journal of Cellular Biochemistry 51:47-54 (1993)
`
`Gene Regulation by Phosphate in Enteric Bacteria
`
`B.L. Wanner
`
`Department of Biological Sciences, Purdue University, West Lafayette, Indiana 47907
`
`The Escherichia coli phosphate (PHO) regulon includes 31 (or more) genes arranged in eight separate
`Abstract
`operons. All are coregulated by environmental (extra-cellular) phosphate and are probably involved in phosphorus
`assimilation. Pi control of these genes requires the sensor PhoR, the response regulator PhoB, the binding protein(cid:173)
`dependent Pi-specific transporter Pst, and the accessory protein PhoU. During Pi limitation, PhoR turns on genes of the
`PHO regulon by phosphorylating PhoB that in turn activates transcription by binding to promoters that share an 18-base
`consensus PHO Box. When Pi is in excess, PhoR, Pst, and PhoU together turn off the PHO regulon, presumably by
`dephosphorylating PhoB. In addition, two Pi-independent controls that may be forms of cross regulation turn on the
`PHO regulon in the absence of PhoR. The sensor CreC, formerly called PhoM, phosphorylates PhoB in response to some
`(unknown) catabolite, while acetyl phosphate may directly phosphorylate PhoB. Cross regulation of the PHO regulon by
`CreC and acetyl phosphate may be examples of underlying control mechanisms important for the general (global)
`control of cell growth and metabolism.
`Q 1993 Wdey-L,ss, Inc.
`
`Key words: acetyl phosphate, cross regulation, phosphate control, phosphorus metabolism, protein phosphorylation,
`two-component regulatory systems
`
`Escherichia coli uses three kinds of phospho(cid:173)
`rus compounds for growth: inorganic phosphate
`(Pi), organophosphates, and phosphonates (Pn).
`When Pi, the preferred phosphorus (P) source,
`is in excess, Pi is taken up by the low affinity Pi
`transporter, Pit. Under these conditions, the
`genes for the high affinity Pi-specific trans(cid:173)
`porter, Pst, and ones for use of alternative P
`sources are repressed. The latter genes are coreg(cid:173)
`ulated as members of the phosphate (PHO) reg(cid:173)
`ulon and are induced more than 100-fold during
`Pi limitation. Altogether 31 genes belonging to
`the E. coli PHO regulon have now been cloned
`and sequenced. Most of the corresponding gene
`products have also been characterized. They are
`transcribed as eight separate genes and operons.
`All probably have a role in the assimilation of
`different P sources from the environment (Ta(cid:173)
`ble 1).
`Cellular P metabolism is complex. Hence, it is
`likely that multiple controls may act on the
`PHO regulon. Basically, the assimilation of any
`P compound involves two early steps. First, Pi
`or an alternative P compound must be taken up.
`And, second, the Pi, or the P in the alternative
`compound must be incorporated into ATP, the
`primary phosphoryl donor in metabolism. Even(cid:173)
`tually, P is incorporated into essential compo-
`
`Received August 25, 1992; accepted August 25, 1992.
`
`co 1993 Wiley-liss, Inc.
`
`nents in membrane lipids, complex carbohy(cid:173)
`drates such as lipopolysaccharides, and nucleic
`acids. P also fo:r;ms high energy bonds and is
`incorporated into many proteins posttranslation(cid:173)
`ally.
`The PHO regulon is controlled by PhoR and
`PhoB that are similar to the respective partner
`proteins in the large family of two-component
`regulatory systems of sensors that act as histi(cid:173)
`dine protein kinases and of response regulators
`[Stock et al., 1989]. PhoR is the sensor and
`PhoB is the response regulator, a DNA binding
`protein that acts as a transcriptional activator.
`Transcription of PHO regulon genes requires
`the phosphorylated form ofPhoB [Makino et al.,
`1989]; and controls that regulate the PHO regu(cid:173)
`lon affect the amount of phosphorylated PhoB.
`Three controls act on the PHO regulon. One
`control is Pi-dependent and requires the sensor
`PhoR. Two others are Pi-independent and acti(cid:173)
`vate transcription in the absence of PhoR. One
`Pi-independent control requires the sensor CreC,
`formerly called PhoM, and the other requires
`acetyl phosphate [Wanner et al., 1988; Wanner
`and Wilmes-Riesenberg, 1992].
`Pi control of the PHO regulon is a form of
`transmembrane signal transduction. It is cou(cid:173)
`pled to the first step in P metabolism, Pi uptake.
