Printed in Great Britain Microbiology (1 996), 142, 203 1-2040 

The phage-like element PBSX and part of the 
skin element, which are resident at different 
locations on the Bacillus subtilis chromosome, 
are highly homologous 

Susanne Krogh,' Mary O'Reilly,' Niamh Nolan112 and Kevin M. Devine' 

Author  for correspondence : Kevin hi. Devine Tel: + 353 1 6081 872. Fax : + 353 1 6798558. 
e-mail: kdevine@tcd.ie 

Department of Genetics1 
and National Pharmaceutical 
Biotechnology Centrez, 
Trinity College, Dublin 2, 
Ireland 

PBSX and skin are two unusual genetic elements resident on the Bacillus 
subfilis chromosome. PBSX is a phage-li ke element located at  approximately 
100" which is induced by the SOS response and results in cell lysis with the 
release of phage-like particles. The phage particles contain bacterial 
chromosomal DNA and kill sensitive bacteria without injecting DNA. The skin 
element is located at  approximately 230" on the chromosome and is  positioned 
within the sigK open reading frame (ORF). It is excised at  a particular stage of 
sporulation, leading to reconstitution of the complete sigK gene. In this paper, 
we show that there are phage-like operons present in the skin element which 
are highly homologous to the region of PBSX comprising part of the control 
region and the late operon. These operons are similar in terms of their gene 
organization, the percentage identity of the products of homologous ORFs and 
the positioning and strengths of ribosome-binding sites for each ORF. 
Although this high degree of conservation suggests that the phage-like 
operons in skin can be expressed, expression of the late operon was not 
detected during exponential growth, during sporulation or after induction of 
the SOS response. However two non-phage-like operons in the skin element 
are expressed and have distinct expression profiles that are dependent on the 
growth and developmental status of the cell. 

Keywords : PBSX, skzn, high level of  similarity, expression of  skin operons 

INTRODUCTION 

Many species of Bacillus have a prophage-like element 
resident on their chromosomes (Steensma e t  al., 1978). 
Expression of this element is initiated by agents which 
induce the SOS response and results in cell lysis with the 
concomitant release of phage-like particles (Okamoto e t  
a/., 1968; Steensma e t  al., 1978). The particles can be 
distinguished serologically, morphologically by the num- 
ber of cross-striations in the tail, and by their range of 
killing activity. Using these criteria, the phage-like 
elements PBSW, PBSX, PBSY and PBSZ have been 
identified in Bacillus subtilis var. vulgdtus and in strains 168, 
S31 and W23, respectively. Bacillzis lichenifarmis harbours 
an apparently related element PBSV. The phage-like 
particles can bind to and kill bacterial cells which carry 

The GenBank accession number for the sequence reported in this paper is 
2701 77. 

heterologous phage-like elements, whereas the host 
bacterium is resistant to their killing activity. For PBSX 
and PBSZ, the type of teichoic acid resident in the cell 
wall determines the adsorption spectrum of the phage and 
hence their killing selectivity (Glaser e t  al., 1966; Kara- 
mata e t  al., 1987; Young e t  al., 1989). PBSX binds to 
erythritol teichoic acids of €3. szzbtilis strain W23, but 
cannot bind to the glycerol teichoic acids present in the 
cell wall of the host B. subtilis strain 168. The reverse holds 
for the PBSZ phage-like element resident in B. subtilis 
strain W23. The mechanism whereby the bound phage 
kills the cell is not known, but it does not involve 
injection of DNA into the cell. An additional unusual 
feature of PBSX is that the head of the phage-like particle 
is capable of packaging only 13 kb of DNA, even though 
the PBSX genome exceeds 33 kb in size. These 13 kb 
fragments found in the phage-like particle head are 
predominantly B. subtilis chromosomal DNA (Okamoto 
e t  a/., 1968; Haas & Yoshikawa, 1969). From these 

0002-0594 0 1996 SGM 203 1 



S. K R O G H  a n d  O T H E R S  

observations it is evident that PBSX induction is suicidal 
both for the phage and for the bacterial cell. Why then 
does the element persist in bacterial cells? All these 
elements encode proteins which can assemble into a 
phage-like particle, suggesting that it is the particle which 
holds the key to their function. One model proposes that 
they are phage-like bacteriocins. This does not exclude the 
possibility that they also encode genetic functions which 
are beneficial for cell physiology. 

PBSX is the most thoroughly investigated of the phage- 
like elements resident in strains of B. subtilis. Mutations 
affecting induction (xin and xhi1479), the ability to form 
phage-like particles (xhd and xtl), and the killing activity 
(xki 1479) were isolated. These mutations all map between 
the metA and metC loci of the B. subtilis chromosome 
(Garro e t  al., 1970; Thurm & Garro, 1975a; Buxton, 
1976). The proteins synthesized upon PBSX induction, 
and the structural proteins of the phage-like particle, 
range in size from 12 to 76 kDa (Thurm & Garro, 1975b; 
Mauel & Karamata, 1984). The size of the PBSX genome 
is not known precisely. However it has been established 
that part of the regulatory region and the entire late 
operon are approximately 33 kb in size (Wood e t  al., 
1990a). The element is maintained on the chromosome by 
a repressor (xre) which binds to operator sites located in 
the vicinity of the promoters for divergent transcription 
units (Wood e t  al., 1990b; McDonnell & McConnell, 
1994). One of these transcription units contains the gene 
p f ,  which encodes a sigma-factor-like protein necessary 
for transcription of the late operon (McDonnell e t  al., 
1994; McDonnell & McConnell, 1994). The structural 
proteins of the phage-like particle are encoded within the 
late operon. The last four cistrons of the late operon 
encode proteins which are involved in host cell lysis and 
include a holin-like protein and an autolysin (N-acetyl- 
muramoyl-L-alanine amidase ; Foster, 1993 ; Longchamp 
e t  al., 1994; S. I irogh & I<. Devine, unpublished). 

