Use of New Methods for Construction of Tightly Regulated Arabinose and Rhamnose Promoter Fusions in Studies of the Escherichia coli Phosphate Regulon

Abstract

Escherichia coli genes regulated by environmental inorganic phosphate (P_i) levels form the phosphate (Pho) regulon. This regulation requires seven proteins, whose synthesis is under autogenous control, including response regulator PhoB, its partner, histidine sensor kinase PhoR, all four components of the P_i-specific transport (Pst) system (PstA, PstB, PstC, and PstS), and a protein of unknown function called PhoU. Here we examined the effects of uncoupling PhoB synthesis and PhoR synthesis from their normal controls by placing each under the tight control of the arabinose-regulated P_araB promoter or the rhamnose-regulated P_rhaB promoter. To do this, we made allele replacement plasmids that may be generally useful for construction of P_araB or P_rhaB fusions and for recombination of them onto the E. coli chromosome at the araCBAD or rhaRSBAD locus, respectively. Using strains carrying such single-copy fusions, we showed that a P_rhaB fusion is more tightly regulated than a P_araB fusion in that a P_rhaB-phoR⁺ fusion but not a P_araB-phoR⁺ fusion shows a null phenotype in the absence of its specific inducer. Yet in the absence of induction, both P_araB-phoB⁺ and P_rhaB-phoB⁺ fusions exhibit a null phenotype. These data indicate that less PhoR than PhoB is required for transcriptional activation of the Pho regulon, which is consistent with their respective modes of action. We also used these fusions to study PhoU. Previously, we had constructed strains with precise ΔphoU mutations. However, we unexpectedly found that such ΔphoU mutants have a severe growth defect (P. M. Steed and B. L. Wanner, J. Bacteriol. 175:6797–6809, 1993). They also readily give rise to compensatory mutants with lesions in phoB, phoR, or a pst gene, making their study particularly difficult. Here we found that, by using P_araB-phoB⁺, P_rhaB-phoB⁺, or P_rhaB-phoR⁺ fusions, we were able to overcome the extremely deleterious growth defect of a Pst⁺ ΔphoU mutant. The growth defect is apparently a consequence of high-level Pst synthesis resulting from autogenous control of PhoB and PhoR synthesis in the absence of PhoU.

The control of the Escherichia coli phosphate (Pho) regulon by environmental inorganic phosphate (P_i) levels is a paradigm of a bacterial signal transduction pathway in which occupancy of a cell surface receptor (the P_i-specific binding protein PstS) regulates gene expression in the cytoplasm (reference 29 and references therein). This signal transduction pathway requires seven proteins, all of which probably interact in a membrane-associated signaling complex. The P_i signaling proteins include (i) two members of the large family of two-component regulatory systems, namely, response regulator PhoB (a transcriptional activator) and its partner, histidine sensor kinase PhoR (itself an integral-membrane protein); (ii) four components of the ATP-binding cassette family P_i-specific transport (Pst) machinery (PstA, PstB, PstC, and PstS); and (iii) a negative regulator of unknown function called PhoU.

We proposed elsewhere that the P_i signaling response involves three processes: activation, deactivation, and inhibition (30). Accordingly, activation occurs under conditions of P_i limitation and requires both PhoB and PhoR; activation involves autophosphorylation of PhoR by ATP, phosphotransfer to PhoB, and transcriptional activation of Pho regulon promoters by phospho-PhoB (P-PhoB). Deactivation is a distinct intermediate step-down process that occurs upon a growth shift from P_i-limiting to P_i excess conditions. It is required to reestablish inhibition and leads to the dephosphorylation of P-PhoB in a process requiring PhoR and an excess of either PhoU, a Pst component(s), or both. Inhibition prevents phosphorylation of PhoB when P_i is in excess; it requires all seven P_i signaling proteins (PhoB, PhoR, PhoU, PstA, PstB, PstC, and PstS) in an “inhibition complex” that, by insulating PhoB, interferes with its phosphorylation.

In order to study how PhoU and the Pst system interact with PhoB and PhoR, we had previously constructed mutants with defined deletions of the pstSCAB-phoU operon. This led to our discovery that a ΔphoU (unlike a phoU missense) mutation causes a severe growth defect, resulting from an apparent P_i sensitivity phenotype (26). Growth inhibition by P_i has also been seen in mutants of the Pst transporter in which the P_i transport channel is permanently “switched on” (open), although in that case the growth defect is much less severe (34). The deleterious growth phenotype resulting from a ΔphoU mutation also leads to the accumulation of compensatory mutants with lesions in phoB, phoR, or a pst gene under normal growth conditions (26). Indeed, the poor growth of a ΔphoU mutant was apparently responsible for another laboratory concluding incorrectly that a ΔphoU mutation abolishes P_i uptake (19), as their “ΔphoU mutant” carries also a linked pst mutation (29). On the contrary, we proved that a ΔphoU mutation has no effect on P_i uptake (26).

We have now found ways to overcome the difficulties of studying phoU and (presumably) open-channel pst mutations as well. The growth defect due to a ΔphoU mutation is apparent only when the Pst system is synthesized at a high level, which in turn requires increased amounts of PhoB and PhoR, whose synthesis is autogenously controlled (10). We uncoupled phoB expression and phoR expression from their normal controls by placing them behind the foreign, arabinose-regulated P_araB or rhamnose-regulated P_rhaB promoter. Upon induction of the corresponding ΔphoB or ΔphoR mutant with arabinose or rhamnose, such strains show nearly normal P_i control of the Pho regulon. However, the fold induction is lowered. Importantly, a ΔphoU mutation has no deleterious effect on growth of these strains, even in the presence of the respective inducer. Our methods for constructing and characterizing P_araB and P_rhaB fusions may also be generally useful.

MATERIALS AND METHODS

Media and culture conditions.

Luria-Bertani broth, tryptone-yeast extract, and M63 were routinely used as complex and minimal media. These were prepared as described elsewhere (28). To maintain plasmids, antibiotics (Sigma, St. Louis, Mo.) were added as follows: ampicillin at 100 μg/ml, gentamicin at 15 μg/ml, kanamycin at 50 μg/ml, and tetracycline at 12.5 μg/ml. Recombinants with a single-copy plasmid were selected with gentamicin at 4 μg/ml or kanamycin at 10 μg/ml. 5-Bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal; Bachem, Torrance, Calif.) was used at 40 μg/ml to detect β-galactosidase. Tetracycline-sensitive (Tet^s) cells were selected on tetracycline-sensitive-selective (TSS) agar as described elsewhere (17). MacConkey agar (Difco, Detroit, Mich.) containing 1% l-arabinose, lactose, or l-rhamnose (Sigma) was used to test for the use of these carbon sources. Ara⁻ and Rha⁻ strains were verified by their inability to grow on M63 media containing 0.2% l-arabinose or l-rhamnose, respectively. Ara⁻ and Rha⁻ recombinants are distinguishable from arabinose- or rhamnose-sensitive ones by the inability of the latter to grow on MacConkey or glycerol agar in the presence of arabinose or rhamnose, respectively. In contrast, Ara⁻ and Rha⁻ recombinants are insensitive to these carbohydrates.

MOPS (morpholinepropanesulfonic acid) media containing different carbon sources were used to study gene regulation in P_araB, P_rhaB, and P_rhaS fusion strains. Cells were grown on MOPS agar containing the same carbon source at 0.2% for d-glucose, d-fructose, and d-mannitol or 0.4% for glycerol, but without an inducer. Fresh isolated colonies (after less than 20 h of growth) were used to inoculate 0.06% glucose–, fructose–, or mannitol–MOPS medium or 0.1% glycerol–MOPS medium without or with an inducer. Cultures were incubated at 37°C for 16 to 24 h prior to the assay. Such carbon-limited cultures yield highly reproducible values that are qualitatively similar to ones obtained for logarithmic-growth phase cultures (27). l-Arabinose, l-rhamnose, and isopropyl-β-d-galactopyranoside (IPTG; Sigma) were used at 1.3, 1.1, and 0.2 mM, respectively, for induction.

