JOURNAL OF BACTERIOLOGY, Mar. 1996, p. 1394-1400 0021-9193/96/$04.00+0 Copyright 0 1996, American Society for Microbiology Vol. 178, No. 5 Inorganic Polyphosphate Supports Resistance and Survival of Stationary-Phase Escherichia coli I NARAYANA N. RAO AND ARTHUR KORNBERG* Department of Biochemistty, Stanford University School of Medicine, Stanford, California 94305-5307 Received 17 July 1995iAccepted 27 December 1995 The Escherichia coli mutant (ppk) lacking the enzyme polyphosphate kinase, which makes long chains of inorganic polyphosphate (poly P), is deficient in functions expressed in the stationary phase of growth. After 2 days of growth in a medium limited in carbon sources, only 7% of the mutants survived compared with nearly 100% of the wild type; the loss in viability of the mutant was even more pronounced in a rich medium. The mutant showed a greater sensitivity to heat, to an oxidant (H,O,), to a redox-cycling agent (menadione), and to an osmotic challenge with 2.5 M NaCI. After a week or so in the stationary phase, mutant survivors were far fewer in number and were replaced by an outgrowth of a small-colony-size variant with a stable genotype and with improved viability and resistance to heat and H,O,; neither polyphosphate kinase nor long-chain poly P was restored. Suppression of the ppk feature of heat sensitivity by extra copies of rpoS, the gene encoding the RNA polymerase sigma factor that regulates some 50 stationary-phase genes, further implicates poly P in promoting survival in the stationary phase. ~ Inorganic polyphosphates (poly P), linear polymers of P, linked by phosphoanhydride bonds, are ubiquitous in nature, having been found in all organisms examined (31, 32, 57). Microbes, including bacteria, cyanobacteria, fungi, and algae, accumulate poly P, which may amount to 10 to 20% of their dry weight. Chain length may vary from 3 to 1,000 P, residues, depending on the organism, its growth, and physiological con- ditions (45, 53, 54). Despite its ubiquity and conservation in evolution, the physiological roles of poly P remain unknown or ignored. Poly P provides a substitute for ATP for sugar and adenylate kinases (8, 13, 22, 42), a pH-stat to counterbalance alkaline stress (43, 44), a phosphate reservoir with osmotic advantages (31, 33), an energy source (31, 57), a chelator of divalent metal ions (4, 14), a membrane component for DNA entry and transformation (46), and a possible regulator in transcription. Poly P appears to have a multiplicity of roles in its many locations (30). To elucidate the functions of poly P in Escherichia colt, we examined mutants that fail to synthesize polyphosphate kinase (PPK) and lack the long-chain poly P. These mutants failed to survive in stationary phase and were less resistant to heat and oxidants (10). In further studies reported here, the mutants, after a few days in the stationary phase, were replaced by a small-colony variant that was genet- ically stable, more viable, and stress resistant. Still other results implicate poly P in the complex responses that promote sur- vival in the stationary phase. MATERIALS AND METHODS Reagents. The sources of the reagents used in this study were as follows: ATP, creatine kinase, catalase, DNase I, and RNase, Boehringer Mannheim; creatine phosphate. 3-(N-morpholino) propanesulfonic acid (MOPS), menadione, kana- mycin, ampicillin, and bovine serum albumin, Sigma; H,O,, Fisher Scientific Co.; Immobilon-N membrane, Millipore; [y-"PIATP, Amersham Corp.; ['H-meth- yl]thymidinc, ICN Biomedicals Iuc.; and polyethylene imine-cellulose F thin- layer chromatography plates, Merck. rPPXl used in the poly P assay was pre- pared and assayed as described previously (59). * Corresponding author. Mailing address: Department of Biochem- istry, Beckman Center, Stanford University School of Medicine, Stan- ford, CA 94305-5307. Phone: (415) 723-6167. Fax: (415) 723-6783. Bacterial strains and plasmids. The following E. coli K-12 derivatives were used: JMlOl {h(luc-proAB) supE tht-IF' [fruD36 proAB+ lncl' lncZA M15]}; CA10, the ppk ppx mutant constructed by transfer of a disrupted ppk gene, insertionally inactivated with a gene for kanamycin resistance, into JMlOl by P1 transduction (10); CAlO pMMkufF3, constructed by transformation of CA38 (CA10 red) with pMmafF3 (40) (obtained from A. Matin, Stanford University, Stanford, Calif.); and pBC29, a multicopy plasmid bearing the gene ppk, which was used for poly P overexpression (2). rpoS::TnlO derivatives of CAlO and the CAlO small-colony phenotype were made by transduction with P1 grown on MC8301 [MC4100 F- uruD139 A (nrg-luc) u169 rp.sL150 relAl fib E5301 deoC ptsF25 rpsR rpoS::TnlO]. Survival assays. Long-term survival of E. coli was assayed as described earlier (10) by growing cells either in MOPS-buffered minimal medium (MOPS medium [41]) containing glucose (0.1%) as the carbon source and K2HP0, (2 mM) as the P, source or in a rich medium, Luria-Bertani (LB) liquid medium. Viable cell counts were determined by plating of cells onto LB agar plates: kanamycin (50 pg/ml) was added to the medium for plating of CA10. To test the sensitivity of the stationary-phase cells to H202, JMlUl and CAlO were grown overnight (about 20 h) in LB medium and LB medium with kana- mycin (50 pg/ml). respectively; the medium for CAlO pBC29 contained 50 kg of ampicillin per ml. The cells were washed and resuspended in 0.85% NaCl to an optical density at 540 nm (ODs4") of 