Download The Possible DNA-binding Nature of the Regulatory Proteins

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Histone acetylation and deacetylation wikipedia , lookup

SR protein wikipedia , lookup

Signal transduction wikipedia , lookup

G protein–coupled receptor wikipedia , lookup

Magnesium transporter wikipedia , lookup

Protein phosphorylation wikipedia , lookup

Protein wikipedia , lookup

Protein moonlighting wikipedia , lookup

Protein domain wikipedia , lookup

Nuclear magnetic resonance spectroscopy of proteins wikipedia , lookup

Cyclol wikipedia , lookup

List of types of proteins wikipedia , lookup

Intrinsically disordered proteins wikipedia , lookup

JADE1 wikipedia , lookup

Artificial gene synthesis wikipedia , lookup

Homology modeling wikipedia , lookup

Protein structure prediction wikipedia , lookup

Silencer (genetics) wikipedia , lookup

Transcript
Journal of General Microbiology (1987), 133, 2381-2391.
Printed in Great Britain
238 1
The Possible DNA-binding Nature of the Regulatory Proteins, Encoded by
spZZD and gevE, Involved in the Sporulation of Bacillus subtilis
BySUSAN K. HOLLAND,’ SIMON C U T T I N G 2 AND
JOEL MANDELSTAM2*
Laboratory of Molecular Biophysics, University of Oxford, South Parks Road,
Oxford OX1 3QU, UK
Microbiology Unit, Department of Biochemistry, University of Oxford, South Parks Road,
Oxford OX1 3QU, UK
(Received 17 December 1986; reuised 25 May 1987)
/
The predicted polypeptide products of two genes, spoIID and gerE, which appear to be
concerned in the regulation of spore formation in BaciZZus subtilis have been compared by
modelling methods with known DNA-binding proteins. The results indicate that both
polypeptides may have DNA-binding properties and the conclusion is drawn that this may
account for their regulatory action.
INTRODUCTION
The formation of spores in Bacillus is the simplest example of cellular differentiation that has
been extensively studied and the process has been divided into six stages on the basis of the
appearance of cell sections in the electron microscope (Ryter et al., 1966). It is now known that
the first two of these stages (0 and I) have, in fact, nothing to do with the developmental process
and can be disregarded as far as this paper is concerned.
For the molecular biologist the questions posed are those that arise in any study of
differentiation. (a) How many operons - gene clusters functioning as units - control the whole
developmental change? (b) What is the order of expression of these operons? (c)What kinds of
molecule are involved in switching on these operons?
The answer to the first question is partly known. About 50 genetic loci, presumed to act as
operons, have been identified (see Piggot & Coote, 1976; Piggot & Hoch, 1985). Some of these
have been shown by RNA-DNA hybridization to produce mRNA transcripts corresponding to
several genes (Savva & Mandelstam, 1986) whereas others apparently code for only one gene
each (see below). New sporulation loci are still being discovered, though at a diminishing rate,
and the total number has been estimated on a statistical basis at about 60 (Hranueli et al., 1974).
It thus appears that over two-thirds of the genetic loci have been identified.
The second question has now also been partly answered. It was evident from early work on
sporulation that the expression of operons must be in some way sequential. This is because in a
mutant blocked at any particular stage, the morphological events and the biochemical events
associated with them occur normally until the block is encountered. The simplest model is the
‘linear dependent sequence’ (see Mandelstam, 1976) in which the turning on of each operon
causes the formation of a molecule that regulates the expression of the next operon. However,
some experiments designed to test the hypothesis suggested that two or more operons might be
activated in parallel (Coote & Mandelstam, 1973). More recent work has shown that the original
model was too simple and that, though there is a dependent sequence, it has at least three
branches, one of which leads to the expression of some of the sporulation operons in the spore
compartment while the other two lie in the mother cell compartment (Turner et al., 1986;
Cutting & Mandelstam, 1987). It is already clear, although the data are far from complete, that
the number of sequential steps is large (Mandelstam & Errington, 1987).
0001-3889 0 1987 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
2382
S . K . HOLLAND, S . CUTTING A N D J . MANDELSTAM
The final question, the nature of the molecules that control the sporulation operons, remains
unsolved. It was proposed mainly by Losick and his colleagues (see e.g. Losick & Sonenshein,
1969; Losick & Pero, 1981) that the successive switching on of operons depended on the
production of a succession of different types of sigma factor subunits for RNA polymerase. So
