Tue Jovgna. or BiovocicaL CHemistry © 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 269, No. 9, Issue of March 4, pp. 6290-6295, 1994 Printed in U.S.A. Genetically Altered Levels of Inorganic Polyphosphate in Escherichia coli* (Received for publication, October 18, 1993, and in revised form, December 13, 1993) Elliott Crooke}, Masahiro Akiyama§, Narayana N. Rao, and Arthur Kornberg From the Department of Biochemistry, Stanford University School of Medicine, Stanford, California 94305 The ppk gene encoding polyphosphate kinase (PPK), the enzyme in Escherichia coli that makes long chains of polyphosphate (polyP) reversibly from ATP, was dis- rupted by insertion of a kanamycin resistance gene. Ex- pression of the exopolyphosphatase gene (ppx) immedi- ately downstream of ppk in the operon was likewise disrupted. Cells were also transformed with a high- copy-number plasmid bearing ppk. Genetically altered polyP levels were estimated in cell extracts by the PPK conversion of ADP to ATP. PolyP levels (ug/10™ cells) near 2.0 were reduced in the ppk~-ppx- mutants to 0.16 and increased more than 100-fold (e.g. 220) in cells trans- formed with multiple copies of ppk. Mutant cells, lack- ing the long polyP chains, showed a growth lag following dilution of a stationary-phase culture. PolyP-deficient cells exhibit a striking phenotype in their failure to sur- vive in stationary phase and loss of resistance to heat (55 °C) and to oxidants (42 mm H,0,). High polyP levels are also associated with reduced survival. Inorganic polyphosphate (polyP),! a long linear polymer of orthophosphates linked by high energy phosphoanhydride bonds, is widely distributed in bacteria, fungi, protozoa, plants, and mammals (1-3). Yet little is known about its cellular func- tions. Potential roles include: (i) energy source (1, 2), Gi) phos- phate reservoir (1), (iii) donor for sugar and adenylate kinases (4-6), (iv) chelator for divalent cations (7, 8), (v) buffer for alkaline stress (9), (vi) regulator of transcription,? and (vii) component in competence for DNA entry and transformation (10). A membrane-associated enzyme in Escherichia coli, poly- phosphate kinase (PPK), catalyzes the synthesis of polyP from the terminal phosphate of ATP in a freely reversible reaction (nATP <— nADP + polyP,,) (11-13). Cloning of the ppk gene (14) also identified an adjacent gene, ppx, which encodes an exo- polyphosphatase (PPX), an enzyme which processively releases orthophosphate from the termini of polyP in the reaction (polyP, —~ pelyP,_1 + P;) (15). Expression of ppx is dependent upon the ppk promoter, indicative of a polyphosphate operon (15). * This work was supported in part by grants from the National In- stitutes of Health and the National Science Foundation (to A. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adver- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. + Supported by American Cancer Society, California Division Senior Postdoctoral Fellowship 8-26-91. Present address: Dept. of Biochemis- try and Molecular Biology, Georgetown University Medical Center, Washington, D. C. 20007. § Present address: Cold Spring Harbor Laboratory, Cold Spring Har- bor, NY 11724. 1 The abbreviations used are: polyP, polyphosphate; PPK, polyphos- phate kinase; PPX, exopolyphosphatase; MOPS, 4-morpholinepropane- sulfonic acid; kb, kilobase pair(s). 2 A. Ishihama, personal communication. In attempts to elucidate the functions of polyP, we have con- structed mutants that fail to express ppk and ppx as well as cells which overproduce PPK. The consequences of altering the level of polyP 1,000-fold have been examined. MATERIALS AND METHODS Reagents—Sources were: ATP, ADP, creatine kinase, DNase I, and RNase IIJa, Boehringer Mannheim; creatine phosphate, MOPS, and polyP glass (type 65), Sigma; [a-?P]dATP and (y-°*P]ATP, Amersham Corp.; restriction endonucleases, New England Biolabs; Immobilon-N membrane, Millipore; and polyethyleneimine-cellulose F thin-layer chromatography