1.0; H,O, was added to a final concentra- tion of 42 mM. At the times indicated, 0.1-ml samples were withdrawn, diluted immediately in 0.85% NaC1, and plated on LB plates to determine viable cell counts. For actively growing cells. JMlOl and CAlU were grown aerobically at 37°C to the mid-exponential phase (ca. 4 x 10" cells per ml) in MOPS medium containing glucose (0.4%) and KZHP04 (2 mM). Two-milliliter samples of these cultures were exposed for 15 min to H,O, (0 to 40 mM) in air at 37°C. To study the lethal effect of H,O,, the exposure was terminated by addition of 2 pg of catalase per ml, followed by dilution into 0.85% NaCI. Cells were then plated onto LB plates to determine the viable cell count. Starvation-induced protection against osmotic challenge in JMlOl and CAlO cells was assessed by comparison of the viabilities of cultures osmotically chal- lenged during mid-log growth or at 3 h after entering the stationary phase. Cultures were grown at 37°C in M9 medium (8 mM NaCI) supplemented with either 0.025% glucose (for starvation at an OD5, of about 1.0 [4 X lo8 cells per ml]) or in 0.4% glucose for sampling during exponential growth at a similar cell density. Cultures were osmotically challenged by addition of solid NaCl to the medium to reach a final concentration of 2.5 M (27). Survival was determined by colony counts on LB plates compared with those of the cultures just before the osmotic challenge. For the heat shock survival assay, cells were grown overnight in LB medium. Stationary-phase cells were washed once and diluted in 0.85% NaCl to a density of about 5 X 10' cells per ml. Samples (2 ml) were put into prewarmed glass tubes (55°C); at the times indicated, 0.1-ml portions were plated directly onto LB plates to determine viable cell numbers. Growth studies. JM101. CAlO small-colony phenotype, and CAlO big-colony phenotype cells were grown overnight in MOPS medium containing glucose (0.4%) and K,HPO, (2 mM). The cultures were then diluted in fresh, pre- 1394 VOL. 178, 1996 INORGANIC POLYPHOSPHATE SUPPORTS SURVIVAL OF E. COLI 1395 012345 Time (min) FIG 1. Heat resistance. Wild-type (WT), CAl0, and CAlO pMMkatF3 cells grown overnight (-20 h) at 37°C in LB medium were exposed to 55T, and their viability was determined by plating on LB agar, with antibiotics wherever nec- essary. Percent survival is determined as the viable cell number at each time point divided by the viable cell number before exposure to heat. warmed (37°C) medium, and samples were removed at hourly intervals to mon- itor growth by measurement of the turbidity at 540 nm. ['Hlthjmidine uptake. Cells were grown to the stationary phase in LB me- dium aerobically at 37°C and then incubated further for assay of long-term survival (10). The number of surviving cells was determined each day by plating on LB agar. From the 3rd to 10th days, cells were labeled by addition of ['Hlthymidine to a final concentration of 1 pM in 100 kl of culture medium and incubation for 60 min at 37°C with shaking. Cells were collected in a microcen- trifuge and washed once with 0.85% NaCI. Cold 10% trichloroacetic acid (1 ml) was added to the pellet; the suspension was vortexed and collected on a What- man GFK Millipore glass fiber filter, washed first with cold 10% trichloroacetic acid and then with ethanol. Radioactivity was measured in a liquid scintillation counter. Resistance to menadione. For growth-inhibition measurements, 0.1 ml of an overnight culture grown in LB medium was added to 3 ml of molten F top agar (39) and poured on a K medium agar plate (12). A solution (100 PI) of mena- dione (2 mg/100 pl of dimethyl sulfoxide) was allowed to saturate a 13-mm- diameter filter disc (Whatman no. 1). The disc was then placed in the center of the solidified top agar plate and allowed to incubate overnight (18 to 24 h). The zone of inhibition was determined by measurement of the diameter of the zone. less that of the disc. Extraction and assay of poly P. Cells were grown in LB medium to the stationary phase (OD,,,, -4.5). Poly P was extracted, and the concentration was estimated by a rapid radioassay technique (36). Other methods. PPK and exopolyphosphatase (PPX) activities in cell extracts of wild-type and CAlO (small- and big-colony phenotype) cells were assayed as described previously (1, 2). Catalase activities were measured in cells grown to the stationary phase in LB medium. Cells permeabilized by toluene were used in the assay, which measures the disappearance of peroxide spectrophotometrically at 240 nm (5). Visualization of catalases (HPI and HPII) on acrylamide gels. Typical hy- droperoxidase 1 (HPI) and HPII bands were visualized by running whole-cell extracts on an 8% polyacrylamide gel and then staining them with a 50:SU solution of 2% K,Fe (CN),, and 2% FeCI, (9,26). Bovine serum catalase (Sigma) was used as a standard. RESULTS Effect of poly P deficiency on long-term survival in the sta- tionary phase: heat resistance of theppk mutant (CA10). Dur- ing survival in the stationary phase or upon starvation, cells acquire ipoS(katF)-mediated resistance to multiple stresses, including heat and oxidants (3, 6, 16, 19, 27, 33-35, 56). When poly P-deficient cells (CA10) were held at 55°C for 2 min, only 2% survived compared with about 90% of the corresponding wild-type cells (JM101) (10). The mutant complemented with multiple copies of the plasmid (pMM katF3) bearing rpoS, the TABLE 1. Menadione sensitivity of E coli with