Discovery and Characterization of Three New Escherichia coli Septal Ring Proteins That Contain a SPOR Domain: DamX, DedD, and RlpA

Abstract

SPOR domains are ∼70 amino acids long and occur in >1,500 proteins identified by sequencing of bacterial genomes. The SPOR domains in the FtsN cell division proteins from Escherichia coli and Caulobacter crescentus have been shown to bind peptidoglycan. Besides FtsN, E. coli has three additional SPOR domain proteins—DamX, DedD, and RlpA. We show here that all three of these proteins localize to the septal ring in E. coli. The loss of DamX or DedD either alone or in combination with mutations in genes encoding other division proteins resulted in a variety of division phenotypes, demonstrating that DamX and DedD participate in cytokinesis. In contrast, RlpA mutants divided normally. Follow-up studies revealed that the SPOR domains themselves localize to the septal ring in vivo and bind peptidoglycan in vitro. Even SPOR domains from heterologous organisms, including Aquifex aeolicus, localized to septal rings when produced in E. coli and bound to purified E. coli peptidoglycan sacculi. We speculate that SPOR domains localize to the division site by binding preferentially to septal peptidoglycan. We further suggest that SPOR domain proteins are a common feature of the division apparatus in bacteria. DamX was characterized further and found to interact with multiple division proteins in a bacterial two-hybrid assay. One interaction partner is FtsQ, and several synthetic phenotypes suggest that DamX is a negative regulator of FtsQ function.

Cell division in Escherichia coli is mediated by a collection of approximately 20 proteins, all of which localize to the midcell, where they form a structure called the septal ring, or divisome. About half of these proteins are essential for cell division. The corresponding temperature-sensitive mutants or depletion strains become filamentous and die under nonpermissive conditions. The remaining proteins are not essential under most laboratory conditions. In some cases null mutations reveal modest division defects, but in other cases division defects become apparent only under certain growth conditions or in combination with mutations in genes for other division proteins. For reviews of this topic, see references 18, 22, 29, and 67.

One of the essential cell division proteins is a bitopic membrane protein named FtsN (see Fig. 1A) (13, 14). How FtsN facilitates cell division is not clear. Because overproduction of FtsN rescues a variety of mutants with lesions in genes for other cell division proteins [ftsA(Ts), ftsI(Ts) ftsQ(Ts), ftsEX null, ftsK null, and ftsP (sufI) null strains], it seems likely that one function of FtsN is to improve the assembly and/or stability of the septal ring (13, 20, 24, 30, 58, 63). Very recent evidence indicates that FtsN plays an important role in triggering constriction, probably by allosteric activation of some other component of the septal ring (26).

FIG. 1.

SPOR domain proteins included in this study. (A) Membrane topology and number of amino acids in each domain as retrieved from UniProt release 15.7 (http://www.uniprot.org) or the GTOP update of 15 December 2008 (http://spock.genes.nig.ac.jp/∼genome/gtop.html). N, amino terminus; CM, cytoplasmic membrane; OM, outer membrane. RlpA and VPA1294 have a covalently attached lipid at their amino termini. (B) Multiple-sequence alignment of SPOR domains shown in the present study to localize to the septal ring of E. coli. Sequences were aligned manually to the position-specific scoring matrix (PSSM) from http://www.ncbi.nlm.nih.gov/Class/Structure/pssm/pssm_viewer.cgi with the SPOR domain (Pfam accession no. 05036) as the PSSM identifier (PSSM ID). Residues with identity to those in the consensus sequence from the PSSM alignment are shaded gray. Numbers to the left refer to the first positions of the SPOR domains in the indicated proteins.

A notable feature of FtsN is that it contains at its C terminus a peptidoglycan (PG) binding domain known as a SPOR domain (Pfam accession no. 05036) (23, 65, 72). SPOR domains are both common and widespread in bacteria. At the time of this writing (August 2009), over 1,500 proteins that contain a SPOR domain are listed in the Pfam database (23). These proteins come from over 500 bacterial species. The domain is named after the founding member of the protein family, a Bacillus subtilis protein named CwlC that is produced relatively late in the process of sporulation (41). CwlC, which comprises an N-terminal amidase domain and a C-terminal SPOR domain, facilitates release of the mature spore by degrading PG in the mother cell (48, 61).

Our interest in SPOR domain proteins was piqued during a study of Vibrio parahaemolyticus (in collaboration with Linda McCarter) when we observed that a gene of unknown function, designated vpa1294, is highly induced in V. parahaemolyticus swarmer cells. The VPA1294 protein was annotated as a “putative DamX-related protein” (44; http://genome.gen-info.osaka-u.ac.jp/bacteria/vpara/). To learn about DamX, we turned to the EcoGene website (http://ecogene.org/) (57), which noted that (i) DamX from E. coli has an essentially unknown function, (ii) overproduction of DamX inhibits cell division (43), and (iii) DamX is one of four E. coli proteins that contain a SPOR domain, the others being the cell division protein FtsN and two proteins of unknown function, DedD and RlpA. Based on this information, we decided to investigate whether DamX, DedD, and RlpA are involved in cell division in E. coli. While this work was in progress, the Thanbichler laboratory demonstrated that Caulobacter crescentus has an FtsN-like protein that is needed for cell division (49) and the de Boer laboratory published a report on DamX, DedD, and RlpA from E. coli (26). We also learned that J. Maddock's laboratory has been investigating DamX, DedD, and RlpA from E. coli (personal communication). Importantly, the major findings from all four laboratories are in general agreement: SPOR domain proteins are widespread in bacteria, many of these proteins are involved in cell division, and SPOR domains are sufficient for septal localization, probably because SPOR domains bind to septal PG.

MATERIALS AND METHODS

Media.

E. coli strains were grown in Luria-Bertani (LB) medium containing 10 g tryptone, 5 g yeast extract, and 10 g NaCl per liter. For LB0N medium, NaCl was omitted. Plates contained 15 g agar per liter. Antibiotics were used at the following concentrations: 200 μg ampicillin/ml, 30 μg chloramphenicol/ml, 40 μg kanamycin/ml, and 100 or 35 μg spectinomycin/ml for plasmids or chromosomal alleles, respectively.

Strains.

Strains used in this study are listed in Table 1. Aquifex aeolicus DNA was obtained from R. Huber, Cytophaga hutchinsonii ATCC 33406 was obtained from the American Type Culture Collection, and V. parahaemolyticus LM5674 was obtained from L. McCarter. Eviction of antibiotic cassettes by using pCP20, integration of CRIM (conditional-replication, integration, and modular) plasmids into the chromosome, and P1-mediated transduction were done according to procedures described previously (11, 15, 34, 47). All kan evictions left behind an frt scar.

TABLE 1.

