in Bacillus subtilis - NCBI

1 downloads 0 Views 2MB Size Report
bypass of forespore mutations at loci called bofA and boJB that relieve the ..... amyE::bofA insertion into the bofA spoIIIGAl mutant. A. I. 11. _____. __. __+. +. 1-.
Vol. 174, No. 10

JOURNAL OF BACTERIOLOGY, May 1992, P. 3177-3184

0021-9193/92/103177-08$02.00/0 Copyright © 1992, American Society for Microbiology

Characterization of bofA, a Gene Involved in Intercompartmental Regulation of Pro-o.K Processing during Sporulation in Bacillus subtilis EZIO RICCA, SIMON CUTTING,

AND

RICHARD LOSICK*

Department of Cellular and Developmental Biology, The Biological Laboratories, Harvard University, Cambnidge, Massachusetts 02138 Received 26 December 1991/Accepted 17 February 1992

Sporulating cells of the gram-positive bacterium Bacillus subtilis are partitioned into two cellular compartments called the mother cell and the forespore. Gene expression in the mother cell and the forespore is regulated differentially by the compartment-specific transcription factors and 0G, respectively. Gene expression between the two compartments is also coordinated by a signal transduction pathway that couples the activation of o-K (by processing of its inactive precursor pro-o&K) in the mother cell to orG-directed gene expression in the forespore. To dissect the signal transduction pathway genetically, we previously isolated bypass of forespore mutations at loci called bofA and boJB that relieve the dependence of pro-oK processing on the action of s7G* bofl mutations were previously shown to be allelic to the two-cistron sporulation operon spoIVF, which encodes the pro-o&K-processing enzyme or its regulator. We now report that bofA mutations are located in a small open reading frame of 87 codons that encodes a putative integral membrane protein with three potential membrane-spanning domains. The possibility is discussed that BofA and the SpoIVF proteins form a heteromeric complex in the mother cell membrane that surrounds the forespore and that this complex mediates the intercompartmental coupling of pro-4" processing to events in the forespore. A central feature of the process of spore formation by the gram-positive bacterium Bacillus subtilis is the formation of a septum near one pole of the sporulating cell (the sporangium) that partitions the cell into two cellular compartments, known as the forespore and the mother cell (19, 23). After the initial events of compartmentalization, the septum migrates around and wholly engulfs the forespore, thereby pinching it off as a free protoplast within the mother cell. During later morphological stages, a cortex of cell wall-like material is deposited in the space between the membranes that separate the mother cell from the forespore and a protein shell known as the coat is deposited from within the mother cell around the developing forespore. The forespore and the mother cell each contain a chromosome; however, the two compartments are distinct cell types, and they follow dissimilar programs of gene expression. At intermediate to late stages of development, that is, after engulfment, gene expression in the forespore is directed by a transcription factor called or' (15, 30), whereas gene expression in the mother cell is under the control of a transcription factor called o.K (16). The or' and or factors are bona fide cell-typespecific regulatory proteins in that both their synthesis and their sites of action are confined to one or the other partment (9, 12, 15, 17). Although the two cell types each contain their own

(sigK) to the mature and active transcription factor (5, 20). The primary product of sigK is an inactive proprotein called pro-(oK bearing an NH2-terminal extension of 20 amino acids (5, 20). The processing of pro4-{ in the mother cell and expression of genes under the control of off are prevented by mutations in the structural gene (spoIIIG) for a' and in other

(e.g., spoIIL4 and spoIIIE) whose products are required for cr-directed gene expression in the forespore (5, 20). Activation of a& is additionally prevented by mutations in a particular c.G_controlled gene called spoIVB (that is, mutations in other known, aG_controlled genes do not block o'-directed gene expression) (4). Thus, the dependence of pro4aK on crG is mediated, at least in part, by spoIVB, whose product communicates (directly or indirectly) with the mother cell to permit processing of pro4oK. The existence of an intercompartmental pathway for regulating o.'-directed gene expression is supported by two additional lines of investigation. The first involves the use of a deletion mutation that precisely removes the pro-amino acid coding sequence from sigK (5). The presence of such a deletion-mutated sigK gene is found to relieve the dependence of o-'-directed gene expression on CrG. The use of the deletion mutation also provides a clue as to the function of intercompartmental coupling; removal of the pro-amino acid coding sequence causes o.K-directed gene expression to commence about 1 h earlier than normal, and in otherwise wild-type cells, the presence of the truncated sigK gene partially impairs spore formation and causes the formation of abnormal (germination-defective) spores. Thus, intercompartmental coupling is a checkpoint that coordinates the timing of oa'-directed gene expression with the action of G in the forespore, and correct timing is evidently important for proper morphogenesis of the spore. genes

comcom-

partment-specific transcription factor, gene expression in the mother cell is not independent of gene expression in the forespore. Rather, the action of the two transcription factors is coupled by an intercompartmental regulatory mechanism in which c-directed gene expression in the mother cell is dependent on cG-directed transcription in the forespore (5). The coupling mechanism operates at the level of the proteolytic conversion of the product of the aoK structural gene

*

The second kind of evidence involves the isolation of mutations that b pass the dependence of o.K-directed gene expression on (5). Such bypass of forespore (bof) muta-

v7

Corresponding author. 3177

3178

RICCA ET AL.