far, however, only one sigma factor, cr29 coded for by spoIIG, has been found to be associated
with core polymerase and also necessary for sporulation. A second sigma factor, also with M ,
29000, has been inferred from the sequence of one of the sporulation operons, spoIIA, whose
nucleotide sequence (see Fort & Piggot, 1984) showed a high degree of homology to known
sigma factors of B. subtilis and Escherichia coli (Errington et al., 1985). Both of these sigma factor
genes are concerned with stage 11, which is (see above) @e first relevant stage of sporulation.
However, two of the genes that have been sequenced in this laboratory and that have been
shown to have a regulatory function in sporulation, gerE and spoIID, appear to code for typical
DNA-binding proteins and this paper describes the observations that point to this conclusion.
METHODS
The predicted amino acid sequences of the Bacillus subtilis gerE and spollD gene products were searched for
similarities with other protein sequences held in the NBRF Database using the SEARCH program (Protein
IdentificationResource, National Biomedical Foundation, Georgetown University Medical Center, Washington
DC, USA) on the Molecular Biophysics VAX computer (Cutting & Mandelstam, 1986; Lopez-Diaz et al., 1986;
Dayhoff et al., 1983). The observed similaritieswere aligned using the ALIGN program (Needleman & Wunch,
1970; Orcutt et al., 1984). The potential DNA-binding regions were modelled upon a DNA-binding protein, the
phage 1 cro structure (Anderson et al., 1981), using the co-ordinates supplied by the Protein Databank and the
Molecular BiophysicsEvans and Sutherland PS300 graphicssystem. The main backbone of the 1 cro structure and
the side chains of the spoZZD and gerE sequences corresponding to those of the 1 cro side sequence were all held
fixed. Where possible, all similar atoms were spatially matched and additional atoms were placed with regard to
ideal stereochemistry as defined by the FRODO dictionary. Further manipulation was required to relieve
unfavourable Van der Waals contacts.
RESULTS A N D DISCUSSION
The amino acid sequences of the B. subtilis gerE and spoIID polypeptides predicted from the
nucleotide sequences are shown in Fig. 1.
The sequences were initially searched for homologies with other proteins by analysis of their
sequences with the NBRF Database using the SEARCH program. The database search
revealed that there were significant similarities with several regulatory proteins. Detailed
inspection of the sequences revealed that both the gerE and spoIID polypeptides contained
regions which showed significant homology with other DNA-binding proteins over the
characteristic helix-turn-helix region, suggesting that these two proteins may also possess a
DNA-binding substructure which interacts with a specific operator sequence. The potential
DNA-binding regions of the gerE and spoIID sequences are shown underlined in Fig. 1 .
The same segments are aligned with several DNA-binding proteins in Fig. 2. We have
included in Fig. 2 the sequence data corresponding to another sporulation gene, spoIIIC (J.
Errington, personal communication), which shows remarkable similarity to trpR.
Characterization of sequence-spec@ DNA-binding proteins
Regulatory proteins which interact with specific DNA sequences have been widely studied at
the tertiary and primary level. The determination of the three-dimensional structures of the E.
coli catabolite activator protein (CAP) (McKay et al., 1982), E. coli trp apo-repressor (Schevitz
et al., 1985), and phage A cI (Pabo et al., 1982; Pabo & Lewis, 1982) and cro repressors (Anderson
et al., 1981) and the phage 434cI repressor (Anderson et al., 1985) has revealed certain
characteristics of DNA-binding proteins ;these have enabled specific models to be proposed for
the recognition of operator DNA sequences by each of the proteins, and a comparison of
primary sequences with other regulatory proteins has suggested similar modes of recognition of
specific DNA sequences (Ptashne et al., 1980; Pabo & Sauer, 1984; Ohlendorf & Matthews,
1983; Sauer et at., 1982; Ohlendorf et al., 1983; Steitz et al., 1982).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
Role of spoIID and gerE in sporulation
2383
S I P V Y R T A N Q S V E N I P L E E Y V I G V V A S E M P
100
A T F K P E A L K A Q A L A A R T F I V R L M V S N S A V E
150
A P K G S L V D D T Q M F Q V Y K S K A E L K K Q W G T S Y
E T K L K K I T D A V A S T Q G K I L T Y N N Q P I E A S F
200
F S T S N G Y T E N A E A Y W T S A I P Y L K S V K S P W D
K K S P K Y K A T K T F T A A E F Q Q K L G V K L D G S S A
250
V G K I T G E T P G H Q V A T A V I N G K T L K G R D I R E
300
K L G L N S A D F E W K R N G D T I T V T T R G F G H G V G
M S Q Y G A N F M A K E G K T V D D I V K Y Y Y Q G T Q I S
E A D A F L N K Y M A K K
- -
1
10
20
30
40
