Phospholipid Metabolism During Penicillinase Production in Bacillus licheniformis

During membrane-bound penicillinase production, Bacillus licheniformis forms vesicles and tubules that do not appear in the absence of penicillinase production. The major lipids of B. licheniformis were shown to be phospholipids. The proportions, metabolism, and the total phospholipid per cell were shown to be essentially the same in the uninduced control, induced and constitutive penicillinase forming cells during both the exponential and stationary growth phases. Membrane phospholipids were not secreted into the medium during penicillinase formation. In the shift from the exponential to the stationary growth phase, there was an accumulation of phosphatidyl glycerol and a marked decrease in cardiolipin. These two lipids had the most active turnover of their phospholipid phosphate of all the lipids studied.     

Membrane interaction of the glycosyltransferase MurG: a special role for cardiolipin

MurG is a peripheral membrane protein that is one of the key enzymes in peptidoglycan biosynthesis. The crystal structure of Escherichia coli MurG (S. Ha, D. Walker, Y. Shi, and S. Walker, Protein Sci. 9:1045-1052, 2000) contains a hydrophobic patch surrounded by basic residues that may represent a membrane association site. To allow investigation of the membrane interaction of MurG on a molecular level, we expressed and purified MurG from E. coli in the absence of detergent. Surprisingly, we found that lipid vesicles copurify with MurG. Freeze fracture electron microscopy of whole cells and lysates suggested that these vesicles are derived from vesicular intracellular membranes that are formed during overexpression. This is the first study which shows that overexpression of a peripheral membrane protein results in formation of additional membranes within the cell. The cardiolipin content of cells overexpressing MurG was increased from 1 +/- 1 to 7 +/- 1 mol% compared to nonoverexpressing cells. The lipids that copurify with MurG were even further enriched in cardiolipin (13 +/- 4 mol%). MurG activity measurements of lipid I, its natural substrate, incorporated in pure lipid vesicles showed that the MurG activity is higher for vesicles containing cardiolipin than for vesicles with phosphatidylglycerol. These findings support the suggestion that MurG interacts with phospholipids of the bacterial membrane. In addition, the results show a special role for cardiolipin in the MurG-membrane interaction.     

Specific extraction of bacterial cardiolipin from sporulating Bacillus subtilis

A method for rapid purification of bacterial cardiolipin is presented. The cardiolipin level was first increased by suspending Bacillus subtilis cells in a buffer containing an uncoupling agent. At least 90% of the phosphatidylglycerol molecules were rapidly converted into cardiolipin. In sporulating strains, the accumulated cardiolipin appeared to be unextractable by conventional phospholipid extraction procedures. Sporulating bacteria were therefore treated first by a classical technique in order to eliminate lipids other than cardiolipin; a second extraction in a highly acidic medium then allowed us to quantitatively extract the remaining cardiolipin. Besides simplicity and rapidity, this method has the advantage of yielding cardiolipin in a nearly pure form from a relatively low number of bacteria.     

Calorimetric studies of cytochrome oxidase-phospholipid interactions

Thermotropic phase transitions in phospholipid vesicles reconstituted with mitochondrial cytochrome oxidase (EC 1.9.3.1) were studied using differential scanning calorimetry. Both dimyristoylphosphatidylcholine (DMPC) and mixtures of DMPC and cardiolipin were used at different lipid-to-protein ratios. The incorporated protein reduces the energy absorbed during phase transitions of DMPC vesicles, and causes a small decrease in the transition temperature (tm). delta H depends on the amount of protein in the vesicles. This dependence indicates that about 72 DMPC molecules are influenced per cytochrome alpha alpha 3 monomer. The transition parameters remain unaffected by changes in ionic strength or by reduction of the enzyme. Incorporation of cytochrome oxidase depleted of subunit III into DMPC liposomes resulted in a larger decrease of tm, but the amount of perturbed phospholipids remains similar to that in the case of the intact enzyme. Incorporation of cytochrome oxidase into DMPC/cardiolipin vesicles counteracts the effect of cardiolipin in decreasing the enthalpy of the DMPC transition. Thus cytochrome oxidase segregates the phospholipids by attracting cardiolipin from the bulk lipid. Cytochrome c does not significantly affect this apparent cardiolipin 'shell' around membranous cytochrome oxidase.     