J. BACTERIOL.

TABLE 1. Strains used in this study Strain

Genotype or description

PY79 SC500 SC572 SC742 SC1329 SC1333 SC1334 SC1337 SC1338 ER28 ER30 ER38 ER45 ER46 ER52 ER77 ER78 ER80 ER88 ER91 ER92

Prototrophic spoIIIGAI spoIIIA59 bofAl spoIIIGAI bofA3 spoIIIGAl bofA9 spoIIIGAl bofA4 spoIIIGAl bofA5 spoIIIGAl bofA6 spoIIIGAl bofA7 bofA::cat spoIIIGAl bofA3 amyE::bofA amyE::bofA spoIlIGAl bofA3 amyE::bofAA25 spoIIIGAl bofA7 amyE::bofA spoIIIGAl amyE::bofAA25 spoIIIGAl bofA8 spoIIIGAl bofA6 amyE::bofA spoIIIGAI bofA8 amyE::bofA amyE::bofA*C spoIIIGAl amyE::bofA *C spoIIIGAl bofA3 amyE::bofA*C

Source (reference) Laboratory stock Laboratory stock Laboratory stock Laboratory stock This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work This work

(32) (5) (5) (5)

tions map to loci called bofA and bofB. Such bof mutations relieve the dependence of pro-ol processing and the expression of oK-controlled genes on spoIVB, spoIIIG, and other genes whose products are needed for cr'-directed transcription. bofB mutations are allelic with the promoter-proximal cistron (spoIVFA) of the two-cistron sporulation operon called spoIVF (7). Interestingly, spoIVF null mutations cause the opposite phenotype, that is, they prevent pro-_K processing and the transcription of genes under CrK control (5, 20). Our interpretation of these findings is that the SpoIVF proteins constitute the pro-aK processing enzyme or its regulator, that the default state of the SpoIVF proteins is inactive, and that bofB mutations are gain of function mutations that lock SpoIVF in the active state. Because spoIVF is under the control of the earlier-acting transcription factor &-- (7), which appears to be a mother cell-specific transcription factor (9, 11), it is likely that spoIVF is expressed in the mother cell. Thus, pro-o' processing in the mother cell appears to be regulated by a signal transduction pathway in which uG turns on the transcription of spoIlVB, whose product sends a signal from the forespore to the mother cell that stimulates SpoIVF-mediated processing of pro-oK (4). Here, we report on the nature of the bofA locus and its possible role in the pro-oK signal transduction pathway. Our results and those of Ireton and Grossman (14), who identified and characterized the same gene from an independent line of investigation, indicate that bofA is identical to a previously sequenced small open reading frame called orf87 (1). The bofA (orf87) gene is transcribed under the control of cE, and its product is inferred to be an integral membrane protein. We propose that BofA, SpoIVFA, and SpoIVFB form a heterotrimeric complex in the mother cell membrane that surrounds the forespore and that this complex couples the processing of pro_uK to the induction of spolVB in the forespore.

MATERIALS AND METHODS Bacterial strains. The strains of B. subtilis used are listed in Table 1. All strains used in this study were isogenic to the

prototrophic wild-type (Spo+) strain PY79, except for the presence of the indicated mutation. General methods. Competent cells were prepared and transformed by the method of Dubnau and Davidoff-Abelson (10). Selection for Cmr was made on agar plates containing 5 ,ug of chloramphenicol per ml. Sporulation, which was carried out at 37°C, was induced on solid medium with Difco sporulation agar and in liquid by suspension in SM medium by the method of Sterlini and Mandelstam (29). o--directed gene expression was monitored by using a gerE-lacZ fusion, which was introduced into the indicated strains by specialized transduction with a fusion-bearing derivative of phage SP3 (SP3::gerE-lacZ) (6). Samples (1.0 ml each) of gerElacZ-bearing cells to be analyzed for ,-galactosidase activity were collected at the indicated times during sporulation by centrifugation, and the cell pellets were stored at -70°C until the time of assay. The specific activity of P-galactosidase was determined as described by Miller (21) with the substrate o-nitrophenol-p-D-galactosidase (ONPG). DNA manipulations were carried out as described by Sambrook et al. (25). Cloning of bofA. Our initial efforts to clone bofA by cloning the bofA region of the chromosome in Eschenchia coli were unsuccessful, probably because of the presence of the rrnA gene and its strong promoter just downstream of bofA. To circumvent the problem, we used the polymerase chain reaction (PCR) to amplify DNA between the dnaX region and sequences just upstream of the rnA promoter. To prime DNA synthesis, we used synthetic oligonucleotides EO (5'CCAATTACAAAGCGCCAGTAGGCAG-3') and E5 (5'GGGGATCCAAATACATAATATCAATT-3') (Fig. 1). EO is just upstream of a natural EcoRI site in the dnaX coding region, whereas E5 contains an artificial BamHI site (underlined bases) adjacent to the rrnA upstream sequence in the primer. The EcoRI-BamHI segment of DNA amplified with the EO and E5 primers was cloned into pER19, a derivative of pUC19 that contains a cat gene capable of being expressed in B. subtilis, to create pER28 (Fig. 1). Subclones of the EcoRI-BamHI insert in pER28 were also cloned into pER19 to create the plasmid inserts diagrammed in Fig. 1, except that of pER44. The insert in this plasmid was created by PCR with oligonucleotides E5 and E4 (5'-GGGAATTCT TGAAGGAAGACGTGAATT-3'; the underlined bases correspond to an artificial EcoRI site), which corresponds to the 3' end of the recM gene (Fig. 1). The eight bofA mutant alleles were amplified by PCR with oligonucleotides E4 and E5 with chromosomal DNAs from bofA mutant derivatives of PY79 as templates. EO, E5, and E4 were designed from the nucleotide sequences of Alonso et al. (1). Construction of strains bearing wild-type and mutant copies of the bofA gene at the amyE locus. To create strains partially diploid for bofA, we introduced a wild-type copy of the cloned bofA gene at the amyE locus (28). We used the plasmid pDG364 (from P. Stragier) as an amyE insertion vector. The bofA coding region and 180 bp of DNA upstream of the translational start site were obtained from plasmid pER44 (Fig. 1) by cleavage with DdeI (which cuts in or74). The ends were rendered blunt with T4 DNA polymerase, and the fragment was cut with BamHI. The resulting 460-bp fragment was then ligated with plasmid pDG364 that had been cut with HindIll, rendered flush at its ends with T4 DNA polymerase, and cut with BamHI. The resulting plasmid, called pER57, was amplified in E. coli. pER57 was linearized, gel purified, and used to transform strains SC572, SC742, SC1337, SC1338, and ER77 containing the bofAl, bofA3, bofA6, bofA7, and bofA8 mutations, respectively, to

