YycH and YycI Interact To Regulate the Essential YycFG Two-Component System in Bacillus subtilis

JOURNAL OF BACTERIOLOGY, Apr. 2007, p. 3280 3289 Vol. 189, No. 8 0021-9193/07/$08.00 0 doi:10.1128/jb.01936-06 Copyright 2007, American Society for Microbiology. All Rights Reserved. YycH and YycI Interact To Regulate the Essential YycFG Two-Component System in Bacillus subtilis Hendrik Szurmant, Michael A. Mohan, P. Michael Imus, and James A. Hoch* Division of Cellular Biology, Department of Molecular and Experimental Medicine, The Scripps Research Institute, 10550 N. Torrey Pines Road, La Jolla, California 92037 Received 21 December 2006/Accepted 7 February 2007 The YycFG two-component system is the only signal transduction system in Bacillus subtilis known to be essential for cell viability. This system is highly conserved in low-g C gram-positive bacteria, regulating important processes such as cell wall homeostasis, cell membrane integrity, and cell division. Four other genes, yychijk, are organized within the same operon with yycf and yycg in B. subtilis. Recently, it was shown that the product of one of these genes, the YycH protein, regulated the activity of this signal transduction system, whereas no function could be assigned to the other genes. Results presented here show that YycI and YycH proteins interact to control the activity of the YycG kinase. Strains carrying individual in-frame deletion of the yyci and yych coding sequences were constructed and showed identical phenotypes, namely a 10-fold-elevated expression of the YycF-dependent gene yoch, growth defects, as well as a cell wall defect. Cell wall and growth defects were a direct result of overregulation of the YycF regulon, since a strain overexpressing YycF showed phenotypes similar to those of yych and yyci deletion strains. Both YycI and YycH proteins are localized outside the cytoplasm and attached to the membrane by an N-terminal transmembrane sequence. Bacterial two-hybrid data showed that the YycH, YycI, and the kinase YycG form a ternary complex. The data suggest that YycH and YycI control the activity of YycG in the periplasm and that this control is crucial in regulating important cellular processes. Two-component systems (TCS) are commonly utilized for signal transduction in bacteria and archaea and are also found in the form of phosphorelays in lower eukaryotes and plants (14, 27). Bacillus subtilis expresses over 30 such TCS, only one of which has proven essential for cell viability, the YycFG system (8). Orthologs of YycFG are found in most low-g C gram-positive bacteria, including many important pathogens. The essentiality of this signal transduction system has been shown for most organisms in which it has been studied, including Staphylococcus aureus, Streptococcus pneumoniae, and Enterococcus faecalis (3, 13, 24). The two proteins have received several different names in different organisms (1, 7, 30). For clarity, we will refer to these proteins here only as YycF (response regulator) and YycG (histidine kinase). As the most highly conserved in amino acid sequence of the two-component proteins in this group of organisms with an average conservation of over 70% of YycF response regulator residues it might be logical to assume that the genes controlled by this system would be similar. Surprisingly however, there is significant variation in the gene regulons for YycF in different bacteria. In general, however, most genes of the regulon are involved in cell wall homeostasis, cell membrane integrity, and cell division processes (6, 10, 15, 28, 31). Two distinct subgroups of YycFG systems have recently been defined: one found in the streptococci (class II) and the * Corresponding author. Mailing address: Division of Cellular Biology, Mail Code MEM-116, Department of Molecular and Experimental Medicine, The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. Phone: (858) 784-7905. Fax: (858) 784-7966. E-mail: hoch@scripps.edu. Manuscript no. 18604 from The Scripps Research Institute. Published ahead of print on 16 February 2007. other in the remainder of the YycFG-expressing organisms (class I) (32, 37). Three main differences distinguish the two systems. First, in class I both kinase and response regulator are essential, whereas in class II, only the response regulator is essential (although the kinase becomes essential in cells depleted for YycF (30). Second, class I kinases have an extracellular sensing domain flanked by two transmembrane regions. Conversely, class II kinases have no significant extracellular sensing domain and generally only have one transmembrane domain. The exception is Lactococcus lactis YycG, which has two transmembrane domains connected via a short extracellular loop region, which is much smaller than the extracellular domains found in class I kinases (32). The third difference is found in the operon structure. In class I systems, three or four genes including yych and yyci are organized immediately downstream of the kinase within the same operon (8). These two genes are absent in class II-expressing bacteria (32). Both YycH and YycI have no homologs in the databases aside from orthologs in the yycfg operons of different organisms. Amino acid conservation of YycH and YycI proteins is quite low, particularly when compared to the high conservation found for YycF and YycG. In a recent genomewide transposon mutagenesis study, disruption of yych was identified to cause suppression of a temperature-sensitive yycf mutation (37). A yych disruption strain had an elevated YycF-dependent expression profile, as measured through a yoch-lacz reporter gene. YycF and YycG protein levels remained unchanged in the mutant strain compared to those in the wild type, suggesting that the activity of the YycFG system had been elevated. An extracellular localization of YycH was demonstrated by a PhoA fusion approach, and therefore the most likely explanation for the results was 3280

VOL. 189, 2007 REGULATION OF THE YycG SENSOR KINASE 3281 TABLE 1. Plasmids used in this study TABLE 2. Strains used in this study Plasmid Description Source or reference Strain Relevant genotype or description Source or reference 37 pjm134 Spc r gene in pbluescript (Stratagene); Amp r Spc r pcr4-topo Cloning vector; Amp r Kan r Invitrogen pjs07 pcr4-topo-yycf This work pjs08 pcr4-topo-yycf(d54h) This work pjs09 pcr4-topo- yycg-yych-yyci This work pjs10 pcr4-topo- yycg- yych-yyci This work pjs11 pcr4-topo- yych-yyci-yycj This work pjs12 pcr4-topo- yych- yyci-yycj This work pjs13 pcr4-topo- yyci-yycj-yyck This work pjs14 pcr4-topo- yyci- yycj-yyck This work pjs15 pcr4topo- yyci-spc r -yyck This work pht315s B. subtilis/e. coli shuttle vector; Amp r Ery r 39 pjs16 pht315s-yycf This work pjs17 pht315s-yycf(d54h) This work pjs18 pht315s-yych This work pjs19 pht315s-yych( 391 455) This work pjs20 pht315s-yyci This work pjs21 pht315s-yyci( 199 280) This work pdg1664 thrc integration vector; Amp r Spc r Ery r 12 pjs22 pdg1664-p spac -yych This work pjs23 pdg1664-p spac -yyci This work pma5 B. subtilis/e. coli shuttle vector; Amp r Kan r 4 pjv217 pma5-kapb(c20p) - phoa 5 pjs02 