Bacteria.

All strains assayed are described in Table 1. Others included BT333 (endA::tetAR; from W. Wackernagel [4]), BW5045 (srlC300::Tn10; [18]), BW8078 (λrecA⁺ recA1; [28]), BW21391 (leu-63::Tn10; [12]), BW22773 (lacI^q rrnB_T14 ΔlacZ_WJ16 proC::Tn5-132; [14]), CA10 (galU95; from M. Berlyn), JW383 (metF159 zii-510::Tn10 thi-1; from M. Berlyn), W3110trpB114 (trpB114::Tn10; from C. Yanofsky), and ZK1001 (cysC95::Tn10 rpoS::kan; from R. Kolter). Only relevant markers are given in parentheses.

TABLE 1.

Bacterial strains

Straina	Genotype	Pedigreeb	Source or reference(s)
BW17142	DE3(lac)X74 ΔphoU559::kan rpoS(Am)	BD792 via BW13711	26
BW17335	DE3(lac)X74 Δ(pstSCAB-phoU)560::kan rpoS(Am)	BD792 via BW13711	26
BW18897	DE3(lac)X74 ΔphoU559 rpoS(Am) phn(EcoB)	BD792 via BW18834	Ilv+ with P1 on BW18833 (26)
BW21578	rrnBT14 ΔlacZWJ16	BD792 via BW21342	15
BW22831	lacIqrrnBT14ParaB-lacZAH31 ΔaraBADAH33rpoS(Am)	BD792 via BW22766	Leu+ Ara− with P1 on BW22826c (12)
BW22861	lacIqrrnBT14PphnC-lacZWJ19 ΔphoR574 ΔcreBCD153 Δ(pta ackA hisQ hisP)TA3516phn(EcoB) ΔaraBADAH33::ParaB-phoR[M1-D431]AH35rpoS(Am)	BD792 via BW22829	Leu+ Ara− with P1 on BW22835
BW22886	rrnBT14PrhaB-lacZLD68 ΔrhaBADLD78	BD792 via BW22779	Met+ Rha− with P1 on BW22875d
BW22887	rrnBT14PrhaB-lacZLD68 ΔrhaBADLD78rpoS(Am)	BD792 via BW22780	Met+ Rha− with P1 on BW22875
BW22888	rrnBT14PrhaS-lacZLD69 ΔrhaBADLD78	BD792 via BW22781	Met+ Rha− with P1 on BW22875
BW23473	Δlac-169 rpoS(Am) robA1 creC510 hsdR514 uidA(ΔMluI)::pir+endABT333recA1	BD792 via BW23438	12, 13
BW24249	lacIqrrnBT14 ΔlacZWJ16 ΔphoBR580 ΔcreABCD154 hsdR514 Δ(pta ackA hisQ hisP)TA3516phn(EcoB) ΔaraBADAH33 ΔrhaBADLD78uidA(ΔMluI)::pir+rposS(Am) endABT333galU95 recA1	BD792 via BW24217	Srl+recA with P1 on BW8078 (28)
BW24486	lacIqrrnBT14PphnC-lacZWJ19 ΔphoBR580 ΔcreBCD153 Δ(pta ackA hisQ hisP)TA3516phn(EcoB) ΔaraBADAH33::ParaB-phoR[M1-D431]AH35rpoS(Am) Δ(pstSCAB-phoU)560::Ω	BD792 via BW23424	Spcr and Strr with P1 on BW17636 (26)
BW24508	Like BW24486 except attλ::pSK58(PphoB-phoB+)	BD792 via BW24486	pSK58 integrant
BW24962	lacIqrrnBT14 ΔlacZWJ16 ΔphoB578 ΔcreABDC154 rpoS(Am) ΔaraBADAH33 ΔrhaBADLD78 Δ(ackA pta)160AH25	BD792 via BW24651	Pro+ with P1 on BW23562 (13)
BW24739	lacIqrrnBT14 ΔlacZWJ16 ΔcreABDC154 rpoS(Am) ΔaraBADAH33 ΔrhaBADLD78 Δ(ackA pta)160AH25 Δ(pstSCAB-phoR)560::Ω	BD792 via BW24713	Ilv+ with P1 on BW17646 (26)
BW24740	lacIqrrnBT14 ΔlacZWJ16 ΔphoB574 ΔcreABDC154 rpoS(Am) ΔaraBADAH33 ΔrhaBADLD78 Δ(ackA pta)160AH25 Δ(pstSCAB-phoU)560::Ω	BD792 via BW24715	Ilv+ with P1 on BW17646 (26)
BW24741	lacIqrrnBT14 ΔlacZWJ16 ΔphoB578 ΔcreABDC154 rpoS(Am) ΔaraBADAH33 ΔrhaBADLD78 Δ(ackA pta)160AH25 Δ(pstSCAB-phoU)560::Ω	BD792 via BW24717	Ilv+ with P1 on BW17646 (26)
BW24768	Like BW24740 except attHK022::pAH151(PrhaB-phoR+)	BD792 via BW24740	pAH151 integrant
BW24770	Like BW24741 except attλ::pAH85(PphoB-phoB+)	BD792 via BW24741	pAH85 integrant
BW24773	Like BW24741 except attλ::pAH136(PrhaB-phoB+)	BD792 via BW24741	pAH136 integrant
BW24774	Like BW24741 except attλ::pAH150(ParaB-phoB+)	BD792 via BW24741	pAH150 integrant
BW24775	Like BW24741 except attλ::pSK49(Ptac-phoB+)	BD792 via BW24741	pSK49 integrant
BW24777	lacIqrrnBT14 ΔlacZWJ16 ΔphoB578 ΔcreABCD154 rpoS(Am) ΔaraBADAH33 Δ(ackA pta)160AH25 Δ(pstSCAB-phoU)560::Ω ΔrhaBADLD78::PrhaB-phoBLD79	BD792 via BW24754	Met+ Rha− with P1 on BW22876
BW24803	Like BW24692 except attλ::pAH150(ParaB-phoB+)	BD792 via BW24692	pAH150 integrant
BW24858	Like BW24740 except attHK022::pAH152(PrhaS-phoR+)	BD792 via BW24740	pAH152 integrant
BW24936	lacIqrrnBT14 ΔlacZWJ16 ΔphoR574 ΔcreABCD154 rpoS(Am) ΔaraBADAH33 ΔrhaBADLD78 Δ(ackA pta)160AH25 ΔphoU559	BD792 via BW24714	Ilv+ Spcs and Strs with P1 on BW24627
BW24937	lacIqrrnBT14 ΔlacZWJ16 ΔphoB578 ΔcreABCD154 rpoS(Am) ΔaraBADAH33 ΔrhaBADLD78 Δ(ackA pta)160AH25 ΔphoU559	BD792 via BW24716	Ilv+ Spcs and Strs with P1 on BW24627
BW24948	Like BW24936 except attHK022::pAH151(PrhaB-phoR+)	BD792 via BW24936	pAH151 integrant
BW24949	Like BW24937 except attλ::pAH136(PrhaB-phoB+)	BD792 via BW24937	pAH136 integrant
BW24950	Like BW24937 except attλ::pAH150(ParaB-phoB+)	BD792 via BW24937	pAH150 integrant

^aAll are derivatives of E. coli K-12, except that the phn(EcoB) locus is from E. coli B (31). 