The B. subtilis chromosome encodes a second unusual 
genetic element called skin (sigK intervening sequence), 
located at approximately 230” on the B. subtiljs chromo- 
some. During sporulation, a genetic rearrangement occurs 
in the mother cell compartment only, which results in the 
excision of the skin element from the chromosome. This is 
catalysed by a site-specific recombinase. Expression of the 
spolL/CA gene which encodes the recombinase 
(spolL’CA is located within the skin element) is activated 
at stage 111 of sporulation (Stragier e t  al., 1989; I<roos e t  
al., 1989; Sat0 e t  al., 1990, 1994; Kunkel e t  al., 1990; 
Popham & Stragier, 1992). Excision of the skin element is 
necessary to reconstitute the sigK open reading frame 
(ORF). SigIi then directs mother-cell-specific gene ex- 
pression. The skin element has been completely sequenced 
(Takemaru e t  al., 1995). The products of ORFs 5 , 7 , 8  and 
50 show homology to proteins encoded by phage 4105, 
suggesting that this element is a remnant of a cryptic 
ancestral phage (Takemaru e t  al., 1995). However, skin 
may be a composite element since it also contains genes 
encoding arsenate resistance which are homologous to 
those resident on the Staph3/lococcus aureus plasmid pI258 
(Takemaru e t  al., 1995). Two findings suggest that 

excision of skin is not critical for developmentally 
regulated expression of sigK: (i) strains engineered such 
that the skin element is excised and the sigK ORF is 
contiguous grow and sporulate normally (Kunkel e t  al. , 
1990); (ii) in closely related Bacillus strains, the sigK ORF 
is not interrupted (Adams etal. ,  1991). These data indicate 
that developmental regulation of sigK expression is 
effected by mechanisms which are independent of skin 
excision (Kunkel et al., 1988 ; Cutting e t  al. , 1990 ; Lu e t  al. , 
1990). These observations raise questions about the origin 
of the skin element and the role it plays (if any) in cell 
physiology. 
In this paper, we show that part of the skin element is 
highly homologous to the control and late operon regions 
of PBSX. We also investigate the expression of skin 
operons and show that two operons are expressed with 
distinct profiles which are dependent on the growth and 
developmental status of the cell. The phage-like late 
operon within skin, however, is not expressed during 
exponential growth, during sporulation, or  during the 
SOS response. 

METHODS 

Bacterial strains and growth conditions. Escherichia coli strain 
TP611 (recBC hsdR M 9 a 6 1 0  pin) was used for cloning large 
chromosomal DNA fragments (Glaser e t  al., 1993). E.  coli strain 
tgl  [I<-12 A(lac pro) supE thi hsdR F’ t raD36 proAB lacI lacZ 
AM151 was used for subcloning and for preparation of 
sequencing templates. B. subtilis chromosomal DNA for se- 
quencing was isolated from strain 168. Studies on skin and PBSX 
expression were performed in B. stibtilis strain JH642 and in B. 
subtilis strain SKPOP9, a derivative of strain L8508 (xhi  1479 &t 
2) from which the distal part of the PBSX late operon (from 
fragment 37 to the distal part of fragment 39, which includes the 
lysis genes; Wood e t  al., 1990a) is deleted. B. subtilis and E .  coli 
strains were routinely grown in Luria-Bertani (LB) medium 
(Miller, 1972). Sporulation in B. subtilis was initiated by 
resuspension in Sterlini-Mandelstam medium as described by 
Nicholson & Setlow (1990). The timing of sporulation was 
monitored by measurement of alkaline phosphatase and dipi- 
colinic acid accumulation. B. stibtilis transformation was carried 
out according to the method of Anagnostopoulos & Spizizen 
(1961). E. coli transformation was carried out according to the 
method of Sambrook e t  al. (1989). 
Plasmid and strain construction. Three oligonucleotide primer 
pairs were made (Applied Biosystems PCR-Mate DNA Syn- 
thesizer) and used to amplify segments of skin DNA by PCR 
(their location in the skin sequence is in parentheses): (1) 5‘ 
GAAGGGTGCTACCATTAC 3’ (24423 + 24440) and 5’ 
ATCTCCTCACCTAGATC 3’ (25 472 25 456) ; (2) 5’ 
AGCCCTTCTCATTGAGAG 3’ (38023 + 38040) and 5’ 
GAAGCATCGACTAACACC 3’ (39018 + 39001); and (3) 5’ 
CTGCGTTGCTGTAGTAGC 3’ (43 543 + 43 560) and 5’ 
GTAATATCACTTGGCGG 3’ (44 530 + 44 514). The DNA 
amplified using each pair of primers was isolated, the ends were 
polished and the fragments were subcloned in pUC19 to give 
plasmids pSkinlF, pSkinlR, pSkin2R, pSkin3F and pSkin3R, 
where F and R represent the forward (the sequence in GenBank) 
and reverse orientations relative to sites within the polylinker. 
The integrity of the cloned fragments was checked by se- 
quencing. Insert DNA with non-compatible ends was prepared 
from each plasmid and ligated directionally to the integrating 
plasmid p JM783 (a plasmid conferring resistance to chloram- 

2032 



Prophage-like elements skin and PBSX are homologous 

phenicol and containing a promoterless IacZ gene; a gift from 
J .  A. Hoch) so that transcriptional fusions are generated upon 
integration of the recombinant plasmid into the chromosome of 
B. stlbtilis. The point of fusion in integrant strains S K l F  and 
SK1R (the insertion is directed by the chromosomal inserts of 
pSkinlF and pSkinlR, respectively) is located within 9qbO 
(ORF 54 ; Takemaru e t  al., 1795) ; the point of fusion in integrant 
strain SK2R (directed by the chromosomal insert in pSkin2R) is 
located immediately 3' to they@ (ORF 95; Takemaru e t  ai., 
1995) cistron and the point of fusion in integrant strains SK3F 
and SK3R (directed by the chromosomal inserts in pSkin3F and 
pSkin3R, respectively) is positioned within the j qcK  (ORF 2 ; 
Takemaru e t  a/., 1995) cistron. The location and structural 
integrity of DNA after integration at each locus was verified by 
Southern analysis using pSkin 1F as probe for strains S K l F  and 
SKlR, pSkin 2R as probe for SK2R and pSkin3R as probe for 
SK3F and SK3R. Fragment 37 from PBSX was also used to 
generate a strain, SK37, which carried a lacZ fusion in the late 
operon of PBSX orientated in the direction of transcription of 
the late operon (Wood e t  al., 1970a). The lacZ fusions resident 
in SKlF, SK2R, SK3F and SK37 were transferred into the 
B. stlbtilis strain SKPOP9 by chromosomal transformation, 
generating strains SKPOPl F, SKPOP2R, SKPOP3F and 
SKPOP37. 

Isolation of B. subtilis chromosomal DNA and sequencing 
strategy. DNA was isolated from the chromosome of B. sztbtiiis 
strain 168 by chromosomal walking. A fragment from the 
vicinity of the PBSX xre gene was cloned into the integrating 
vector pDIA5304 (Glaser ef al., 1993), giving plasmid pDIArep. 
This plasmid was then inserted into the chromosome of B. 
subtiik strain 168 (giving strain SKREP) with selection for 
resistance to chloramphenicol (3 pg ml-'), Southern analysis 
was performed on restricted chromosomal DNA from integrant 
strain SKREP using pDIArep as a probe and it was determined 
that digestion with BamHI releases a fragment of 12 kb in size 
adjacent to the insertion site. Chromosomal DNA from SKREP 
was restricted with BamHI, ligated and transformed into E. colz 
strain TP611. Transformants containing the 12 kb chromo- 
somal insert subcloned in plasmid pDIA5304 (pBaml2) were 
identified. This chromosomal walking procedure was repeated, 
giving pSall4, which is contiguous with, and overlaps, pBaml2. 
A 5-5 kb chromosomal fragment, pEco5, was isolated from a 
lambda clone of chromosomal DNA from B. stlbtilis strain 
SO113 (Wood e t  ai., 1990a). Parts of pEco5 were sequenced 
directly using oligonucleotides (S. Krogh, unpublished results). 
This sequence was used to design further oligonucleotides 
through which the sequence of this region was obtained from B. 
sztbtilis strain 168. Southern analysis was performed to verify the 
integrity of pBaml2, pSall4 and pEco5. After sequencing, the 
actual insert sizes were: pBaml2, 13832 bp; pSall2, 11 483 bp 
and pEco5, 5497 bp. The overlap between pBaml2 and pSall2 
is 898 bp and that between pSall2 and pEco5 is 468 bp. 

