. x ‘ .
m J 4,. .afiké . , V .
. V ,. . . : .

a 35%..“ aka. .
U. Vt.- 1 . . n
. r . q! :31:— _
. . In» (4.. . . . .
. .
A . . _h ‘TA,
V any. .
.5: .9?

a
i may
:finfim‘m. my
mmiwwu .4

.. r" ...,..i.,..#.
fix. . New... :dauunumflwm

fig?
rfinfih.“ . .4
313%.?

.vflr

emu. , , v.

V. . .

2m,

. r

. xr

any... , 31... 5
. .. ea, ‘ at
a. a. 4m xm .mmfim;
,.

er. v.1

: 1 .
. v1
‘ :fixfigh t
.z .
3 In...» .. v
2'
29:13.. ‘

. . “flaw .,
was»

.3... .
an“... w...

. iswm

frl! :.
I

, .
.2 .

, . ‘ > .. :.

. . .5335:

1.x ‘52:.”

 

 

.gCflbl

This is to certify that the

dissertation entitled

TRANSCRIPTIONAL REGULATION OF GENES ENCODING SPORE
COAT PROTEINS BY MOTHER-CELL SPECIFIC SIGMA K RNA
POLYMERASE DURING BACILLUS SUBTILIS SPORULATION

presented by
Hiroshi Travis Ichikawa

has been accepted towards fulfillment
of the requirements for

Ph .D. degree in War

 

/,
£1192, (Ami;

Major professor

 

Date g-lV’OQ

 

MS U is an Affirmative Action/Equal Opportunity Institution 0- 12771

 

LIBRARY
Mlchlgan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11m clam-gasp.“

 

TRANSCRIPTIONAL REGULATION OF GENES ENCODING SPORE COAT
PROTEINS BY MOTHER-CELL SPECIFIC oK RNA POLYMERASE DURING
BACILLUS SUBTILIS SPORULATION
By

Hiroshi Travis Ichikawa

A DISSERTATION
Submitted to
Michigan State University
in partial fulfilment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Biochemistry

2000

ABSTRACT
TRANSCRIPTIONAL REGULATION OF GENES ENCODING SPORE COAT
PROTEINS BY MOTHER-CELL SPECIFIC 0“ RNA POLYMERASE DURING
BACILLUS SUBTILIS SPORULATION
By

Hiroshi Travis Ichikawa

Sporulation of the Gram-positive bacterium Bacillus subtilis is a well established
model system to study temporal and spatial gene regulation. Upon starvation, B. subtilis
initiates sporulation. Asymmetrical cell division into the larger mother cell and the
smaller forespore is the first easily observed morphological change. Each compartment
expresses different sets of sporulation specific genes from the identical genome. As
sporulation proceeds, the forespore is engulfed by the mother cell, forming a free
protoplast inside the mother cell. Maturation of the forespore involves deposition of
spore coat proteins, which are encoded in cat genes. Later, the mature spore is released
by lysis of the mother cell. These interesting morphological changes are a consequence
of cascade activation of RNA polymerase 0 subunits, which allows each cell type to
sequentially express specific genes for sporulation. In the mother cell, in addition to CE
and OK, the sequential appearance of two DNA-binding proteins, SpoIIID and GerE,
regulates gene expression.

The research presented in this dissertation is focused on understanding the
regulatory role of GerE and SpoIIID in transcription by oK-containing RNA polymerase
during late stages of sporulation in the mother cell. GerE was demonstrated to be a
sequence-specific DNA-binding protein by DNase I footprinting, and mapping of these

sites revealed a consensus GerE binding sequence. Based on in vitro transcription

experiments, GerE directly activates transcription of several cot genes by (JK RNA
polymerase, and this activation, in part, involves interaction with the C-terminal domain
of the or subunit of RNA polymerase.

In addition to its positive effects, GerE also initiates a negative feedback loop that
inhibits transcription of SigK, the gene that encodes UK. This was demonstrated by
comparing expression of the sigK-lacZ fusion in wild-type cells and in a gerE mutant,
and by comparing the SigK protein level of these strains by Western blot analysis. GerE
was found to bind within the SigK promoter region near the transcriptional start site and
repress transcription by (JK RNA polymerase.

The combined action of GerE and SpoIIID regulates expression of some cot
genes. Analysis of lacZ fusions and mRNA levels showed that cotB is expressed slightly
earlier than cotX, whereas cotC expression lags behind that of cotX. This pattern of
expression can be explained by different levels of GerE activation and/or SpoIIID
repression of the three cot genes, as observed in in vitro transcription experiments. The
results support a model in which a decreasing level of SpoIIID and an increasing level of
GerE during sporulation set the timing and the level of expression of these cot genes,

which may be important for proper assembly of the spore coat.

TABLE OF CONTENTS

LIST OF FIGURES .................................................................................................... viii
LIST OF ABBREVIATIONS .................................................................................... xi
INTRODUCTION ...................................................................................................... 1
CHAPTER I
Overview of Morphological Changes during Sporulation ............................. 5
Initiation of Sporulation ................................................................................. 8
Asymmetric Septum Formation and Chromosome Segregation .................... 8
Activation of (TF .............................................................................................. 10
Activation of GE .............................................................................................. 14
Phagocyte-like F orespore Engulfment ........................................................... 18
Activation of 0G .............................................................................................. 19
Activation of oK .............................................................................................. 21
Hierarchical Regulatory Cascade in the Mother Cell ..................................... 25
CHAPTER II
Abstract .......................................................................................................... 29
Introduction .................................................................................................... 29
Materials and Methods ................................................................................... 30
Results ............................................................................................................ 33
Purification of GerE ........................................................................... 33
Mapping of the 5' terminus of cotC mRNA ....................................... 33
GerE binds to specific sequences ....................................................... 35

iv

GerE stimulates cotB and cotC transcription in vitro ......................... 35

Effects of GerE on in vitro transcription of other

mother-cell-expressed genes ................................................... 37
Discussion ...................................................................................................... 38
CHAPTER III
Abstract .......................................................................................................... 44
Introduction .................................................................................................... 44
Materials and Methods ................................................................................... 45
Results ............................................................................................................ 46
Organization and expression of transcripts from the
cot VWXYZ gene cluster .......................................................... 46
Regulation of transcription of the cot VWXYZ cluster ........................ 49
In vitro transcription of genes in the cot VWXYZ cluster .................... 49
GerE binding sites in the vax, PX and PYZ promoter regions ............ 49
Discussion ....................................................................................................... 51
CHAPTER IV
Abstract .......................................................................................................... 56
Introduction .................................................................................................... 56
Materials and Methods ................................................................................... 57
Results ............................................................................................................ 57
Location of the GerE-binding site in the SigK promoter region ......... 57
GerE inhibits sigK expression in vivo ................................................ 58
GerE inhibits cotD transcription in vitro ............................................ 58
Location of the GerE-binding site in the cotD promoter region ........ 58

The upstream GerE-binding site in the cotD promoter is

necessary for activation of transcription in vitro

but not for repression .............................................................. 59
Discussion ...................................................................................................... 59
CHAPTER V

Abstract .......................................................................................................... 63
Introduction .................................................................................................... 63
Materials and Methods ................................................................................... 63
Results ............................................................................................................ 64

Role of upstream GerE binding sites in cat gene transcription .......... 64

Cat genes exhibit different patterns of expression ............................. 64

A lower concentration of GerE activates cotB transcription
than cotX or cotC transcription ............................................... 65

SpoIII is a potent repressor of cotC transcription ............................... 66

SpoIIID binds to specific sites in the cotC and cotX

promoter regions .................................................................... 66
Discussion ...................................................................................................... 67
CHAPTER VI

APPENDIX A

Abstract .............................................................................................. 76

Introduction ........................................................................................ 77

Materials and Methods ....................................................................... 79

Results ................................................................................................ 81

aCTD plays a role in GerE activation of cotB,
cotC, and cotX transcription ....................................... 81

aCTD is not required for activation of SigK
transcription by SpoIIID ............................................. 82

vi

Discussion .......................................................................................... 83

BIBLIOGRAPHY

vii

CHAPTER I
Figure 1.
Figure 2.

Figure 3.

Figure 4.

CHAPTER 11
Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

CHAPTER III

Figure 1.

Figure 2.

LIST OF FIGURES

The morphological stages of sporulation ...........................................
Activation of oF ..................................................................................

Processing of pro-oE in the mother cell depends on a signal
from the forespore ..................................................................

Processing of pro-oK ...........................................................................

The GerE binding sites in the 5' regions of cotB and cotC .................
Production of GerE in E. coli .............................................................
Mapping the 5' terminus of cotC mRNA............. ..............................
GerE footprints in cotB and cotC DNAS ............................................
GerE stimulates cotB and cotC transcription in vitro .........................

Effects of GerE on cotD, SigK, gerE and COM
transcription in vitro ..............................................................

Alignment of promoters transcribed by GK RNA polymerase ............

Regulatory effects of SpoIIID and GerE during stages IV and V
of sporulation ..........................................................................

Regulatory interactions controlling the levels of SpoIIID ..................

Northern hybridization analysis of the cotVWXYZ gene cluster .........

Reverse transcription mapping of the 5' termini of the 1.6 kb

cotVWX mRNA (Prl), the0.6 kb cotX mRNA (Pr2),

viii

7

12

16

23

33

33

34

36

37

37

39

40

40

47

Figure 3.

Figure 4.

Figure 5.

Figure 6.

CHAPTER IV

Figure 1.
Figure 2.

Figure 3.

Figure 4.
Figure 5.

Figure 6.

Figure 7.

CHAPTER V

Figure 1.
Figure 2.

Figure 3.

0.6 kb cotY mRNA (Pr3) and

the 1.4 kb cotYZ mRNA (Pr4) ................................................ 48
cot VWXYZ gene transcription in B. subtilis wild-type strain

JH642 (WT) and gerE mutant lAA-l (gerE) ......................... 49
Effects of GerE on transcription of promoters in the

cotVWXYZ cluster by (JK RNA polymerase in vitro ............... 50
GerE footprints in the vax, Px and PYZ promoter regions ................. 50

Alignment of promoters in the cot VWXYZ cluster with the consensus
sequence for oK-transcribed promoters (A) and alignment

of GerE binding sites (B) ........................................................ 51
GerE footprints in the SigK promoter region ...................................... 57
sigK-lacZ expression in wild-type and gerE mutant cells .................. 58

Levels of pro-oK and 0K during sporulation of Wild-type

and gerE mutant cells ............................................................. 58
Effect of GerE on cotD transcription in vitro ..................................... 59
GerE footprints in the cotD promoter region ...................................... 59

Effect of GerE on transcription in vitro of a cotD template lacking
the upstream GerE binding site ............................................... 59

Regulatory interactions between mother cell-specific transcription
factors and a model for regulation of cotD transcription

at different times during sporulation ...................................... 59
GerE binding sites in the cotB, cotC, and cotX promoter regions ...... 64
Expression of cot-IacZ fusion ............................................................. 65
Levels of cotB, cotC, and cotX mRNA during sporulation ................ 65

ix

Figure 4. Effect of GerE on cotB, cotC, and cotX transcription in vitro ............ 65

Figure 5. Effect of SpoIIID on cotB, cotC, and cotX transcription in vitro ....... 66
Figure 6. SpoIIID footprints in the cotC promoter region ................................. 67
Figure 7. SpoIIID footprints in the cotX promoter region ................................. 67

Figure 8. A model showing how the combined action of SpoIIID and GerE may
regulate cot gene context of interactions between mother cell-specific

transcription factors ................................................................ 68
APPENDIX A
Figure 1. In vitro transcription of cotB, cotC, and cotX and gerE by heterologous
RNA polymerase with or without intact aCTD ................................. 87
Figure 2. Effect of aCTD deletion on transcriptional activation ....................... 89

Figure 3. In vitro transcription of SigK and gerE by heterologous RNA
polymerase with or without intact aCTD ............................... 91

ADP
AMV

ATP

A660
bp
BRL
BSA
C-terminus
CTP
Da
dATP
dCTP
DNA
DNase
aNTD
dNTP
DPA
DSM
DTT
ECL
EDTA

GDP

LIST OF ABBREVIATIONS

adenosine-5'-diphosphate

avian myeloblastosis virus
adenosine-5'-triphosphate
absorbency at 660 nm

base pair

Bethesda Research Laboratories
bovine serum albumin

carboxy terminus
cytosine-5'-triphosphate

dalton
deoxyriboadenosine-S'-triphosphate
deoxyribocytosine-S'-triphosphate
deoxyribonucleic acid
deoxyribonuclease

N-terminal domain of a-subunit
deoxyribonuceotide

dipicolinic acid

Difco sporulation medium
dithiothreitol

Enhanced Chemiluminescence
(Ethylenedinitrilo)tetraacetic acid

guanosine-5'-diphosphate

xi

GF P

GTP

IPTG

kb

kDa

LB

mRNA

O.D.

ONPG

ORF

PAGE

PMSF

RNase
SDS

SM

Tx

tRNA
Tris-HCl
UTP
vol.

v/v

w/v

green flourescent protein
guanosine-5'-triphosphate

isopropyl B-D thiogalactopyranoside
kilobases

kilodalton

Luria-Bertani

messenger ribonucleic acid

optical density
o-nitrophenol-B-D-galactoside

open reading frame

polyacrylamide gel electrophoresis
phenylmethylsulfonyl fluoride
ribonucleic acid

ribonuclease

sodium dodecylsulfate
Sterlini-Mandelstam

x hours after the onset of sporulation
transfer ribonucleic acid
tris(hydroxymethyl)aminomethane hydrochloride
uracil-5'-triphosphate

volume

volume per volume

weight per volume

xii

INTRODUCTION

Upon starvation, B. subtilis initiates sporulation. Asymmetrical cell division into
a larger mother cell and a smaller forespore is the first easily observed morphological
change. Each compartment expresses different sets of sporulation-specific genes. As
sporulation proceeds, the forespore is engulfed within the mother cell, forming a free
protoplast. Maturation of the forespore involves deposition of spore coat proteins, which
are encoded in cot genes that are expressed in the mother cell. Later, the mature spore is
released by mother cell lysis. These interesting morphological changes are a consequence
of cascade activation of RNA polymerase 0 subunits, which allows each cell type to
sequentially express specific genes for sporulation. In the mother cell, in addition to oE
and OK, the sequential appearance of two DNA-binding proteins, SpoIIID and GerE,
regulates gene expression.

The experiments in Chapter II demonstrate that GerE is a sequence-specific DNA-
binding protein that directly activates transcription of cotB and cotC and inhibits
transcription of sigK (encoding OK) and COM, by oK-containing RNA polymerase. This
work was a collaboration between L. Zhang, S. Roels and R. Losick at Harvard
University and R. Halberg, L. Kroos, and myself at Michigan State University. I was
responsible for constructing a plasmid containing the cotC promoter region and
performing DNase I footprinting experiments to map a GerE binding site that is close to
the cotC transcriptional start site. This work was published in the Journal of Molecular
Biology.

Experiments in Chapter III demonstrate that GerE binds to and activates

transcription from the promoters of the cotVWXYZ cluster. We also established the GerE

binding site consensus sequence from the GerE binding sites that were mapped. This
work was a collaboration between J. Zhang and A. Aronson at Purdue University and R.
Halberg, L. Kroos, and myself at Michigan State University. I was responsible for
preparing GerE from an E. coli strain, performing DNase I footprinting of GerE on the
cot VWX, cotX, and cotYZ promoter regions, and measuring GerE activation of
transcription from the cotVWX, cotX and cotYZ promoters using in vitro transcription
assays. This work was published in the Journal of Molecular Biology.

The effects of transcription inhibition by GerE are described in Chapter IV. These
experiments provide evidence that the level of Si gK products is negatively regulated by
GerE . A feedback loop is initiated when gerE is transcribed by oK-containing RNA
polymerase because GerE represses sigK transcription. The experiments in Chapter IV
also demonstrate that GerE binding to a site centered at -25.5 relative to the cotD
transcriptional start site is involved in repression of cotD transcription. This work was a
collaboration with R. Halberg, a former graduate student in the Kroos lab, who
contributed mapping of one of the two GerE binding sites in the cotD promoter region.
This work was published in the Journal of Biological Chemistry.

The timing and level of gene expression of individual genes during sporulation is
tightly controlled by transcription factors. GerE binds to two sites in each of several cot
gene promoters. The importance of GerE binding to the upstream site in the cotB, cotC,
and cotX promoters was examined by fusing a lacZ gene to promoters with or without the
upstream GerE binding site. Also in Chapter V, I demonstrate that the pattern of
expression from the cotB, cotC, and cotX promoters appears to be regulated by the
combined action of SpoIIID and GerE. The order of appearance of transcription factors

in the mother cell is SpoIIID first, followed by GK and finally GerE. In vitro transcription

experiments showed that, of the three genes, cotB is most sensitive to activation by GerE,
and cotC is most sensitive to repression by SpoIIID. This may explain why, as the level
of GerE rises, cotB is expressed slightly earlier than cotX and cotC, and as the level of
SpoIIID falls, cotC is the last of the three genes to be fully expressed. This work was
published in the Journal of Biological Chemistry.

In vitro transcription experiments in Appendix A were performed with
heterologous RNA polymerase consisting of B. subtilis oK and E. coli core RNA
polymerase with or without the C-terminal domain of the a subunit (E. coli core subunits
were provided by A. Ishihama of the National Institute of Genetics in Japan). The levels
of cotB, cotC, and cotX transcription in the presence of GerE were lower when
heterologous RNA polymerase was missing the C-terminal domain of the a subunit,
indicating that GerE transcriptional activation of these genes may involve interaction with
the C-terminal domain of a. This dependency on the C-terminal domain of a was not
detected when sigK transcription was tested in the presence of SpoIIID, suggesting that

SpoIIID transcriptional activation of sigK involves a different mechanism.

CHAPTER I

Literature Review

Literature review

Gene regulation is one of the fundamental activities that governs life. Living
organisms regulate gene expression spatially and temporally during development and in
response to numerous internal and external cues. Therefore, programmed gene regulation
is a subject of great interest. Sporulation of Bacillus subtilis is an excellent system to
study spatial and temporal gene regulation because of its well-established experimental
tractability and a large database of genetic information.

Overview of Morphological Changes during Sporulation.

Nutrient depletion, cell density, the Krebs cycle, DNA damage and DNA
synthesis generate signals that trigger initiation of sporulation of the Gram-positive soil
bacterium Bacillus subtilis. The first morphological change of sporulation is
characterized by formation of an axial filament in which the two chromosomes from the
last round of DNA replication become aligned across the long axis of the cell (Stage I)
(Figure 1). Next, an asymmetrical septum divides the cell into a lager mother cell and a
smaller forespore (Stage 11). Each compartment receives a copy of the genome. The
septum migrates toward the forespore pole to engulf and finally pinchs off the forespore
as a free protoplast within the mother cell (Stage III). Maturation of the forespore occurs
while it resides in the mother cell. Cell-wall like material known as cortex, which is
thought to be involved in attaining or maintaining the dehydrated and heat-resistant state
of the spore, is synthesized in the space between the two membranes surrounding the
forespore (Stage IV). As the final step of spore maturation in the mother cell, spore coat
proteins are synthesized and deposited around the surface of the forespore (Stage V). The

spore coat consists of a lamellar inner layer and an electron-dense outer layer, providing a

Figure l. The morphological stages of sporulation. The stages are designated by
Roman numerals. The wavy lines are chromosomes. The sporangia are surrounded by
cytoplasmic membrane (thin line) and a cell wall (thick line). The developing spore
(stage IV-VI) is encased in a layer of cortex (light stippling) and a coat layer (dark
stippling). The four specific sporulation sigma factors are shown in the cells where they

become active.

thick, protective barrier that encases the mature spore. Maturation of the forespore also
includes coating the forespore chromosome with small acid-soluble proteins and
condensation of the DNA into a doughnut-like structure, conferring resistance to UV
radiation (Stage VI). Finally, lysis of the mother cell releases the spore (Stage VII). The
sporulation process takes six to ten hours under laboratory condition. This series of
complex morphological changes is controlled by sequential activation of five sporulation-
specific RNA polymerase 0 factors (reviewed in Haldenwang 1995; Stragier and Losick
1996)
Initiation of Sporulation

Environmental and physiological signals culminate in phosphorylation of a key
regulatory protein, SpoOA. SpoOA is a member of the response regulator family of two-
component regulatory systems. There are at least three proteinqkinases that transfer
phosphate to SpoOF. The phosphate of SpoOF~P is transferred to SpoOA via SpoOB.
SpoOA~P can be dephosphorylated by SpoOE and SpoOF~P can be dephosphorylated by
RapA (The phosphorelay is reviewed in Grossman 1995). SpoOA~P binds to various
promoter regions to activate transcription of genes that are important to initiate
sporulation. Activation of transcription of the spoIIE gene and the spoIIA operon leads to
forespore-specific gene expression (Bramucci et al. 1995). Activation of the two-cistron
operon spoIIG leads to mother-cell specific gene expression (Baldus et al. 1994; Bird et _
al. 1996). There is also one or more unknown gene(s) regulated by SpoOA~P that
determines the switch of the position of cell division from medial to polar.
Asymmetric Septum Formation and Chromosome Segregation

During binary fission, daughter chromosomes first segregate and then a

symmetric septum forms where FtsZ molecules assemble into a ring structure (Z-ring)

(Levin and Losick 1996). Upon entering sporulation, daughter chromosomes align along
the long axis of the cell (forming the axial filament) and Z-rings are formed near both
poles of the cell (Levin and Losick 1996). One or more unknown gene(s) under SpoOA
control is responsible for Z-ring formation (Levin and Losick 1996). The product of
spoOH, a sigma subunit (oH ) controls the next set of genes, whose products make the
polar septum. However, no genes in this group have yet been discovered. Several genes
(ftsZ, divIB, divIC and ppr ) involved in binary fission during grth are also involved
in asymmetric septum formation during sporulation (Beall and Lutkenhaus 1989; Levin
and Losick 1994; Yanouri et al. 1993).

The polar sporulation septum forms around the axial filament and then one of the
chromosomes is translocated across the septum into the forespore by the product of the
SpoIIIE gene, which resembles proteins that mediate conjugational plasmid transfer in
Streptomyces (Wu et al. 1995). A chromosomal position effect on transcription by the
forespore-specific oF RNAP in spoIIIE mutant cells can be explained by the finding that
only 30% of the origin-proximal region of one chromosome is present in the forespore
after the septum forms (Wu and Errington 1994). The product of spoOJ may play a role in
anchoring the origin region of the chromosomes to the cell poles (Sharpe and Errington
1996). In disporic mutant sporangia, there are two polar forespore-like compartments
into which chromosomes are translocated, leaving an anucleate compartment in the
middle (Setlow et al. 1991). During vegetative growth, polar division is prevented by the
products of the last two ORFs of the div] VB locus that are homologous to E. coli minC
and minD (Barak et al. 1998). In a minCD mutant, however, SpoIIIE pumps DNA out of
the polar compartment, resulting in production of anucleate minicells (Sharpe and

Errington 1995). During sporulation, there is a thin mid-cell septum found in minD

mutants (Barak et al. 1998). Therefore, MinCD complex may play an important role in
preventing symmetric septum formation during sporulation.

Effects of spoIIE on sporulation septum formation have been reported.
Mutations in the extreme C-terminal portion of SpoIIE make the sporulation septum
develop an aberrantly thicker layer of peptidoglycan and the time of septum formation is
delayed (Feucht et al. 1996). In a minD mutant, the thin septum is co-localized with
SpoIIE-GFP by fluorescent microscopy (Barak et al. 1998). In wild-type cells, SpoIIE is
required for the SpoOA-directed formation of polar Z-rings composed of FtsZ (Khvorova
et al. 1998). Likewise, F tsZ is required for formation of polar E-rings composed of
SpoIIE (Levin et al. 1997). These rings form early in sporulation and one becomes the
site of septum formation whereas the other ring disappears (Bouche and Pichoff 1998).
Activation of (IF

After the asymmetric division, OFRNAP is responsible for transcribing genes in
the forespore. Several lines of evidence indicate that CF is produced prior to septation
(Arigoni et al. 1996; Gholamhoseinian and Piggot 1989; Lewis et al. 1994). How its
activity is confined to the forespore has been a focus of great interest. The activity of (IF
is key event in CF activation (Garsin, Duncan et al. 1998). Once phosphorylated,
SpoIIAA cannot bind to SpoIIAB; SpoIIAB remains bound to 0F and keeps it inactive.
Proteins similar to SpoIIAA and SpoIIAB regulate the stress response 0 factor, 08, in
response to cellular ATP levels (Voelker et al. 1995). It is unknown whether the
ATP/ADP ratio is different in the forespore and mother cell compartment when oF
becomes active in the forespore (Magnin et al. 1997).