and without poly P Strain Genotype Killing zone (rnm?)" JMlOl PPk + CAlO PPk CAlO pBC29 PPk+ CA38 ppk recA CA38 pUC18 ppk recA 31 5 13 200 5 74 210 2 24 284 t 6 13 t- 6 a The killing zone measurements represent the average of SIX separate exper- iments (see Matenals and Methods). stationary-phase RNA polymerase sigma factor, proved to be as heat resistant as wild-type cells (Fig. 1). Similar results were obtained when the mutant was complemented with a plasmid which contained ppk (data not shown). Resistance to oxidative stress. Preliminary studies had noted that poly P-deficient cells in the stationary phase were more sensitive than wild-type cells to 42 mM H,O, (10). When the viability of exponentially growing cells was determined after exposure to various levels of H,O, for 15 min, both wild-type and mutant cells showed similar degrees of sensitivity to low levels (2.5 to 10 mM) of H,O, (Fig. 2). However, the mutants were far more sensitive to higher concentrations of H,O, (15 to 45 mM) than the wild type. Complementation ol the mutant in the stationary phase with a plasmid bearing the ppk gene restored H,O, resistance to near wild-type levels (data not presented here). Exposure of cells to the quinone menadione leads to oxida- tive stress due to the production of oxygen radicals (51); cells lacking catalase, peroxidase, and superoxide dismutase show increased sensitivity (11, 17). Theppk mutant in the stationary phase also proved to be highly sensitive to menadione toxicity and could be restored to resistance by complementation with the ppk gene (Table 1). Kesistance to osmotic challenge. Stationary-phase cells show enhanced osmotic resistance compared with cultures in the exponential growth phase or preadapted to osmotic stress (27). The levels of osmotic resistance of stationary-phase cells 4 h after they were depleted of a carbon source (glucose) were compared with those of cells in the mid-log phase by exposure to 2.5 M NaC1. Viability in the stationary-phase cultures con- 100 10 1 0.1 0.01 0 F 10 20 30 40 H202 (mM) FIG. 2. Resistance to HZO,. Wild-type (WT) and CAlO cells were grown exponentially in MOPS medium. The cells were exposed to various concentra- tions of HZOz for a period of 15 min at 2O"C, and the viable cell numbers were determiiied by plating onto LB agar. Survival of 100% corresponds to the viable cell number determined immediately before the addition of H,O,. 1396 RAO AND KORNBERG J. BACTEKIOL. .001 - 100 10 1 Fl CAI 0 .001 -01 i 0 5 10 15 20 Time (hours) FIG. 3. Osmotic resistance to 2.5 M NaCI. Wild-type (WT) and CAlO cells were examined in the exponential phase (A) or in the stationary phase after 3 h of starvation for glucose (B). sistently indicated higher osmotic resistance than did that in the cultures challenged during the log phase of growth (Fig. 3). However, the level of protection conferred on the mutant by carbon starvation was lower than that observed in the wild-type cells. After exposure of the starving stationary-phase cells to 2.5 M NaCl for 21.5 h, about 10% of the wild-type cells and only about 1.0% of the mutant cells survived (Fig. 3B). The sensitivity towards osmotic challenge was more pronounced in the exponential cultures, and the absence of long-chain poly P in theppk mutant resulted in even lower levels of viability (Fig. 3A). Mutants fail to survive in the stationary phase and are replaced by a small-colony variant. Wild-typc and CAlO mu- tant cells were grown to the stationary phase in a synthetic MOPS medium limited in glucose (0.1%) and incubated aer- obically at 37°C for several days. Viable cell counts measured after 1 day in the stationary phase indicated a 93% loss for the mutant compared with no loss in the wild type (Fig. 4A). After the second day, the mutant showed equal numbers of novel small colonies and the standard larger type. Both the small- and big-colony phenotypes of the mutant decreased in number upon prolonged incubation. About 3% of the ppk cells (both small- and big-colony phenotypes) and 20% of the wild-type cells survived after 10 days in the stationary phase (Fig. 4A). Both phenotypes of the mutant exhibited a growth lag when overnight (-20-h) cultures were diluted and grown in a rich medium (LB); they were also more heat sensitive (10). In LB medium, survival of the mutant was 0.001 to 1.0% during the first 4 days of the stationary phase, dropping to values as low as 100 CFU/ml of culture. At this point, thc big-colony phenotype predominated, but when cultures grew back to about 5 X lo7 CFU/ml, 90% or more of the population .were of the small-colony phenotype. Despite variations in loss of viability in several trials (Fig. 4B), over 90% of the mutants were of the small-colony variety after 5 to 6 days. The viability of the wild type remained constant after 3 days of incubation with no visible change in colony morphology. Thymidine uptake and stationary-phase survival. The dy- namic nature of stationary-phase cultures incubated for pro- longed periods is manifested in the appearance of revertants and suppressors (52, 61). In order to study the rapid takeover of the mutant population by a small-colony stable phenotype, both the wild type and the mutant were grown to the stationary phase in LB medium and incubated aerobically at 37°C. The viability of the cultures was determined daily for 10 days (Fig. 5). Starting from the third day of incubation, when the viability of both the wild type and the mutant started to decline, thc cultures were pulse-labeled for 1 h with ['Hlthymidine. A loss in viability of the wild-type cells was observed until the third day of incubation, and then viability remained almost un- changed thereafter; the uptake of [3H]thymidine was not sig- nificant (Fig. 5A). The viability of CAlO cells decreased con- siderably with time to about 10" CFU/ml on thc scvcnth day of incubation (Fig. 5B). The drop in viability of the mutant on the seventh day coincided with a burst of thymidine uptake and also coincided with a rapid increase in the number of cells of 100 10 1 n CAI0 (big) ?