( b ) M K E K E F Q S K P S L T K R E R e V F E L L V Q D K T T K E I A S E L F I S E K T
------
Turn
-
a
-
Turn
b
-
50
60
70
V R N H I S N A M Q K L G V K G R S Q A V V E L L R M G E L E L
Turn
Turn
d
Fig. 1. Predicted primary sequencesof the polypeptideproducts of spollD and gerE. (a)spollD ;broken
underlining indicates a strongly hydrophobic region (see Discussion); the continuous line indicates the
proposed DNA-binding region. (b) gerE; the consensus of secondary structure predictions using the
methods of Chou & Fasman (1974), Lim (1974) and Gamier et al. (1978) is shown under the sequence.
Alpha helices are shown as bold arrows and turns are labelled.
C
The proteins interact as dimers, with an operator sequence exhibiting two-fold symmetry.
Although their overall tertiary structures are different, they all possess a common structure
consisting of two consecutive helices separated by a tight turn. The second helix protrudes from
the surface of the molecule and fits into the major groove of B-DNA (Warrant & Kim, 1978).
The recognition of the operator is by multiple hydrogen bonds or hydrophobic interactions
between amino acid side-chains of this helix and the exposed edges of the bases in the major
groove (Seeman et al., 1976; Ebright et al., 1984; von Hippel et al., 1986). The phosphate groups
of the DNA backbone are also involved, interacting mainly with the first helix of the structural
unit, adding to the total binding energy. The DNA-binding sequences of several DNA-binding
proteins are shown in Fig. 2 and the conserved characteristics of each position within the
20-residue unit are shown. The most strongly conserved residues flank the tight turn, and the
residues at positions 5 and 15, which contribute t o the structural integrity of the unit by
hydrophobic interactions, maintain the angle between the helices.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
2384
S . K . HOLLAND, S . CUTTING AND J . MANDELSTAM
Helix 1
0
0
Turn
0
0
0
0
0
0
0
Helix 2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
DNA-contacting
Hydrophobic
0
Hydrophilic
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
33Q
E
S
V
A
D
K
M
G
M
G
Q
S
G
V
G
A
L
F
NAcl*
1 6 Q T K T A K D
L
G
V
Y
Q
S
A
I
N
K
A
I
HAcro*
169R
Q
E
I
G
Q
I
V
G
C
S
R
E
T
V
G
R
I
L
KCAP*
68Q
R
E
L
K
N
E
L
G
A
G
I
A
T
I
T
R
G
S
NtrpR*
l8Q
A
E
L
A
Q
K
V
G
T
T
Q
Q
S
I
E
Q
L
E
N434cl*
28R
A
E
I
A
Q
R
L
G
F
R
S
P
N
A
A
E
E
H
LlexA
292T
R
G
F
G
H
G
V
G
M
S
Q
Y
G
A
N
F
M
A
Ks~oZZD
471
S
N
A
M
Q
K
L
G
V
K
G
R
S
Q
A
V
V
E
LgerE
95Q
R
E
I
A
K
E
L
G
I
S
R
S
Y
V
S
R
I
E
KspoZZZC
Fig. 2. Diagram showing the characteristics of the residues of the DNA-binding unit in several
sequence-specificregulatory proteins and the alignment with the potential DNA-binding regions of
spoZZD and gerE. An asterisk indicates that the three-dimensional structure of the protein is known.
References for primary sequences: 1 cZ (Sauer & Anderegg, 1978), A cro (Hsiang et al., 1977), CAP
(Aiba et al., 1982),trpR (Gunsalus& Yanofsky, 1980),434 cZ(Grossched1& Schwarz, 1979),lexA (Horii
et al., 1981), spoZZD (Lopez-Diaz et al., 1986), gerE (Cutting & Mandelstam, 1987), spollZC
(J. Errington, personal communication).
The possible DNA-binding region of spoIID
The spoIID segment (residues 292-311) contains similar residues to those observed at the
strongly conserved positions 8-9-10 (valine-glycine-methionine) about the tight turn, and also
at positions 5 and 15 there are two residues (glycine and alanine) which maintain the angle
between the two helices. The numbering of all DNA-binding segments is as given in Fig. 2. The
interior facing triplets of the two helices (4-5-8 and 15-18-19) all contain non-polar residues,
indicating a conserved hydrophobic environment in the interior of the unit. Comparison of the
spoIID sequence with the known DNA-binding segments suggests that residues 11 (serine), 12
(glutamine), 13 (tyrosine), 16 (asparagine), 17 (phenylalanine) and 20 (lysine) are capable of
interacting with the nucleotide bases of a DNA operator (Wharton & Ptashne, 1985;
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
Role of spoIID and gerE in sporulation
2385
Hochschild & Ptashne, 1986). In particular, the residues serine and glutamine are often found at
positions 11, 12 and 13. The models of the DNA-protein complexes suggest that residues at
these positions form bidentate hydrogen bonds with bases at position 1 and 2 of the
operator. By analogy, residues 13 (tyrosine) and 16 (asparagine) may contact positions 4 and
5 of an operator sequence. These are only tentative assignments for specific interactions
because the recognition code is complex and the operator sequence for the spoIID protein is
unknown. Other non-specific interactions may be formed between residues 2 (arginine), 6
(histidine), 11 (serine) and 16 (asparagine) and the phosphate groups of the backbone of the
DNA.