`The Pi-independent controls are examples of
`cross regulation. They appear to be coupled to
`
`BEQ 1024
`Page 1
`
`

`
`48
`
`Wanner
`
`TABLE I. Phosphate-Starvation-Inducible
`(psi) Genes of the E. coli PHO Regulon *
`Name
`Map
`Function
`
`phnCDEFGHIJKLMNOP 93.3' Pn uptake and
`breakdown
`(psiD)
`phoA (psiA)-psiF
`8.7' Bap, unknown
`9.0' Regulator, sensor
`phoBR
`5.8' Polyanion porin
`phoE
`phoH(psiH)
`23.6' Unknown
`91.5' Unknown
`psiE
`84.0' Pi uptake, Pi
`pstSCAB-phoU
`repression
`75.8' G-3-P uptake,
`phospho-
`diesterase
`
`ugpBAECQ (psiB, psi C)
`
`*Only sequenced psi genes are listed.
`
`subsequent steps in P metabolism. Both may be
`connected to central metabolic pathways for the
`incorporation of intracellular Pi into ATP. The
`PHO regulon and its control are summarized
`below. See Wanner [1987a], Wanner [1990], and
`Wanner [1992] for earlier reviews.
`
`THE PHO REGULON AND P ASSIMILATION
`
`Depending upon the external Pi concentra(cid:173)
`tion, Pi is taken up by the Pit or Pst transporter
`[Rosenberg, 1987]. The Pit transporter is a sin(cid:173)
`gle component, proton motive force-driven trans(cid:173)
`porter similar to LacY [Kaback, 1990]. Pit has a
`10 1-1M~' is made constitutively, and is not part
`of the PHO regulon. Also, Pit has no role in Pi
`control since pit mutations are without effect on
`gene regulation. In contrast, the Pst transporter
`is a multicomponent system that is similar to
`ones for histidine, maltose, ribose, and other
`bacterial periplasmic transport systems [Land(cid:173)
`ick et al., 1989]. Pst has a 0.8 1-1M ~' is made
`during Pi limitation, and is a member of the
`PHO regulon. PstS is the peri plasmic Pi binding(cid:173)
`protein [Luecke and Quiocho, 1990], PstC and
`PstA are integral membrane proteins, and PstB
`is the permease. Like perm eases for similar bind(cid:173)
`ing protein-dependent transporters, PstB shares
`sequence similarities at the protein level to the
`products of the mammalian multidrug resis(cid:173)
`tance and cystic fibrosis genes [Hyde et al., 1990].
`The Pst system is encoded by an operon that
`also encodes a protein called PhoU, which has
`no role in transport. Even a complete deletion of
`the phoU gene is without effect on Pi uptake by
`the Pst system (PM Steed and BLW, unpub(cid:173)
`lished data). Both Pst and PhoU are required for
`
`repression, but not for activation, of the PHO
`regulon.
`Under conditions of Pi limitation, both trans(cid:173)
`portable and nontransportable organophos(cid:173)
`phates can serve as P sources. sn-glycerol-3-
`phosphate (G-3-P) is taken up by the binding
`protein-dependent Ugp transporter of the PHO
`regulon. U gpB is the peri plasmic G-3-P binding
`protein, U gpE and U gpA are integral membrane
`proteins, and UgpC is the permease. G-3-P trans(cid:173)
`ported by the U gp system may directly enter the
`biosynthetic pool for membrane phospholipid
`biosynthesis, without the release of free Pi in(cid:173)
`side the cell. Glycerophosphoryl diesters (de(cid:173)
`acylated phospholipids) are also transported by
`the U gp system, but they are immediately hydro(cid:173)
`lyzed by the U gpQ phosphodiesterase, which is
`closely associated with the Ugp transporter
`[Brzoska and Boos, 1988].
`Nontransportable organophosphates are hy(cid:173)
`drolyzed in the periplasm by the non-specific
`phosphomonoesterase bacterial alkaline phos(cid:173)
`phatase [Bap; Heid and Wilson, 1971; Coleman,
`1992], thephoA gene product of the PHO regu(cid:173)
`lon, and the Pi released is taken up by the Pit or
`Pst transporter. Complex P anions probably dif(cid:173)
`fuse into the periplasm through the PhoE (or
`another) porin, where they are broken down by
`degradative enzymes such as Bap [Rao and Tor(cid:173)
`riani, 1988]. Also, other periplasmic phosphata(cid:173)
`ses, e.g., an acid phosphatase, AppA, and a glu(cid:173)
`cose-1-phosphate phosphatase, Agp [Dassa et
`al., 1990], can hydrolyze these and other esters.
`In addition, some organophosphates, e.g., G-3-P
`and hexose phosphates, are taken up by other
`transport systems. However, these other phos(cid:173)
`phatases and transport systems probably have
`no role in P assimilation because they are not
`under PHO regulon control [Wanner, 1990].