Strains used in this study

Strain	Relevant featuresa	Source or reference
BL21(λ DE3)	ompT gal dcm hsdSB(rB− mB−) λ (PlacUV5::T7 gene 1)	Novagen
BW25113	rrnB3 ΔlacZ4787 hsdR514 Δ(araBAD)567 Δ(rhaBAD)568 rph-1	2
DHM1	glnV44(AS) recA1 endA gyrA96 thi-1 hsdR17 spoT1 rfbD1 cya-854	38
EC251	K-12 wild type MG1655	Lab collection
EC295	EC251 ftsI23(Ts) leu::Tn10	This study
EC297	EC251 ftsA12(Ts) leu::Tn10	21
EC303	EC251 ftsQ1(Ts) leu::Tn10	This study
EC309	EC251 ftsZ84(Ts) leu::Tn10	63
EC1793	BW25113 ΔdamX	This study
EC1794	BW25113 ΔdedD	This study
EC1795	BW25113 ΔrlpA	This study
EC1908	EC251 PBAD::ftsN (Kmr)	63
EC1910	EC251 damX<>kan	This study
EC1912	EC251 dedD<>kan	This study
EC1914	EC251 rlpA<>kan	This study
EC1916	BW25113 ΔdamX rlpA<>kan	This study
EC1918	BW25113 ΔdedD rlpA<>kan	This study
EC1920	BW25113 ΔrlpA damX<>kan	This study
EC1922	BW25113 ΔrlpA dedD<>kan	This study
EC1924	BW25113 ΔdamX dedD<>kan	This study
EC1926	BW25113 ΔdedD damX<>kan	This study
EC1928	BW25113 ΔdamX/pDSW998 (PBAD::dedD Cmr)	This study
EC1929	BW25113 ΔrlpA/pDSW998 (PBAD::dedD Cmr)	This study
EC1930	BW25113 ΔdamX ΔrlpA	This study
EC1932	BW25113 ΔdamX ΔrlpA	This study
EC1934	BW25113 ΔdamX dedD<>kan/pDSW998 (PBAD::dedD Cmr)	This study
EC1936	BW25113 ΔrlpA dedD<>kan/pDSW998 (PBAD::dedD Cmr)	This study
EC1938	BW25113 ΔdamX ΔrlpA/pDSW998 (PBAD::dedD Cmr)	This study
EC1939	BW25113 ΔdamX ΔrlpA/pDSW998 (PBAD::dedD Cmr)	This study
EC1940	BW25113 ΔdamX ΔdedD/pDSW998 (PBAD::dedD Cmr)	This study
EC1941	BW25113 ΔrlpA ΔdedD/pDSW998 (PBAD::dedD Cmr)	This study
EC1942	BW25113 ΔdamX ΔrlpA dedD<>kan/pDSW998 (PBAD::dedD Cmr)	This study
EC1944	BW25113 ΔdamX ΔrlpA dedD<>kan/pDSW998 (PBAD::dedD Cmr)	This study
EC1946	BW25113 ΔdamX ΔdedD rlpA<>kan/pDSW998 (PBAD::dedD Cmr)	This study
EC1948	BW25113 ΔrlpA ΔdedD damX<>kan/pDSW998 (PBAD::dedD Cmr)	This study
EC1986	EC251 leu::Tn10 damX<>kan	This study
EC1988	EC251 ftsI23(Ts) leu::Tn10 damX<>kan	This study
EC1990	EC251 ftsA12(Ts) leu::Tn10 damX<>kan	This study
EC1992	EC251 ftsQ1(Ts) leu::Tn10 damX<>kan	This study
EC1994	EC251 ftsZ84(Ts) leu::Tn10 damX<>kan	This study
EC1996	W3110 zipA1(Ts) damX<>kan	This study
EC2130	BW25113 damX<>kan φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2132	BW25113 dedD<>kan φ80::pDSW984 (P204::gfp-dedD Spr)	This study
EC2138	EC251 φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2139	EC251 φ80::pDSW984 (P204::gfp-dedD Spr)	This study
EC2140	EC251 ftsZ84(Ts) leu::Tn10 φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2142	EC251 ftsA12(Ts) leu::Tn10 φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2144	EC251 ftsQ1(Ts) leu::Tn10 φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2146	EC251 ftsI23(Ts) leu::Tn10 φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2148	EC251 PBAD::ftsN kan φ80::pDSW983 (P204::gfp-damX Spr)	This study
EC2150	BW25113 ΔdamX ΔrlpA dedD<>kan/pDSW390	This study
EC2152	BW25113 ΔdamX ΔrlpA dedD<>kan/pDSW390	This study
EC2154	BW25113 ΔdamX ΔdedD rlpA<>kan/pDSW390	This study
EC2156	BW25113 ΔrlpA ΔdedD damX<>kan/pDSW390	This study
EC2177	BW25113 ΔdamX dedD<>kan φ80::pDSW983(P204::gfp-damX Spr)	This study
JW3351	BW25113 damX<>kan	2
JW3578	BW25113 dedD<>kan	2
JW0628	BW25113 rlpA<>kan	2
PS223	W3110 zipA1(Ts)	54

^aAS, amber suppressor.

(i) Strains for phenotypic analysis of damX, dedD, and rlpA mutants.

The Keio collection strains JW3351 (damX<>kan), JW3578 (dedD<>kan), and JW0628 (rlpA<>kan) were obtained and verified by diagnostic PCR as described previously (2, 15). Strains EC1793, EC1794, and EC1795 were constructed using pCP20 to evict kan from JW3351, JW3578, and JW0628. EC1910, EC1912, and EC1914 were constructed using transduction to introduce damX<>kan, dedD<>kan, or rlpA<>kan from JW3351, JW3578, or JW0628 into EC251. Strains EC1916, EC1918, EC1920, EC1922, EC1924, and EC1926 were constructed using transduction to introduce damX<>kan, dedD<>kan, or rlpA<>kan from JW3351, JW3578, or JW0628 into EC1793, EC1794, or EC1795. Strains EC1928 and EC1929 were constructed by transforming EC1793 and EC1795 with pDSW998. EC1930 and EC1932 were constructed using pCP20 to evict kan from EC1916 and EC1920. Strains EC1934 and EC1936 were constructed using transduction to introduce dedD<>kan from JW3578 into EC1928 and EC1929. EC1938 and EC1939 were constructed using electroporation to introduce pDSW998 into EC1930 and EC1932. Strains EC1940 and EC1941 were constructed using pCP20 to evict kan from EC1934 and EC1936. Strains EC1942, EC1944, EC1946, and EC1948 were constructed by transduction to introduce dedD<>kan, rlpA<>kan, or damX<>kan from JW3351, JW3578, or JW0628 into EC1938, EC1939, EC1940, or EC1941. Strains EC2150, EC2152, EC2154, and EC2156 were constructed using pDSW390 to bump out pDSW998 from EC1942, EC1944, EC1946, and EC1948.

(ii) Strains for localization of green fluorescent protein (GFP)-DamX or GFP-DedD in wild-type or mutant backgrounds.

The CRIM plasmids pDSW983 (P₂₀₄::gfp-damX) and pDSW984 (P₂₀₄::gfp-dedD) were integrated into the chromosomes of EC251, EC295, EC297, EC303, EC309, EC1908, EC1924, JW3351, and JW3578. The resulting strains were designated EC2138, EC2139, EC2140, EC2142, EC2144, EC2146, EC2148, EC2130, EC2132, and EC2177.

(iii) Strains to screen for synthetic phenotypes of damX<>kan with other fts mutations.

P1 transduction was used to move damX<>kan from JW3351 into EC294, EC295, EC297, EC303, EC309, and PS223. The resulting strains were designated EC1986, EC1988, EC1990, EC1992, EC1994, and EC1996, respectively.

(iv) Other strains.

EC295 and EC303 were constructed by transduction to introduce ftsI23(Ts) together with leu::Tn10 and ftsQ1(Ts) together with leu::Tn10, respectively, into EC251. The donor strains were LMG64 and MJC129 (33).

Plasmids.

Plasmids used in this study are listed in Table 2. In the following descriptions, primer-carried restriction sites are underlined.

TABLE 2.

Plasmids used in this study

Plasmid	Relevant featuresa	Source or reference
pBAD33	Arabinose regulation (PBAD); p15A ori Cmr	32
pDSW204	IPTG regulation (P204); lacIq pBR ori Apr	68
pDSW206	IPTG regulation (P206); lacIq pBR ori Apr	68
pDSW207	gfp fusion vector (pDSW204-gfp-MCS)	68
pDSW390	gfp fusion vector (P206::gfp-MCS lacIq p15A ori Apr)	Lab collection
pDSW433	pBAD33-ftsN	Lab collection
pDSW913	rfp fusion vector (pDSW206-MCS-mcherry)	This study
pDSW918	P204::gfp-damX lacIqbla pBR ori	This study
pDSW929	P206::vpa1294-rfp lacIqbla pBR ori	This study
pDSW931	P206::rlpA-rfp lacIqbla pBR ori	This study
pDSW937	P204::gfp-dedD lacIqbla pBR ori	This study
pDSW962	TTgfp fusion vector (P204::sstorA-gfp lacIqbla pBR ori)	63
pDSW966	Plac::cya1-675-damX kan p15A ori (pKT25-damX)	This study
pDSW967	Plac::cya1-675-dedD kan p15A ori (pKT25-dedD)	This study
pDSW968	Plac::cya675-1197-damX bla pBR ori (pUT18c-damX)	This study
pDSW969	Plac::cya675-1197-dedD bla pBR ori (pUT18c-dedD)	This study
pDSW983	P204::gfp-damX lacIq Spr (pJC69 derivative)	This study
pDSW984	P204::gfp-dedD lacIq Spr (pJC69 derivative)	This study
pDSW992	P204::TTgfp-ftsN(240-319) lacIqbla	This study
pDSW994	P204::TTgfp-dedD(141-220) lacIqbla	This study
pDSW995	P204::TTgfp-rlpA(283-362) lacIqbla	This study
pDSW997	P204::TTgfp-damX(338-428) lacIqbla	This study
pDSW998	pBAD33-dedD	This study
pDSW1000	pQE80L-damX(338-428)	This study
pDSW1002	pQE80L-dedD(141-220)	This study
pDSW1003	pQE80L-rlpA(283-362)	This study
pDSW1113	pQE80L-chu2221(161-249)	This study
pDSW1114	pQE80L-aq1869(254-342)	This study
pDSW1115	pQE80L-vpa1294(37-134)	This study
pDSW1067	P204::TTgfp-aq1896(254-342) lacIqbla	This study
pDSW1069	P204::TTgfp-chu2221(161-249) lacIqbla	This study
pJC69	oriR6Kγ attP φ80 Spr (CRIM vector)	8
pET-3Z+	PT7gene10::ftsZ Apr pBR ori	55
pKT25	BACTH vector for fusion of target proteins to Bordetella pertussis cya gene T25 fragment; Plac::cya1-675 p15 ori Kmr	40
pKT25-ftsA	Plac::cya1-675-ftsA	38
pKT25-ftsB	Plac::cya1-675-ftsB	38
pKT25-ftsI	Plac::cya1-675-ftsI	38
pKT25-ftsL	Plac::cya1-675-ftsL	38
pKT25-ftsN	Plac::cya1-675-ftsN	38
pKT25-ftsQ	Plac::cya1-675-ftsQ	38
pKT25-ftsW	Plac::cya1-675-ftsW	38
pKT25-ftsX	Plac::cya1-675-ftsX	38
pKT25-ftsZ	Plac::cya1-675-ftsZ	38
pKT25-zip	Plac::cya1-675-zip (construct encoding leucine zipper from GCN4)	39
pQE80L	Carries T5 promoter and lac operators; lacIq ColE1 ori Apr	Qiagen
pUT18c	BACTH vector for fusion of target proteins to B. pertussis cya gene T18 fragment; Plac::cya675-1197-MCS pUC ori Apr	40
pUT18c-ftsA	Plac::cya675-1197-ftsA	38
pUT18c-ftsB	Plac::cya675-1197-ftsB	38
pUT18c-ftsI	Plac::cya675-1197-ftsI	38
pUT18c-ftsL	Plac::cya675-1197-ftsL	38
pUT18c-ftsN	Plac::cya675-1197-ftsN	38
pUT18c-ftsQ	Plac::cya675-1197-ftsQ	38
pUT18c-ftsW	Plac::cya675-1197-ftsW	38
pUT18c-ftsX	Plac::cya675-1197-ftsX	38
pUT18c-zip	Plac::cya675-1197-zip (construct encoding leucine zipper from GCN4)	39