+? - .1 E 100 10 1 .I .01 .001 Y IIIIIIII 13 5 7 911 Time (days) FIG. 4. Long-term survival. Wild-type (WT) and CAlO cells (small and big) were incuhated in MOPS medium with limited glucose (0.1%). (A) Percent survival is expressed as the viable cell number at each time point divided by the viable cell number of the same culture after 24 h after initial inoculation of the within 4 days of incubation, whereas the wild-type cell number remained about 20777 of the initial value pig, 4~). in the a small-colony phenotype appeared in the mutant cultures. Mutants lost about 99% of their viability culture. Similar results were obtained in three independent experiments. (B) Wild-type and CAIU cells were kept in LB medium. Curves 1, 2, 3, and 4 represent four different CAlO cultures showing fluctuations in survival due to time of appearance of the small-colony variant. Percent survival is calculated as for panel A VOL. 178, 1996 INORGANIC POLYPHOSPHATE SUPPORTS SURVIVAL OF E. COLf 1397 3H-Thy 100 .1 lL1 10 1 .1 1357911 Time (days) FIG. 5. Thymidine (Thy) uptake during long-term incubation of stationary- phase wild-type (W1') (A) and CAlO (B) cells. 'The cells were pulse-labeled with ['Hlthymidine from the 3rd to 10th days. Both survival of the cells and the incorporation of ['Hlthymidine into cells were measured (Materials and Meth- ods). the small-colony phenotype. The viable cell number increased to about 5 X lo7 CFU/ml after 9 days; most o€ the cells (-90%) observed in four separate experiments were the small- colony variants. Takeover of the cell population by the small- colony phenotype was invariably accompanied by a burst of thymidine uptake. Characteristics of the small-colony variant. (i) Stable phe- notype. CAlO cells were grown to the stationary phase in LB medium with kanamycin, incubated aerobically at 37°C for 5 days, and plated on LB plates containing kanamycin. The sta- tionary-phase culture contained small- and big-colony pheno- types. The big colonies and the more slowly growing small colonies were isolated and maintained as stable phenotypes. The colony size was well defined for these phenotypes. When grown in MOPS-buffered medium for 48 h at 37"C, the colony diameters were 1.0 ? 0.5 mm for the small-colony phenotype and 3.0 2 0.5 mm for the big-colony phenotype. A few of the characteristics of these variants were studied relative to the role of poly P during survival in the stationary phase. (ii) Growth lag. When wild-type cells and the small- and big-colony variants of CAlO were grown in MOPS medium containing 2 mM K2HP0, and 0.4% glucose, the small-colony phenotype cells consistently showed a longer growth lag (Fig. 6). Generation times were 56 min for the wild-type and the CAlO big-colony phenotype cells; the CAlO small-colony cells grew with a generation time of 60 min. (iii) Stationary-phase survival. The big- and small-colony variants of CAlO were grown in LB medium to the stationary phase and incubated aerobically at 37°C for 10 days (Fig. 7). t I,IIIIII 2 4 6 8 Time (hours) FIG. 6. Growth of wild-type (WT) cells and CAlO cells exhibiting either the big-colony phenotype or the small-colony phenotype. Prewarmed (37°C) MOPS medium was inoculated with overnight (-20 h) cultures (diluted 1 to 100); growth was monitored by measuring the turbidity at The big-colony phenotype, like its parent strain, showed a considerable decrease in viability; after 5 days, only about 1% of the original population had survived. On the other hand, the small-colony phenotype showed an improved viability with long-term incubation; about 10% of the population survived until 8 days (Fig. 7). (iv) Heat resistance. Upon entry into the stationary phase, wild-type cells develop a pronounced tolerance to heating at 55 to 57°C (28). Stationary-phase cells of the wild-type and big- and small-colony phenotypes of CAlO were exposed to 55°C and their viability was assessed (Fig. 8). The survival profile of the big-colony phenotype was similar to that of the parent CAlO strains; less than 1% of the cells survived heating for 2 min compared with over 90% of the wild-type cells (JM101). The small-colony variant showed a significantly higher level of resistance towards heat shock than did the big-colony pheno- type; about 60% of the small-colony cells survived the 2-min 0.1 2 4 6 8 10 Time (days) FIG. 7. Comparison of rates of long-term survival of big- and small-colony phenotypes of CAlO cells in the stationary phase. The cells were grown and incubated aerobically in LB medium at 37°C. Percent survival IS calculated as described in the legend to Fig. 1. 