`Pn are structural analogs of organophos(cid:173)
`phates that have a direct carbon-phosphorus
`(C-P) bond in place of the carbon-oxygen-phos(cid:173)
`phorus ester linkage. Many bacteria can use
`natural or synthetic Pn as a sole P source under
`conditions of Pi limitation. The use of Pn as a P
`source requires breakage of the C-P bond, for
`which two pathways exist, a phosphonatase path(cid:173)
`way and a C-P lyase pathway. These pathways
`are distinguishable by their strikingly different
`substrate specificities, which results from differ(cid:173)
`ences in the biochemical mechanism for break(cid:173)
`age of the C-P bond. The phosphonatase path(cid:173)
`way has a narrow substrate specificity, while the
`C-P lyase pathway has a very broad substrate
`
`BEQ 1024
`Page 2
`
`

`
`Gene Regulation by Phosphate in Bacteria
`
`49
`
`specificity. Some bacteria such as E. coli and
`Salmonella typhimurium have only one path(cid:173)
`way, while other bacteria such as Enterobacter
`aerogenes have both pathways. When cloned
`into E. coli, genes for both pathways are under
`PHO regulon control [Metcalf and Wanner, 1991;
`Lee et al., 1992; K-S Lee, WW Metcalf, and
`BLW, unpublished data].
`In E. coli, all genes for Pn uptake and degrada(cid:173)
`tion are in the fourteen gene phnC-to-phnP
`operon, which encodes a binding protein-depen(cid:173)
`dent Pn transporter and a C-P lyase [Wackett et
`al., 1987; Wanner and Boline, 1990; Chen et al.,
`1990; Metcalf et al., 1990; Wanner and Metcalf,
`1992; WWM and BLW, unpublished data]. The
`transport system is composed of three proteins:
`PhnC, PhnD, and PhnE. PhnD is the periplas(cid:173)
`mic binding protein, PhnE is an integral mem(cid:173)
`brane protein, and PhnC is the permease. The
`C-P lyase is an apparent membrane-associated
`enzyme complex composed of nine polypeptides:
`PhnG, PhnH, Phni, PhnJ, PhnK, PhnL, PhnM,
`PhnN, and PhnP. Two other gene products,
`PhnF and PhnO, have no (apparent) role in
`uptake or degradation. On the basis of sequence
`similarities, they may have roles in gene regula(cid:173)
`tion. Actually, Pn utilization is cryptic in the
`common laboratory strain E. coli K-12, though
`it is functional in most natural strains of E. coli,
`and in many other members ofthe family Entero(cid:173)
`bacteriaceae [Wanner and Boline, 1990]. The
`cryptic Pn phenotype for E. coli K-12 is due to a
`frameshift mutation in the phnE (EcoK) gene
`[Makino et al., 1991].
`The phoBR operon [Wanner and Chang, 198 7]
`is autogenously regulated [Guan et al., 1983].
`Thus, the synthesis of the regulatory proteins
`PhoB and PhoR is under PHO regulon control.
`On the basis of studies using genetic fusions, the
`amounts of PhoB and Ph oR increase more than
`100-fold during Pi limitation. PhoB synthesis is
`also subject to translational control by a process
`that may involve down regulation of the phoBR
`operon by an antisense RNA [Wanner, 1990;
`B-D Chang and BLW, unpublished data]. In(cid:173)
`creased PhoB synthesis during induction of the
`PHO regulon is probably responsible, at least in
`part, for an unusual phenomenon associated
`with certain regulatory mutants. NullphoR mu(cid:173)
`tations [Wanner, 1987b] as well as particular
`mutations of the pstSCAB-phoU operon [Wan(cid:173)
`ner, 1986] lead to a "clonal variation" pheno(cid:173)
`type in which induction leads to an "induced
`state." In such mutants, conditions that turn on
`
`the PHO regulon lead to the formation of cells in
`which induction is maintained in the absence of
`the inducing signal.