The spoIID segment (residues 292-31 1) was modelled upon the backbone of the phage J. cro
structure civer the DNA-binding region, substituting the corresponding side chains of the J. cro
for those of spoIID. The superimposed structure is shown in Fig. 3. The A cro DNA-binding
structure is shown in blue and the spoIID structure in orange. Overlap of similar atoms is shown
in white. Several of the contacting residues are visible on the exterior of the unit. Examination of
the internal packing between the two helices, based on Van der Waals radii, revealed a good fit
between interface residues with no room for solvent molecules within the interior of the unit.
Notably, the planar ring of residue 4 (phenylalanine) packed in between 15 (alanine) and the
aliphatic part of 1 (threonine), provided a good hydrophobic contact.
+
+
+
+
The possible DNA-binding region of gerE
ThegerEsegment (47-66) shows a good correspondence with the residues of the known DNAbinding sequences, with similar characteristics in most positions. Of the most strongly
conserved residues in the unit, the tight turn appears to be conserved in the gerE sequence at
positions 8-9-10 (leucine, glycine, valine) and methionine at position 5. The last of these
residues does not appear to be unacceptable, as the trp apo-repressor contains a lysine at this
position. However, at position 15 only hydrophobic residues have previously been observed and
here the gerE sequence contains a glutamine. To test the feasibility of the gerE polypeptide as a
DNA-binding protein the segment was imposed upon the iZ cro structure - as was done with the
spoIID segment - again with the main backbone held fixed (Fig. 3).
Apart from the glutamine side chain the remainder of the gerE sequence fitted well onto the
bi-helical unit with no steric hindrance between residues. A study of 200 determined protein
structures showed that 7% of all glutamines are found 95% or more buried within a protein
(Chothia, 1976). It is therefore possible to accommodate this residue in the interface between the
helices, provided that the polar head group can be hydrogen-bonded. Possible hydrogen bonds
could be achieved between the carbonyl of the glutamine and the hydroxyl group of residue 14
(serine). Also, the amide of the glutamine could interact with the main chain carbonyl of residue
15 without induced stress in the conformation of the side chains. In the gerE structure this may
indicate that the maintenance of the angle between helices is contributed to by these hydrogen
bonds in addition to hydrophobic interactions.
From the comparison of the segment with the consensus of DNA-binding characteristics,
there are several potential DNA-contacting residues. Those involved in specific base
interactions may be 11 (lysine) which by analogy with the proposed complexes may contact
position 1 of the unknown operator. Residue 13 (arginine) may contact position +4, with
residues 14 (serine) and 17 (valine) in the vicinity of position + 5 . Residue 19 (glutamic acid)
may also contact an exposed base. Non-specific interactions with the phosphate backbone could
occur with residues 2 (serine), 3 (asparagine), 6 (glutamine), 7 (lysine), 11 (lysine) and 19
(glutamine).
Further study of the gerE primary sequence reveals that the amino-terminal arm is highly
charged like that of the phage J. cI repressor, containing two glutamic acid residues and two
lysines. In the 1 cI repressor there are three lysines, a serine and a threonine which have been
shown to contact the far side of the DNA, interacting with the bases and phosphates and aiding
specificity (Pabo et a f . , 1982; Eliason et al., 1985). It is possible that the gerE protein may also
similarly contact DNA, possessing a flexible arm to interact with the bases of the operator
sequence.
+
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
2386
S . K . H O L L A N D , S . C U T T I N G A N D J . MANDELSTAM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
Role of spoIID and gerE in sporulation
2387
Secondary structure predictions, based upon the methods of Chou & Fasman (1974), Lim
(1974) and Garnier et al. (1978) suggest, with good agreement, that the gerE gene product is an
all-helical protein, containing four helices of which the third and fourth (c and d) comprise the
DNA-binding unit (see Fig. 1).
The NBRF protein database search also indicated several similarities between the gerE
sequence and other regions of DNA-binding proteins. Significant similarities were found by
running the ALIGN program for the complete gerE sequence and the whole of the aminoterminal DNA-binding domain of the E. coli lexA repressor (residues 1-58) (Pabo & Sauer,
1984; Brent 8z Ptashne, 1981;Hurstel eta]., 1986) as shown in Fig. 4. This may indicate that the
two structures adopt similar conformations.
Other similarities were found for residues 10-31 of the gerE sequence and a region in the