bofA REGULATION OF PRO-ca PROCESSING IN B. SUBTILIS

VOL. 174, 1992

->

11 tIO7

raN

Pstl

MspI

-

-

ll-1

3179

lt-

I

~~~IIorf74 I

rinA

bqfA)

// EcoRI EO

P'vuII MspI

Ddel NMpI E4

___

_

_

+

pER28

pER36

pER44 pER37

pER35

FIG. 1. The bofA region of the chromosome. The figure is a physical map of the bofA (orf87) region of the chromosome. A dotted line is used to indicate the downstream boundary of the recM open reading frame, which overlaps the 5' end of the orf74 gene (1). The horizontal lines below the map identify the inserts in the indicated plasmids. + and indicate whether integration of the indicated plasmids corrected or did not correct, respectively, the bofA3 mutation. The arrows labeled EO, E4, and E5 identify the locations of sequences corresponding to the synthetic oligonucleotides used for the PCR amplification of chromosomal DNA (see Materials and Methods). The arrows above the map indicate the extents and orientations of the nucleotide sequence determinations of DNA templates overlapping the bofA region. Nucleotide sequencing was carried out by the dideoxy chain termination method (27). -

the PCR product with EcoRI and BamHI, the digested PCR product was ligated to plasmid pDG364 that had been cut with EcoRI and BamHI. Ligation at the two BamHI sites created an in-frame fusion of bofA with plasmid sequences, thereby generating a hybrid bofA open reading frame, in which codons 84 through 87 were replaced with codons from the vector DNA specifying the residues R I L S A G R. The resulting plasmid (pER32) was subjected to nucleotide sequence analysis to verify that the PCR synthesis had not introduced nucleotide substitutions and that the ligation between the two BamHI sites had created the expected fusion. Plasmids pER66 containing bofAA25 and pER32 containing bofA *C were linearized, gel purified, and used to transform strains PY79, SC500, and SC742 to resistance to chloramphenicol. The resulting transformants were tested for their Amy phenotype to verify that the insertion had occurred by marker replacement (double) recombination at amyE. Construction of a bofA null mutation. A bofA null mutation

resistance to chloramphenicol. The resulting transformants tested for their Amy phenotype to verify that the insertion had occurred by marker replacement (double) recombination at amyE. bofAA25 was constructed by cutting pER44 at the EcoRI and BamHI sites that flank the insert in the plasmid. The resulting 610-bp fragment was gel purified, one portion of which was digested with Dral and another portion of which was cut with NlaIV. The EcoRI-NlaIV fragment of about 335 bp was then joined to the DraI-BamHI fragment of about 200 bp by blunt-end ligation. The resulting EcoRI-BamHI fragment was ligated to pDG364 that had been cut with EcoRI and BamHI. Nucleotide sequence analysis confirmed that the resulting plasmid (pER66) contained a deletion in bofA that fused codon 2 in frame with codon 28 (Fig. 2). bofA*C was created by PCR with oligonucleotides EO and El (5'-GGG.iAICCGCTLlAATGACGACTAACGC-3'). The underlined bases correspond to a BamHI site which was inserted into the coding region of bofA. After treatment of

were

K

E

S

L

P

E

S

A

E

L

V

E

Y

L

K

A

V

A

E

K

F

Y

F

R

L

Y

A

E

K

R

Q

L

N

K

1

GAAAAGCTTAGAACCGTCTGCTGAAGTGTTATACGAACTGAAAGTAGCTGAAGCGAAkGTATTTTTTTTACTTAAGGGAAGCGAAGCAGCGAAATTTAA

101

AAATCAGCCGGTGGAAGTAAGTGGTCTAAACTCCTGGATCTTCTCATAAGCTTGTACTAGAACAAGCGAAmAOATGAGAAGATTCATGGAGCCTATTTT (bofA6)A A(bofA7, 9) MapI

201

TATTATTGGGATTATTTTAGGACTGGTTATTCTTCTTTTTTTATCAGGTTCAGCAGCGAAGCCTTTAAAGTGGATTGGCATCACAGCTGTTAAATTTGTG

I

S

I

I

K

W

R

I

G

S-D

*

I

L

G

I

V

L

L

F

L

L

S

G

S

A

A

P

K

L

K

M

I

W

I

G

G

A

L

A

L

L

V

C

V

N

M

G

G

F

S

G

L

I

H

V

P

I

N

L

V

A

T

(bofA3,4,5)A

P

E

I

K

V

F

orJ74

F

bofA

V

PvuII. T

T

A

I

S

G

I

L

G

301

GCAGGTGCTTTGCTGCTGGTTTGTGTAAATATGTTTGGCGGCAGTCTTGGCATTCATGTGCCGATTAATCTGGTTACAACAGCTATCAGCGGAATTTTAG

401

_A _-r I I Mea GAATACCCGGAATAGCTGCGTTAGTCGTCATTAAGCAATTTATCATTTAATTGATATTATGTATZfzAaZTAGACAACTGAAGGTGTI&TZQT.AATATACG

I

P

G

I

A

A

L

V

V

I

K

Q

F

I

I

-35

*

(bofA8)AA (bofAl) -35

S 01

-10

TCGCTGATGACGAACAGCTT TTTTGAAAAGAAAATGCTAAAAAGT TGZTTGACAGTAGCGGCGGTAAATGTTATGAAATAAAGTCGCTTAAACGAGCGGT

FIG. 2. Nucleotide sequence of the bofA gene and adjacent DNA. The nucleotide sequence is as reported by Alonso et al. (1 [accession number X17014]), except for the substitution of a G for a T at position 349 (see Results). Deduced amino acid sequences are shown above the nucleotide sequence. The proposed Shine-Dalgarno (S-D) sequence for bofA mRNA is underlined. The indicated MspI site is the upstream functional boundary of the bofA transcription unit (see Results). Nucleotide substitutions corresponding to the bofA mutations indicated in parentheses are in boldface type. The start sites and proposed -10 and -35 sequences for the tandem o--recognized promoters governing rmA transcription are from Ogasawara et al. (22).

3180

RICCA ET AL.

J. BACTERIOL.

was created by inserting a cat gene into the coding region of