`Three additional sequenced genes (psiE, psiF,
`and phoH) whose functions are unknown are
`also coregulated as members of the PHO regu(cid:173)
`lon. These genes were originally identified in a
`set of 55 mutants made with the transposon Mu
`d1 (lacZ) that showed a phosphate-starvation(cid:173)
`inducible (psi) Lac+ phenotype. The psiE gene
`corresponds to an open reading frame between
`the malG and xylE genes near 91.5'. Two mu(cid:173)
`tants had insertions in the psiE gene. The psiF
`gene corresponds to an open reading frame im(cid:173)
`mediately downstream of the phoA gene, which
`showed that the phoA gene itself was in an
`operon. One mutant had an insertion in the psiF
`gene. The phoH gene corresponds to a new gene,
`formerly called psiH [Metcalf et al., 1990; PMS,
`WWM, and BLW, unpublished data]. Three mu(cid:173)
`tants had insertions in the phoH gene. Also,
`seven psiA mutants had insertions in the phoA
`gene, one psiB and one psiC mutant had inser(cid:173)
`tions in the ugpB gene, and three psiD mutants
`had insertions in the phnD gene. Early studies
`on these and other psi genes also showed that
`several were induced by other stresses, in addi(cid:173)
`tion to Pi limitation. For example, the ugpB
`(psiB and psiC) fusions were induced by both Pi
`and carbon limitations [Wanner and McSharry,
`1982]. In agreement, two laboratories recently
`showed that the ugpBAECQ operon has two
`promoters. One is induced by Pi limitation and
`the other, by carbon limitation [Kasahara et al.,
`1991; Suet al., 1991].
`
`Pi-REGUlATED PROMOTERS
`
`All PHO regulon promoters except one have
`an 18-base consensus "PHO Box." No PHO Box
`lies immediately upstream of the psiE gene; it
`may be a distal gene in an operon, however. The
`in vitro transcription of the phnC, phoA, phoB,
`phoH, pstS, and ugpB promoters requires phos(cid:173)
`phorylated PhoB, which on the basis ofDNAse I
`protection and methylation interference studies
`binds to the corresponding PHO Boxes (Fig. 1 ).
`In vitro studies on the phoE and psiE genes
`have not been done. Also, the in vivo mRNA
`start sites were determined for the phoA, phoB,
`and pstS promoters [Makino et al., 1986], for
`the phoE promoter [Tommassen et al., 1987],
`and for the phoH promoter (H. Shinagawa, per(cid:173)
`sonal communication). That the PHO Box is
`required for transcriptional activation in vivo
`
`BEQ 1024
`Page 3
`
`

`
`50
`
`Wanner
`
`PHO Box Consensus: CTGTCATA(ATlAITAlCTGT(CA)A(CT)
`
`phnC ( -41) CTGTTAGTCAGTTTAATTAACCAAATCGTCACAATAATCC~
`phoA ( -40) CTGTCATAAAGTIGTCACGGCCGAGACTIATAGTCGCTTT~
`phoB ( -40) TTTICATAAATGGTCATAAATCTGACGCATAATGACGTC~
`phoE ( -41) CTGTAATATATGTTAACAATCTCAGGTIAAAAACTTICGG~
`phoH ( -37) CTGTCATCACTGGTCATCTTICCAGTAGAAACTAAT~
`pstS ( -63) CTGTCATAAAACTGTCATATTC-
`~
`CTTACATATAAGGTCACCTGTTTGTCGATITIGCTICTCG
`ugpB ( -95) CC'Gi"C'ACCGCC-
`TTGTCATCTTICTGACACGTA(cid:173)
`CTATCTTACAAATGTAACAAAA(cid:173)
`AAGTTATTTITCTGTAATTCGAGCATGTCATGTTACCCC~
`Information Content, PHO Box
`7 -bp Direct Repeat
`A 2 1 1 2 5 20 0
`A 1 1 1 1 1 9 0 6 4 7 3 1 0 0 1 4 10 0
`c 15 1 0 0 14 0 6
`(700070022117 0006 05
`G007000100010 0800 00
`G 0 0 16 0 0 0 1
`T 2 9 2 9 2 1 9 2 4 2 5 2 10 2 9 0 0 5
`T 4 19 4 19 2 1 14
`
`Fig. 1. Sequences of PhoB-activatable promoters. The mRNA
`start sites were determined for the phnC promoter [Makino et
`al., 1991], the phoA, phoB, and pstS promoters [Makino et al.,
`1986], the phoE promoter [Tommassen et al., 1987], and the
`ugpB promoter [Kasahara et al., 1991]. Phosphorylated PhoB
`protects bases -55 to -12 of the phnC promoter [Makino et
`al., 1991], bases -44 to -11 of the phoB promoter [Makino et
`al., 1988], bases -65 to -16 of the pstS promoter [Makino et
`al., 1988], and bases -100 to -14 of the ugpB promoter
`
`[Kasahara et al., 1991]. The complete DNA sequence of the
`phoH gene, its mRNA start site and protection of bases -41 to
`-14 of the phoH promoter by phosphorylated PhoB were
`determined by Kim, Makino, Amemura, Shinagawa, and Nakata
`(H Shinagawa, personal communication). The PHO Box regions
`are marked with lines above them. The -10 regions are under(cid:173)
`lined. The information content is described by Schneider et al.
`[1986]. The -35 region in the PHO Box consensus is under(cid:173)
`lined.