Erythrocyte morphology reflects the transbilayer distribution of incorporated phospholipids

The transbilayer distribution of exogenous phospholipids incorporated into human erythrocytes is monitored through cell morphology changes and by the extraction of incorporated 14C-labeled lipids. Dilauroylphosphatidylserine (DLPS) and dilauroylphosphatidylcholine (DLPC) transfer spontaneously from sonicated unilamellar vesicles to erythrocytes, inducing a discocyte-to-echinocyte shape change within 5 min. DLPC-induced echinocytes revert slowly (t1/2 approximately 8 h) to discocytes, but DLPS-treated cells revert rapidly (10-20 min) to discocytes and then become invaginate stomatocytes. The second phase of the phosphatidylserine (PS)-induced shape change, conversion of echinocytes to stomatocytes, can be inhibited by blocking cell protein sulfhydryl groups or by depleting intracellular ATP or magnesium (Daleke, D. L., and W. H. Huestis. 1985. Biochemistry. 24:5406-5416). These cell shape changes are consistent with incorporation of phosphatidylcholine (PC) and PS into the membrane outer monolayer followed by selective and energy-dependent translocation of PS to the membrane inner monolayer. This hypothesis is explored by correlating cell shape with the fraction of the exogenous lipid accessible to extraction into phospholipid vesicles. Upon exposure to recipient vesicles, DLPC-induced echinocytes revert to discoid forms within 5 min, concomitant with the removal of most (88%) of the radiolabeled lipid. On further incubation, 97% of the foreign PC transfers to recipient vesicles. Treatment of DLPS-induced stomatocytes with acceptor vesicles extracts foreign PS only partially (22%) and does not affect cell shape significantly. Cell treated with inhibitors of aminophospholipid translocation (sulfhydryl blockers or intracellular magnesium depletion) and then incubated with either DLPS or DLPC become echinocytic and do not revert to discocytic or stomatocytic shape for many hours. On treatment with recipient vesicles, these echinocytes revert to discocytes in both cases, with concomitant extraction of 88-99% of radiolabeled PC and 86-97% of radiolabeled PS. The accessibility of exogenous lipids to extraction is uniformly consistent with the transbilayer lipid distribution inferred from cell shape changes, indicating that red cell morphology is an accurate and sensitive reporter of the transbilayer partitioning of incorporated exogenous phospholipids.     

Solubilization of proteins from bovine brain coated vesicles by protein perturbants and Triton X-100

To identify integral and peripheral membrane proteins, highly purified coated vesicles from bovine brain were exposed to solutions of various pH, ionic strength, and concentrations of the nonionic detergent Triton X-100. At pH 10.0 or above most major proteins were liberated, but four minor polypeptides sedimented with the vesicles. From quantitative analysis of phospholipids in the pellet and extract, we determined that at a pH of up to 12 all phospholipids could be recovered in the pellet. Electron microscopic examination of coated vesicles at pH 12.0 showed all vesicles devoid of coat structures. Treatment with high ionic strength solutions (0-1.0 M KCl) at pH 6.5-8.5 also liberated all major proteins, except tubulin, which remained sedimentable. The addition of Triton X-100 to coated vesicles or to stripped vesicles from which 90% of the clathrin had been removed resulted in the release of four distinct polypeptides of approximate Mr 38,000, 29,000, 24,000 and 10,000. The 38,000-D polypeptide (pK approximately 5.0), which represents approximately 50% of the protein liberated by Triton X-100, appears to be a glycoprotein on the basis of its reaction with periodic acid-Schiff reagent. Extraction of 90% of the clathrin followed by extraction of 90% of the phospholipids with Triton X-100 produced a protein residue that remained sedimentable and consisted of structures that appeared to be shrunken stripped vesicles. Together our data indicate that most of the major polypeptides of brain coated vesicles behave as peripheral membrane proteins and at least four polypeptides behave as integral membrane proteins. By use of a monoclonal antibody, we have identified one of these polypeptides (38,000 mol wt) as a marker for a subpopulation of calf brain coated vesicles.     