plates, Merck. Bacterial Strains, Plasmids, and Phages—E. coli K12 derivatives were: DH5a (F-, supE44, A(lacZYA-argF)U169[80lacZAM15], hsdR17, recAl, endA1, gyrA96, thi-1, relAl, deoR), SC864 (F-, polA12(ts), thi-1, thr-1, leuB6, lacY1, fhuA21, supE44, fad751::Tn10, d—); JM101 (Aflac- proAB), supE, thi-1/F'{traD36, proAB*, lacI*, lacZAMI15Jj; CP1648 (F-, A-); NK6056 (AC gpt-lac)5, purC80::Tn10, relAl, spoT 1, thi-1, A~); N3007 (7, gua-26::Tn 10 (in guaA or guaB), IN(rrnD-rrnE)1). DHSa was the host strain for plasmid preparations. SC864 was used to construct ppk- mutants by inserting, through homologous recombination, the kan gene into ppk on the E. coli chromosome, as described below (Fig. 2). The mutated ppk genes were transferred into JM101 by Pl transduction (16); CA10 contains the kan gene in the same orientation as ppk, whereas CA11 contains kan in the opposite orientation. NK6056 and N3007 were used to map the integration site of kan on the chromosome. Phage d 10H6 (14, 17) was from the Kohara A library (17). Plasmid pBC8 (Figs. 2 and 4) was used as a probe for Southern blotting. Plasmid pBC29 used for polyP overexpression is a multicopy plasmid bearing ppk (15). Survival Assays—Long-term survival of FE. coli was assayed by grow- ing cells in MOPS-buffered minimal medium (18) containing glucose (0.1%) as the carbon source and K,HPO, (2 mm) as the P, source. Cul- tures (3 ml) incubated at 37 °C for prolonged periods in glass test tubes (18 by 200 mm) were aerated by rotation in a New Brunswick gyratory wheel. Viable cell counts were determined by plating onto MOPS-buff- ered minimal agar medium containing 0.4% glucose and 2 mm K,HPO,; kanamycin (50 pg/ml) was added to the medium on which the mutant CA10 was plated. For the heat-shock survival assay, cells were grown overnight (~20 h) in LB. The stationary-phase cells were washed and diluted in 0.9% NaCl to a density of about 5 x 10° cells/ml. Samples (2 ml) were put in glass tubes prewarmed to 55 °C and at times indicated, aliquots (0.1 ml) were plated directly on LB plates to determine viable cell numbers. To test sensitivity to H2Oz, cells were grown overnight (~20 h) in LB, washed, and resuspended in 0.9% NaCl to an ODg4o of 1.0. H,O2 waa added to a final concentration of 42 mm. At the times indicated, 0.1-ml samples were withdrawn, diluted immediately in 0.9% NaCl and plated onto LB plates to determine viable cell numbers. Extraction of PolyP from E. coli-LB medium (1,500 m!) was inocu- lated to an optical density (Agoo nm) Of 0.05 with a sample of an overnight culture and incubated at 37 °C. Cells were harvested by centrifugation (6,000 x g, 7 min, 2 °C) when the culture reached mid-log phase growth. The resulting cell pellet was resuspended and lysed in ice-cold 2% trichloroacetic acid (0.6 ml); the suspension was maintained at 0 °C for 30 min with occasional vigorous mixing. Acid-insoluble material was collected by centrifugation (3 min, 14,000 x g, 2 °C), and the pellet was washed with ice-cold 67% acetone (1 ml) and collected by centrifugation, aa above. The pellet was resuspended in 50 mm HEPES-KOH, pH 7.5 (0.5 ml), and the pH was adjusted to neutrality by the addition of 0.2 m KOH. MgCl, (5 mo), DNase I, and RNase ila (300 g/m! of each) were added, and the mixture was incubated at 37 °C for 30 min. Proteinase 6290 Inorganic Polyphosphate in E. coli 100 ATP Formed (pmol) jh 4 } 1 0 100 200 300° “ 500 PolyP Added (pmol of Pi) Fic. 1. Enzymatic assay for polyP. Synthetic polyP glass (type 65) solutions of known concentration (see “Materials and Methods”) were incubated with [1*C-UJADP and purified PPK. The level of ATP formed was determined as described (see “Materials and Methods”). K was added (600 pg/ml) and the incubation was continued for an additional 45 min. The mixture (approximately 1 ml) was extracted with phenol:chloroform (0.5 ml, 1:1, equilibrated with 0.1 mM ammonium acetate), the phases were separated by centrifugation (5 min, 14,000 x g&, room temperature), only the