] obtained from Burton Tropp (City University of New York) (Heber and Tropp, 1991). WC3899 (FnadB7 supE rfbD1 cLs::Tn10dTet3), obtained from Dr W. Dowhan (University of Texas), is a tetracycline-resistant P1 transductant of W3899 (FnadB7 supE rfbD1) from SOH9 (Heber and Tropp, 1991). WG350 [Ftrp lacZ rpsL thiΔ(putPA)101Δ(proU)600Δ(proP-melAB)212] lacks all known proline transporters of E. coli (Culham et al., 1993). WG980 was isolated as a tetracycline-resistant P1 cml clr100 transductant of WG350 from WC3899 that harbours the cls::Tn10dTet3 insertion. Plasmid pDC79 (Culham et al., 2000), encoding wild-type ProP, was introduced to WG350 and WG980 to create strains WG709 and WG983 respectively. MVCCPGCC-ProP is encoded by plasmid pDC232 which was created by site-directed mutagenesis of pDC79 as previously described (Culham et al., 2003a). Plasmid pDC232 was introduced to strains WG350 and WG980 to create strains WG994 and WG1008 respectively. In these plasmids, proP and its derivatives are expressed from the araBAD promoter of vector pBAD24 (Guzman et al., 1995). Basic molecular biological techniques were as described by Sambrook et al. (1989) and the PCR was carried out as described by Brown and Wood (1993).

The bacteria were cultivated in LB (Miller, 1972) or in NaCl-free MOPS medium (Neidhardt et al., 1974) with glycerol [0.4% (v/v)] as carbon source, NH4Cl (9.5 mM) as nitrogen source, tryptophan (245 μM) and thiamine (1 mg ml−1) to meet auxotrophic requirements and ampicillin (100 μg ml−1) to maintain plasmids. NaCl or sucrose was added as indicated to adjust the osmolality and osmolalities were measured with a Wescor vapour pressure osmometer (Wescor, Logan, UT, USA). To monitor growth in liquid medium, cultures (24 ml) were prepared in 125 ml sidearm flasks and incubated at 37°C with rotary shaking at 200 r.p.m. ODs were monitored with a Bausch and Lomb Spectronic 88 spectrophotometer.

β-Galactosidase assays

The β-galactosidase activities of E. coli strains SOH92 and SOH93 were determined as previously described (Wood et al., 2005) with the following modifications. Bacteria were cultivated to late exponential phase (OD of 0.8 at 550 nm) in MOPS media of the indicated osmolalities, without IPTG (isopropyl-β-d-thiogalactopyranoside), as described for transport assays (Culham et al., 2003b). The protein contents of the bacterial suspensions were determined with the BCA assay (Smith et al., 1985). Each experiment (four replicates per sample) was repeated at least twice and representative data are expressed as nmoles o-nitrophenylgalactoside hydrolysed per minute per mg of cell protein.

Transport assays and determination of ProP protein levels by Western blotting

Published procedures were used to cultivate bacteria in NaCl-free MOPS medium, to measure proline uptake (Culham et al., 2003b) and to determine the expression levels of ProP and FlAsH-tagged ProP in intact bacteria by Western blotting (Culham et al., 2000). Transport assay media were adjusted to the indicated osmolalities with NaCl. All transport measurements were performed in triplicate. Figure 7 shows the means and standard errors of the mean for representative triplicate assays.

Phospholipid analysis

The proportions of CL, PG and PE in exponential-phase bacteria were determined as described previously (Tsatskis et al., 2005), with the exception that 33P-orthophosphate was used in place of 32P-orthophosphate. The proportions of the phospholipids in stationary-phase bacteria were determined by following the same procedure but allowing cultures to incubate overnight (24 h) before the bacteria were harvested and phospholipids were extracted.

Fluorescence microscopy

Fluorescent dyes were obtained from Molecular Probes (Eugene, OR). Membranes were stained by adding FM 4-64 [N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl) hexatrienyl)pyridinium dibromide] to 0.1 μM. To stain CL and cell nucleoids, samples from liquid cultures were made 0.2 μM in NAO [10-N-nonyl-3,6-bis (dimethylamino)acridine] and incubated for 1 h at 37°C with rotary shaking at 200r.p.m. DAPI (4′,6-diamidino-2-phenylindole) was added to 1 μg ml−1 for the last 5 min. To stain FlAsH-tagged ProP and cell nucleoids, samples from liquid cultures were made 100 μM in FlAsH-EDT2 and 0.5 mM in BAL (2,3-dimercapto-1-propanol, to suppress labelling of endogenous cysteine pairs), then incubated for 3 h at 37°C with rotary shaking. DAPI was added to 1 μg ml−1 for the last 5 min. Theseadditions were timed so that cells could be stained and then their fluorescence could be examined at a culture OD of 1 (exponential phase) and after approximately 24 h growth (stationary phase). Incubation with NAO did not inhibit bacterial growth. After staining, cells in 4 μl of each sample were immobilized on object slides coated with a flat layer of agarose [40 μl of 2% agarose (w/v) in the corresponding medium].