Phospholipid Alterations During Growth of Escherichia coli

As cultures of Escherichia coli progressed from the exponential growth phase to the stationary growth phase, the phospholipid composition of the cell was altered. Unsaturated fatty acids were converted to cyclopropane fatty acids, and phosphatidyl glycerol appears to have been converted to cardiolipin. With dual isotope label experiments, the kinetics of synthesis of cyclopropane fatty acid for each of the phospholipids was examined in vivo. The amount of cyclopropane fatty acid per phospholipid molecule began to increase in phosphatidyl ethanolamine at a cell density below the density at which this increase was observed in phosphatidyl glycerol or cardiolipin. The rate of this increase in phosphatidyl glycerol or in cardiolipin was faster than the rate of increase in phosphatidyl ethanolamine. After a few hours of stationary-phase growth, all the phospholipids were equally rich in cyclopropane fatty acids. It is suggested that the phospholipid alterations observed are a mechanism to protect against phospholipid degradation during stationary phase growth. Cyclopropane fatty acid synthetase activity was assayed in cultures at various stages of growth. Cultures from all growth stages examined had the same specific activity in crude extracts.     

Growth of the mitochondrial inner membrane in synchronous cultures of Tetrahymena pyriformis: an examination of phospholipid accumulation

Based on morphological evidence, mitochondrial inner membrane growth has been reported to be discontinuous in heat shock-synchronized Tetrahymena pyriformis. As a biochemical measure of membrane growth under these conditions, we have examined phospholipid accumulation in the cell. No marked modulation of the accumulation of any of the major phospholipids could be detected through the cell cycle. At least 89% of the cardiolipin in the cells is restricted to the mitochondria, and we have used it as a marker for the growth of the mitochondrial inner membrane. During the heat shock synchrony, cardiolipin accumulates uniformly in parallel with the exponential rate of increase of total cellular phospholipids. These results suggest that at least the phospholipid component of all membrane systems in the cell grow continuously and uniformly. Additionally, we have shown that the total phospholipid content of Tetrahymena increases by a factor of 2.4 per generation following a series of heat shocks. No such net overaccumulation is observed for protein content.     

Phospholipids as ionophores.

The ionophoretic capabilities of phospholipids have been examined by direct measurement in a Pressman cell of the phospholipid-mediated translocation of cations across an organic phase separating two aqueous phases. Cardiolipin and phosphatidic acid were the most active inonophores among the phospholipids tested, with activities comparable to that of X537A in respect to the translocation of divalent cations. Cardiolipin translocates both divalent and monovalent cations at approximately equal rates. The ionophoretic activity of cardiolipin could be modulated by other phospholipids (inhibition), by butacaine (stimulation), by complexation with cytochrome c (inhibition), and by ruthenium red and lanthanum (inhibition). The rate of translocation of cations mediated by cardiolipin was independent of pH over a wide pH range (5.4 to 8.3). The same general pattern of properties observed for cardiolipin applied to phosphatidic acid except for stimulation by butacaine. Complexation of phospholipid mixtures, such as asolectin or mitochondrial lipid, with reduced cytochrome c, enhanced the ionophoretic capability of these phospholipids by 1 order of magnitude. The complex thus formed has the properties of a polyionophore. The possible physiological significance of this enormous ionophoretic potential of phospholipids is examined.     

Bromopyruvate mediates autophagy and cardiolipin degradation to monolyso-cardiolipin in GL15 glioblastoma cells

The GL15 glioblastoma cell line undergoes viability loss upon treatment with bromopyruvate. The biochemical mechanisms triggered by the antiglycolytic agent indicate the activation of an autophagic pathway. Acridine orange stains acidic intracellular vesicles already 60 min after bromopyruvate treatment, whereas autophagosomes engulfing electron dense material are well evidenced 18 h later. The autophagic process is accompanied by the expression of the early autophagosomal marker Atg5 and by LC3-II formation, a late biochemical marker associated with autophagosomes. In agreement with the autophagic route activation, the inhibitory and the activator Akt and ERK signaling pathways are depressed and enhanced, respectively. In spite of the energetic collapse suffered by bromopyruvate-treated cells, MALDI-TOF mass spectrometry lipid analysis does not evidence a decrease of the major phospholipids, in accordance with the need of phospholipids for autophagosomal membranes biogenesis. Contrarily, mitochondrial cardiolipin decreases, accompanied by monolyso-cardiolipin formation and complete cytochrome c degradation, events that could target mitochondria to autophagy. However, in our experimental conditions cytochrome c degradation seems to be independent of the autophagic process.     