aqueous phase was removed and dis- carded. Additional portions of phenol:chloroform (0.2 ml) and 50 mm HEPES-KOH, pH 7.5 (0.5 ml), along with EDTA (to 10 mm) were added to the interfacial material and organic phase. The sample was mixed vigorously, incubated at 37 °C for 5 min, and the phases were separated by centrifugation, as above. The aqueous phase was removed and re- tained, Extraction of the organic phase was repeated three more times with 50 mm HEPES-KOH, pH 7.5, 10 mm EDTA (0.25 ml/extraction), and all of the aqueous phases containing EDTA were pooled. The pooled sample was treated twice with water-saturated chloroform (1 m/l/ extraction), followed by the addition of ice-cold sec-butanol (1.0 ml). The mixture was kept at -20 °C for 30 min or more, and the precipitated polyP was collected by centrifugation (15 min, 14,000 x g, 2 °C). The pellet was washed with ice-cold acetone (1 ml), dried under vacuum, and resuspended in water (0.2 ml). Assay for PolyP—The ability to serve as a substrate for the PPK- catalyzed conversion of ADP to ATP was used as an assay to quantitate levels of polyP. The reaction mixture (10 pl) contained: 50 mm HEPES- KOH, pH 7.2, 40 mm (NH,).S0,, 4 mu MgCly, and 11.6 ps [?4C-UJADP (0.1 Ci/mmo}), 2,000 units PPK, and polyP as indicated. The reaction (37 °C for 45 min) was terminated by chilling to 0 °C and the addition of ADP and ATP (to 5 mm of each). Aliquots were spotted onto a poly- ethyleneimine-cellulose F thin-layer chromatography plate; polyP, ADP, and ATP were resolved using 0.4 m LiCl, 1.0 m HCOOH as a solvent system and visualized with UV-irradiation. The radiolabel that re- mained at the origin and that which migrated with ADP, ATP, and the solvent front was quantitated by liquid scintillation counting. With ADP in excess, PPK catalyzed the almost complete consumption of polyP, yielding an equivalent amount of ATP (Fig. 1). Other Procedures—PPK and PPX were assayed as described (14, 15). PPK (3 x 10” units/mg) was purified as described (14). [°?P]PolyP was synthesized with PPK from [y-22P]ATP and purified (15). The concen- trations of polyP glass (type 65) solutions were determined by measur- ing total phosphate (19). RESULTS Disruption of the ppk Gene—A 9-kb fragment of E. coli DNA spanning both ppk and ppx was cloned from phage A10H6 (14, 17) into the vector pBR322 (Fig. 2A). The EcoRI site near the middle of the fragment is approximately 0.3 kb downstream of the start codon for ppk (14). A 1.3-kb DNA fragment of plasmid pUC4K (Pharmacia LKB Biotechnology Inc.) containing the kan gene was inserted at this EcoRI site in either orientation relative to the ppk gene to produce plasmids pBC22 and pBC24 (Fig. 2B). The wild-type ppk gene on the E. coli chromosome was re- 6291 placed with the disrupted ppk gene by exploiting the require- ment for functional DNA polymerase I to replicate plasmids with a ColE1 origin. A strain (SC864) with a temperature- sensitive DNA polymerase I was transformed with pBC22 and pBC24 at a permissive temperature (30 °C). In addition to con- taining the inserted gene for kanamycin resistance, the plas- mids also confer resistance to ampicillin through their pBR322 component. During growth at the permissive temperature, plasmids may integrate into the host chromosome by recombi- nation between homologous regions of the plasmid and the chromosome (integration), or a part of the chromosome may be replaced with the corresponding cloned DNA by a double cross- over event (replacement) (Fig. 3). When such cells are shifted to the nonpermissive temperature (42 °C) for DNA polymerase I function, plasmids are lost. Only those cells with a kan gene inserted into their chromosome can survive exposure to kana- mycin. Ampicillin- and kanamycin-resistant transformants were se- lected at the permissive temperature of 30°C. Cultures of transformants were grown overnight at 30°C in LB medium (16) containing ampicillin (50 pg/ml) and kanamycin (50 pg/ ml). Aliquots of each culture were used to