carboxyl domain of the E . coli lexA repressor and the C helix of the E. coli trp repressor (Fig. 4)
(Schevitz et al., 1985). The phage repressors and lexA repressor exhibit good homology over their
carboxyl domains, which are known to contain strong oligomerization sites that create
additional stabilizing dimer contacts (see Fig. 5) (Sauer et al., 1982). Similarly helix C of the trp
repressor is known to form dimer contacts and the repressor itself contains only a-helices
(Schevitz et al., 1985). These similarities between gerE and the DNA-binding proteins suggest
not only that the gerE protein contains a DNA-binding unit, but that it also utilizes the rest of its
relatively short length to form regions of inter-domain contacts, along one or more helices (a
and/or b) (see Fig. 1b). As little of the chain extends beyond the potential DNA-binding region,
it would appear likely that any dimer contacts would lie in the amino-terminal portion of the
polypeptide. As the polypeptide is quite short it appears unlikely that the protein contains more
than one domain. This homology strongly suggests that the gerE and lexA repressor contain
similar helices to the helix C of the trp repressor and that the lexA repressor proteins and phage
repressors use a similar helical region in the carboyxl domain to form dimer contacts.
The regulatory function of the B . subtilis potential DN A-binding proteins is indeterminable
from sequence studies alone and this point will be considered in more detail below.
Negative gene regulation involves the repressor molecule binding at, or near to, the origin of
transcription, and thus sterically preventing the access of the RNA polymerase to the initiation
site (Maurer et al., 1980; Takeda et af., 1983). There is some evidence for a ‘direct contact’
mechanism in positive gene regulation by 1 cIIand P22 1 cI(Johnson et al., 1981). This suggests
that negatively charged residues clustered on the surface of the activator are in close contact
with a positive area on the RNA polymerase. Mutations in P22 cI at 42 (glutamic acid) prevent
stimulation of transcription. Comparable residues occur in CAP at 192 (aspartic acid) and 191
(glutamic acid).
The spoIID and gerE sequences contain glutamic acid residues in similar locations to those in
the P22 cI and CAP sequences. Glutamic acid residues occur at position 312 in spoIID and at
positions 71 and 73 in gerE. Like the CAP protein, the DNA-binding units are towards the Cterminal of the chain.
As described in the Introduction, sporulation operons in general are expressed in a dependent
sequence which has been partly elucidated. The stage 0 operons, designated as spoOA, spoOB etc.,
are now known to be expressed in the vegetative cell and can therefore be excluded in a
consideration of the developmental process. They must be presumed to determine basic
structures and functions of the cell which do not affect growth but which are a pre-requisite for
sporulation although they are not specifically concerned with it. It is known that mutations in
stage 0 operons have widely varying pleiotropic effects, for example, on competence of cells in
genetic transformation, on susceptibility to infection of the cells by a variety of phages, on the
secretion of exo-enzymes, etc.
Fig. 3 (on facing page). Structures based on modelling studies of protein products of spoZZD and gerE.
(a)The modelled structure of the spZZD DNA-binding region based upon the 1 cro structure. Overlap
of similar atoms is shown in white; the 1 cro structure is in blue, and the spoIZD structure in orange.
(b)The modelled structure of the gerE DNA-binding region based upon the 1 cro structure. Overlap of
similar atoms is shown in white; the 1 cro structure is in blue, and the gerE structure orange,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
2388
S . K . H O L L A N D , S . C U T T I N G A N D J . MANDELSTAM
9
gerE K P S L T K R E R E V F E L L V Q D K T T K E I A S E L F I S E K T V R N H I S N ~ Q K L G ~ G R S Q A - - - ~ L L ~ G E ~ L
O . . . o ~ O
..000
00
0
0 0 0 0 o m
lexA MKALTARQQEVFDLIRDH- -- -- ISQTGMPPTRAEIAQRLGFRSPNAAEEHLKaLARKGVIEI
1
1
.
o.
o
--- - -
.
..
gerE MKEKEFQSKPSLTKREREVFELLVQDKTTKEIAS---ELFISEKT
0. O
.O
...
0.0.0
0.0
E GHYQVDPSLFKPNAD-FLLRVS GMGMKDI GIMDGDLLAVHKT
lexA
95
1
.
gerEMKEKEFQSKPSLTKREREVFELLVQDKTTKEIASEL
ooao
OO0.0.
1rP
fPEREALGTRVRIVEELLRGEMSQmLmL
0
Identical residue
0 Similar residue
Fig. 4. Homology between the amino acid sequences of gerE and other DNA-binding proteins. The
numbers indicate the position of the first residue of the segment in each sequence.
Dimer contacts recA cleavage
Carboxyl domain
Fig. 5. An overall scheme to show the homologies between gerE and the E. coli lexA repressor. The
diagram is drawn to scale based upon the polypeptide length. Significantly homologous regions are
indicated by a row of dots between adjacent sequences and boxes indicate different functional domains