`
`was shown by studying the effects of 5' deletions
`on the expression of the phnC promoter (D
`Agrawal and BLW, unpublished data), thephoE
`promoter [Tommassen et al., 1987], and the
`pstS promoter [Kimura et al., 1989].
`The PHO Box in the phnC, phoA, phoB, phoE,
`and phoH promoters is composed of two 7 -bp
`direct repeats separated by 4-bp that is part of
`the -35 region (Fig. 1). The pstS promoter has
`an additional PHO Box that was also protected
`by PhoB; the pstS upstream region contains
`three additional half sites that were not pro(cid:173)
`tected (not shown). The ugpB promoter has one
`additional PHO Box and one additional half site
`that were protected. Interestingly, the phoE pro(cid:173)
`moter has an additional PHO Box in the oppo(cid:173)
`site orientation more than 100 bases upstream
`of the mRNA start site (not shown). This addi(cid:173)
`tional PHO Box may act as a transcriptional
`enhancer [Tommassen et al., 1987].
`
`Pi CONTROL BY TRANSMEMBRANE SIGNAL
`TRANSDUCTION
`
`Pi control of the PHO regulon involves two
`processes: repression by the Pst system, PhoU,
`
`and PhoR when Pi is in excess and activation by
`PhoR when Pi is limiting. That Pi control in(cid:173)
`volves transmembrane signal transduction is
`evident by the requirement for the Pst trans(cid:173)
`porter. The vast majority ofpst mutations simul(cid:173)
`taneously abolish both transport and repres(cid:173)
`sion. Missense changes in PstS and PstB,
`including ones in the nucleotide binding domain
`of the PstB permease [Cox et al., 1989], abolish
`both processes .. Yet, repression is independent of
`transport per se because a missense change in
`PstA abolishes transport without affecting re(cid:173)
`pression [Cox et al., 1988]. Further, Pi repres(cid:173)
`sion involves detection of environmental Pi. An
`extracellular Pi concentration of 4 f.LM represses
`PHO regulon gene expression. This corresponds
`to an amount that would fully saturate the Pi
`binding-protein PstS (0.8 f.LM K.J). Derepression
`by Pi limitation is also not accompanied by a
`lowering of the internal Pi concentration. Fur(cid:173)
`ther, PhoU is required for repression, although
`PhoU has no role in transport. Since PhoU is a
`product of the pstSCAB-phoU operon and is
`loosely associated with the inner membrane, it
`is reasonable to suppose that PhoU interacts
`with the Pst system. An interaction of PhoU
`
`BEQ 1024
`Page 4
`
`

`
`Gene Regulation by Phosphate in Bacteria
`
`51
`
`Unshaded:
`Pit
`Pi transporter
`Pst Pi-specific transporter
`S
`Pi binding protein
`A. C Integral membrane proteins
`Pi permease
`B
`U Negative regulator PhoU
`Shaded:
`B
`Response regulator PhoB
`R
`Pi sensor PhoR
`
`Fig. 2. Pi control of PHO regulon by transmembrane signal transduction. See text.
`
`with both Pst and PhoR may be important in Pi
`control [Wanner, 1990].
`PhoR has a dual role in Pi control. Therefore
`PhoR probably exists in two forms, a repressor,
`PhoRR, and an activator, PhoRA, form [Wanner,
`1987a]. These two forms may correspond to
`monomer and dimer forms, though no evidence
`indicates which is the repressor or activator
`form. In any case, PhoRA probably activates
`PhoB by phosphorylation during Pi limitation
`and PhoRR probably inactivates PhoB by dephos(cid:173)
`phorylation when Pi is in excess. Accordingly, Pi
`control would involve the interconversion of
`. PhoR between its PhoRR and PhoRA forms by a
`mechanism involving environmental Pi, the Pst
`system, and Pho U.
`A model for Pi transmembrane signal trans(cid:173)
`duction is shown in Figure 2, in which PhoRR
`and PhoRA are arbitrarily indicated as monomer
`and dimer forms, respectively. This model de-
`
`picts three events important in Pi repression:
`saturation of PstS by environmental Pi, binding
`of Pi bound PstS with PstABC complexes in the
`membrane, and the formation of a "repressor
`complex" containing Pi bound PstS, PstABC,
`PhoU, and PhoR. Such a complex would main(cid:173)
`tain PhoR in its repressor form that would de(cid:173)
`phosphorylate PhoB. Accordingly, Pi limitation,
`or mutations in the Pst system or PhoU, would
`cause PhoR to be released from the complex.
`This would lead to its conversion to PhoRA that
`would autophosphorylate and would phosphory(cid:173)
`late PhoB .