inoculate LB medium containing kanamycin (50 pg/ml) for overnight growth at 42 °C. Appropriate dilutions of the cultures were spread onto LB agar plates containing kanamycin (50 pg/ml), and the plates were incubated at 42 °C. The resulting kanamycin-resistant colonies were screened for sensitivity to ampicillin. Kanamycin-resist- ant, ampicillin-sensitive cells (ppk::kan) were isolated from SC864/pBC22 and SC864/pBC24, and the disrupted ppk genes were transferred to the strain JM101 by P1 transduction yield- ing strains CA10 and CA11, respectively. Confirmation of Disrupted ppk Genes—A 4.4-kb SalI frag- ment upstream of ppk was subcloned from A10H6 into the vector pUC18, and the resulting plasmid, pBC8, was used as a probe for Southern blotting. The ppk- mutants were examined by Southern blotting (Fig. 4). The radiolabeled probe pBC8 hybridized to a 4.4-kb SalI fragment of genomic DNA from JM101 (Fig. 4A, Jane 1), but not from CA10 or CA11. Instead, CA10 and CA11 yielded a 5.1-kb fragment (Fig. 4A, lanes 2 and 3) as expected from examination of their physical map (Fig. 3B). The smaller band in lane 4 corresponds to vector plasmid pBR322. Since HindIII endonuclease digests the kan gene into 0.6- and 0.7-kb fragments, the orientation of the inserted kan gene relative to ppk was confirmed by Southern blotting of SalI and HindIlI-digested genomic DNA (Fig. 4B). Whereas the 4.4-kb fragment from JM101 was not affected by the additional digestion with HindIII (Fig. 4B, lane 1), the 5.1-kb fragments from CA10 and CA11 were altered, confirming the insertion of the kan gene into ppk; the size difference of the resulting frag- ments demonstrates that kan is in the same transcriptional orientation as ppk in CA10, whereas in the opposite orientation in CA11. Following a Sall-HindII digestion, the expected 0.9-kb fragment from CA10 and 1.0-kb fragment from CA11 were found to hybridize to pBC8 with a longer exposure of the gel (data not shown). Additionally, transduction frequencies with P1 phage revealed that the gene for kanamycin resistance in CA10 is closely linked to guaAB and purC, genetic markers close to ppk (20) on the E. coli chromosome (data not shown). PPK and PPX Activities in Mutant CA10 and CA1I Cells— Lysates prepared from JM101, CA10, and CA11 cells were as- sayed for PPK and PPX activities (Table I). JM101 cells con- tained PPK and PPX in quantities comparable with other wild- type cells (14, 15). However, in CA10 and CA1I, cells in which the ppk gene has been disrupted, the levels of both PPK and PPX were below the limit of detection. Immunoblot analysis using anti-PPK serum also failed to detect PPK in the ppk~ strains (Fig. 5); analysis with anti-PPX serum also confirmed 6292 Inorganic Polyphosphate in E. coli Fic. 2, Maps of phage and plasmid 52 38 DNAs. A, a region of phage A1OH6 (14, 7). The EcoR! site shown at the junction of the E&. coli DNA insert and the A DNA ppk pex kan come from the » DNA. The thick line above the map shows the 4.4-kb Sal] frag- ment cloned into pBCS. B, plasmids BE Ss EB HES E SB pBC22 and pBC24. Plasmids were con- B | Z ra ena) — BCR — ee structed as described in the text. Arrows DG indicate the direction of transcription. DNA fragment lengths are shown in kilo- base pairs. Symbols are: unfilled box, EB. coli DNA; filled hex, ppk gene: cross- hatched box, ppx gene; stippled box, 1.3-kb kan gene fragment; wavy line, A DNA; thin line, vector pBR322 DNA. Ab- breviations are: B, BamHlL E, EcoRI: H, E EBS E Hindil; S. Sali. C Plnnamnnnnnt BBORS ~~ eee bemnnd $Q : ~ tkb si Genet co ; ‘ ( ) Sal f Sal | -Hind 1 . Km’ Apr SS 2 Sot ® SS se foe $858 y VOR eS VM ae Ssé&s FFF 3 HD ££ $+ 9-30 & FEE & FEE & & & esd & Saga ogee 4123 4 1623 4 kb 8.0 74 — 6.0 —— see ati: om 4 are 3.0 --- pron a ~- vector 2 (ee ? Integration Km’ Ap‘ 2.0 -— we eee Replacement Km! Ap’ 1.0-~- Fic. 3. Strategy for the isolation of ppknkan mutants. Strains disrupted for ppk expression were isolated as described in the text. 0.5 — Symbols used are: unfilled box, cloned E. colt chromosomal DNA; cross- hatched box, region of the E. coli chromosome corresponding to the cloned DNA; *’, kan gene inserted into the cloned ppk gene; thick line, bacterial chromosome; thin line, vector pBR322; Km,, kanamycin re- sistance; Ap,, ampicillin resistance; Ap,, ampicillin sensitivity. the absence of PPX (data not shown). PolyP Content in Wild-type Cells, ppk-,ppx” Mutants, and Cells Overproducing PPK---PolyP was extracted from cells with wild-type (JM101 and JM101/pUC18) and mutant (CA10 and CALD levels of PPK and PPX and with overproduced (JM10V pBC29) (15) levels of PPK. The extraction procedure separated polyP from other phosphate-containing compounds in the cell lysate by its insolubility in cold acid and organic solvents, re- sistance to nucleases and proteases, and chelation by Mg** (see “Materials and Methods”). ["*P|PolyP, added to a portion of the initial cell lysates as a marker, was recovered with a yield of 90% or greater. The polyP content of extracts prepared from the various cell types was measured using the PPK conversion of ADP to ATP, Fic. 4. Southern blot analysis of ppk- mutant genomic DNA. E. coli genomic DNA was prepared from the ppk” strain (JM101) and the ppk- mutant strains (CA10 and CA11) as described (30). Genomic DNA and plasmid pBC8 were digested with: SalI (Aj or Sail and Hindi] endonucleases. Resulting fragments were separated by electrophoresis (B) through an agarose gel (1%), transferred to an Immobilon-N filter, and probed with radiolabeled pBC8 DNA. Lane 7,.JM101 genomic DNA: lane 2, CALO genomic DNA; lane 3, CA11 genomic DNA, lane 4, plasmid pBC8 DNA. as described above. Cellular polyP concentration was about 2.0 g/10" cells in the wild type cells (JM101) and wild-type cells transformed with the vector plasmid (Table ID). Another EF. coi wild-type strain CF1648 contained 0.44 pg of cellular polyP. Cells overexpressing PPK had polyP at levels over 100-fold Inorganic Polyphosphate in E. coli Tarr | Lysates from ppk~ strains lack PPK and PPX activities Strains (JM101, CA10, and CALL were grown in LB medium at 37 °C to an optical density (Ago nm! of 1. Cells were harvested by centrifuga- tion and lysed as described (14, 15). PPK activity was measured for the sonicated lysate pellet fraction (14), and PPX activity was determined for the soluble lysate fraction (15). Activities of: Strain PPK PPX units x 10 ‘ig cell JM101 320 2590 CAO <3.0 <8.3 CALL «2.4 <3.8 Purified PPK (ng) Culture (yb) 101 CAIN 6 18 60780 50 200 800 50 200 800 Mr ~ , oT — 7. 6e~ ie atl 43 —- 29 —- 1B Fic. 5. Cellular abundance of PPK. Cells (JM101, CA10) were grown at 37 °C in LB medium to an optical density (ODgoa nm! of 1.8. Total cell protein was precipitated from 1.3 ml of the culture with trichloroacetic acid (10% final concentration). The precipitate was har- vested in a Microfuge, washed twice with 1 ml of ice-cold acetone, and solubilized in SDS-PAGE buffer. Samples were subjected to SDS-PAGE (15%). PPK was determined by immunoblotting with PPK antiserum. Purified PPK served as a standard. Tasre Il PolyP content of cells with altered expression of ppk and ppx Extracts enriched for polyP were prepared (see “Materials and Meth- ods”) and their polyP content was determined as for solutions of polyP glass (Fig. 1). Strain Genotype PolyP content? ug? 10? cell JM101 Wild type 2.00 JIM101/pUCci18 Wild type 2.25 CA10 ppk” ppx” 0.16 CALI ppk ppx- 6.16 JM10UpBC29 ppk plasmid 220 7 Assays of PolyP extracted from three different cultures were aver- aged. higher than wild-type cells. PolyP levels in mutant cells (CALO and CA11) that lacked a functional ppk and ppx operon were 0.16 pg/10"" cells. The extracts from these mutant cells were not inhibitory to the PPK-catalyzed production of ATP, in that a mixture of CA10 extract and synthetically prepared polyP was indistinguishable from that of polyP alone (data not shown). Preliminary analysis of the polyP from the mutant cells indicated that they were