1398 RAO AND KORNBERG J. BACTERIOL. I 12345 Time (min) FIG. 8. Heat resistance. Wild-type (WT) cells and big- and small-colony phenotypes of CAlO were grown overnight (-20 h) at 37°C in LB medium and exposed to 55°C; percent survival is calculated as described in the legend to Fig. 1. No viable cells were detected beyond the 3-min timc point for the CAlO big-colony phenotype. (v) PPK and PPX activities. To determine whether the en- hanced survival of the small-colony phenotype was due to a reversion or suppression of the PPK and PPX mutant pheno- type, the enzyme activities were compared (Table 2). CAlO cells still showed very low levels of both PPK (-2% of the wild-type level) and PPX activity. The residual PPX activity (about 6% of the wild-type level) in the mutant could be attributed to the guanosine pentaphosphate poly Pase (29). No significant difference was observed in the enzyme activities of the mutant strains showing either the big-colony or the small- colony phenotype (Table 2). To confirm that the PPX protein was not being expressed in the mutants, particulate fractions obtained after lysis of the cells by lysozyme were analyzed by immunoblotting with an anti-PPX antibody. Neither the big- colony phenotype nor the small-colony phenotype of CAlO showed the presence of PPX (data not shown). Similarly, a Western blot (immunoblot) analysis of a trichloroacetic acid precipitate of the total cell protein of the mutant phenotypes of CAlO failed to indicate the presence of PPK (data not shown). (vi) Poly P content. The poly P content of the stationary- phase wild-type cultures was 678 ? 19 pmol/mg of protein; the levels for the CAlO big- and small-colony phenotypes were at the limit of detection (about 100 pmol/mg of protein). Poly P levels in an rpoS knockout mutant were compared with those of a corresponding isogenic wild-type strain, during both the exponential and stationary phases of growth. No significant difference in poly P contents was noticed between the wild type TABLE 2. PPK and PPX activities Enzyme activity (Uimg of protein)" PPK PPX JMlOl (wild type) 424 248 CA10-s (small colony) 8.0 1.5 CA10-b (big colony) 12 6.0 " The PPK assay measured the production of acid-insoluble [BzP]poly P from [y3'P]ATP (1). One unit of enzyme is defined as the amount incorporating 1 pmol of phosphate into acid-insoluble poly P per min at 37°C. The assay for PPX measured the loss of acid-insoluble ['zP]poly P. One unit of enzyme is defined as the amount that decreases the acid-insoluble poly P by 1 pmol of phosphate per min at 37°C. 80 20 CAlO (small) / (Small) rpoS::TnlO CA10 rpoS::TnlO /I 0 - 024601012 Time (rnin) FIG. 9. Heat resistance. CAlO small-colony phenotype, CAlO smallcolony phenotype rpoS::TnlO, and CAlO rpoS::TnlO cells were grown overnight (-20 h) at 37°C in LB medium and exposed to 55°C; percent survival is cdkuhted as described in the legend to Fig. 1. and the rpoS::TnlO strain, especially in cells during stationary- phase growth (data not shown). (vii) Effect of rpoS on heat resistance. The enhanced heat resistance acquired by the stationary-phase small-colony phe- notype might be attributed to a secondary mutation in rpoS. In order to test this hypothesis, rpoS mutant derivatives of the small-colony phenotype, as well as that ol the original ppk strain (CAlO), were constructed. The stationary-phase cells of these rpoS knockout mutants and the cells of the small-colony phenotype were tested for their heat resistance (Fig. 9). Intro- duction of the rpoS mutation to the small-colony ppk mutant resulted in reduced heat tolerance at 55T, but this effect was less pronounced than that observed with the rpoS mutant of the CAlO cells, thus failing to identify an rpoS locus for the small-colony phenotype; rpoS knockout mutant derivatives of both the small-colony phenotype and the CAlO cells were found to be starvation sensitive, and only about 10% of the stationary-phase cells survived alter 20 h of incubation in LB medium. (viii) Catalase activity. Stationary-phase cells exhibit strong rpoS-dependent resistance to H,O,. Catalase HPII, the rpoS- controlled katE gene product, is induced in stationary-phase cells as a defense against H,O, (37). Catalase activities of the wild type and of the small- and big-colony phenotypes of CAlO cells were measured in cells grown to the stationary phase in LB medium (Table 3). The small-colony phenotype contained about 60% of the activity seen in the wild-type cells; the big- colony phenotype contained only 35%. When the small-colony variant was complemented with a plasmid, pBC29, which car- ries the ppk gene, the catalase activity was restored to the wild-type level. However, these catalase activity measurements TABLE 3. Catalase activities of strains used in this study Strain Phenotype Activity (Ulmg of protein)" JMlOl PPK+ (wild type) 32 CAlO PPK- (small colony) 19 CAlOpBC29 PPK+++ (small colony) 31 CAlO PPK- (big colony) 11 * The activities of permeabilized stationary-phase cells were measured. One unit is equal to 1 kmol of hydrogen peroxide decomposed per min at 25°C. VOL. 178, 1996 INORGANIC POLYPHOSPHATE SUPPORTS SURVIVAL OF E. COLI 1399 fail to differentiate between the rpoS-independent HPI and the tpoS-dependent HPII activities. Therefore, extracts of station- ary-phase cells, analyzed qualitatively on a nondenaturing polyacrylamide gel, were stained for catalase activity in the HPI and HPII bands. No significant difference in HPI activity was observed between the wild-type and mutant cells (both small- and big-colony phenotypes). The stationary-phase wild- type cells possessed a high level of HPII activity; however, the mutants showed only trace amounts of this enzyme