^aResidues encoded by cya gene fragments are indicated by superscript numbers. ^sstorA, TorA signal sequence construct; BACTH, bacterial two-hybrid system.

(i) Plasmids that encode GFP or red fluorescent protein (RFP) fusions to full-length SPOR domain proteins.

pDSW918 (P₂₀₄::gfp-damX) was constructed by amplifying damX from EC251 chromosomal DNA with primers P1136 (CAGCAATTGAACAACAACATGGATGAATTCAAACCAGAAGAC) and P1137 (CTGAAGCTTACTTCAGATCGGCCTGTACCT). The 1,312-bp product was cut with MfeI and HindIII and ligated into pDSW207 that had been digested with EcoRI and HindIII. The fusion joins GFP to DamX via a 5-amino-acid (5-aa) linker, ELNNN. pDSW929 (P₂₀₆::vpa1294-mcherry) was constructed by amplifying vpa1294 from chromosomal DNA of V. parahaemolyticus strain LM5674 with primers VPA1294Up (CAGGAATTCGTGCACAAGATTTTATCTTTGACAT) and VPA1294Down (GTCTCTAGAGTTGTTGTTGCGCATTCTCACGACACTAGTT). The 429-bp product was cut with EcoRI and XbaI and ligated into the corresponding sites of pDSW913. The fusion construct changes the N terminus of VPA1294 from MHK to MEFVHK, and there is a 5-aa linker, SRNNN, between VPA1294 and mCherry. pDSW931 (P₂₀₆::rlpA-mcherry) was constructed by amplifying rlpA from EC251 chromosomal DNA with primers P1140 (CAGGAATTCATGCGTAAGCAGTGGCTCGGGA) and P1141 (GTCTCTAGAGTTGTTGTTCTGCGCGGTAGTAATAAATGAC). The 1,113-bp product was cut with EcoRI and XbaI and ligated into pDSW913 that had been cut with the same enzymes. The fusion construct adds the tripeptide MEF to the N terminus of RlpA, and there is a 5-aa linker, SRNNN, between RlpA and mCherry. pDSW937 (P₂₀₄::gfp-dedD) was constructed by amplifying dedD from EC251 chromosomal DNA with primers P1138 (CAGGAATTCAACGTGGCAAGTAAGTTTCAGAATCG) and P1139 (GACAAGCTTAATTCGGCGTATAGCCCATTAC). The 682-bp product was cut with EcoRI and HindIII and ligated into pDSW207 that had been digested with the same enzymes. The fusion joins GFP to DedD via a one-Asn-residue linker and replaces the initiating Met of DedD with Val. pDSW983 allows for integration of P₂₀₄::gfp-damX into the chromosome at the φ80 att site. It was constructed by digesting pDSW918 with SphI and ScaI. The 4,322-bp fragment that carried lacI^q and P₂₀₄::gfp-damX was ligated into pJC69 that had been cut with SphI and HincII. pDSW984 (P₂₀₄::gfp-dedD) was constructed similarly using a 3,692-bp SphI-ScaI fragment from pDSW937.

(ii) Plasmids that encode fusions of the TorA signal sequence and GFP to isolated SPOR domains.

pDSW992, expressing a fusion protein comprising Tat-targeted GFP (^TTGFP) and FtsN residues 240 to 319 under the control of P₂₀₄ [P₂₀₄::^TTgfp-ftsN(240-319)], was constructed by amplifying the last 80 codons of ftsN with primers P1123 (GCCGGATCCAACAACAACGCGGAGAAAAAAGACGAA) and P761 (CCCAAGCTTTCAACCCCCGGCGGCGAGCCG) and plasmid pDSW433 as the template. The 270-bp product was cut with BamHI and HindIII and ligated into the corresponding sites of pDSW962. pDSW994 [P₂₀₄::^TTgfp-dedD(141-220)] was constructed by amplifying the last 80 codons of dedD from plasmid pDSW937 with primers P1125 (GCCGGATCCAACAACAACGGTAAAGCCTATGTTGTG) and P1126 (GCCAAGCTTTAATTCGGCGTATAGCCCATT). The 269-bp product was cut with BamHI and HindIII and ligated into the corresponding sites of pDSW962. pDSW995 [P₂₀₄::^TTgfp-rlpA(283-362)] was constructed by amplifying the last 80 codons of rlpA from plasmid pDSW931 with primers P1127 (GCCAGATCTAACAACAACGTCTCGCAAAGCGCCAGC) and P1128 (GCCAAGCTTTACTGCGCGGTAGTAATAAAT). The 269-bp product was cut with BglII and HindIII and ligated into the BamHI and HindIII sites of pDSW962. An analogous fusion protein comprising GFP and the last 80 residues of DamX did not localize to the septal ring, so we constructed a longer fusion protein that included the last 91 residues of DamX, which localized well. The corresponding plasmid, pDSW997 [P₂₀₄::^TTgfp-damX(338-428)], was constructed using primers P1144 (GCCGGATCCAACAACAACGGTTCGTTGAAATCGGCA) and P1137 (CTGAAGCTTACTTCAGATCGGCCTGTACCT) with pDSW918 as the template. The 301-bp product was cut with BamHI and HindIII and ligated into the corresponding sites of pDSW962. pDSW1067 [P₂₀₄::^TTgfp-aq1896(254-342)] was constructed by amplifying the last 89 codons of aq1896 from A. aeolicus chromosomal DNA with primers P1223 (GCCGGATCCAACAACAACATCCCAAGAGTGCATAAA) and P1224 (CTGAAGCTTATCACTTGATTTCAACGACGAA). The 298-bp product was cut with BamHI and HindIII and ligated into the corresponding sites of pDSW962. pDSW1069 [P₂₀₄::^TTgfp-chu2221(161-249)] was constructed by amplifying the last 89 codons of chu2221 from chromosomal DNA of C. hutchinsonii ATCC 33406 with primers P1229 (GCCGGATCCAACAACAACTTCTATTCTACTAAATTGGAG) and P1230 (CTGAAGCTTATTTTGGAGATAGGATCACACT). The 295-bp product was cut with BamHI and HindIII and ligated into the corresponding sites of pDSW962.

(iii) Plasmids for overproduction of His₆-tagged SPOR domains.

Plasmids for the overproduction of His₆-tagged SPOR domains were constructed as described above for ^TTGFP-SPOR domain plasmids except that the region encoding the SPOR domain of VPA1294 was amplified using primers P1317 (GCCGGATCCAACAACAACGCATACGGCTACCTGAATCCC) and P1318 (CTGAAGCTTAGCGCATTCTCACGACACTAG). The PCR products were ligated into pQE80L. All constructs carry the vector-derived sequence MRGSHHHHHHGSNNN at the N terminus.