Cells were viewed in the presence of the dyes with an Imaging RetigaEX CCD camera mounted on an Axiovert 200M inverted fluorescence microscope (Carl Zeiss Microimaging) equipped with a Zeiss Plan Neofluor 100× oil NA1.3 objective. The fluorescence was excited and detected with a 100 W halogen lamp (N HBO 103) in combination with different filter sets: for blue fluorescence (excitation filter BP365/12 nm, dichroic mirror FT395, emission filter LB397), for green fluorescence (excitation filter BP470/40 nm, dichroic mirror FT495, emission filter LB525/50), and for red fluorescence (excitation filter BP560/40 nm, dichroic mirror FT585, emission filter BP630/75). Images were obtained and processed using Openlab (Improvision).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank William Dowhan (University of Texas) and Burton Tropp (Queens College, City University of New York) for E. coli strains, Joseph Lam (University of Guelph) for the use of the Axiovert 200M inverted fluorescence microscope and RetigaEX CCD camera and Zoya Ignatova (Max Planck Institute for Biochemistry, Martinsreid) and Lila Gierasch (University of Massachusetts, Amherst) for introducing us to FlAsH labelling. This research was supported by Research Grant OPG0000508, awarded to J.M.W., and Undergraduate Student Research Awards, given to C.G. and L.S., by the Natural Sciences and Engineering Research Council of Canada.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Adams, S.R., Campbell, R.E., Gross, L.A., Martin, B.R., Walkup, G.K., Yao, Y., et al. (2002) New biarsenical ligands and tetracysteine motifs for protein labeling in vitro and in vivo: synthesis and biological applications. J Am Chem Soc 124: 60636076.
  • Bolen, D.W. (2001) Protein stabilization by naturally occurring osmolytes. Methods Mol Biol 168: 1736.
  • Brown, G.R., Sutcliffe, I.C., Bendell, D., and Cummings, S.P. (2000) The modification of the membrane of Oceanomonas baumannii when subjected to both osmotic and organic solvent stress. FEMS Microbiol Lett 189: 149154.
  • Brown, G.R., Sutcliffe, I.C., and Cummings, S.P. (2003) Combined solvent and water activity stresses on turgor regulation and membrane adaptation in Oceanimonas baumannii ATCC 700832. Antonie Van Leeuwenhoek 83: 275283.
  • Brown, E.D., and Wood, J.M. (1993) Conformational change and membrane association of the PutA protein are coincident with reduction of its FAD cofactor by proline. J Biol Chem 268: 89728979.
  • Cayley, D.S., Guttman, H.J., and Record, M.T., Jr (2000) Biophysical characterization of changes in amounts and activity of Escherichia coli cell and compartment water and turgor pressure in response to osmotic stress. Biophys J 78: 17481764.
  • Cayley, S., and Record, M.T., Jr (2003) Roles of cytoplasmic osmolytes, water and crowding in the response of Escherichia coli to osmotic stress: biophysical basis of osmoprotection by glycine betaine. Biochemistry 42: 1259612609.
  • Cronan, J.E. (2003) Bacterial membrane lipids: where do we stand? Annu Rev Microbiol 57: 203224.
  • Cronan, J.E., and Vagelos, P.R. (1972) Metabolism and function of the membrane phospholipids of Escherichia coli. Biochim Biophys Acta 265: 2560.
  • Culham, D.E., Lasby, B., Marangoni, A.G., Milner, J.L., Steer, B.A., van Nues, R.W., and Wood, J.M. (1993) Isolation and sequencing of Escherichia coli gene proP reveals unusual structural features of the osmoregulatory proline/betaine transporter, ProP. J Mol Biol 229: 268276.
  • Culham, D.E., Dalgado, C., Gyles, C.L., Mamelak, D., MacLellan, S., and Wood, J.M. (1998) Osmoregulatory transporter ProP influences colonization of the urinary tract by Escherichia coli. Microbiology 144: 91102.