pma5-yych - phoa 37 pjs24 pma5-yyci - phoa This work pjm103 Suicide integration vector; Amp r Cm r 33 pjm103(i-scei) pjm103 with I-SceI restriction site M. Perego pjs25 pjm103(i-scei)- yycg- yych-yyci This work pjs26 pjm103(i-scei)- yych- yyci-yycj This work pjs27 pjm103(i-scei)- yyci- yycj-yyck This work pjs28 pjm103(i-scei)- yycg- yychi-yycj This work pjm117 pmutin4 derivative; Amp r Cm r M. Perego pbkj223 I-SceI expression vector; Amp r Tet r 19 put18-c T18 expressing two-hybrid vector; Amp r Hybrigenics pjs29 put18-c-yycg This work pjs30 put18-c-yycg(1 210) This work pjs31 put18-c-yych This work pjs32 put18-c-yyci This work pkt25 T25 expressing two-hybrid vector; Kan r Hybrigenics pjs33 pkt25-yycg This work pjs34 pkt25-yycg(1 210) This work pjs35 pkt25-yych This work pjs36 pkt25-yyci This work that YycH inhibits the autokinase activity of YycG by either direct interaction with the extracellular sensing domain or by production of a signaling molecule (37). This idea is consistent with the observation that in class II YycFG-expressing bacteria, the kinase is missing an extracellular sensing domain and the yych gene is also absent. We previously reported that disruption of any of the yych distal genes (yyci, yycj, and yyck) had no effect on YycFdependent expression (37). A yycj homolog is found in every YycFG operon (class I and class II) and is homologous to proteins with a -lactamase fold (38). YycK is a protease and is not always present, even in class I organisms. For instance, the yyc operon in Enterococcus faecalis does not contain a yyck gene. The yyci gene, on the other hand, always appears in a pair with yych. Furthermore, the amino acid sequences of YycH and YycI suggest a similar topology, namely an N-terminal transmembrane helix, suggesting that both proteins are localized on the periplasmic face of the cellular membrane. Here we show that deletion of either yych or yyci results in identical phenotypes, which are similar to those already observed for the original yych disruption strain. We present evidence that YycH, YycI, and YycG form a ternary complex as determined via a bacterial two-hybrid system. Taken together, E. coli TG1 General cloning strain Lab stock BTH101 Adenylate cyclase-deficient strain for two-hybrid assay Hybrigenics B. subtilis JH642 phea1 trpc2 Lab stock JH25001 amye::(p yoch -lacz aph3-a) 37 JH25002 yych::pjm103 amye::(p yoch -lacz aph3-a) 37 JH25003 yyci::pjm103 amye::(p yoch -lacz aph3-a) 37 JH25011 yych::pjm117 amye::(p yoch -lacz aph3-a) 37 JH25020 amye::(p yoch -lacz aph3-a) thrc::(ery r ) This work JH25021 yych1 amye::(p yoch -lacz aph3-a) This work JH25022 yyci1 amye::(p yoch -lacz aph3-a) This work JH25023 yych1 amye::(p yoch -lacz aph3-a) This work thrc::(ery r ) JH25024 yych1 amye::(p yoch -lacz aph3-a) This work thrc::(p spac -yych Ery r ) JH25025 yych1 amye::(p yoch -lacz aph3-a) This work thrc::(p spac -yyci Ery r ) JH25026 yyci1 amye::(p yoch -lacz aph3-a) This work thrc::(ery r ) JH25027 yyci1 amye::(p yoch -lacz aph3-a) This work thrc::(p spac -yych Ery r ) JH25028 yyci1 amye::(p yoch -lacz aph3-a) This work thrc::(p spac -yyci Ery r ) JH25029 yycj1 amye::(p yoch -lacz aph3-a) This work JH25030 yycj1::spc r amye::(p yoch -lacz aph3-a) This work JH25031 yychi101 amye::(p yoch -lacz aph3-a) This work MH3402 phea1 trpc2 phoa::cat phob::mtn10 18 our data strongly suggest that YycI and YycH function together to control the activity of the essential histidine kinase YycG. MATERIALS AND METHODS Growth media and conditions. All strains were grown in Luria-Bertani (LB) broth unless otherwise indicated. Medium was supplemented with appropriate antibiotics whenever necessary. The antibiotic concentrations for Escherichia coli strains were 100 g/ml ampicillin, 100 g/ml spectinomycin, or 30 g/ml kanamycin, and for B. subtilis, they were 5 g/ml kanamycin, 0.5 g/ml erythromycin, 12.5 g/ml lincomycin, 5 g/ml chloramphenicol, 10 g/ml tetracycline, and 100 g/ml spectinomycin. Strain and plasmid construction. All plasmids were cloned in E. coli strain TG1. All B. subtilis strains were derived from strain JH642. Plasmids and strains are listed in Table 1 and Table 2, respectively. Oligonucleotide sequences and names are listed in Table 3. To clone yycf, the gene was amplified from chromosomal DNA by PCR with oligonucleotide primers ON1 and ON2 and ligated with the vector pcr4-topo (Invitrogen) according to the manufacturer s recommendations, yielding plasmid pjs07. A D54H mutation was introduced in the yycf gene by subjecting vector pjs07 to a PCR with inverted mutagenic oligonucleotide primers ON3 and ON4 and also introducing a silent unique BglII restriction site. The PCR product was digested with BglII and ligated, and template DNA was removed by DpnI digestion. The resulting clones were sequenced to confirm the correct mutation, yielding plasmid pjs08. To create multicopy overexpression vectors, the yycf and yycf(d54h) genes were subcloned into KpnI and HindIII sites of the shuttle vector pht315s, placing the genes under control of the P spac promoter and yielding plasmids pjs16 and pjs17, respectively. Similarly, yych (oligonucleotide primers ON05 and ON06), yych( 391 455) (ON05 and ON07), yyci (ON08 and ON09), and yyci( 200 280) (ON08 and ON10) were PCR amplified and cloned into pht315s using SmaI and BamHI sites for yych, SmaI and PstI sites for yych( 391 455), SacI and BamHI sites for yyci, and SacI and XbaI sites for yyci( 200-280) and yielding plasmids pjs18 to pjs21, respectively. Subcloning the genes, including the P spac promoter, from vectors pjs18 and pjs20 into the thrc integration vector pdg1664 using restriction enzymes EcoRI

3282 SZURMANT ET AL. J. BACTERIOL. TABLE 3. Oligonucleotide primers used in this study Name Restriction site Sequence a Plasmid constructed ON01 KpnI 5 -TACATGGTACCTAAGTGATGAAACGAATG-3 pjs07 ON02 HindIII 5 -CAACCTTAAGCTTATGGGCATTAGTCCTG-3 pjs07 ON03 BglII 5 -TCCAAAGATCTAATTCTTTTACATATTATGCTCCC-3 pjs08 ON04 BglII 5 -TCCAAAGATCTGGCTGAAGCTCTTCTACCATTTC-3 pjs08 ON05 SmaI 5 -CAGGCGCCCGGGGGAGATATTTG-3 pjs18/pjs19 ON06 BamHI 5 -GAGAGGATCCGTCTCACTCCATTATTTTCATTGATC-3 pjs18 ON07 PstI 5 -TTGTTTCTGCAGCTACATCTTCACTTCGCTG-3 pjs19 ON08 SacI 5 -GAGAGAGCTCGAAGGAGGGGGCTAACAGT-3 pjs20/pjs21 ON09 BamHI 5 -GAGAGGATCCGTCTCACTCCATTATTTTCATTGATC-3 pjs20 ON10 XbaI 5 -TTAAATCTAGATCATTCTTTCTGGATTTGCTTGAAGGTC-3 pjs21 ON11 EcoRV 5 -CAACATCAACCAACGATATCCTCTCGTTG-3 pjs24 ON12 EcoRV 5 -CGTCCTCGATATCCTTCTTTGATAGTGCG-3 pjs24 ON13 I-SceI 5 -GATCTAGGGATAACAGGGTAAT-3 pjm103(i-scei) ON14 I-SceI 5 -GATCATTACCCTGTTATCCCTA-3 pjm103(i-scei) ON15 SalI 5 -CCGGGAGTCGACTCAGATTGTCCG-3 pjs09 ON16 SmaI 5 -CATATTCCCGGGATTGTCGGTCTTTTGG-3 pjs09 ON17 MscI 5 -GGAAAAATTTGGCCAATTACAAAGGATTTATTG-3 pjs10 ON18 MscI 5 -CGTCCATATTGGCCATGTAAAAACAAGACTG-3 pjs10 ON19 PstI 5 -CAATTTGCTGCAGAGCACATCCTATCAAAC-3 pjs11 ON20 SalI 5 -CCGGTCGTCGACATATCCCGTATCTG-3 pjs11 ON21 EcoRV 5 -GCTTTGGATATCACGATTTTAGATACAGATCAATG-3 pjs12 ON22 EcoRV 5 -GCCTAAGAAGATATCTAAAATGAGGAAGGCAACG-3 pjs12 ON23 PstI 5 -GAGGACGCTGCAGAGGCACTTG-3 pjs13 ON24 SalI 5 -GTCGCTGTCGACAAGATCAGCTGATAC-3 pjs13 ON25 EcoRV 5 -CTTATACGATATCGATCCGAAGAAGGCC-3 pjs14 ON26 EcoRV 5 -CTCCCGATATCAAGTACGCTAAATTGCAAG-3 pjs14 ON27 XbaI 5 -ACTAATGCTCTAGAGAATAAGGTTGGTTTTTTTCGG-3 pjs29/pjs33 ON28 BamHI 5 -GTTTTTGGATCCTCACGCTTCATCCCAATC-3 pjs29/pjs33 ON29 XbaI 5 -ATGATTCTAGAGAAGCGTGAAAATATAAAAACG-3 pjs31/pjs35 ON30 BamHI 5 -GAGAGGATCCCTTATTCCACTCCACTGT-3 pjs31/pjs35 ON31 XbaI 5 -GGGCTAACTCTAGAGTGGAATAAGACAAAATC-3 pjs32/pjs36 ON32 BamHI 5 -GAGAGGATCCGTCTCACTCCATTATTTTCATTGATC-3 pjs32/pjs36 ON33 None 5 -TAAAGAGGAACAAGAGGATGATTGGGATG-3 pjs30/pjs34 ON34 None 5 -AAGCGGGTGGGTAATGGTTCTTG-3 pjs30/pjs34 a Introduced restriction sites are in boldface, and mutated amino acid codons are underlined. and BamHI, generating plasmids pjs22 and pjs23, respectively, created plasmids suitable for single-copy expression of yych and yyci in B. subtilis. Bacterial two-hybrid vectors were created for YycG, YycH, and YycI. The respective coding sequences were PCR amplified, introducing 5 XbaI and 3 BamHI sites utilizing oligonucleotide primers ON27 and ON28 (yycg), ON29 and ON30 (yych), or ON31 and ON32 (yyci). The resulting DNA fragments were cloned into the same sites of vector put18-c, resulting in pjs29, pjs31, and pjs32, and into vector pkt25, resulting in pjs33, pjs35, and pjs36, respectively. pjs29 and pjs33 were subjected to PCR with inverted primers ON33 and ON34, excluding the part of yycg that codes for the cytoplasmic region. The PCR fragments were blunt-end ligated, and template DNA was removed by DpnI digest, resulting in vectors pjs30 and pjs34. All constructs were verified by DNA sequencing. All of these constructs expressed the respective genes or gene fragments fused C terminally to the Bordetella pertussis adenylate cyclase gene fragment T18 or T25 (23). Construction of markerless in-frame deletion strains. Markerless in-frame yych, yyci, and yycj deletion strains were constructed adopting the Bacillus anthracis method of Janes and Stibitz with slight modifications (19). This method utilizes the Saccharomyces cerevisiae endonuclease I-SceI, which has an 18-bp recognition sequence (29). Briefly, in a first step, an I-SceI restriction site is introduced in the chromosome along with the desired in-frame deletion by single-crossover recombination of a suicide vector. In a second step, a plasmid expressing the human endonuclease I-SceI is introduced, resulting in a doublestranded break in the DNA, which is repaired by homologous recombination, resulting in either a markerless deletion or wild-type strain (Fig. 1). The suicide vector pjm103 was modified to include an I-SceI site distant from the multiple cloning site (M. Perego, unpublished). Two complementary oligonucleotide primers, ON13 and ON14, with BglII-compatible 5 overhangs were designed and annealed. The vector pjm103 was digested with the restriction enzyme BglII. The annealed oligonucleotides were ligated with BglII-digested pjm103, yielding pjm103-i-scei. The final vector was confirmed by I-SceI digestion as well as by sequencing. In three separate PCRs, the yych, yyci, and yycj genes including about 500 bp of upstream and downstream regions were amplified using oligonucleotide primers ON15 and ON16 for yych, ON19 and ON20 for yyci, and ON23 and ON24 for yycj. The PCR fragments were ligated in the vector pcr4-topo (Invitrogen), yielding plasmids pjs09, pjs11, and pjs13, respectively. The yych, yyci, and yycj genes were deleted in frame from these vectors by PCR amplification of the vectors and flanking regions, excluding genes subjected for deletion using ON17 and ON18 for yych, ON21 and ON22 for yyci, and ON25 and ON26 for yycj. The PCR fragments were digested with MscI (yych)or EcoRV (yyci and yycj) and ligated to yield plasmids pjs10, pjs12, and pjs14. The deletion constructs were subcloned into pjm103-i-scei using restriction enzymes SalI and SmaI for yych and PstI and SalI for yyci or yycj and yielding vectors pjs25, pjs26, and pjs27, respectively. A yychi double-deletion plasmid was constructed by exchanging the yych downstream region in pjs25 with the yyci downstream region from vector pjs26, utilizing SphI and MscI sites for pjs25 and EcoRV and SphI sites for pjs26, resulting in plasmid pjs28. Vectors pjs25, pjs26, pjs27, and pjs28 were transformed into B. subtilis strain JH25001, selecting for Cm r. The resulting strains were transformed with the I-SceIexpressing plasmid pbkj223, selecting for Tet r. One colony was streaked twice over LB broth containing tetracycline but lacking chloramphenicol. Single colonies were screened for a Cm s phenotype indicating that a DNA double-stranded break followed by a homologous recombination event had occurred. Cm s colonies were subjected to diagnostic PCR to identify a correct homologous recombination event resulting in deletion of yych, yyci, yycj, or yychi in the respective strains. Roughly 50% of all colonies contained the correct deletion, whereas the rest were wild type, as expected for this technique. Finally, strains were cured for the pbkj223 plasmid by serial transfer in LB broth in the absence of tetracycline, followed by screening for a Tet s phenotype. This procedure yielded the yych

VOL. 189, 2007 REGULATION OF THE YycG SENSOR KINASE 3283 FIG. 1. Construction of markerless yych, yyci, yycj, and yychi deletion strains. Deletion strains JH25021 ( yych), JH25022 ( yyci), JH25029 ( yycj), and JH2531 ( yychi) were constructed by adopting the Bacillus anthracis method of Janes and Stibitz to Bacillus subtilis (19). This method involves single-crossover integration of a suicide plasmid featuring chromosomal regions flanking the gene to be deleted ( 500 bp each site) and an I-SceI site. The Cm r pjm103 integrative plasmid was modified by introducing an I-SceI site (M. Perego, unpublished). Following singlecrossover integration (shown here for yych) and selection for Cm r, the I-SceI gene was expressed from a second plasmid pbkj223 conferring Tet r. This introduced a DNA double-stranded break, which was repaired by homologous crossover recombination resulting in Cm s strains. Roughly 50% of all Cm s transformants contained the correct deletion, whereas the other 50% are wild type. Deletion strains were identified by colony PCR. The pbkj223 plasmid was lost by repeated growth without antibiotic selection in liquid media. strain JH25021, the yyci strain JH25022, the yycj strain JH25029, and the yychi double-deletion strain JH25031. To complement the yych and yyci strains, either pdg1664 or its derivatives pjs22 and pjs23 were linearized by AatII digestion and transformed into strains JH25021 and JH25022 selecting for Ery r /Linc r. Correct double-crossover integration into the thrc locus was confirmed by screening for threonine auxotrophy and an Spc S phenotype. This procedure yielded strains JH25023 to JH25028. In addition, the wild-type strain JH25001 was transformed with empty pdg1664 vector to generate a control strain with an isogenic background, JH25020. A complementing yycj expression vector could not be generated, probably due to toxicity of the gene product. Therefore a second yycj deletion strain was constructed which replaced the yycj gene with an Spc r marker. This strain was generated by excising the Spc