(iv) Other plasmids.

pDSW433 (P_BAD::ftsN) was constructed by amplifying ftsN from a pBAD24-ftsN construct (lab collection) with primers P353 (GACTCTAGATCAACCCCCGGCGGC) and P354 (CTTGAGCTCGAGGAATTCACCATGGCACAAC). The 990-bp product was cut with SacI and XbaI and ligated into the corresponding sites of pBAD33. The resulting construct endowed ftsN with a Shine-Dalgarno sequence and a start codon derived in part from pBAD24. pDSW913 (P₂₀₆::multiple-cloning site [MCS]-mcherry) is a vector for fusing the mCherry variant of RFP to the C termini of target proteins. It was constructed by amplifying mcherry from plasmid pMalp2-mCherry (a gift from Greg Phillips, Iowa State University) with primers P991 (GTCAAAGCTTTTACTTGTACAGCTCGTCCATG) and P992 (GTCATCTAGAATGGTGAGCAAGGGCGAGGA). The 731-bp fragment was digested with HindIII and XbaI and ligated into the corresponding sites of pDSW206. To construct pDSW966 (P_lac::T25-damX) and pDSW968 (P_lac::T18-damX), damX was amplified by PCR from EC251 chromosomal DNA with primers P1076 (CTAGAGGATCCCATGGATGAATTCAAACCAG) and P1077 (CTTAGGTACCTTACTTCAGATCGGCCTG). The 1,286-bp fragment was cut with BamHI and KpnI and ligated into pKT25 and pUT18c that had been cut with the same enzymes. Likewise, to construct pDSW967 (P_lac::T25-dedD) and pDSW969 (P_lac::T18-dedD), dedD was amplified by PCR using primers P1078 (CTAGAGGATCCCGTGGCAAGTAAGTTTCAG) and P1079 (CTTAGGTACCTTAATTCGGCGTATAGCC). The 663-bp PCR product was cut with BamHI and KpnI and ligated into the corresponding sites of pKT25 and pUT18c. pDSW998 is a pBAD33 derivative that expresses dedD. It was constructed using primers P1153 (GTGTCTAGAGTCGCACATGTCATGGAAGTG) and P1154 (GTGAAGCTTAATTCGGCGTATAGCCCATTACC) to amplify dedD from BW25113 chromosomal DNA. The resulting 719-bp product was cut with XbaI and HindIII and ligated into the corresponding sites of pBAD33.

Division phenotypes of damX, dedD, and rlpA mutants.

Cultures grown overnight in LB medium were diluted 1:500 in the same medium and grown for 4 h at 30°C to an optical density at 600 nm (OD₆₀₀) of ∼0.5, at which point the cells were fixed. To better observe chaining and invaginations, cells were stained with the membrane dye FM4-64. To assay for deoxycholate sensitivity, overnight cultures grown in LB medium were adjusted to an OD₆₀₀ of 1.0 in LB0N medium and then fourfold serial dilutions were made and 3-μl aliquots were spotted onto an LB0N agar plate containing 0.01 mM IPTG (isopropyl-β-d-thiogalactopyranoside) and 0.1% deoxycholate. Plates were photographed after 18 h at 30°C.

Rescue of ftsQ1(Ts) mutant by damX<>kan.

Plating efficiency was assessed by spotting dilutions of cultures onto LB0N medium. One milliliter of overnight culture grown in LB medium was centrifuged, and the resulting cell pellet was taken up into 1 ml of LB0N medium. This culture was then diluted to a starting OD₆₀₀ of 1.0, and a series of 10-fold serial dilutions in LB0N medium was prepared. Four microliters of each dilution was spotted onto plates containing LB0N medium, and the plates were incubated at 42°C for 16 h before being photographed. Division efficiency was assessed by phase-contrast microscopy. Overnight cultures were diluted 1:50 into LB medium and grown for 2 h at 30°C. These cultures, now in exponential growth, were then diluted 1:6 into LB0N medium, grown for 1 h at 42°C, fixed, and photographed.

Protein localization.

Live cells were used in all cases for studies of protein localization. (i) Strains expressing full-length SPOR proteins fused to GFP carried IPTG-inducible fusions stably integrated into the chromosome, so antibiotic selection was not necessary. Cultures grown overnight in LB medium were diluted 1:500 into LB medium containing 0.1 mM IPTG and grown for 4 h at 30°C to an OD₆₀₀ of ∼0.5, at which point they were spotted onto thin agarose pads (63). (ii) Strains expressing full-length SPOR proteins fused to RFP were cultured similarly except that ampicillin was included to maintain plasmids and 1 mM IPTG was used to induce production of the fusion protein. (iii) Strains expressing fusions of GFP to isolated SPOR domains were grown similarly except that ampicillin was included to maintain plasmids, IPTG was omitted because basal expression proved sufficient, and overnight cultures were diluted 1:50 and grown for 2 h before microscopy.

Localization dependency studies.

Live cells were used in all cases for localization dependency studies. To localize GFP-DamX in temperature-sensitive strains, overnight cultures were diluted 1:50 into LB medium and grown for 2 h at 30°C. These cultures were then diluted 1:6 under permissive conditions (LB medium at 30°C) or nonpermissive conditions (LB0N medium at 42°C) and grown for 60 to 90 min. To visualize GFP-DamX in cells depleted of FtsN, cultures were grown overnight in LB medium containing 0.2% l-arabinose. The next morning, cultures were diluted 1:50 into LB medium containing 0.2% d-glucose and grown for 2 h at 30°C to an OD₆₀₀ of ∼0.5. These cultures were then diluted 1:50 into permissive medium (LB medium with arabinose and 0.1 M IPTG) or nonpermissive medium (LB medium with glucose and 0.1 M IPTG) and grown for 3 to 4 h, by which time the glucose-grown cells were filamentous. To localize GFP-DamX in a damX dedD<>kan double mutant, overnight cultures were diluted 1:500 into LB medium containing various concentrations of IPTG and grown to an OD₆₀₀ of ∼0.5.

General microscopy methods.

Our microscope, camera, and software have been described previously, as has the immobilization of live cells on agarose pads (1, 46, 63). The mCherry variant of RFP was visualized using the filter set we used previously for Texas Red and FM4-64 (70).

Bacterial two-hybrid assay.

Transformants of DHM1 carrying appropriate plasmid pairs (derivatives of pUT18c and pKT25) were streaked onto plates of LB medium containing ampicillin (200 μg/ml) and kanamycin (40 μg/ml), and the plates were incubated for 2 or 3 days at 30°C. Three to five colonies were used to inoculate 5 ml of LB medium containing ampicillin and kanamycin plus 0.5 mM IPTG and grown on a roller at 30°C for ∼16 h to an OD₆₀₀ of 0.6 to 0.8. Then duplicate 15-μl culture samples were assayed for β-galactosidase activity by standard procedures (47). All assays were repeated at least three times (on different days).

Western blotting.

For Western blotting, typically 1 ml of culture at an OD₆₀₀ of ∼0.5 was centrifuged and the resulting cell pellet was taken up into ∼0.5 ml of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer to achieve a sample concentration of 4.0 OD₆₀₀ units per ml. Samples were boiled, and 10-μl aliquots were subjected to SDS-10% PAGE. Proteins were then transferred onto nitrocellulose and detected by standard methods. Rabbit anti-GFP antibodies were obtained from W. Margolin and C. Ellermeier and used at dilutions of 1:2,500 and 1:10,000, respectively. Rabbit anti-RFP serum (a gift from L. Shapiro) was used at a 1:10,000 dilution, and affinity-purified rabbit anti-FtsQ (10) was used at a 1:4,000 dilution. The secondary antibody was horseradish peroxidase-conjugated goat anti-rabbit antibody (1:8,000; Pierce, Rockford, IL), which in turn was detected with SuperSignal Pico West chemiluminescent substrate (Pierce, Rockford, IL). Blots were visualized with an LAS-1000 luminescent imager from Fuji (Stamford, CT).

Purified proteins.

MBP2* was purchased from New England Biolabs (Beverly, MA). FtsZ was overproduced and purified as described previously (55), except that Q-Sepharose Fast-Flow was substituted for DEAE-Sephacel. His₆-tagged SPOR domains were overproduced in and purified from E. coli BL21(DE3) transformants by nickel affinity chromatography on Talon affinity resin according to instructions from the manufacturer (Clontech, Mountain View, CA). The purified proteins were dialyzed into storage buffer (50 mM Tris-HCl, 200 mM NaCl, 5% glycerol, pH 7.5), and aliquots were stored at −80°C until needed. A 0.5-liter culture yielded ∼4 mg purified protein as determined by a bicinchoninic acid assay (Pierce) with bovine serum albumin as the standard. The purity was >95% as judged by SDS-PAGE.

Purification and quantification of PG.

Whole PG sacculi were isolated from 1-liter cultures of EC251 after being boiled in SDS essentially as described previously (28). To quantify muramic acid, samples of purified PG were suspended in 6 N HCl and hydrolyzed at 95°C for 4 h. Hydrolysates were subjected to amino sugar analysis (31) and quantified relative to purified standards processed in parallel.

PG binding assays.

Binding assays (65) were conducted with a solution of 25 mM potassium phosphate, pH 7.5, and 200 mM NaCl. A standard assay mixture contained 12 μg protein and 75 μg PG in a total volume of 100 μl. As a control, assay mixtures that lacked PG were also prepared. Mixtures were incubated for 1 h on ice and then centrifuged at 4°C in a Beckman TLA-55 rotor at 50,000 rpm (average g force, 112,000 × g) for 45 min. The PG pellet was washed by suspending it in 100 μl of cold assay buffer and centrifuging as described above. The PG pellet was again suspended in 100 μl of binding buffer. The supernatant from the initial binding step, the wash fluid, and the resuspended pellet were analyzed by SDS-15% PAGE. Gels were stained with GelCode blue (Pierce, Rockford, IL) and scanned using a Typhoon 8610 imager with the following instrument settings: excitation laser, 532 nm; emission filter, 560 nm, long pass; photomultiplier, 600 V; and pixel size, 100 μm. ImageQuant software was used to quantify fluorescence signals.