predominantly short chain poly- mers with an average chain length of 60 (data not shown). It is likely that these short chain polyP are synthesized by an inde- pendent pathway. Cells Deficient in PolyP Exhibit a Growth Lag—-When JM101 (PPK*) and CA10 (PPK”) cells were grown on minimal medium (M9 + 0.2% glucose) plates (16}, the mutant cells showed slow growth (small colony size) during the early stages of the incu- bation (data not shown). However, with further growth the A Cells/ml Cells/ml 0.8 1.0 2c 3.9 4.0 Hours Fic. 6. Growth of wild-type cells and mutants lacking long chain polyP. Cells (dM101 and CA10) were grown in the indicated media to saturation. Prewarmed (37 °C) media were inoculated with samples of the stationary cultures and incubated at 37 “C. Appropriate dilutions of the cultures were plated onto media to monitor growth. Media were: A, LB (16) and B, M9 + 0.2% glucose (16). difference in colony size decreased. The initial growth of the two cell types was measured following the inoculation of vari- ous media with stationary-phase cultures. In a rich medium (LB), the cell number for wild-type cells started to increase after 1h, whereas the growth lag for mutant cells was repro- ducibly longer (Fig. GA); this lag was also observed using a minimal glucose medium (Fig. 6B). Effects of PolyP Deficiency on Long-term Survival in Station- ary-phase, Thermal Resistance, and Hy»O»g Resistance—E. coli mutant strain CA10 which lacks PPK and PPX exhibited re- duced survival in stationary phase (Fig. 7A). After 2 days in stationary phase in minimal medium with a limited carbon source (0.1% glucose), the viability of the mutants dropped to ~1% of the original value (Fig. 7A). The corresponding wild- type strain (JM101) did not show any significant loss in viabil- ity during this period. In addition to the loss in viability, a 6294 A 0 = = JM101 $ 10 > 5 0 CA1O IT T T_T T 1 3 § T 9 "i Time (days) B 100 = 7 e | JMi01 g ¥y ae 7 Cato 1 “T T T T T 0 1 2 3 4 § 6 Time (minutes) Cc wo = M101 z Ji g 2 3 CA10 + tT T T 0 20 40 60 80 Time (minutes) Fic. 7. Physiological tests. A, long-term survival of wild-type E. coli (JM101) and the polyP-deficient mutant (CA10) in MOPS-buffered minimal medium with limited glucose (0.1%, w/v). Percent survival represents the viable cell number at each time point divided by the viable cell number of the same culture after 24 h in stationary phase. Similar results were obtained in three experiments. B, heat-shock sur- vival. Cells were grown overnight (~20 h) at 37°C in LB medium, exposed to 55 °C, and the viability determined by plating on LB agar; 100% is equivalent to ~5 x 10° cells/ml. C, HO, sensitivity; stationary phase cultures were exposed to 42 mm H,Og, and viable cell numbers were determined as described under “Materials and Methods”; 100% corresponds to the viable cell number determined immediately before the addition of H,0,. stable small colony phenotype emerged from stationary-phase CA10 cells after 48 h in stationary phase (data not shown). During survival in stationary phase or starvation, £. coli cells acquire rpoS (kat F)-mediated resistance to multiple stresses, including heat and oxidants (25-28). When polyP-deficient cells (CA10) were held at 55 °C for 2 min, only 2% survived, as compared with about 90% of the wild type (JM101) (Fig. 7B). Similarly, of the polyP-deficient E. coli cells in stationary phase exposed to 42 mm H2Oz, only 40% survived as compared with about 85% of the wild type (Fig. 7C). DISCUSSION The ease with which orthophosphate condenses to form in- organic polyphosphate (polyP), a polymer linked by phospho- anhydride bonds, argues for a prominent role for the energy- rich compound in prebiotic evolution. In view of the ubiquity and abundance of polyP throughout nature, presumptions of multiple roles for this molecular fossil are reasonable, but ex- actly what these are must vary depending on where and when polyP serves in cellular functions. Inorganic Polyphosphate in E. coli In our