Serine phosphatase activity of SpoIIE dephosphorylates SpoIIAA~P, resulting in

activation of CF. SpoIIE has membrane spanning segments in its N-terminal domain and

10

Figure 2. Activation of OF. A. The anti-o factor SpoIIAB(IIAB) binds to CF in the
presence of ATP, but ADP favors binding to SpoIIAA (IIAA) and release of OF. B.
SpoIIAA can be phosphorylated (IIAA-P) by SpoIIAB in the presence of ATP, resulting
in CF inhibition, as SpoIIAA-P can not bind to SpoIIAB. SpoIIE (IIE) preferentially
dephosphorylates SpoIIAA in the forespore, allowing it to bind to SpoIIAB, which

permits OF activation as shown in (A). Adopted from Kroos et al., 1999.

11

 

ll

A ADP
IIAA + llAB-oF—A llAA-llAB + 0‘

v.
ATP
8 IIE phosphatase
m
OFINHIBITION "AA-P IIAA OFACTIVATION
V

"AB “"1888

12

the C-terminal portion extends into the cytosol (Arigoni et al. 1999). SpoIIE-mediated
dephosphorylation is the rate-limiting step in activation of OF. Overexpression of SpoIIE
triggers premature activation of (TF (Arigoni et al. 1996; F eucht et al. 1996). Moreover,
active OP is found in the mother cell compartment of a SpoIIIE null mutant in which
SpoIIE persists in the mother cell longer than usual (Sun et al. 1991). One group has
presented evidence that SpoIIE is sequestered to the forespore face of the sporulation
septum (F eucht et al. 1996; Wu et al. 1998). This would explain why oF becomes active
only in the forespore. However, another group has presented evidence that SpoIIE is
equally distributed to both sides of the septum (Duncan et al. 1995). These investigators
suggest that the smaller size of the forespore leads to a more rapid increase in the
concentration of unphosphorylated SpoIIAA, allowing oF activation in the foresore.
SpoIIE may require two checkpoints for indirect activation of CF by
dephosphorylating SpoIIAA~P. The first checkpoint is the formation of Z-rings at the
two potential polar division sites. SpoIIE co-localizes with FtsZ, forming E-rings.
SpoIIE phosphatase activity requires F tsZ and there is evidence that SpoIIAA~P is
already dephosphorylated by SpoIIE at the stage of ring formation since
unphosphorylated SpoIIAA accumulates in a divIC mutant which undergoes ring
formation but not septation (King et al. 1999). However, oF remains inactive in the divC
mutant, suggesting that septum formation serves as a second checkpoint. Based on the
effects of modified and mutant forms of SpoIIE, it has been proposed that afier
SpoIIAA~P is dephosphorylated by SpoIIE, SpoIIAA is retained by SpoIIE (or by
another protein in the septum) so that SpoIIAA is unable to react with the SpoIIAB-oF
complex to activate oF (Kinget al. 1999). Release of SpoIIAA is proposed to be

dependent on completion of the septum (King et al. 1999).

13

 

Activation of GE

Soon after the activation of CF in the forespore, OE becomes active only in the
mother cell. Compartmentalization of CE activity has been demonstrated directly
visualizing B-galactosidase or the green fluorescent protein expressed under the control of
a GE dependent promoter, using immunoelectron and immunofluorescence microscopy
(Driks and Losick 1991; Harry et a1. 1995; Webb et al. 1995). Also, disruption of the
forespore copy of (IE controlled sporulation genes has been shown to have no effect on

sporulation (Errington 1993). The inactive form of GE is called pro-aE and is 27 amino

 

acids longer at the N-terminal end than active OE. Pro-oE is the product of spoIIGB, the
downstream gene in the in the two-cistron spoIIG operon. SpoOA~P activates
transcription from the spoIIG promoter, stimulating the vegetive RNAP containing oA
prior to septation. Activation of pro-oE occurs only on the mother cell side and involves
SpoIIGA, the product of the first gene in the spoIIG operon. SpoIIGA is believed to be a
protease that processes pro-oE to (IE (Stragier et al. 1988).

Some of the genes required to activate oE include spoIIAA, SpoIIAB, spoIIE,ftsZ
and divIC, which are also involved in the activation of Up. This suggested a requirement
for an additional gene that is controlled by OF. The gene was found to be spoIIR (or cst)
(Karow et a1. 1995; Londono-Vallejo and Stragier 1995). When spoIIR, SpoIIGA and
spoIIGB are expressed during exponential growth, efficient pro-oIE processing is observed
(Londono-Vallejo and Stragier 1995). SpoIIR has an apparent signal sequence and is
secreted from B. subtilis if expressed during the exponential phase of grth (Hofrneister
et a1. 1995). When produced in the forespore, SpoIIR is thought to cross one septal
membrane and act as signal molecule in the space between the two septal membranes

(Hofineister et a1. 1995; Karow et al. 1995; Londono-Vallejo and Stragier 1995;

14

Figure 3. Processing of pro-oE in the mother cell depends on a signal from the
forespore. The top part depicts a sporulating cell just after polar septation. Pro-OE and
SpoIIGA (IIGA) made earlier are thought to associate with both septal membranes. The
bottom part shows the cell slightly later. 0F RNAP in the forespore transcribes the gene
encoding SpoIIR (IIR). SpoIIR probably crosses the membrane surrounding the
forespore and we speculate that it activates SpoIIGA by promoting dimerization, resulting
in pro-ol3 processing in the mother cell. Conceivably, SpoIIR might also activate
SpoIIGA in the septal membrane adjoining the forespore (not shown in the bottom part),
but normally this does not occur or, if it does, another mechanism inhibits OE activity in

the forespore. Adopted from Kroos et al., 1999.

15

Londono-Vallejo 1997). There, it activates pro-oE processing by SpoIIGA, an integral
membrane protein (In et a1. 1997; Hofmeister 1998). SpoIIGA is believed to be a
receptor/protease with a receptor domain that interacts with SpoIIR and a protease
domain, that cleaves pro-oE in the mother cell (Londono-Vallejo 1997).

How is the mother-cell specific activity of GE established? Three patterns of
subcellular localization of the spoIIGB product have been observed by
immunofluorescence microscopy (Hofmeister 1998). Pro-oE is found first in association
with the cytoplasmic membrane. After asymmetric division, pro-~oE accumulates at the
septum. Finally, the processed active GE is found only in the mother cell cytoplasm.
Amino acids of the pro-sequence are responsible for membrane association (Hofmeister
1998). Pro-oE is in a complex with SpoIIGA or with a protein that depends on SpoIIGA;
however, pro-oE associates with membranes even in the absence of SpoIIGA (Hofineister
1998). Interestingly, the product of gp fused to the pro-sequence of sigE is found to be
localized on the mother-cell side of the septum (In and Haldenwang 1999). This
localization is pro-sequence dependent as deletion of the first 15 amino acids of pro-oE
abolishes the specific localization of the GF P fusion protein (In and Haldenwang 1999).
Localization of pro-oE to the mother-cell side of the septum could explain how oE activity
is confined to the mother cell. The SpoIIIE DNA translocase is somehow involved in
confining 0'3 activity to the mother cell because mutations in SpoIIIE allow oE activity to
also accumulate in the forespore (Pogliano et al. 1997). One hypothesis is that SpoIIIE is
necessary for translocation of pro-oE to the mother-cell side of the septum, as well as for
DNA translocation into the forespore (Hofmeister 1998; Ju and Haldenwang 1999).
Alternatively or in addition, SpoIIIE mediated DNA translocation into the forespore may

activate a protease that destroys pro-OE, SpoIIGA, and any OE inadvertently formed in the

17

forespore (Ju et al. 1998; Ju and Haldenwang 1999).
Phagocytic-like Forespore Engulfment.

Several morphological steps are observed during the process of forespore
engulfment (Perez et al. 2000; Sharp and Pogliano 1999). First, peptidoglycan between
the septal membranes becomes thinner or is removed. This septal thinning proceeds from
the middle toward the edges of the septum. The elimination of this rigid structure leads
to the bulging of the forespore compartment into the mother cell, presumably reflecting
the higher osmolarity of the forespore cytoplasm. Next, the comers of the septum
migrate toward the forespore pole and the membrane fuses there. This process results in
a protoplast-like forespore within a double membrane which freely floats in the mother
cell. Abundant de novo synthesis of membrane components is required for forespore
engulfment by the mother cell and this is likely to be under the control of GE and OF.

The spoIIB and spa VG genes are directly or indirectly controlled by oH (Resnekov
et al. 1995). SpoIIB is found in the sporulation septum during septal biogenesis, but is
degraded once the septum is complete (Perez et al. 2000). The sporangia of spoIIB
mutant displays a transient engulfment defect in which the forespore pushes through the
septum and bulges into the mother cell, however, the sporangia completes engulfment in
slower speed compared to the wild-type sporangia (Perez et al. 2000). The investigators
suggest that SpoIIB facilitates the rapid and spatially regulated dissolution of septal
peptidoglycan (Perez et al. 2000). SpoIIB, in conjunction with SpoVG, is required for
initiation of degradation of the septal peptidoglycan. But the synergistic mechanism of
SpoIIB and SpoVG action during engulfment is not clearly understood. Inactivation of
spa VS compensates for the absence of spoIIB and spa VG, suggesting an additional level

of regulation (Resnekov et al. 1995). A possible role for SpoIIB/SpoVG is to antagonize

18

the action of SpoVS, which may prevent cell wall degradation (Stragier and Losick
1996). Inactivation of oE-controlled genes such as spoIID, spoIIM and spoIIP leads to a
blockage in the completion of cell wall dissolution at the periphery of the septum
(Frandsen and Stragier 1995; Lopez-Diaz et al. 1986; Smith and Youngman 1993). The
biochemistry of these gene products is not well understood, however, structural analysis
predicts that these proteins are membrane-associated.

Although the initiation of migration of the mother-cell membrane around the
forespore is OF dependent, all other genes required for degradation of septum
peptidoglycan (other than spoIIB/spa VG) are controlled by GE (Stragier and Losick 1996).
One of the genes controlled by 0': is spoIIQ (Londono-Vallejo 1997). Its product is
involved in completion of engulfment process. Most of the SpoIIQ protein is located
outside of the forespore membrane (Londono-Vallejo 1997). SpoIIQ appears to facilitate
the wrapping movement of the mother-cell membrane around the forespore (Londono-
Vallejo 1997). It is also possible that SpoIIQ is involved in dissociating the connection
between the forespore membrane and the polar cell wall at the initiation of engulfment
(Londono-Vallejo 1997).

The final stage of engulfment is to pinch off the forespore as the migrating
membrane meets and fuses at the forespore pole. SpoIIIE may be involved in this process
based on its localization pattern (Sharp and Pogliano 1999). Moreover, the two functions
of SpoIIIE, DNA translocation and completion of engulfment, are separable by
mutational analysis (Sharp and Pogliano 1999).

Activation of 0“
The spoIIIG gene encodes 06. 0G controls the expression of a large set of genes in

the engulfed forespore, including the ssp genes whose products are the so-called small

19

acid soluble proteins (SASP) (Cabrera-Hemandez and Setlow 2000; Sun et al. 1989).
Transcription from the spoIIIG promoter is initially oF-controlled but oG RNAP also
recognizes this promoter and thus, autoregulates its own transcription (Partridge and
Errington 1993).

There is about an hour delay between activation of 0F and initiation of spoIIIG
expression (Partridge and Errington 1993). This time lag is not due to the weakness of
the spoIIIG promoter. Rather, it depends on (IE-controlled events in the mother cell (Sun
et al. 1991). Although a spoIIB spa VG double mutant arrests sporulation immediately
after polar septation, there is no effect on oF-dependent spoIIIG transcription, suggesting
that no further morphological development is required for activation of spoIIIG
transcription (Stragier and Losick 1996).

Placing the spoIIIG coding sequence under the control of the oF-controlled
spoIIQ promoter uncouples oG synthesis from dependency of SpoIIIG transcription on
OE. In this system, there was no premature 06 activity (Stragier and Losick 1996). There
is growing evidence that 00 activity is antagonized by SpoIIAB, the same anti-o factor
that inhibits OF activity (Kellner et al. 1996). In addition to SpoIIAB and completion of
engulfment, the vegetatively expressed spoIIIJ and the eight products of the G's-controlled
spoIIIA operon are required for 0‘3 post-transcriptional activation (Kellner et al. 1996).
Some evidence suggests that SpoIIAB might be preferentially degraded in the forespore
in a spoIIIA-dependent fashion (Kirchman et al. 1993).

do RNAP transcribes spo VT, whose product appears to be a DNA-binding protein.
SpoVT regulates transcription of genes under the control of CG. SpoVT also negatively

regulates spoIIIG transcription (Bagyan et al. 1996).

20

Activation of 0".

The last sporulation-specific 0 factor to be activated is UK, which is first expressed
as an inactive precursor called pro-oK (Lu et al. 1990). Processing of pro-oK is blocked in
mutants that fail to activate 00 (Lu et al. 1990). However, expressing spoI VB in the
forespore using a oF-dependent promoter eliminates the need for 0G and is sufficient to
allow processing of the N-terminal 20 amino acids of pro-oK to produce active 0"
(Gomez, Cutting et al. 1995). One of the products of the (IE-controlled spoI VF operon,
SpoIVF B, contains amino acid sequences conserved in a newly recognized family of
putative metalloproteases, and mutations in these sequences inhibit pro-oK processing
(Rudner et a1. 1999; Kroos et al. 1999). SpoIVFB is believed to be a thermolabile protein
as it requires the other product of the spa] VF operon, SpoIVFA, to function at 42°C but
not at 30°C (Cutting et al. 1990). The requirement for the SpoIVB-dependent forespore
signal in processing of pro-oK can be bypassed by some mutation in the C-terminal
portion of SpoIVFA (Cutting, Oke et al. 1990). Therefore, in addition to its role in
stabilizing SpoIVFB, SpoIVFA appears to inhibit SpoIVFB activity until the inhibition is
released by the signal from the forespore (Resnekov and Losick 1998). Another inhibitor
of SpoIVFB activity is the product of the (IE-controlled bofA gene (Resnekov and Losick
1998). There is evidence that the concentration of SpoIVB is critical for releasing
SpoIVFB inhibition since modification of the concentration of SpoIVB in the forespore
influenced the timing of pro-oK processing (Stragier and Losick 1996). SpoIVFA,
SpoIVFB and BofA are thought to be localized to on the membrane surrounding the
forespore (Green and Cutting 2000; Kroos et al. 1999). SpoIVB might reside on the outer
face of the forespore membrane, anchored by a single membrane-spanning domain.

SpoIVB has a putative serine protease catalytic site and may be self-processed to smaller

21

Figure 4. Processing of pro-0K. The upper diagram show a sporangium in which
engulfment of the forespore has been completed. Black dots represent pro-OK, which
associates with the mother cell membrane and the outermost membrane surrounding the
forespore. The latter membrane is thought also to contain SpoIVFB-SpoIVFA-BofA
complexes, shown only in the enlarged view (lower diagram). oG RNAP in the forespore
transcribes the gene encoding SpoIVB, a protein thought to cross the innermost
membrane surrounding the forespore or, perhaps, to be anchored in this membrane.
SpoIVB both signals processing of pro-oK to begin and plays a role in formation of the
germ cell wall. SpoIVFB appears t be the protease that processes pro-OK' oK RNAP
transcribes a gene(s) necessary for cortex formation and many genes encoding proteins

that form the spore coat (not shown). Adopted from Kroos et al., 1999.

22

 

 

 

 

   
 
  
  

{Sty
as:

J

    

8,
*Q

 

"V-

be

\\germ cell wall
cortex

23

forms that freely diffuse in the space between the two membranes surrounding the
forespore (Oke et a1. 1997). It is possible that activation of SpoIVFB (and therefore
processing of pro-OK) occurs as the mature SpoIVB cleaves the external domains of
SpoIVFA or BofA (Stragier and Losick 1996). The coupling between oK activation and
forespore signaling is very important because premature oK activation in the mother cell
leads to decreased efficiency of sporulation and release of germination defective spores
(Cutting et al. 1990).

Unlike SpoIVFB, a SpoIVFB-GF P fusion protein is able to accumulate in
vegetatively growing cells and this fusion protein alone is sufficient to process pro-oK to
0K. Expressing SpoIVFA in this system increases the SpoIVFB-GE P level, presumably
by stabilizing the fusion protein. Addition of BofA to this system inhibits pro-oK
processing and this depends upon SpoIVFA. Recent results with this system suggested
that SpoIVFA is the primary inhibitor of SpoIVFB and that BofA stabilizes SpoIVF A
(Resnekov 1999).

A gene expressed in the forespore called bofC contains a putative signal sequence
and is also implicated in the regulation of pro-oK in the absence of 00, but SpoIVB is still
required (Gomez and Cutting 1996). It is thought that a very small amount of SpoIVB
possibly produced by orF RNAP in spoIIIG mutant cells is enough to activate pro-oK
processing, but only if BofC is absent. Hence, BofC appears to be an inhibitor of spoIVB
activity (Gomez and Cutting 1996). BofC appears to play an accessory role, at least
under laboratory conditions, because mutation of bofC does not affect the timing of oK
activation or the production of heat-resistant spores in cells with a functional spoIIIG

gene (Gomez and Cutting 1996).

24

Hierarchical Regulatory Cascade in the Mother Cell.

Once OE is activated, a hierarchical regulatory cascade of gene expression is
played out over the course of the next five to six hours in the mother cell. The
appearance of four transcription factors in the order OE, SpoIIID, 0K and finally GerE
constitutes the cascade (Zheng and Losick 1990). SpoIIID and GerE are mother-cell
specific DNA-binding proteins (Halberg and Kroos 1994; Zhang et al. 1997; Zheng et al.
1992; Zhang et al. 1994). The spoIIID gene is transcribed by (JE RNAP (Kunkel et al.
1989; Halberg and Kroos 1994). The product, SpoIIID, recognizes specific DNA
sequences in the promoter regions and open reading frames of bafA, spoIVCA, sigK, catD
and spa VD (Halberg and Kroos 1994). SpoIIID activates spoIVCA and sigK transcription
but represses bafA and spa VD transcription by GE RNAP in vitro (Halberg and Kroos
1994). SpoIIID also activates and represses sigK and catD transcription, respectively, by
oK RNAP (Halberg and Kroos 1994).

Production of oK involves multiple steps before the forespore-dependent
processing of pro-oK discussed earlier can occur (Kroos 1991). The sigK gene (encoding
pro-0K) is generated as a consequence of a sporulation specific chromosomal
rearrangement (Kunkel et al. 1990). The sigK gene is interrupted by an intervening
sequence of 48 kb of DNA known as skin, which contains the spoIVCA gene (Stragier
and Losick 1996). The product of spoIVCA is a site-specific recombinase that rearranges
mother-cell chromosomal DNA to generate the intact sigK gene (Sato et al. 1994).
Because of the DNA sequence similarity to bacteriophage PBSX, skin is thought to be a
cryptic prophage (Krogh et al. 1996). Excision of this DNA sequence is mandatory in
order to successfully complete sporulation (Kunkel et al. 1990). Transcription of sigK is

initially controlled by GE, but active oK also contributes to transcription of its own gene

25

(setting up a positive feedback loop) (Kroos et al. 1989; Halberg and Kroos 1994).

The mother cell is responsible for the production of spore cortex and coat
proteins. A model for coat morphologenesis may help explain the existence of the
hierarchical regulatory cascade in the mother cell. The (IE-controlled spaI VA gene
product is assembled on the forespore surface and is proposed to recruit other proteins to
form a scaffold structure. Another oE-controlled gene product, CotE, assembles on the
outside surface of the scaffold. Then, other coat proteins expressed under the control of
oK and in some cases GerE are deposited within the scaffold, forming the inner coat, or
are recruited by CotE to form the outer coat (reviewed in Driks 1999).

How does the newly activated a factor in a cascade replace the previous 0 factor
from the core RNA polymerase in order to redirect gene transcription? Such a problem
arises only if the RNA polymerase core enzyme is in limiting supply. A possible solution
is for the new a factor to have a higher affinity for core RNA polymerase than the
previous 0 factor. Such competition between 0 factors for a limiting amount of core
RNAP has been demonstrated between the vegetative oA and the minor factor oH in
growing cells and at the initiation of sporulation (Hicks and Grossman 1996). Some
evidence also suggests that oE displaces o" from core RNAP and that UK displaces oE as
sporulation proceeds (Ju et al. 1999). In addition, oK appears to initiate two negative
feedback loops that limit further production of oE as well as its own synthesis. One
feedback loop leads to inhibition of sigE transcription by 0" RNA polymerase (Zhang
and Kroos 1997; Zhang et al. 1999). One or more oK-controlled gene products are
thought to be involved in this loop. The second feedback loop involves the oK-dependent
GerE protein repressing sigK transcription (Zheng et al. 1992). A similar feedback loop

is also found in the forespore, as the gene encoding 0G is negatively regulated by the OO-

26

controlled SpoVT gene product (Bagyan et al. 1996).

27

 

CHAPTER II

Sporulation Regulatory Protein GerE from Bacillus subtilis Binds to and Can Activate or

Repress Transcription from Promoters for Mother-cell-specific Genes

28

Sporulation Regulatory Protein GerE from Bacillus subtilis
Binds to and Can Activate or Repress Transcription from
Promoters for Mother-cell-specific Genes

Liangbiao Zheng‘, Richard Halberg“, Steven Roels', Hiroshi Ichiltawa2
Lee Kroosz and Richard Losick‘

1Department of Cellular and Developmental Biology
The Biological Laboratories
Harvard University '
Cambridge, Massachusetts 02138, U .S.A.

2Department of Biochemistry
Michigan State University
East Lansing, Michigan 48824, U .S.A.

(Received 26 December 1991, accepted 3 April 1992)

The mother-cell line of gene expression during sporulation in Bacillus subtilis is a
hierarchical cascade consisting of at least four temporally controlled gene sets. the first three
of which each contain a regulatory gene for the next gene set in the pathway. gerE, a
member of the penultimate gene set, is a regulatory gene whose product is required for the
transcriptional activation of genes (coat protein genes cotB and cotC) in the last gene set.
The gerE product also influences the expression of other members of the penultimate gene
set (coat protein genes cotA and catD appear to be repressed and activated, respectively).
We now report that the purified product of gerE (GerE) is a DNA-binding protein that
adheres to the promoters for cotB and cotC. We also show that GerE stimulates cotB and
cotC transcription in vitro by RNA polymerase containing the mother-cell sigma factor 0".
These findings support the view that GerE is a positively acting, regulatory protein whose
appearance at a late stage of development directly activates the transcription of genes in the
last known temporal class of mother-cellexpressed genes. In addition, GerE stimulates catD
transcription and inhibits cotA transcription in vitro by a“ RNA polymerase. as expected
from in vivo studies, and. unexpectedly, profoundly inhibits in vitro transcription of the gene
(sigK) that encodes a". The effects of GerE on catD and sigK transcription are just the
opposite of the effects exerted by the earlier-appearing, mother-cell regulatory protein
SpoIIID. suggesting that the ordered appearance of first SpoIIID. then GerE, ensures
proper flow of the regulatory cascade controlling gene expression in the mother cell.

Keywords: sporulation; sigma factor; regulatory protein; Bacillus subtilis

 

1. Introduction

Following the formation of a transverse septum
at morphological stage II, gene expression during
the process of sporulation in Bacillus subtilis is
regulated differentially between the forespore and
mother-cell chambers of the developing sporangium
(De Lencastre & Piggot. 1979; Losick & Stragier,
1992). Thus. the transcription of certain genes is
restricted to the forespore, whereas the transcrip-
tion of other genes is limited to the mother cell.
Although some heterogeneity in the time of induc-
tion of gene expression in the forespore has been

(NIH-2836]”! rower-rs “AD/0

29

reported (Panzer et al., 1989: Sussman & Setlow,
1991), most forespore-expressed genes are switched
on co-ordinately and only a single regulatory gene,
spoIIIG, which encodes the forespore sigma factor
a‘3 (Karmazyn-Campelli et al.. 1989; Sun et al.,
1989), is known to be exclusively transcribed in the
forespore chamber of the sporangium. Gene
expression in the mother cell. in contrast. is rela-
tively complex, involvingthe expression of at least
four co-ordinately controlled gene sets. which are
switched on successively during the course of
sporulation (Cutting et al., 1989; Kunkel et al.. 1988.
1989: Sandman et al., 1988; Zheng & Losick, 1990).

