Molecular Microbiology (1996) 19(5), 1047–1060
Structure and function of the Bacillus SpoIIE protein and its localization to sites of sporulation septum assembly Imrich Bara´k,1,† Jaideep Behari,1 Gabriela Olmedo,2 Plinio Guzma´n,2 David P. Brown,1 Elda Castro,2 DeEtte Walker,1 Janet Westpheling1 and Philip Youngman1* 1 Department of Genetics, University of Georgia, Athens, Georgia 30602, USA. 2 Departamento de Ingenierı´a Gene´tica, CINVESTAV, Unidad Irapuato, GTO, Mexico.
tum, perhaps determining the special properties of the structure that permit intercompartment signalling during development.
Introduction At stage II in the morphogenetic programme of sporulation in Bacillus species the developing cell is partitioned into compartments of unequal size as the result of an asymmetric septation event (Piggot et al., 1994). Except for the fact that the sporulation septum contains significantly less peptidoglycan, it resembles the symmetrically positioned septum that defines the location of the cell-division cleft during growth by binary fission (Illing and Errington, 1991). At least some of the genes required for synthesis of the cell division septum are also required for synthesis of the sporulation septum (Lutkenhaus, 1994), indicating that both septum types may be fundamentally similar in their structure and mode of assembly. Nevertheless, the sporulation septum clearly has properties which distinguish it in structural and functional terms from the division septum. The division septum is a participant in cytokinesis, with the peptidoglycan component splitting bilaterally as daughter cells separate from one another. In contrast, the sporulation septum is a participant in the engulfment process, during which most or all of the peptidoglycan that is initially detectable by electron microscopy within the septum is removed (Holt et al., 1975). Asymmetric septation at stage II coincides with, and may actively initiate, the establishment of separate programmes of gene expression in the two sporangium compartments (Losick and Stragier, 1992). The foresporespecific programme is apparently initiated by the activation of sF, a sigma factor which is present in both compartments but which is held inactive in the mother cell through an association with the anti-sigma factor SpoIIAB (Schmidt et al., 1990). In the forespore, SpoIIAB is prevented from inhibiting sF by another protein, SpoIIAA, which is an alternative binding partner for SpoIIAB (Duncan and Losick, 1993). Results from in vitro studies have led to the hypothesis that choice of binding partners for SpoIIAB is controlled by the ratio of ATP:ADP in the two sporangium compartments (Alper et al., 1994). How a differential between the mother cell and forespore in ATP:ADP ratio might be created is unknown. Although mutations in several genetic loci are known
Summary Functioning of the spoIIE locus of Bacillus subtilis is required for formation of a normal polar septum during sporulation and for activation of the transcription factor rF, which directs early forespore-specific gene expression. We have determined the DNA sequence of the wild type and several mutant alleles of the spoIIE gene of B. subtilis and sequenced a substantial portion of its presumptive homologue in Bacillus megaterium. We show that the spoIIE locus encodes a single large protein with a predicted molecular mass of 92 kDa. Each of five point-mutation alleles, which have traditionally defined the locus, and two transposon-generated mutations were shown to fall within the coding sequence for the 92 kDa gene product or within sequences expected to be required for its expression. The amino-terminal portion of the predicted SpoIIE gene product, comprising approximately 40% of the protein, is extremely hydrophobic and is expected to contain up to 12 membrane-spanning segments. The remainder of the protein contains no hydrophobic segments long enough to span a lipid bilayer and is therefore presumed to comprise one or more globular, aqueous-phase exposed domains. An in-frame fusion joining the 3' end of the B. megaterium spoIIE coding sequence to the 5' end of gfp, a gene encoding the green fluorescent protein (GFP) of Aquorea victoria, resulted in a strong, sporulationspecific fluorescent signal localized to the sites of sporulation septum assembly. We speculate that SpoIIE plays a role in assembling the sporulation sepReceived 30 July, 1995; revised 10 October, 1995; accepted 13 October, 1995. †Present address: Institute of Molecular Biology, Slovak Academy of Sciences, Bratislava, Slovak Republic. *For correspondence. E-mail
[email protected]; Tel. (706) 542 1417; Fax (706) 542 1417.