approach to identify and examine polyP functions, we have purified enzymes responsible for the synthesis and utili- zation of polyP. The homogeneous enzyme provides a route to the discovery of the gene which encodes it and a means of modulating gene expression from depletion to overproduction of the enzyme. Phenotypes created by this “reverse genetics” may supply clues to the physiologic functions of polyP in cellu- lar growth, metabolism, and development. Often unappreci- ated is the immediate utility of the purified enzyme as a rea- gent to prepare labeled well defined substrates and as an analytical tool to determine the features and abundance of polyP in extracts of various cells and organisms. Previously, we identified the E. coli genes ppk for a polyP kinase (PPK) (14) and ppx for an PPX (15). These adjacent genes form an operon, the functions of which in polyP metabo- lism needs to be explored. To this end, we have constructed the mutant strains, reported in this study, in which the expression of ppk and ppx have been disrupted (Table I); overexpression of these genes, introduced in high copy number plasmids, has been obtained (14, 15). By these manipulations, the ijevels of polyP have been reduced more than 10-fold in the mutants and raised about 100-fold in the overproducer (Table ID, an overall spread in the level of 1,000-fold. The mutant cells are viable and show no striking phenotype as judged by growth in various media. Rates do not differ from wild-type cells in rich media at temperatures between 23 and 42 °C, in media limited in carbon (16) or in phosphate (21), at high ionic strength (e.g. 330-800 mm) or at decreased oxygen tensions (22). Nor were any growth rate differences observed between wild-type, mutant, and overproducer cells in a mini- mal medium containing a nonfermentable carbon source (e.g. succinate) or in the presence of dinitrophenol or azide at levels sufficient to reduce cellular ATP. The polyP content of wild-type cells grown on succinate was also similar to that of cells grown on glucose. The most suggestive indications of a deficiency in the ppk mutant are in the adjustments made in response to nutrient deprivation and for survival in the stationary phase. Guano- sine pentaphosphate hydrolase, one of the two enzymes (the other being RelA which synthesizes pppGpp) essential for mak- ing ppGpp in response to the stringency of amino acid starva- tion (23), has unexpectedly turned out to be a potent exopoly- phosphatase, different from PPX (24). Furthermore, polyP levels increased 10-fold or more in cells treated with serine hydroximate, an amino acid analog that induces (p)ppGpp pro- duction.? How polyP is involved in regulating promoter selec- tion in this and other circumstances remains to be determined. The ppk mutant is impaired in responses to stress and dep- rivation. There is a striking lability to heat and to hydrogen peroxide; survival in the stationary phase is also affected (Fig. 7), After the second day in a minimal medium with a limited carbon source at 37 °C, there is not only a greater loss of vi- ability in the mutant strain, but also the emergence of a small colony variant. These and other changes are suggestive of a selection for rpoS (katF) alleles that direct patterns of gene expression essential for adjustments to remaining viable in the stationary phase (25-28). Cells with elevated polyP levels (JM101/pBC29, Table II) attained only one doubling in cell density when grown in minimal medium. The viable cell num- ber at this stage was less than 4% of the wild type at stationary phase (data not shown). PolyP has been identified as a component of a complex with polyhydroxybutyrate and Ca?* in the membrane of bacteria competent for DNA transformation (10). The mutant strain with reduced levels of polyP can still acquire competence, al- 3 E. Crooke and A. Kornberg, unpublished data. Inorganic Polyphosphate in E. coli though with less efficiency than the wild type.