r gene from the plasmid pjm134 via an EcoRV digest and subcloning this DNA fragment into the EcoRV site of the vector with yycj deleted, pjs14. This vector was linearized with SphI to prevent singlecrossover recombination and transformed into B. subtilis strain JH25001, selecting for Spc r. Colonies were PCR screened for correct double-crossover insertion of the Spc r marker, replacing the yycj gene and resulting in strain JH25030. Unlike the in-frame deletion constructs, this deletion is transferable; utilizing the Spc r gene, however, might have a polar effect on the expression of the downstream yyck protease gene. -Galactosidase assay. To analyze time-dependent lacz expression profiles, cultures were grown in LB broth and -galactosidase activity was determined as described previously (9). Bacterial two-hybrid assay. The bacterial two-hybrid assay was performed essentially as described previously (21, 22). To probe for interaction between YycG, YycH, and YycI, all combinations of put18-c-derived plasmids and pkt25-derived plasmids were cotransformed into the adenylate cyclase-deficient E. coli strain BTH101. Transformants were plated on LB broth containing 30 g/ml kanamycin, 100 g/ml ampicillin, and 40 g/ml 5-bromo-4-chloro-3-indolyl- -D-galactopyranoside. Individual colonies were scored for a Lac (blue colonies) or Lac (white colonies) phenotype. A Lac phenotype indicates complementation of adenylate cyclase activity and therefore interaction between the proteins under investigation. To quantify the strength of interaction, -galactosidase specific activity of overnight cultures derived from individual colonies was determined and calculated according to Miller (26). Analysis of cell wall defects. Cell wall defects as evidenced by enhanced susceptibility to sodium dodecyl sulfate (SDS)-induced lysis were detected as previously described (37). Briefly, strains were grown at 37 C until early stationary phase. Cells were collected by centrifugation and washed and diluted to an A 525 of 1.0. Suspensions were treated with SDS following 30 min of incubation with or without lysozyme. The suspensions or solutions were subjected to SDSpolyacrylamide gel electrophoresis followed by Coomassie staining. Detectable levels of cytoplasmic proteins for samples not treated with lysozyme were interpreted as a direct result of a cell wall defect. Immunoblot analysis. YycF and YycG were detected immunologically as described previously (37). RESULTS Deletion of either yych or yyci causes overexpression of the YycFG-dependent gene yoch. We previously reported that disruption of yych but not yyci caused a 10-fold elevation of transcription of the YycF-dependent gene yoch (37). This interpretation was based on strains that had yych and yyci disrupted by integrative vector single-crossover recombination. A similar yych disruption strain utilizing a pmutin derivative, designed to prevent any polarity effects on expression of downstream genes, gave identical results (37). The recent development of a method to generate markerless deletions in Bacillus anthracis (19) allowed us to construct markerless plasmidless in-frame deletions of yych, yyci, and yycj in B. subtilis and retest their effects. The background strain for these deletions carried the YycFdependent yoch-lacz reporter gene in single copy at the amye locus. YycF-dependent yoch expression levels were determined in a liquid assay measuring -galactosidase activity. The markerless deletions of yych and yyci were designed to delete all but the first 15 amino acids and the last 10 amino acids to ensure a ribosome binding site for the deletion distal gene. Both yych and yyci deletion strains showed identical expression profiles, namely up to 10-fold overexpression of yoch. Furthermore, both strains entered stationary phase early, reaching only about 50% of the cell density of the wild-type strain (Fig. 2). These results are identical to those obtained for the original yych disruption strain, JH25002, but not the original yyci disruption strain, JH25003, which did not show any change in YycF-dependent expression (37). Further experiments were undertaken to resolve the differences in results using different methods. yyci and yych deletions cannot be cross-complemented. The first experiments were designed to determine if both YycH and

3284 SZURMANT ET AL. J. BACTERIOL. FIG. 2. Deletion of either yych or yyci causes elevated YycF-dependent expression. Growth curves measured in optical density (OD) at a wavelength of 525 nm (A) and -galactosidase specific activity in Miller units (B) were determined for strains harboring the YycFdependent reporter gene yoch-lacz in the amye locus. A time of 0 h was defined as the onset of stationary phase. Strains were either wild type (JH25001 [diamonds]) or yych::pjm103 (strain JH25002 [squares]), yych (strain JH25021 [triangles]), or yyci (JH25022 [circles]) mutants. YycI proteins were necessary for normal activity of the YycFG two-component system. In order to investigate the reason for the identical phenotypes observed in the yych and yyci deletion strains, we carried out a complementation assay. The yych and yyci genes were cloned in the thrc integration vector pdg1664 under control of the constitutive P spac promoter. The respective constructs were transformed in the yych and yyci deletion strains. Expression of yych but not yyci in single copy from the thrc locus in the yych deletion strain resulted in full complementation of the growth defect, and YycF-dependent expression was restored to the wild-type level (Fig. 3A and B). Similarly, expression of yyci but not yych from the thrc locus in the yyci deletion strain resulted in full complementation of both phenotypes (Fig. 3C and D). The yych and yyci genes were also expressed from the multicopy plasmid pht315 under control of the identical constitutive P spac promoter. Again, deletion strains could be fully complemented with their homologous gene, whereas no cross-complementation was observed (not shown). These results demonstrate that both YycH and YycI are involved in regulating the activity of the YycFG system. Since the yoch expression patterns were identical in both deletion strains, the results suggested that YycH and YycI either constitute a pathway or act independently but equally to regulate YycF-dependent expression. Both proteins are essential for this function since overexpression of yych does not enable cross-complementation of the yyci deletion strain and overexpression of yyci does not cross-complement the yych deletion strain. In order to determine whether YycH and YycI functioned together on the same pathway or independently of each other, we constructed a yychi double-deletion strain, JH25031. We argued that if the yych and yyci deletion phenotypes were additive, the two gene products could be functioning independently, whereas if the phenotype for the strain with both genes deleted mimicked the phenotype seen for strains with individual deletions, the two proteins likely constitute a pathway or function as a protein complex. Indeed, the double-deletion strain JH25031 showed a growth defect and yoch