activity. Complementation of either CAlO cells or the CAlO small- colony phenotype with the ppk gene (pBC29) or an rpoS plas- mid (pMMkatF3) elevated synthesis of HPII to the wild-type level (data not shown). DISCUSSION The stationary phase of growth studied in the laboratory can resemble the stressful and deprived state that characterizes the natural habitat of most bacteria (47). To cope with adversity, cells entering the stationary phase undergo drastic changes in cell physiology and morphology (48). Many genes not required for exponential growth are induced in order that cells may survive in the stationary phase (6, 19, 20, 34, 50). Mutations in these genes in E. coli decrease viability during the prolonged periods of nutritional deprivation in a stressful environment. Among them, as the present studies show, isppk, the gene that encodes PPK, the enzyme responsible for making long-chain Stresses that E. coli responds to or anticipates include oxi- dants, heat, and high salt levels (27, 28). Aerobic metabolism produces reactive byproducts such as superoxide and hydrogen peroxide (15, 17) that are involved in the generation of cyto- toxic hydroxyl radicals (7, 23, 24). In log-phase cultures, genes induced to cope with these oxidants are found in the SOXR and oxyR regulons (11,17, 18). We report here that theppk mutant, lacking poly P, is more sensitive to hydrogen peroxide than the wild type (Fig. 2). Also, in stationary-phase cultures, the heightened resistance to hydrogen peroxide is diminished in the ppk mutant (10). This resistance to hydrogen peroxide is largely restored to the wild-type level when the mutant is complemented with a plasmid bearing the ppk gene. Another example of oxidant stress is the toxicity of menadi- one, the one-electron redox agent that generates cytotoxic oxidants as in the redox cycling of respiratory chain compo- nents during aerobic metabolism (58). Catalase activities in- duced to scavenge oxidants such as hydrogen peroxide include HPI in log-phase cells and HPII in stationary-phase cells (37). The ppk mutant proved to be hypersensitive to menadione (Table l), presumably because of a deficiency in catalase ac- tivities (Table 3), particularly that of HPII because of the lack of its induction in the stationary phase (17, 26). In addition to increased sensitivity to oxidants, responses to heat and high salt levels were also examined in theppk mutant, and resistance to these agents was found to be greatly dimin- ished. After exposure to 55°C for 2 min, only 2% of the mutant cells in the stationary phase survived compared with about 90% of the wild-type cells (Fig. 1). With regard to the chal- lenge of a medium with 2.5 M NaCl, the mutants lost viability to a greater extent than wild-type cells in both the log and stationary phases (Fig. 3). The most striking phenotype of the ppk mutant is its poor rate of survival in stationary-phase cultures (10) (Fig. 4). Via- ble cell counts decreased drastically within a few days upon aerobic incubation, the more so in a rich medium. Of mutants in a medium limited in carbon source (i.e., glucose), about 1% survived after 9 days, whereas in a rich medium only about 1 in poly P. lo6 cells survived after 4 days. A similar influence of growth medium on the long-term survival of wild-type cells has also been observed (52). Although the bases for the dramatic poly P effects in pro- moting resistance to stressful agents and prolonging survival in the stationary phase are still unexplained, some insight may be gained from the suppressive influence of rpoS on the ppk mu- tation. The product of rpoS is the a" subunit of RNA poly- merase responsible for the expression of nearly 50 genes in- volved in adjustments in the stationary phase to nutrient deprivations, high osmolarity (21), and other stressful agents. When multiple-copy rpoS plasmids were introduced into the ppk mutant, heat resistance was elevated to the wild-type level (Fig. 1). RpoS plays a central role in the development of stress resistance in E. coli. Modulation of the concentration of this sigma factor brought about either by the exposure of the wild- type cells to various stresses or by the presence of a plasmid bearing rpoS in the ppk mutant may be responsible for the induction of proteins essential lor heat tolerance. Less direct, but suggestive of a relationship between rpoS and poly P, is the emergence of a novel small-colony phenotype inppk mutants (10). After maintenance of theppk mutant for several days in the stationary phase, the colonies of normal size are largely replaced by small colonies which are genetically stable. Their emergence coincides with a burst of DNA syn- thesis indicated by thymidine incorporation (Fig. 5). The small- colony variant demonstrates increased resistance to heat (Fig. 8) and an improved rate of survival (Fig. 7). Inasmuch as the levels of PPK activity and poly P in the variant remain at the mutant cell levels, some other genetic alteration can be in- ferred. The altered gene in the small-colony variant was initially assumed to be allelic to rpoS, since this locus is involved in adaptive mutations (49, 52, 61). Besides, rpoS plays a major regulatory role during survival in the stationary phase (25, 38, 55, 60). The rpoS-negative mutants did not lack poly P during both exponential and stationary-phase growth. Introduction of rpos:TnlO into the stable, small-colony variant did not result in the complete restoration of heat sensitivity in stationary-phase cells (Fig. 9). It appears that the second site suppressor muta- tion in the small-colony phenotype may act independently of tpos. ACKNOWLEDGMENTS This work was supported by a grant from the National Institutes of Health. REFERENCES 1. Akiyama, M., E. Crooke, and A. Kornberg. 