bofA. The B. subtilis DNA insert in pER30 (Fig. 1) was released by cleavage at SphI and BamHI sites flanking the insert in the plasmid. The resulting SphI-BamHI fragment was then cloned into plasmid pER54, a derivative of pBR322 deleted of the NruI-PvuII region, so as to remove the PvuII site. The resulting plasmid (pER56) contains a unique PvuII site within the bofA open reading frame, into which was inserted a SmaI cassette containing the cat gene (from plasmid pMI1101). Plasmid pER58, containing the cat insertion in bofA, was linearized and used to replace the bofA gene on the chromosome of strain PY79 by marker replacement (double) recombination. To verify that a double-recombination event had occurred, a PCR analysis with oligonucleotides E4 and E5 was carried out; with chromosomal DNA from strain PY79 as template, a PCR product of about 600 bp was obtained, whereas with chromosomal DNA from strain ER28 as template, a PCR fragment of about 2 kb was obtained, a size that corresponds to that of the bofA region (600 bp) plus that of the cat gene cassette (1,400 bp). RESULTS of mutations Bypass forespore map to only two loci. To investigate whether bofA and bofB (spoIVF) are the only sites of mutations that relieve the dependence of pro-or processing on the action of or', we isolated and characterized 15 additional mutations by previously described methods (5). All 15 mutations rendered spollIG mutant cells Pig+, a phenotype diagnostic of the expression of the oacontrolled cotA gene (8, 26), and advanced their morphological stage of blockage to stage V. Using chromosomally inserted chloramphenicol resistance genes known to be highly linked (by DNA-mediated transformation) either to spoIVF or to bofA as described by Cutting et al. (5), we determined that nine of the mutations were at or near spoIVF and that six were at or near bofA. Thus, no additional bof loci were identified among the 15 presently and 5 previously characterized bof mutations. The bofA gene is a previously identified open reading frame. In previous work, bofB was cloned and sequenced and shown to be identical to the promoter-proximal cistron of the sporulation operon spoIVF. To clone and characterize bofA, we first carried out mapping experiments to define more precisely its position on the chromosome. Previous work had shown that bofA is located at about 20 on the genetic map, between the guaB and abrB loci (5). Additional mapping experiments showed that bofA was highly linked by transformation to the xpaC locus (13), which is immediately adjacent to the rRNA operon, rnLA. The gene order in this region of the chromosome is dnaX-orf1O7-recM-orf74-orf87-

rrnA-xpaC (1, 13) (Fig. 1).

To locate bofA on the physical map of this region of the chromosome, we cloned 1.9 kb of DNA extending from the

3' end of the dnaX gene to the 5' end of rrnA. The 1.9-kb DNA was amplified from the chromosome by PCR and by the use as primers of two oligonucleotides (EO and E5) designed on the basis of the published nucleotide sequence of the dnaX-rrnA region (1) (Fig. 1; see Materials and Methods). The PCR-amplified DNA was then cloned into an integrational vector to create plasmid pER28. Using this recombinant integrational plasmid (pER28), we transformed competent cells of the Pig+ strain SC742 (bofA3 spoIIIGA1) with selection for Cmr (conferred by the integrational plasmid). Cmr transformants were expected to arise by integration of pER28 into the chromosome of SC742 by single-

TABLE 2. Nature of the bofA mutations Position

Base

Strain

Mutation

change

Altered sequence

(bp)a

SC572 SC742 SC1333 SC1334 SC1337 SC1338 ER77 SC1329

bofAl bofA3 bofA4 bofAS bofA6 bofA7 bofA8 bofA9

G-*A G-+A G--A G-)A G-*A G--A G-*A G-*A

Gly--Glu Trp--Stop Trp--Stop Trp-*Stop Shine-Dalgarno sequence Shine-Dalgarno sequence Gly--Arg Shine-Dalgarno sequence

410 273 273 273 172 174 409 174

a See Fig. 2 for base pair positions.

reciprocal (Campbell-type) recombination between the B. subtilis DNA insert in the plasmid and the corresponding region of the chromosome. Because most of the resulting Cmr transformants had been converted to Pig-, we conclude that the insert in pER28 contained the wild-type allele of the bofA3 mutation present in SC742. Further localization of bofA was achieved by transformation of SC742 with subcloned fragments of the 1.9-kb DNA. On the basis of the results that pER30, pER35, and pER44 could rescue the bofA3 mutation and that pER36 and pER37 could not (Fig. 1), we infer that the bofA gene must lie immediately adjacent to rnnA. Thus, bofA seems to correspond to a previously identified, small open reading frame of unknown function, called orf87. To confirm that orf87 is bofA, we cloned and sequenced orf87 from eight strains, each bearing one of the bofA mutations isolated in this study or one of the two bofA mutations (bofAl and bofA3) identified earlier (5) (Table 2). As shown in Fig. 2 and summarized in Table 2, orf87 from the bofA mutants SC742, SC1333, and SC1334 (of which only the last two may have been siblings) contained the same G-to-A transition, which converted codon 29 to a nonsense codon; that from strains SC572 and ER77 contained nucleotide substitutions that converted glycine codon 75 to glutamic acid and arginine codons, respectively; that from strains SC1329 and SC1338 contained identical G-to-A transitions in the putative orf87 ribosome-binding site; and that from strain SC1337 contained a G-to-A mutation in the ribosome-binding site. Because the mutant alleles were obtained by PCR amplification and because PCR frequently introduces nucleotide substitutions, the identities of the bofA mutations were verified by amplifying, cloning, and sequencing each mutant allele more than once. Thus, our original eight bofA mutants represented five different mutations in orJ87, which we call bofAl, bofA3, bofA6, bofA7, and bofA8. Together with our previously reported results (5), Fig. 3 shows that all five mutations had similar effects in relieving the block in the transcription of a acr-controlled gene fusion, gerE-lacZ, caused by spoIIIGA1. Also, in confirmation of previous observations (5), the bofA mutations caused the transcription of gerE-lacZ to commence 30 to 60 min earlier than that observed in spoIIIG+ cells. Our nucleotide sequence analysis also revealed a discrepancy in the sequence of orJ87 from that reported by Alonso et al. (1); namely, the G at position 349 in Fig. 2 had been reported to be a T. As a consequence, Fig. 2 shows a glycine codon (CGC) instead of the originally inferred cysteine codon (IGC) at this position (1). Once again, the discrepancy could not have been due to a PCR artifact in our