# 1996 Blackwell Science Ltd
g
1048 I. Bara´k et al. that block development from progressing beyond stage II, mutations in the spoIIE locus are unique in two respects. First, the asymmetric septum formed in a spoIIE mutant is abnormally thick, resembling more closely the thickness of a typical division septum (Illing and Errington, 1991). Second, activation sF (Margolis et al., 1991) but not its expression (Errington and Mandelstam, 1986) is prevented in a spoIIE mutant. A mutation in spoIIAB can restore activity of sF in a spoIIE mutant, however, raising the possibility that SpoIIE function might be required, directly or indirectly, for creating a differential in the ATP:ADP ratio between the sporangium compartments (Alper et al., 1994; Margolis et al., 1991). To begin to understand SpoIIE function, we have determined the DNA sequence of the wild type and several mutant alleles of the Bacillus subtilis spoIIE gene and a substantial portion of the homologue of this gene from Bacillus megaterium. We found that SpoIIE is a relatively large protein which is probably anchored in a cellular membrane by means of a hydrophobic amino-terminal domain. Using antibody raised against the carboxy-terminal domain, we obtained preliminary evidence for association of SpoIIE with the sporulation septum in B. subtilis. Using a SpoIIE–GFP fusion in B. megaterium, we have obtained clear evidence for the localization of SpoIIE to actual or potential sites of septum synthesis during sporulation. We speculate that the primary function of SpoIIE may be to play a role in septum assembly, and that SpoIIE may be responsible for the structural and functional properties of the sporulation septum which distinguish it from the division septum.
Results
Comparison of the B. subtilis and B. megaterium spoIIE gene products The spoIIE locus is contained within a 180 kb segment of the B. subtilis chromosome recently sequenced by Ogasawara et al. (1994). Our independent determination of 2920 bp of DNA sequence spanning what we have determined to be the functional boundaries of the spoIIE locus (Guzma´n et al., 1988) shows 100% identity to that sequence (GenBank Accession Number U26835). To help identify the most highly conserved regions within the C-terminal domain of the spoIIE gene product and to facilitate the construction of in-frame fusions involving the spoIIE homologue of B. megaterium, we also determined the DNA sequence of a 1780 bp segment of the B. megaterium chromosome spanning the 3' end of the coding sequence (GenBank Accession Number U26836). This segment is estimated to include about 70% of the coding sequence of the B. megaterium spoIIE homologue and was retrieved from the B. megaterium
Fig. 1. Alignment of the peptide sequences deduced by translation from the sequenced portions of the B. subtilis and B. megaterium spoIIE homologues. Peptide sequence files compiled from translation of the entire B. subtilis coding sequence and the last 585 codons of the presumptive B. megaterium homologue were subjected to a pairwise comparison and alignment by the method of Needleman and Wunsch (1970) using the program GAP (Genetics Computer Group) with a Gap Weight of 3.0.
chromosome by polymerase chain reaction (PCR) amplification, taking advantage of the presence of another open reading frame downstream from spoIIE known to be conserved between B. subtilis and B. megaterium (see the Experimental procedures). An alignment of the peptide sequences deduced by translation from the sequenced portions of B. subtilis and B. megaterium spoIIE homologues is given in Fig. 1. Overall, the aligned sequences are identical at 65% of positions and include 11 separate regions consisting of eight or more contiguous identical residues. The basis for deducing the probable translation initiation site for the B. subtilis peptide is explained below.
Interpretation of the spoIIE sequence The B. subtilis spoIIE locus includes a single large open reading frame spanning nearly the entire region deter-
# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 1047–1060
Structure and function of SpoIIE 1049 mined previously to contain the functional boundaries of the locus (Guzma´n et al., 1988). This open reading frame has a potential coding capacity of 840 amino acids (aa), if actual translation were to begin at the first possible initiation codon. However, based upon our previous determination of the transcription start site (Guzma´n et al., 1988) and subsequent analysis of mutations that affect promoter activity (York et al., 1992), and constrained by the requirement for an appropriately positioned sequence that might serve as a ribosome-binding site, we conclude that the most likely start site for expression of the SpoIIE protein in vivo would be at the 23rd codon (an ATG) in the putative SpoIIE open reading frame (ORF), as indicated by header annotations in the GenBank sequence file (U26835). Therefore, we predict that the expressed product of the spoIIE locus would be a peptide of 827 aa residues having a molecular mass of about 92 kDa. Our assignment of codon 23 of the spoIIE ORF as the likely start site for translation is also supported by our analysis of a mutant allele of spoIIE (spoIIE180), which we interpret as a deleterious substitution in the ribosome-binding site (see below). The spoIIE coding sequence is preceded upstream by a tRNA gene which appears to be expressed independently of spoIIE (Levin and Losick, 1994), and the promoter region responsible for regulated expression of spoIIE was the subject of previous studies (Guzma´n et al., 1988; York et al., 1992). Centred 7 bp downstream from the end of the spoIIE coding sequence is a sequence of hyphenated dyad symmetry which could serve as a transcription terminator. Beginning 87 bp downstream from the spoIIE coding sequence in B. subtilis and 71 bp downstream from its counterpart in B. megaterium is another gene, referred to previously in B. megaterium as orf32 (GenBank Accession Number X64338), which appears from an examination of its sequence to have its own promoter and ribosome-binding site. Therefore, although no attempt has been made to map the 3' end of the spoIIE transcript, it seems likely that spoIIE is a monocistronic gene.