`That Pi control probably involves protein(cid:173)
`protein interactions with PhoR is supported by
`its (apparent) domain structures. PhoR is 431
`amino acids in length and probably consists of
`three domains (Fig. 3). It has near its N··
`terminus a highly hydrophobic region of about
`50 amino acids. This segment may form a do··
`
`PhoR 431 a a.
`
`H
`
`213
`
`H
`Cree 474 a a. -t••------11••---+1--------
`266
`
`Fig. 3. Simple diagrams for the sensors PhoR and CreC. Filled boxes are hydrophobic segments. The
`sensor Cree is like many sensors in that it has two transmembrane domains and a short stretch of about 50
`amino acids after its second transmembrane domain and before its kinase domain. The histidine residues
`that are probably phosphorylated are shown. See text.
`
`BEQ 1024
`Page 5
`
`

`
`52
`
`Wanner
`
`Oxidative phosphorylation
`
`ADP + Pi AtpiBEFHAGDG t> ATP
`PMF
`
`Glycolysis
`Ga
`Glyceraldehyde-3-Pi + Pi + NAD ~1 ,3-diphosphoglycerate + NADH
`1 ,3-diphosphoglycerate + ADP ~ 3-phosphoglycerate + ATP
`
`Tricarboxylic acid cycle
`SucG,D
`Succinyi-GoA + ADP + Pi ----1> Succinate + A TP + GoA
`
`Mixed -acid fermentation
`Acetyl-GoA + Pi ~ Acetyl phosphate + GoA
`AckA
`Acetylphosphate + ADP ----1> Acetate + ATP
`
`Fig. 4. Pathways for incorporation of Pi into ATP. Gene symbols are used as enzyme names. AckA,
`acetate kinase; AtplBEFHAGDC, ATP synthase; Gap, glyceraldehyde 3-phosphate dehydrogenase; Pgk,
`phosphoglycerate kinase; Pta, phosphotransacetylase; SucC, D, succinyi-CoA synthetase.
`
`main with two a-helices spanning the mem(cid:173)
`brane and a short six or seven amino acid region
`exposed to the periplasm, or, it may form a
`domain that is entirely in the membrane. Regard(cid:173)
`less, the bulk of Ph oR is in the cytoplasm and is
`likely to consist oftwo additional domains. Many
`sensors have N-terminal domains followed by
`regions of about 50 amino acids preceding their
`kinase domains, e.g., see the structure for CreC
`in Figure 3. In contrast, PhoR has an unusually
`large region of about 150 amino acids immedi(cid:173)
`ately following its membrane domain. Further,
`mutational studies imply that this "linker
`domain" in PhoR is both necessary for repres(cid:173)
`sion and important for protein-protein interac(cid:173)
`tion(s). The changes L146P, L147P, and R148C
`in this region eliminate repression and are par(cid:173)
`tially dominant to the wild-type protein. The
`dominance of such mutant proteins implies that
`this region has a role in protein-protein interac(cid:173)
`tions, which may involve interactions with Pho U
`or PhoR dimerization. The carboxyl terminal
`domain of Ph oR is its kinase domain, in which a
`conserved histidine, H213, is probably the site of
`autophosphorylation. The changes T217A,
`P218L, P218S, T220N, and Y225C in this re(cid:173)
`gion also abolish repression (but not activation)
`and are recessive to wild-type. They may affect
`
`the PhoR kinase or phosphatase function (B-DC
`and BLW, unpublished data).
`
`Pi-INDEPENDENT CONTROLS AND CROSS
`REGULATION
`
`Pi control is abolished and two Pi-indepen(cid:173)
`dent controls are apparent in phoR mutants.
`These Pi-independent controls are the basis for
`cross regulation. That they are likely important
`in wild-type c:ells is evident from physiology and
`mutational studies. Both are regulated by the
`carbon and energy source, but in different ways.
`One is induced by glucose and involves the sen(cid:173)
`sor CreC; the other is induced by pyruvate and
`involves acetyl phosphate synthesis. Both ap(cid:173)
`pear to be coupled to steps in central metabolism
`that are also steps in intracellular Pi metabo(cid:173)
`lism. CreC may phosphorylate PhoB in response
`to a signal for the incorporation of Pi into ATP
`by oxidative phosphorylation or substrate-level
`phosphorylation, in glycolysis or the TCA cycle.
`Acetyl phosphate may directly phosphorylate
`PhoB when large amounts are made by sub(cid:173)
`strate-level phosphorylation in mixed-acid fer(cid:173)
`mentation. In this regard, acetyl phosphate is an
`intermediate in a pathway for the entry of Pi
`into ATP. Therefore acetyl phosphate synthesis
`
`BEQ 1024
`Page 6
`
`

`
`Gene Regulation by Phosphate in Bacteria
`
`53
`
`is formally a pathway for both carbon and Pi
`metabolism [Fig. 4; Wanner, 1992].