bofA REGULATION OF PRO-ouK PROCESSING IN B. SUBTILIS

VOL. 174,- 1992

3181

A 11

I _____

__

__+

1-

+

MEPIFIIGIILGLVILLFLSGSAAKPLKWIGITAVKFVAGALLLVCVNMFGGSLGI

80.

II HVPINLVTTAISGILGIPGIAALVVIKQFII RILSAGR

40-

+

+

B *1

IFM

'a C)

1556-00"y-

-- .....

112 IN t-N -insommem OFM ffz= II II I I li;mlllllllllw N

160'

*X K COOH +

120

C 40IFM

2

3

4

5

6

7

8

9

2- 3'

4

6

_

7

Time (hours) FIG. 3. bofA mutations are recessive. Shown are patterns of gerE-directed I-galactosidase synthesis in spo+ cells (O [strain PY79]), in spoIIIGAl-bearing cells (O [strain SC500]), and in spoIIIGAI-bearing strains containing a bofA mutation (-) or the same bofA mutation and a wild-type copy of the bofA gene (amyE::bofA) inserted at amyE (-). The bofA mutations were bofA3 (A [strains SC742 and ER30]), bofA 7 (B [strains SC1338 and ER46]), bofA6 (C [strains SC1337 and ER78]), and bofA8 (D [strains ER77 and ER80]). The open and filled circles in panels C and D are coincident and are not distinguished.

experiments, because position 349 was a G in all of our sequence determinations of orf87 and its mutant alleles. The deduced amino acid sequence of the predicted bofA product did not reveal any significant similarities to those of proteins present in the GenBank and EMBL data bases. A computer analysis that used the Kyte and Doolittle (18) hydropathy program (window of 21 residues) identified three hydrophobic stretches of amino acids that could correspond to membrane-spanning domains (Fig. 4A; see Discussion). bofA mutations are recessive. To determine whether these bofA mutations were dominant or recessive, we constructed merodiploids containing the mutant bofA allele at its normal position on the chromosome and a wild-type copy of bofA at the amyE locus. For these experiments, we used plasmid pER57, in which a wild-type copy of the bofA gene and a cat gene are flanked by the right and left arms of the amyE gene (see Materials and Methods). We introduced the wild-type bofA gene into a collection of bofA-bearing spoIIIG mutants by transforming each mutant with linearized pER57 DNA and then by selection for Cmr and screening for Amycolonies. Amy- transformants were expected to arise by the replacement of the chromosomal amyE gene with the amyEbofA fragment from pER57 by a double-crossover recombination between the left and right arms of amyE in pER57 and amyE on the chromosome. The resulting partial-diploid strains were lysogenized with an SPI bacteriophage containinggerE-lacZ. Figure 3 shows that the bofA3, bofA6, bofA7, and bofA8 mutations were all complemented by the wildtype bofA gene; that is, in all cases the introduction of the amyE::bofA insertion into the bofA spoIIIGAl mutant

t ,#

Il -1 1imm

OFMI1 NH2

COH

r _.

N x2 lX

I iI

_ .|

+COOH

bofA*C bofA 425 FIG. 4. A model for the arrangement of BofA in the outer membrane that surrounds the forespore. (A) Three potential membrane-spanning domains. Charged amino acid residues are indicated by + and -. The dashes indicate residues absent from the product of the bofAA25 mutant gene. The panel also shows the altered COOH-terminal sequence of the bofA*C mutant gene product. (B) Topological model for the arrangement of BofA in the outer membrane surrounding the forespore. IFM, the inner forespore membrane; OFM, the outer forespore membrane. (C) Topological model for the arrangement in the membrane of the products of the bofAA2S5 and bofA*C mutant genes.

strains abolished transcription of the gerE-lacZ fusion. Similar results were obtained with bofAl (data not shown). Construction of a bofA null mutation. Sequencing of the bofA mutations suggested that the Bof phenotype (that is, the capacity to bypass the requirement for the spoIIIG gene product) might be due to the loss of function of the bofA gene product because some of the mutations were in the putative ribosome-binding site or were nonsense mutations. To verify that the Bof+ phenotype is due to the absence of the BofA protein, we created a null mutation in the bofA gene. As described in Materials and Methods, we introduced the cat gene into a unique PvuII endonuclease restriction site within the bofA coding region. When the bofA::cat insertional mutation was introduced into the chromosome, it caused a Bof phenotype, as judged by the following criteria: (i) bofA::cat spoIHIGA1 mutant cells formed dark-brown (Pig') colonies on Difco sporulation agar plates; (ii) the bofA::cat mutation restored the capacity of spoIIIGAl mutant cells to express gerE-lacZ (Fig. 5); and (iii) the timing of gerEdirected P-galactosidase synthesis in bofA: :cat spoIIIGAl cells was 30 to 60 min earlier than that observed in spoIIIG+ cells (Fig. 5). As with the previously characterized bofA mutations -1 and -3 (5), bofA: :cat, when present in otherwise wild-type cells, caused the formation of irregularly shaped spores that were partially defective in their capacity to

3182

J. BACTERIOL.

RICCA ET AL. 200

01-b

0-b

4i

"4

4i

"4

1

to

160

4.0

C)

49 0 4a 4) .4 0 1 Cd

V "W

C)

co

0

120

1 w V 0 .4

k

0=

4)

09 4i X

0

80

4i9 0

a

co

co

40

I

C)

P-4

CL

CL

2

3

4

5 6

7

8

9 2

3

4

5

6

7

8 9

Time (hours) Time (hours) FIG. 5. Effect of a null mutation in bofA on expression of gerE-lacZ. The figure shows the patterns of gerE-directed P-galactosidase synthesis in spo+ cells (El [strain PY79]), in spoIIIGAlbearing cells (O [strain SC500]), and in cells bearing spoIIIGAl and bofA::cat (- [strain ER28]).