^bPedigree gives the parental strain from another laboratory and its most immediate ancestor in this laboratory. Many of our strains are descendents of E. coli K-12 strain BD792 (CGSC6159), which, like MG1655 (CGSC6300), is an unmutagenized and direct descendent of E. coli W1485 (CGSC5024) (1, 2). 

^cThe ara and leu-63::Tn10 markers are about 55% linked in P1kc crosses (data not shown). 

^dThe relevant gene order is zii-510::Tn10 metF159 ΔrhaBAD_LD78 in which Tn10 and metF are ca. 70% linked and metF and rha are ca. 10% linked (data not shown).  

Plasmids and phage.

pAH85, pSK49, and pSK58 were constructed in an earlier study; each has the same backbone as pSK50ΔuidA2 (13). pAH85 and pSK58 are similar, except that the former has a silent SpeI site at codon 113 of phoB. Each expresses phoB from its native promoter. pSK49 expresses phoB from P_tac. pAH136 and pAH150 are derivatives of pAH85 and pSK49 carrying P_rhaB and P_araB in place of P_phoB and the lacI^q-P_tac region, respectively; pAH151 and pAH152 are similar plasmids carrying P_rhaB-phoR⁺ and P_rhaS-phoR⁺ except that they have the attachment site of HK022 and encode gentamicin resistance (14). pBAD32 and pBAD33 (11) were from L.-M. Guzman. They are similar except that pBAD32 has an undefined deletion of ca. 0.6 kbp between the polylinker and the cat gene of pBAD33 (data not shown). pBC35 is a derivative of pBGS19+ (25) carrying the phoBR operon within a 4.7-kbp PstI-to-EcoRI fragment (32) (Fig. 1). pSK47 (16) has the same phoBR fragment cloned into the backbone of pSK50ΔuidA2 (13). pSLF13 and pSLF22 were constructed in an earlier study (8). pSLF13 is a derivative of pSPORT1 (Gibco-BRL, Bethesda, Md.) that encodes a segment of VanS within a 0.4-kbp EcoRI-to-BamHI insert that has a start codon overlapping an NdeI site immediately downstream of a synthetic ribosome-binding site. pSLF22 is a derivative of pBAD32 containing a similar insert. pWJ17 and pWJ18 are derivatives of the lacZ transcriptional fusion vector pWJ13 containing a kanamycin resistance cassette as a PstI fragment, whose loss facilitates recognizing promoter-lacZ fusion plasmids as kanamycin-sensitive (Kan^s) derivatives (15). Fusions made with these plasmids were recombined onto the chromosome by allele replacement as described below.

FIG. 1.

Structure of the phoBR chromosomal region. The 4.7-kbp PstI-to-EcoRI fragment in pBC35 and pSK47 is shown. Asterisks mark the NdeI and BamHI sites that were previously introduced by site-directed mutagenesis upstream and downstream, respectively, of the phoB coding region. The ΔphoB578 mutation was made in pLD82 (Table 2) and recombined onto the chromosome by allele replacement as described in the text. The construction of ΔphoR574 and ΔphoBR580 is described elsewhere (12). Arrows show gene orientations.

All plasmids constructed in this study are described in Table 2. Many have the γ replication origin of R6K, denoted oriR_R6Kγ, which requires the Π protein (encoded by pir) for replication as a plasmid. Many also contain tetAR, which can be deleterious when present at high copy number. Therefore, these plasmids were routinely maintained in BW23473 or BW24249 or in similar pir⁺ hosts (17). λRZ5lacP-phoU⁺ (PS15) has been described previously (26). Generalized transduction was carried out with P1kc from a laboratory stock.

TABLE 2.

Plasmid constructions

Plasmid	Descriptiona	Constructionb
pAH16	Plac-phoR+bla oriRMB1lacI+	Replaced 0.4-kbp EcoRI-to-BamHI vanS fragment of pSLF13 with similar 1.3-kbp phoR+ fragment generated by using P109 and P110 with pBC35 as the template
pAH21c	ParaB-phoR+cat oriR15AaraC+	Replaced 0.4-kbp NdeI-to-SacI vanS fragment of pSLF22 with similar 1.3-kbp phoR+ fragment from pAH16
pAH31	ParaB-lacZ tetAR bla oriRR6Kγ	Replaced 0.5-kbp XhoI-to-PstI kan fragment of pWJ18 with 0.5-kbp ParaB PCR fragment generated by using P125 and P126 with pBAD33 as the template
pAH33d	ParaB-′araD′ tetAR bla oriRR6Kγ	Replaced 3.0-kbp SalI-to-BsiWI lacZ fragment of pAH31 with 0.6-kbp ′araD′ PCR fragment generated by using P127 and P128 with BW13711 chromosomal DNA as the template
pAH35	ParaB-phoR+-′araD′ tetAR bla oriRR6Kγ	Subcloned 1.4-kbp BamHI phoR+ fragment from pAH21 into pAH33
pAH54	Like pAH33 except lacking NdeI2	Digested pAH33 with BsiWI and partially with NdeI followed by filling-in with T4 DNA polymerase and religation
pLD68	PrhaB-lacZ tetAR bla oriRR6Kγ	Replaced 1.3-kbp XhoI-to-SalI kan fragment of pWJ18 with similar 0.3-kbp PrhaSB fragment generated by using P119 and P120 with pLD1 (3) as the template
pLD69	PrhaS-lacZ tetAR bla oriRR6Kγ	Same as pLD68, except fragment is in opposite orientation
pLD70	′rhaD tetAR bla oriRR6Kγ	Replaced 3.0-kbp BamHI-to-BsiWI lacZ fragment of pWJ17 with 0.6-kbp ′rhaD fragment generated by using P121 and P122 with BW13711 chromosomal DNA as the template
pLD71	PrhaB-′rhaD tetAR bla oriRR6Kγ	Subcloned 0.7-kbp XhoI-to-BamHI PrhaB fragment from pLD68 into pLD70
pLD77	Like pWJ17 except lacking NdeI2	Digested pWJ17 partially with NdeI followed by filling-in with T4 DNA polymerase and religation
pLD78	Like PLD71 except lacking NdeI2	Replaced 4.3-kbp XhoI-to-BsiWI lacZ fragment of pLD77 with similar 0.8-kbp PrhaB-′rhaD fragment from pLD71
pLD79	PrhaB-phoB+-′rhaD tetAR bla oriRR6Kγ	Subcloned 0.7-kbp NdeI-to-BamHI phoB+ fragment from pSK2 (9) into pLD78 digested with BamHI and partially with NdeI
pLD82	ΔphoB578 kan oriRR6Kγ	Digested pSK47 with NdeI and BamHI followed by filling-in with T4 DNA polymerase and religation
pLD83	ΔphoB578 tetAR bla oriRR6Kγ	Subcloned 3.4-kbp NcoI fragment from pLD82 that was made blunt with T4 DNA polymerase into SmaI-cut pLD55 (17)

^aThe replication origins of pMB1 and p15A are denoted oriR_MB1 and oriR_15A, respectively. 

^bAdditional information is given in the text. Oligonucleotide primers are shown in Table 3. 

^cpAH21 has a unique NdeI site at the phoR start codon and unique downstream SacI, SphI, PstI, and HindIII sites. 

^dRestriction analysis showed that the SphI and BglII sites in P127 were absent from the cloned fragment. 

Molecular biology methods.

PCR amplifications of ara, rha, and phoR sequences were carried out with Vent DNA polymerase (New England Biolabs, Beverly, Mass.) and oligonucleotide primers (Table 3; IDT Inc., Coralville, Iowa). Other enzymes were from New England Biolabs or Promega (Madison, Wis.). QIAGEN (Hilden, Germany) products were used for isolation of plasmid DNA, extraction of DNA fragments from agarose gels, and purification of PCR fragments. The phoR, P_araB, and P_rhaSB fragments were sequenced on both strands at the Dana Farber Cancer Institute Molecular Biology Core Facility (Harvard Medical School, Boston, Mass.). The ‵araD′ and ‵rhaD fragments (the prime indicates that a gene portion is missing on the side of the prime) were verified only by restriction enzyme analysis.