Predicted features of the SpoIIE gene product Beginning about 37 aa residues from its predicted amino terminus and extending through residue 324, the deduced peptide sequence of SpoIIE is extremely hydrophobic. The hydrophobicity profile of this region of the protein displayed on a Kyte–Doolittle scale is shown in Fig. 2A. That this hydrophobic portion of the peptide can function as a membrane-spanning segment in bacteria is supported by the fact that PhoA fusions to certain sites within the spoIIE coding sequence are strongly expressed in Escherichia coli (G. Olmedo, unpublished). The remaining Cterminal 60% of the deduced peptide sequence of SpoIIE
# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 1047–1060
contains no extended stretches of hydrophobic residues, and is therefore predicted to comprise one or more globular domains extending into the aqueous phase. A hypothetical topological model for the SpoIIE protein is shown in Fig. 2B. In this model, the highly hydrophobic portion of the SpoIIE amino terminus comprising residues 37 through to 324 is suggested to consist of 12 membranespanning segments, with both the short hydrophilic Nterminal tail and the putative C-terminal globular domain(s) presented to the interior side of the plasma membrane. This hypothetical topology was found to be a satisfactory fit to our determination of residue hydrophobicities (Hopp and Woods, 1981), surface probabilities (Emini, 1985) and a good accommodation to the ‘positive inside’ rule of von Heijne (1992). In this model, the C-terminal globular domain of SpoIIE (residues 325–827) would be a polypeptide of 57 kDa with a pI of 5.56. A model in which the amino-terminal domain contains 10 membrane-spanning segments also provides a reasonable fit with available information and theoretical considerations (not shown).
Known mutant alleles of spoIIE map to the presumptive spoIIE coding sequence The spoIIE locus has been the subject of considerable previous genetic analysis (Piggot and Coote, 1976) and is defined by a collection of mutations which were reported in some previous studies to be associated with different phenotypic classes distinguishable at the level of electron microscopy (Waites et al., 1970), perhaps indicating that spoIIE actually consists of a cluster of closely linked genes (Hranueli et al., 1974; Piggot, 1973). To determine whether the previously characterized mutant alleles of the spoIIE locus available to us mapped within the DNA interval corresponding to the long open reading frame that we have designated as the spoIIE coding sequence, we carried out a series of marker-rescue integrative recombination experiments using integrational vectors containing different portions of the coding sequence. The vectors were used to transform competent cells of strains containing the alleles spoIIE20, spoIIE21, spoIIE48, spoIIE64, spoIIE71, spoIIE172, and spoIIE180 and transformants were scored for the percentage of Spo+ segregants. The chromosomal segments cloned into these vectors each included one or the other end of the gene and extended various distances into the coding sequence. Therefore, we could deduce from the numbers of Spo+ segregants obtained not only whether a given mutation fell within the coding sequence, but also its approximate location. The results (Fig. 3) indicated that all of the previously studied mutant alleles of spoIIE are within, or in one case (spoIIE180) very near to, the beginning of the coding sequence. The results obtained with spoIIE180 suggested
7
+
Fig. 2. Topological model of the B. subtilis SpoIIE protein. A. Hydrophilicity profile of the first 340 residues of B. subtilis SpoIIE on a Kyte–Doolittle Scale (1982). A sequence file compiled from translation of the first 340 codons of the spoIIE coding sequence was subjected to a graphical hydropathy index calculation using the program MACVECTOR (International Biotechnologies, Inc.) with a window size of seven residues. MACVECTOR output was edited to improve graphic display using the program CANVAS (Deneba Corp.). Numbers above each hydrophobic segment (1–12) correspond to putative membrane-spanning elements in the proposed topological model. B. Hypothetical membrane-spanning topology of the SpoIIE protein. In the interval representing residues 1–330: , residues carrying a positive charge at neutral pH (H,K,R); , residues carrying a negative charge at neutral pH (D,E). Inside the rectangles representing each membrane-spanning segment (Tm1,2 . . . 12) numbers at the top and bottom indicate the first or last residue suggested to be within the membrane-penetrating portion of the segment.