germinate but caused little impairment in the efficiency of spore formation (data not shown). Dissecting BofA function by the use of in vitro-constructed mutations. To further define functionally important features of BofA, we constructed, in vitro, two mutations that altered the structure of the BofA protein. bofAA25 is an in-frame deletion of residues 3 to 28, whereas bofA*C is an alteration of the 3' end of the bofA gene that produces a modified protein in which the four terminal amino acid residues of BofA are replaced with seven heterologous amino acid residues (see Materials and Methods). We introduced the bofAA25 or bofA*C mutation into the chromosome of spoIIIGAl cells at the amyE locus (see Materials and Methods). Next, we measured the level of gerE-directed ,B-galactosidase synthesis in these cells. As shown in Fig. 6, gerE-lacZ expression was not restored, a finding that suggests either that both mutations were recessive to the wild-type bofA gene present on the chromosome or that the mutant genes were functionally unimpaired. To distinguish between these possibilities, we introduced the two mutant genes at the amyE locus of cells containing both the spoIIIGAl mutation and the bofA3 mutation, and gerE-lacZ expression was measured (Fig. 6). The bofAA25 deletion mutation substantially restored the inhibition of gerE-lacZ expression observed in cells bearing wild-type bofA, whereas the bofA *C mutant gene did not. We infer that residues 3 through 28 of BofA are largely dispensable (i.e., the deletion-mutated protein is substantially functional) but that the COOH terminus of BofA is critical for the function or stability of the protein. An alternative but less likely explanation for the ability of the deletion-mutated gene to remain functional is that the NH2 terminally truncated product of the bofAA25 mutant protein is complemented by the COOH terminally truncated product of the bofA3 mutant gene, which contains a nonsense mutation at codon 29 (as described above). Upstream boundary of the bofA transcription unit. In the experiment described in the legend to Fig. 3, a bofAcontaining DNA fragment that had been inserted into the chromosome at amyE was shown to complement bofA mutations. The DNA fragment used in these experiments extended to a DdeI site located 180 bp upstream of the open

FIG. 6. Effect of bofAA25 and bofA*C on expression of gerElacZ. Shown are the patterns of gerE-directed p-galactosidase synthesis in spo+ cells (l [strain PY79]), in spoIIIGAl-bearing cells (O [strain SC500]), and in spoIIIGAI-bearing cells containing bofA3 (U [strain SC742]). In addition, panel A shows enzyme synthesis in spoIIIGAl-bearing cells containing a copy of the bofAA25 mutant gene (amyE::bofAA25) inserted at amyE (0 [strain ER45]) and in spoIIIGAl-bearing cells containing bofA3 and the amyE::bofAA25 insertion (A [strain ER52]), and panel B shows enzyme synthesis in spoIIIGAl-bearing cells containing a copy of the bofA*C mutant gene (amyE::bofA*C) inserted at amyE (0 [strain ER91]) and in spoIIIGAl-bearing cells containing bofA3 and the amyE::bofA*C insertion (A [strain ER92]). The open and filled circles in both panels are coincident and are not distinguished.

reading frame (see Materials and Methods). We infer from this result that the DNA fragment contained a functional bofA gene and, therefore, that the promoter for bofA must precede the open reading frame by no more than 180 bp. To determine the upstream boundary of the bofA transcription unit more precisely, a disruption experiment was carried out in which pER35 (Fig. 1), which contains part of bofA and 75 bp of upstream DNA (extending to an MspI site identified in Fig. 2 that is located near the 3' end of orf74), was integrated by single-reciprocal recombination into the chromosome of the Pig- strain SC500 (spoIIIGAI). If integration of pER35 had blocked transcription of bofA, then the resulting integrants should have exhibited a Pig' phenotype. However, because the integrants were, in fact, Pig-, we infer that the sequences upstream of the translational start site contained in plasmid pER35 were sufficient for normal transcription of the bofA gene. This 75-bp region (Fig. 2) contains at a spacing of 14 bp the sequences CATAag-T and GtcTAaa, which conform well (matching in five out of seven positions) and moderately well (matching in three out of seven positions), respectively, to the consensus -10 (CATACA-T) and -35 (kmATATT, where k is a T or a G and m is a C or an A) sequences for oF-recognized promoters (24). Experiments by Ireton and Grossman (14) directly demonstrate the existence of a o--recognized promoter governing the transcription of bofA from this site. DISCUSSION Mutations that bypass the dependence of pro-o- processing on crG-directed gene expression in the forespore are located at two loci, called bofA and bofB. Bypass of forespore mutations at the B locus were previously shown to be located in the promoter-proximal member (spoIVFA) of the two-cistron sporulation operon called spoIVF (7). Our present investigation shows that mutations at theA locus are alleles of a previously identified open reading frame of 87

VOL. 174, 1992

bofA REGULATION OF PRO-&K PROCESSING IN B. SUBTILIS