TABLE 3.

Oligonucleotide primers

Name	Description	Sequencea
P109	phoR N terminus	5′-GGAATTCGTCATATGCTGGAACGGC-3′
P110	phoR C terminus	5′-CGGGATCCGCTGAGCTCAGTCGCTGTTTTTGGC-3′
P119	PrhaSB upstream	5′-CCGCTCGAGGCCATGGTGGCCTCCTGATGTCG-3′
P120	PrhaSB downstream	5′-CCGGTCGACATATGGTGATCCTGCTGAATTTC-3′
P121	rhaD′ upstream	5′-CGGGATCCGCATGCGTCGACAGCGACGG-3′
P122	rhaD′ downstream	5′-CGTACGTACGCAGCGCCAGCGCAC-3′
P125	ParaB downstream	5′-GCGCGCTGCAGGTATGGAGAAACAGTAGAG-3′
P126	ParaB upstream	5′-GGTCCTCGAGCGAATGGTGAGATTG-3′
P127	′araD′ upstream	5′-GACACGTCGACGCATGCAGATCTCAAACGCCAG-3′
P128	′araD′ downstream	5′-CGTACGTACGACCTCTTCCAGCAC-3′

^aUnderlined bases are complementary to the template(s). Those in boldface indicate restriction enzyme sites. 

Molecular genetics.

Many new strains were constructed by generalized P1 transduction as described elsewhere (28). BW23321 (Δlac-169 hsdR514 uidA(ΔMluI)::pir⁺ endA_BT333) was made by transduction of BW21116 (17) to tetracycline resistance (Tet^r) with P1 grown on BT333 followed by transformation of resultant transductant BW23298 with FLP expression plasmid pCP20 and excision of the tetAR genes as described elsewhere (4). The notation endA_BT333 signifies the resultant allele. In order to construct strains carrying a minimum number of antibiotic resistance markers, new mutations were usually introduced in two steps. In the first step, a linked auxotrophic marker was introduced by selecting Tet^r transductants. These were then made prototrophic with P1 grown on an appropriate donor; the resultant transductants were scored for loss of antibiotic resistance and other relevant phenotypes. Several new mutations and fusions were constructed on plasmids as described above and then recombined onto the chromosome by allele replacement as described elsewhere (17). The resultant chromosomal allele is often given a designation corresponding to the plasmid used in its construction. As necessary, transductants carrying a ΔphoU mutation were verified by complementation with λRZ5lacP-phoU⁺ upon introduction of phoB⁺ or phoR⁺ or by backcross of the ΔphoU mutation into an appropriate recipient.

The ΔphoB578 mutation (Fig. 1) was recombined onto the chromosome of BW21016 [DE3(lac)_X74] by using pLD83 to make BW22901. The P_rhaB-lacZ_LD68 and P_rhaS-lacZ_LD69 fusions were recombined onto the chromosome of BW21578 (rrnB_T14 ΔlacZ_WJ16) by using pLD68 and pLD69 to make BW22716 and BW22721, respectively. The P_araB-lacZ_AH31 fusion was recombined onto the chromosome of BW21480 (lacI^q rrnB_T14 ΔlacZ_WJ16) by using pAH31 to make BW22746. The desired lacZ fusion recombinants were recognized as ones showing a rhamnose- or arabinose-dependent Lac⁺ phenotype on X-Gal agar. These fusions have the following on the chromosome at the lac locus in a counterclockwise orientation: lacI, four tandem copies of the rrnB transcriptional terminator (denoted rrnB_T14), and the foreign promoter preceding the lacZYA operon. The construction of these and similar strains with lacI^q rrnB_T14 P_phnC-lacZ_WJ19, lacI^q rrnB_T14 ΔlacZ_WJ16, rrnB_T14 ΔlacZ_WJ16, and other promoter-lacZ fusions is unpublished (15). Site-specific recombination of oriR_R6Kγ attP plasmids onto the chromosome has been described previously (13). All integrants were verified by PCR as described elsewhere (12).

The P_araB and P_rhaB fusions were recombined onto the chromosome by using derivatives of new allele replacement plasmids pAH33, pAH54, and pLD78 (Fig. 2). This was done with an Ara⁺ or Rha⁺ parental strain, so that the resulting segregants with an allele replacement were recognizable as Ara⁻ or Rha⁻ recombinants, respectively. Because these plasmids contain a segment of araD or rhaD, most integrants formed by the initial recombination event are AraD⁻ or RhaD⁻. Such recombinants are arabinose or rhamnose sensitive, respectively. All selections are therefore carried out in the absence of arabinose or rhamnose.

FIG. 2.

Structures of the P_araB and P_rhaB allele replacement plasmids. Arrows show gene and promoter orientations. Open boxes in pAH33, pAH54, and pLD78 are promoter regions; hatched boxes are coding regions. All sites are shown for the enzymes indicated. NdeI sites are subscripted to correspond to those indicated in Table 2. pAH33 and pAH54 contain unique PstI and SalI sites downstream of P_araB for construction of a P_araB fusion. Genes cloned into these sites require also a ribosome binding site. pLD78 has SalI, XbaI, BamHI, and SphI sites downstream of P_rhaB for construction of a P_rhaB fusion. pLD78 has the native rhaB ribosome-binding site upstream of the NdeI₂ site, which corresponds to the normal met start codon of rhaB. Therefore, genes cloned into this site do not require a ribosome-binding site, although partial digestions are required to use this site. An asterisk marks SphI and BglII sites that were lost during cloning of the ‵araD′ fragment. Arrows show gene and promoter orientations. bla, β-lactamase gene; tetA and tetR, tetracycline resistance and repressor genes, respectively; oriT, origin of transfer from RP4; T1₄ and T1T2₂, transcription terminators; kan, kanamycin resistance gene; lacZ (op), promoterless lacZ gene for construction of transcriptional, i.e., operon (op), fusions (17). See text.

The ΔaraBAD_AH33 mutation was recombined onto the chromosome of BW13711 [DE3(lac)_X74] by using pAH33. Integrants carrying pAH33 were selected as Tet^r transformants; they were purified once nonselectively, after which Tet^s segregants were selected on TSS agar. One showing the expected Ara⁻ phenotype was verified and named BW22826. Similarly, the P_araB-phoR[M1-D431]_AH35 fusion was recombined onto the chromosome of BW21555 (ΔphoR574) by using pAH35 to make BW22835. Recombination of the P_araB-phoR[M1-D431]_AH35 fusion onto the chromosome results also in removal of the same sequences eliminated by the ΔaraBAD_AH33 mutation. Hence, the resultant allele is denoted ΔaraBAD_AH33::P_araB-phoR[M1-D431]_AH35. The ΔrhaBAD_LD78 mutation and ΔrhaBAD_LD78::P_rhaB-phoB_LD79 fusion were recombined onto the chromosome of BW22860 [Δ(phoBR brnQ)525 Δcya-161] by using pLD78 and pLD79 to make BW22875 and BW22876, respectively. The desired recombinants were recognized as Rha⁻ ones.

Enzyme assays.

β-Galactosidase and bacterial alkaline phosphatase (BAP) assays were carried out as described elsewhere (28).

RESULTS

Construction of allele replacement vectors for tightly regulated protein synthesis from arabinose- and rhamnose-regulated promoters.