expression FIG. 3. Complementation studies in yych and yyci deletion strains. To determine whether the yych and yyci strains would be complemented by expressing the respective genes from elsewhere in the chromosome, strains containing either yych or yyci or empty pdg1664 vector integrated in the thrc locus were grown in LB broth. Shown are (A) growth curves and (B) -galactosidase specific activity in Miller units of JH25020, a wild-type strain harboring pdg1664 vector in the thrc locus (diamonds), and yych strains JH25023 through JH25025 harboring either pdg1664 (squares), pdg1664-yych (triangles), or pdg1664-yyci (circles) in the thrc locus, respectively. Also shown are (C) growth curves and (D) -galactosidase specific activity in Miller units of wild-type strain JH25020 harboring pdg1664 vector in the thrc locus (diamonds) or yyci strains JH25026 through JH25028 harboring either pdg1664 (squares), pdg1664-yych (circles), or pdg1664-yyci (triangles) in the thrc locus, respectively. pattern identical to those of the individual deletion strains (data not shown). We concluded that YycH and YycI are likely to be acting together to regulate the YycFG system. yych and yyci deletions cause a cell wall defect. The original yych disruption strain showed susceptibility to SDS-induced lysis compared to a wild-type strain, which is resistant to SDSinduced lysis and only lyses in the presence of SDS following a lysozyme treatment (37). The yych disruption strain phenotype was interpreted as a cell wall defect. To investigate whether the yych and yyci deletion strains had a similar cell wall defect, they were subjected to SDS treatment in the presence or absence of lysozyme (Fig. 4). As expected, both strains lysed, whether or not they were treated with lysozyme. Expression of yych from the thrc locus complemented the phenotype of the yych strain; similarly, expression of yyci complemented the yyci strain (Fig. 4). No cross-complementation could be observed (data not shown). Therefore, deletion of either yyci or yych causes a cell wall defect. To investigate the nature of the cell wall defect further, we subjected the wild-type strains and the yych and yyci deletion strains to electron microscopy analysis. If the nature of the cell wall defect was a thinner cell wall, it would be apparent in an electron micrograph, whereas if no effect was seen, the cross-

VOL. 189, 2007 REGULATION OF THE YycG SENSOR KINASE 3285 FIG. 4. A cell wall defect is apparent for all strains that overexpress the YycF-regulon. Strains were grown to mid-stationary phase and subjected to SDS treatment. Whole-cell protein extracts were separated by SDS-polyacrylamide gel electrophoresis and visualized by Coomassie staining. Extracts were from wild-type strain JH25001 (lane 1), yych disruption strain JH25002 (lane 2), yych strain JH25021 (lane 3), yyci strain JH25022 (lane 4), yych strain JH25024 harboring yych (lane 5), yyci strain JH25028 harboring yyci (lane 6) under P spac control in the thrc locus, and wild-type strain JH25001 harboring either multicopy expression vector pht315s (lane 7) or the yycf overexpression plasmid pht315s-yycf (lane 8). Wild-type cells did not lyse under the treatment conditions, as evidenced by the absence of detectable levels of cytoplasmic protein. linking rate or the chemical composition of the cell wall must have changed. Electron micrographic analysis of individual cells did not reveal a thinner cell wall for the yych and yyci deletion strains compared to the wild-type strain, suggesting that the latter hypothesis is likely the case (data not shown). Overexpression of yycf and deletion of yych or yyci produce similar phenotypes. We designed experiments to examine whether deletion of yych and yyci caused growth and cell wall defects (i) because of overexpression of the YycF regulon or (ii) because YycH and YycI were directly involved in these phenotypes. Multicopy plasmids for both yycf and a previously reported constitutively active yycf(d54h) phosphorylation point mutant (10) were constructed. The strains were subjected to an immunoblot analysis to ensure that YycF was indeed overexpressed (Fig. 5A). Overexpression of wild-type YycF caused a yoch expression profile similar to that seen for yych and yyci deletion mutants (Fig. 5). Additionally, both cell wall (Fig. 4) and growth (Fig. 5) defects were apparent in the overexpression strains. Therefore, all observed phenotypes in the deletion strains are likely caused by overexpression of the YycF regulon, and therefore the main function for YycH and YycI would seem to be to regulate the activity of this essential TCS. Despite a similar yoch expression profile, the growth defect for the wild-type YycF overexpression strain was more severe than that for the yych strain, indicating that this overexpression does not merely mirror enhanced phosphorylation of YycF but has other effects. Interestingly, overexpression of the previously reported YycF(D54H) mutant protein affected growth in LB medium even more severely than overexpression of wild-type yycf, yet YycF-dependent expression was only marginally elevated (not shown). This suggests that this mutant not only mimics phosphorylation of YycF but also interferes with other, unknown, processes in the cell. Deletion of yych or yyci has no effect on YycFG protein levels. YycF and YycG protein levels in the original yych disruption strain JH25002 remained unchanged compared to those of a wild-type strain, indicating that the activity rather than the expression level of the YycFG two-component system FIG. 5. Overexpression of yycf in B. subtilis has effects similar to deletion of yych. (A) Wild-type strain JH25001 harboring either pht315s or pht315s-yycf was grown in LB broth, and expression of yycf was immunologically detected at the indicated times to confirm yycf overexpression in the appropriate strain. Shown are (B) growth curves and (C) -galactosidase specific activity in Miller units of wildtype strain JH25001 (diamonds) or yych strain JH25021 (squares) and strains harboring multicopy plasmid pht315s or wild-type strain JH25001 harboring the yycf overexpression vector pht315s-yycf (triangles). must have been altered (37). To verify that the yych and yyci deletion strains generated in this study also had unaltered YycF and YycG protein levels, these proteins were visualized immunologically at different times throughout growth. As expected, protein levels remained unchanged, similarly to the original disruption strain, and this confirmed that YycFG activity rather than expression was altered (data not shown). YycG, YycH, and YycI interact in a two-hybrid assay. Given the fact that YycF and YycG protein levels did not change in the yych and yyci deletion strains, phenotypes of these strains were likely due to overactivity of the YycFG two-component signaling system. YycH was demonstrated in a phoa fusion approach to face the periplasmic space. YycI had a similar topology and also faced the periplasmic space addressed in the same way (not shown). Furthermore, immunodetection utilizing anti-yych or anti-yyci antibodies detected full-length proteins in the membrane fraction. YycH or YycI proteins were not found in the medium fraction or the soluble fraction (data not shown). Therefore, both proteins are periplasmic proteins tethered to the membrane via a single N-terminal transmembrane helix. Since YycH and YycI are physically separated from the cytoplasm, and hence the YycF response regulator, we argue that the most likely target for YycH and YycI is the periplasmic sensing domain of the YycG kinase and/or the transmembrane helices. Generation of a signaling molecule by YycH and YycI or a direct interaction with the kinase YycG could achieve this regulation. To investigate whether YycH and YycI are capable of interacting with YycG and with each other, we used a bacterial two-hybrid system developed by Ladant and colleagues (21). This system has been successfully applied to the study of interacting transmembrane proteins (20). It relies on the B. pertussis adenylate cyclase enzyme that is inactive when its two