© 1992 Academic Press Limited

 

These gene sets constitute a hierarchical cascade in
that the first three gene sets each contain a regula-
tory gene that governs the expression of the next
gene set in the pathway (Zheng & Losick, 1990).
spoIIID, a member of the earliest regulon, is
a regulatory gene whose product (a small,
DNA-binding protein; RE. is L.K, unpublished
results) turns on the transcription of the next set of
genes (Halberg & Kroos, 1992; Kroos et al., 1989;
Kunkel et al., 1989; Stevens 8!. Errington, 1990). One
of the genes in the spa] I I D-dependent class of co
ordinately controlled genes is sigK (Kroos et al..
1989; Kunkel et al., 1988, 1989), a composite gene
(generated from two truncated genes by a DNA
rearrangement in the mother-cell chromosome;
Kunkel et al., 1990; Stragier et al., 1989) that
encodes the mother-cell sigma factor a" (Kroos et
al., 1989). The sigK gene product (after a regulatory

step involving the conversion of its primary .

product, pro-rt". to the mature sigma factor;
Cutting et al., 1990; Lu et al., 1990) then turns on
the penultimate class of genes. This set of genes
includes the spore coat protein genes cotA and catD
(Sandman et al., 1988; Zheng & Losick, 1990) and
the regulatory gene gerE (Cutting et al., 1989;
Holland et al.. 1987), which encodes a product
(GerE) that is, in turn, required for the expression of
genes in the last known gene set in the mother-cell
line of gene expression (Zheng & Losick, 1990). Two
members of this gerE-dependent gene set are the
coat protein genes cotB and 0000 (Donovan et al.,
1987).

Here we are concerned with the role of gerE in the
hierarchical cascade of mother-cell gene expression.
gerE is inferred to be a transcriptional regulatory
gene because: (1) a gerE nonsense mutation (gerE36;
Cutting, 1986) has highly pleiotropic effects on
sporulation, causing the production of spores that
are lysozyme-sensitive, germination-defective and
aberrant in coat ultrastructure and protein com-
position (Feng & Aronson, 1986; Jenkinson & Lord,
1983; Moir, 1981); (2) gerE36 partially inhibits
expression of catD (Zheng & Losick, 1990) and
causes overexpression of cute! in rich sporulation
medium (Cutting et al., 1989; Sandman et al., 1988);
(3) as indicated above, yerE'36 prevents the
expression of cot genes B and 0 (Zheng &. Losick,
1990); and (4) the predicted product of gerE (GerE),
an 85 kDa polypeptide (Cutting 8t Mandelstam.
1986; Hasnain et al., 1985), contains a region of
similarity to the a-helix—B-tum—a-helix motif of
many procaryotic transcriptional regulatory pro-
teins (Holland et al., 1987) and exhibits high overall
similarity to the COOK-terminal region of certain
regulatory members (the ngA sub-group) of the
family of two-component sensor-regulator systems
in bacteria (Gross et al., 1989; Kahn & Ditta, 1991),
which includes the B. subtilis regulatory proteins
DegU and ComA (Henner et al., 1988; Kunst et al.,
1988; Weinrauch et al., 1989). GerE is also similar to
the COOK-terminal region of the Escherichia coli
regulatory protein MalT (Gross et al., 1989) and to
the COOK-terminal region of sigma factors, which is

30

involved in the recognition of the —35 region of
bacterial promoters (Kahn & Ditta, 1991).

On the basis of DNase I footprinting experiments
with purified GerE, we now report that the gerE
gene product is a DNA-binding protein that adheres
to the regulatory regions for cotB and cotC. We also
show that GerE greatly stimulates cotB and cotC
transcription in vitro by a‘ RNA polymerase, a
finding in support of the view that the appearance
of GerE at a late stage of sporulation directly acti-
vates transcription of these genes. Moreover, we
show that GerE affects the in vitro transcription of
several other mother-cell-expreswd genes, either

positively or negatively, findings that suggest a

functional analogy between GerE and the mother-
cell regulatory protein SpoIIID (Kroos et al., 1989).
However, the effects of GerE on catD and sick
transcription in vitro are just the opposite of the
efi‘ects exerted by SpoIIID. Because production of
a" RNA polymerase appears to cause a decrease in
the level of SpoIIID (Halberg & Kroos, 1992) and is
required for the transcription of gerE (Cutting et al.,
1989; and this work), we propose that a declining
level of SpoIIID and a rising level of GerE produce
a reinforced switch in the pattern of mother-cell
gene expremion during sporulation.

2. MaterialsandMethods

(a) Struinsaadplamids

E. coli K38 (HfrC trp thi 1*) carrying plasmid pGP1-2
(Tabor a Richardson, 1986) was maintained at 30°C in
LB medium containing 25 ug of ksnamycin/rnl.

The source of the gerE open reading frame was
pSGMU 101 (provided by J. Errington of Oxford
University; (hitting & Mandelstam. 1986). The gerE open
reading frame was released as a 0'5 kbf EcoRV/Xbal
fragmem and was cloned into the SmI/Xbai site of
p'l'll3 (Bethesda Research Laboratories) to create
plZ304. The EcoRV site is 67 bp upstream from the gerE
coding sequence. whereas the Xbal site is a polylinlrer site
adjacent to a B. subtilis Mbol site located 22"! bp down-
stream from the earl? open reading frame.

Plasmids pLZ304 and pT713 were introduced into K38
cells containing pGPl-2 by transformation and selection
on LB medium containing 50 pg of ampicillin/ml and
25ug of kanamycin/ml at the permissive temperature
(30°C).

Plasmids pLRKlOO and pBK18 containing the catD
and sigK promoter regions, respectively. served as
templates for in vitro transcription and have been
described previously (Kroos at al.. 1989). A 08 kb
EcoRI/PvuII fragment (Fig. 1) containing the cotB
promoter region (Zheng a Losick, 1990) was subcloned
into EcoRI/SmaI-digested replicative form Ml3mp18
(Yanisch-Perron et al.. 1985) and replicative form DNA of
the recombinant phage was used for in vitro transcription.
Plasmid pBD281 containing the cotC promoter region was
constructed as follows: the lacZ-cabcontaining BamHI
fragment of pSGMU3l (Errington. 1988) was cloned into
the Bell site of pBD95 (Zheng & Losick. 1990). creating
an in-frame fusion of cotC to lch in pBD239, then the

 

TAbbreviations used: kb, 10’ bases or base-pairs; bp,
baseopairs; nt, nucleotides. '

 

 

 

 

 

 

 

 

“I I.” - ”In .
m L /: <
r
'-:- -.;. «:- * «:- ,
mum . l————-1 I"
. .....
1 2 ,. '
we r I I 4—1
find!!! “at! .' Ill! 8.! N8 users:
100 h

Figure 1. The GerE binding sites in the 5' regons ofch and cotC. The Figure shows restriction maps ofDNA in the
vicinity of cotB and 0000. based on a previous report (Donovan er al., 1987) and other unpublished analysis. (Note that
the map of Saa3A sites is incomplete and only a single site is shown.) The positions of the open reading frames for each
gene are shown by the filled bars (the map includes only part of the cotB open reading frame). Also shown are the
nucleotide sequences of the promoter regions of both genes. The startsites of transcription are indicated by the arrows.
Regions of DNA that were protected by GerE from the action of DNase I are indicated above and below each strand.
'1'he2GerEbindingsitesinootCaredesignatedland2intheFigure.

eatC-iacZ-ealocontaining BamHI/Bgln fragment of
pBD239 was cloned into the Baal-II site of pBR322
(Bolivar et al.. 1977) to construct pBD259 in which the
HiadIII site upstream from the cotC promoter is proximal
to the IliadIII site of the vector. and finally pBD259 was

' with IliadIII and recircularired to construct
pBD281 in which the small HiadIII fragment was
deleted. pHIl was constructed by subcloning a 06 kb
HindIII/Xbal fragment from pBD281 (extending from
the HisdIII site upstream from 00:0 to an XbaI site in the
polylinker downstream from the former Bell site of cotC
(Fig. 1) into HindIII/XbaI-digested pUCl9 (Yanisch-
Perron at al., 1985). Restriction fragments from pHIl
were purified after electrophoresis on an agarose gel and
used as templates for in vitro transcription of cotC. The
gerE promoter-containing plasmid, p801“, was
constructed by 8. Cutting as follows: the 268 hp .4131
fragment from pSGMUIOI (encompassing the gerE
promoter region; Cutting & Mandelstam, 1986) was
subcloned into the 81361 site in the polylinker of
pSGMU31 (Errington. 1980), then excised as a
KpaI/BamHI fragment and inserted into
KpaI/BamHI-digested pUCl9 (Yanisch-Perron at al..
1985). Three com promoter-containing plasmids were
constructed by K. Sandman as follows: (1) pKS22 was
constructed by ligating HiadIII linkers to a 08 kb
[linen/Aral fragment from pKSll (encompassing the
co“ promoter region; Sandman er al., 1988). cleaving the
linkers with HiadIII. and subcloning the fragment into
HiadIII-digeated pIBI76 (International Biotechnologies.
Inc.); (2) pKS23 was constructed by digesting pKSl9
(Sandman er al.. 1988) with PstI and ligating to delete
B. subtilis DNA beyond 55 bp upstream from the com
tramriptional startsite: (3) pKS24 was constructed by
digesting pKSl9 (Sandman at al., 1988) with £1:on and

31

ligating to delete B. was. DNA beyond 115 bp
upstream from the com transcriptional startsite.

(b) Production of GerE is E. coli

Cultures of K38 cells containing pGPl-2 (bearing the
T7 RNA polymerase gene) alone. pGPl-2 and
pT713. or pGPl-2 and pLZ304 were grown at 30°C to an
A... of 03. Cells were induced by a temperature shift to
62°C for 20 min. Rifampicin was added to a final concen-
tration of 200 ug/ml. and the cells were incubated at 42°C
for 10 min and then at 30°C for 30 min. Cells were
collected by centrifugation and disso1ved in sample buffer
(Laemmli, 1970). The sample was denatured at 90°C for 2
min and fractionated by electrophoresis through an
SDS/polyacrylamide gel containing 15% acrylamide.
For the preparation of GerE, protein from induced K38
cells containing pGP1-2 and pLZ304 was subjected to
electrophoresis and the putative GerE band was cut from
the gel. GerE was then eluted from the gel slice and
renatured as described (Hager & Burgen. 1980). For the
preparation of control protein, protein from induced K38
cells containing pGPl-2 and p'l'713 was subjected to
electrOphoresis. A slice from the position corresponding to
thatofGerEwascutfromthegel.Proteinwasthen
eluted and renatured from the gel slice as described for
GerE.

(c)PnparationofDNAprobeslabeledatoaiyoseer-d

For the preparation of radioactive cotC probes, a 358 bp
[fluff/Bell fragment (Fig. 1) whose Hinfl terminus
had been rendered flush by use of the Klenow fragment
of DNA polymerase I was cloned into
HiacII/BamHI-digested pUC’l8 (Yanisch-Perron er al..

1985). This created plasmid pLZl275 in which the
Hindi“ site of the vector was upstream from the cotC
promoter. Plasmid pLZl275 was linearized with HiadIII
and then treated with alkaline phosphatase. Next. the
cotC-containing fragment was released from the pUC18
vector by digestion with Ele, which cuts at the end of
the polylinker next to the BamHI site. A probe labeled at
the HiadIII site in the non-transcribed strand was
prepared using phage T4 polynucleotide kinase and
[7-3 P]ATP. To prepare a probe labeled at the HiadIII
site in the transcribed strand. cotC was released as a
IliadIII/Smal fragment (SmaI also cuts in the polylinker
next to BamHI site) and was labeled by end-filling the
HiadIII terminus usin the Klenow fragment of DNA

polymerase I and [er-3 P}dATP. Additional cotC probes '

were prepared by digesting pHIl (see above) with Earl,
which cleaves in the 7th codon of cotC (Donovan et al.,
1987), labeling in the non-transcribed strand using the fill-
in reaction of the Klenow fragment of DNA polymerase I
and [a-"PldCI‘P or labeling in the transcribed strand by
treatment with alkaline phosphatase followed by phage
T4 polynucleotide kinase and [y-“P1ATP. then digesting
with EcoRV and EcoRI. and purifying the 239 bp
EcoRV/Earl fragment encompassing the cotC promoter
region after electrophoresis on a non-denaturing, poly-
acrylamide gel containing 8% acrylamide (Maniatis er al.,
1982).

For preparation of radioactive cotB probes, we took
advantage of a plasmid pUClB derivative called pBD136
(constructed by W. Donovan. unpublished results), which
contains a 08 kb Saa3AI/Sau3AI fragment that includes
the promoter and the NI-Iz-terminal coding region of the
cotB open reading frame cloned into the Basil-II site in an
orientation such that the polylinlrer EcoRI site was proxi-
mal to and upstream from the promoter. pBD136 was
digested with EcoRI. treated with alkaline phosphatase.
and the cotB-containing fragment was released by diges-
tion with HindIII. which cuts at the opposite end of the
polylinker. The non-transcribed strand probe was labeled
at the EcoRI site by the T4 polynucleotide kinase reac-
tion. To prepare the. transcribed strand probe. a cotB-
containing. EcoRI/PvaII fragment was purified (see
Fig. 1) and was then labeled by end-filling using the
Klenow fragment of DNA polymerase I and [a-“PldATP.

(a) DNase I footprinting

Tw0 different methods were used to carry out the
DNase I footprinting experiments. In method (1), the
conditions for the binding of GerE were the same as
described by Strauch et al. (1989). except that poly(dI-dC)
was added to a final concentration of 30 ug/ml. DNA
fragments labeled at one end were incubated in separate
experiments without protein. with control protein. or
with different amounts of GerE protein in 20 n1 reaction
mixtures for 15 min. Then lul of a 001 mg/ml DNase I
solution (BRL) was added to each reaction for l min. The
digests were stopped by adding 5 ul of stop solution
(01 rr-EDTA and 0-5 % SDS) and chilled on ice. The DNA
in each reaction was precipitated with 1 ml of ethanol and
with 02 ug of poly(dI-dC) as carrier. The precipitates
were dissolved in formamide loading buffer (Maniatis at
al.. 1982) and denatured at 90°C for 90 s. The samples
were then subjected to electrophoresis in an 8 rr-urea-
containing polyacrylamide gel.

In method (2) the conditions for the binding of GerE
were the same as in the in vitro transcription experiments
(see below), except that poly(dI-dC) was added to a final
concentration of 1-2 rag/ml. DNA fragments labeled at one
end were incubated at 37°C in separate experiments

without protein. with control protein. or with different
amounts of GerE in 42 ul reaction mixtures for 10 min.
Then 3 ul of 00004 mg DNase I/ml (Boehringer-
Mannheim) solution (prepared by diluting a stock solution
(Davis at al.. 1980) with buffer (20 msr-Tris-HCI. pH 8-0.
20 mu-MgClz. 20 mu-CaClz) was added to each reaction.
After 1 min. the digests were stopped by adding 50 ul of
buffer (100 mu-Tris'HCl. pH 8-0. 50 mrr-EDTA. 200 pg
yeast tRNA/m1) and incubating for 2 min at 65°C. The
DNA in each reaction was precipitated with 250 ul
ethanol. The precipitates were dissolved in formamide
loading bufl‘er (Maniatis er al.. 1982) and denatured by
boiling. The samples were then subjected to electro-
phoresis in an 8 rr-urea/polyacrylamide gel containing 69/0
acrylamide.

(e) DNA sequencing

End-labeled DNA probes were subjected to the
chemical cleavage reactions of Maxam in Gilbert (1980)
with a kit from New England Nuclear or as described
previously (Maniatis et al.. 1982).

. (f) In vitro transcription
a" RNA polymerase was partially purified from

, B. subtilis strain 80104 (trpCZ gerE.” SPfl::coM-lacZ) as

described (Kroos el al.. 1989). This enzyme was compar-
able in protein composition and in coiD- and sigK-tran-
scribing activities to fraction 24 shown in Fig. 2 of Kroos
el al. (1989). a" RNA polymerase was reconstituted from
gel-purified. renatured a" and B. subtilis core RNA poly-
merase as described previously (Kroos er al.. 1989).
Transcription reactions were performed as described
previously (Kroos er al.. 1989) except that heparin (6 ug)
was added 2 min after the addition of nucleotides to
prevent reinitiation. and after the reactions were stopped
10 rd of the reaction mixture was subjected to electro-
phoresis. [ts-”HUI? was the labeled nucleotide unless
indicated otherwise. After gel electrophoresis. transcripts
were detected by autoradiography and the signals were
quantitated using a Visage 110 Image Analyser
(BioImage).

(g) Primer extension analysis

RNA was prepared from sporulating cells as described
by Cutting er al. (1991a). Ia vitro synthesised cotC tran-
scripts were generated as described above (but without
radiolabeled nucleotide) and then precipitated with
ethanol and suspended in 25 pl of diethylpyrocarbonate-
treated water. In preparation for primer extension analy-
sis. a sample of the in vitro synthesized RNA (4 [11) was
treated with 5 units of DNase I (Pharrnacia) in buffer (20
msr-Tris-l-ICI. pH 7'6. 10 mu-MgClz) in a total reaction
volume of 5 al. The reaction was incubated at 37 °C for 10
min and then at 90°C for 10 min.

Primer extension was carried out by use of the 6000-
specific oligonucleotides Prl and Pr2 (Zheng a Losick.
1990). The oligonucleotides were 5'-endolabe1ed using
[y-"P]ATP and T4 polynucleotide kinase (BRL) as
described by Sambrook et al. (1989). For analysis of in
viva synthesized RNA. from 2 to 6 pmol of 5'-end labeled
oligonucleotide was incubated with 5 ug of total RNA.
and reactions were carried out as described by Roels et al.
(1992). For analysis of in vitro—generated cotC transcripts.
12 pmol of 5'-end-labeled Pr2 oligonucleotide was incu-
bated with 4 pl in vitro synthesised RNA (from above) in
10 pl of annealing buffer (50 mu-Tris-l-ICI. pH 7-6.
100 mu-KCI) at 90°C for 2 min and then at 47°C for

32

30 min: 5 ul of the primer2RNA hybrid solution was then
incubated with 5 units of phage M-MuLV reverse tran-
scripture (Pharmacia) at 47 °C for 45 min in a final volume
of 10 ul of reverse transcriptase buffer (50 msr-Tris'HCl.
pH 7'6. 60 mu—KCI. 10 mnrngClz. l msr—dithiothreitel
(D'I'I'). 2'5 ml of each dNTP. 1 unit placental RNase
inhibitor/rd (Pharmacis)) after which 7 ul of 95% form-
amide loading dye was added.

The 5'-end-labeled oligonucleotides were also used to
generate sequence ladders by the dideoxy chain termina-
tion method of Sanger et at. (1977). The products of
primer extension were subjected to electrophoresis in 6%
polyacrylamide slab gels containing 8 ill-urea.

3. Results
(a) Purification of GerE

GerE was purified by engineering E. coli cells to
express a cloned copy of the gerE gene using the T7
RNA polymerase/T7 promoter system of Tabor &
Richardson (1985). A DNA fragment containing the
peril open reading frame was placed under the
control of a phage '17 promoter by inserting into the
expression vector pT713 a gerE-containing segment
of B. subtilis DNA that extended 67 bp upstream
and 227 bp downstream from the per]? open reading
frame to create plasmid pLZ304 (see Materials and
Methods). Cells containing pLZ304, the vector
pT713. or neither plasmid, were grown at 30°C and
were then shifted to 42°C to induce transcription
from the phage T7 promoter. Total cellular proteins
from induced and uninduced cells were displayed by
electrophoresis through an SDS/polyacrylamide gel
(Fig. 2). In addition to normal cellular proteins.
induced cells of the pLZ304-containing strain (lane
F) produced a protein of 6 to 8 kDa, which was

  
 

A B C D E F
a ‘>.s .
$13332“;
..-é§r£§»
29>éfla—A‘:=_.f
.:=x
“main-A
18> annual-“m "
‘_‘.—
|4>E~_
mar-n
-aa-
6» .'
I

Figure 2. Production of GerE in E. coli. Total cellular
proteins were extracted from cells grown at 30°C (lanes A
to C) or from cells that had been shined to 42°C (lanes -D
to F) and were then resolved by electrophoresis in a
SDS/12% polyacrylamide gel. The cells were derivatives
of E. coli strain K38 containing pGPl-2 alone (lane A and
D). pGP1-2 and pT713 (lane B and E). or pGP1-2 and
pLZ304 (lane C and F). (pGPls2 contains the phage T7
RN A polymerase gene.) The positions of the molecular
mass markers (in kDa) are shown on the left.

33

absent in induced cells that lacked the gerE-bearing
plasmid (lane D) or that contained the plasmid
vector, pT713 (lane E). This protein was also absent
in cells that were not heat-induced (lanes A to C).
The size of the protein was consistent with that (8-5
kDa) deduced from the nucleotide sequence of the
gerE open reading frame (Cutting a Mandelstam,
1986). Because there was no other open reading
frame in the gerE-bearing B. subtilis insert in
pLZ304 that could encode a protein of this size. we
presume that the 6 to 8 kDa protein was GerE.

To purify GerE, the 6 to 8 kDa protein band was
excised from an SDS/polyacrylamide gel displaying
proteins from pLZ304-containing bacteria. Protein
was eluted from the gel slice and renatured as
described by Hager & Burgess (1980). As a control,
a gel slice corresponding in position to GerE was cut
from an SDS/polyacrylamide gel displaying total
cellular proteins from induced cells of pT713-
containing bacteria. Protein was eluted from the gel
slice and. renatured and is referred to as “control
protein”.

(b) Mapping the 6' terminus of cotC mRN A

To study the interaction of GerE with cotB and
cotC, it was first necessary to know the precise sites
from which transcripts from these genes originate.
Previously reported primer extension experiments
established the location of the 5' terminus of cotB
mRNA (see Fig. 1) but only provided a tentative
assignment for the 5’ terminus of cotC mRNA
(Zheng & Losick, 1990); an extension product of
127 nt that would correspond to an apparent 5'
terminus located 120 bp upstream from the cotC
initiation codon was obtained with an oligonucleo-
tide primer called Prl but no extension products
were observed with two other oligonucleotides
called Pr2 and Pr3 (see Fig. 1 of Zheng & Losick
(1990) for a description of the primers). Alerted
from the results of in vitro transcription experi-
ments (below) that the true'startsite of transcrip-
tion was actually very close to the beginning of the
cotC open reading frame (and hence that relatively
short extension products were to be expected). we
repeated the primer extension analysis and found
that Prl and Pr2 generated extension products of
33 and 68 nt, respectively, that corresponded (as
judged by electrophoresis alongside Prl and
Pr2-generated nucleotide sequencing ladders:
Fig. 3(a) and 3(b)) with a 5’ terminus that preceded
the initiation codon by only 26 bp (Fig. 1). These
extension products were specific to gerE” cells at a
late stage of sporulation in that Prl and Pr2 gener-
ated little or no extension products with RNA from
wild-type cells at an early stage of development or
with RNA from gerE mutant cells at a late stage of
development (Fig. 3(d)). Consistent with our earlier
results (Zheng 6t Losick, 1990). Prl generated the
previously observed 127 nt extension product in
addition to the 33 nt product (Fig. 3(a)), but direct
nucleotide sequencing of this extension product by
the use of dideoxynucleotides in the primer exten-

    
  

GATCP: GATCP: 12 CTAG
3? g“ ‘
*r‘a‘g
gu‘ O. 1
3 ' .
:f‘,.;_=--<112"v ..
~54; ' =3" ‘
7;; 3‘ a?“
i g. 3.: . S ..
a i A ‘3'. “t
.. A 3 O - A
_. , G :3 ‘-- A
_. c ,g a- G
" . ._ .G ; 32<6o , . C
= =‘ '1' -" 68> . 3 Go
r _ A g .-‘ 'r
.. G " ‘ A
_ - T =5. g a
- — I-« .‘ 1'
-
.- ‘ ;" .
" :2 s'
I '.
8
”<3 . " '
.
.. . - ‘ ‘
- - . .
.nr --

(d)
123 456 787919

Fig. 3.

34

sion reaction established that the 127 nt product
was not copied from 0010 mRNA but rather from
the transcript of another, similarly regulated gene
(data not shown). Finally, the previous failure to
observe an extension product with Pr3 is now
explained by the fact that this primer corresponds
to a sequence (see Fig. l of Zheng 8t Losick, 1990)
that is located upstream from the 5' terminus of
cotC mRN A.