The isolated chromosomal DNA was sequenced using a shotgun 
strategy (pBam12 and pSall4) and by a directed approach using 
oligonucleotides (pEco5). Plasmid DNA (30 pg in 120 p1 TE 
buffer) (TE buffer is 10 mM Tris/HCl; 1 mM EDTA) was 
randomly sheared either by sonication (Braun Labsonic : 7 
pulses/0.22 cycles/0*25 W) or using DNaseI in the presence of 
manganese (Sambrook e t  al., 1989). The sheared DNA was 
resolved on agarose gels and fragments of 2-5 kb in size were 
isolated. Fragment ends were polished using Klenow enzyme 
and T4 DNA polymerase, ligated to pUC18 and cloned into E. 
coli strain tgl  . Template DNA for sequencing was prepared by 
two methods : (1) the miniprep procedure of Holmes & Quigley 
(1981) and (2) using magnetic beads. In the latter method, 
inserts were amplified using universal forward and reverse 

primers, one of which was biotinylated. An aliquot of amplified 
DNA was then added to 2 p1 of a streptavidin-conjugated 
magnetic bead suspension (Dynabeads ; Dynall AS) and incu- 
bated for 10 min at 37 "C. The beads were then washed 
extensively, resuspended in TE buffer and used as templates for 
sequencing. Sequencing reactions were carried out using 
GENPAK fluorescent dye-primer sequencing kits according to 
the manufacturer's instructions. Reactions were resolved on an 
ABI Automated Sequencer model 373A. Gaps in the sequence 
were filled using custom synthesized oligonucleotides (Applied 
Biosystems PCR-Mate DNA Synthesizer) and the Applied 
Biosystems dye-terminator sequencing kit. 
Computer sequence analysis. Sequence alignment and editing 
were performed using the XBAP program of the STADEN package. 
Conceptual translation of the sequence and other sequence 
analyses were performed using the NIP program of the STADEN 
package. The GenBank database was accessed using ACNUC 
(Gouy e t  al., 1985) and homology searches of this database were 
performed using the TBLASTN program (Altschul e t  al., 1990). 
Multiple sequence alignments were performed using CLUSTALV 
(Higgins e t  al., 1992). 
General molecular biological methods. Minipreps of plasmid 
DNA from E. coli were made according to the method of 
Holmes & Quigley (1981). Restriction analysis, nick translation, 
Southern analysis and other general molecular biological 
methods were performed as detailed in Sambrook e t  al. (1989). 
Enzyme assays. Alkaline phosphatase and dipicolinic acid 
accumulation during sporulation were measured as described by 
Nicholson & Setlow (1990). Expression of iacZ was measured as 
described by Ferrari e t  al. (1986) with the activity being 
expressed in Miller units (Miller, 1972). 

RESULTS AND DISCUSSION 

Sequence comparison between PBSX and skin 

T h e  chromosomal  region of  B. stlbtitis strain 168 extending 
f r o m  xre (100") t o  the vicinity of xpa (103"), encoding par t  
of the control  region a n d  the  entire late operon  of  the 
phage-like element PBSX,  has been sequenced. This  
region of PBSX is approximately 28 kb in  size. Two 
DNA fragments f r o m  within this region of B. stlbtilis 
strain SO113 were previously sequenced:  a 5.5 k b  seg- 
ment comprising par t  of the  PBSX control  region 
(accession number  234287, McDonnel l  et al., 1994) and  a 
2.5 kb segment  encoding the  PBSX lysis genes (accession 
number  L25924, Longchamp e t  al., 1994 ; accession 
number  236941, S. I<rogh & K. Devine,  unpublished). 
T h e  PBSX sequence shows a very h igh  level o f  similarity 
t o  part of the  sequence of  the  skin element located at  230" 
on the  B. sabtilis chromosome.  T h e  similarity between the 
two elements is evident in  their operon  organizations and  
in  the gene content of each operon  (Fig. 1). There  are 
three operons within this region of  PBSX : the  first (which 
for  ease of discussion is called the early operon)  is 
comprised of  xre (the phage repressor) and  x k d A ;  the  
second (called the  middle operon)  contains the genes 
xkdB, xkdC, x k d D ,  x t r A  and xpf(formerlypg,  the sigma- 
factor-like transcription factor) ; a n d  the  third (called the 
late operon)  includes the 25 genes f r o m  x t m A  to xlyA 
which encode the head and tail proteins of the  phage 
particle, the  lysis proteins and  perhaps o ther  phage 
functions. There  are three phage-like operons within the 
skin element, corresponding t o  each of the  PBSX operons, 

2033 



S. K R O G H  a n d  OTHERS 

xtrA 
XB XC xD xof xtmA 

0 kb 4-6 kb PBSX 

XA xre 

bA aC aD aF aG aH al aJ aK aL aM aN a 0  a0 aR aS aT 

0 kb 15-0 kb skin 

XF XF XG xH X /  XJ xK X M  xN xo  XP XQ 

hA hR hC hD bE bFbG bH bl bl bK bL bM bN bO bP 
~~ - - -. - - -. . . . . . -. . - - - - - - 

I t t  30 kb skin 
............. ............ ............. ............. .. :.:.: ...:..:. :.:.:.:j ... .. ..................... . ............ ............ ............ 

n 
xR xS xT XU 

XX xhlA 
xv XW xeDA xhlB xlvA 

27.3 kb PBSX 

bP bQ bR bS bT cA CB CC cDcE yqxG yqxH cwlA K 
37-7 kb skin 

Fig. 1. A comparison of the ORFs identified in PBSX and in the skin element. Each ORF is indicated by a box with a 
distinctive shading. ORFs above the line are transcribed from lef t  to right while those below the line are transcribed from 
right to left. Homologous ORFs in the two elements have the same shading. The name of each skin ORF is that assigned 
by MCdigue eta / .  (1995), with the 'yq' part of the name omitted for clarity of presentation. The full gene name and the 
corresponding number assigned to each gene by Takemaru eta / .  (1995) are shown in Table 1. Likewise, the PBSX genes are 
indicated with the ' k d '  part of the name omitted (the full gene name and the corresponding skin ORF are shown in Table 
1). ORFs which do not have a homologue in the other element are shown as open boxes. Adjacent ORFs which overlap 
have an open triangle located at the junction between them. The stem-loop structures signify putative transcriptional 
terminators positioned after an autolysin gene in both elements. 

and their organization is similar to that found in skin (Fig. 
1). In both elements, the middle and late operons are 
orientated in the same direction while the early operon is 
adjacent to the middle operon but orientated in the 
opposite direction. There are also a number of differences 
in the operon organization of the elements: (i) there is a 
single ORF located between the middle and late operons 
of the skin element (and oriented in the opposite direction 
to them) which is not present in PBSX and (ii) xkdA and 
xre are adjacent in PBSX, whereas in the skin element 
there are two interposed ORFs and these are transcribed 
in the opposite direction. 