Since expression of the phoBR operon is subject to positive autogenous control (10), we considered that it would be advantageous to uncouple PhoB synthesis and PhoR synthesis from their normal controls for new studies on the Pho regulon. We therefore constructed a derivative of P_araB plasmid pBAD32 (11) carrying a P_araB-phoR⁺ fusion (pAH21; Table 2). However, in preliminary studies, we found that substantial amounts of PhoR were apparently made in the absence of inducer. A ΔphoR mutant carrying pAH21 synthesized BAP upon P_i limitation in the absence of arabinose even under conditions of catabolite repression due to glucose (data not shown). No doubt the “leakiness” of P_araB under these conditions reflects the small amount of PhoR required for activation (phosphorylation) of PhoB. To further reduce PhoR synthesis, we assembled an allele replacement, “suicide” vector system to recombine the P_araB-phoR⁺ fusion onto the chromosome in single copy at the ara locus. We also constructed an analogous allele replacement vector carrying the rhamnose-regulated P_rhaB promoter (6) for construction and recombination of a P_rhaB-phoB⁺ fusion onto the chromosome in single copy at the rha locus.

Our vectors for constructing chromosomal P_araB and P_rhaB fusions are illustrated in Fig. 2. pAH33 and pAH54 are useful for making a P_araB fusion; pLD78 is useful for making a P_rhaB fusion. These plasmids contain sequences of the araCBAD or rhaRSBAD locus flanking cloning sites for construction of the respective promoter fusion. pAH33 and pAH54 have a 0.5-kbp fragment containing P_araB and a 0.6-kbp fragment of araD upstream and downstream of the cloning region, respectively. They differ in that pAH33 has two NdeI sites (one in tetR and the other in a segment of lacZ) and pAH54 has only one NdeI site (in tetR; Table 2). Derivatives of the latter have facilitated the use of NdeI in particular plasmid constructions (data not shown). pLD78 has a 0.3-kbp fragment containing P_rhaB and a 0.6-kbp fragment of rhaD upstream and downstream of its cloning region, respectively. The flanking upstream and downstream sequences provide homologous regions for recombining the fusions onto the chromosome by allele replacement.

pAH33, pAH54, and pLD78 are derivatives of pWJ18 (or pWJ17; Fig. 2), which is in turn a derivative of pir-dependent, counterselectable (tetAR) allele replacement vector pLD53 (17). Upon introduction of these plasmids into a normal (non-pir) Ara⁺ host (by transformation, electroporation, or conjugation), they cannot replicate and therefore integrate into the chromosome via homologous recombination. Figure 3 shows the integration of P_araB-phoR⁺ plasmid pAH35 (Table 2) at the araCBAD locus. Integrants are selectable as Tet^r colonies. A subsequent recombination event occurring in the absence of selection leads to the loss of plasmid sequences; this event results in restoration of the parental (wild-type) chromosomal structure or an allele replacement. Recombinants that have lost the plasmid backbone are selectable as Tet^s ones. Those carrying the desired chromosomal P_araB or P_rhaB fusion have the fused gene in single copy at the araCBAD or rhaRSCAB locus in place of araBAD or rhaBAD sequences (Fig. 4); they are therefore recognizable as Ara⁻ or Rha⁻ ones, respectively. Ones that are also antibiotic sensitive are then verified by genetic linkage or PCR tests. Because the appropriate segregants are Ara⁻ or Rha⁻, the recombinants provide the additional advantage of not being able to catabolize the respective inducer.

FIG. 3.

Recombination of a P_araB-phoR⁺ fusion onto the chromosome at the araCBAD locus. The construction of pAH35 is described in Table 2. pAH35 can integrate into the chromosome via either of two homologous recombination events (event A or event B). Only an event A integrant is shown for simplicity. Subsequent recombination events result in loss of the plasmid, regenerating the parental strain (event A) or an allele replacement (event B segregant). These are selectable on TSS agar; ones with an allele replacement are recognizable as Ara⁻ colonies (see Materials and Methods). A ΔphoR mutant was used to recombine the P_araB-phoR⁺ fusion onto the chromosome to avoid recombination at the phoBR locus; a wild-type strain can also be used. A black arrow shows the orientation of phoR. Grey arrows show the orientations of other genes. Grey boxes show araC or araD gene fragments and the pir-dependent vector origin (oriR_R6Kγ). A prime indicates that a gene portion is missing on the side of the prime. An arrowhead marks the nick site of oriT, the RP4 conjugative transfer origin.

FIG. 4.

Allele replacements at the chromosomal araCBAD and rhaRSBAD regions. (A) Structure of the ΔaraBAD mutant and a P_araB-phoR⁺ fusion at the araCBAD locus. (B) Structure of the ΔrhaBAD mutant and a P_rhaB-phoB⁺ fusion at the rhaRSBAD locus. Black arrows represent P_araB-phoR and P_rhaB-phoB fusions. Light grey arrows signify genes belonging to the ara and rha gene clusters. Dark grey arrows show the yabI and polB genes flanking the ara locus and the sodA and yiiL genes flanking the rha locus. Dotted lines indicate regions where homologous recombination between the plasmid and the chromosome can occur. See text.

Measurements of arabinose and rhamnose regulation of single-copy promoter-lacZ fusions.

To judge the regulatory capabilities of the arabinose- and rhamnose-regulated promoters in this system, we constructed P_araB-, P_rhaB-, and P_rhaS-lacZ fusions and recombined them onto the chromosome in single copy at the lac locus (Materials and Methods). We then examined lacZ expression by measuring β-galactosidase levels when strains were grown in the presence and absence of the respective inducer on different carbon sources. Different carbon sources were used because these promoters are subject to catabolite repression. The results are shown in Table 4. In brief, arabinose or rhamnose led to induction of >100-fold to several thousandfold, depending upon the promoter, inducer, and carbon source. As expected, the lowest expression levels were obtained during growth on glucose, for which catabolite repression is the strongest; intermediate expression levels were obtained during growth on fructose, for which catabolite repression is less severe; the highest expression levels were obtained during growth on glycerol, for which catabolite repression is the weakest (7). Differences in both the basal (uninduced) and induced levels exist. In the absence of inducer, the basal level was about 10-fold higher for P_araB than for P_rhaB or P_rhaS. The induced levels were similar for P_araB and P_rhaB (compare BW22831 with BW22887), while the induced levels for P_rhaS were ca. 25% of those for P_rhaB (compare BW22886 with BW22888).

TABLE 4.

Regulation of P_araB-, P_rhaB-, and P_rhaS-lacZ fusions

Strain	Relevant genotypea	Carbon source	β-Galactosidase sp act onb:		Fold induction
			−Ind	+Ind
BW22831	ParaB-lacZ rpoS(Am)	d-Glucose	6.6 ± 0.1	1,270 ± 114	190
		d-Fructose	6.7 ± 0.6	3,861 ± 313	580
		Glycerol	6.5 ± 0.4	3,853 ± 119	590
BW22887	PrhaB-lacZ rpoS(Am)	d-Glucose	0.5 ± 0.0	743 ± 44	1,600
		d-Fructose	0.5 ± 0.0	3,300 ± 67	6,900
		Glycerol	0.5 ± 0.0	3,689 ± 72	7,800
BW22886	PrhaB-lacZ rpoS+	d-Glucose	0.4 ± 0.0	185 ± 10	430
		d-Fructose	0.4 ± 0.0	1,158 ± 33	2,700
		Glycerol	0.4 ± 0.0	2,209 ± 50	5,800
BW22888	PrhaS-lacZ rpoS+	d-Glucose	0.4 ± 0.0	56 ± 2	130
		d-Fructose	0.5 ± 0.1	253 ± 5	490
		Glycerol	0.8 ± 0.1	619 ± 21	810

^aAll strains are also ΔaraBAD or ΔrhaBAD, as appropriate; complete genotypes are given in Table 1. 