1992. The polyphosphate kinase gene of Escherichia coli. Isolation and sequence of the ppk gene and mem- brane location of the protein. J. Biol. Chem. 267:22556-22561. 2. Akiyama, M., E. Crooke, and A. Koroberg. 1993. An exopolyphosphatase 01 Escherichia coli. The enzyme and itsppx gene in a polyphosphate operon. J. Biol. Chem. 268:633-639. 3. Almirh, M., A. J. Link, D. Furlong, and R. Kolter. 1992. A novel DNA- hinding protein with regulatory and protective roles in starved Escherichia coli. Genes Dev. 6~2646-2654. 4. Archibald, F. S., and I. Fridovich. 1982. Investigations of the state of the manganese in Lacfobacillus plantarum. Arch. Biochem. Biophys. 215:589- 596. S. Beers, R. F., Jr., and 1. W. Sizer. 1952. A spectrophotometric method for measuring the breakdown of hydrogen peroxide by catalase. J. Biol. Chem. 195133-135. 6. Bohannon, D. E., N. Connell, J. Keener, A. Tormo, M. Espinosa-Urgel, M. M. Zambrano, and R. Kolter. 1991. Stationary-phase-inducible "gear- box" promoters: differential effects of kulF mutations and role of u'". J. Bacteriol. 173:4482-4492. 7. Bol, D. IC, and R. E. Yasbin. 1990. Characterization of an inducible oxidative stress system in Bacillus suhfilis. .I. Bacteriol. 1723503-3506. 8. Bonting, C. F. C., G. J. J. Kortstee, and A. J. B. Zehnder. 1991. Properties of 1400 RAO AND KORNBERG J. BACrERIOL. po1yphosphate:AMP phosphotransferase of Acinetobacter strain 210A. J. Bacteriol. 173:64844488. 9. Clare, D. A, N. Duong, D. Darr, F. Archibald, and I. Fridovich. 1984. Effects of molecular oxygen on the detection of superoxide radical with nitro blue tetrazolium and an activity stain for catalase. Anal. Biochem. 140532-537. 10. Crooke, E., M. Akiyama, N. N. Rao, and A. Kornberg. 1994. Genetically altered levels of inorganic polyphosphate in Escherichia coli. J. Biol. Chem. 2696290-6295. 11. Demple, B. 1991. Regulation of bacterial oxidative stress genes. Annu. Rev. Genet. 25:315-337. 12. Demple, B., and J. Halbrook 1983. Inducible repair of oxidative DNA damage in Eschenchra coli. Nature (London) 304:46h-468. 13. Dirheimer, G., and J. P. Ebel. 1964. Sur le m&taholisme des polyphosphates inorganiques chez Coynrbacterium xerosis. C. R. SCances Soc. Biol. 158: 1948-1951. 14. Dum, T., K. Gable, and T. Beeler. 1994. RegukdtioII of cellular Ca'+ by yeast vacuoles. J. Bid. Chem. 2697273-7278. 15. Fridovich, I. 1978. The biology of oxygen radicals. Science 201:875-880. 16. Gentry, D. R., V. J. Hernandez, L. H. Nguyen, D. B. Jensen, and M. Cashel. 1993. Synthesis of the stationary-phase sigma factor us is positively regulated by ppGpp. J. Bacterid. 1757982-7989, 17. Greenbere, J. T.. and B. Demple. 1988. Overproduction of peroxide-scav- 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28 29 30 31 32 33 34 35 enging enzymes in Lscherichiu coli suppresses spontaneous mutagenesis and sensitivity to redox-cycling agents in oxy R- mutants. EMBO J. 7:2611-2617. Greenberg, J. T., and B. Demple. 1989. A global response induced in Esch- mri~hru ~uli by redox-cycling agents overlaps with that induced by peroxide stress. J. Bacterid. 171:3933-3939. Hengge-Aronis, R. 1993. Survival of hunger and stress: the role of rpoS in early stationary phase gene regulation in E. coli. Cell 72165-168. Hengge-Aronis, R., W. Klein, R. Lange, M. Rimmele, and W. Boos. 1991. Trehalose synthesis genes are controlled by the putative sigma factor en- coded by rpoS and are involved in stationary-phase thermotolerance in Eschericliiu coli. J. Bacterid. 173:7918-7924. Hengge-Aronis, R., R. Lange, N. Henneberg, and D. Fischer. 1993. Osmotic regulation of rpoS-dependent genes in Escherichia coli. J. Bacterid. 175: 259-265. Hsieh, P.-C., B. C. Shenoy, J. E. Jentoft, and N. F. Phillips. 1993. Purification of polyphosphate and ATP glucose phosphotransferase from Mycobacterium tubercrrlosLs H37 Ra: evidence that poly (P) and ATP glucokinase activities are catalyzed by the same enzyme. Protein Expr. Purif. 476-84. Imlay, J. A, S. M. Chin, and S. Linn. 1988. Toxic DNA damage by hydrogen peroxide through the Fenton reaction in vivo and in vitro. Science 240 640442. Imlay, J. A., and S. Linn. 1988. DNA damage and oxygen radical toxicity. Science 24031302-1309. Ishihama, A. Personal communication. Ivanova, A,, C. Miller, G. Glinsky, and A. Eisenstark 1994. Role of rp~S (kulF) in oxyR-independent regulation of hydroperoxidase I in Escherichiu ~~li. Mol. Microbiol. 12571-578. Jenkins, D. E., S. A. Chaisson, and A. Matin. 1990. Starvation-induced cross protection against osmotic challenge in Escherichia coli. J. Bacteriol. 172: 2779-2781. Jenkins, D. E., J. E. Schnltz, and A. Matin. 1988. Starvation-induced cross protection against heat or H,O, challenge in Escherichiu ~li. J. Bacteriol. 170:3910-3914. Keasling, J. D., L. Bertsch, and A. Kornberg. 1993. Guanosine pentaphos- phate phosphohydrokase of Escherichia coli is a long chain exopolyphos- phatase. Proc. Natl. Acad. Sci. USA 9k7029-7033. Kornberg, A. 1995. Inorganic polyphosphate: toward making a forgotten polymer unforgettable. J. Bacterid. 177:491496. Knlaev, I. S. 1979. The biochemistry of inorganic polyphosphates. Wiley. New York. Kulaev, I. S., and V. M. Vagabov. 