`Mutational studies showed that these Pi(cid:173)
`independent controls are separate and that they
`are coupled to central metabolism. CreC is a
`product of the creABCD operon, in which muta(cid:173)
`tions are without effect on control by acetyl
`phosphate. Conversely, mutations in the pta or
`ackA gene, for acetyl phosphate metabolism, are
`without effect on CreC-dependent control [Wan(cid:173)
`ner et al., 1988; Wanner and Wilmes-Riesen(cid:173)
`berg, 1992]. Also, mutations in other genes that
`are likely to affect central metabolism alter Pi(cid:173)
`independent control of the PHO regulon involv(cid:173)
`ing CreC or acetyl phosphate synthesis. These
`other genes include ones for aerobic respiratory
`·control (areA), adenylate cyclase (cya), cAMP
`receptor protein (crp ), isocitrate dehydrogenase
`(icd), malate dehydrogenase (mdh), outer mem(cid:173)
`brane regulator (ompR ), exopolysaccharide pro(cid:173)
`duction (ops), phosphotransferase enzyme I, and
`the phosphohistidinoprotein HPr · (ptsHI), and
`others. Further, there may be a regulatory cou(cid:173)
`pling between genes for acetyl phosphate metab(cid:173)
`olism and Pi uptake, at least under some growth
`conditions. Even though ackA and pta muta(cid:173)
`tions are without effect on the use of Pi as a P
`source, a mutation that abolishes ackA gene
`expression simultaneously abolishes the use of
`Pi asaP source [Wanner et al., 1988; Wanner,
`1992; MR Wilmes-Riesenberg and BLW, unpub(cid:173)
`lished data]. Accordingly, there may exist a com(cid:173)
`mon regulator for genes for acetyl phosphate
`metabolism and Pi utilization.
`
`PROSPECTS
`
`Many PHO regulon genes have been analyzed
`in great detail. Yet, the roles of some (phoH,
`psiE, and psiF and perhaps others) are not un(cid:173)
`derstood. The biochemical mechanism fortran(cid:173)
`scriptional activation of these genes is also fairly
`well established. In response to Pi limitation,
`the sensor PhoR autophosphorylates and phos(cid:173)
`phorylates PhoB that in turn activates transcrip(cid:173)
`tion by binding to the respective promoter sites.
`How PhoR detects environmental Pi levels is
`particularly unclear, although both the Pst
`transporter and Pho U are known to be required.
`Somehow the Pst system and PhoU must com(cid:173)
`municate a signal for Pi repression to PhoR.
`This process probably involves protein-protein
`interactions. Future studies on signal transduc(cid:173)
`tion in the PHO regulon should include determi(cid:173)
`nation of which proteins (PstA, PstB, PstC,
`
`PstS, PhoU, and PhoR) interact and how they
`interact in regards to Pi control.
`In addition, the PHO regulon is controlled by
`two forms of cross regulation that are apparent
`in phoR mutants. Both are connected to carbon
`and energy metabolism in ways that imply a role
`for cross regulation in the overall control of the
`PHO regulon. CreC autophosphorylates and
`phosphorylates PhoB in response to some (un(cid:173)
`known) catabolite, while high levels of acetyl
`phosphate may directly phosphorylate PhoB.
`These controls by CreC and acetyl phosphate
`synthesis may be coupled to different pathways
`for intracellular Pi metabolism. Regulatory inter(cid:173)
`actions of this sort may be especially important
`in the coordinate control of cell metabolism and
`growth in regard to Pi availability. Cross regula(cid:173)
`tion may also be important as a general (global)
`control in other gene systems regulated by pro(cid:173)
`tein phosphorylation, as it seems to be in the
`PHO regulon. Future studies on cross regula(cid:173)
`tion of the PHO regulon should include determi(cid:173)
`nation of the signal for the sensor CreC as well
`as the role for CreC and acetyl phosphate in
`wild-type cells.
`
`ACKNOWLEDGMENTS
`
`This laboratory is supported by Public Health
`Service grant GM35392 from the National Insti(cid:173)
`tutes of Health. I especially thank H. Shinagawa
`for personal communications and W.W. Metcalf
`for helpful comments on the manuscript.
`
`REFERENCES
`
`Brzoska P, Boos W (1988): Characteristics of a ugp-encoded
`and phoB-dependent glycerophosphoryl diester phospho(cid:173)
`diesterase which is physically dependent on the Ugp trans(cid:173)
`port system of Escherichia coli. J Bacterial 170:4125-
`4135.