^bCells were assayed after 18 h of growth in 0.06% glucose–, 0.06% fructose–, or 0.1% glycerol–MOPS–2 mM P_i without (−Ind) or with (+Ind) the inducer arabinose or rhamnose. Values are nanomoles of o-nitrophenol made per unit of cell culture optical density at 420 nm (means ± standard deviations). Strains were grown and assayed in triplicate. 

In the course of this study, we found that several of our strains carry an rpoS(Am) mutation (Table 1), which is also present in many lines of E. coli K-12, including progenitor K-12 strains EMG2 and W1485 (1, 22). As shown in Table 4, the induced level of P_rhaB expression is significantly lower in an rpoS⁺ strain. Apparently, this rpoS effect is related to catabolite repression, as the magnitude varies with the carbon source. The reason for this is unknown. The induction levels in this study are therefore directly comparable only among strains having the same rpoS allele.

Strategy for using P_araB and P_rhaB fusions to study regulation of the Pho regulon.

We constructed a number of ΔphoB and ΔphoR mutants in which PhoB or PhoR is synthesized under the control of a foreign inducible promoter(s). We did this for two reasons. First, we considered that such strains may be useful in our studies on protein-protein interactions between their gene products, as reported elsewhere (13). Second, we suspected that the deleterious effects of a ΔphoU mutation were in part due to high-level expression of the Pho regulon. Because expression of the phoBR operon is under positive autogenous control, high-level expression of the Pho regulon probably requires also high-level synthesis of PhoB and PhoR. We therefore expected to overcome these deleterious effects by down regulating PhoB or PhoR synthesis by using these foreign-regulated promoters.

Arabinose-independent expression of P_araB fusions.

The above results show the P_araB promoter to be tightly regulated for expression of the lacZ structural gene. To assess whether P_araB is useful for tight control of a regulatory gene under similar conditions, we constructed strains carrying chromosomal P_araB-phoB⁺ and P_araB-phoR⁺ fusions. Each strain has also a precise deletion of the respective regulatory gene. They have in addition a deletion of creC (ΔcreBCD or ΔcreABCD) as well as genes for acetyl phosphate synthesis [Δ(ackA pta)], thus eliminating activation of PhoB by the kinase CreC or acetyl phosphate (33). We then examined PhoB- and PhoR-dependent control by measuring BAP levels under conditions of P_i limitation in the absence of arabinose during growth on different carbon sources. As shown in Table 5, P_araB-phoB⁺ ΔphoB strain BW24803 exhibits a null phenotype in the absence of arabinose on all carbon sources. In contrast, P_araB-phoR⁺ ΔphoR strain BW22861 exhibits a PhoR⁺ phenotype in the absence of arabinose, even during growth on glucose. These results are consistent with less PhoR than PhoB being required for transcriptional activation of phoA. Therefore, P_araB appears to be sufficiently tightly controlled for the expression of phoB, but not for the expression of phoR. Yet, phoR expression is clearly limiting under these conditions, as much higher BAP levels are seen in the presence of arabinose (data not shown).

TABLE 5.

Inducer-independent expression of P_araB-phoB⁺ and P_araB-phoR⁺ fusions

Strain	Relevant genotypea	BAP sp act onb:
		d-Glucose	d-Fructose	Glycerol
BW24803	ParaB-phoB+ ΔphoB phoR+	0.3 ± 0.0	0.4 ± 0.0	0.4 ± 0.0
BW22861	ParaB-phoR+phoB+ ΔphoR	26.4 ± 4.9	47.7 ± 5.1	58.8 ± 4.4

^aStrains are ΔcreABCD Δ(ackA pta) ΔaraBAD; complete genotypes are given in Table 1. The P_araB-phoB⁺ fusion is in single copy at att_λ; the P_araB-phoR⁺ fusion is in single copy at the araCBAD locus. 

^bCells were assayed after 18 h of growth in 0.06% glucose–, 0.06% fructose–, or 0.1% glycerol–MOPS medium containing 0.1 mM P_i without arabinose. Values are nanomoles of p-nitrophenol made per unit of cell culture optical density at 420 nm (means ± standard deviations). Strains were grown and assayed in triplicate. 

An eventual goal was to express both phoB and phoR independently and simultaneously from a foreign promoter(s). In order to find appropriate conditions to do this, we used strains with a deletion of the pstSCAB-phoU operon that leads to constitutive expression of the Pho regulon. Individual ones are also ΔphoB or ΔphoR as well as ΔcreC and Δ(ackA pta), as described above. As shown in Table 6, phoA expression is abolished in the absence of PhoB or PhoR (compare BW24741 and BW24740 with BW24739). Introduction of a P_phoB-phoB⁺ fusion in single copy elsewhere on the chromosome restores normal phoA expression (compare BW24770 with BW24739). Introduction of a P_tac-phoB⁺ fusion in single copy restores phoA expression partially in the absence of IPTG and fully in the presence of IPTG (compare BW24775 without and with IPTG). This apparent leakiness of P_tac was expected, as proper regulation by LacI requires additional upstream and downstream operator sequences (20) that are absent in the P_tac-phoB⁺ fusion. Yet partial inducer dependence is observed even for expression of a regulatory gene from P_tac.

TABLE 6.

Inducer-dependent control by using a P_tac-phoB⁺ fusion

Strain	Relevant genotypea	BAP sp act onb:
		−IPTG	+IPTG
BW24739	phoBR+ Δ(pstSCAB-phoU)	376 ± 19	N.D.
BW24741	ΔphoB phoR+ Δ(pstSCAB-phoU)	0.2 ± 0.0	N.D.
BW24740	phoB+ ΔphoR Δ(pstSCAB-phoU)	0.2 ± 0.0	N.D.
BW24770	PphoB-phoB+ ΔphoB phoR+ Δ(pstSCAB-phoU)	376 ± 17	N.D.
BW24775	Ptac-phoB+ ΔphoB phoR+ Δ(pstSCAB-phoU)	21.6 ± 0.9	355 ± 12

^aAll strains are lacI^q ΔcreABCD Δ(ackA pta); complete genotypes are given in Table 1. The P_phoB-phoB⁺ and P_tac-phoB⁺ fusions are in single copy at att_λ. 

^bCells were assayed after 18 h of growth in 0.06% glucose–MOPS–2 mM P_i without (−) or with (+) IPTG. BAP specific activity values are as defined for Table 5. N.D., not determined. 

Arabinose- and rhamnose-dependent expression of P_araB, P_rhaB, and P_rhaS fusions.

We compared strains carrying chromosomal P_araB-phoB⁺, P_rhaB-phoB⁺, P_araB-phoR⁺, P_rhaB-phoR⁺, and P_rhaS-phoR⁺ fusions to determine which promoter fusion(s) and growth conditions were appropriate for studying gene regulation in the Pho regulon. This was done with strains that are otherwise similar to ones described above. We examined PhoB- and PhoR-dependent control of phoA expression by measuring BAP levels when strains were grown on one of four carbon sources (glucose, mannitol, fructose, or glycerol), which were expected to result in different levels of catabolite repression (7). As shown in Table 7, strains carrying a P_araB-phoB⁺ or P_rhaB-phoB⁺ fusion express a null phenotype in the absence of inducer. The same ones synthesized ca. 12-fold to 2,400-fold more BAP in the presence of inducer, depending upon the carbon source and promoter. The relative expression levels correlate well with expectation in regards to catabolite repression. Catabolite repression is more severe with glucose than mannitol, more severe with mannitol than fructose, and more severe with fructose than glycerol. The BAP levels in these P_araB-phoB⁺ and P_rhaB-phoB⁺ fusion strains (ca. 280 to 490 U; BW24774, BW24777, and BW24773) (Table 7) during growth on glycerol are similar to that in an otherwise isogenic wild-type strain (ca. 375 U; BW24739) (Table 6). Strains with a P_rhaB-phoR⁺ fusion at two chromosomal sites were examined; no difference between them was seen.