(c) GerE binds to specific sequences

Preliminary gel retardation experiments
indicated that GerE binds to a HindIII/Sspl frag-
ment of 500 bp containing the 5’ region of cotC and
to an E’coRI/PvuII fragment of 635 bp containing
the 6' region of cotB (Fig. l; and Zheng, I990). To
localize the binding of GerE to 00:8 and 0000 more
precisely, DNase I protection experiments were
carried out with radioactive DNA probes separately
end-labeled on one or the other DNA strand. The
radioactively labeled DNAs were incubated with
GerE and then mildly treated with DNase I to
generate a spectrum of fragments. After the enzyme
digestion step, the DNA fragments were fraction-
ated by gel electrophoresis. Figure 4(a) displays the
pattern of fragments generated by enzyme treat-
ment of GerE-bound cotC DNA that had been
labeled on the non-transcribed strand (the upper
strand in Fig. l) at a HindIII site located upstream
from the promoter. The Figure shows that GerE
caused protection from the action of DNase I along
an approximately 16 bp stretch of DNA extending
from position - 126 to position - 141 relative to the
5' terminus of cotC RNA. No protection was
observed with control protein (lane 1). Likewise,
Figure 4(b) shows that GerE protected a 19 bp
stretch of DNA extending from position -l29 to
position — 147 on the transcribed strand (the lower
strand in Fig. l) of cotC and that no protection was
observed with control protein.

A second GerE binding site was mapped in the
0000 promoter region using DNA probes labeled at
the Earl site located downstream from the
promoter. Figures 4(c) and (d) show that GerE
protected an approximately 22 bp stretch of DNA
extending from position —56 to position —77 on

the non—transcribed strand (the upper strand in
Fig. l) and an approximately 21 bp stretch of DNA
extending from position —58 to position —78 on
the transcribed strand (the lower strand in Fig. 1),
respectively. In both cases, no protection was
observed with control protein. The regions
upstream from cotC that were protected from
DNase I digestion by GerE binding are indicated in
Figure l.

Analogous experiments. showed that GerE
protected a wide region of cotB from DNase I diges-
tion. Figure 4(a) shows that the GerE-protected
region on the non-transcribed strand (the upper
strand in Fig. l) was 40 to 50 bp in length, with the
strongest protection occurring between position
~41 and position -81. Figure 4(t) shows that a
similar region was protected on the transcribed
strand, the protected region being 47 bp in length
and extending from position -36 to position -82.
The region upstream from cotB that was protected
from DNase I digestion by the binding of GerE is
indicated in Figure 1.

(d) GerE stimulates cotB and cotC
mm in vitro

To test for efi'ects of GerE on transcription of 0018
and 0000 in vitro, linearised DNA templates were
transcribed in the presence of GerE or control pro-
tein with a" RNA polymerase partially purified
from a gerE mutant (see Materials and Methods). a"
RNA polymerase produced run-off transcripts of the
expected sizes from cotB in the presence of GerE
(Fig. 5(a), lanes 3 and 4). The signal was sevenfold
weaker in the presence of control protein (Fig. 5(a),
lane 2) or with no addition (Fig. 5(a), lane 1). 0‘
RNA polymerase reconstituted from gel-purified a"
and purified B. subtilis core RNA polymerase was
also stimulated by GerE to transcribe cotB (data not
shown). -

Partially purified 0‘ RNA polymerase failed to
transcribe a linearized plasmid (pBD261) bearing
the cotC promoter, even in the presence of GerE,
apparently because sequences located in the vector
portion of the plasmid compete with the sequences
located upstream from 0000 for binding to GerE
(data not shown). However, when a 0'5 lrb restric-

 

Plgun 3. Mapping the 5' terminus of cotC mRNA. (a) and (b) Results of high-resolution mapping of the 5' terminus of
calf} mRNA using the oligonucleotide primers Prl (a) and Pr2 (b) to prime cDNA synthesis from total RNA from PY79
(1130*) cells harvested 10 h after the onset of sporulation (tm). (c) Results of high resolution mapping of the 5' terminus of
cotC mRNA using the oligonucleotide primer Pr2 to prime cDNA synthesis from RNA transcribed in vitro with
o‘-polymerase in the absence (lane 1) or presence (lane 2) of GerE protein. The products of primer extension were
subjected to electrophoresis alongside nucleotide sequencing ladders generated with either the Prl or Pr2 primer. The
arrows indicate the position and size of the principal extension product(s) obtained with each oligonucleotide primer. The
open arrow in (a) indicates the position of an artifact band. seen with Prl but not Pr2. that is not the result of extension
of 0000 mRNA (see the text). The sequences shown in the vicinity of the 5' terminus correspond to the non-transcribed
DNA strand and the filled circles indicate the nucleotide corresponding to the 5' terminus. (d) The time course and
genetic dependence of the appearence of 00:0 mRNA. The primers Prl (lanes 1 to 6) and Pr2 (lanes 7 to 10) were used to
generate extension products from total RNA from sporulating PY79 (spo‘) cells harvested at 1, (lanes 4 and 8), t. (lanes
5 and 9), or in (lanes 6 and 10) or from K8450 (”536) cells harvested at 1, (lane 1), 1. (lane 2). or t", (lanes 3 and 7).

35

(o)

§G123456
z’iiilti

o- i
n

a: l
-__....

 

 

 

6123456
III!!!
:l..pn-
28:!!I!

om

'93.:3

-129> ..'

‘147> ..a.-.

.-.-.. 3'

—77>

-56>

-7s>

-ss>

(cl

123456

_-iQ

all?“

II" 0

It.
ll
.5531].

mm
mama‘s

“Osman;
Iltszrl
Il‘lhtllllll
IWOs

I I vl I-lIll
Imus

(d)

allllllll”
wriulllll‘”

~«u- assess-seese

i It! Ill. 4W6 -mi.~3wl 0‘
9.1%. “at I m

I

IIIIO
marinasw

N u Iv" at.“

'8!” I!”

l u art.- 1.

antral-M M“

P.“

l as auburn-Illossasuis+=;:f

G123456
"u-
5' ii-
‘. ----
41> -.I
.3 :t,
.: :-
‘a e
.- O
‘C '0.
I: E
-3 . . I '
1’3; ::§:
0: ,'.
t - .
3' ‘-o. 0
0‘ .
(f)
G
GA23456
3.

 

(a)

 

 

Figure 4. GerE footprints in cotB and 6060 DNAs. Radioactive DNA fragments separately end-labeled on the
transcribed or non-transcribed strand were incubated in separate reactions without protein (lane 6). with control protein
(lane 1), or with 1 ug (lane 2). 05 ug (lane 3), 01 ug (lane 4). and 0-05 rig (lane 5) ofGerE. After treatment with DNase I.
the partially digested DNAs were separated by electrophoresis through an 8 M-urealpolyacrylamide gel alongside a
sequencing ladder generated by chemical cleavage of the respective end-labeled DNAs. (a) and (b) Footprints of the non-
transcribed (upper strand in Fig. l) and the transcribed (lower) strands of cotC. respectively. using probes labeled at the
IliudIII site located upstream from the promoter. (c) and (d) Footprints of the non-transcribed (upper) and transcribed
(lower) strands of cotC. respectively, using probes labeled at the Earl site located downstream from the promoter. (e) and
(t) Footprints of the non-transcribed (upper) and transcribed (lower) strands of cotB. respectively. using probes labeled at
the EcoRI site located upstream from the promoter. The experiments of (a). (b). (e) and (l) were carried out using method
(I) in Materials and Methods. and the experiments of (c) and (d) were carried out using method (2). although in other
experiments (data not shown) similar footprinting results to those seen in (a) and (b) were obtained using method (2).

36

(o) (D)

234

-<
'4

Figure 5. GerE stimulates cotB and cod? transcription
in vitro. Template DNA (04 pmol) was transcribed with
partially purified a" RNA polymerase (02 ug) alone, or
with control protein or GerE (0.4 ug) added immediately
after the addition of RNA polymerase. Run-off tran-
scripts were electrophoresed in 5% polyacrylamide gels
containing 8 u-urea and were detected by autoradio-
graphy. Arrowheads denote the positions of run-off tran-
scripts of the expected sizes in each panel, as judged from
the migration of end-labeled DNA fragments of
Hopi-digested pBR322. (a) cotB transcription from repli-
cative form DNA of a recombinant phage containing the
cotB promoter region. The phage DNA was linearized with
Basil-ll (lanes 1 to 3, l77-base transcript) or HiadIII
(lane 4, 207-base transcript) and transcribed with 0‘ RNA
polymerase alone (lane 1), or with control protein (lane 2)
or GerE (lanes 3 and 4) added. (b) cotC transcription
from restriction fragments isolated from pill]. The
HindIII/Xbal fragment (lanes 1 to 3, 174cbase transcript)
or the HindIII/SaII f ent (lane 4, 180-base transcript)
were transcribed with RNA polymerase alone (lane 1).
or with control protein (lane 2) or GerE (lanes 3 and 4)
added.

tion fragment was used as the template for in vitro
transcription. 0‘ RNA polymerase produced run-off
transcripts of the expected sizes from cotC in the
presence of GerE (Fig. 5(b), lanes 3 and 4). A very
weak signal was observed in the presence of control
protein (lane 2) or with no addition (lane 1) in a
longer autoradiographic exposure than that shown
in Figure 5(b). Primer extension analysis of the
GerE-stimulated cotC transcript produced by 0‘
RNA polymerase in vitro demonstrated that it had
the same 5' terminus as cotC mRNA produced in
vivo (Fig. 3(c)). - -

(e) Efects of GerE on in vitro transcription
of other Wall-expressed genes

0" RNA polymerase partially purified from a gerE
mutant has been shown to transcribe from the 00:0
and sigK (previously called spa! VCB) promoters in
vitro (Kroos at al., 1989). GerE stimulated catD
transcription two- to threefold (Fig. 6(a)) and com-
pletely inhibited sigK transcription (Fig. 6(b)) by
partially purified a" RNA polymerase. Although the
effect of GerE on catD transcription was modest, a
two- to threefold stimulation was consistently
observed in four independent experiments (data not
shown). The level of stimulation was not further
enhanced by the use of twice as much GerE as that
employed in the experiment of Figure 6(a) (data not
shown).

The partially purified 0‘ RNA polymerase
produced run-off transcripts of the expected sizes
from gerE (Fig. 6(c), lanes 1 and 2) and from cold
(Fig. 6(d), lanes 1 and 2). These transcripts were
also produced by a" RNA polymerase reconstituted

(c) (D) (c) (a)
l 2 l 2 l 2 3 4 I 2 3 4 5 6
.5; >. i ‘v
; >13 an”
H -;‘ 5 Q ”9"
List “.rusr

Flgure 6. Effects of GerE on catD. sigK. gerE and com transcription in vitro. Linearized plasmid DNA (2 pg) was
transcribed with partially purified 0‘ RNA polymerase (02 ug) alone. or with control protein or GerE (04 ug) added
immediately after the addition of RNA polymerase. Run-off transcripts were electrophoresed in 5% polyacrylamide gels
containing 8 source and were detected by autoradiography. Arrowheads denote the positions of run-off transcripts of the
expected sizes in each panel. as judged from the migration of end-labeled DNA fragments of Hazel-digested pBR322. (a)
roll) transcription from HindIII-digested pLRKlOO (225-base transcript) with 0‘ RNA polymerase and control protein
(lane 1) or GerE (lane 2). (b) sigK transcription from XbaI-digested pBKl6 (l70-base transcript) with control protein
(lane 1) or GerE (lane 2). (c) gerE transcription from pSCl46 digested with BamHI (lane 1, l74-base transcript) or
HiadIII (lanes 2 to 4, 204-base transcript) with a‘ RNA polymerase alone (lanes 1 and 2), or with control protein (lane 3)
or 0er (lane 4) added. (d) cotzl transcription from pKS23 digested with .VcoI (lane 1, l3l-base transcript) or 53le
(lane 2, l49-base transcript), from Neal-digested pKS22 (lanes 3 and 4. 13l-base transcript). and from NcoI-digested
pKS24 (lanes 5 and 6. l31-base transcript) with 0‘ RNA polymerase alone (lanes 1 and 2). or with control protein (lanes
3 and 5) or GerE (lanes 4 and 6) added. [a-"P]UTP was the labeled nucleotide in the experiments shown in (c) and (d).

37

from gel-purified a" and B. subtilis core RNA poly-
merase (data not shown). GerE protein had no effect
on gerE transcription in vitro by partially purified
0‘ RNA polymerase (Fig. 6(c), lane 4). The effect of
GerE on cotA transcription in vitro varied
depending on the particular DNA template used
(Fig. 6(d)). GerE had little effect on cotA transcrip-
tion from a template containing 115 bp of DNA
upstream from the cotA transcriptional startsite
(lane 6); however, GerE inhibited cotA transcription
approximately twofold (in 2 independent experi-
ments) from a template containing approximately
430 bp of upstream DNA (lane 4).

Discussion
(a) Transcriptional activation by GerE

We have identified binding sites for GerE at the 5'
ends of cotB and cotC, coat protein genes whom
transcription depends on the appearance of GerE
during sporulation. We have also shown that GerE
stimulates cotB and cotC transcription in vitro by 0"
RNA polymerase and that the gerE gene itself can
be transcribed in vitro by a" RNA polymerase.
These results support the view (Zheng & Losick,
1990) that in the mother cell, a" RNA polymerase
first directs the transcription of gerE, then acts in
conjunction with the product of gerE to direct the
transcription of cotB, cotC and, perhaps, other late-
activated sporulation genes.

Interestingly, the region of GerE-conferred
protection from DNase I action in cotB (41 to 47 bp)
was approximately twice the length of the two
separate protected regions in cotC (16 to 19 bp for
binding site 1 and 21 to 22 bp for binding site 2).
Our interpretation of this observation is that cotB
contains tandem GerE binding sites and that cotC
contains two separate GerE binding sites.
Inspection of the sequences in the protected regions
reveals three similar 5 bp sequences, two (TGGGT
and TAGGC) in cotB and one (TGGGC. found in
binding site 1) in cotC. If these are recognition
sequences for GerE, then a possible consensus
sequence for the GerE binding site is TPuGGPy.
The closest match to this consensus in cotC binding
site 2 is the sequence TGGAC. Nevertheless, binding
site 2 appears to be sufficient to mediate
GerE-stimulated transcription of cotC, since a DNA
template with less than half of binding site 1
(produced by cleavage with HaeIII at position
-— 133) retains the ability to be transcribed in vitro
by a" RNA polymerase in the presence of GerE
(RH. & L.K.. unpublished results).

The downstream boundaries of the GerE binding
sites in cotB (position —36) and in cotC (position
-56 for binding site 2) are immediately adjacent to
or near regions of DNA that are expected to interact
with RNA polymerase (that is, the promoters). By
analogy with the positioning of the binding sites for
several well-characterized, positively acting regula-
tory proteins. the DNA-bound GerE can be con-
sidered to be appropriately" positioned to stimulate
RNA polymerase to transcribe from the cotB and

38

cotC promoters. For example, the binding region
(the tandem operator sites 0" and O") for the
phage A repressor is located between 34 and 74 bp
upstream from the transcriptional startsite for the
promoter (P...) from which the repressor stimulates
transcription of its own structural gene (CD by E.
coli RNA polymerase (Johnson et al., 1979; Meyer 8t
Ptashne, 1980). As another example, the catabolite
gene activator protein (CAP) binds to sequences
(typically protecting about 25 bp) centered from
about 41 to 107 bp upstream from the transcrip-
tional startsite of genes whose transcription it
stimulates (de Crombrugghe et al., 1984).

' An added significance of our demonstration in
vitro that GerE can bind to and stimulate transcrip-
tion from the regulatory regions of genes under its
control is that GerE is the prototypical representa-

. tive of the putative regulatory domain of a large

and diverse group of procaryotic transcriptional
activator proteins. Thus, GerE exhibits high overall
similarity to the COOK-terminal regions of certain
regulator members (the ngA sub-group, which
includes DegU, ComA and FixJ) of the family of
two-component sensor and response regulator
systems in bacteria (Gross et al., 1989; Kahn in
Ditta, 1991). The NIL-terminal region of the
response regulator proteins contains a site for phos-
phorylation by the sensor component of the two-
component system, and its phosphorylation state
influences the activity of the COOK-terminal
domain with respect to transcriptional activation
(Kofoid & Parkinson, 1988; Nixon et al., 1986).
Strikingly, GerE, which lacks the NHZ-terminal
domain, is highly similar along its entire 72 amino
acid length to the COOK-terminal domain of the
ngA sub-group of response regulator proteins. No
member of the ngA sub-group has (to our know-
ledge) been shown to bind to DNA or activate
transcription in vitro, but our results reinforce the
view that the GerE-like, COOK-terminal domain of
these proteins is directly responsible for the tran-
scriptional activation of genes under the control of
this sub-group of response regulator proteins.
Likewise, the similarity of GerE to the COOH
terminus of E. coli MalT, a large regulatory protein
whose NIL-terminal region is dissimilar to the
phosphorylation domain of the two-component
regulator proteins (Gross et al., 1989), once again
suggests that the GerE-like region of MalT could be
responsible for DNA-binding and transcriptional
activation by this regulatory protein (Richet et al..
1991; Vidal-Ingigliardi et al., 1991). Finally, the
similarity of GerE to the COOK-terminal domain
(region 4) of sigma factors (Kahn in Ditta, 1991)
reinforces the view that this domain mediates the
recognition of the -35 region of promoters
(Gardella et al., 1989; Siegele et al., 1989).

(b) Consensus .nce for promoters recognized
by RNA polymerase

a" RNA polymerase has been shown to transcribe
from the cotD and sigK (previously called spol VCB)

-35 -10

Consensus & - 11 bp - ass-«n

319x ceqtscaqacmsqscaqcctccceqtcaCAflcetfllcetatewc
coca sttttthtaACcatcacqtccttsttqtcartaac‘raaaqtaccaat
coco ttecatcsquCatqtaccccttatttttCATAact‘rnqtattntsat
gar: tqtasacqtcACctcctchcccttcttaCA‘rAtqa'rAtcteaactat
coca ttqasttaqtt.Cascaaataaatqtqacva‘rAtatatqcaqtuqe
eecc aactqtccsqucqcaaastc tacthcctr‘l'ntaa‘rmeeaeaqea

Figure 7. Alignment of promoters transcribed by a‘
RNA polymerase. The nucleotide sequences of the sigK,
cotA. catD and gerE promoter regions (see the text for
references) are aligned with respect to conserved nucleo-
tides (capital letters) in the - 10 and -35 regions relative
to the transcriptional startsites (underlined). Shown
above are the consensus -10 and -35 sequences,
separated by 17 bp, and shown below are the cotB and
cotC promoter regions with matches to the consensus
indicated by capital letters (a 1 hp gap was introduced
into the cotC sequence between the ‘10 and -35

regions).

 

promoters in vitro (Kroos et al., 1989). Efficient
transcription of sigK by a" RNA polymerase also
required a small, DNA-binding protein that is the
product of the spoIIID gene (Kroos et al., 1989;
Kunkel et al., 1989). Using 0‘ RNA polymerase
reconstituted from gelopurified a‘ and B. subtilis
core RNA polymerase, we find that gerE and cotA,
like catD, are transcribed efficiently by a" RNA
polymerase in the absence of SpoIIID, findings that
confirm and extend the results of studies on the
regulation of these genes in viva (Cutting et al., 1989;
Sandman et al., 1988). As shown in Figure 7 and as
noted previously (Foulger 8t Errington. 1991; Zheng
& Losick, 1990), the promoter regions of cotD
(Zheng & Losick, 1990), gerE (Cutting et al., 1989),
cotA (Sandman et al., 1988) and sigK (Kunkel et al.,
1988) each contain sequences similar to CATA---TA
at about position -10 relative to their transcrip-
tional startsites. Figure 7 also shows that the cotB
and cotC promoters, which were transcribed weakly
by a" RNA polymerase in the absence of GerE,
display some similarity to the CATA---TA sequence.
Interestingly, the putative - 10 consensus sequence
for a"-recognized promoters is similar to the
sequence CATACA-T, which is conserved in the
— 10 region of promoters transcribed by RNA poly-
merase containing the related sporulation sigma
factor. a" (see Roels et at. (1992) and Foulger &
Errington (1991) for recent compilations of
promoters recognized by a" RNA polymerase).
Unlike promoters recognized by a"3 RNA poly-
merase, however, promoters recognized by a" RNA
polymerase that have been characterized to date
display only a limimd region of similarity to each
other in their -35 regions. The sequence AC is,
however, found 17 bp upstream from the —10
region in the four promoters transcribed by 0'" RNA

39

polymerase in the absence of GerE. but only the C is
found at the corresponding position in cotB and
cotC. We note that three other promoters that are
inferred to be transcribed by a" RNA polymerase in
the absence of GerE, namely, spoVJ P2 (Foulger &
Errington, 1991), cotEP2 (Zheng & Losick, 1990),
and the promoter for the newly discovered coat
protein gene cotF (Cutting et al., 1991b), contain
-35 and -10 sequences that strongly conform to
the sequences AC and CATA-o-TA, respectively, at a
spacing of 16 to 17 bp. Mutational analyses will be
needed to determine whether the AC and
CATA---TA sequences are important for promoter
recognition by 0“ RNA polymerase.

(c) Eject of GerE on the transcription
of cotD and sigK

' GerE stimulated catD transcription in vitro by 6‘
RNA polymerase two. to threefold (Fig. 6(a)). This
result is in qualitative agreement with the finding
that cotD-lacZ expression is mduced about sevenfold
in gerE mutant cells (Zheng a Losick, 1990).
Inspection of the catD promoter region (Zheng &
Losick, 1990) reveals a 12 bp sequence (AAAA-
TAGGTCI'I‘) at positions -43 to -54 with ten
matches to a sequence (positions -69 to -80)
protected by GerE in the cotB promoter region
(Fig.1). Within the 12 bp sequence in the catD
promoter region is a 5 bp sequence (TAGGT) that
conforms to _ the putative consensus binding
sequence for GerE described above. It will be inter—
esting to determine whether GerE binds to this
sequence, since it would appear to position GerE
appropriately to stimulate RNA polymerase, as
discussed above.

GerE completely inhibited sigK transcription by
partially purified 0‘ RNA polymerase (Fig. 6(b)).
The partially purified 6" RNA polymerase used in
this experiment contained a small amount of
SpoIIID, thus permitting adequate transcription of
sigK. Inhibition of sigK transcription by GerE was
unexpected, since expression of a sigK-lacZ fusion
was about normal in gerE mutant cells (Kunkel et
al., 1988). Inspection of the sigK promoter region
(Kunkel et al., 1988) reveals a 15 bp sequence
(ACATATAGGCT'I'I'I‘G) at positions -—4 to +11
with 12 matches to a sequence (positions -41 to
-55) protecwd by GerE in the cotB promoter
region (Fig. 1). Within the 15 bp sequence in the
sigK promoter region is a 5 bp sequence (TAGGC)
that conforms to the putative consensus GerE
binding sequence. In addition, this 5 bp sequence is
repeated in inverted orientation at positions +11 to
+15. Thus, there may be two GerE binding
sequences near the start-site of sigK transcription
and GerE bound at these sites may prevent a“ RNA
polymerase from transcribing sigK. If these sites do
mediate repression of sigK transcription by GerE, it
could explain why a sigK-lacZ fusion was expressed
equally in wild-type or gerE mutant cells. since the
fusion was created by insertion of a transposon
(Tn917lac) 4 bp downstream from the sigK tran.

Figure 8. Regulatory effects of SpoIIID and GerE
during stages IV and V of sporulation. The effects of
SpoIIID and GerE on transcription by a‘ RNA poly-
merase in vitro are illustrated. As noted in the text,
studies in vivo also support some of the regulatory effects
depicted. (a) During stage IV of sporulation, SpoIIID
stimulates transcription of sigK and inhibits transcription
of catD. The gerE gene is transcribed by 0" RNA poly-
merase. (b) During stage V of sporulation, GerE inhibits
transcription of sigK and cotA, and stimulates transcrip.
tion of catD. cotB and cotC.

 

scriptional startsite (Kunkel et al., 1988), and this
would have presumably disrupted the putative
GerE binding site.

(d) Opposite ejects of GerE and Spoil ID help to
drive the mother-cell regulatory cascade

GerE and SpoIIID exert opposite effects on
H-diwcted transcription of both catD and sigK
(Fig. 8). It has been shown that SpoIIID stimulates
sigK transcription and inhibits catD transcription in
vitro (Kroos at al., 1989; and Fig. 8(a)). These
properties of SpoIIID led to the suggestion that
inactivation or sequestering of SpoIIID during
sporulation causes a switch from transcription of
sigK (and perhaps other stage IV sporulation genes)
to transcription of catD (and perhaps other stage V
genes) (Kroos et al., 1989). Recently, it has been
shown that the level of SpoIIID decreases at the
appropriate time during sporulation to cause such a
switch (Halberg & Kroos, 1992). Furthermore, the
decrease in the level of SpoIIID correlates with the
:ppearance of 0‘, suggesting that the appearance of

initiates the switch. We have shown here that 0"
RNA polymerase transcribes gerE (Fig. 6(c)). Thus,
the appearance of 0" RNA polymerase beginning at
about hour 4 of sporulation (Cutting at al., 1989; Lu
st al., 1990) would result not only in a declining level
of SpoIIID, but also in a rising level of GerE. We
have also shown here that GerE inhibits sigK tran-
scription (Fig. 6(b)) and stimulates catD transcrip-
tion (Fig. 6(a)) by a‘ RNA polymerase in vitro

40

' SpoIIID—D 0" ——>6er

/

Stage IV —>8tloe V

Figure 9. latory interactions controlling the levels
of SpoIIID, and GerE govern the stage IV to V
transition in the mother cell. SpoIIID stimulates sigK
transcription, leading to 0" production (Kroos et al..
1989). a" RNA polymerase transcribes gerE, leading to
GerE production. The appearance of o" causes a decrease
in the level of SpoIIID (Halberg & Kroos, 1992). GerE
inhibits transcription of sigK. down-regulating a‘ produc-
tion. Thus. 0‘ RNA polymerase functions during both
stage IV and stage V, but a declining level of SpoIIID and
a rising level of GerE switch the pattern of o‘-directed
gene expression from the stage IV pattern to the stage V

-pa_ttern.