3286 SZURMANT ET AL. J. BACTERIOL. FIG. 6. A bacterial two-hybrid assay reveals interactions between YycG, YycH, and YycI. (A) -Galactosidase activity was determined for adenylate cyclase-deficient E. coli strain BTH101 harboring plasmids expressing B. pertussis adenylate cyclase fragment T18 or T25 either unfused or fused to yycg, yych, or yyci in all possible combinations as indicated. -Galactosidase activity above background levels indicates an interaction between the coexpressed hybrid constructs. Error bars indicate standard deviation derived from three independently grown cultures. (B) Similarly, interaction between a truncated YycG(1 210) fragment (white columns) corresponding to the periplasmic sensing domain and the two transmembrane helices and either itself, YycH, or YycI was compared to interaction achieved for full-length YycG kinase (gray columns) utilizing T18 and T25 fusions. Interaction of hybrid constructs was quantified by measuring -galactosidase activity. Error bars indicate standard deviation derived from three independently grown cultures. domains, T18 and T25, are expressed individually in E. coli. However, when interacting proteins are fused to the individual adenylate cyclase domains, activity is reconstituted. False positives are rare in this assay and can be eliminated through suitable controls (22). Adenylate cyclase activity was measured indirectly in an adenylate cyclase deletion E. coli strain, BTH101, by measuring -galactosidase expressed from the lacz gene under control of the catabolite activator protein CAP, which is activated by cyclic AMP, the product of adenylate cyclase reaction (for review, see reference 11). In the first experiments, full-length yycg, yych, and yyci gene constructs were produced in both T18 and T25 plasmids and coexpressed in E. coli BTH101 (Fig. 6A). Coexpression of T18-yycG, T18-yycH, or T18-yycI constructs with unfused T25 did not result in activity above T18/T25 coexpression background levels. Similarly, coexpression of unfused T18 with T25-yycG, T25-yycH, or T25-yycI did not result in -galactosidase activity above the control background level. Coexpression of all fusion constructs did, however, result in -galactosidase activity at least sixfold over background (Fig. 6A). Remarkably, both possible combinations for each hetero-interaction rendered very similar activities. We found that all three proteins were capable of forming homo-dimers with a strong interaction found for histidine kinase YycG protomers as expected. A strong response was found for the interaction between YycH and YycI in both possible hybrid construct combinations, suggesting that YycH and YycI form a tight complex. Most importantly, both proteins individually were also capable of interacting with the YycG kinase. The data suggest that YycG, YycH, and YycI form a ternary complex. Therefore, a likely function of YycH and YycI is that of a signal sensor capable of modulating YycG activity through direct interaction. The periplasmic domain of YycG is sufficient for YycH and YycI interaction. Our data thus far support the notion that YycH and YycI are facing the periplasmic space, and therefore interaction with the YycG kinase would have to be stabilized either through interaction of the periplasmic domains of these proteins with the periplasmic sensing domain of YycG or through transmembrane interaction. If this is true, we would expect a YycG kinase construct, truncated for its C-terminal cytoplasmic domain, to still be able to interact with YycH and YycI in the bacterial two-hybrid assay. To test this possibility, we constructed both T18-YycG and T25-YycG truncation fusions (coding for amino acids 1 to 210). Coexpression of these constructs with the appropriate unfused fragments resulted in -galactosidase activity similar to background. Coexpression of the two truncated YycG fusion constructs resulted in dimerization of similar strength to full-length YycG, suggesting that a stable homodimer was still formed. Interaction with both YycH and YycI was even stronger than with full-length YycG (Fig. 6B). This could be either due to a change in conformation of the truncated YycG protein or due to reduced toxicity of this construct in E. coli. We conclude that the N-terminal sensing and transmembrane domains of YycG are sufficient for interaction with YycH and YycI, indirectly confirming the periplasmic localization of the proteins under investigation. These interpretations are consistent with PhoA fusions to YycH and YycI indicating a periplasmic location (37; data not shown). Deletion of yycj results in altered colony morphology and occurrence of sporulation-deficient colonies at a high frequency. An in-frame deletion mutant of yycj, coding for a protein in the -lactamase fold family and ubiquitously found in all Yyc operons, did not have any effect on YycF-dependent yoch-lacz expression, consistent with a previously constructed yycj disruption strain (37). We did however observe two peculiar phenotypes for the yycj deletion strain JH25029. First, the colony morphology for the yycj strain changed significantly compared to the wild type. Whereas a wild-type B. subtilis colony shows a relatively round shape when streaked on an LB or SM plate, yycj colonies consistently showed a riffled edge and no clear shape (Fig. 7A and B). Furthermore, when streaked on a Schaeffer sporulation plate (SM), about 2% of colonies showed a sporulation-deficient phenotype, as evidenced by transparent colonies and microscopic analysis of cells derived from these colonies (Fig. 7C and D). Since the yycj deletion strain was created on LB medium, where colony phenotype is not easily discernible, it was not clear whether Spo or Spo colonies constituted the original phenotype. To investigate this, a second yycj deletion strain, JH25030, was constructed in which the yycj gene was replaced with an Spc r marker. It should be noted that this strain could potentially have a polarity effect on the expression of the down-