RESULTS

Overview of the E. coli SPOR domain proteins DamX, DedD, and RlpA.

The E. coli SPOR domain proteins are diagrammed in Fig. 1A. DamX, DedD, and FtsN have similar overall structures. Each is a bitopic inner membrane protein with an N-terminal cytoplasmic domain, a single transmembrane helix, and a relatively large periplasmic domain, the last ∼70 aa of which comprises a SPOR domain. RlpA, in contrast, is predicted to be an outer membrane lipoprotein. Alignments of the four proteins indicate that their homology is confined to the SPOR domains, which are only 18% identical in pairwise comparisons. Figure 1B compares the four E. coli SPOR domains and three heterologous SPOR domains included in this study.

Except for FtsN, the E. coli SPOR domain proteins have unknown functions. None of the proteins appear to be essential, as null mutants exist in the Keio collection and the corresponding genes did not turn up in a previous screen for essential genes in E. coli (2, 25). damX is named for its location next to the dam gene, which encodes DNA adenine methyltransferase (37). Overexpression of damX has been reported to interfere with cell division and alter biofilm formation (43, 64). DedD is named for downstream E. coli DNA, reflecting the discovery of this protein by sequencing downstream of folC (52). To our knowledge, no phenotypes have been associated with dedD mutants. RlpA is named for rare lipoprotein A (62), and rlpA is located in an operon with genes encoding PBP2 and RodA; the downstream gene encodes PBP5. All of these proteins are involved in cell wall metabolism and/or cell shape, suggesting that RlpA may be involved in these processes as well (3). Nevertheless, the only previously published data regarding a phenotype associated with RlpA is the finding that a truncated rlpA gene is a multicopy suppressor of a prc mutation (3). Prc, also known as Tsp, is a periplasmic protease that processes a variety of substrates that have nonpolar C termini, including the cell division protein FtsI (51).

DamX, DedD, and RlpA localize to the septal ring.

To determine whether the uncharacterized E. coli SPOR domain proteins localize to the septal ring, we constructed IPTG-inducible fusions to the mut2 variant of GFP (12, 68) or the mCherry variant of RFP (9, 42, 59). We also constructed an RFP fusion to VPA1294, the V. parahaemolyticus protein that launched this investigation. The gfp-damX and gfp-dedD fusions were integrated into the chromosome in single copies, but our camera had difficulty detecting RFP when produced from the chromosome, so rlpA-rfp and vpa1294-rfp were expressed from plasmids.

Cells in exponential growth were immobilized on an agarose pad and visualized by fluorescence microscopy. Each of the SPOR domain proteins appeared as a bright band of fluorescence at the midcell, but in deeply constricting cells, the fluorescent signal was more reminiscent of a bright dot at the point of constriction (Fig. 2A and B). In the case of DamX, ∼80% of the cells exhibited septal localization, suggesting that DamX is an early recruit to the septal ring. But only ∼40% of cells exhibited septal DedD or R1pA, implying that these proteins are late recruits. In summary, we infer that the SPOR domain proteins localize to the septal ring prior to or at the onset of constriction and remain associated with the ring until division is complete.

FIG. 2.

Full-length SPOR domain proteins localize to the septal ring. (A) Fluorescence micrographs of live cells producing the indicated GFP or RFP fusion protein (left) and frequencies of septal localization (in percentages for ≥200 cells analyzed) in these cells (right). Bar, 5 μm. Cells were grown to an OD₆₀₀ of ∼0.5 in LB medium with 0.1 or 1 mM IPTG for GFP or RFP fusions, respectively. (B) Localization of GFP-DamX in a deeply constricted cell. Phase-contrast (left), fluorescence (middle), and overlay (right) micrographs. (C) Analysis of GFP and RFP fusion proteins by Western blotting. Whole-cell extracts were subjected to SDS-PAGE, transferred onto nitrocellulose, and probed with anti-GFP or anti-RFP. Predicted masses are as follows: GFP-DamX, 73.4 kDa; GFP-DedD, 50.1 kDa; RlpA-RFP, 65.2 kDa; and VPA1294-RFP, 40.6 kDa. For RlpA-RFP and VPA1294-RFP, these masses apply to the constructs after processing of the type II signal sequences. Molecular mass markers in kilodaltons are indicated to the left. The strains used were EC2130 (damX<>kan P₂₀₄::gfp-damX), EC2132 (dedD<>kan P₂₀₄::gfp-dedD), JW0628/pDSW931 (rlpA<>kan P₂₀₆::rlpA-rfp), and EC251/pDSW929 (wild type carrying P₂₀₆::vpa1294-rfp plasmid).

Samples of the cultures examined by microscopy were also analyzed by Western blotting with anti-GFP and anti-RFP antibodies to verify that the fusion proteins were largely intact. As expected, the full-length fusion protein was the predominant species in all cases, although VPA1294-RFP was somewhat prone to proteolysis (Fig. 2C). For reasons that are not known, GFP-DamX consistently migrated more slowly than expected based on its predicted mass. We also verified that the GFP-DamX and GFP-DedD fusion proteins function in cell division (see below), but the functionality of RlpA-RFP could not be tested because we were unable to find any phenotypes associated with the loss of RlpA.

SPOR domains themselves localize to septal rings.

The observation that DamX, DedD, RlpA, and VPA1294 localize to septal rings suggested that SPOR domains themselves localize to septal rings. This inference was particularly strong in the case of VPA1294, since it consists of a SPOR domain and little else (Fig. 1A). Moreover, FtsN′s SPOR domain was said to localize to division sites (P. de Boer, personal communication). We therefore constructed fusions of GFP to the SPOR domains of DamX, DedD, FtsN, and RlpA. These fusions were targeted to the periplasm by using a Tat signal sequence from TorA (4). As expected, septal localization was readily apparent in every case, and there was a tendency for production of GFP-SPOR domain constructs to interfere with cell division (Fig. 3A). Septal localization depended on the presence of a SPOR domain because ^TTGFP (Tat-targeted GFP without a SPOR domain) exhibited only uniform periplasmic localization and did not impair division (data not shown, but see references 4 and 63). It should be noted that for the SPOR domains of DamX, DedD, and RlpA, localization was observed in null mutants, so the possibility that full-length DamX, DedD, or RlpA is needed to recruit the cognate SPOR domain to the septal ring can be excluded.

FIG. 3.

SPOR domains localize to the septal ring and bind PG sacculi. (A) Fluorescence micrographs of live cells producing the indicated GFP fusion protein from a plasmid (left) and frequencies of septal localization (in percentages for ≥200 cells analyzed) in these cells (right). Cells were grown to an OD₆₀₀ of ∼0.5 in LB medium without IPTG. Bar, 5 μm. The strains shown are EC251/pDSW992 (expressing the fusion comprising GFP and the SPOR domain of FtsN [GFP-FtsN^SPOR]); JW3351/pDSW997 (expressing GFP-DamX^SPOR); JW3578/pDSW994 (expressing GFP-DedD^SPOR); JW0628/pDSW995 (expressing GFP-RlpA^SPOR); EC251/pDSW1069 (expressing GFP-CHU2221^SPOR); and EC251/pDSW1067 (expressing GFP-AQ1896^SPOR). (B) Analysis of GFP fusion proteins by Western blotting. Whole-cell extracts were subjected to SDS-PAGE, transferred onto nitrocellulose, and probed with anti-GFP. Molecular mass markers in kilodaltons are indicated to the left. The predicted masses of the various GFP-SPOR domain fusion proteins after export range from 36 to 38 kDa. The strains analyzed are as listed in the legend to panel A except that (−) indicates EC251 and GFP corresponds to EC251/pDSW962. (C) PG binding assay. Purified proteins were mixed with purified E. coli sacculi and subjected to ultracentrifugation to pellet the sacculi, which were washed and centrifuged again. Samples of the supernatant, the wash fluid, and the final pellet were analyzed by SDS-PAGE and Coomassie staining to determine what fraction of the input protein was in the final pellet. Bars indicate the averages and standard deviations of results from three independent experiments. MBP, maltose binding protein.

To determine how general this phenomenon might be, we investigated the localization of SPOR domains from two very distant relatives of E. coli. The organisms chosen were C. hutchinsonii, a soil bacterium belonging to the Flavobacteria-Bacteroides group (roughly as distant from E. coli as gram-positive organisms like B. subtilis), and A. aeolicus, which grows in hot springs (optimal temperature, 80°C) and is a member of the deepest branch in the domain Bacteria (16, 71), although there is evidence for extensive lateral gene transfer (5). Remarkably, both foreign SPOR domains localized efficiently and impaired cell division somewhat when produced in E. coli (Fig. 3A). Production of each SPOR domain construct was verified by Western blotting (Fig. 3B).