1050 I. Bara´k et al.
# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 1047–1060
Structure and function of SpoIIE 1051
Fig. 3. Schematic representation of the results of marker-rescue integrative recombination experiments carried out to determine the approximate locations of previously isolated mutations that define the spoIIE locus genetically. Top. Physical map of the spoIIE locus showing the positions of landmark restriction sites. Open rectangle, spoIIE orf; numbered rectangles, deduced location of mutant alleles with corresponding numbers. Bottom. Extent of chromosomal DNA cloned into the indicated integrational vector. Right. Proportion of Spo+ transformants among drug-resistant transformants obtained with indicated integrational vector.
that this mutation might be upstream of the actual coding sequence and within a DNA interval essential for efficient expression of the gene, as confirmed by additional work (see below).
DNA sequence of six mutant alleles of spoIIE In the course of our previous characterization of deletions into the spoIIE promoter region, we found that a severely deleterious effect on promoter activity was required to produce a Spo phenotype (Guzma´n et al., 1988). Subsequent analysis of point mutations within the 0A box sites of the spoIIE promoter indicated that as little as 5% of wildtype transcription of this gene will still support substantially normal levels of sporulation (I. Bara´k, unpublished). Therefore, if the previously studied point-mutation alleles of spoIIE included missense mutations, we expected that these mutations might identify regions of the protein critical for function. To determine the specific nature of 7
Table 1. DNA sequence analysis of spoIIE alleles.
some of the point mutation alleles of spoIIE, the DNA interval to which they were assigned on the basis of integrative marker-rescue experiments was amplified by PCR and the product was sequenced on both strands (see the Experimental procedures). The results are summarized in Table 1. Mutant alleles spoIIE48, spoIIE64 and spoIIE71 each proved to be transitions resulting in missense substitutions. Allele spoIIE20 is a transversion which results in a termination codon. Allele spoIIE180 is a substitution affecting what was predicted to be the ribosome-binding site of spoIIE. If that sequence truly functions as the ribosome-binding site, we estimate that the spoIIE180 mutation would reduce the free energy of association between the site (when transcribed) and its partially complementary sequence in 16S ribosomal RNA from 12.0 kCal M 1 to 7.0 kCal M 1. Two Tn917-generated mutations were also shown previously to fall within the spoIIE locus (Sandman et al., 1987). One of these insertions, spoIIE::Tn917 OHU7, was estimated to be approxi7
7
Allele
Mutated basea/codonb
DNA sequence change
Functional consequence
spoIIE180 spoIIE48 spoIIE71 spoIIE64 spoIIE20 spoIIE::Tn917 OHU181
18/– 1111/361 1855/609 1965/646 2121/698 2345+/772+
GGAGA to GAAGA TCC to TTC GGT to GAT CTT to TTT GGA to UGA Tn insertion
Poor RBSc S to F G to D L to F G to STOP Null
a. Numbered with respect to transcription start (Guzma´n et al., 1988). b. Numbered with respect to inferred translation start (see text). c. RBS = ribosome-binding site.
# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 1047–1060
1052 I. Bara´k et al. Table 2. Complementation of spoIIE alleles with a C-terminal truncated form of SpoIIE. Strain
spoIIE allele
% Sporulationa When Complemented With: no phage SPb::Rsa I– Bgl IIb SPb::Rsa I– Bcl Ib
BD170 PY180 PY394 PY813 PY811 PY812 PY507 PY1165
wt spoIIE::Tn917 OHU7 spoIIE::Tn917 OHU181 spoIIE20 spoIIE21 spoIIE48 spoIIE64 spoIIE71
30