  • Culham, D.E., Tripet, B., Racher, K.I., Voegele, R.T., Hodges, R.S., and Wood, J.M. (2000) The role of the carboxyl terminal α-helical coiled-coil domain in osmosensing by transporter ProP of Escherichia coli. J Mol Recog 13: 114.
  • Culham, D.E., Hillar, A., Henderson, J., Ly, A., Vernikovska, Y.I., Racher, K.I., et al. (2003a) Creation of a fully functional, cysteine-less variant of osmosensor and proton-osmoprotectant symporter ProP from Escherichia coli and its application to assess the transporter’s membrane orientation. Biochemistry 42: 1181511823.
  • Culham, D.E., Henderson, J., Crane, R.A., and Wood, J.M. (2003b) Osmosensor ProP of Escherichia coli responds to the concentration, chemistry and molecular size of osmolytes in the proteoliposome lumen. Biochemistry 42: 410420.
  • Danevcic, T., Rilfors, L., Strancar, J., Lindblom, G., and Stopar, D. (2005) Effects of lipid composition on the membrane activity and lipid phase behaviour of Vibrio sp. DSM14379 cells grown at various NaCl concentrations. Biochim Biophys Acta 1712: 18.
  • Guo, D., and Tropp, B.E. (2000) A second Escherichia coli protein with CL synthase activity. Biochim Biophys Acta 1483: 263274.
  • Guzman, L.-M., Belin, D., Carson, M.J., and Beckwith, J. (1995) Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol 177: 41214130.
  • Haines, T.H., and Dencher, N.A. (2002) Cardiolipin: a proton trap for oxidative phosphorylation. FEBS Lett 528: 3539.
  • Heber, S., and Tropp, B.E. (1991) Genetic regulation of cardiolipin synthase in Escherichia coli. Biochim Biophys Acta 1129: 112.
  • Houssin, C., Eynard, N., Shechter, E., and Ghazi, A. (1991) Effect of osmotic pressure on membrane energy-linked functions in Escherichia coli. Biochim Biophys Acta 1056: 7684.
  • Koppelman, C.-M., den Blaauwen, T., Duursma, M.C., Heeren, R.M.A., and Nanninga, N. (2001) Escherichia coli minicell membranes are enriched in cardiolipin. J Bacteriol 183: 61446147.
  • Lai, E.-M., Nair, U., Phadke, N.D., and Maddock, J.R. (2004) Proteomic screening and identification of differentially distributed membrane proteins in Escherichia coli. Mol Microbiol 54: 10291044.
  • López, C.S., Heras, H., Ruzal, S.M., Sánchez-Rivas, C., and Rivas, E.A. (1998) Variations of the envelope composition of Bacillus subtilis during growth in hyperosmotic medium. Curr Microbiol 36: 5561.
  • López, C.S., Alice, A.F., Heras, H., Rivas, E.A., and Sánchez-Rivas, C. (2006) Role of anionic phospholipids in the adaptation of Bacillus subtilis to high salinity. Microbiology 152: 605616.
  • Machado, M.C., López, C.S., Heras, H., and Rivas, E.A. (2004) Osmotic response in Lactobacillus casei ATCC 393: biochemical and biophysical characteristics of membrane. Arch Biochem Biophys 422: 6170.
  • MacMillan, S.V., Alexander, D.A., Culham, D.E., Kunte, H.J., Marshall, E.V., Rochon, D., and Wood, J.M. (1999) The ion coupling and organic substrate specificities of osmoregulatory transporter ProP in Escherichia coli. Biochim Biophys Acta 1420: 3044.
  • Matsumoto, K., Kusaka, J., Nishibori, A., and Hara, H. (2006) Lipid domains in bacterial membranes. Mol Microbiol 61: 11101117.
  • Meury, J. (1994) Immediate and transient inhibition of the respiration of Escherichia coli under hyperosmotic shock. FEMS Microbiol Lett 121: 281286.
  • Mileykovskaya, E., and Dowhan, W. (2000) Visualization of phospholipid domains in Escherichia coli by using the cardiolipin-specific fluorescent dye 10-N-nonyl acridine orange. J Bacteriol 182: 11721175.
  • Mileykovskaya, E., and Dowhan, W. (2005) Role of membrane lipids in bacterial division-site selection. Curr Opin Microbiol 8: 135142.