`Chen C-M, Ye Q, Zhu Z, Wanner BL, Walsh CT (1990):
`Molecular biology of carbon-phosphorus bond cleavage:
`Cloning and sequencing of the phn (psiD) genes involved
`in alkylphosphonate uptake and C-P lyase activity in
`Escherichia coli B. J Bioi Chern 265:4461-447L
`Coleman JE (1992): Structure and mechanism of alkaline
`phosphatase. Annu Rev Biophys Biophys Chern 21:441-
`483.
`Cox GB, Webb D, Godovac-Zimmermann J, Rosenberg H
`(1988): Arg-220 of the PstA protein is required for phos(cid:173)
`phate transport through the phosphate-specific transport
`system in Escherichia coli but not for alkaline phospha(cid:173)
`tase repression. J Bacteriol170:2283-2286.
`Cox GB, Webb D, Rosenberg H (1989): Specific amino acid
`residues in both the PstB and PstC proteins are required
`for phosphate transport by the Escherichia coli Pst sys(cid:173)
`tem. J Bacteriol171:1531-1534.
`Dassa J, Marek C, Boquet PL (1990): The complete nucleo(cid:173)
`tide sequence of the Escherichia coli gene appA reveals
`
`BEQ 1024
`Page 7
`
`

`
`54
`
`Wanner
`
`significant homology between pH 2.5 acid phosphatase
`and glucose-1-phosphatase. J Bacterio\172:5497-5500.
`Guan C-D, Wanner B, Inouye H (1983): Analysis of regula(cid:173)
`tion of phoB expression using a phoB-cat fusion. J Bacte(cid:173)
`riol156:710-717.
`Hyde SC, Emsley P, Hartshorn MJ, Mimmack MM, Gileadi
`U, Pearce SR, Gallagher MP, Gill DR, Hubbard RE, Hig(cid:173)
`gins CF (1990): Structural model of ATP-binding proteins
`associated with cystic fibrosis, multidrug resistance and
`bacterial transport. Nature 346:362-365.
`Kaback HR (1990): The lac permease of Escherichia coli: A
`prototypic energy-transducing membrane protein. Bio(cid:173)
`chim Biophys Acta Bio-Energetics 1018: 160-162.
`Kasahara M, Makino K, Amemura M, Nakata A, Shinagawa
`H (1991): Dual regulation of the ugp operon by phosphate
`and carbon starvation at two interspaced promoters. J
`Bacterio\173:549-558.
`Kimura S, Makino K, Shinagawa H, Amemura M, Nakata A
`(1989): Regulation of the phosphate regulon of Esche(cid:173)
`richia coli: Characterization of the promoter of the pstS
`gene. Mol Gen Genet 215:374-380.
`Landick R, Oxender DL, Ames GF-L (1989): Bacterial amino
`acid transport systems. In Martonosi AN (ed): "The En(cid:173)
`zymes of Biological Membranes Second Edition," Volume
`3. New York: Plenum Press, pp 577-615.
`LeeK-S, MetcalfWW, Wanner BL (1992): Evidence for two
`phosphonate degradative pathways in Enterobacter aero(cid:173)
`genes. J Bacteriol174:2501-2510.
`Luecke H, Quiocho FA (1990): High specificity of a phos(cid:173)
`phate transport protein determined by hydrogen bonds.
`Nature 347:402-406.
`Makino K, Shinagawa H, Amemura M, Nakata A (1986):
`Nucleotide sequence of the phoB gene, the positive regula(cid:173)
`tory gene for the phosphate regulon of Escherichia coli
`K-12. J Mol Biol190:37-44.
`Makino K, Shinagawa H, Amemura M, Kimura S, Nakata A,
`Ishihama A (1988): Regulation of the phosphate regulon
`of Escherichia coli: Activation of pstS transcription by
`PhoB protein in vitro. J Mol Bioi 203:85-95.
`Makino K, Shinagawa H, Amemura M, Kawamoto T, Yamada
`M, Nakata A (1989): Signal transduction in the phosphate
`regulon of Escherichia coli involves phosphotransfer be(cid:173)
`tween PhoR and PhoB proteins. J Mol Bio\210:551-559.
`Makino K, Kim S-K, Shinagawa H, Amemura M, Nakata A
`(1991): Molecular analysis of the cryptic and functional
`phn operons for phosphonate use in Escherichia coli K-12.
`J Bacteriol173:2665-2672.
`MetcalfWW, Steed PM, Wanner BL (1990): Identification of
`phosphate-starvation-inducible genes in Escherichia coli
`K-12 by DNA sequence analysis of psi::lacZ(Mu d1) tran(cid:173)
`scriptional fusions. J