In terms of gene content, the similarity between the 
elements is most evident in the late operon (Fig. 1). The 
homologous ORFs are similarly shaded and the number 
assigned to each ORF of the skin element by Takemaru e t  
dl .  (1995) is indicated in parentheses after the name of the 

gene (Mkdigue e t  aZ., 1995) in Table 1. Twenty-three 
of the ORFs present in the late operon of PBSX have 
a homologue in the skin late operon. These genes are 
present in the same order in both operons and the 
positioning of start and stop codons for many adjacent 
ORFs is very similar. There are also a number of 
differences between the late operons: (i) there are no 
homologous ORFs in the late operon of PBSX for the skin 
genesyqbB (66),yqbC (84),jqbF (77),yqbG (76) andyqcB 
(1 30) ; (ii) there is a single ORF present in the lysis gene 
cluster of skin CyqxH, 121), whereas there are two smaller 
ORFs (xhZA and xhlB) present at the corresponding 
position in PBSX which do not show homology toyqxH; 
and (iii) there are two instances of rearrangements within 
ORFs in the late operons. The XMV gene from PBSX 
contains a 963 nucleotide segment which is absent from 
the homologous ORFjqcC (131) in the skin element and 
theyqbO (54) from skin contains a 765 nucleotide segment 

2034 



Prophage-like elements skin and PBSX are homologous 

Table 1. Percentage identity and similarity between ORFs from PBSX and from skin 

Operon Gene* skin homologuet Identity Similarity 

(W (Yo) 

Early 

Middle 

Late 

Gene is incomplete 
35.3 66.4 
19.5 44.0 
45.7 64.5 
36.7 73.5 
11.5 39.5 
16.6 46.8 
74.3 82.2 
67.2 84.9 
48-7 67.4 
79.7 93.9 
54.1 77.5 
49.7 74.2 
51.7 77.9 
95.2 98.0 
82.8 95.7 

20.7 
77-3 
82.9 
81-6 
72.7 
85.8 
88.8 
86.4 

39.9 
91.8 
92.7 
94-5 
90.9 
98.6 
95.4 
94-2 

83.5 90.7 
53.7 70.4 
63.9 75.7 
68.5 87.0 
54.9 75.8 
39.0 61.0 

* Constituent genes of the PBSS early, middle and late operons. 
t The nomenclature of the skirr genes is taken from MCdigue e t  a/ .  (1995) and from Takemaru e t  a/. (1995). 
$ Contains 255 aa which are absent in the homologous PBSX protein. 
$ Contains 321 aa which are absent in the homologous skin protein. 

which is absent from the homologous ORF in PBSX. The 
differences in gene content of the early and middle operons 
of the elements are more striking. There are seven 
contiguous ORFs in skin betweenyqaE (7) a n d y q a l  (33) 
(the PBSX homologues are xre and xkdB, respectively) 
which are not present in PBSX. There are also three 
additional ORFs, yqaN (39),yqaP (40) andyqaR (42), in 
this region of skin which do not have homologues in 
PBSX. One of these,JyqaP (40), is oriented in the opposite 
direction to the others within this region. In summary, it 
is clear that the phage-like operons of skin are similar, in 
both organization and gene content, to those of PBSX. 
However, significant rearrangements have occurred since 
the elements diverged, with the result that there are now 
17 fewer ORFs within this region of PBSX than are found 
in the corresponding region of skin. 

The percentage identity and similarity (at the amino acid 
level) between products of homologous ORFs from the 
skin and PBSX elements are presented in Table 1. The 
percentage identities between the gene products of the 
homologous ORFs of the late operons are very high and 
range (with two exceptions detailed below) from 48.7 ‘10 
(674 Oh similarity) to 95.2 % (98 % similarity). The 
exceptions are the x t m A  (1 6.6 YO identity/46.8 ‘/O simi- 
larity) and xbA (39 YO identitp/bl 5’0 similarity) gene 
products. The x t m A  gene encodes the small subunit of 
the terminase (the large subunit being encoded by xtmB) 
and the x&A gene encodes N-acetylmuramoyl-L-alanine 
amidase, an activity which functions in lysis of the 
bacterial cell upon PBSX induction. The identities 
between the products of the genes which show intragenic 
rearrangement are interesting: (i) in the case of xkdwand 

2035 



S. K R O G H  a n d  O T H E R S  

Gene skin 

A T T G T C G T C T T T m T T T T C A K . T T G  
GAATAAAAACAAAAGTCA=&TAAT. ATG 
C A T A A C T G C A C T C T G T m a G . G T G  
T T C A C T C A A ? i A G ~ C M l T .  ATG 
TTCTATCATAACATEGECWTTTGA.GTG 
T G A A T T C A T T m C A T A A T C C A G . T T A  
ATCT-CGGCAGTGAATGTAG.ATG 
A T C A C C A T T G T A A A G A C A  . GTG 
AAGATCAAGACCAGGmL-GC . ATG 
CTTTTCTTATTTT-GMTAAC . ATG 
G T T G T T T T A A T G G A T W A T A C G A . A T G  
TTGAAGCGWTGCmTGAWAAT.ATG 
TGACGGCAGTACGGGA(%TTGAATACTT.ATaA 
TTATGAAGAATTTTmCWCTGCC.ATG 
TCAGTTTCTT-TCAAATAG . ATG 
TCTACTTTAATGTTaGGTfwTmA.ATG 
AACTWGT-TTAAT?GUC. ATG 
AAGAAGC-GAAG-TTAACT . ATG 
GTAGCTAAGATTUUCCCTTGTCG.ATG 
AAAAATACCGCGTEC-GAC . ATG 
AAAATTCTATCCCTGATGGTGTTTCCTC.ATG4 
TCTCGACAAAATATmBTTTGGA.ATG 
T T G G A G A T C G m C T T G G C ! G A G G . A T G  
ATCCCTAAATTGGGGC&GC,TGMCATCA.TTG 
TTAGGATTACTTTTFAGGTGGTTTAACA.ATG 
AAACTATCWTTTTmGAC&TAACG.ATG 
TGCTTACTCTACAAGGAGGGCAAGGTTC.ATDA 
TGATCTACAATGCBBS;CWTTTATT.ATG 
ATCTTACTGACAGaAGAGGGTAT.ATG 