Specific lipids enhance the activity of UDP-GlcNAc: dolichol phosphate GlcNAc-1-phosphate transferase in rat liver endoplasmic reticulum membrane vesicles.

The rate of the reaction catalyzed by UDP-N-acetylglucosamine (GlcNAc):dolichol phosphate GlcNAc-1-phosphate transferase in rat liver endoplasmic reticulum vesicles was shown to be influenced by particular lipids. Utilizing in vitro assay conditions where the membrane vesicles retained latency of glucose-6-phosphatase activity, the addition of phosphatidylethanolamine, cardiolipin, or monogalactosyldiglyceride resulted in severalfold increases in the rate of dolichol pyrophosphate N-acetylglucosamine synthesis. Other phospholipids were not stimulatory. These rates were dependent on the concentrations of the exogenous lipids and of the substrate dolichol phosphate. In the presence of cardiolipin, the membrane-bound enzyme became more susceptible to inactivation by protease K and to inhibition by tunicamycin. Titration of cardiolipin-containing endoplasmic reticulum vesicles with adriamycin indicated that the majority of the cardiolipin was exposed on the outer surface. These results suggest that the particular lipids altered membrane structure in a way that allowed further access of the enzyme to substrate, inhibitor, and other molecules. Lipids observed in these studies to be stimulatory are known to exist in the macromolecular hexagonal phase and may therefore be affecting the GlcNAc-1-phosphate transferase by locally disrupting the bilayer structure of the membrane. As other dolichol-utilizing enzymes have been previously observed by other investigators to be similarly influenced by such lipids, the effects may be common to enzymes of the dolichol cycle.     

Specific binding of mitochondrial protein precursors to liposomes containing cardiolipin

In vitro synthesized precursors of several mitochondrial proteins, including P-450(SCC), adrenodoxin, and malate dehydrogenase, bound to liposomes prepared from mitochondrial phospholipids, but not to those from microsomal phospholipids. When liposomes were prepared from various pure phospholipids, adrenodoxin precursor was bound only to the liposomes that contained cardiolipin. The liposomes containing other phospholipids did not show the binding affinity for the precursor. The binding was observed only with the precursor peptides of adrenodoxin and malate dehydrogenase, and their mature forms were not bound to the liposomes. The binding of the precursors was dependent on the concentration of cardiolipin in the liposomes. Liposomes containing various cardiolipin derivatives with modified polar head groups showed very different binding affinity for adrenodoxin precursor, suggesting the importance of the structure of the polar head of the cardiolipin molecule. Two or three positively charged amino acid residues in the extension peptide of P-450(SCC) precursor were replaced by neutral amino acid residues by site-directed mutagenesis. The mutated P-450(SCC) precursors did not bind to the liposomes containing cardiolipin. The results indicated that mitochondrial protein precursors have specific affinity for cardiolipin, and the affinity was due to the interaction between the extension peptides of the precursors and the polar head of the cardiolipin molecule.     

Evidence for a lipid dependence of mitochondrial nicotinamide nucleotide transhydrogenase

1. The lipid dependence of mitochondrial nicotinamide nucleotide transhydrogenase from beef heart was investigated. With submitochondrial particles digestion of phospholipids by phospholipases A and C led to a partial inhibition that could not be readily reversed by phospholipids.

2. Extraction of neutral lipids including ubiquinone from lyophilized submitochondrial particles with pentane did not inhibit the transhydrogenase, whereas further extraction with water/acetone led to a complete and apparently irreversible inhibition.

3. A partially purified preparation of transhydrogenase, depleted of lipids (and inactivated) by treatment with cholate and ammonium sulphate, was reactivated by various purified phospholipids but not by detergents or triacylglycerols.

4. It is concluded that mitochondrial transhydrogenase, catalyzing the non-energy-linked transhydrogenase reaction, requires phospholipids specifically for its catalytic activity and not as dispersing agents. A mixture of phospholipids appears to fulfill this requirement better than the individual phospholipids.     