 

(Fig. 8(b)). Because GerE exerts the opposite effects
of SpoIIID on a‘-directed transcription of sigK and
cotD, the appearance of GerE would reinforce the
switch in the pattern of mother-cell gene expression
previously postulated to be brought about by
inactivation or sequestering of SpoIIID. The reason
for the apparent redundancy in the switch is
unclear. Perhaps SpoIIID prevents premature
expression of cotD (and perhaps other stage V genes)
during the stage (IV) of spore cortex formation so
that the a" preduced initially directs expression of
sigK (autogenous regulation), gerE, and genes
involved in cortex formation (Halberg & Kroos.
I992). Accumulation of GerE would terminate this
period by diverting a" RNA polymerase away from
transcription of sigK (and perhaps other stage IV
genes) and would initiate the stage (V) of spore coat
formation by directing a" RNA polymerase to tran-
scribe genes encoding spore coat proteins. According
to this model, the regulatory interactions illustrated
in Figure 9 co-ordinate theJevels of SpoIIID. 0'"
and GerE so as to produce a molecular switch
governing the transition from the stage IV pattern
of mother-cell gene expression to the stage V
pattern.

The effects of GerE on transcription of gerE and
cotA by o'K RNA polymerase in vitro are consistent
with the effects of a gerE mutation on expression of
these genes in viva. GerE had no effect on transcrip-
tion of the gerE gene in vitro (Fig. 6(a)) and
expression of a gerEJacZ fusion is normal in a get!)
mutant (Cutting et al., 1989). The effect of GerE on
cotA transcription in vitro varied from little effect
with a template containing 115 bp of DNA
upstream from the transcriptional startsite to a
modest (but reproducible), twofold inhibition with a
template containing approximately 430 bp of
upstream DNA (Fig. 6(d)). If GerE inhibits soul
transcription by binding to DNA. this result
suggests that it must do so by binding to a site(s)
more than 115 bp upstream from the startsite of
transcription. In a previous study (Cutting st al..

1989), a cotA-lacZ fusion containing as little as
300 bp of DNA upstream from the colA transcrip-
tional startsite was found to be expressed about
threefold higher in a gerE mutant relative to wild-
type cells.

In summary, we have presented biochemical
evidence that GerE is a regulatory protein capable
of either stimulating or inhibiting transcription of
particular genes in the mother-cell line of gene
expression. In this respect, GerE appears to be
analogous to SpoIIID; however, the ordered
appearance of first SpoIIID. then GerE, and the
opposite effects of these two proteins on the tran-
scription of genes like sigK and catD presumably
ensures proper flow of the regulatory cascade
(Zheng & Losick, I990) controlling mother-cell gene
expression. In the cases of cotB and cotC, GerE binds
to specific sequences immediately adjacent to the
promoter and stimulates transcription by a1 RNA
polymerase. This finding supports the previous pro-
posal (Zheng 8t Losick, 1990) that GerE directly
activates the expression of genes in the terminal
temporal class of mother-cell expressed genes.

We thank Simon Cutting and Kathleen Sandman for
providing plasmids. This work was supported by the
Michigan Agricultural Experiment Station. by NIH grant
63143585 to L. K. and by NIH grant GMIS568 to R. L.

References

Bolivar. F.. Rodriguez. R. L.. Greene. P. J.. Betlach.
M. C.. I-Ieyneker. H. L. a Boyer. H. W. (1977).
Construction and characterization of new cloning
vehicles: II a multipurpose cloning system. Gene 2.
95—113.

Cutting. S. (1986). Genetics and properties of germination
mutants of Bacillus subtilis. D.Phil. thesis, Oxford
University.

Cutting. S. & Mandelstam. J. (1986). The nucleotide
sequence and the transcription during sporulation of
the gerE gene of Bacillus subtilis. J. Gen. Microbial.
132. 3013—3024.

Cutting. S. M. a Vander Horn. P. B. (1990). Genetic
analysis. In Molecular Biological Methodsfor Bacillus.
pp. 27-7-1. John Wiley & Sons Ltd. Chichester. UK.

Cutting. S.. Panzer. S. a Losick. R. (1989). Regulatory
studies on the promoter for a gene governing synthe-
sis and assembly of the spare coat in Bacillus subtilis.
J. Mal. Biol. 207. 393-404.

Cutting. 8.. Oke. V. Driks. A.. Losick. R. Lu, S. tKroos.

L.(1.990) A forespore checkpoint for mother cell gene
expression during development in B. subtilis. Cell. 62.
239—250.

Cutting. S.. Roels. S. & Losick. R. (1991a). Sporulation
operon spa! VF and the characterization of
mutations that uncouple mother-cell from forespore
gene expression in Bacillus subtilis. J. Mol. Biol. 221.
1237—1256.

Cutting. 8.. Zheng, L. & Losick. R. (1991b). Gene
encoding two alkali-soluble components of the spare
coat from Bacillus subtilis. J. Bacteriol. 173.
2915—2919.

Davis. R. w.. Botstein. D. a Roth. J. R. (1980).
Advanced Bacterial Genetics. Cold Spring Harbor
Laboratory Press. Cold Spring Harbor. NY.

41

de Crombrugghe, B., Busby. S. & Buc. II. (1984). Cyclic
AMP receptor protein: role in transcription activa-
tion. Science, 224, 831—838.

De Lencastre. I-I. & Piggot. P. (1979). Identification of
different sites of expression for spa loci by trans-
formation of Bacillus subtilis. J. Gen. Microbiol. 114.
377-389.

Donovan. W., Zheng, L, Sandman. K. & Losick. R.
(1987). Genes encoding spore coat polypeptides from
Bacillus subtilis. J. Mol. Biol. 196. 1—10.

Errington. J. (1986). A general method for fusion of the
Escherichia coli lacZ gene to chromommal genes in
Bacillus subtilis. J. Gen. Microbial. 132. 2953-2906.

Feng, P. & Aronson. A. I. (1986). Characterization of a
Bacillus subtilis germination mutant with pleiotropic
alterations in spore coat structure. Curr. Microbial.
13, 221-226.

Foulger. D. a Errington, J. (1991). Sequential activation

, of dual promoters. by different sigma factors main-
tains spoVJ expression during successive develop-
mental stages of Bacillus subtilis. Mal. Microbiol. 5,
1363—1373. ‘

Gardella. T., Moyle. 1!. & Susskind, M. M. (1989).
Amutant E. coli a” subunit of RNA polymerase
with altered promoter specificity. J. Mol. Biol. 2“.
579-590.

Gross, R.. Arico. B. a Rappuoli. R. (1989). Families of
bacterial signal-transducing proteins. Mol. Microbiol.
3. 1661—1667.

Hager, D. & Burgess. R. (198)). Elution of proteins from
SDS polyacrylamide gels. removal of SDS. and

/ renaturation of enzymatic activity: results with
sigma subunit of E. coli RNA polymerase. wheat-
germ topoisomerase. and other enzymes. Anal.
Bioclrern. 109. 76-86.

Halberg, R. & Kroos. L. (1992). Fate of the SpoIIID
switch protein during Bacillus subtilis s rulation
depends on the mother-cell sigma factor . J. Mol.
Biol. In the press.

1-Iasnain. S., Sammons. R.. Roberts. 1. at Thomas. C. M.
(1985). Cloning and deletion analysis of a genomic
segment of Bacillus subtilis coding for the all A, B. C'
(succinate dehydrogenase) and gerE (spore germina-
tion) loci. J. Gen. Microbial. 131. 2269—2279.

Henner. D. J .. Yang, M. & Ferrari. E. (1988). Localization
of Bacillus subtilis sacU(H Y) mutations to two
linked genes with similarities to the conserved pro-
caryotic family of two-component signalling systems.
J. Bacteriol. 170. 5102-5109.

Holland. 8. K.. Cutting. S. & Mandelstam. J. (1987).
Possible DNA-binding nature of the regulatory pro-
teins encoded by spoll D and gerE. involved in
sporulation of Bacillus subtilis. J. Gen. Microbial.
133. 2381—2391.

Jenkinson. H. F. & Lord. H. (1983). Protease deficiency
and its association with defects in spore coat struc-
ture. germination and resistance properties in a
mutant of Bacillus subtilis. J. Gen. Microbial. 129.
2727-2737.

Johnson, A. D.. Meyer. B. J. & Ptashne, M. (1979).
Interactions between DNA-bound repressor-s govern
regulation by the i. phage repressor. Prac. Nat. Acad.
Sci., U.S..-l. 76, 5061—5085.

Kahn. D. & Ditta. G. (1991). Modular structure of FixJ:
homology of the transcriptional activator domain
with the —35 domain of sigma factors. Mal.
Microbial. S. 987-997.

Karmazyn-Campelli. C.. Bonamy. C.. Savelli. B. &
Stragier. P. (1989). Tandem genes encoding a-factors

for consecutive steps of development in Bacillus
subtilis. Genes . 3, 150-157.

Kofoid. E. C. A: Parkinson, J. S. (1988). Transmitter and
receiver modules in bacterial signalling proteins.
Prac. Nat. Acad. Sci., U.S.A. 85, 4981—4985.

Kroos, L., Kunkel. B. & Losick, R. (1989). Switch protein
alters specificity of RNA polymerase containing a
compartment-specific sigma factor. Science, 243,
526—529.

Kunkel, B., Sandman, K., Panzer, 8., Youngman, P. &
Losick. R. (1988). The promoter for a sporulation
gene in the spa! VG locus of Bacillus subtilis and its
use in studies of temporal and spatial control of gene
expression. J. Bacterial. 170, 3513-3522.

Kunltel. B., Kroos, L., Path, 11., Youngman. P. & Losick,
R. (1989). Temporal and spatial control of the
mother-cell regulatory gene spa! I I D of Bacillus
subtilis. Genes Develop. 3, 1735—1744.

Kunkel. 3., Losick, R. a Stragier, P. (1990). The Bacillus
subtilis no for the developmental transcription
factor is generated by excision of a dispensable
DNA element containing a sporulation recombinase
gene. Genes Develop. 4, 525—535.

Kunst. F., Debarbouille, M., Msadek, T., Young, M.,
Mauel, C., Karamata. D., Klier, A., Rapoport, G. s
Dedonder. R. (1988). Deduced polypeptides encoded
by the Bacillus subtilis sacU locus share homology
with two-component sensor-regulatory systems.
J. Bacteriol. 170. 5093—5101.

Laemmli. U. K. (1970). Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature (London), 227. 680—685.

Losick. R. a Stragier, P. (1992). Crimcross regulation of
cell-type~specific gene expression during development
in Bacillus subtilis. Nature (London), 355, 601—804.

Lu. 8.. Halberg, R. a Kroos. L. (1990). Processing of the
mother-cell a factor, a‘, may depend on events
occurring in the forespore during Bacillus subtilis
development. Proc. Nat. Acad. Sci., U.8.A. 87.
9722-9726.

Maniatis. T., Fritsch. E. F. a Sambrook. J. (1982).
Molecular Cloning, A Laboratory Manual. Cold
Spring Harbor Laboratory Press, Cold Spring
Harbor. NY.

Maxarn. A. s Gilbert, W. (1980). Sequencing end-labeled
DNA with base-specific chemical cleavages. Methods
Ensyrnal. 65. 499—560.

Meyer. B. J. & Ptashne, M. (1980). Gene regulation at the
right operator (0.) of bacteriophage 1. III. 3.
repressor directly activates gene transcription.
J. Mol. Biol. 139, 196—205.

Mair. A. (1981). Germination properties of a spare coat-
defective mutant of Bacillus subtilis. J. Bacterial.
146.1106—1116.

Nixon. T., Ronson, C. W. a Ausubel. F. M. (1986).
Two-component regulatory systems responsive to
environmental stimuli share strongly conserved
domains with the nitrogen assimilation genes ntrB
and ntrC. Prac. Nat. Acad. Sci., 0.8.21. 83,
7850-7854.

Panzer. 8., Losick, R.. Sun, D. a Setlow, P. (1989).
Evidence for an additional temporal class of gene
expression in the forespore compartment of sporu-
lating Bacillus subtilis. J. Bacteriol. 171, 561-564.

Richet. E.. Vidal-Ingigliardi. D. a Raibaud. O. (1991).
A new mechanism for coactivation of transcription
initiation: repositioning of an activator triggered by
the binding of a second activator. Cell, 66.
11854195.

Roels, 8., Driks, A. a Losick, R. (1992). Characterization
of spa! VA, a sporulation gene involved in coat
morphogenesis in Bacillus eubtilis. J. Bacteriol. 174.
575-586.

Sambroolt, J.. Fritsch, E. F. a Maniatis. T. (1989).
Molecular Cloning: A laboratory Manual, Cold
Spring Harbor Laboratory Press, Cold Spring
Harbor, NY.

Sandman. K., Kroos, L., Cutting, 8., Youngman, P. &
Losick. R. (1988). Identification of the promoter for a
spare 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.

Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA
sequencing with chain-terminating inhibitors. Prac.
Nat. Acad. Sci., U .S.A. 74, 5463-5467.

Siegele, D..A, I-Iu,J..C, Walter, W. A. aGromHC A.
(1989). Altered opromoter recognition by mutant
forms of the a7 °subunit of Escherichia coli RNA
polymerase. J. Mal. Biol. 206, 59l—603.

Stevens, C. M. & Errington, J. (1990). Differential gene

- expression during sporulation in Bacillus subtilis:
structure and regulation of the spolIlD gene. Mol.
Microbiol. 4, 543—552.

Stragier, P., Kunkel, B., Kroos, L. & Losick, R. (1989).

Chromosomal rearrangement generating a composite
for a developmental transcription factor.
Science, 243, 507-512.

Strauch, M. A., Spiegelman, G. B., Perego, M., Johnson.
W. C., Burbulys, D. a Hoch, J. A. (1989). The
transition state transcriptional regulator aer of
Bacillus subtilis is a DNA binding protein. EMBO J.
8, 16154621.

Sun, D., Stragier. P. a Setlow, P. (1989). Identification of
a new c-factor involved in compartmentalized gene
expreuion during sporulation of Bacillus subtilis.
Genes Develop. 3, 141-149.

Summan. M. D. & Setlow, P. (1991). Cloning, nucleotide
sequence, and regulation of the Bacillus subtilis 9pr
gene which codes for the protease that initiates
degradation of small, acid-soluble. proteins during
spore germination. J. Bacteriol. 173, 291-300.

Tabor. S. & Richardson. C. C. (1985). A bacteriophage T7
RNA polymerase/promoter system for controlled
exclusive expression of specific genes. Proc.’ Nat.
Acad. Sci. U.S..A 821074-1078.

Vidal Ingigliardi. D. Richet. E. a Raibaud. 0. (1991).
Two MalT- binding sites in direct repeat. a structural
motif involved in the activation of all the promoters
of the maltose regulons in Escherichia coli and
Klebsiella pneumoniae. J. Mol. Biol. 218. 323-334.

Weinrauch. Y., Guillen. N. a Dubnau, D. A. (1989).
Sequence and transcription mapping of Bacillus
subtilis competence genes carnB and cornA. one of
which is related to a family of bacterial regulatory
determinants. J. Bacteriol. 171, 5362-5375.

Yanisch-Perron. C.. Vieira, J. a Messing, J. (1985).
Improved M13 phage cloning vectors and host
strains: nucleotide sequences of the M13mpl8 and
pUCl9 vectors. Gene. 33, 103-119.

Zheng, L. (I990). Spore coat protein genes and their
regulation in Bacillus subtilis, Ph.D. thesis. Harvard
University.

Zheng. L. a Losick. R. (1990). Cascade regulation of spare
coat gene expression in Bacillus subtilis. J. Mal. Biol.
212. 645—660.

Edited by M. Gottcsman

42

CHAPTER III

Regulation of the Transcription of a Cluster of Bacillus subtilis Spore Coat Genes

43

Regulation of the Transcription of a Cluster of Bacillus subtilis

Spore Coat Genes

Jlanke Zhang‘)‘, Hiroshi Ichikawa’, Richard Halberg’, Lee Kroos2 and
Arthur I. Aronson‘t

1Department of Biological Sciences, Purdue University
Wm Lafayette, IN 47907, U.S.A.

1Department of Biochemistry, Michigan State University
East Lansing, MI 48824. U .S.A.

The pattern of transcription has been examined for a cluster of genes encoding polypeptides
some or all of which are assembled into a crou- -linked component of the Bacillus subtilis
spore coat. Three promoters, designated wa, Px and P", were indicated by reverse
transcriptase mapping. On the basis of Northern hybridization, it appeared that the cot V, W
and X genes were transcribed as a polycistronic mRNA from P"; as well as a
monocistronic cotX mRNA from Px. The cat Y and call genes are cotranscribed from the P"
promoter with a smaller cat 1' mRN A resulting from premature termination or RNA
proceuing. All four transcripts were synthesized late during sporulation and were not
produced in mutants lacking sigma K, which directs RNA polymerase to transcribe genes in
the mother-cell compartment of sporulating cells. The DNA-binding protein GerE, which
affects transcription of many genes in the mother cell during the late stages of sporulation,
was also shown to be involved. There was essentially no cotX mRNA in a gerE mutant and
the amounts of cotVWX, cotY Z and cat Y mRNAs were somewhat reduced. In vitro run-off
transcription studies with a“ RNA polymerase and GerE confirmed the presence of the three
promoters, and directly showed that GerE was necessary for transcription from Px as well as
enhanced transcription from the wa and P" promoters. The DNase I footprints of GerE
for all three promoters were immediately upstream of the —35 regions. These GerE binding
sites were compared to those in other GerE-responsive promoters and a larger consensus
sequence for GerE binding was recognized. This complex transcriptional pattern of the
cat VWX YZ cluster is probably necessary to ensure that an optimal amount of each protein
is made for the assembly of the spare coat.

Keywords: spore coat protein gene cluster and operons; spore promoters;
in vitro transcription; DNA-binding protein; footprinting

 

1. Introduction

During sporulation of Baciaus subtilis, a number
of small polypeptides are synthesized within the
mother-cell chamber and deposited on the deve-
loping forespore to form a thick, multilayered cost
which has a protective function (Aronson & Fitz-
James, 1976; Pandey & Aronson, 1979; Goldman &
Tipper, 1978; Jenkinson. 1981) and may also have
an indirect role in germination (Mair, 1981; Bourne
et al., 1991; Zhang et al., 1993). There are at least
fifteen soluble spore coat polypeptides and about
30% of the spare coat protein is resistant to a
variety of solubilization treatments. This “insoluble

 

fPresent addrem: Department of Biochemistry
Purdue University, West Lafayette, IN 47907, USA.

tAuthor to whom all correspondence should be
addremed.

fraction"probab1y consists of cram-linked poly-
peptides (Pandey a Aronson, 1979; Zhang et al.,
1993).

Seven genes encoding soluble coat proteins, cotA
through F and catT, have been cloned and analyzed
(Donovan et al., 1987; Zheng et al., 1988; Aronson et
al., 1989; Cutting et al., 1991). The transcription
units for these genes are monocistronic and scat-
tered widely on the B. subtilis chromosome. The
expression of these genes is restricted to the mother
cell chamber and transcription involves both
mother cell- -specific as and a" RNA polymerases
(Sandman et al.,1988;Kroos et al.,1989;Zheng &
Losick, 1990; Zheng et al., 1992). In addition, two
sporulation-specific DNA-binding proteins, SpoIIID
and GerE, function as regulators of cat gene
expression by acting as transcriptional activators or
repressors (Kroos et al., 1989; Kunkel et al., 1989;
Zheng et al., 1992; Halbergfi Kroos, 1992).

The spore coat insoluble fraction accounts for a
large proportion of the total coat protein. These
cross-linked proteins could be responsible for the
resilience of the spore coat as well as the hydropho-
bicity of the spore (Zhang et al., 1993). A cluster of
five genes, cot VWX Y2, some or all of which encode
polypeptides present in the spore coat insoluble
fraction of B. subtilis, has been cloned and the
products characterimd (Zhang et al., 1993). The
cotX gene encodes a glutamine-, lysine- and
cysteine-rich protein which appears to be exten-
sively cross-linked and is probably a major com-
ponent of the coat insoluble fraction. Two other
proteins, CotY and CotZ, are similar in sequence
with 9-3% and 67% cysteine, respectively, some or
all of which are disulfide cross-linked. The clustering
of these spore coat protein genes and the phenotypic
effects of null mutants (Zhang et al., 1993) indicated
that these genes may be coordinately controlled in
order to produce the appropriate amounts of the
proteins necessary for the interactions required to
form the coat insoluble fraction. In this report, we
present results which show that the genes in the
cotVWX Y Z cluster are organized into two operons
as well as two overlapping monocistronic transcrip
tion units. Expression of the cotVWX Y2 genes
involves the mother-cell-specific sigma factor, 0“.
the DNA-binding protein GerE, and probably an
antiterrnination mechanism.

2. Materials and Methods
(a) Strainsandplasrnids

The wild-type strain, B. subtilis JH642 (trp02, pheAl).
was obtained from Dr J. Hoch (Scripps Research
Institute). Strain 1AA-1 (gerEJB, leuB) was described by
Feng & Aronson (1986). Strains 55.3 (spoIIGBJ5, trpC'2;
Errington a Mandelstam, 1986) and BK410 (spoIIIC94;
Kunkel et al., 1989) and plasmids containing the catD
(Donovan et al.. 1987) and call! (Zheng et al., 1988) genes
were obtained from Dr R. Losick (Harvard University).

Plasmid N26 was previously described (Zhang et al.,
1993). Plasmid N222 was derived from N26 by
removing a 1175 bp BglII fragment from the 1771 bp
HindIII insert. and was used for in vitro transcription
from promoter Pr Plasmid N223 was constructed by
digesting N26 with BstXI (9 bp downstream of the cotX
initiation codon) and SphI (on the pUC18 vector) to
remove a 1378 bp fragment, and then religation with T4
ligase after treatment with Klenow fragment in the pre-
sence of 50 M each of the deoxynucleoside triphosphates.
This plasmid was used for the experiment of
GerE-footprinting on the P, promoter. A 2 lrb HindIII
fragment containing the entire cot V coding region and the
first 13 codons of the cotW gene (Zhang et al., 1993) was
cloned into the HindIII site of a modified pUClS lacking
the EcoRI site. The resulting plasmid was digested with
EcoRI (a unique site 57 bp downstream of the cat V initiao
tion codon) and PstI (in the multicloning site) to remove a
410 bp fragment. Blunt-ends were generated by filling in
with Klenow fragment and 50 uM each of the deoxynuc-
leoside triphosphates and the construct was then ligated.
This plasmid, N230, wu used for in vitro transcription
and for GerE-footprinting on promoter P“...

45

' (b) Total RNA preparation

Cells of various B. subtilis strains were grown in ‘
nutrient sporulation .medium (Shaeffer et al., 1983) at
37°C in a New Brunswick rotary shaker at 250 rpm.
Growth was monitored by measuring the A,“ in .
Perkin-Elmer Junior Model 35 spectrophotometer. Cells
(30 to 50 ml) were harvested at the end of expanemm
growth and at hourly intervals (designated t,, t2 etc.)
thereafter by centrifugation at 5000 g for 10 min. TOtal
RNA was extracted as described by Wu et al. (1989) and
further treated with ribonuclease-free DNase I (Boehr.
inger-Msnnheim) at 37°C for 1 h. The phenol/chloroform]
isoamyl alcohol (25 :24: 1) extraction was repeated 3 to ‘
times followed by 2 extractions with chloroform/my)
alcohol (24:1). The RNAs were precipitated with 2'0]
ethanol and one tenth vol of 3 M Na acetate (pH 5). Th.
final precipitaws were redissolved in diethylpyrocu-bo.
nate-treated water (Maniatis et al., 1982). The concentra.
tions were determined (Am, ..,) and 40 units of the RNase
inhibitor RNasin (Promega) was added prior to storage at
—70°C.