AGO 

-20.0 
-15.6 
-14.2 
-18.4 
-7.2 
-4.6 

-19.0 
-1 5.6 
-20.0 
-24.4 
-17.8 
-11.6 
-1 6.6 
-1 5.8 
-21.2 
-1 6.0 
-14.0 
-1 6.6 
-8.6 

-17.2 
-9.4 

-21.0 
-21.0 
-1 3.8 
-1 3.8 
-18.8 
-1 7.8 
-1 8.0 
-12.8 

AGO PBSX Gene 

-1 1.8 
-9.4 

-1 4.0 
-9.4 

-1 4.2 
-16.6 
-18.8 
-8.4 

-23.4 
-21.0 
-1 8.6 
-15.2 
-9.4 

-19.0 
-20.0 
-1 6.6 
-9.4 

-1 6.6 
-8.6 

-1 6.6 
-12.8 
-14.0 
-25.0 
-13.8 
-14.6 
-1 8.8 
-16.2 
-19.0 
-12.8 

GAACACACGTTC-GAGTATTCAA-TTG 
AGGGAGTTTAAAAAAGAUGC?iTAGT.ATG 
ATGATGTT-GCGGCGTACA.ATG 
GCAATTCAGCATGAG-TGWC.ATG 
CTGCACCGAAAAGGGCACGAAATWC . ATGA 
ATTTGCCAAAAATG&XAGGATCATAGA.ATG 
ATGCTGGAGGTGGCGGTGATGCAGTAGC.ATG 
TTA-CGAAAAGAWGCAAGTC . ATGA 

ACGTCCG-AATCAAGCAG . GTG 
GAACCG-GGAGCfiGGTCATMATC . ATG 

T A A A A C C A G T C A A G G ~ P A W ~ ~ A A  . TTG 
AATCCTCC~CTG~GGTGAGATCTTT.ATG 

GCTGAAGAAGCTGTwiAGG.AGCAGG.ATG 
mCGTCGCAGTCAGGGATGAAAGCCT.ATGA 

CATTTCTAACCmTCATGTCATC.ATG 
TTCTACTTTAATGTmAAACTAAT.ATG 
CAAACAATAGACAAGaTTTTTTTAC.ATG 
GAATCAAGGAGAQ3iGGGAACTAATATC.GTG 
GCGGCGAAGATCUUCGCTTATAG.ATG 
GATACCGCAATGAAAC&XmTATG. 
AGAATACGATGCCGGUGCTTTCCTG. 
TATTTAGCATGT-TGGGCAAC . 
CTTGAGATTG-TGGCGAAG . 
ATTCCTAAGCTTGGGC&XTGUGATCA. 
T C A G A T T T A C T T T T G E G T A A A C .  
ATTACCCGAGAGTTTWGAGGTACCTTT. 
T A C T T T C C T T C A T T A T G C T G A .  
CCTGTTTTTAGATCBBAGCGGATGAAG. 
TAATCTGACGAAAT-AGATGAAA. 

ATG 
ATG 
ATG 
ATG 
TTG 
ATG 
ATG 
ATGA 
ATG 
ATG 

xkdA 
xre 
xkdB 
xkdC 
xtrA 

xtmA 
xtmB 
xkdE 
xkdF 
xkdG 
xk dH 
xkdI 

xk dK 
Xk M 
xkdN 
xkdO 
xkdP 
xk dQ 
xk dR 
xk 
xk dT 
xkdU 
xkdV 
xkdW 
xkdX 
xepA 
x&A 

xpf 

Xkdj 

jqcC, the identities of the amino- and carboxy-terminal 
regions of their products are very high, but a deletion has 
occurred within the skin gene ; (ii) in the case of xkdO and 
yq60, the level of identity between the carboxyl halves of 
the proteins is high (77.3 YO identity/91.8 YO similarity), 
whereas the amino-terminal regions are quite dissimilar. 
The percentage identities between the products of hom- 
ologous ORFs of the early and middle operons of the two 
elements are significantly lower than those observed for 
the products of genes located within the late operons, 
ranging between 11.5 YO identity (39-9 YO similarity) for 
the xpf/’qaO products to 45.7% identity (64.5% simi- 
larity) for the xkdC/yqaM products. 

The observed similarities between these regions of PBSX 
and the phage-like operons of the skin element clearly 
show that they have shared a common ancestor. That this 
ancestor was probably a phage is suggested by the 
similarities of the two elements to phages in terms of their 
gene organization and by the presence within the elements 
of genes homologous to phage repressors, terminases and 
lysis genes (Wood e t  a/., 1990a; McDonnell e t  al., 1994; 
Midigue e t  al., 1995; Takemaru e t  al., 1995). The highest 
level of divergence is seen in the early and middle operons, 
suggesting that the regulation of expression of these two 
elements may now differ. There are differences in the 
overall structure of the elements which may yield insight 
into their origins and possible contributions to cell 
physiology. The PBSX element terminates at the late 
operon terminator. The genes located downstream of the 

PBSX terminator are different to those positioned down- 
stream of the terminator in the skin element and they are 
not homologous to phage genes (S. Krogh, M. O’Reilly 
& I<. Devine, unpublished). The genes located down- 
stream of the terminator in skin include a homologue of B. 
suhtilisgsiA, an arsenate-resistance operon and the recom- 
binase s p o W C A  (Takemaru e t  al., 1995) and do not 
appear to be of phage origin. These observations suggest 
that skin is a composite genetic element composed both of 
phage- and non-phage-derived DNA. 

Investigation of the expression signals for the ORFs 
of PBSX and skin 

The high level of identity between the homologous 
proteins from the PBSX and skin elements suggests that 
the phage-like operons within skin may be expressed and 
be functional. The late operon of PBSX is a single 
transcription unit (Wood e t  al., 1990a; McDonnell e t  al., 
1994) and the similar organization of the skin late operon 
suggests that it may also be similarly expressed. Therefore 
the ribosome-binding site strength of each ORF will be an 
important determinant of the levels of head, tail and lysis 
protein production. The positioning and strength of the 
ribosome-binding sites for each of the ORFs of PBSX and 
skin are presented in Table 2. The strengths of the 
ribosome-binding sites were calculated based on comple- 
mentarity with the 3’-end of the 16s ribosomal RNA (not 
allowing bulges) according to the rules of Tinoco e t  al. 
(1973). It is evident that the ORFs located within the late 

2036 



Prophage-like elements skin and PBSX are homologous 

(a) 
skin element 1R 4- 

4 1F 
3R 4- 

2R 4- 4 3F 

4-- IVCA 
I IllC I I I I 1 I I I IVCB 1 

36 42 48 0 6 12 18 24 30 - 
6 kb 

2 20 !! U '"1 
U 

2 10 
u 
Q k n  - 

E O  2 4 6 8 1 0 1 2 1 4  E O  2 4 6 8 10 12 14 
Time (h) 