Many SPOR domains bind PG.

We favor the notion that SPOR domains localize to division sites because they bind preferentially to septal PG (see Discussion). Nevertheless, to our knowledge, only the SPOR domains from FtsN proteins of E. coli and C. crescentus have actually been shown to bind PG (49, 65). In both cases, binding was demonstrated by showing that the purified SPOR domain copelleted with isolated PG sacculi upon ultracentrifugation. The SPOR domain from CwlC of B. subtilis has been reported previously to bind PG on the basis of nuclear magnetic resonance chemical shifts in certain protein residues elicited by the addition of PG fragments obtained by digesting sacculi with an amidase (48), but our analysis of the SPOR domain of DamX suggests that those chemical shifts were misinterpreted (K. Williams, A. Fowler, and D. Weiss, unpublished data).

To determine whether PG binding is a general property of SPOR domains, we purified His₆-tagged SPOR domains from DamX, DedD, RlpA, and the VPA1294, CHU2221, and AQ1896 proteins. The purified proteins were mixed with isolated PG sacculi and subjected to ultracentrifugation. Approximately 80% of the SPOR domain protein was recovered in the PG pellet (Fig. 3C), whereas all of the protein remained in the soluble fraction when the sample was centrifuged in the absence of PG (data not shown). As a negative control, we performed similar pulldown assays with maltose binding protein and FtsZ. In both cases, only ∼20% of the protein copelleted with PG (Fig. 3C), and neither protein was detected in the pellet when the sample was centrifuged in the absence of PG (data not shown). We conclude that the SPOR domains under study are indeed PG binding proteins, as their affinities for PG in the pulldown assay are much higher than expected for proteins that bind PG only nonspecifically.

Placement of DamX in the recruitment hierarchy.

Studies of Fts protein localization in cells with various fts mutant backgrounds have revealed a remarkably linear set of dependencies in E. coli (29, 66, 67). Although the physiological and mechanistic significance of these observations remains unclear, it is possible that the dependencies reflect, at least in part, a pathway for assembly of proteins into the septal ring. The dependencies also suggest which proteins are likely to interact with each other.

We introduced our IPTG-inducible gfp-damX fusion into the chromosomes of a set of strains in which division proteins other than DamX could be inactivated or depleted. Cells were grown under both permissive and nonpermissive conditions and immobilized on agarose pads, and fluorescence microscopy was used to assess the localization of GFP-DamX. The results are presented in Fig. 4 and Table 3. Inactivation of FtsZ prevented recruitment of GFP-DamX to potential division sites (Fig. 4A; Table 3). Western blotting revealed that the ftsZ84(Ts) mutant had appropriate amounts of GFP-DamX (Fig. 4B), so the lack of fluorescent bands can be attributed to a localization defect rather than a problem with the stability of the fusion protein. GFP-DamX localization was readily apparent in cells of all other mutant backgrounds tested (Fig. 4A; Table 3), including ftsA12(Ts) cells, although the fluorescent bands were sometimes ill defined, as if GFP-DamX was efficiently recruited to potential division sites but the septal rings were not well organized. These results imply that DamX fits into the dependency hierarchy at the same stage as FtsZ binding proteins such as FtsA, ZipA, and ZapA. Although we did not explicitly test the dependency of downstream division proteins on DamX, the fact that damX mutants are normal in length makes it very unlikely that DamX plays an important role in recruiting any of the essential downstream division proteins to the septal ring.

FIG. 4.

Dependency of GFP-DamX localization on other septal ring proteins. (A) Fluorescence micrographs of live cells with the indicated genotypes or phenotypes grown under permissive conditions (30°C for temperature-sensitive mutants, with arabinose for the FtsN depletion [FtsN^DEP] strain) or nonpermissive conditions (42°C for temperature-sensitive mutants, with glucose for the FtsN depletion strain). IPTG was used at 0.1 mM in all cases to induce expression of gfp-damX. Bar, 5 μm; WT, wild type. (B) Western blot showing steady-state levels of GFP-DamX in cells pictured in panel A. Molecular mass markers in kilodaltons are indicated to the left of the blot. The black arrow points to the band corresponding to the GFP-DamX fusion. Strains used for analyses presented in panels A and B were as follows: WT, EC2138; ftsZ(Ts), EC2140; ftsA(Ts), EC2142; ftsQ(Ts), EC2144; ftsI(Ts), EC2146; and FtsN^DEP, EC2148. (C) Localization of GFP-DamX in a damX dedD double mutant. Phase-contrast and fluorescence micrographs of strain EC2177 (ΔdamX dedD<>kan P₂₀₄::gfp-damX) showing that GFP-DamX localization is independent of DedD and that gfp-damX rescues division when induced with 0.01 mM IPTG but causes striking filamentation in cells with this background grown with 0.1 mM IPTG. Bar, 5 μm.

TABLE 3.

Localization of GFP-DamX in strains with fts mutant backgrounds

Strain backgrounda	Temp (°C)	NaClb	Sugarc	No. of cells evaluated	Avg length (μm) (SD)	Total no. of rings	% of cells with ring(s)	Ring spacingd
WT	30	+	−	164	4.1 (1.1)	136	82	5
	42	−	−	178	4.5 (1.6)	233	92	3
ftsZ(Ts)	30	+	−	258	7.7 (2.2)	220	76	9
	42	−	−	194	34.2 (18.5)	7	4	947
ftsA(Ts)	30	+	−	168	7.1 (2)	217	96	6
	42	−	−	160	33.5 (7.3)	446	98	12
ftsQ(Ts)	30	+	−	128	13.9 (5.9)	339	99	5
	42	−	−	126	28.9 (14.2)	350	90	10
ftsI(Ts)	30	+	−	134	7.8 (2.8)	145	94	7
	42	−	−	117	29.8 (9.3)	368	97	10
FtsNDEP	30	+	Ara	208	5.4 (1.5)	187	83	6
	30	+	Glu	155	36.6 (15.2)	744	99	8

^aStrains used were EC2138, EC2140, EC2142, EC2144, EC2146, and EC2148. WT, wild type; FtsN^DEP, FtsN depletion.

^bEither 1 or 0% NaCl.

^cEither 0.2% l-arabinose (Ara) or 0.2% d-glucose (Glu).

^dTotal length of cells (or filaments) divided by total number of GFP-DamX rings. This normalization procedure facilitates comparison of wild-type and mutant strains.

Two other noteworthy observations emerged from this set of experiments. First, GFP-DamX caused a mild chaining phenotype in cells with the ftsQ1(Ts) background, as if DamX is an inhibitor of FtsQ (Fig. 4A). Second, gfp-damX eliminated the division defect of a ΔdamX dedD<>kan mutant, indicating that the gfp-damX fusion is functional (Fig. 4C). However, rescue from the defect was observed only at low levels of gfp-damX expression. If IPTG was increased to 0.1 mM to drive higher-level expression, gfp-damX caused striking filamentation in the ΔdamX dedD<>kan mutant. Note that induction of gfp-damX with 0.1 mM IPTG was well tolerated in cells with other genetic backgrounds (Fig. 2A and 4A).

Cell division phenotypes of damX, dedD, and rlpA mutants suggest functional redundancy.

We obtained and characterized damX, dedD, and rlpA null mutants from the Keio collection (2). In LB medium, the dedD<>kan mutant was somewhat elongated and had a slight chaining phenotype (Fig. 5A; Table 4). The gfp-dedD fusion used to demonstrate septal localization abolished both phenotypes (compare the mutant in Fig. 5A to the complemented mutant illustrating GFP-DedD localization in Fig. 2A). The division defects were milder in LB0N medium (data not shown). The dedD<>kan mutant was not sensitive to deoxycholate when plated on LB0N medium (Fig. 5B). This result implies that the outer membrane is intact. Deoxycholate sensitivity in PG hydrolase mutants with defects in septal PG synthesis and cell separation has been reported previously (35). The damX<>kan mutant did not have a visible division phenotype when grown on LB or LB0N medium (Fig. 5A; also data not shown) but was mildly sensitive to deoxycholate on LB0N medium, and the gfp-damX fusion abolished that phenotype (Fig. 5B). The gfp-damX fusion also eliminated the striking division defects (see below) observed in a damX dedD double mutant, provided the fusion was not too highly expressed (Fig. 4C). We were unable to find any division-related phenotype for the rlpA<>kan mutant, which did not elongate, chain, or exhibit sensitivity to deoxycholate (Fig. 5). Owing to the lack of a phenotype due to rlpA<>kan, we could not test whether the RlpA-RFP fusion protein was functional.

FIG. 5.