as labelled. The single dotted line in the lexA sequence represents gaps incorporated to allow for
maximum alignment. The 1 clsequence is drawn to indicate the alignment of the phage repressors with
the lexA repressor. The dotted boxes indicate homologous regions between the gerE sequence and the
carboxyl region of the lexA repressor.
Coming now to the ‘real’ developmental operons, i.e. those that determine events from stage
I1 onwards, we find that spoIIA, consisting of three genes (Fort & Piggot, 1984), has one which
codes for a protein having strong homology with known sigma factors in both B . subtilis and E.
coli (Errington et al., 1985). This seems to be one of the earliest operons in the sequence and the
proper functioning of all the subsequent operons in the dependent sequence is blocked partly or
completely by mutations in it. Another operon, spoIIG, codes for a product which, after
proteolytic processing, is converted to the sigma factor
Whereas the polypeptide coded for
by one of the genes in spoIIA has been presumed to code for a sigma factor on the basis of
homology, the gene in spoIIG encodes a product that has actually been found in association with
RNA polymerase (Losick & Pero, 1981). It is curious that the only two sigma factors that are
apparently concerned with sporulation are both involved in the earliest stage, stage 11, and are
expressed independently of one another (Clarke & Mandelstam, 1987).
To explain the successive expression of a large number of operons we either have to assume
that there are a large number of other sigma factors which are non-homologous and whose
amino acid sequences have therefore as yet given no clue to their function (see Doi & Wang,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
Role of spoIID and gerE in sporulation
2389
1986), or that other types of molecule are involved in the transcription of sporulation operons. It
is in this context that the findings reported in this paper acquire significance. Of some 15
sporulation genes sequenced in this laboratory three, spoIID, gerE and spoIIIC, appear to have
the requisite properties of DN A-binding proteins. The simplest assumption is that they might
function as activators rather than repressors. The first of these, the spoIID product, is of special
interest because, at its N-terminal end, 20 residues of 26 are hydrophobic (see Fig. l), which
suggests that the polypeptide might enter or cross membrane structures. In addition, the Nterminal sequence has the characteristics of a signal peptide that is cleaved from proteins that
are exported from the cell. Since we know that some sporulation operons are expressed in the
spore compartment it follows that a molecule of some type must cross from the mother cell into
the forespore where it sets in train a separate branch of the dependent sequence. From its
structure it seems that the protein encoded by spoIID might have the requisite properties.
The operon that appears to code for the DNA-binding protein, gerE, is also monocistronic. It
constitutes part of the dependent sequence of events expressed in the mother cell, and mutations
in it affect some of the properties of the spore that develop at a relatively late stage (S. Cutting &
J. Mandelstam, unpublished).
In summary, the predicted protein products of three genes (spoIID,spoIIICand gerE) that are
known to be regulatory share the helix-turn-helix structure at the C-terminal end that
characterizes known activators and repressors binding to DNA. We have already emphasized
that the structure of a protein cannot be accurately predicted from its sequence, and the
probability of error in modelling of the kind we have used is generally estimated at 45 % to 50%
(say, 47.5%). In fact, when the sequences are treated by an alternative method (J. Wootton,
personal communication) the spoIID product does not appear to have the helix-turn-helix
motif; gerE, on the other hand, is a potential DNA-binding protein but the relevant domain on
this model contains the residues 29 to 48 and is thus in a different part of the molecule to that
indicated by us. However, other parts of the gerE product show homology to portions of the trpR
repressor outside the DNA-binding motif. The regulatory protein encoded by spoIIIC has not
been analysed by either method, but its close homology to trpR makes it very likely that it would
score as a DNA-binding protein in both programs.
The probability of mis-classification in all three proteins is only about 0.1 (0.4753).On this
basis, we propose as a hypothesis that the helix-turn-helix structure is responsible for the
regulatory function of these molecules and that its activity is exerted by DNA-binding. It is in
any case apparent that these proteins are distinct from the homologous set of sigma factors
described by Doi & Wang (1986) and that they account, at least in part, for the sequential