Peroxidative modification of phospholipids in myocardial membranes

Rat heart myocardial membranes exposed to the free radical generating system, Fe2+/ascorbate, undergo lipid peroxidation as evidenced by the accumulation of thiobarbituric acid-reactive substances, loss of polyunsaturated fatty acids from phospholipids, and formation of conjugated dienes and fluorescent substances. In addition, the treated membranes exhibit a dramatic decrease in extractable phospholipids. This decrease is even more pronounced in individual phospholipid classes isolated by high-performance liquid chromatography. The decrease in lipid phosphorus under oxidant stress is accompanied by an increase in the phosphorus content of the aqueous phase after Folch extraction and by an even greater increase of phosphorus in the protein residue. In addition, increased amounts of saturated and monounsaturated fatty acyl groups are found in the protein residue of Fe2+/ascorbate-treated membranes. Extraction of the oxidant-treated membranes with acidic solvents does not enhance the recovery of phospholipids and neither does treatment with detergents, trypsin, and chymotrypsin prior to lipid extraction. However, treatment with the bacterial protease, Pronase, markedly enhances the recovery of phospholipids from the peroxidized membranes. These results indicate that membrane phospholipids undergoing free radical-induced peroxidation may form lipid-protein adducts, which renders them inextractable with lipid solvents.     

Phospholipid accumulation during the cell cycle in synchronous cultures of the yeast, Saccharomyces cerevisiae

Phospholipid concentrations have been examined throughout successive cell cycles in synchronously growing cultures of the yeast, Saccharomyces cerevisiae. Total phospholipid phosphorus, as well as lecithin and phosphatidylethanolamine levels, exhibited stepwise increases during the cell cycle with step increments beginning just prior to new rounds of bud formation. Phosphatidylinositol and phosphatidylserine levels, on the other hand, showed what have been interpreted to be peak concentrations near the time of bud formation. Cardiolipin content varied considerably and was dependent upon the carbon source of the growth medium. Glucose-grown cells exhibited peak concentrations of cardiolipin near the time of bud formation, with marked decreases after this time. In contrast, galactose-grown synchronous cells exhibited stepwise increments in cardiolipin content, with step increases occurring near the time of new rounds of bud formation. Step or peak increases in cardiolipin, as well as all other phospholipids, were found to coincide with the time of stepwise increases in cytochrome c oxidase activity in these cells. No correlations were observed between the elaboration of mitochondrial membranes during the synchronous cell cycle and the observed patterns of phospholipid increase.     

The monocarboxylate carrier from bovine heart mitochondria: partial purification and its substrate-transporting properties in a reconstituted system

The monocarboxylate (pyruvate) carrier from bovine heart mitochondria was extracted from submitochondrial particles with Triton X-114 in the presence of cardiolipin. By a single hydroxylapatite chromatography step a 125-fold purification of the carrier protein could be achieved. High pyruvate/pyruvate-exchange activity was recovered, when the protein was reconstituted into phospholipid vesicles. No transport activity was observed, when the isolation occurred in the absence of phospholipids. The 2-cyano-4-hydroxycinnamate sensitive pyruvate exchange reaction was strongly temperature sensitive and dependent on the amount of protein reconstituted. Other 2-ketoacids caused competitive inhibition of the pyruvate uptake. Inhibitors of other mitochondrial carries, however, had very low or no effect on the monocarboxylate exchange. The influence of different -SH group reagents on the measured pyruvate/pyruvate-exchange in the reconstituted system was similar to the one observed with intact mitochondria. It is concluded that the described procedures for extraction, purification and reconstitution of the mitochondrial monocarboxylate carrier conserved the functional properties of the protein.     

Binding of cardiolipin/lecithin and monoamine oxidase to lipid-depleted mitochondrial membrane structure.

Lipid-depleted pig liver mitochondrial residues were incubated with different proportions of the acidic phospholipid cardiolipin and the zwitterionic phospholipid lecithin in either separate or mixed liposomes. When cardiolipin and lecithin were present in separate liposomes all of the cardiolipin but no lecithin bound to the residues. When present in the same liposomes, cardiolipin also caused binding of lecithin to the mitochondrial residues. When monoamine oxidase solubilized from pig liver mitochondria by extraction of the phospholipids was included in the incubation, binding of the enzyme to the residues occurred in the presence of cardiolipin. The percentage of enzyme bound followed the same trend as the binding of phospholipids to the mitochondrial residues.     

Effect of the purified (Mg2+ + Ca2+)-activated ATPase of sarcoplasmic reticulum upon the passive Ca2+ permeability and ultrastructure of phospholipid vesicles.