Phenotypes of damX, dedD, and rlpA mutants. (A) Micrographs of FM4-64-stained cells with the indicated genotypes. The white bar represents 5 μm. Cells were grown in LB medium at 30°C for at least six mass doublings and fixed. The strains shown are BW25113 (WT), JW3351 (damX<>kan), JW3578 (dedD<>kan), JW0628 (rlpA<>kan), EC1926 (ΔdedD damX<>kan), and EC2152 (ΔdamX ΔrlpA dedD<>kan). (B) Sensitivity to deoxycholate. Stationary-phase cultures were adjusted to an OD₆₀₀ of 1.0, and then fourfold serial dilutions were spotted onto LB0N agar plates containing 0.1% deoxycholate and 10 μM IPTG. Strains shown are BW25113, JW3351, EC2130, JW3578, and JW0628.

TABLE 4.

Division phenotypes of mutants lacking damX, dedD, and/or rlpA^a

Strain	Genotype	No. of cells evaluated	Avg length (μm) (SD)	% of cells >10 μm long	% of cells with no. of constrictions:
					0	1	>1
BW25113	Wild type	241	3.7 (1.0)	0	77	23	0
JW3351	damX<>kan	306	3.9 (1.0)	0	78	22	1
JW3578	dedD<>kan	301	6.1 (7.9)	7	58	35	7
JW0628	rlpA<>kan	297	3.9 (1.1)	0	81	18	1
EC1916	ΔdamX rlpA<>kan	310	3.8 (0.9)	0
EC1920	ΔrlpA damX<>kan	335	4.0 (1.1)	0
EC1918	ΔdedD rlpA<>kan	235	6.3 (4.7)	10
EC1922	ΔrlpA dedD<>kan	296	7.0 (4.5)	8
EC1924	ΔdamX dedD<>kan	268	18 (15)	61	45	37	18
EC1926	ΔdedD damX<>kan	286	15 (14)	46	50	34	16
EC2150	ΔdamX ΔrlpA dedD<>kan	379	15 (13)	45
EC2152	ΔdamX ΔrlpA dedD<>kan	284	19 (18)	53	47	35	18
EC2154	ΔdamX ΔdedD rlpA<>kan	273	19 (15)	62	45	32	23
EC2177	ΔdamX dedD<>kan P204::gfp-damX	277	5.0 (1.8)	4
EC251	Wild type	539	3.4 (1.0)	0
EC1910	damX<>kan	528	3.8 (1.0)	0
EC1912	dedD<>kan	467	6.7 (5.1)	7
EC1914	rlpA<>kan	525	3.9 (1.1)	0

^aCells were grown for six doublings to an OD₆₀₀ of ∼0.5, fixed, stained with FM4-64, and photographed using fluorescence microscopy.

These mild phenotypes are not due to the Keio strains' having acquired suppressor mutations, because we observed similar phenotypes after using P1 transduction to move the alleles into cells with the MG1655 strain background from our laboratory collection (Table 4). Moreover, the phenotypes observed before and after eviction of the kan cassette were identical (data not shown).

These findings distinguish DamX, DedD, and RlpA from the only previously studied E. coli SPOR domain protein, FtsN, which is essential, although its SPOR domain is not (13, 26, 65). Could the importance of the new SPOR domain proteins be masked because they compensate for one another? To pursue this possibility, we constructed a redundant set of double and triple mutants. (By redundant, we mean that damX<>kan was introduced by transduction into a ΔdedD mutant and dedD<>kan was introduced into a ΔdamX mutant, etc.) The damX dedD double mutants had striking division defects, with an average length of ∼16 μm and over 15% of the cells exhibiting multiple constrictions (Fig. 5A; Table 4). Adding an rlpA mutation to the mix had no discernible effect on cell length or propensity for chaining (Fig. 5A; Table 4). This pattern is unlikely to be an artifact of the mutants' having acquired suppressor mutations because we also constructed ΔdamX ΔrlpA dedD<>kan/pBAD33-damX depletion strains, which upon extensive depletion, behaved like the simple triple mutants (data not shown).

In the course of these studies, we noticed that the division defects in the mutants were growth phase dependent. The mutants continued to divide in stationary phase, so they were about the same length as wild-type cells after being grown overnight. The cell lengths reported in Table 4 were recorded after cells were grown for six doublings in LB medium at 30°C. Maintaining the cultures in exponential growth by repeated dilution into fresh medium did not result in further elongation (data not shown).

DamX interacts with multiple Fts proteins in a bacterial two-hybrid system.

Using a bacterial two-hybrid system based on reconstitution of an adenylate cyclase (38, 39), we observed that DamX interacts strongly with itself, FtsQ, and FtsN (Fig. 6). Weaker interactions with FtsZ, FtsA, FtsI, and possibly FtsL were detected (Fig. 6). It should be noted that the interactions of DamX with itself and FtsZ could be tested only in one configuration; in the case of FtsZ, this is because the gene is tolerated only in the low-copy-number pKT25 plasmid.

FIG. 6.

Findings from the bacterial two-hybrid assay of DamX interactions with itself and other division proteins. Transformants of DHM1 harboring derivatives of pUT18c and pKT25 were grown in LB medium containing ampicillin and kanamycin plus 0.5 mM IPTG to an OD₆₀₀ of 0.6 to 0.8 and assayed for β-galactosidase activity. Bars represent the means and standard deviations of results from at least three independent experiments. Black bars indicate results for the empty vector control harboring pUT18c and pKT25 (−/−), the positive control harboring pKT25-zip and pUT18-zip (zip/zip), and the homodimerization sample with pKT25-damX and pUT18c-damX (damX). Results for other samples are shown as pairs of gray and white bars reflecting the two potential genetic configurations, pKT25-damX/pUT18c-fts (gray) and pKT25-fts/pUT18c-damX (white). (−), empty vector control; nd, not determined (because the high-copy-number ftsZ plasmid is not stably maintained).

DamX antagonizes FtsQ.

Because cells with a damX null mutation had no obvious division defect, we introduced the mutation into various fts mutant backgrounds by transduction in the expectation that some combinations might prove synthetically lethal or cause cell sickness. The mutations combined with damX<>kan were the ftsZ84(Ts), ftsA12(Ts), zipA1(Ts), ftsQ1(Ts), and ftsI23(Ts) mutations. The double mutants were assayed for viability by spotting serial dilutions onto plates containing LB or LB0N medium, and the plates were incubated overnight at 30, 37, and 42°C. To our surprise, the only synthetic phenotype we observed was rescue of an ftsQ1(Ts) mutant. We got the same result with four independent damX<>kan ftsQ1(Ts) isolates, which argues that the loss of damX rather than an unrecognized suppressor mutation is responsible for amelioration of the temperature-sensitive phenotype. The damX<>kan allele improved the plating efficiency of an ftsQ1(Ts) strain by about 10⁶-fold on LB0N medium at 42°C (Fig. 7A). Moreover, when liquid cultures where shifted from 30 to 42°C, the ftsQ1(Ts) single mutant exhibited more extensive filamentation than the damX<>kan ftsQ1(Ts) double mutant (Fig. 7B). The two strains grew at the same rate after the shift to 42°C, indicating that the double mutant was shorter because it divided more efficiently, not because it grew more slowly. Western blotting revealed that the presence or absence of DamX had no effect on the level of FtsQ1(Ts) protein (Fig. 7C), which was present in very small amounts, as reported previously (6). Efforts to introduce an ftsQ::TnphoA50 (Kan^r) null mutation (7) into a damX strain were unsuccessful, suggesting that the loss of DamX facilitates division by resulting in small amounts of FtsQ rather than by bypassing the requirement for FtsQ altogether (data not shown). This possibility was confirmed by constructing a ΔdamX ftsQ::TnphoA50 (Kan^r)/pBAD33-ftsQ depletion strain. The depletion strain required arabinose for viability, and the terminal phenotype when the strain was grown in the presence of glucose to prevent ftsQ expression was extensive filamentation (data not shown).

FIG. 7.

Loss of damX ameliorates the ftsQ1(Ts) phenotype. (A) Plating efficiency. Tenfold serial dilutions of cells with the indicated genotypes were spotted onto LB0N agar and incubated at 42°C. (B) Cell division. Cells of EC303 or EC1992 were fixed and photographed 1 h after being shifted to LB0N medium at 42°C. Average lengths are based on measurements of 120 cells from each strain. Bar, 5 μm. (C) Western blot showing levels of FtsQ 1 h after shifting of EC251, EC1910, EC303, and EC1992 cells to LB0N medium at 42°C. The 30-kDa mass marker is indicated to the left of the blot.

DISCUSSION

DamX, DedD, and RlpA are new septal ring proteins.

This study has identified three new E. coli septal ring proteins: DamX, DedD, and RlpA (Fig. 2). We have also shown that damX and dedD mutants have division defects, either alone or in combination with mutations in other division genes, demonstrating that DamX and DedD are division proteins (Fig. 5 and 7; Table 4). In the case of RlpA, however, we have so far failed to uncover a phenotype that proves the protein is actually involved in cytokinesis. Follow-up experiments revealed that numerous SPOR domains localize to division sites and bind PG (Fig. 3). Remarkably, even a SPOR domain from Aquifex localizes to division sites when produced in the heterologous host E. coli. This finding argues that septal localization of SPOR domains extends even to the deepest branch in the bacterial tree and that whatever targets SPOR domains to the septal ring is conserved in species from E. coli to Aquifex. Finally, we have characterized the DamX protein in some detail, showing that it interacts with several division proteins in a bacterial two-hybrid system and appears to antagonize FtsQ (Fig. 4, 6, and 7; Table 3).