expression of sporulation operons.
We are indebted to Dr J. Errington for the sequence data for spolIIC, to Dr Colin Blake for advice on the
preparation of the manuscript, and to Dr R. Pickersgill for help with the graphics.
This work was done by S. K. H. on an MRC Research Studentship and by S. C. on an SERC-CASE award; it
was supported by the SERC.
REFERENCES
AIBA,H., FUJIMOTO,
S. & OZAKI,N. (1982). Molecular
cloning and nucleotide sequencingof the gene for E.
coli CAMP receptor protein. Nucleic Acids Research
10, 1345-1 361.
ANDERSON,
J. E., PTASHNE,
M. & HARRISON,
S. C.
(1985). A phage repressor-operator complex at 7 A
resolution. Nature, London 316, 596-601.
ANDERSON,
W. F., OHLENDORF,
D. H., TAKEDA,
Y. &
MATTHEWS,
B. W. (1981). Structure of the cro
repressor from bacteriophage I and its interaction
with DNA. Nature, London 290, 754-758.
BRENT,
R. & PTASHNE,
N. (1981). Mechanism of action
of the lexA gene product. Proceedings of the National
Academy of Sciences of the United States of America
78, 4204-4208.
C. (1976). The nature of the accessible and
CHOTHIA,
buried surfaces in proteins. Journal of Molecular
Biology 105, 1-14.
CHOU,P. Y. & FASMAN,
G. D. (1974). Prediction of
protein conformation. Biochemistry 13, 222-245.
'S. & MANDELSTAM,
J. (1987). Regulation of
CLARKE,
stage I1 of sporulation in Bacillus subtilis. Journal of
General Microbiology 133,2371-2380.
COOTE,J. G. & MANDELSTAM,
J. (1973). Use of
constructed double mutants for determining the
temporal order of expression of sporulation genes in
Bacillus subtilis. Journal of Bacteriology 114, 12541263.
J. (1986). The nucleotide
CWING, S. & MANDELSTAM,
sequence and transcription during sporulation of the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
2390
S . K . HOLLAND, S. CUTTING A N D J . MANDELSTAM
gerE gene of Bacillus subtilis. Journal of General
repressor and cro - components of an efficient
Microbiology 132,3013-3024.
molecular switch. Nature, London 294,217.
DAYHOFF,MrO., BARKER,
W. C. &HUNT,L. T. (1983).
LIM,V. I. (1974). Structural principles of the globular
Establishing homologies in protein sequences.
organization of protein chains, a stereochemical
Methods in Enzymology 91,524-545.
theory of globular protein secondary nature. Journal
DOI,R. H. & WANG,L.-F. (1986). Multiple procaryotic . . of Molecular Biology 88, 857-872.
ribonucleic polymerase sigma factors. Microbiologi- LOPEZ-DIAZ,
I., CLARKE,
S. & MANDELSTAM,
J. (1986).
spoIID operon of Bacillus subtilis: cloning and
cal Reviews 50, 227-243.
EBRIGHT,R. H., COSSART,P., GIQUEL-SANZEY,
B. &
sequence. Journal of General Microbiology 132,341BECKWITH,J. (1984). Molecular basis of DNA354.
LOSICK,R. & PERO, J. (1981). Cascades of sigma
sequence recognition by the catabolite gene activafactors. Cell 25, 582-584.
tor protein and detailed inferences from 3 mutations
that alter DNA-sequence specificity. Proceedings of
A. L. (1969). Change in the
LOSICK,R. & SONENSHEIN,
template specificity of RNA polymerase during
the National Academy of Sciences of the United States
of America 81,7274-7278.
sporulation of Bacillus subtilis. Nature, London 224,
ELIASON,
J. L., WEISS,M. A. & PTASHNE,
M. (1985).
35-37.
NH2-terminal arm of phage l repressor contributes
MANDELSTAM,
J. (1976). Bacterial sporulation : a
energy and specificity to repressor binding and
problem in the biochemistry and genetics of a
determines the effects of operator mutations. Proprimitive developmental system. Proceedings of the
ceedings of the National Academy of Sciences of the
Royal Society B193, 89-106.
J. (1987). Dependent
MANDELSTAM,
J. & ERRINGTON,
United States of America 82,2339-2343.
sequences of gene expression controlling spore
J. (1985).
ERRINGTON,
J., FORT,P. & MANDELSTAM,
formation in Bacillus subtilis. Microbiological SciDuplicated sporulation genes in bacteria : implicaences 4,238-244.
tions for simple developmental systems. FEBS
Letters 161, 556-562.
MAURER,R., MEYER,B. J. & PTASHNE,M. (1980).
FORT,P. & PIGGOT,P. J. (1984). Nucleotide sequence
Gene regulation of the right operator (0,) of
of sporulation locus spoIIA in Bacillus subtilis.
bacteriophage l. Journal of Molecular Biology 139,
Journal of General Microbiology 130,2 147-2 153.
147- 161.
GARNIER,
J. D., OSGUTHORPE,
J. & ROBSON,B. (1978).
MCKAY,D. B., WEBER,I. T. & STEITZ,T. A. (1982).
Analysis of the accuracy and implications of simple
The structure of the catabolite gene activator protein
methods for predicting the secondary structure of
at 2.9 8, resolution. Journal of Biological Chemistry
globular proteins. Journal of Molecular Biology 120,
257,95 18-9524.
97-120.
NEEDLEMAN,
S. B. & WUNCH,C. D. (1970). A general
GROSSCHEDL,
R. & SCHWARZ,E. (1979). Nucleotide
method applicable to the search for similarities in
sequence of the cro-cII-oop region of bacteriophage
the amino acid sequence of two proteins. Journal of