TABLE 7.

Inducer-dependent control by using P_araB-phoB⁺, P_rhaB-phoB⁺, P_araB-phoR⁺, P_rhaB-phoR⁺, and P_rhaS-phoR⁺ fusions

Strain	Relevant genotypea	Inducerb	BAP sp act onc:
			d-Glucose	d-Mannitol	d-Fructose	Glycerol
BW24774	ParaB-phoB+ ΔphoB Δ(pstSCAB-phoU)	None	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0
		Ara	4.0 ± 0.4	103 ± 0	174 ± 6	281 ± 5
BW24777	PrhaB-phoB+ ΔphoB Δ(pstSCAB-phoU)	None	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0
		Rha	2.6 ± 0.3	120 ± 23	350 ± 36	442 ± 5
BW24773	PrhaB-phoB+ ΔphoB Δ(pstSCAB-phoU)	None	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0
		Rha	2.4 ± 0.3	104 ± 9	359 ± 11	486 ± 2
BW24508	ParaB-phoR+ ΔphoR Δ(pstSCAB-phoU)	None	24.8 ± 0.5	36.3 ± 1.6	75.5 ± 7.4	118 ± 3
		Ara	149 ± 6	483 ± 9	336 ± 29	400 ± 27
BW24768	PrhaB-phoB+ ΔphoR Δ(pstSCAB-phoU)	None	0.2 ± 0.0	0.2 ± 0.0	0.3 ± 0.0	0.2 ± 0.0
		Rha	3.4 ± 0.8	108 ± 13	357 ± 4	464 ± 2
BW24858	PrhaS-phoR+ ΔphoR Δ(pstSCAB-phoU)	None	78.5 ± 5.4	85.7 ± 8.1	161 ± 11	257 ± 15
		Rha	92.4 ± 4.5	175 ± 13	398 ± 14	673 ± 5

^aAll strains are ΔcreABCD Δ(ackA pta) and ΔaraBAD or ΔrhaBAD, as appropriate; complete genotypes are given in Table 1. The P_araB-phoB⁺ and P_rhaB-phoB⁺ fusions are in single copy at att_λ; the P_rhaB-phoB⁺ and P_araB-phoR⁺ fusions are in single copy at the rha and ara loci, respectively; the P_rhaB-phoR⁺ and P_rhaS-phoR⁺ fusions are in single copy at att_HK022. 

^bCells were assayed after 18 h of growth in 0.06% glucose–, 0.06% mannitol–, 0.06% fructose–, or 0.1% glycerol–MOPS medium containing 2 mM P_i without (none) or with arabinose (Ara) or rhamnose (Rha) for induction. 

^cBAP specific activity values are as defined for Table 5. 

Of the three phoR fusion strains examined, only one (P_rhaB-phoR⁺ fusion strain BW24768) showed a null phenotype in the absence of inducer. Substantial activation was apparent in both the P_araB-phoR⁺ and P_rhaS-phoR⁺ fusion strains (BW24508 and BW24858) in the absence of inducer even during growth on glucose, indicating that both P_araB and P_rhaS are somewhat leaky. Their basal levels of expression are also subject to catabolite repression, as they vary with the carbon source. Upon induction with rhamnose, the P_rhaB-phoR⁺ fusion strain synthesized ca. 17-fold to 2,300-fold more BAP; these BAP levels also correlate well with the carbon source. The addition of the respective inducer resulted in increased BAP synthesis in the P_araB-phoR⁺ and P_rhaS-phoR⁺ fusion strains as well.

Use of foreign promoters to bypass growth defect due to a ΔphoU mutation.

We showed above that we can modulate expression of the Pho regulon in our P_araB-phoB⁺, P_rhaB-phoB⁺, and P_rhaB-phoR⁺ fusion strains by growing them on different carbon sources in the presence of the respective inducer. We therefore tested whether these conditions can be used to overcome the deleterious effect of a ΔphoU mutation. Whereas a ΔphoU mutant such as BW17142 grows extremely poorly under all conditions tested, an otherwise isogenic Δ(pstSCAB-phoU) mutant such as BW17335 grows reasonably well (Table 8) (26). These strains have a kanamycin resistance cassette in place of phoU and pstSCAB-phoU sequences, respectively. A similar strain with an unmarked ΔphoU mutation (BW18897; Table 8) also shows a severe growth defect. Yet all these strains synthesize similar amounts of BAP. To determine whether modulating expression of phoB or phoR can be used to overcome the growth inhibition due to a ΔphoU mutation, we constructed otherwise isogenic ΔphoU and Δ(pstSCAB-phoU) strains with a ΔphoB and ΔphoR mutation carrying the respective fusions. These strains have an unmarked ΔphoU mutation, as each has a kanamycin resistance marker elsewhere (Table 1).

TABLE 8.

Pho regulon control in ΔphoU mutants by using P_araB-phoB⁺, P_rhaB-phoB⁺, and P_rhaB-phoR⁺ fusions

Strain	Relevant genotypea	Growthb	BAP sp act onc:
			−Ind	+Ind
BW17142	ΔphoU	+/−	419d	N.D.
BW17335	Δ(pstSCAB-phoU)	+++	428d	N.D.
BW18897	ΔphoU	+	444 ± 33	N.D.
BW24739	Δ(pstSCAB-phoU) Δ(ackA pta)	+++	443 ± 20	N.D.
BW24774	ParaB-phoB+ ΔphoB Δ(pstSCAB-phoU) Δ(ackA pta)	++++	0.2 ± 0.0	94.8 ± 3.5
BW24950	ParaB-phoB+ ΔphoB ΔphoU Δ(ackA pta)	++++	0.2 ± 0.0	113 ± 3
BW24773	PrhaB-phoB+ ΔphoB Δ(pstSCAB-phoU) Δ(ackA pta)	++++	0.2 ± 0.0	109 ± 27
BW24949	PrhaB-phoB+ ΔphoB ΔphoU Δ(ackA pta)	++++	0.2 ± 0.0	89.7 ± 6.4
BW24768	PrhaB-phoR+ ΔphoR Δ(pstSCAB-phoU) Δ(ackA pta)	++++	0.2 ± 0.0	104 ± 2
BW24948	PrhaB-phoR+ ΔphoR ΔphoU Δ(ackA pta)	++++	0.2 ± 0.0	83.4 ± 8.9

^aComplete genotypes are given in Table 1. The P_araB-phoB⁺ and P_rhaB-phoB⁺ fusions are in single copy at att_λ; the P_rhaB-phoR⁺ fusion is in single copy at att_HK022. 

^bColony size after 24 h of incubation +/−, barely detectable; +, tiny; +++, ca. 1 mm in diameter; ++++, ca. 1.5 mm in diameter. 

^cCells were assayed after 18 h of growth in 0.06% mannitol–MOPS–2 mM P_i without (−Ind) or with (+Ind) arabinose or rhamnose for induction. BAP specific activity values are as defined for Table 5. N.D., not determined. 

^dBAP specific activity value is an average of four independent determinations (unpublished data cited in reference 26). 

As shown in Table 8, our ΔphoU ΔphoB, Δ(pstSCAB-phoU) ΔphoB, ΔphoU ΔphoR, and Δ(pstSCAB-phoU) ΔphoR strains carrying a P_araB-phoB⁺, P_rhaB-phoB⁺, or P_rhaB-phoR⁺ fusion grow equally well in the absence or presence of the respective inducer. The P_araB-phoB⁺ and P_rhaB-phoB⁺ fusion strains (BW24774, BW24950, BW24773, and BW24949) show a null phenotype in the absence of inducer. Each synthesizes ca. 500-fold more BAP in the presence of the respective inducer. Likewise, the P_rhaB-phoR⁺ fusion strains (BW24768 and BW24948) show a null phenotype in the absence of rhamnose; each synthesizes ca. 500-fold more BAP in its presence. Upon induction these strains synthesize ca. 25% of the normal amount of BAP (see BW24739 for comparison), suggesting that this level of Pho regulon gene expression is apparently insufficient to result in a severe growth defect due to a ΔphoU mutation.