Implications of SPOR domain localization to septal rings.

The finding that SPOR domains themselves localize to division sites has several interesting implications. First, we infer that many of the >1,500 SPOR domain proteins in the Pfam database are likely to be involved in cell division. Indeed, it was recently shown that SPOR domain proteins localize to division sites in C. crescentus, Burkholderia thailandensis, and Myxococcus xanthus (49). Moreover, we have observed that VPA1294-RFP localizes to division sites when produced in V. parahaemolyticus swimmer cells (R. Kustusch and D. Weiss, unpublished data). Nevertheless, we doubt that all proteins that contain SPOR domains are involved in cell division because, for example, CwlC of B. subtilis has been implicated in sporulation (41, 61). Second, as discussed in more detail below, the SPOR domains studied here probably localize by binding to septal PG, which in turn implies that septal PG is different from lateral wall PG. Following up on this observation may lead to novel insights into PG metabolism during cytokinesis. Third, the SPOR domains are quite divergent at the level of amino acid sequence. An alignment of the seven SPOR domains investigated here revealed that only the amino acids at one position, which corresponds to residue Q251 in FtsN, are identical in all proteins (Fig. 1B). Yet these seven SPOR domains all localize to the midcell. It will be interesting to learn how proteins with such different sequences bind to the same PG structure. Alternatively, should it turn out that different SPOR domains bind to different sites, there must be a multitude of differences between septal and lateral wall PGs.

Several of the preceding remarks are predicated on the inference that SPOR domains localize by binding to septal PG, and we realize that this inference remains to be proven. The current paradigm is that septal ring assembly is driven by a complex web of protein-protein interactions (29, 38, 66). We do not think SPOR domains are likely to conform to this paradigm because it is difficult to imagine what septal ring protein in E. coli could be conserved highly enough to be recognized by SPOR domains from B. subtilis (26), C. hutchinsonii, and A. aeolicus. Genes for widely conserved potential interaction partners such as FtsQ and FtsI are difficult or impossible to identify in the Cytophaga and Aquifex genomes (45; S. J. Arends and D. Weiss, unpublished data). Moreover, the analysis of the SPOR domain from DamX under way in our laboratory indicates that mutations that impair septal localization also impair binding to PG sacculi in the pulldown assay (K. Williams and D. Weiss, unpublished data). For these reasons, we suspect that SPOR domains target a form of PG that is transiently present at the division site, as also suggested by Gerding et al. (26).

This possibility raises the question of how septal PG differs from PG in the lateral wall or at the cell poles. Published analyses have come to different conclusions, but the general consensus appears to be that any differences are minor (17, 36, 53, 56, 60). Nevertheless, there is a clear consensus that the division site is a zone of intense PG synthesis during constriction (19, 69). It is therefore likely that some forms of PG, such as glycan strands that have been extended but not yet cross-linked, pentapeptide side chains, PG with a paucity of covalently attached Lpp, intermediates in PG turnover, and multilayered PG, may be more abundant in septal regions.

Gerding et al. suggested that SPOR domains target a specific PG turnover intermediate, namely, glycan strands that lack peptide side chains (26). Naked glycan strands would arise if amidases ran ahead of lytic transglycosylases during PG degradation. Gerding et al. based this suggestion on their finding that SPOR domains did not localize to septal regions in an E. coli strain devoid of amidases (26) and on a report from Vollmer's group that the purified SPOR domain from FtsN binds to naked glycan strands generated by amidase treatment of isolated sacculi (65). In that study, FtsN′s SPOR domain did not bind to muropeptides or to free peptides (65), although it should be noted that glycan strands that carried peptides were not tested. It is also worth noting that a subsequent paper from Vollmer's group reported binding of FtsN′s SPOR domain to muropeptides (50). The basis for this contradiction was not discussed but may have to do with the fact that in the second study the muropeptides contained 1,6-anhydromuramic acid ends owing to the use of a different enzyme to degrade the sacculi. Finally, to our knowledge, disaccharides indicative of naked glycan strands have never turned up in analyses of the muropeptide composition of PG, implying either that they do not exist or that the analytical methods are not adequate to detect them. Clearly, more work will be needed to test the hypothesis that SPOR domains bind preferentially to glycan strands that lack peptide side chains.

DamX.

The first report on damX noted that overexpression inhibits cell division (43). This characteristic appears to distinguish DamX from other bitopic membrane proteins involved in cell division in E. coli. Why is DamX (apparently) different? One possibility is that DamX's SPOR domain impairs processing of septal PG when the domain is present in excess, as indicated by the fact that GFP-SPOR domain fusion proteins impair division to some extent (Fig. 3). But this hypothesis does not readily explain why excess FtsN is well tolerated. A second possibility is that excess DamX excludes much of the FtsN from the septal ring. A third potential explanation is that DamX sequesters FtsQ. Consistent with this hypothesis, we found that DamX interacted strongly with FtsQ in a bacterial two-hybrid system (Fig. 6), the expression of a gfp-damX fusion in cells with an ftsQ1(Ts) background impaired division even at the permissive temperature (Fig. 4), and a damX null mutation rescued an ftsQ1(Ts) mutant at the nonpermissive temperature (Fig. 7). The same damX null mutation did not rescue any of the other temperature-sensitive mutants tested—the ftsZ84, ftsA12, zipA1, and ftsI23 mutants (data not shown). The damX mutation had no obvious effect on the stability of the FtsQ1(Ts) protein. Taken together, these findings suggest that DamX antagonizes FtsQ function. Nevertheless, the fact that the loss of damX exacerbates the division defect of dedD mutants implies that DamX also contributes positively to cytokinesis (Fig. 5).

Complementary approaches point to DamX being an early recruit in the hypothetical pathway for assembly of the septal ring. Roughly 80% of the cells in a population growing in LB with a doubling time of ∼35 min exhibited septal localization of DamX-GFP (Fig. 2A; Table 3). These frequencies are similar to what has been reported for FtsZ, FtsA, and ZipA. For comparison, late recruits such as FtsQ, FtsI, and FtsN typically localize to the midcell in only 40 to 50% of the population. We also observed that recruitment of DamX-GFP to the septal ring required FtsZ but not FtsA or proteins downstream of FtsA (Fig. 4; Table 3). These findings, together with our observation of weak interactions of DamX with FtsZ and FtsA in the two-hybrid system (Fig. 6), raise the intriguing possibility that DamX links the Z ring directly to the PG layer. In this context, it is worth noting that the cytoplasmic domain of DamX is predicted to be both large (Fig. 1A) and highly charged—103 aa, with 34 Asp or Glu and 24 Arg or Lys.

One context in which the SPOR domain of DamX may cause misinterpretation of experimental results is the two-hybrid system. We observed that DamX interacts with itself and FtsN (Fig. 6). This pattern may reflect colocalization to a PG site rather than direct protein-protein interaction, although if this is the case, it is interesting that DamX did not interact with DedD. The DedD fusions used in the two-hybrid study are at least partially functional in the sense that they interact with some Fts proteins (Williams and Weiss, unpublished).

DedD.

Of the three new septal ring proteins described here, DedD is arguably the most important in that a simple dedD null mutant had discernible division defects whereas mutants lacking either damX or rlpA were normal in length (Fig. 4). DedD had interesting genetic interactions with DamX. The deletion of damX in a dedD background accentuated the division defect, as did the expression of a gfp-damX fusion. One potential explanation is that DedD protects FtsQ from DamX by competing with DamX for localization to the septal ring, perhaps because DedD and DamX target the same sites on PG. Nevertheless, DedD did not interact with DamX in the two-hybrid system (Fig. 6), even though such an interaction would be expected if DedD and DamX are attracted to the same PG sites.

RlpA.

RlpA is perhaps the most enigmatic of the E. coli SPOR domain proteins because null mutants have no obvious phenotype and even combining an rlpA mutation with mutations in damX and dedD had no discernible effect. Nevertheless, we doubt that RlpA is an innocent bystander at the septal ring. The most obvious role for RlpA would be to link the outer membrane to the PG during constriction. As this role is also fulfilled by Pal (27) and perhaps by other proteins such as OmpA and Lpp, it is possible that synthetic phenotypes will emerge if rlpA mutations are combined with mutations affecting these other outer membrane proteins.

Gerding et al. observed localization of RlpA both to the septal ring and to foci at various cites along the cell cylinder (26), whereas we observed RlpA only in septal rings. We think it is important to note that our charge-coupled device camera is inferior to theirs for imaging of RFP. A role for rlpA in elongation would make sense because it resides on the chromosome next to other genes that encode proteins involved in morphogenesis, namely, pbpA, rodA, and dacA (62).