VOL. 189, 2007 REGULATION OF THE YycG SENSOR KINASE 3287 FIG. 7. Colony phenotypes of the yycj strain. Shown are pictures of a representative colony of (A) wild-type strain JH25001 as well as three different colonies from the yycj strain JH25030. (B)The predominant form of yycj colonies is able to sporulate and is opaque but erose on SM medium. Two different forms of sporulation-deficient yycj colonies evidenced by a more transparent colony appearance were also observed: (C) one that heavily segregates Spo colonies and (D) one that shows a stable Spo phenotype. stream yyck protease gene. Spc r colonies obtained upon transformation of the wild-type strain were plated on SM agar. The large majority of colonies showed a Spo phenotype, suggesting that the initial yycj-inactive strain is capable of sporulation. However, Spo colonies were observed at a rate of roughly 2%. Furthermore, at least two different sporulation-defective colonies were observed: one that immediately segregated Spo colonies and one that appeared stable as Spo (Fig. 7). To identify whether the stable Spo phenotype was linked to the Spc r marker and therefore in close proximity on the genome to yycj, chromosomal DNAs of both a Spo colony and a Spo colony were backcrossed into a wild-type strain and plated on SM medium. Essentially all transformants from both were Spo. Spo colonies were observed at about 2%, suggesting that the Spo mutation is not linked to the yyc operon. We conclude that deletion of yycj has no effect on YycFG activity; however, the deletion does have an effect on colony morphology and induces an additional mutation resulting in Spo colonies. DISCUSSION The present evidence indicates that YycH and YycI are two membrane-bound periplasmically located proteins that form a ternary complex with the essential YycG histidine kinase. Deletion of either yych or yyci caused greatly elevated expression of the YycFG-dependent gene yoch, while protein levels of YycG and YycF remained unchanged. Phenotypes of the deletion strains namely an increased susceptibility to SDS-induced lysis as well as a growth defect matched those observed for a yycf overexpression strain and are consistent with a role for YycFG in regulating cell wall metabolism and cell divisional processes. The matching phenotypes also indicate that the entire YycF regulon is deregulated in the yych and yyci deletion strains. Given the interaction with the YycG kinase and the physical separation of YycH and YycI from YycF by the cell membrane, the simplest conclusion for our results is that YycH and YycI regulate the activity of the YycG kinase, as previously suggested for YycH only (37). At this point, we can only speculate as to the precise biochemical function of the YycH and YycI proteins. Crystal structures presented in the accompanying article demonstrate that these two proteins arose from gene duplication, as evidenced by a shared unique protein fold (34). A current absence of similar structures in the databases implies an unusual function. The inability to identify an active site and the low conservation of these proteins among different organisms suggest a role in protein interaction rather than an enzymatic activity. The bacterial two-hybrid system revealed that YycH and YycI interact strongly with each other and also interact with the YycG kinase. These results suggest that these proteins may exist as a ternary complex held together by interaction of either their periplasmic domains, their transmembrane helices, or both. Since all phenotypes associated with deletion of yych and yyci can be explained by an overactive YycFG system, it appears that a sensing mechanism and regulation of the YycG kinase activity may be the main function for YycH and YycI. Essential operons in B. subtilis such as the ftsaz and tagd/ taga operons have been demonstrated to contain a YycF binding site in their promoter region (15, 16). Other genes regulated by YycF such as yoch and ykvt have a likely role in cell wall metabolism (10). Several genes with similar functions contain putative YycF binding sites in their promoter regions. The essential nature of the YycFG system underscores the importance of the multiplicity of genes under YycF expression control. Our results on yych and yyci deletion strains which could have the entire regulon misregulated demonstrate the importance of keeping the expression of the regulon within certain limits. Therefore, the YycFG system seems to be a leading contributor to maintaining cell wall homeostasis. Numerous two-component sensing systems regulate genes involved in the utilization and metabolism of specific nutrients (e.g., CitST) (25). The signals sensed by these systems are of course these nutrients, a concept easily understood. On the contrary, the signals sensed by the essential YycFG two-component system are not as obvious. While it is possible that some environmental signals might alter the activity of this sensing system, it seems more likely that cellular signals are being sensed to allow for adjustments in expression levels of the regulon. Given the role of YycFG in regulating genes for divisional processes (i.e., ftsz) and for cell wall homeostasis, it is likely that the periplasmic domains of this TCS are transmitting signals that derive from the cell wall, cell membrane, or specific cellular systems located at these positions. These could be small compounds such as cell wall metabolites or other proteins. In any case, the periplasmic ternary complex of YycG, YycH, and YycI may allow for the integration of multiple signals. Despite the high conservation of the YycFG system among

3288 SZURMANT ET AL. J. BACTERIOL. the low-g C gram-positive bacteria, the streptococci are missing the three-tier regulation observed in B. subtilis and most other YycFG-containing bacteria (32, 37). YycH and YycI as well as the periplasmic domain of YycG are absent in these bacteria. Therefore, only intramembrane or cytoplasmic signals may be sensed in the streptococci. In addition, current knowledge suggests significant differences in the regulon. Despite the sequence conservation, it remains to be seen how similar the role of the YycFG system is in these distinct groups of Firmicutes. The YycFG system given its conservation and essentiality serves a special role in the low-g C gram positives. B. subtilis and many other Firmicutes express two phylogenetically closely related sensing systems: the PhoPR system and the ResDE system. The PhoPR system is important under phosphate limitation conditions, whereas the ResDE system is essential in the absence of glucose as a carbon source (17, 36). Several investigators have demonstrated interconnections of processes regulated by this triad of sensing systems. For instance, the ResDE system contributes significantly to the phosphate limitation response indirectly by controlling the redox state of the quinone pool (35). Devine and colleagues showed apparent cross phosphorylation between the PhoR histidine kinase and the YycF response regulator (16). On a transcriptional level, expression of the phopr operon is under direct YycF control and expression of the resde operon is under direct PhoP control (2, 16). The interconnection of these sensing systems is now further complicated by the fact that additional