* The competent mutant strains do possess a chloroform-extractable polyP chain of about 60 residues, presumably synthesized by a pathway other than PPK.5 Aremarkable ecological contribution made by PPK is in the bacterial removal of phosphate which pollutes waterways and causes algal blooms. In current sanitary engineering practice, an aerobic fermentation fixes phosphate in ATP, which is then converted by PPK to polyP and removed with the bacterial sludge (29). By transforming one of the bacterial strains with a high copy number plasmid bearing the ppk gene, phosphate removal from the surrounding medium became far more rapid and complete.® With a similar plasmid bearing both the ppk and ppx genes, overproduction of the PPX counteracts the ef- fectiveness of PPK, a clear demonstration of the actions in vivo of each of these enzymes. Acknowledgments—We thank Dr. Stanley Cohen (Stanford Univer- sity) for P1 phage and strain SC864, Dr. Barbara Bachmann (£. coli Genetic Stock Center, Yale University) for strains NK6056 and N3007, Dr. Michael Cashel (National Institutes of Health) for strain CF1648, and Dr. Yuji Kohara (National Institute of Genetics, Japan) for the Kohara phage library. REFERENCES 1. Kulaey, I. S. (1979) The Biochemistry of Inorganic Polyphosphates, John Wiley & Sons, New York 2. Wood, H. G., and Clark, J. E. (1988) Annu. Rev. Biochem. 87, 235-260 3. Pisoni, A. L., and Lindley, E. R. (1992) J. Biol. Chem. 267, 3626-3631 4B. Crooke, C. E. Castuma, and A. Kornberg, unpublished data. 5 ¢. E. Castuma, R. Reusch, and A. Kornberg, unpublished data. 6H. Ohtake, personal communication. 4 5. ll. 12. 13. 14. 15. 16. 17, 19. 20. 21. 22. 23, 24. 25. 26, 27. 28. 29. 30. 6295 . Hsieh, P-C., Shenoy, B. C., Jentoft, J. E., and Phillips, N. F. B. (1993) Protein Expression Purif. 4, 76-84 Phillips, N. F., Horn, P. J., and Wood, H. G. (1993) Arch. Biochem. Biophys. 300, 309-319 . Bonting, C. F. C., Kortstee, G. J. J., and Zehnder, A. J. B. (1991) J. Bacteriol. 173, 6484-6488, . van Veen, H. W., Abee, T., Kortstee, G. J. J., Konings, W. N., and Zehnder, A. J. B. (1993) J. Biol. Chem. 268, 19377-19383 . Archibald, F. S., and Fridovich, I. (1982) Arch. Biochem. Biophys. 215, 589-596 . Pick, U., and Weiss, M. (1991) Plant Physiol. 97, 1234-1240 . Reusch, R. N., and Sadoff, H. L. (1988) Proc. Natl. Acad. Sci. U.S. A. 85, 4176-4180 Kornberg, A., Kornberg, S. R., and Simms, E. S. (1956) Biochim. Biophys. Acta 20, 215-227 Kornberg, S. R. (1957) Biochim. Biophys. Acta 26, 294-300 Ahn, K., and Kornberg, A. (1990) J. Biol. Chem, 265, 11734-11739 Akiyama, M., Crooke, E., and Kornberg, A. (1992) J. Biol. Chem. 267, 22556— 22561 Akiyama, M., Crooke, E., and Kornberg, A. (1993) J. Biol. Chem. 268, 633-639 Miller, J. H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Kohara, Y., Akiyama, K., and Isono, K. (1987) Cell 50, 495-508 . Neidhardt, F.C., Block, P. L., and Smith, D. F. (1974) /. Bacteriol. 119, 736-747 Ames, B., N., and Dubin, D. T. (1960) J. Biol. Chem. 235, 769-776 Bachmann, B. J. (1990) Microbiol. Rev. 54, 130-197 Levinthal, C., Signer, E. R., and Fetherolf, K. (1962) Proc. Natl. Acad. Sci. U.S. A. 48, 1230-1237 Rao, N. N., Roberts, M. F., and Torriani, A. (1985) J. Bacteriol. 162, 242-247 Cashel, M., and Rudd, K. E. (1987) in Escherichia coli and Salmonella typhi- murium: Cellular and Molecular Biology, pp. 1410-1438, American Society of Microbiology, Washington, D. C. Keasling, J. D., Bertsch, L., and Kornberg, A. (1993) Proc. Natl. Acad, Sci. U.S. A. 90, 7029-7033 McCann, M. P,, Kidwell, J. P, and Matin, M. (1991) J. Bacteriol. 173, 4188— 4194 Siegele, D. A., and Kolter, R. (1992) J. Bacteriol. 174, 345-348 Hengge-Aronis, R. (1993) Cell 72, 165-168 Zambrano, M. M., Siegele, D. A., Almiron, M., Torma, A., and Kolter, R. (1993) Science 269, 1757-1760 Auling, G., Pilz, F., Busse, H. J., Karrasch, S., Streichan, M., and Schon, G. (1991) Appl. Environ. Microbiol. 57, 3583-3592 Ausubel, F. M. (ed) (1991) Current Protocols in Molecular Biology, Vol. 1, John Wiley & Sons, New York