E O  2 4 6 8 1 0 1 2 1 4  

Fig. 2. Expression of lacZ fusions located within the skin element during vegetative growth and sporulation. (a) Three 
phage-like operons within the skin element which are homologous to the control and late operon regions of PBSX are 
indicated by arrows immediately above the scale bar. The length and direction of each arrow signify the size and 
transcriptional direction of the operon, respectively. The position and orientation of the lacZ fusions are shown by small 
arrows followed by the fusion-identifying code: insertions 1 F and 1 R are positioned within the y9bO (54) gene; insertion 
2R is  positioned immediately 3' to the y9cF (95) gene; and insertions 3F and 3R are positioned within the yqcK (2) gene. 
The letters b, c and d at  each insertion locus refer to the panel of this figure in which expression of lacZ a t  that locus i s  
shown. The stem-loop structure indicates the putative transcriptional terminator present a t  the end of the phage-like 
late operon. The recombinase gene spolVCA (IVCA) is transcribed from left to right and the two halves of the sigK gene 
(IVCB and IIIC) are transcribed from right to left (not drawn to scale). (b) Accumulation of P-galactosidase in strains SKlF 
(0-El), SKIR (A-A) and JH642 (0-0); (c) accumulation of P-galactosidase in strains SK2R (O-n), SK37 (fusion in 
PBSX, 0-0) and JH642 (0-0); (d) accumulation of P-galactosidase in strains SK3F (A-A), SK3R (0--0) and JH642 
(0-0). E signifies exponential growth and to represents the time at which sporulation was initiated by resuspension as 
described. In all the time courses, accumulation of alkaline phosphatase initiated a t  tz and accumulation of dipicolinic acid 
initiated between t4 and t,. 

operons of the PBSX and skin elements have ribosome- 
binding sites which are appropriately positioned for 
translation. They show extensive complementarity with 
the 16s ribosomal RNA and this is reflected in ribosome- 
binding site strengths ranging from AGO = -8.6 to 
-25 kcal mol-' ( -  36.12 to - 105 kJ mol-'). There is a 
striking similarity in the ribosome-binding site strengths 
of homologous ORFs from the late operons of PBSX and 
skin. The ribosome-binding site strengths for 13 of the 23 
pairs of homologous ORFs in the late operons only differ 
by approximately 1 kcal mol-I (4.2 kJ mol-'), while the 
ribosome-binding site strengths of five additional ORFs 

(e.g. the stop codon designated in bold characters within 
the ribosome-binding site of xkd] is that of the preceding 
ORF, xkdl). For genes xkdl, xkd], xkdO, xkdR, xkdW 
and xkdX, the stop codon used and its positioning are 
conserved between the homologous ORFs of the two 
elements. In total, the stop codons for 11 PBSX ORFs and 
13 skin ORFs are located within, or in close proximity to, 
the ribosome-binding site of the downstream ORF. These 
analyses support the hypothesis that the phage-like 
operons located within the skin element are expressed. 

Are the phage-like operons of skin expressed? 
only differ by apGoximately3 kcal mol-' (1 2.6 k J mol-'). 
Correspondence analysis between the ribosome-binding 
site strengths of each ORF in the two elements gives a 
correspondence coefficient of 0.69, indicating a significant 
correlation between the two sets of data. This similarity 
contrasts with the early and middle operons, where the 
ribosome-binding site strengths are not significantly 
similar. 

The positioning of the stop codon for the ORF preceding 
each ribosome-binding site is also indicated in Table 2 

To investigate the patterns of expression of operons 
within the skin element, three IacZ fusions were generated 
at the positions outlined in Fig. 2(a): (1) a fusion within 
the phage-like late operon at position b [the point of 
fusion is positioned within the amino-terminal region of 
yq60 (54)] ; (2) a fusion at position c [the point of fusion is 
positioned immediately 3' to theyqcF (95) gene] ; and (3) a 
fusion at  position d [the point of fusion is located within 
theyqcK (2) gene]. The orientation of each insertion is 
indicated in Fig. 2(a). The progression of sporulation was 

2037 



S. K R O G H  a n d  O T H E R S  

established by measuring alkaline phosphatase accumu- 
lation (which begins to accumulate at stage I1 of 
sporulation, t ,  in our experiments) and dipicolinic acid 
(which begins to accumulate at stage V of sporulation, t,-, 
in our experiments). The pattern and level of P-galacto- 
sidase accumulation for each fusion are shown: in Fig. 
2(b), fusions 1F and 1R at position b ;  in Fig. 2(c), fusion 
2R; and in Fig. 2(d), fusions 3F and 3R. The expression 
profile for the two fusions 1F  and 1R positioned at b 
within the skin late operon is the same as that seen in the 
control strain JH642, which has no IacZ fusion, indicating 
that it is not expressed during vegetative growth nor 
during sporulation (Fig. 2b). The fusion 2R positioned at 
c shows a very high level of P-galactosidase during 
vegetative growth, followed by a steady decline in levels 
up to f, of sporulation (Fig. 2c). Expression of a IacZ 
fusion positioned within the late operon of PBSX was 
also examined during vegetative growth and sporulation 
and the profile observed was the same as the background 
(Fig. 2c). The insertion at position d in the 3F orientation 
was moderately expressed during vegetative growth, 
decreased slightly between to and t, of sporulation and 
then increased to a peak at t4 (Fig. 2d). The insertion in the 
reverse orientation at this position (3R) shows only 
background levels of P-galactosidase activity. These data 
show that there are two operons within the skin element 
which are transcriptionally active, and each has a dis- 
tinctive expression profile. However, it is clear that neither 
the phage-like late operon of skin nor the late operon of 
PBSX are expressed during either vegetative growth or  
sporulation under these conditions. We then tested the 
hypothesis that expression of the phage-like skin operons 
may be under the control of the SOS response in a manner 
similar to PBSX. The strains harbouring the lacZ fusions 
lF ,  2R and 3F within the skin element were exposed to 
mitomycin C and the levels of accumulated P-galactosidase 
were determined. These l a Z  fusions were also transferred 
to a strain in which the distal part of the PBSX late 
operon, including the lysis genes, is deleted. PBSX is 
induced normally in this strain but the cells do not lyse, 
which facilitates the measurement of the intracellularly 
located P-galactosidase. None of the skin fusions examined 
were expressed after exposure of the cells to mitomycin C 
(results not shown). The same results were obtained with 
the fusions in the lysis-negative genetic background. In 
order to verify that the SOS response was induced by 
treatment with mitomycin C, we examined the profile of 
P-galactosidase accumulation in strains with a lacZ fusion 
positioned within the PBSX late operon. In this case, 
intracellular P-galactosidase activity increased after ex- 
posure to the inducing agent and the cells subsequently 
lysed, confirming that the SOS response was induced 
(results not shown). These data demonstrate that ex- 
pression of the three phage-like transcriptional units 
within the skin element is not under the control of the 
SOS response. 