The passive Ca2+ permeability of fragmented sarcoplasmic reticulum membranes is 10(4) to 10(61 times greater than that of liposomes prepared from natural or synthetic phospholipids. The contribution of membrane proteins to the Ca2+ permeability was studied by incorporating the purified [Ca2+ + Mg2+]-activated ATPase into bilayer membranes prepared from different phospholipids. The incorporation of the Ca2+ transport ATPase into the lipid phase increased its Ca2+ permeability to levels approaching that of sarcoplasmic reticulum membranes. The permeability change may arise from a reordering of the structure of the lipid phase in the environment of the protein or could represent a specific property of the protein itself. The calcium-binding protein of sarcoplasmic reticulum did not produce a similar effect. The increased rate of Ca2+ release from reconstituted ATPase vesicles is not a carrier-mediated process as indicated by the linear dependence of the Ca2+ efflux upon the gradient of Ca2+ concentration and by the absence of competition and countertransport between Ca2+ and other divalent metal ions. The increased Ca2+ permeability upon incorporation of the transport ATPase into the lipid phase is accompanied by similar increase in the permeability of the vesicles for sucrose, Na+, choline, and SO42- indicating that the transport ATPase does not act as a specific Ca2+ channel. Native sarcoplasmic reticulum membranes are asymmetric structures and the 75-A particles seen by freeze-etch electron microscopy are located primarily in the outer fracture face. In reconstituted ATPase vesicles the distribution of the particles between the two fracture faces is even, indicating that complete structural reconstitution was not achieved. The Ca2+ transport activity of reconstituted ATPase vesicles is also much less than that of fragmented sarcoplasmic reticulum. The density of the 40-A surface particles visible after negative staining of native or reconstituted vesicles is greater than that of the intramembranous particles and the relationship between these two structures remains to be established.     

Cyclopropane fatty acid synthase of Escherichia coli. Stabilization, purification, and interaction with phospholipid vesicles.

The cyclopropane fatty acid (CFA) synthase of Escherichia coli catalyzes the methylenation of the unsaturated moieties of phospholipids in a phospholipid bilayer. The methylene donor is S-adenosyl-L-methionine. The enzyme is loosely associated with the inner membrane of the bacterium and binds to and is stabilized by phospholipid vesicles. The enzyme has been purified over 500-fold by flotation with phospholipid vesicles and appears to be a monomeric protein having a molecular weight of about 90 000. The enzyme binds only to vesicles of phospholipids which contain either unsaturated or cyclopropane fatty acid moieties. CFA synthase is active on phosphatidylglycerol, phosphatidylethanolamine, and cardiolipin, the major phospholipids of E. coli, and also has some activity on phosphatidylcholine. The enzyme is equally active on phospholipid vesicles in the ordered or the disordered states of the lipid phase transition. Studies with a reagent that reacts only with the phosphatidylethanolamine molecules of the outer leaflet of a phospholipid bilayer indicate that CFA synthase reacts with phosphatidylethanolamine molecules of both the outer and the inner leaflets of phospholipid vesicles.     

Activity of N-acetylmuramoyl-L-alanine amidase in phospholipidic environments

A purified preparation of N-acetylmuramoyl-L-alanine amidase (EC 3.5.1.28), a murein hydrolase from Escherichia coli, was found to lose its activity during incubation in the presence of bacterial phospholipid suspensions. Whether it was co-dispersed with the phospholipids or added to sonicated phospholipid suspension, the enzyme was inhibited (or inactivated) from the first minutes of incubation at 37 degree C. As phosphatidylglycerol/cardiolipin ratio of the phospholipid suspension as increased (all other things being equal), a further decrease of amidase activity was observed. The highest losses of activity were found after co-dispersion of the enzyme and the substrate together with the phospholipids, the resulting suspension being formed of larger multilayered vesicles, as revealed by electron microscopy. In these conditions, the effect on enzyme activity was only partially accounted for by the proportion of the enzyme that was entrapped in the vesicles. The entrapment capacity of the enzyme (using a 35S-labelled enzyme preparation) and of the substrate (3H-labelled) by the multilamellar phospholipidic vesicles did not significantly change as a function of their relative content of phosphatidylglycerol and cardiolipin. The possible physiological meaning of the results is discussed is connection with our previous data and with other works related to membranous phospholipid distribution and to septum formation control in bacteria.