codons (called orf87) located immediately upstream of the rRNA operon rrnA (1). Our results also show that the phenotype of bofA mutations is due to the absence of the orf87 gene product, because some of the bofA mutations are apparent null mutations in orf87 and because an in vitroconstructed null mutation of orf87 caused the Bof mutant phenotype. The presence of three stretches of hydrophobic residues in its predicted amino acid sequence suggests that BofA is an integral membrane protein with three membrane-spanning domains (I, II, and III; Fig. 4A). A model for the disposition of BofA in the membrane is presented in Fig. 4B. Because transcription of bofA is under the control of &- (14) and because at least some cr'-directed gene expression is confined to the mother cell (9), we assume in our model that BofA is produced in the mother cell from which it is inserted into the mother cell membrane. As BofA is involved in the coupling of pro-a" processing to events occurring in the forespore, we further hypothesize that the 87-residue polypeptide is located in the mother cell membrane that surrounds the forespore protoplast, that is, the other forespore membrane as depicted in the Fig. 4. Taking into account the asymmetric distribution of positively charged amino acids in membrane proteins of known orientation (31), we infer that BofA is oriented in the membrane with its NH2 terminus in the space between the inner and outer membranes surrounding the forespore and with its COOH terminus exposed in the mother cell (Fig. 4B). Because an in vitro-constructed deletion mutation (bofAA2S) that removes the membrane-spanning domain (I) proximal to the NH2 terminus did not seem to eliminate the function of BofA, the second (II) and third (III) putative transmembrane segments are evidently sufficient for function (Fig. 4C). On the other hand, as evidence that domain III is required for BofA function, bofAl and bofA8, which are expected to introduce charged amino acids into the third membrane-spanning domain, inactivate BofA, that is, they cause a bypass of forespore phenotype. Finally, as evidence that the COOH terminus plays a critical role in BofA function or stability, a substitution of several amino acids (created by bofA*C) at the terminus of the protein also caused a bypass of forespore phenotype (Fig. 4C. How is BofA involved in the coupling of pro- processing to au-directed gene expression? In other work, we have shown that spoIVF is transcribed under the control of a- and that both its A and B cistrons are inferred to encode integral membrane proteins (7). As with BofA, SpoIVFA and SpoIVFB are likely to be located in the mother cell membrane that surrounds the forespore protoplast. Our previous evidence indicated that SpoIVFB is the pro-ca-processing enzyme or its regulator and that SpoIVFA is an inhibitor of SpoIVFB (5, 7). We now suggest that BofA and the two spoIlF-encoded proteins interact with each other to form a heteromeric complex in the membrane and that both BofA and SpoIVFA are involved in inhibiting the action of SpoIVFB. Inhibition of SpoIVFB requires both BofA and SpoIVFA because a bofA or bofB mutation alone is sufficient to relieve the dependence of pro-ao processing on ofG_ directed gene expression. The inhibition of SpoIVFB by BofA and SpoIVFA is normally relieved in response to a signal from the forespore under the control of or. The nature of the signal is not known, but it evidently involves the product of a gene called spoIVB, which is transcribed in the forespore under the control of aG (4). Conceivably, the spoIVB gene product inserts into the membrane of the forespore protoplast, from

3183

which it interacts directly with BofA and/or SpoIVFA to overcome the inhibition of SpoIVFB. Remarkably, bofA (orf87) is the site of a recently isolated mutation called ski4, which was isolated by a strategy unrelated (or seemingly unrelated) to the bypass of forespore phenotype (14). The sporulation gene spoIIJ (kinA) encodes a histidine protein kinase that activates the early-acting sporulation regulatory protein SpoOA via a phosphate relay (2, 3). Mutations in spoIIJ impair sporulation only weakly but cause a more severe block in sporulation when combined with ski mutations, which by themselves have little effect on sporulation (14). Thus, spoIIJ and ski mutations act synergistically to cause a strong block in sporulation. One such ski mutation, ski4, is located in orf87 (bofA). As evidence that the effect of the ski4 mutation is due to the involvement of orf87 in the processing of pro-&r, a deletion mutation that removes the pro-amino acid coding sequence from the oK structural gene (5) also causes a ski phenotype (14). Ireton and Grossman (14) speculate that the severe sporulation defect of spoIIJ ski4 double mutants is due to the effect of these mutations on the timing of morphogenesis; spoIIJ mutations retard the early stages of sporulation, whereas bofA mutations (and other bypass of forespore mutations) advance late morphogenetic events. Genetic studies had identified four genes that are involved in the signal transduction pathway that couples the processing of pro-4r in the mother cell to crG-directed gene expression in the forespore (4, 5, 7). These genes are spoIVB, spoIVFA, spoIVFB, and bofA. The identification of the bofA gene in the present work completes the physical characterization of each of the four signal transduction genes. We are now left with the more daunting challenge of understanding whether and how the products of these genes interact with each other to coordinate gene expression between the two cellular compartments of the sporangium. ACKNOWLEDGMENTS We thank Amy Bosma for her participation in the isolation of additional bof mutations, Karen McGovern for advice on the possible membrane topology of BofA, and L. Kroos and P. Stragier for advice on the manuscript. This work was supported by a grant from the Human Frontier Science Program to R.L. E.R. was supported in part by a fellowship from the Consiglio Nazionale delle Ricerche (IABBAM) of Italy. REFERENCES 1. Alonso, J. C., K. Shirahige, and N. Ogasawara. 1990. Molecular cloning, genetic characterization and DNA sequence analysis of the recM region of Bacillus subtilis. Nucleic Acids Res. 18: 6771-6777. 