input proteins, YycH and YycI, regulate the YycFG system. The function of the YycJ protein whose gene is found adjacent to all YycFG TCS remains unknown. It was suggested to have a role in regulating YycG in YycF-depleted S. pneumoniae (30), but in B. subtilis, we have not found conditions that indicate a similar role. The YycJ protein is clearly a member of the -lactamase fold family of proteins,which includes many proteases. This may be its function as yycj with a promoter cannot be cloned in E. coli despite many attempts. Equally mysterious is accumulation of Spo segregants in a strain with yycj deleted. ACKNOWLEDGMENTS The authors would like to thank Scott Stibitz for furnishing the I-SceI expression plasmid pbkj223 and for providing his protocol for construction of in-frame deletions prior to its publication. We also thank Marta Perego for plasmid pjm103-i-scei. This study was supported in part by grant GM019416 from the National Institute of General Medicine Sciences and grant AI055860 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, USPHS. The Stein Beneficial Trust supported in part oligonucleotide synthesis and DNA sequencing. REFERENCES 1. Bent, C. J., N. W. Isaacs, T. J. Mitchell, and A. Riboldi-Tunnicliffe. 2004. Crystal structure of the response regulator 02 receiver domain, the essential YycF two-component system of Streptococcus pneumoniae in both complexed and native states. J. Bacteriol. 186:2872 2879. 2. Birkey, S. M., W. Liu, X. Zhang, M. F. Duggan, and F. M. Hulett. 1998. Pho signal transduction network reveals direct transcriptional regulation of one two-component system by another two-component regulator: Bacillus subtilis PhoP directly regulates production of ResD. Mol. Microbiol. 30:943 953. 3. Clausen, V. A., W. Bae, J. Throup, M. K. Burnham, M. Rosenberg, and N. G. Wallis. 2003. Biochemical characterization of the first essential two-component signal transduction system from Staphylococcus aureus and Streptococcus pneumoniae. J. Mol. Microbiol. Biotechnol. 5:252 260. 4. Dartois, V., J.-Y. Coppée, C. Colson, and A. Baulard. 1994. Genetic analysis and overexpression of lipolytic activity in Bacillus subtilis. Appl. Environ. Microbiol. 60:1670 1673. 5. Dartois, V., T. Djavakhishvili, and J. A. Hoch. 1997. KapB is a lipoprotein required for KinB signal transduction and activation of the phosphorelay to sporulation in Bacillus subtilis. Mol. Microbiol. 26:1097 1108. 6. Dubrac, S., and T. Msadek. 2004. Identification of genes controlled by the essential YycG/YycF two-component system in Staphylococcus aureus. J. Bacteriol. 186:1175 1181. 7. Echenique, J. R., and M.-C. Trombe. 2001. Competence repression under oxygen limitation through the two-component MicAB signal-transducing system in Streptococcus pneumoniae and involvement of the PAS domain of MicB. J. Bacteriol. 183:4599 4608. 8. Fabret, C., and J. A. Hoch. 1998. A two-component signal transduction system essential for growth of Bacillus subtilis: implications for anti-infective therapy. J. Bacteriol. 180:6375 6383. 9. Ferrari, E., D. J. Henner, M. Perego, and J. A. Hoch. 1988. Transcription of Bacillus subtilis subtilisin and expression of subtilisin in sporulation mutants. J. Bacteriol. 170:289 295. 10. Fukuchi, K., K. Kasahara, K. Asai, K. Kobayashi, S. Moriya, and N. Ogasawara. 2000. The essential two-component regulatory system encoded by yycf and yycg modulates expression of the ftsaz operon in Bacillus subtilis. Microbiology 146:1573 1583. 11. Green, J., C. Scott, and J. R. Guest. 2001. Functional versatility in the CRP-FNR superfamily of transcription factors: FNR and FLP. Adv. Microb. Physiol. 44:1 34. 12. Guerout-Fleury, A. M., N. Frandsen, and P. Stragier. 1996. Plasmids for ectopic integration in Bacillus subtilis. Gene 180:57 61. 13. Hancock, L. E., and M. Perego. 2004. Systematic inactivation and phenotypic characterization of two-component signal transduction systems of Enterococcus faecalis V583. J. Bacteriol. 186:7951 7958. 14. Hoch, J. A. 2000. Two-component and phosphorelay signal transduction. Curr. Opin. Microbiol. 3:165 170. 15. Howell, A., S. Dubrac, K. K. Andersen, D. Noone, J. Fert, T. Msadek, and K. Devine. 2003. Genes controlled by the essential YycG/YycF two-component system of Bacillus subtilis revealed through a novel hybrid regulator approach. Mol. Microbiol. 49:1639 1655. 16. Howell, A., S. Dubrac, D. Noone, K. I. Varughese, and K. Devine. 2006. Interactions between the YycFG and PhoPR two-component systems in Bacillus subtilis: the PhoR kinase phosphorylates the non-cognate YycF response regulator upon phosphate limitation. Mol. Microbiol. 59:1199 1215. 17. Hulett, F. M. 1996. The signal-transduction network for Pho regulation in Bacillus subtilis. Mol. Microbiol. 19:933 939. 18. Hulett, F. M., J. Lee, L. Shi, G. Sun, R. Chesnut, E. Sharkova, M. F. Duggan, and N. Kapp. 1994. Sequential action of two-component genetic switches regulates the PHO regulon in Bacillus subtilis. J. Bacteriol. 176:1348 1358. 19. Janes, B. K., and S. Stibitz. 2006. Routine markerless gene replacement in Bacillus anthracis. Infect. Immun. 74:1949 1953. 20. Karimova, G., N. Dautin, and D. Ladant. 2005. Interaction network among Escherichia coli membrane proteins involved in cell division as revealed by bacterial two-hybrid analysis. J. Bacteriol. 187:2233 2243. 21. Karimova, G., J. Pidoux, A. Ullmann, and D. Ladant. 1998. A bacterial two-hybrid system based on a reconstituted signal transduction pathway. Proc. Natl. Acad. Sci. USA 95:5752 5756. 22. Karimova, G., A. Ullmann, and D. Ladant. 2000. A bacterial two-hybrid system that exploits a camp signaling cascade in Escherichia coli. Methods Enzymol. 328:59 73. 23. Karimova, G., A. Ullmann, and D. Ladant. 2001. Protein-protein interaction between Bacillus stearothermophilus tyrosyl-trna synthetase subdomains revealed by a bacterial two-hybrid system. J. Mol. Microbiol. Biotechnol. 3:73 82. 24. Martin, P. K., T. Li, D. Sun, D. P. Biek, and M. B. Schmid. 1999. Role in cell permeability of an essential two-component system in Staphylococcus aureus. J. Bacteriol. 181:3666 3673. 25. Mascher, T., J. D. Helmann, and G. Unden. 2006. Stimulus perception in bacterial signal-transducing histidine kinases. Microbiol. Mol. Biol. Rev. 70:910 938. 26. Miller, J. H. 1972. Experiments in molecular genetics, p. 352 355. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 27. Mizuno, T. 2005. Two-component phosphorelay signal transduction systems in plants: from hormone responses to circadian rhythms. Biosci. Biotechnol. Biochem. 69:2263 2276. 28. Mohedano, M. L., K. Overweg, A. de la Fuentes, M. Reuter, S. Altabe, F. Mulholland, D. de Mendoza, P. Lopez, and J. M. Wells. 2005. Evidence that the essential response regulator YycF in Streptococcus pneumoniae modulates expression of fatty acid biosynthesis genes and alters membrane composition. J. Bacteriol. 187:2357 2367. 29. Monteilhet, C., A. Perrin, A. Thierry, L. Colleaux, and B. Dujon. 1990. Purification and characterization of the in vitro activity of I-Sce I, a novel and highly specific endonuclease encoded by a group I intron. Nucleic Acids Res. 18:1407 1413.