(0) Northern hybridization analysis

A procedure modified from that of Miller (1987) for ‘
RNA agarose-formaldehyde gel electrophoresis was used,
Thirty ug of RNA was denatured by heating at 65°C for
5 min in 15 to 20 pl of a solution containing gel running
buffer (05 M 3-(N-morpholino) propanesulfonic acid
(Mops), 001 M Na, EDTA (pH 7), 3-3 M formaldehyde,
50% formamide) and resolved on a 1% agarose gel
containing 2-2 M formaldehyde in running buffer. Two ug
of an RNA ladder (0-24 to 949 kb; BRL) was loaded on
the gel and served as size standards. RNAs were blotted
onto a 0-45 am BA-S85 reinforced nitrocellulose
membrane (Schleicher a Schuell) using a VacuGene
vacuum blotting system (Pharrnacia) in 20 x SSC (1 x SSC
contains 015 M NaCl, 0015 M Na citrate) for l h. DNA
fragments prepared as described below were labeled with
[a- 2PldC'I'I’ by using a Multiprime DNA labeling kit
(Amersham). Hybridization was carried out by the
method of Mahmoudi a Lin (1989).

(d) Preparation of Northern hybridisation probes

Phage mpJZ3(7A) contains a deletion of the 17 kb
HindIII fragment (see Figure 1) generated by exonuclsase
III (Zhang et al., 1993). This deleted fragment containing
the C terminus of cotW (lacking the first 12 codons, but
including 12 bp downstream of the catW stop codon; see
Figure 1) was excised by HindIII-EcoRI digestion and
used as probe A. Probe B was a 1'3 lrb fragment
containing the sequence encoding residues 9 to 54 of CotX
(see Figure l) and was isolated from N213 (Zhang et al..
1993) by EcaRI-PstI digestion. Probe C includes codons
55 to 146 of the cotY coding region and was isolated by
digestion of phage mN24(6C) (another deletion of the
17 kb HindIII fragment derived as described above) with
EcoRI-HincII. Probe D is a 154 bp PvuII-HindIII frag-
ment containing 52 bp upstream of the start oodon plus
the first 34 codons ofcotZ (see Figure 1), and was isolated
from plasmid N28 (which contains the 1771 bp HindIII
fragment (see Figure 1) in pBR322).

A 201th HindIII fragment containing the cot]! gene
was isolated from plasmid pL2100,(2heng et al.. 1988).
A 1'8 ltb HindIII fragment containing the catD gene was
from plasmid pBD156 (Donovan et al., 1987) and a 1-2 kb
HindIII fragment containing the cotT gene was from a

pHP13 clone (Aronson et al.. 1989). All of these DNA
fragments were labeled with [a-“PldC'I‘P and Klenow
fragment employing a Multiprime labeling kit
(Amersham).

(e) Primer extension assay

The primer extension method used was modified from
that of Ferrari et at. (1988). Eighteen or nineteen‘base
oligonucleotide primers were synthesized in the Purdue
University Laboratory for Macromolecular Structure and
labeled at their 5'.termini by incubation for 2 h at 37°C
with T4 polynucleotide kinase and [y-“P]ATP. Reactions
were terminated by heating at 65°C for 20 min. About
30 pmol of labeled primer was mixed with 30 ug of total
RNA. NaCl was added to a final concentration of 0-4 M
followed immediately by 20 units of RNasin. Annealing
was carried out by incubation at 37 °C for 2 h followed by
precipitation with 2 vol ethanol at —20°C for 1 h. After
centrifugation. the precipitates were washed with 70%
ethanol. air-dried and redissolved in 50 ul AMV reverse
transcriptase bufi'er (Promega) plus 05 mM each of the
deoxynucleoside triphosphates and 20 units of RNasin.
The reaction was initiated by the addition of 20 units of
AMV reverse transcriptase (Promega). After incubation at
42°C for 2 h, the reaction was terminated by extraction
with phenol/chloroform/isoamyl alcohol (25:24: 1) and
the nucleic acids were precipitated with 2 vol ethanol and
one tenth vol of 3 M Na acetate (pH 5) at -70°C for'l h.
After centrifugation, the precipitates were washed with
70% ethanol. dried and redissolved in fiul Sequenase
buffer plus 4 pl Sequenase stop solution (0.8.
Biochemical). After heating at 90°C for 3 min. 2 to 3 ul of
the mixture was fractionated on a 6% sequencing gel. The
same primers were used in sequencing reactions from M13
templates with [a-“P]dATP and the Sequenase kit (0.8.
Biochemical) or with [a-"PldCI‘P and a modified labeling
mixtureintheSequenasekitThesereactionswerelosded
on the same gel and served as siae standards.

(f) Production of GerE in E. coli

GerEandcontrclproteinwerepreparedasdescr-ibed
previously (Zheng et al.. 1992) with the following modifi-
cations. E. coli cultures were grown at 30°C to an A,” of
0-5 in LB medium (Maniatis et al., 1982) containing kans-
mycin sulfate (50 ug/ml) and ampicillin (75 ug/ml). Cells
were induced by a temperature shift to 42 °C for 25 min.
Rifampicin was added to a final concentration of
200 ug/ml. and the cultures incubated at 42°C for 10 min
and then at 30°C for 60 min. Cells were collected by
centrifugation, resuspended in 005 vol of lysis bufi'er
(10 mM Tris-H01 (pH 8-4), 10 mM MgCl,, 1 mM EDTA,
03 mg/ml phenylmethanesulfonyl fluoride, 05 mg/m1
lysoayme. 0-2 ml! dithiothreitol, and 01 mg/ml DNase I)
and incubated for 10 min at 37°C. After addition of 05
vol of sample bufier (0375 M Tris-H01 (pH 68). 15%
2omercaptoethanol, 60 mg/ml SDS, 30% glycerol and
3 mg] ml bromophenol blue) and immersion in boiling
water for 3 min. the proteins were separated by
SDS/PAGE (18% polyacrylamide gel; Thomas &
Kornberg, 1978). GerE was excised from the gel, eluted
from the gel slice and renatured as described (Hagar &
Burgess. 1980). Control protein was prepared by excising
the region of the gel corresponding to GerE from a lane
containing proteins from E. coli cells not expressing GerE,
as described previously (Zheng et al.. 1992). Control pro-
tein was then eluted from the gel slice and renatured as
dmcribed above.

46

(g) In vitro transcription

0" RNA polymerase was partially purified from gerE
mutant cells as described previously (Kroos et al., 1989).
The enzyme was comparable in protein composition and
in cotD- and sigK-transcribing activities to fraction 24
shown in Figure 2 of Kroos et al. (1989). 6‘ RNA poly~
merase was reconstituted from B. subtilis core RNA poly-
merase and gel-purified, renatured 0‘ as described
previously (Kroos et al., 1989). Transcription reactions
(45 ul) were performed as described previously (Carter &
Moran, 1986) except that RNA polymerase was allowed
to bind to the DNA template for 10 min at 37 °C before
the addition of nucleotides (the labeled nucleotide was
[a-”P]CI‘P). Six ug of heparin was added 2 min after the
addition of the nucleotides to prevent reinitiation. Afier
the reactions were stopped, 10 ul of the reaction mixtures
was subjected to electrophoresis and transcripts were
detected by autoradiography. The signal intensities were
quantified using a Visage 110 Image Analyzer
(BioImage).

(h) DNase I footprinting

DNase I footprinting experiments were performed
accordingtomethod(2)ssdescn‘bedby2hengetat.
(1992), except 005 pmol of probe (prepared as described
below) was used and a five-fold (w/w) excess of poly-
(dI-dC) as compared to probe was added as competitor. In
order to generate markers, the probes were subjected to
the chemical cleavage reactions of Maxam a Gilbert
(1980) as described previously (Maniatis et al., 1%2).
DNA probes labeledatonlyoneendwerepreparedas
follows; for analysis of the cotX promoter region, N223
was digested with XbaI, which cleaves 54 bp downstream
of the transcriptional start site of the Px promoter. and
labeled either in the nontranscribed strand using the fill-
in reaction of the Klenow fragment of DNA polymerase I
and [a-“P]dCI‘P, or labeled in the transcribed strand by
treatment with alkaline phosphatase followed by
T4 polynucleotide kinase and [y-“PlATP. In both cases,
the labeled DNA was digested with HindIII and the
409 bp HindIII-Xbal fragment was purified after electro-
phoresis in a nondenaturing polyacrylamide gel using the
crush and soak method (Sambrook et al., 1989). For
analysis of the cotVWX promoter regon. N230 was
digested with Sail which cleaved 93 bp downstream of
the transcriptional start site of the PM promoter. and-
labeled as described above, digested with HindIII. and
the [-6 kb HindIII-Sall singly, end-labeled fragments
were purified as described above. For analysis of the
cotYZ promoter region, N26 was digested with NeoI
which cleaves 60 bp downstream of the transcriptional
start site of the Pa promoter. end-labeled as described
above. digested with EcoRV, and the 240 bp EcoRV-NcoI
n'ngly, end-labeled fragments were purified as described
above.

3. Results

(a) Organization and expression of transcripts from
the cotVWXYZ gene cluster
Total RNA prepared from samples of B. subtilis
JH642 cultures, collected at one hour intervals after
the cells had entered the stationary phase, was
fractionated in a 1% agarcse gel and blotted onto
nitrOcellulose (see Materials and Methods). The

ll 12 (3'14 [5 to t7 is 19110

(1.6 kb -

 

1.7 kb probe

 

 

 

"immfi ...x )im E ...y )ri rmz Hi

 

 

 

«m lfdm
lid HcPv
06 kb _ 0.6 is
1.6 kb _ 1.4 kb
— - — -
Prom: A B C D

la is lb 17 is Is lo [7 t4 is is 17 his 1617

rem- -- e! ‘3
1! ~ ., ., «.2

mi "4"
0.6 kb- . C . ‘

‘.-r"

  

A B C D

Figure 1. Northern hybridization analysis of the catVWX YZ gene cluster. Total RNA was prepared from B. subtilis
JH6-l2 cells at hourly intervals after cells had entered the stationary phase (designated t‘ to tw). Thirty ug of RNA was
fractionated in 1% formaldehyde-agarcse gels and blotted to nitrocellulose (Materials and Methods). Top: hybridization
with a 17 kb HindIII fragment. Bottom: hybridization with probes A. B. C and D prepared as described in Materials
and Methods. The sizes and relative locations within the gene of each probe are shown with thick bars. The diagram of
the organization of the genes is based on the results reported here. primer extension mapping and in vitro transcription
experiments (Figures 2 and 4). The 5 open reading frames are shown as boxes. Three promoters. Pm. P’x and P", and
the orientation of transcription are marked with bent arrows. Three hairpin loops indicate potential transcription
terminators (Zhang et al., 1993). The 4 arrows immediately below the diagram show the orientation and sizes of
transcripts produced from this gene cluster. The sizes of hybridizing RNA bands are indicated to the left of the gels.
RNA size standards of 0-24 to 949 kb (BRL) were used. Orientations and relative binding sites of oligonucleotides

(marked Prl to PM) used for primer extension experiments are indicated with arrow heads. Abbreviations: Hd. HindIII:
Hc. HincII; Pv, PvuII.

1-7 kb HindIII fragment containing the C terminus Transcription initiating from a promoter located
(149 codons) of cotW, all of cotX and 000’, and the immediately upstream of the cotX coding region
N terminus (34 codons) of cotZ hybridized to at least could produce the 06 kb mRNA if there was termi-
three mRNAs of 16 kb, 14 kb and 06 kb, which nation at a potential Rho-independent terminator
were first detected at t‘ to t, (Figure 1, top). The immediately downstream of the cotX coding region
cotE mRNA, which is expressed starting at about (Figure 1; Zhang et al., 1993). There are no potential
stage II of sporulation (Zheng & Losick, 1990), was stem-loop structures between the cotV and anti?
first detected at tt, whereas transcripts of the cotD genes nor between cotW and cotX (Zhang et al..
and cot’I' genes appeared at about the same time as 1993). Assuming that the only site of termination is
those of the cotVWX Y2 cluster (data not shown). after cotX, the cotV. cotW and cotX genes are prob-
In order to delineate the transcription pattern in ably cotranscribed, therefore, from another
detail, four subfragments were prepared from the promoter upstream of cotV to produce the 16 kb
1 7 kb probe and each was labeled with [a-“P]dCTP mRNA.
and Klenow fragment and used as a Northern The existence of two promoters, designated Px
hybridization probe (see Materials and Methods). As and PW“, was suggeswd by primer extension
depicted in Figure 1, probe B which contains the N mapping (Figure 2). The 5' end of the 16 kb mRNA
terminus of the cotX gene hybridized to mRNAs of was mapped to 24 bp upstream of the first codon of
16 kb and 06 kb. Probe A, from the C terminus of 000’ (Pym in Figure l) with primer Prl (which
catW, hybridized only to a 16 kb mRNA. corresponds to the region between codons 37 and 42

47

Pr2
A T C tstr

Prl
A T C tell

(i G

 

r
A
T
c
T
c
A
at (
A
L. A
C A
T 1'
A
T T
c
c
an
a
A
r
T
1'
c
" A
A t
r a.
T .
c =- ..
T ‘f..
A -
an -.-- (
; I"
“r
c
r
T

Pr3 Pr-i
GATClell (iATClbll
f;

. '13-

'I'
>n-l-l-l>C-it‘.f‘.-l-i>>

   

Figure 2. Reverse transcri ptase mapping of the 5’ termini of the l 6 kb cot VWX mRNA (Prl), the 06 kb cotX mRNA
(Pr2). the 06 kb rat 1" mRNAP (Pr3) and the l 4 kb cot YZ mRNA (Pr4). RNA was prepared from B. subtilis JH642 I and
6 h after the onset of the stationary phase (tl and t5; Materials and Methods). The products of primer extension were
fractionated in a 6% sequencing gel. Sequencing reaction mixtures with the same primers were loaded in adjacent lanes
and served as size markers. Prl corresponds to the region between codons 37 and 42 of cotV, Pr2 to the region between
codons l5 and 20 of cotX. Pr3 corresponds to the region between codons 9 and 14 of wt Y: and Pr4 to the region between
codons 20 and 25 of cotZ (Figure 1: Zhang et al., 1993). Arrow-heads indicate the positions and sizes of major extension

products. ‘ identifies the 5’ termini of the mRNAs.

of cotV). The 5' end of the 06 kb cotX mRNA was
mapped with primer Pr2 (which corresponds to
codons 15 to '20 of cotX) and several major primer
extension products of 80 to 87 bases in length with
single base difi‘erences among them were generated.
This cluster could correspond to a 5' terminus of the
06 kb cotX mRNA located about 21 bp upstream of
the first codon of cotX (Figures 1 and 2). In addi-
tion. larger products were detected. some of which

48

could be prematurely terminated reverse transcript-
ase extensions of the 16 kb cotVWX mRNA. The
longest could correspond to the 5' terminus of this
mRNA which was not well resolved in this gel. The
in vitro transcription studies presented below indi-
cate that the synthesis of cotX mRNA resulted from
transcription from an internal promoter. Px, rather
than from processing of the [-6 kb cot VWX mRNA.

A DNA fragment from the middle portion of the

cotY coding region. C. hybridized to 14 kb and
06 kb mRNAs. while probe D, from the N-terminal
fragment of cotZ, hybridized only to a 14 kb mRNA
(Figure 1). There is a potential Rho-independent
terminator (a stern—loop followed by an A-T rich
sequence) immediately downstream of cotZ coding
region (Figure 1; Zhang et al., 1993). Another poten-
tial stem-loop structure could be formed in the
region between cotY and cotZ coding regions and
may also function as a transcriptional terminator or
a site for RNA processing (Figure 1; Zhang et al.,
1993). Primer extension with Pr3 (which corre-
sponds to codons 9 to 14 of cotY) mapped a 5'
terminus 39 bp upstream of the first codon of cotY
(Figure 2). In addition, primer Pr4 (which corre-
sponds to codons 20 to 25 of cotZ) produced a
731 bp extension product, indicating that the 5'
terminus of the [-4 kb mRNA is the same as that of
the 06 kb cotY mRNA. No potential internal
promoters were found in this region. It appears that
both the 14 kb and 06 kb mRNAs are transcribed
from the same promoter. Pu, immediately
upstream of the cot Y coding region. Termination at
the downstream stem-loop structure could produce
the 14 kb mRNA and termination or prowssing at
the internal stem-loop structure could generate the
06 kb mRNA (Figure 1). Total RNA from t1 cells
was used as a control in each of the primer extension
experiments and in no case were any of these exten-
sion products found (Figure 2).

(b) Regulation of transcription of the cotVWXYZ
cluster

None of the four cotVWX YZ mRNAs was
detected in RNA preparations from B. subtilis
strain 55-3 which has a mutation in the sigE gene
(spoIIGB) or from strain BK-tlo, which contains a
deletion (sp0111094; Errington et al.. 1988) encom-
passing the C-terminal half of the sigK gene (data
not shown). The absence of detectable transcripts in

   

gerE

ts ts (1 tr
corVWX g A? -l.6 Irb

Probe s .. J . _
th . ' —0.6 kb
cotYZ Q -l.4 kb

Probe C _ J», ’
cotZ "’ Q Q —0.6 kb

Figure 3. CotVWXYZ gene transcription in B. subtilis
wild-type strain .11qu (WT) and gerE mutant lAA-l
(gerE). Preparation of total RNA, fractionation and
hybridization to probe B (from the N terminus of the cotX
gene) and probe C (from the middle of the cot Y gene) were
asdescribedinthelegendtoFigurel.

49

the sigK mutant, which still produces a“ RNA
polymerase (Trempy at al., 1985), suggests that 0‘
RNA polymerase, and not a'E RNA polymerase,
transcribes the cot VWX Y Z gene cluster. The lack of
transcription in the sigE mutant presumably results
from the failure to produce a" (Kunkel et al., 1988;
Stragier et al., 1989; Lu et al., 1990). In a gerE
mutant strain (lAA-l), the 1-6 kb cotVWX, the
l-4kb cotYZ and the Mkb cotY mRNAs were
present, albeit in slightly reduced amounts relative
to the wild-type, but the 06 kb cotX mRNA was
essentially absent (Figure 3). These results suggest
that GerE is required for transcription from the P,
promoter and may enhance transcription from the
Pm and Pa promoters. The RNA hybridizing to
probe B was smeared, probably due to the instabi-
lity of the 1-6 kb cotVWX mRNA. Such smearing
was not observed when the same RNA was hybrid-
ized to probe C.

(c) In vitro transcription of genes in the
'cotVWXYZ cluster

In order to determine directly the effects of a"
RNA polymerase and GerE on transcription from
promoters in the cot V WX YZ cluster, linearized
DNA templates were transcribed with 0" RNA
polymerase partially purified from a gerE mutant in
the presence or absence of gel-purified GerE (see
Materials and Methods; Figure 4). Use of 0" RNA
polymerase resulted in run-off transcripts of the
expected sizes from Pu“ and there was a two- to
threefold stimulation of transcription when GerE
was present (Figure 4A). This level of stimulation
by GerE was reproducible in three separate experi-
ments. The larger of the two transcripts in lane 4 of
Figure 4A is probably an artifact resulting from
run-off transcription of DNA templates cleaved
with that particular restriction enzyme (SocI) since
only transcripts of the expected sizes were observed
in the other lanes. Run-off transcripts of the
expected sizes from PX were produced only when 0"
RNA polymerase was supplemented with GerE
(Figure 4B), consistent with the finding that the
06 kb cotX transcript was not detected in a gerE
mutant (Figure 3). 0" RNA polymerase produced
run-ofi' transcripts of the expected sizes from P"
and there was a twofold stimulation of transcription
when GerE was preunt (Figure 4C). 0" RNA poly-
merase reconstituted from B. subtilis core RNA
polymerase and gel-purified 0" also produced run-
off transcripts of the expected sizes from Pm. PX,
and Pa in the presence of GerE (data not shown).
These results support the existence of three
promoters in the cotVWX YZ cluster and each is
transcribed in vitro by a" RNA polymerase with
different levels of dependence upon GerE.

(d) GerE binding sites in the PH", P, and P"
promoter regions
GerE was shown previously to bind. to DNA at
particular sites upstream of the transcriptional start

 

 

Figure 4. Effects of GerE on transcription of promoters
in the cot VWX YZ cluster by a“ RNA polymerase in vitro.
Linearized lasrnid DNA was transcribed with partially
purified RNA polymerase (02 ug) alone. or with
control protein or GerE added immediately alter the
addition of RNA polymerase. Run-06' transcripts were
electrophoresed in 5% polyacrylamide gels containing
8 M urea and were detected by autoradiography.
Arrowheads denote the positions of run-off transcripts of
the expected sires. as judged from the migration of end-
labeled DNA fragments of MspI-digested pBR322. A.
Transcription of vax from pJZ30 (005 pmol) digested
with Sml (lanes 1 to 3. ll0base transcript) or SacI (lane
4. Ila-base transcript) with a" RNA polymerase alone
(lane 1). or with control protein (lane 2) or 075 ug of GerE
(lanes 3 and 4) added. B. Transcription of PX from N222
(005 pmol) digested with BgIII (lanes 1 to 3. 188-bsse
transcript) or EcoRI (lane 4. 288-base transcript) with 6‘
RNA polymerase alone (lane 1). or with control protein
(lane 2) or 075ug of GerE (lanes 3 and 4) added. C.
Transcription of P" from pJZG (005 pmol) digested with
Neal (lanes 1 and 2. 62-base transcript) or HincII (lanes
3. 481-bsse transcript) with a‘ RNA polymerase alone
(lane 1) or with l-0 ug of GerE added (lanes 2 and 3).

 

sites of genes such as 0018 and cotC which are
transcriptionally stimulated by this protein (Zheng
el al.. l992). Since GerE enhanced transcription
from all three promoters in the colVW X YZ cluster,
GerE binding in these promoter regions was exam-
ined by DNase I footprinting experiments carried
out with DNA probes end-labeled on each of the
strands (Materials and Methods: Figure 5). GerE
protected an approximately 24 bp stretch of DNA
extending from position —27 to position —50
upstream of the vax transcriptional start site on
the nontranscribed strand (Figure 5A). The protec-
tion may extend to position —54 but the absence of
DNase I digestion products from position —50 to
position —54 makes the boundary on this strand
uncertain (see broken line in Figure 6A). GerE also
protected an approximately 19 bp stretch extending
from position -—32 to position -50 on the tran-
scribed strand in the vax region (Figure 58), with
uncertainty in the upstream boundary of protection
between -—50 and —56 (Figure 6A). GerE protected
a large stretch of approximately 35 bp extending
from position —28 to position —62 upstream of the
PX transcriptional start site on the nontranscribed
strand (Figure 5C), and this protection may extend
to position —69 (Figure 6A). An approximately

50

C)
GI

V w?

H t I
“I... I 9'"-
la-a‘i .0-
1 21m
~ Inn-u
was"
IMOOw

I.“
"we ‘fi’? u n.,._

...-Oil 1 Mr» m

M"'O :4

i

...-W

m ’5
mm»

G
5‘.
I
9
a
-
a
i

   

S
V

i.-

.

V
I—III- ’°

-i

 

Figure 5. GerE footprints in the Pm. P, and Pu
promoter regions. DNA fragments end-labeled on the
nontranscribed or transcribed strand were incubated in
se te reactions: lane 1 without protein; lane 2 with
025ug; lane 305m; lsne4 lug; or lane 5 2ugofGerE;
or lane 6 with control protein. After DNase I treatment.
the partially digested DNAs were electrophoresed in 6%
polyacrylamide gels containing 8 M urea alongside a
sequencing ladder generated by chemical cleavage of the
respective end—labeled DNA. A and B. Footprints of the
nontranscribed (nucleotide sequence shown in Figure 6A)
and transcribed strands. respectively, of the Pm
promoter region. using probes labeled at the Self site
locawd downstream of the promoter in pJZ30. C and D.
Footprints of the nontranscribed (nucleotide sequence
shown in Figure 6A) and transcribed strands. respec-
tively, of the PX promoter region. using probes labeled at
the XbaI site located downstream of the promoter in
pJZ23. E and F. Footprints of the nontranscribed strand
(nucleotide sequence shown in Figure 0A) and transcribed
strands. respectively, of the P" promoter region. using
probes labeled at the N001 site located downstream of the
promoter in p.126. Arrowheads denote the boundaries of
protection by GerE and numbers refer to positions
relative to the transcriptional-start site.

-35 -10 _
an -17 hp- can-44a +1

_s_qaa_sattqqttatttttattattquct.ctqcaccccatttchrtstangaqta ’vux

 

 

taaaaastsqqqttcttcatcaggatatatgactcagtCaaaatsagaqqcthctCII'tt“Incaqta !‘

 

tgsatstatagacfitcacccacflcasqtqqqqcacgqqtammflqttssqqa P"

"66!