434 DNA. Nucleic Acids Research 6,867-88 1.
Molecular Biology 48, 443-453.
C. (1980). Nucleotide OHLENDORF,D. H. & MATTHEWS,B. W. (1983).
GUNSALUS,
R. P. & YANOFSKY,
sequence and expression of E. coli trpR; the
Structural studies of protein-nucleic acid interacstructural gene for the trp aporepressor. Proceedings
tions. Annual Review of Biophysics and Bioengineering
of the National Academy of Sciences of the United
12,259-284.
States of_America 77,7 118-7 12 1.
OHLENDORF,
D. H., ANDERSON,
W. F., LEWIS,M.,
PABO,C. 0. & MATTHEWS,
VON HIPPEL, P. H. & BERG, 0. G. (1986). On the
B. W. (1983). Comparispecificity of DNA-protein interactions. Proceedson of the structures of cro and repressor proteins
ings of the National Academy of Sciences of the United
from bacteriophage A. Journal of Molecular Biology
States of America 83, 1608-1 6 12.
169,757-769.
HOCHSCHILD,
A. & PTASHNE,M. (1986). Homologous
ORCUTT, B. C., DAYHOFF,M., GEORGE,D. G. &
interactions of repressor and l c r o with the l
BAKER,W. C. (1984). ALIGN Score Program of
operator. Cell 44, 925-933.
Protein Identijcation Resource, National Biomedical
Foundation, Georgetown University Medical Center,
HORII,T., OGAWA,
T. & OGAWA,
H. (1981). Nucleotide
Washington, USA.
sequence of the lexA gene of E. coli. Cell 23,689PAM,C. 0.& LEWIS,M. (1982). The operator-binding
697.
J.
HRANUELI,D., PIGGOT,P. J. & MANDELSTAM,
domain of repressor: structure and DNA recognition. Nature, London 298,443-447.
(1974). Statistical estimate of the total number of
operons specific for Bacillus subtilis sporulation. PABO,C. 0. & SAUER,R. T. (1984). Protein-DNA
recognition. Annual Review of Biochemistry 53,293Journal of Bacteriology 119,684-690.
HSIANG,M. W., COLE,R. D., TAKEDA,Y. & ECHOLS,
321.
H. (1977). The amino acid sequence of cro regulatory
PABO, c . o., KROVATION,
w . , JEFFREY,A. & SAUER,
protein from bacteriophage lambda. Nature, London
R. T. (1982). The N-terminal arms of 1 repressor
270,275-277.
wrap around the operator DNA. Nature, London 298,
HURSTEL,
S., GRANGER-SCHNARR,
M., DAUNE,M. &
441-443.
SCHNARR,M. (1986). In vitro binding of lexA
PIGGOT,P. J. & COOTE,J. G. (1976). Genetic aspects
repressor to DNA: evidence for the involvement of
of bacterial endospore formation. Bacteriological
the amino-terminal domain. EMBO Journal 5, 793Reviews 40,908-962.
PIGGOT,P. J. & HOCH,J. A. (1985). Revised genetic
798.
JOHNSON,
A. D., POTEETE,
A. R., LAUER,G., SAUER,
linkage map of Bacillus subtilis. Microbiological
Reviews 49, 158-179.
R. T., ACKERS,G. K. & PTASHNE,M. (1981). l
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01
Role of spoIID and gerE in sporulation
239 1
PTASHNE,M., JEFFREY, A., JOHNSON, A . D., MAURER, National Academy of Sciences of the United States of
America 73, 804-808.
R., MEYER,B. J., PABO,C. O., ROBERTS,T. M. &
SAUER,R. T. (1980). How 1 repressor and cro work. STEITZ,T. A., OHLENDORF,D. H., MCKAY, D. B.,
ANDERSON,W. F. & MA'ITHEWS,B. W. (1982).
Cell 19, 1-1 1.
Structural similarity in the DNA-binding domains
RYTER,A., SCHAEFFER,P. & IONESCO,H. (1966).
of the catabolite gene activator and cro repressor
Classification cytologique, par leur stade de blocage,
proteins. Proceedings of the National Academy of
des mutants de sporulation de Bacillus subtilis
Sciences of the United States of America 79, 3097Marburg. Annales d l Institut Pasteur 110, 305-31 5 .
3100.
SAUER, R. T. & ANDEREGG,R. (1978). Primary
D . H., ANDERSON,W. F. &
structure of 1repressor. Biochemistry 17, 1092-1 100. TAKEDA,Y., OHLENDORF,
MATTHEWS,B. W. (1983). DNA-binding proteins.
SAUER,R. T., YOCUM,R. R., DOOLITTLE,
R. F., LEWIS,
Science 221, 1020-1026.
M. & PAW, C. (1982). Homology among DNAbinding proteins suggests the use of a conserved TURNER,S. M., ERRINGTON,J. & MANDELSTAM,J.
(1986). Use of a lacZ gene fusion to determine the
super-secondary structure. Nature, London 298,447dependence pattern of sporulation operon spoIIIC in
451.
spo mutants of Bacillus subtilis: a branched pathway
SAVVA,D. & MANDELSTAM,J . (1986). Synthesis of
of expression of sporulation operons. Journal of
spoIIA and spoVA mRNA in Bacillus subtilis. Journal
General Microbiology 132, 2995-3003.
of General Microbiology 132, 3005-301 1.
R. W. & KIM,S. H. (1978). a-Helix-double
SCHEVITZ,R. W., OTWINOSKI,Z., JOACHIMIAK, A., WARRANT,
helix interaction shown in the structure of a
LAWSON,C. L. & SIGLER,P. B. (1985). The threeprotamine-transfer R N A complex and a nucleodimensional structure of trp repressor. Nature,
protamine model. Nature, London 271, 13C135.
London 317, 782-786.
SEEMAN,N. C., ROSENBERG,
J . M. & RICH,A. (1976). WHARTON,R. P. & PTASHNE,M. (1985). Changing the
binding specificity of a repressor by re-designing an
Sequence-specific recognition of double helical
a-helix. Nature, London 316, 601-605.
nucleic acids by proteins. Proceedings of the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 02:20:01