DISCUSSION

Our primary goal was to develop a method(s) to study PhoU function. The severe growth defect of a ΔphoU mutant is especially problematic due to the rapid accumulation of compensatory mutants with lesions in phoB, phoR, or a pst gene. A Pst⁺ ΔphoU mutant is also exquisitely sensitive to extracellular P_i, suggesting that PhoU has, in addition to its function in P_i signaling, a role as an enzyme in intracellular P_i metabolism (26). As shown here, we were able to overcome the deleterious effect of a ΔphoU mutation by uncoupling PhoB or PhoR synthesis from its normal autogenous control and expressing phoB or phoR from a foreign promoter(s). To do this, we constructed strains carrying a P_araB-phoB⁺, P_rhaB-phoB⁺, P_araB-phoR⁺, or P_rhaB-phoR⁺ fusion on the chromosome. We used strains with single-copy chromosomal fusions to avoid problems resulting from the use of plasmids, such as an antibiotic requirement, variable plasmid copy number, and even plasmid loss. We also examined ways for modulating expression levels from the P_araB and P_rhaB promoters. We already reported work with similar strains and growth conditions to study the regulatory genes of the Enterococcus faecium vancomycin resistance gene cluster in an E. coli model system (12).

We constructed the corresponding P_araB and P_rhaB fusion strains with conditionally replicative (pir-dependent) allele replacement plasmids pAH33, pAH54, and pLD78 (Fig. 2). Strains carrying these fusions are made via a simple two-step procedure as illustrated in Fig. 3. The resulting P_araB and P_rhaB fusions reside on the chromosome at the araCBAD and rhaRSBAD loci, respectively, as shown in Fig. 4. We made similar strains with the respective ΔaraBAD_AH33 and ΔrhaBAD_LD78 chromosomal mutations as controls. In these ways, we constructed a number of strains that express various normal or mutant pho, cre, or van genes under the control of P_araB or P_rhaB in single copy on the chromosome at the respective loci (12–14, 16).

The P_araB promoter (also called P_BAD) has been shown elsewhere to be tightly regulated (11, 21). In those studies, genes that are normally expressed at moderate levels were examined. In preliminary studies, we found that a pBAD plasmid (11) carrying phoR did not provide sufficiently tight control for our purposes, suggesting that these plasmids are somewhat leaky (14). This is consistent with phoR normally being expressed at a very low level. In order to reduce this basal level of expression, we recombined several P_araB-phoR fusions onto the chromosome by using derivatives of P_araB fusion allele replacement plasmid pAH33. When single-copy fusion strains were used, particular P_araB-phoR fusions, e.g., one expressing only the C-terminal kinase domain of PhoR (13), no longer appeared to be leaky. Yet a single-copy P_araB-phoR⁺ fusion strain synthesizes sufficient PhoR for (partial) activation in the absence of arabinose even in the presence of glucose (Table 5). These results suggest that full-length PhoR is a more active kinase than its C-terminal domain. However, in the absence of arabinose, the amount of PhoR synthesis from a P_araB-phoR⁺ fusion is also clearly limiting because upon P_i limitation only ca. 10% as much BAP is made in this strain as in a wild-type strain. In contrast, a single-copy P_araB-phoB⁺ fusion strain shows no leakiness as it exhibits a PhoB⁻ phenotype in the absence of inducer on all carbon sources tested (Table 7). Therefore, more PhoB than PhoR is apparently required for expression of the Pho regulon. These data are consistent with PhoR acting catalytically as an autokinase and phosphotransferase in the activation (phosphorylation) of PhoB, which in turn acts as a transcription factor.

In order to control PhoB and PhoR synthesis independently, we also constructed strains with a P_rhaB-phoB⁺ or P_rhaB-phoR⁺ fusion. Although both l-arabinose and l-rhamnose act directly as inducers for expression of regulons for their catabolism, important differences exist in regard to the regulatory mechanisms (Fig. 5). l-Arabinose acts as inducer with the activator AraC in the positive control of the arabinose regulon (23). However, the l-rhamnose regulon is subject to a regulatory cascade; it is therefore subject to an even tighter control. l-Rhamnose acts as an inducer with the activator RhaR for synthesis of RhaS, which in turn acts as an activator in the positive control of the rhamnose regulon (6). As shown in Table 7, both the P_rhaB-phoB⁺ and P_rhaB-phoR⁺ fusion strains show a PhoB⁻ and PhoR⁻ phenotype, respectively, in the absence of rhamnose. We also examined a P_rhaS-phoR⁺ fusion. Like the P_araB-phoR⁺ fusion strain, the P_rhaS-phoR⁺ fusion strain synthesized a substantial, though limited, amount of BAP in the absence of induction.

FIG. 5.

Transcriptional activation of the arabinose and rhamnose regulons. (A) Positive control of the l-arabinose regulon. The araCBAD region near 1.5 min on the E. coli chromosome is shown. The unlinked araE and araFGH genes encode arabinose transport systems (23). (B) Positive control of the l-rhamnose regulon. The rhaRSBAD region near 88.2 min is shown. The nearby rhaT gene is transcribed towards rhaR (Fig. 4) and encodes a rhamnose permease (3). Thick arrows show gene orientations. Curved ones with circled pluses indicate sites for transcriptional activation by AraC (A) or RhaR and RhaS (B). The lower bars show the ara and rha sequences in the respective P_araB and P_rhaB fusion allele replacement vectors illustrated in Fig. 2. The dashed lines show the chromosomal regions deleted in the resultant recombinants. See text.

The l-arabinose and l-rhamnose regulons are also regulated by catabolite repression. We therefore modulated expression of P_araB, P_rhaB, and P_rhaS fusions by using various carbon sources (glucose, mannitol, fructose, and glycerol) that lead to different levels of catabolite repression (7). When strains are grown on these carbon sources with the inducer in excess, the expression of the Pho regulon can be modulated 70- to 200-fold in our P_araB-phoB⁺, P_rhaB-phoB⁺, and P_rhaB-phoR⁺ fusion strains (Table 7). As expected, the differences are much smaller in the P_araB-phoR⁺ and P_rhaS-phoR⁺ fusion strains as the expression of these fusions is also leaky under these conditions. In preliminary experiments, we had also attempted to modulate expression levels by using different inducer concentrations. However, we were unable to maintain a constant, steady state level of induction (data not shown) (5). An inability to maintain steady-state induction at intermediate levels by the limiting inducer concentration is expected as the presence of arabinose and rhamnose results also in increased synthesis of their respective transport systems (Fig. 5). This has now been substantiated by monitoring P_araB expression levels in cells grown in the presence of subsaturating inducer concentrations (24). Importantly, as shown here, expression of these promoters can be modulated over a wide range by using different carbon sources in the presence of saturating inducer concentrations.

Both PhoU and the Pst transporter have been implicated in the negative control of the Pho regulon, as mutations of either result in high constitutive Pho regulon gene expression. Presumably, they somehow act together. To study how they interact with the PhoB-PhoR system, we made strains that synthesize PhoB or PhoR under control of the tightly regulated P_araB and P_rhaB promoters. By using these strains, we were able to express phoB or phoR at a reduced level and thereby overcome the harmful effect of a ΔphoU mutation. In these ways, it should now be possible to study further the function(s) of PhoU.