-19 Warm -CI not)
-53 sass-mecca: -42 act)

4.40 enemas: ~12! «cc st“ 1
-14 anrccacaqcc -‘3 cotC st:- 2
-53 ans-coon": 42 coem
-5. “macaque: -55 not:

-32 GAc'IGaG'rca‘la -43 out:

-52 511216“:th -41 «an

-33 flqtmrqssc -4‘ curs

"mm-ax Consensus

Figure 6. Alignment of promoters in the col VWX YZ cluster with the consensus sequence for d"- transcribed promoters
(A) and alignment of GerE binding sites (B). A. Nucleotide sequences (Zhang et al.. 1993) upstream of the transcriptional
start sites of the PH“, Px and P" promoters (Figure 2) are aligned with respect to conserved nucleotides (boldface.
capital letters) found in the — l0 and —35 regions of promoters transcribed by a" RNA polymerase (Foulger and
Errington l99l: Zheng at al.. 1992). shown at the top. Overlining and underlining indicate regions on the nontranscribed
and transcribed strands. respectively. protected by GerE from digestion with DNase I (Figure 5). The broken portions of
lines indicate regions of uncertain protection due to a lack of DNase I digestion in these regions. B. Nucleotide sequences
protected by GerE from digestion with DNase I are aligned with respect to the consensus proposed previously (Zheng a
01.. 1992). shown at the top. GerE binding sites in the cotB and cotC promoter regions are from Zheng el al. (1992) and in
the PM. PX and P" promoter regions are from Figure 5. Numbers refer to positions relative to the transcriptional start
site. Note that the sequences shown for the binding sites between —32 and —43 of the Px and between -33 and -« of
the P" promoter are from the opposite DNA strands as shown in part A of this Figure. The bottom line shows an
enlarged consensus sequence for GerE binding based on the sequences shown. B means purine. W means A or T. and Y
means pyrimidine. Nucleotides that match the consensus (at least 7 out of 9 at each position) are shown as boldface,
capital letters.

39 bp stretch extending from position -32 to posi- and cotX coding regions. A 14 kb mRNA is the
tion --70 on the transcribed strand in the 1’; region product of cotranscription of the cot Y and catZ
was also protected (Figure 5D). The strongest genes, while the 06 kb cotY mRNA appears to
protection by GerE was observed in the P" region, result from transcription from the same promoter
where an approximately '25 bp stretch of DNA from (Pu) and termination or processing at a potential
position -29 to position —53 (or possibly position stem-loop structure located between the colY and
-55) was protected on the nontranscribed strand eotZ coding regions (Zhang el al., 1993). The poten-
(Figure 5E) and an approximately 24 bp stretch tial stem-loop structure between the colY and cotZ
extending from position —34 to position —57 was genes is not a typical Rho-independent terminator
protected on the transcribed strand (Figure 5F). since there is not a stretch of T nucleotides on the
These results show that GerE binds immediately nontranscribed strand following the stem-loop
upstream and partially overlaps the -—35 region of structure. Thus, there is likely to be some trans-
all three promoters in the cot VWX YZ cluster. acting factor involved in the control of termination
or processing at this internal stem-loop structure
resulting in the accumulation of about equal
‘- m amounts of cotYZ and cotY mRNA. The

The cotl'WX Y Z gene cluster is organized into cotVWX Y Z cluster comprises the first example of
two multicistronic operons plus two overlapping genes involved in spore 00“ synthesis 01' assembly
monocistronic transcription units (Figure l). The WhiCh 8'0 organized into operons 38 well 88 over“?-
colV. W and X genes are cotranscribed as a [’6 kb ping single transcriptional units. Seven other genes
mRNA from the most upstream promoter. wa. In encoding spore coat proteins, COM through 1" Md
addition. a 06 kb cotX mRNA is synthesized from colT, are expressed from monocistronic transcrip-
an internal promoter, PX. located between the cotW tion units and are scattered widely on the chromOc

51

some (Donovan et al.. 1987; Zheng et al.. 1988:
Aronson st al.. 1989; Cutting et al.. 1991).

The clustering of the col VWX Y Z genes probably
reflects related functions and interactions among
their protein products in the formation of the spore
coat insoluble fraction. The CotX protein is a major
component of the spore coat insoluble fraction and
is probably extensively cross-linked (Zhang el al.,
1993). The CotY and Z proteins are similar to each
other in sequence and are present as disulfide cross-
linked multimers in spore coat extracts. There is
also evidence of interactions among CotX, Y and Z
in the spore coat since deletion of the cotX gene
resulted in an incIease in the amount of soluble
CotY and CotZ proteins (Zhang at al., 1993).

As with several other spore coat protein genes.
transcription of the cot VWX YZ gene cluster is
dependent upon a" RNA polymerase and to
different extents upon the GerE protein. Transcripts
were first detected at about stage IV-V of sporula-
tion which coincides with the onset of transcription
of two other spore coat genes. catD and cot’I' (data
not shown). The catD gene is known to be tran-
scribed by a“ RNA polymerase in the mother cell
(Kroos et al.. 1989; Zheng & Losick. 1990; Driks &
Losick. 1991) as are the cotVWX YZ genes. The
cotVWX YZ transcripts were not present in a sigK
mutant (data not shown) and partially purified 0"
RNA polymerase generated transcripts in vitro from
the vax and Pyz promoters, as well as from the P,‘
promoter when the GerE protein was present
(Figure 4). In general, therefore, the transcription of
the cotVWX YZ cluster fits the pattern of several
other spore coat genes. The absolute dependence of
the Px promoter on GerE for transcription in vitro
was consistent with the observation that there was
no 06 kb cotX mRNA in a gerE mutant (Figure 3).
The cotC promoter also exhibits almost absolute
dependence on GerE for transcription by 0" RNA
polymerase in vitro (Zheng et al.. 1992) and a cotC-
lacZ fusion was not expressed in a gerE mutant
(Zheng a Losick. 1990).

III vitro. transcription from the wa and Pyz
promoters was similar to that from the catD
promoter in that there was considerable transcrip-
tion in the absence of GerE and this basal level of
transcription was stimulated two- to threefold by
the addition of GerE (Figure 4; Zheng el al., 1992).
In vivo, the amounts of 1-6 kb colVWX mRNA,
14 kb cotYZ mRNA and 06 kb cotY mRNA in a
gerE mutant were less than in the wild-type (Figure
3) but the effect of GerE on accumulation of these
mRNAs appeared to be rather small compared to
the sevenfold enhancement of cotD-lacZ expression
provided by GerE in viva (Zheng & Losick, 1990).

Examination of the vax: P,‘ and Pyz promoter
regions reveals nucleotide sequence similarities to
other promoters recognized by a" RNA polymerase.
The three promoter regions are aligned with a
proposed 0‘ RNA polymerase recognition sequence
(Foulger & Errington 1991; Zheng et al., 1992) in
Figure 6A. All three promoters have five out of six
matches to the consensus in their -10 regions. Only

52

the Pyz promoter, which showsythe least dependence
upon GerE in vitro, has the conserved AC in its -35
region. The wa promoter has CC, rather than AC,
in its —35 region, indicating that CC can be toler-
ated since wa was transcribed appreciably by a"
RNA polymerase in vitro in the absence of GerE
(Figure 4A). The PX promoter has TC in its -35
region, as does the cotB promoter, while the cotC
promoter has AG at these positions (Zheng at al..
1992). These three promoters exhibit little or no
transcription by a“ RNA polymerase in vitro in the
absence of GerE, with strong stimulation upon the
addition of GerE (Figure 4B; Zheng et al., 1992).

One function . of GerE bound to these promoter
regions may be to relieve partially or completely a
requirement for interactions between a‘ RNA poly.
merase and nucleotides in the -35 region. For
example, the cII protein of bacteriophage .1 is a
transcriptional activator that apparently modifies
07° RNA polymerase recognition of the -35 legion
of the 1P“ promoter, and it does so by binding to
the —35 region on the opposite face of the DNA
helix from 07° RNA polymerase (Ho at al., 1983;
Shih 8t Gussin, 1983). Other examples include class
II transcription factors of E. coli which bind to the
-35 region of a promoter and activate transcription
apparently by contacting conserved region 42
(which is normally involved in recognition of the
-35 region of promoters) of 67° (reviewed in
Ishihama, 1993; Gardella at al.. 1989; Siegele et al..
1989). Despite. an extended region of similarity
between GerE and region 4 of sigma factors (Kahn
& Ditta, 1991), GerE does not appear to recognise
the same sequence as RNA polymerase in the
-35 region of promoters (see below). Rather, GerE
binds upstream of (and in some cases partially over-
lapping) the -35 region in the promoters examined
to date (Figure 6A; Zheng 2101., 1992). This position
of binding is typical for class I transcription factors
of E. coli which contact the C-terminal domain of
the a subunit of RNA polymerase when they stimu-
late transcription (Ishihama. 1993), although the 01
protein of .1 binds in this position and activates the
1P”. promoter apparently by contacting 07° (Li et
al.. 1994). '

Inspection of the sequences protected from
DNase I digestion by GerE in the Pm, Px and P"
promoter regions reveals similarities to the consen-
sus sequence for GerE binding. TPuGGPy, proposed
previously (Zheng et al.. 1992). The protected region
in the wa promoter region was similar to binding
site 2 in the cotC promoter region (Zheng et al., 1992)
in that GerE binding was relatively weak and 19 to
24 bp were protected. In both cases a sequence with
four out of five matches to the previously proposed
consensus (TtGGT at position —50 to position -48
for the wa region) is present in the protecwd
region (Figure 6A; Zheng et al., 1992). The analogy
extends further in that a second GerE binding site is
located- about 150 bp upstream,(and hence beyond
the known sequence; Zhang et al.. 1993) of the Pm
transcriptional start site (II. Ichikawa & L. Kroos.
unpublished results), whereas binding site 1 in the

cotC promoter region is centered about 135 bp
upstream of the transcriptional start site (Zheng el
al., 1992). Both GerE binding sites were present in
the vax-bearing template uwd for in vitro tran-
scription (Figure 4A) and we have not yet investi-
gated whether the upstream site is necessary for
stimulation of transcription by GerE. The upstream
site in the cotC promoter region was not necessary
for stimulation of transcription by GerE in vitro (R.
Halberg a L. Kroos. unpublished results cited in
Zheng el al.. 1992).

The protected region in the PK promoter (31 to
44 bp) was approximately twice the length of each
of two separate protected regions in the cotC
promoter region (Zheng el al., 1992) and is similar in
length to the protected region in the cotB promoter
which was proposed to encompass two GerE binding
sites (Zheng el al., 1992). Within this region of the
cotX promoter, two sequences separated by 19 hp
with four out of five matches to the proposed
consensus are found in inverted orientation with
respect to each other. The sequences are TAGGg
(position —63 to position —59 shown in Figure 6A)
and TGaGT (position -35 to position -39 on the
DNA strand opposite that shown in Figure 6A).
Thus, the PX promoter region may contain two
adjacent GerE binding sites in inverted orientation.
In the 00:8 promoter region, the two sequences
matching the proposed GerE binding site consensus
are in the same orientation and are separated by
21 bp (Zheng el al., 1992).

Relatively strong GerE binding was observed in
the Pyz promoter region (Figure 5) and the length of
the protected region (20 to 29 bp) is suggestive of a
single site. However. within the protected region
one sequence with four out of five matches to the
previously proposed consensus (TAGaC at positions
-49 to -45; Figure 6A) and another sequence
which matches the consensus perfectly (TGGGT at
positions -36 to -40 on the DNA strand opposite
that shown in Figure 6A) are found in inverted
orientation with respect to each other and are
separated by 4 bp. An analogous arrangement of
two 5 bp sequences (both TAGGC) in inverted
orientation and separated by 4 bp was noted in the
sigK promoter region (Zheng el al., 1992) and GerE
binds strongly to this region (R. Halberg, H.
Ichikawa & L. Kroos, unpublished results). An
alignment of sequences bound by GerE is shown in
Figure BB. The additional binding sites in the Pm,
PX and Pyz promoter regions permit definition of a
larger consensus sequence for GerE binding,
RWWTRGGY--YY (R means purine, W means A
or T and Y means pyrimidine).

The complex transcriptional pattern of the
colVWX YZ cluster, including overlapping operons
with monocistronic transcriptional units and differ-
ential regulation by GerE, is apparently important
for ensuring that an optimal amount of each protein
is made at the appropriate times for assembly into
the spore coat insoluble fraction. The actual protein
ratios are difficult to establish because of the insolu-
bility but are probably reflected in the steady state

53’

amounts of the various mRNAs. If that were the
case, there would be relatively more of the CotX
and CotY proteins, each of which is pivotal for the
cross-linking and assembly of the coat insoluble
fraction.

J. Zhang was supported by a fellowship from the
Purdue University Research Foundation. Research at
Purdue was supported by grant GM48019 from the
National Institutes of Health. and that at Michigan State
University by the Michigan Agricultural Experiment
Station and grant GM43585 from the National Institutes
of Health.

References

Aronson, A. I. a Fits-James, P. (1976). Structure and
morphogenesis of the bacterial spore coat. '

Aronson. A. 1.. Song. H.-Y. a Bourne. N. (1989). Gene
structure and precursor processing of a novel Bacillus
subtilis spore coat protein. Mal. Microbial. 3,
‘37-444.

Bourne. N.. Fits-James, P. C. a Aronson. A. I. (1991).
Structural and germination defects of Bacillus
subtilis spores with altered contents of s spore coat
protein. J. Bacterial. 173. 6618-6625.

Carter. H. L. a Moran. C. P. (1986). New RNA
polymerase a factor under spoo control in Bacillus
subtilis. Pros. Nat. Acad. Sci. U SA. 83. 9438—9442.

Cutting. S.. Zheng. L. & Losick. R. (1991). Gene encoding
two alkali-soluble components of the spore coat of
Bacillus subtilis. J. Bacterial. 173. 2915-2919.

Donovan. W.. Zheng. L., Sandman. K. & Losick. R.
(1967). Genes encoding spore coat polypeptides from
Bacillus subtilis. J. Mal. Biol. 196. 1-10.

Driks. A. & Losick. R. (1991). Compartmentaliaed
expression of a gene under the control of sporulation
transcription factor 0" in Bacillus subtilis. Proc. Nat.
Acad. Sci., USA. 88. 9934-9938.

Errington. J. & Mandelstam, J. (1986). Use ofa lacZ gene
fusion to determine the dependence pattern of
sporulation operon spoIIA in spa mutants of Bacillus
subtilis. J. Gen. Microbial. 132. 2967-2976.

Errington. J .. Rong. S.. Rosenkrants. M. S. & Sonenshein.
A. L. (1988). Transcriptional regulation and structure
of the Bacillus subtilis sporulation locus spoIIIG.
J. Bacterial. 170. 1162-1167. .

Feng. P. a Aronson. A. I. (1986). Characterization of a
Bacillus subtilis germination mutant with pleiotropic
alterations in spore cost structure. Curr. Microbial.
13. 221-226.

Ferrari. E.. Henner. D. J.. Perego. M. & Koch. J. A.
(1988). Transcription of Bacillus subtilis subtilisin
and expression of subtilisin in sporulation mutants.
J. Bacterial. 170. 289—295.

Foulger, D. a Errington. J. (1991). Sequential activation
of dual promoters by different sigma factors
maintains spaVJ expression during successive
developmental stages of Bacillus subtilis. Hal.
Microbial. 5. 1363-1373.

Gardella, T., Moyle. I-I. & Susslrind. M. M. (1989).
A mutant Escherichia coli 07° subunit of RNA
polymerase with altered promoter specificity. J. Mal.
Biol. 206. 579—590.

Goldman. R. C. a Tipper. D. J. (1978). Bacillus subtilis
spore coats: complexity and purification of a unique
polypeptide component. .7. Bacterial. 135, 1091-1106.

Hager. D. A. a Burgess. R. (1980). Elation of proteins
from sodium dodecyl sulfate-polyacrylamide gels.
removal of sodium dodecyl sulfate. and renaturation
of enzymatic activity: results with sigma subunit of
E. coli RNA polymerase. wheat-germ DNA topo-
isomerase. and other enzymes. Anal. Biacltcrn. 109,
76—86.

Halberg, R. a Kroos. L. (1992). Fate of the SpoIIID

switch protein during Bacillus subtilis s rulation

depends on the mother-cell sigma factor. . J. Mal.

Biol. 228. 840—849.

Y., Wulfi'. D. L. a Rosenberg. M. (1983).
Bacteriophage 1 protein cII binds promoters on the
opposite face of the DNA helix from RNA
polymerase. Nature (London). 304. 703-708.
Ishihama. A. (1993). Protein protein communication

within the transcription apparatus. J. Bacterial. 175,
2483-2489.

Jenkinson. H. F. (1981). Germination and resistance
defects in spores of a Bacillus subtilis mutant
lacking a coat polypeptide. J. Gen. Microbial. 127.
81-91.

Kahn. D. a Ditta. G. (1991). Modular structure of W:
homology of the transcriptional activator domain
with the -35 domain of sigma facters. Mal.
Microbial. 5. 987-997.

Kroos. L.. Kunkel. B. a Losick. R. (1989). Switch protein
alters specificity of RNA polymerase containing a
compartment-specific sigma factor. Science. 243.
626-529.

Kunkel. 3.. Sandman. K.. Panzer. S.. Youngman, P. &
Losick. R. (1988). The promoter for a sporulation
gene in the spolVC locus of Bacillus subtilis and its
use in studies of temporal and spatial control of gene
expression. J. Bacterial. 170. 3513-3522.

Kunkel. E.. Kroos. L., Poth. 11.. Youngman. P. & Losick.
R. (1989). Temporal and spatial control of the
mother-cell regulatory gene spoIIID of Bacillus
subtilis. Genes Develop. 3. 1736-1744.

Li. M.. Moyle. H. & Susskind. M. M. (1994). Target of the
transcriptional activation function of phage .1 cl
protein. Science. 263. 76-77.

Lu. 8.. Halberg. R. a Kroos. L. (1990). Processing ofthe
mother-cell a factor. a‘. may depend on events
occurring in the forespore during Bacillus subtilis
development. Prac. Nat. ‘Acad. Sci., U.S.A. 87,
9722-9726.

Mahmoudi. M. J; Lin. V. K. (1989). Comparison of two
different hybridisation systems in Northern transfer
analysis. BiaTeclrniqucs. 7. 331-334.

Maniatis. T., Fritsch. E. F. & Sambrook. J. (1982).
Molecular Cloning. A laboratory Manual. Cold
Spring Harbor Laboratory Press. Cold Spring
Harbor. NY.

Maxam. A. in Gilbert. W. (1980). Sequencing end-labeled
DNA with base—specific chemical cleavages. Methods
lnzyrnal. 65. 499-560.

Miller. K. (1987). Gel electrophores'u of RNA. Focus. 9.
14-16.

Ho.

Moir.A .(1961). Germination properties of a spore coat.
defective mutant of BaciUus subtilis. J. Bacterial.
146,1106-1116. -

Pandey, N. K. a Aronson. A. I. (1979). Properties of the
Bacillus subtilis spore coat. J. Bacterial. 137.
1208-1218.

Sambrook. J.. Fritsch. E. F. a Mauiatis. T. (1989).
Molecular Cloning. A Laboratory Manual. 2nd edit..
Cold Spring Harbor Laboratory Press. Cold Spring
Harbor. NY.

Sandman. K., Kroos, L., Cutting. 8.. Youngman. P. &
Losick, R. (1988). Identification of the promoter for a
spare coat protein gene in Bacillus subtilis and
studies on the regulation of its induction at late stage
of sporulation. J. Mal. Biol. zoo, 461—473.

Schaeffer. P., Ionesco. R.. Ryter. A. a Balassa. G. (1963).
La sporulation dc Bacillus subtilis: etude genetiqus et

. physiologique. Calloq. Int. Cent. Nat. Beck. Sci. 124.
663—563

Shih, M. a Gussin. G. N. (1%3). Difi'erential efi'ects of
mutations on discrete steps in transcription initiation
' at the AP“ promoter. Cell. 34, 941-949.
Siegele. D. A. Hu.J. 0.. Walter, W. A. tGrcm,C. A.
' (1989). Altered opromoter recognition by mutant
formsol'thea"o subunitoflscltsricbraaoliRNA
polymerase. J. Mal. Biol. ”6. 591-603.

Stragier. P., Kunkel. B., Kroos, L. & Losick. R. (1989).
Chromosomal t generating a composite
gene for a developmental transcription factor.
Science. 243. 507-612.

Thomas. J. O. & Kornberg. R. D. (1978). The study of
histone-histone a-ociations by chemical cross-
linking. Methods Cell Biol. 18, 429-440.

Trempy. J. E. Morrison- Plummer. J. & Haldenwang,
W. G. (1986). Synthesis of a”. an RNA polymerase
specificity determinant. is a developmentally
regulated event in Bacillus subtilis. J. Bacterial. 161.
360-346.

Wu. J.. Howard. M. G. a Piggot. P. J. (1969). Regulation
oftranscriptiou ofthe Bacillus subtilis spoIIA locus.
J. Bacterial. 171. 692-698.

Zhang. J.. Fits-James. P. C. & Aronson. A. I. (1993).
Cloningandcharseterisationofaclusterofgenes
encoding polypeptides poses“ in the imoluble
fraction of the spore coat of Bacillus subtilis.
J. Bacteriol. 175. 3757-3766.

Zheng. L. aliasick. R. (1990). Cascaderegulation ofspore
coat gene expression in Bacillus subtilis. J. Mol. Biol.
211646-660.

Zheng. L., Donovan. W. P. Fitz-James. P. C. & Losick.
R. (1988). Gene encoding a morphogenic protein
required in the a-embly of the outer coat of the
Bacillus subtilis endospore. Genes Develop. 2.
1047-1054.

Zheng. L., Halberg. R.. Roels. 8.. Ichikawa. K., Kroos. L.
a Losick. R. (1992). Sporulation regulatory protein
GerE from Bacillus subtilis binds to and can activate
or rcpt-ea transcription from promoters for mother-
cell-specific genes. J. Mol. Biol. 226. 1037-1060.

Edited by M. Gatlcsman

(Received 8 December 1993; accepted 5 M ay 1994)

54

 

CHAPTER IV

Negative Regulation by the Bacillus subtilis GerE Protein

55

Till JOURNAL” aim CW
01999 by The American Society for Biochemistry and “decals: Biolog, inc.

Vol. 274. No. 12. issue oil-sch 19. pp. 8322-8327, 1990
Printed in USA.

Negative Regulation by the Bacillus subtilis GerE Protein“

Waived for publication, October 15, 1998, and in revised form, December 24, 1998)

Hiroshi Ichikawa, Richard Halbergt, and Lee Kroosi
From the Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824

GerE is a transcription factor produced in the mother
cell compartment of sporulating Bacillus subtilis. It is a
critical regulator of cot genes encoding proteins that
form the spore coat late in development. Most cot genes,
and the gerE gene, are transcribed by on RNA polymer-
ase. Previously, it was shown that the GerE protein in-
hibits transcription in vitro of the sigK gene encoding
0". Here, we show that GerE binds near the sigK tran-
scriptional start site, to act as a repressor. A sigK-lacz
fusion containing the GerE-binding site in the promoter
region was expressed at a 2-fold lower level during spor-
ulation of wild-type cells than gerE mutant cells. Like-
wise, the level of SigK protein (Le. pro-4r“ and 0“) was
lower in sporulating wild-type cells than in a gerE mu-
tant. These results demonstrate that «ix-dependent tran-
scription of gerE initiates a negative feedback loop in
which GerE acts as a repressor to limit production of 0‘.
in addition, Geri-3 directly represses transcription of
particular cot genes. We show that GerE binds to two
sites that span the -35 region of the catD promoter. A
low level of GerE activated transcription of cotD by 01‘
RNA polymerase in vitro, but a higher level of GerE
repressed catD transcription. The upstream GerE-bind-
ing site was required for activation but not for repres-
sion. These results suggest that a rising level of GerE in
sporulating cells may first activate cotD transcription
tom the upstream site then repress transcription as the
downstream site becomes occupied. Negative regulation
by GerE, in addition to its positive eflects on transcrip-
tion, presumably ensures that o‘ and spore coat pro-
teins are synthesized at optimal levels to produce a ger-
mination-competent spore.