It is evident that the expression patterns for each of the 
skin operons are distinct and that only one of the three 
operons showed a pattern of expression which was related 
to sporulation. Thus the location of the skin element 

within a sporulation gene may be fortuitous. There does 
not appear to be any obvious hierarchy or  inter- 
dependency of expression among the operons, indicating 
that the phage and non-phage-like operons of skin are 
unrelated. The fact that the late operon is not expressed 
during exponential growth, sporulation or  during the 
SOS response is consistent with the observations of 
Foster (1993). In his study, expression of the c w l A  gene 
(the location of cwlA was unknown at that time) was 
investigated. The lncZ fusions in the Foster study were 
located in the c w l A  gene, 12 kb downstream of the IacZ 
fusion point b in this study. However expression of 
neither insertion was observed under any of the conditions 
used in these studies. It is interesting to note, however, 
that Foster (1993) observed amidase activity when the 
structural gene for cwlA was inserted into an expression 
vector and cloned in E. coli. This demonstrates that at 
least one gene within the skin late operon is functional. 
The cwlA/xbA homologous gene pair is only 39% 
identical, the second lowest level of identity observed for 
the 23 homologous gene pairs of the late operon, 
suggesting that the remaining 22 genes may also be 
functional. 
Is there a condition under which the phage-like operons 
within skin are expressed? The level of conservation 
which exists between the homologous ORFs in PBSX and 
skin, and the demonstration that c w l A  encodes a functional 
enzyme (Foster, 1993), strongly suggest that the skin 
ORFs are functional and can be expressed. The sequence 
of genes which are not expressed over long periods of 
time will degenerate. This is the source of pseudo-genes in 
eukaryotes and there is evidence that a similar process has 
occurred in B. subtilis (Lazarevic e t  d., 1995). Thus it is 
certain from our analysis that the constituent ORFs in the 
skin late operon must have been functional until very 
recently. It is possible that they are still functional and that 
a stimulus other than those tested in this study and in 
Foster’s study triggers their expression. However, it is 
also possible that mutations, generated only very recently, 
have inactivated the induction mechanism (which may 
have responded to SOS- or sporulation-inducing signals) 
so that expression of the phage-like operons can no longer 
be triggered. These mutations could have occurred either 
during prolonged growth under laboratory conditions or 
during the process (mutagenesis by X-rays) whereby B. 
subtilis strain 168 was generated from the parental 
Marburg strain (Anagnostopoulos, 1990). If the ability to 
induce the skin phage-like operons does not provide a 
survival advantage during growth under laboratory 
conditions, then such mutations would not have been 
detected. To test these ideas, we are undertaking (i) to 
place the skin phage-like late operon under the control of 
the PBSX induction system to see if a functional phage 
particle can be formed and (ii) to examine expression of 
the skin late operon in the parental Marburg strain. 

ACKNOWLEDGEMENTS 

This work was supported by EU grant BI02-CT93-0272, by 
Forbairt grant SC/95/124 and by BioResearch Ireland through 
the National Pharmaceutical Biotechnology Centre at Trinity 

2038 



Prophage-like eIements skin and PBSX are homologous 

College, Dublin. We would like to thank Andrew Lloyd at the 
Irish National Centre for Bioinformatics for assistance with the 
use of their facilities and Des Higgins for ribosome-binding site 
strength calculations. We thank Maria Bradfield for technical 
assistance and David McConnelI and Ken Wolfe for stimulating 
discussion and for critical reviews of the manuscript. 

REFERENCES 

Adams, L. F., Brown, K. L. & Whiteley, H. R. (1991). Molecular 
cloning and characterisation of two genes encoding sigma factors 
that direct transcription from a Bacilltrs thtlringiensis crystal protein 
gene promoter. J Bacterioll73, 3846-3854. 
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. 
(1990). Basic local alignment search tool. J Mol Biol2'15. 1.03-410. 
Anagnostopoulos, C. (1 990). Genetic rearrangements ;r, Bacillzis 
subtilis. In The Bacterial Chromosome, pp. 361-371. Edited by X. Riley 
& C. Drilica. Washington, DC: American Society for Micro- 
biology. 
Anagnostopoulos, C. & Spizizen, J. (1961). Requmments for 
transformation in Bacillus subtilis. J Bacteriol 81, 741-746. 
Buxton, R. 5. (1 976). Prophage mutation causing heat inducibility 
of defective Bacillus subtilis bacteriophage PBSX. J Virol20, 22-28. 
Cutting, 5.8 Oke, V., Driks, A., Losick, R., Lu, 5. & Kroos, L. (1990). 
A forespore checkpoint for mother cell gene expression during 
development in Bacillus szlbtilis. Cell 62, 239-250. 
Ferrari, E., Howard, 5. M. H. & Hoch, J. A. (1986). Effect of stage 0 
sporulation mutations on subtilisin expression. J Bacteriol 166, 

Foster, 5. J. (1993). Analysis of Bacillus subtilis 168 prophage- 
associated lytic enzymes ; identification and characterization of 
CWLA-related prophage proteins. J Gen Microbial 139, 3177-31 84. 
Garro, A. J., Leffert, H. & Marmur, J. (1970). Genetic mapping of a 
defective bacteriophage on the chromosome of BaciliyIr subtilis 168. 
J I4rol6,  340-343. 
Glaser, L., lonesco, H. & Schaeffer, P. (1966). Teichoic acid as 
components of a specific phage receptor in Baci//w . ~ h : i i i .  Biochim 
Biopbys A c t a  124, 415-417. 
Glaser, P., Kunst, F., Arnaud, M., Coudart, M.-P., Gonzales, W., 
Hullo, M.-F., lonescu, M., Lubochinsky, B., Marcelino, L., Moszer, 
I., Presecan, E., Santana, M., Schneider, E., Schweiter, J., Vertes, 
A., Rapoport, G. & Danchin, A. (1993). Bacillus subtilis genome 
project : cloning and sequencing of the 97 kilobase region from 325" 
to 333". Mol Microbiof 10, 371-384. 
Gouy, M., Gautier, C., Attimonelli, M., Lanave, C. & diPaola, G. 
(1985). ACNUC - a portable retrieval system for nucleic acid 
sequence databases : logical and physical designs and usage. 