1983. Polyphosphate metabolism in mi- croorganisms. Adv. Microbiol. 15731-738. Lange, R., M. Barth, and R. Hengge-Aronis. 1993. Complex transcriptional control of the &-dependent stationary-phase-induced and osmotically regu- lated osniY (csi-j) gene suggests novel roles for Lrp, cyclic AMP (CAMP) receptor protein-cAMP complex. and integration host factor in the station- ary-phase response of Escherichia coli. J. Bacteriol. 175:791(!-7917. Lange, R., and R. Hengge-Aronis. 1991. Growth phase-regulated expression of bo& and morphology of Stationary-phase Escherichiu coli cells are con- trolled by the novel sigma factor 0`. J. Bacteriol. 173:44744481. Li, C., and S. Clarke. 1992. A protein methyltransferase specific for altered aspartyl residues is important in Eschenchia coli stationary-phase survival and heat-shock resistance. Proc. Natl. Acad. Sci. USA 89:9885-9889. 36. Liu, S., and A. Kornberg. Unpublished results. 37. Loewen, P. C., J. Switala, and B. L. Triggs-Raine. 1985. Catalases HPI and HPII in Eschenchiu coli are induced independently. Arch. Biochem. Biophys. 243:144-149. 38. McCann, M. P., J. P. Kidwell, and A. Matin. 1991. The putative u factor KdtF has a central role in development of starvation-mediated general re- sistance in Escherichia coli. J. Bacteriol. 173:41884194. 39. Miller, J. H. 1972. Experiments in molecular genetics. Cold Spnng Harbor Laboratory Press, Cold Spring Harhor;N.Y. 40. Mnlvey, M. R., P. A. Sorby, B. L. Triggs-Raine, and P. C. Loewen. 1988. Cloning and physical characterization of katE and htF required for catalase HPII expression in Escherichia coli. Gene 73:337-345. 41. Neidhardt, F. C., P. L. Bloch, and D. F. Smith. 1974. Culture media for enterobacteria. J. Bacteriol. 119736-747. 42. Phillips, N. F., P. J. Horn, and M. G. Wood. 1993. The polyphosphate- and ATP-dependent glucokinase from Propionibacterium shermanir: both activi- ties are catalyzed by the same protein. Arch. Biochem. Biophys. 300309-319. 43. Pick, U., M. Bental, E. Chitlarn, and M. Weiss. 1990. Polyphosphate hydro- lysis-a protective mechanism against alkaline stress? FEBS Lett. 27415-18. 44. Pick, U., and M. Weiss. 1991. Polyphosphate hydrolysis within acidic vacu- oles in response to amine-induced alkaline stress in the halotolerant alga Dundiella salmu. Plant Physiol. 911'234-1240. 45. Rao, N. N., M. F. Roberts, and A. Torriani. 1985. Amount and chain length of polyphosphates in Escherichiu coli depend on cell growth conditions. J. Bacteriol. 162242-247. 46. Reusch, R. N., and H. L. Sadoff. 19x8. Putative structure and functions of a poly-P-hydroxybutyrateicalcium polyphosphate channel in bacterial plasma membranes. Proc. Natl. Acad. Sci. USA 8541764180. 47. Roszak, D. B., and R. R. Colwell. 1987. Survival strategies of bacteria in the natural environment. Microhiol. Rev. 51:365-379. 48. Siegele, D. A,, and R. Kolter. 1992. Life after log. J. Bacterid. 174345-348. 49. Siegele, D. k, and R. Kolter. 1993. Isolation and characterization of an Eschrrichiu mutant defective in resuming growth after starvation. Genes Dev. 12629-2640. 50. Tanaka, K., Y. Takayangi, N. Fujita, A. Ishihama, and H. Takahashi. 1993. Heterogeneity of the principal u factor in Escherichia coli: the rpoS gene product, 038. is a second principal u factor of RNA polymerase in stationary- phase Escherichia coli. Proc. Natl. Acad. Sci. USA W3511-3515. 51. Thor, €1.. M. T. Smith, P. Hartzell, G. Bellomo, S. A. Jewell, and S. Orrenius. 1982. `The metabolism of menadione (2-methyl-l,4-naphthoquinone) by iso- lated hepatocytes. J. Bid. Chem. 257:12419-12425. 52. Tormo, A, M. AlmirBn, and R. Kolter. 1990. srrrA, an Esclierichiu colt gene essential for survival in stationary phase. J. Bacteriol. 17243394347. 53. Van Groenestijn, J. W., M. Zuidemd, J. J. M. Van De Worp, M. H. Deinema, and A. J. B. Zehnder. 1989. Regulation of polyphosphate accumulation in Aciriefobacter sp. Antonie Leeuwenhoek 5567-82. 54. Van Veen, H. W., T. Abee, G. J. J. Kortstee, W. N. Konings, and A. J. B. Zehnder. 1993. Characterization of two phosphate transport systems inAcin- etobucter johru.oilii 210A J. Bacterid. 175:200-216. 55. Wada, A., Y. Yamazaki, N. Fujita, and A. Ishihama. 1990. Structure and probable genetic location of a "ribosome modulation factor" associated with 100s ribosomes in stationary phase Escherichiu coli cells. Proc. Natl. Acad. Sci. USA 87:2657-2661. 56. Weichart, D., R. Lange, N. Henneberg, and R. Iiengge-Aronis. 1993. Iden- tification and characterization of stationary-phase induclblc genes in Esch- erichio wli. Mol. Microbiol. 10407420. 57. Wood, H. G., and J. E. Clark 1988. Biological aspects of inorganic polyphos- phates. Annu. Rev. Biochem. 57335-260. 58. Wu, J., and B. Wiess. 1991. Two divergently transcribed genes, soxR and soxS, control a superoxide response regulon of Escherichiu cdi. J. Bacterrol. 173:2864-2871. 59. Wurst, H., T. Shiba, and A. Kornberg. 1995. The gene for a major ex- opolyphosphatase of Sacclirirumycrs urevisiue. J. Bacterid. 177:898-906. 60. Yamagishi, M., H. Matsushima, A. Wada, M. Sakagami, N. Fujita, and A. Ishihama. 1993. Regulation of the Escherichia coli rnf gene encoding the ribosome modukdtion factor: growth phase and growth rate-dependent con- trol. EMBO J. 12:625-630. 61. Zambrano, M. M., D. A. Siegele, M. Almiron, A. Tormo, and R. Kolter. 1993. Microbial competition: Escherichiu coli mutants that take over stationary phase cultures. Science 2591757-1760.