2. Antoniewski, C., B. Savelli, and P. Stragier. 1990. The spoIIJ gene, which regulates early developmental steps in Bacillus subtilis, belongs to a class of environmentally responsive genes. J. Bacteriol. 172:86-93. 3. Burbulys, D., K. A. Trach, and J. A. Hoch. 1991. Initiation of sporulation in B. subtilis is controlled by a multicomponent phosphorelay. Cell 64:545-552. 4. Cutting, S., A. Driks, R. Schmidt, B. Kunkel, and R. Losick. 1991. Forespore-specific transcription of a gene in the signal transduction pathway that governs pro-cr processing in Bacillus subtilis. Genes Dev. 5:456-466. 5. Cutting, S., V. Oke, A. Driks, R. Losick, S. Lu, and L. Kroos. 1990. A forespore checkpoint for mother cell gene expression during development in B. subtilis. Cell 62:239-250. 6. Cutting, S., S. Panzer, and R. Losick. 1989. Regulatory studies on the promoter for a gene governing synthesis and assembly of the spore coat in Bacillus subtilis. J. Mol. Biol. 207:393-404. 7. Cutting, S., S. Roels, and R. Losick. 1991. Sporulation operon

3184

8. 9. 10. 11.

12.

13.

14. 15. 16. 17.

18. 19.

RICCA ET AL.

spoIVF and the characterization of mutations that uncouple mother-cell from forespore gene expression in Bacillus subtilis. J. Mol. Biol. 221:1237-1256. Donovan, W., L. Zheng, K. Sandman, and R. Losick. 1987. Genes encoding spore coat polypeptides from Bacillus subtilis. J. Mol. Biol. 196:1-10. Driks, A., and R. Losick. 1991. Compartmentalized expression of a gene under the control of sporulation transcription factor (rE in Bacillus subtilis. Proc. Natl. Acad. Sci. USA 88:9934-9938. Dubnau, D., and R. Davidoff-Abelson. 1971. Fate of transforming DNA following uptake by competent Bacillus subtilis. J. Mol. Biol. 56:209-221. Errington, J., D. Foulger, N. Illing, S. R. Partridge, and C. M. Stevens. 1990. Regulation of differential gene expression during sporulation in Bacillus subtilis, p. 257-267. In M. M. Zukowski, A. T. Ganesan, and J. A. Hoch (ed.), Genetics and biotechnology of bacilli, vol. 3. Academic Press, San Diego, Calif. Francesconi, S. C., T. J. MacAlister, B. Setlow, and P. Setlow. 1988. Immunoelectron microscopic localization of small, acidsoluble spore proteins in sporulating cells of Bacillus subtilis. J. Bacteriol. 170:5963-5967. Hulett, M., K. Jensen, C. Bookstein, D. Owen, C. Edwards, and N. Kapp. 1988. In vitro derived phosphatase mutants of Bacillus, p. 325-329. In A. T. Ganesan and J. A. Hoch (ed.), Genetics and biotechnology of bacilli, vol. 2. Academic Press, San Diego, Calif. Ireton, K., and A. D. Grossman. 1992. Interactions among mutations that cause altered timing of gene expression during sporulation in Bacillus subtilis. J. Bacteriol. 174:3185-3195. Karmazyn-Campelli, C., C. Bonamy, B. Savelli, and P. Stragier. 1989. Tandem genes encoding a-factors for consecutive steps of development in Bacillus subtilis. Genes Dev. 3:150-157. Kroos, L., B. Kunkel, and R. Losick. 1989. Switch protein alters specificity of RNA polymerase containing a compartment-specific sigma factor. Science 243:526-529. Kunkel, B., K. Sandman, S. Panzer, P. Youngman, and R. Losick. 1988. The promoter for a sporulation gene in the spoIVC locus of Bacillus subtilis and its use in studies of temporal and spatial control of gene expression. J. Bacteriol. 170:3513-3522. Kyte, J., and R. F. Doolittle. 1982. A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157:105-132. Losick, R., P. Youngman, and P. J. Piggot. 1986. Genetics of endospore formation in Bacillus subtilis. Annu. Rev. Genet.

J. BACTERIOL.

20:625-669. 20. Lu, S., R. Halberg, and L. Kroos. 1990. Processing of the mother-cell cr factor, oK, may depend on events occurring in the forespore during Bacillus subtilis development. Proc. Natl. Acad. Sci. USA 87:9722-9726. 21. Miller, J. H. 1972. Experiments in molecular genetics, p. 352-355. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 22. Ogasawara, N., S. Moriya, and H. Yoshikawa. 1983. Structure and organization of rRNA operons in the region of the replication origin of the Bacillus subtilis chromosome. Nucleic Acids Res. 11:6302-6317. 23. Piggot, P. J., and J. G. Coote. 1976. Genetic aspects of bacterial endospore formation. Bacteriol. Rev. 40:908-962. 24. Roels, S., A. Driks, and R. Losick. 1992. Characterization of spoIVA, a sporulation gene involved in coat morphogenesis in Bacillus subtilis. J. Bacteriol. 174:575-585. 25. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 26. Sandman, K., L. Kroos, S. Cutting, P. Youngman, and R. Losick. 1988. Identification of the promoter for a spore coat protein gene in Bacillus subtilis and studies on the regulation of its induction at a late stage of sporulation. J. Mol. Biol. 200:461-473. 27. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 28. Shimotsu, H., and D. J. Henner. 1986. Construction of a single-copy integration vector and its use in analysis of regulation of the tip operon of Bacillus subtilis. Gene 43:85-94. 29. Sterlini, J. M., and J. Mandelstam. 1969. Commitment to sporulation in Bacillus subtilis and its relationship to development of actinomycin resistance. Biochem. J. 113:29-37. 30. Sun, D., P. Stragier, and P. Setlow. 1989. Identification of a new a-factor involved in compartmentalized gene expression during sporulation of Bacillus subtilis. Genes Dev. 3:141-149. 31. von Heine, G. 1986. The distribution of positively charged residues in bacterial inner membrane proteins correlates with the trans-membrane topology. EMBO J. 5:3021-3027. 32. Youngman, P., J. B. Perkins, and R. Losick. 1984. A novel method for the rapid cloning in Escherichia coli of Bacillus subtilis chromosomal DNA adjacent to Tn917 insertions. Mol. Gen. Genet. 195:424-433.