Haas, M. & Yoshikawa, H. (1969). Defective bacteriophage PBSH 
in Bacillus szjbtilis. I. Induction, purification and physical properties 
of the bacteriophage and its deoxyribonucleic acid. J Virol 3, 

Higgins, D. G., Bleasby, A. J. & Fuchs, R. (1992). CLUSTAL v :  
improved software for multiple sequence alignment. C A  BIOS 8, 

Holmes, D. 5. & Quigley, M. (1981). A rapid boiling method for the 
preparation of bacterial plasmids. Anal Biochem 114, 193-1 97. 
Karamata, D., Pooley, H. M. & Monod, M. (1987). Expression of 
heterologous genes for teichoic acid in Bacillus subtilis 168. Mol Gen 
Genet 207, 73-81. 
Kroos, L., Kunkel, B. & Losick, R. (1989). Switch protein alters 

173-1 79. 

CABIOS 1, 167-172. 

233-247. 

189-191. 

specificity of RNA polymerase containing a compartment-specific 
sigma factor. Science 243, 526-529. 
KUnkel, B., Sandman, K., Panzer, 5.8 Youngman, P. & Losick, R. 
(1988). The promoter for a sporulation gene in the spoIVC locus of 
Bacillus szlbtilis and its use in studies of temporal and spatial control 
of gene expression. J Bacteriol 170, 3513-3522. 
Kunkel, B., Losick, R. & Stragier, P. (1990). The Bacillus stibtilis gene 
for the developmental transcription factor oK is generated by 
excision of a dispensable DNA element containing a sporulation 
recombinase gene. Genes Dev 4,  525-535. 
Lazarevic, v.8 Mauel, C., Soldo, B., Freymond, P.-P., Margot, P. & 
Karamata, D. (1995). Sequence analysis of the 308" to 31 1' segment 
of the Bacillus subtilis 168 chromosome, a region devoted to cell wall 
metabolism, containing non-coding grey holes which reveal 
chromosomal rearrangements. MicrobioloQ 141, 329-335. 
Longchamp, P. F., Mauel, C. & Karamata, D. (1994). Lytic enzymes 
associated with defective prophages of Bacillzls subtilis: sequencing 
and characterization of the region comprising the N-acetyl- 
muramoyl-L-alanine amidase gene of prophage PBSX. Microbiology 

Lu, S., Halberg, R. & Kroos, L. (1990). Processing of the mother-cell 
CT factor, crK, may depend on events occurring in the forespore 
during Bacillus subtilis development. Proc Natl  Acad Sci U S A  87, 

Mauel, C. & Karamata, D. (1984). Characterisation of proteins 
induced by mitomycin C treatment of Bacillus subtilis. J Virol 49, 

McDonnell, G. & McConnell, D. J. (1994). Overproduction, iso- 
lation and DNA-binding characteristics of Xre, the repressor 
protein from the Bacillus subtilis defective prophage PBSX. J 
Bacterioll76, 5831-5834. 
McDonnell, G., Wood, H., Devine, K. M. & McConnell, D. 1. 
(1994). Genetic control of bacterial suicide: the PBSX system of 
Bacillzlx wbtilis. J Bacteriol 176, 5820-5830. 
MBdigue, C., Moszer, I., Viari, A. & Danchin, A. (1995). Analysis of 
a Bacillus subtilis genome fragment using a co-operative computer 
system prototype. Gene 165, GC377GC51. 
Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring 
Harbor, NY: Cold Spring Harbor Laboratory. 
Nicholson, W. L. & Setlow, P. (1990). Sporulation, germination and 
outgrowth. In Molecular Biological Methods for Bacillus, pp. 391-450. 
Edited by C. R. Harwood & S. M. Cutting. Chichester: J .  Wiley 
and Sons. 
Okamoto, K., Mudd, J. A. 81 Marmur, J.  (1968). Conversion of 
Bacillgs subtilis DNA to phage DNA following mitomycin C. J Mol 

Popham, D. L. & Stragier, P. (1992). Binding of the Bacillus subtilis 
spoIVCA product to the recombination sites of the element 
interrupting the oK-encoding gene. Proc Natl Acad Sci U S A  89, 

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: 
a Laborator_y Manzial, 2nd edn. Cold Spring Harbor, NY: Cold 
Spring Harbor Laboratory. 
Sato, T., Samori, Y. & Kobayashi, Y. (1990). The c isA cistron of 
Bacillzis subtilis sporulation gene spolVC encodes a protein hom- 
ologous to a site-specific recombinase. J Bacterioll72, 1092-1098. 
Sato, T., Harada, K., Ohta, Y. & Kobayashi, Y. (1994). Expression 
of the Bacilkus subtilis spoI V C A  gene, which encodes a site-specific 
recombinase, depends on the spoIIBG product. J Bacteriol 176, 
935-937. 
Steensma, H. Y., Robertson, L. A. & van Elsas, J. D. (1978). The 

140, 1855-1867. 

9722-9726. 

806-812. 

Bi0l34, 429-437. 

5991-5995. 

2039 



S. K R O G H  a n d  OTHERS 

occurrence and taxonomic value of PBSX-like defective phages in 
the genus Bacillus. Antonie Leeuwenhoek 44, 353-366. 
Stragier, P., Kunkel, B., Kroos, L. & Losick, R. (1989). Chromosomal 
rearrangement generating a composite gene for a developmental 
transcription factor. Science 243, 507-51 2. 

Takernaru, K., Mizuno, M., Sato,T., Takeuchi, M. & Kobayashi, Y. 
(1995). Complete nucleotide sequence of a skin element excised by 
DNA rearrangement during sporulation in Bacillus subtilis. Micro- 
biology 141, 323-327. 
Thurrn, P. & Garro, A. 1. (1975a). Isolation and characterisation of 
prophage mutants of the defective Bacillus subtilis bacteriophage 
PBSX. J L’ir0ll6, 184-191. 
Thurrn, P. & Carro, A. 1. (1975b). Bacteriophage-specific protein 
synthesis during induction of the defective B a c i l h  subtilis bacterio- 
phage PBSX. J Viro/  16, 179-183. 
Tinoco, I., Borer, P. N., Dengler, B., Levine, M. D., Uhlenbeck, 
0. C., Crothers, D. M. & Gralla, 1. (1973). Improved estimation of 

secondary structure in ribonucleic acids. Nature New Biol 246, 

Wood, H. E., Dawson, M. T., Devine, K. M. & McConnell, D. 1. 
(1 990a). Characterisation of PBSX, a defective prophage of Bacillus 
subtilis. J Bacteriol172, 2667-2674. 
Wood, H. E., Devine, K. M. & McConnell, D. 1. (1990b). Charac- 
terisation of a repressor gene ( m e )  and a temperature sensitive allele 
from the Bacillus subtilis prophage PBSX. Gene 96, 83-88. 
Young, M., Mauel, C., Margot, P. & Karamata, D. (1989). Pseudo- 
allelic relationship between non-homologous genes concerned with 
biosynthesis of polyglycerol phosphate and polyribitol phosphate 
teichoic acids in Bacillus subtilis strains 168 and W23. Mol Microbiol 

40y4 1. 

3, 1805-1812. 

Received 8 December 1995; revised 1 March 1996; accepted 15 March 
1996. 

2040