  • Mileykovskaya, E., Dowhan, W., Birke, R.L., Zheng, D., Lutterodt, L., and Haines, T.H. (2001) Cardiolipin binds nonyl acridine orange by aggregating the dye at exposed hydrophobic domains on bilayer surfaces. FEBS Lett 507: 187190.
  • Mileykovskaya, E., Zhang, M., and Dowhan, W. (2005) Cardiolipin in energy transducing membranes. Biochemistry (Mosc) 70: 154158.
  • Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
  • Nagamori, S., Smirnova, I.N., and Kaback, H.R. (2004) Role of YidC in folding of polytopic membrane proteins. J Cell Biol 165: 5362.
  • Neidhardt, F.C., Bloch, P.L., and Smith, D.F. (1974)Culture medium for enterobacteria. J Bacteriol 119: 736747.
  • Nishijima, S., Asami, Y., Uetake, N., Yamagoe, S., Ohta, A., and Shibuya, I. (1988) Disruption of the Escherichia coli cls gene responsible for cardiolipin synthesis. J Bacteriol 170: 775780.
  • Petit, J.M., Maftah, A., Ratinaud, M.H., and Julien, R. (1992) 10N-nonyl acridine orange interacts with cardiolipin and allows the quantification of this phospholipid in isolated mitochondria. Eur J Biochem 209: 267273.
  • Pluschke, G., Hirota, Y., and Overath, P. (1978) Function of phospholipids in Escherichia coli. Characterization of a mutant deficient in cardiolipin synthesis. J Biol Chem 253: 50485055.
  • Racher, K.I., Voegele, R.T., Marshall, E.V., Culham, D.E., Wood, J.M., Jung, H., et al. (1999) Purification and reconstitution of an osmosensor: transporter ProP of Escherichia coli senses and responds to osmotic shifts. Biochemistry 38: 16761684.
  • Ragolia, L., and Tropp, B.E. (1994) The effects of phosphoglycerides on Escherichia coli cardiolipin synthase. Biochim Biophys Acta 1214: 323332.
  • Rilfors, L., and Lindblom, G. (2002) Regulation of lipid composition in biological membranes – biophysical studies of lipids and lipid synthesizing enzymes. Colloids Surf B Biointerfaces 26: 112124.
  • Russell, N.J. (1989) Adaptive modifications in membranes of halotolerant and halophilic microorganisms. J Bioenerg Biomembr 21: 93113.
  • Russell, N.J. (1993) Lipids of halophilic and halotolerant microorganisms. In The Biology of Halophilic Bacteria. Vreeland, R.H., and Hochstein, L.I. (eds). Boca Raton: CRC Press, pp. 163210.
  • Sambrook, J., Frisch, E.F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
  • Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K., Gartner, F.H., Provenzano, M.D., et al. (1985) Measurement of protein using bicinchoninic acid. Anal Biochem 150: 7685.
  • Tropp, B.E. (1997) Cardiolipin synthase from Escherichia coli. Biochim Biophys Acta 1348: 192200.
  • Trotsenko, Y.A., and Khmelenina, V.N. (2002) Biology of extremophilic and extremotolerant methanotrophs. Arch Microbiol 177: 123131.
  • Tsatskis, Y., Khambati, J., Dobson, M., Bogdanov, M., Dowhan, W., and Wood, J.M. (2005) The osmotic activation of transporter ProP is tuned by both its C-terminal coiled-coil and osmotically induced changes in phospholipid composition. J Biol Chem 280: 4138741394.
  • Vargas, C., Kallimanis, A., Koukkou, A.I., Calderon, M.I., Canovas, D., Iglesias-Guerra, F., et al. (2005) Contribution of chemical changes in membrane lipids to the osmoadaptation of the halophilic bacterium. Chromohalobacter salexigens. Syst Appl Microbiol 28: 571581.
  • Wood, J.M. (1999) Osmosensing by bacteria: signals and membrane-based sensors. Microbiol Mol Biol Rev 63: 230262.
  • Wood, J.M. (2006) Osmosensing by bacteria. Sci STKE 357: pe43.
  • Wood, J.M., Culham, D.E., Hillar, A., Vernikovska, Y.I., Liu, F., Boggs, J.M., and Keates, R.A.B. (2005)Structural model for the osmosensor, transporter, and osmoregulator ProP of Escherichia coli. Biochemistry 44: 56345646.