 

Starvation induces the Gram-positive bacterium Bacillus
subtilis to initiate a series of morphological changes that result
in the formation of a dormant spore (1). Early in the sporula-
tion process a septum forms that divides the cell into a larger
mother cell compartment and a smaller forespore compart-
ment. Each compartment contains a copy of the genome, and
different genes are expressed in each compartment. Gene ex-
pression drives further morphogenesis, including migration of
the septum to engulf the forespore in a double membrane,
deposition of cell wall-like material called cortex between the
membranes, and synthesis in the mother cell of proteins that
assemble on the surface of the forespore to produce a tough
shell known as the coat. The developmental process culminates

 

‘ This research was supported by Grant GM43585 from the National
institutes of Health and by the Michigan Agricultural Experiment
Station. The costs of publication of this article were defrayed in part by
the payment of page charges. This article must therefore be hereby
marked 'advertiscment’ in accordance with 18 0.8.0. Section 1734
solely to indicate this fact.

it Present address: McArdle Laboratory for Cancer Research, Univer-
sity of Wisconsin, Madison. WI 53706.

i To whom correspondence should be addressed. Tel.: 517-355-9726;
Fax: 517—353-9334; E-mail: kroostilot.msu.edu.

8322

with lysis of the mother cell to release a mature spore. When
nutrients become available again, the spore germinates, pro-
ducing a cell that resumes growth and division.

The program regulating transcription of sporulation genes is
exceptionally well understood (2). it involves the synthesis and
activation of four compartment-specific 0 subunits of RNA po-
lymerase (RNAP),‘ each of which directs the enzyme to tran-
scribe a particular set of genes. 0’ and 0° control forespore-
specific gene expression. in the mother cell, activation of o‘ is
followed by the synthesis and activation of a". In addition, two
small, DNA-binding proteins, SpoIIID and GerE, activate or
repress transcription of many mother cell-specific genes (3-6).
The mother cell transcription factors form a hierarchical regu-
latory cascade in which the synthesis of each factor depends
upon the activity of the prior factor, in the order 0', SpoIIID,
a", and finally GerE (7).

In addition to positive regulation between one transcription
factor and the next in the mother cell cascade, there is evidence
of negative regulation as well. a“ RNAP initiates a feedback
loop that inhibits transcription of the sigE gene encoding as (8).
Since an RNAP transcribes the spoIIID gene (9-11), production
of SpoIIID is also negatively regulated (8, 12). This facilitates
the switch from the early 0’:- and SpoIIID-directed pattern of
gene expression to the late 0"- and GerE-directed pattern.
GerE is not a component of this feedback loop (8); however.
there was reason to believe that GerE might initiate a second
feedback loop. GerE was shown previously to inhibit transcrip-
tion in vitro of the sigK gene encoding (7K (4). Whether GerE
inhibits sigK transcription in viva was in doubt, though. be-
cause a sigK-lacZ fusion was not overexpressed in gerE mutant
cells (13). We have resolved this paradox by mapping a GerE-
binding site in the sigK promoter region. This showed that the
sigK-Inez fusion examined previously did not contain the entire
GerEbinding site. Here, we demonstrate that GerE represses
sigK expression about 2-fold in vivo, and we discuss the impli-
cations of this negative feedback.

Overexpression of sigK is not the only defect in gerE mutant
cells. Expression of some cot genes encoding spore coat proteins
is reduced or absent, whereas expression of other genes is
increased. The resulting spores have defective coats and ger-
minate inefficiently (14). Previously, it was shown that purified
GerE binds to DNA sequences matching the consensus RW-
WTRGGY-YY (where R is purine; W is A or T; and Y is pyrim-
idine) and activates transcription in vitro from the cotB and
cotC promoters (4) and from three promoters in the cotVWXYZ
cluster (3). The position of binding ranges in the different
promoters from well upstream of the -35 region to partially
overlapping it. Here, we show that GerE binds to the cotD
promoter region at a position typical for activation of transcrip-
tion, and to a position downstream, which may cause repres-
sion of transcription. Our mapping of GerE-binding sites in the

 

‘ The abbreviations used are: RNAP, RNA polymerase; 'I‘SS. tran-
scriptiOnal start site; PCR, polymerase chain reaction; Tricine, N-[2-
hydroxy-1,l-bis(hydroxymethyl)ethyllglycine; bp. base pair.

“is paper is available on line at http://www.ibcprg

56

B. subtilis GerE Protein Is a Transcriptional Repressor

catD and sigK promoters provides the first information about
how GerE acts as a transcriptional repressor.

EXPERIMENTAL PROCEDURES

DNase I Footprinting—DNA fragments" labeled at only one end were
region, a Hindlll-
Xborl fragment of pBKltind (15) was labeled at the 3’ ends by the Klenow
enzyme fill- -in reaction and lo-J’PldCTP or at the 5’ ends by treatment
with alkaline phosphatase followed by T4 polynucleotide kinase and
lleA’l‘P. in both cases. the labeled DNAw ”digested with Perl
which cleaved off a small fragment containing the labeled Hindlil end,
so it did not interfere with subsequent DNase I footprint ing. The la-
bel ed Xbal and“ Is 164 bp downstream of the sigK transcriptional start
site (TSS). For analysis of the cotD promoter region, pLRKIOO (15) was
ted with Ecol-ll. which cleaves 227 bp upstream of the cotD T58,
and labeled either at the 3' end by the Klenow enzyme fill- no reaction
and [o-“PldA’l'P or at the 5' end by treatment with alkaline phospha-
tase followed by T4 polynucleotide kinase and [y-"PIATP in both
cases, the labeled DNA was digested with Hindlll, which cleaves down-
stream of the curb promoter, to produce eat that was
purified by elution from an 8% polyacrylamide gel (16). Labeled DNA
ants were incu with different amounts of GerE gel purified
from Escherichia coli engines eetored tooverproduce the protein as de-
scribed previously (3) and than mildly digested with DNase 1 according
to method 2 as described previously (4). except 3 pmol ofprobe was used
and a 1-fold (wlw) excess of poly(dI-dC) as compared with probe was
added as competitor. After DNase 1 treatment, the partially digested
DNAs were electrophoresed' In a 7% polyacrylamide.
urea alongside a sequenm'ng ladde appro-
teend -laboled DNA to the chemical cleavage reactions of Maxam

and Gilbert as described previously (16).
Construction of o sigK- -lacZ Fusion—DNA between -115 and +28

relative to the sigKTE'...‘

(PCR) using pBKlG (15) as the template The upstream primer was 5'
GA TGAAGAA TATI'I'l'l‘AAC 3’ and the downstream
primer was 5' 006 WCACAAAAGTATGTIAA 3'. Eco!!! and Hin-
dlll restriction mm toths 5' ends of the
upstream and downstream primers, respectiv ely, to allow directional
subcloning of the PCR product into EcoRl- -Hindlll-digested pTKlac
(17).“
and transformed into B. subtilis 23307. in which marker replacement-
type recombination tad an SP8: sigK- locZ specialised transducing
phageasdeacribedpreviously(18‘ ‘ ‘ it...
' , trpC2) )and 522.2

(germs trpC2) (19) with selection for resistance to chloramphenicol (5
rag/ml) on agar as described previously (20).

Measurement of B—Goloctosidase Activity—Sporulation was induced
by nutrient exhausn’on in DSM at 37 'C as described previously (20).
Samples (1 ml) were collected at hourly intervals during sporulation.
cells were pelleted. and pellets were stored at -70 ‘C prior to the assay.

 

 

I- nnnn

Miller(21). L... " ,Lm'g ‘ , “ ‘L ‘
One unit of enzyme hydrolyzes 1 pmol of substrate per min per unit of
initial cell absorbance at 595nm

Western Blot Analysis—Cells were induced to sporulate by nutrient
exhaustion In DSM at 37 ‘C as described previously (20) Samples (1
ml) were collected at hourly intervals during sporulation, and whole-
csll extracts were prepared- as described previously (22). Protein con-

Pmtn nslfi
pg) were _sepantod by sodium dodecyl sulfate (SDS)-l4% Prosieve
(FMC BioProducts) polyacrylamide gel electrophoresis with Trial
Tricins electrode bufi'er (0.1 II Tris. 0. I II 'l‘I-icine, 0.1% SDS) and
electroblotted onto lmmobilon-P membranes (Millipore). Blots were
incubated as described previously with polyclonal antioproo" anti
ies thatdetectboth panclro-oK 22).The darysntibodybd was

horseradishpe roxidase-coniuga ted anti- rabbit immunoglobulin G (Bio-
Rad). Chemiluminescence detection (ECL) was performed accordingto
the manufacturer‘s instructions (Amenham Pharmacis Biotech) The
signal intensities were quantified using a 68 505 Molecular lmager
System (Bio-Rad).

In Vitro Mnsrfiptron—rf‘ RNAP was partially purified from gerE
mutant cells as described previously (15). The enzyme was comparable
in protein composition ' catD and sigK-transcribing activities to
fraction 24 shown in Fig. 2 of Kroos er al. (15). Transcription reactions
(45 Id) were performed» descn bed previous usly (24). except that RNAP
was allowed to bind to the DNA template for 10 min at 37 'C before the
addition of nucleotides (the labeled nucleotide was [Ir-”PICTP). Hope

57

8323
A o B a
A612345 A012345
-‘— —- 3';*'-- O
' a:lllim ‘m‘
.. 3 -
91
+1 .
- l
- l
, -..-I
nentranscrlbeo 5 transcribes
air "and strand
C
... nah-w- ...
D

-“ "m '0
eilh::flmml .7
ml--Itm

AL

550.1." "‘ ' ‘ regl Radioactive
DNA fragments separately end-label; on the nontranscribed (A) or
transcribed mm) strand were incubatedin “reactions with-

In(bovinese rumalbumin, 310 pmol) only (lane 5) or with
6 pmol (lane 1), 12 pmol (lane 2), 60 pmol (lane 3), or 120 pmol (lane 4)
of gel-purified GerE in addition to the carrier protein and then ulr
jectodto toaseDN aselfootprintinginatotalvolumeof“ul.flotchedboau
indicate the region protected from DNase l digestion by GerE. Arrow-

 

to positions relative to the TSS.m as “deduced from sequencing ladders
generated by chemical cleavage of the respectivee rid-labeled DNA at
purines (lane 6 oucrA) )or guanines (lone G). C. position of the GerE-

 

ontranscribod strand” of the sigK promoter region is shown (13). Over-
lining and underlining indicate regi and
transcribed ’

. consensus
nce are shown as capital learn. and numbers refer to positions
relative to the 188 Note that the sequence shown for the site
between +18and t7isfromthe strandoppositothatshowninc.

rin (6 pg) was added 2 min alter the addition of nucleotide-to tops-vent
reinitiation After the reactions were stopped, 20 Id of each reaction
mixture was subjected to electrophoresis In a 5% polyacrylamide gel
y no
The signal intensities were quantified using a Storm 820 Phosphorlm-
agar (Molecular Dynami , _., the
catD promoter and lackingths upstream nGerE- -binding site, DNA be-
tween —44 and +227 relative to the 138 was amplified by the PCR
using pLRKlOO (15) as the template. The upstream primer was 5’
3' (the underlined portion corresponds

 

MW TGC
lined portion corresponds to a Hindlll site downstream of the cotD
promoter in pLRKlOO
RESULTS
Location of the GerE-binding Site in the sigK Promoter Re-
gion—Purified GerE was shown previously to strongly inhibit
transcription in vitro of the sigK gene by a" RNAP (4). How-
ever, expression of a sigK-locZ fusion in viva was unaffected by
a gerE mutation (13). The discrepancy between the two results
could be explained if the sigK~lacZ fusion did not contain a
binding site for GerE that mediates repression. To see if GerE
binds specifically in the sigK promoter region and, if so, to
determine the position of binding. we performed DNase I foot-
printing experiments.
Fig. 1 shows that GerE protected a stretch of DNA from
DNase l digestion that included the T38 of sigK and extended

on
CD
N)
I“

 

 

 

1501
a
a
I
2
a
9A
as. a:
3:: l”
'55
I l
5
v:
a! ...
as.
'95 I
8-
23
a.
0" s
0 8

Time (hr)

1750.2. .ng-qu expression in wild-type and gerE mutant
cells. sigK-directed fl-galacteeidase activity was measured at the indi-
cated times during sporulation of conganic wild-type (8638. U) and
gerE mutant (522.2, A) strains. Points on the graph are averages for
isolatssofeachtype.andenorborssbow 18D.oftbedata.

downstream. The protection spanned positions -4 to +19 on
the nontranscribed strand (Fig. 1A) and positions +1 to +20 on
the transcribed strand (Fig. 18). Complete protection from
DNase I digestion was observed at the lowest concentration of
GerE tested, indicating that GerE binds with relatively high
afinity to this site as compared with other GerE-binding sites
mapped previously (3, 4). Fig. 10 shows the sequence of the
sigK promoter in the region protected by GerE. Within the
protected region are two sequences in inverted orientation that
overlap slightly and match the consensus sequence for Ger-E
binding (Fig. ID).

We conclude that GerE binds to a site in the sigK promoter
region that overlaps the T88 and extends downstream. Pre-
sumably, GerE binding to this site represses sigK transcription
in vitro (4) by interfering with RNA polymerase binding and/or
a subsequent step in initiation. The location of the GerEobindo
ing site provides a plausible explanation for the lack of an efi'ect
of agerE mutation on sigK-lacZ expression reported previously
(13). The sigK portion of the fusion extended only to +4, so it
did not contain the entire GerE-binding site.

GerE Inhibits sigK Expression in Viva—To determine
whether GerE affects sigK expression in vivo, we constructed a
new sigK-Incl transcriptional fusion that includes the GerE-
binding site. DNA between -115 and +28 relative to the sigK
TSS was fused to tool, and the fusion was recombined into
phage SP3. The reeulh'ng SPflzzsigK-lacz phage was trans-
duced into wild-type and gerE mutant B. subtilis, creating
lysogens. Fig. 2 shows the average B-galactoaidase activity
during sporulation of three isolates of each type. The average
maximum activity was 2-fold higher in gerE mutant cells than
in wild-type cells, demonstrating that GerE inhibits sigK ex-
pression in vivo.

We also compared the level ofsi'gK gene products in devel-
oping wild-type andan mutant cells. The primary translation
product of sigK is pro-0", an inactive precursor that is protec-
lytically processed to active a“ (22). We used anti-pro-a" anti-
bodies to detect both pro-0" and a" in extracts of cells subjected
to Western blot analysis. Fig. 3 shows that the levels of pro-0’t
and a" are higher in gerE mutant cells than in wild-type cells
late in sporulation. Quantitation of the combined pro-aK plus
0" signal for the experiment shown in Fig. 3 and three addi-
tional experiments showed that the maximum level of SigK
gene products in wild-type cells during sporulation. on average,
reached 57% ($61». 1 SD.) of the level inan mutant cells. In

B. subtilis GerE Protein Is a Transcriptional Repressor

 

villa-type gerE
23458782345678 K
/pf0-O
...-v- ‘ “-\°K

Fro. 3. Levelsofpro-o‘and o‘duringsporulationofwild—typs
and gerE mutant calla Whole-cell extracts were prepared from wild-
type (8638) and gerE mutant (522.2) cells collected at the indicated
number of hours after the onset of sporulation in DSM. Proteins were
fractionated by SDS-PAGE and subjected to Western blot analysis with
anti-prov" antibodies, which detect both pro-0" and 0".

all four experiments, the level of both pro-oK and a" was
elevated in T., and T. samples from the gerE mutant compared
with wild type. These results show that GerE normally inhibits
the accumulation of pro-0" and a" during the late stages of
sporulation. It. is likely that GerE represses transcription of the
sigK gene, reducing the synthesis of pro-(rK and o". This would
explain the similar 2-fold decrease of sigK-directed B-galacto-
sidase activity (Fig. 2) and pro-o" plus trK (Fig. 3) in wild-type
cells compared with gerE mutant cells. The alternative expla-
nation that GerE in wild-type cells causes increased turnover of
pro-o" and o‘K and a similar increase in turnover of B-galacto-
sidase is unlikely because B-galactosidase activity from a sigK-
locZ fusion lacking the GerE-binding site identified in Fig. 1 is
similar during sporulation of wild-type and gerE mutant cells
(13), as is B—galactosidase activity from lacZ fusions to many
other genes.

GerE Inhibits catD Transcription in Vitro—In addition to
possible inhibitory effects of GerE on transcription of genes in
the o" regulon (due to the inhibition of o" accumulation by
Ger-E), GerE stimulates expression of certain genes in the o"
regulon (3, 4, 7). Expression of a cotD-lacZ transcriptional
fusion was 7-fold higher in developing wild-type cells than in
gerE mutant cells (7). It was shown previously that purified
GerE stimulates catD transcription by 0" RNAP 2—3-fold (4).
We discovered that at a higher concentration GerE inhibits
catD transcription in vitro. Fig. 4 shows the result of an in vitro
transcription experiment with a mixture of DNA templates
containing the catD or cotC promoter. The cotC promoter was
included as an internal control because GerE has been shown
to bind to a site centered at -68.5 relative to the T88 and to
activate transcription by 0" RNA polymerase (4). Lower levels
of GerE stimulated catD transcription about 2-fold (based on
quantitation of signals in the experiment shown and one addi-
tional experiment), as reported previously (4), but higher con-
centrations of GerE inhibited cotD transcription. The inhibition
was specific to catD, since cotC transcription was activated at
the highest concentration of GerE tested.

Location of the GerE-binding Site in the catD Promoter Rs-
gion—To understand how GerE both positively and negatively
afl'ects cotD transcription, we tested whether GerE binds in the
promoter region by performing DNase I footprinting experi-
ments. Fig. 5 shows that GerE protected a stretch of DNA from
DNase I digestion that included the ~35 region of the cotD
promoter and extended upstream. 0n the nontranscribed
strand. the protection spanned positions -39 to -25 relative to
the T88 (Fig. 5A), whereas on the transcribed strand positions
-60 to -23 were protected. We could not be certain whether
protection extends farther upstream on the nontranscribed
strand due to the absence of DNase I cleavage in this region
even in the absence of GerE. Gar-E appears to bind to the cotD
promoter region with lower affinity than to the sigK promoter
region, since a higher concentration of GerE was required to
observe protection from DNase I digestion (compare Figs. I and
5). Fig. 50 shows the sequence of the catD promoter in the
region protected by GerE. Typically, GerE protects a stretch of

58

B. subtilis GerE Protein Is a Transcriptional Repressor

12345

“‘0‘ ‘—

""‘--cotc

Pro. 4. fleet of Carl on eotD transcription in vitro. DNA tem-
plates (0.2 pmol each template) were transcribed with partially purified
o" RNAP (0.2 ug) alone (lane 1) or with 50 (lane 2). 100 (lane 3). 200
(lane 4), or 400 pmol (lane 5) of gel-purified GerE added immediately
after the aK RNAP. DNA templates were pLRKlOO ( l5) linearized with
HindIII (225-base cotD transcript) and a HaeIII-EcoRl restriction frag-
ment from lel(4) (196-base cotC transcript). The positions of run-oi!"
transcripts of the expected sizes, as judged from the migration of end-
labeled DNA fragments of MspI-digested pBR322. are indicated.

A B

3.1zsssizsssoss.
a a
:- - . .
- . g: I I 4. I ~I‘
as ‘i..‘ ':C.~ .2:
r :- 9 ' ‘
o O O
as ' ’
Q
.- - t .‘ . . I. ‘ . 1‘ .‘o
- ~ “
‘ "'1" .; . - .j, '
s -‘ 3s
”INPIIIOCHM transcribed
Otllfld .Iflll‘
C use.
45 ng'fééiiém- flimsyone as
D

~ssaaasaoarcrrr -‘2
-20 Wf-OI' -JI
W--?! cssesasu

has. Mlootprlntsln thscotDpso-oteeregionkadiosc-
tive DNA fragments separately end-labeled on the nontranscribed (A)
or transcribed (B) strand were incubated in separate reactions with a
carrier protein (bovine serum albumin, 310 pmol) only (lane I ) or with
12 (lane 2). 25 (lane 3). 50 (lane 4). or 100 pmol (lane 5) of gel-purified
GerE in addition to the carrier protein and then subjected to DNase I
footprinting in a total volume of 45 ul. See Fig. l legend for explanation
of bares. arrowheads, and numbers. C, position of the GerE-binding site
in the cotD promoter region. The nucleotide sequence of the nontran-
scribed strand of the cotD promoter region (7) is aligned with respect to
conserved nucleotides found in the -35 regions of promoters tran-
scribed by 0" RNAP (34). shown above the sequence. Nucleotides in the
catD -35 region that match the consensus are shown as boldface capital
letters. Overlining and underlining indicate regions on the nontran-
scribed and transcribed strands. respectively, protected by GerE from
DNase I digestion. The dashed lines indicate regions of uncertain pro-
tection due to a lack of DNase I digestion in these regions. Numbers
refer to positions relative to the T88. M indicates A or C. D. nucleotide
sequences within the GerE-protected region of the catD promoter are
aligned with the consensus sequence for GerE binding (3). Matches to
the consensus sequence are shown as capital letters, and numbers refer
to positions relative to the T88. Note that the sequence shown for the
binding site between -20 and -31 is from the strand opposite that
shown in C.

about 20 bp from DNase I digestion (3. 4). The long region of
protection (nearly 40 bp) observed on the transcribed strand in
the cotD promoter region suggests that GerE binds to two sites.
Within this region is a perfect match (positions -53 to ~42) to
the consensus sequence for GerE binding (Fig. 50). A second
match (7 out of 10) to the consensus sequence is present at
positions -20 to ~31. in inverted orientation with resmct to

59

8325
12345

catD-..-- ...
‘h‘Q—cotc

F106. Meetotaerlontranscriptioninottmolaootbte-o
plate lacking the upstream GerE-binding site. The amounts of
DNA templates. or“ RNAP. and GerE were the same as in the Fig. 4
legend. The cotD template was a 275-bp PCR fragment containing 44 bp
upstream of the T88 (228-base transcript) prepared as described under
“Experimental Procedures." The cotC template was described in the
Fig. 4 legend. The positions of run-00' transcripts of the sizes,
as judged from the migration of end-labeled DNA fragments of Mapl-
digested pBR322. are indicated.

the first match (Fig. 50), which probably accounts for GerE
binding in this region.

The Upstream GerE-binding Site in the catD Promoter Is
Necessary for Activation of Transcription in Vitro but Not for
Repression—We hypothesized that GerE bound at the up-
stream site in the cotD promoter (i.e. recognising the perfect
match to the consensus centered at -47.5) activates transcrip-
tion, because GerE binds at a similar position in the cotB (4)
and cotVWX (3) promoters and activates transcription. GerE
bound at the cotD promoter downstream site (centered at
-25.5) might repress transcription, explaining the inhibition of
cotD transcription we observed (Fig. 4). To test these ideas, we
repeated the in vitro transcription experiments with a catD
DNA template lacking the upstream GerE-binding site. As
shown in Fig. 6, GerE failed to activate transcription of this
template. However, GerE still repressed transcription of this
template, under conditions that permitted GerE to activate
cotC transcription. We conclude that GerE binding to the up-
stream site in the catD promoter is required to activate tran-
scription. GerE binding to the downstream site probably rs-
presses transcription, although we cannot rule out the
possibility that GerE binding farther downstream is also nec-
essary and was not detected by DNase I footprinting.

DISCUSSION

Our results strongly support the model that GerE is a re-
pressor of sigK transcription which lowers the level of pro-0"
and o" in cells during the late stages of sporulation. The
previous finding that GerE inhibits sigK transcription in vitro
(4) can be explained by GerE binding near the T83 (Fig. I) and
acting as a repressor. The previous observation that expression
of a sigK-lacZ fusion lacking the GerE-binding site in the
promoter region is unaffected by a gerE mutation (13), together
with our finding that expression of a sigK-lacZ fusion contain-
ing the GerE-binding site is 2-fold lower in wild-type cells than
in gerE mutant cells (Fig. 2). demonstrates the importance of
the GerE-binding site for sigK regulation in vivo. The lowering
of sigK expression by GerE in wild-type cells appears to result
in a comparable decrease in the level of SigK gene products
(Fig. 3). The simplest interpretation of these results is that
GerE represses sigK transcription, limiting the synthesis of
pro-ax and 0" during sporulation.

The negative effect of GerE on the 0" level during sporula-
tion would lower expression of genes for which 0" is the limit-
ing factor for transcription. Four genes in the o" regulon show
a lower level of expression in wild-type cells than in gerE
mutant cells. These are cotA (25). spoVF (26), csk22 (27). and
cotM (28). In the case ofcotA. there is evidence ofa direct effect
of GerE on transcription. Purified GerE inhibits cotA transcrip-
tion in vitro (4). It is unknown how much of the increased cotA
expression observed in a gerE mutant (25) results from loss of

 

8326 B. subtilis GerE Protein Is a Transcriptional Repressor

direct inhibition by GerE and how much results from an ele-
vated level of 0". Also unknown is whether GerE. represses
spoVF, csk22, or cotM directly. or whether these genes are
overexpressed in a gerE mutant because GerE fails to repress
sigK. It should be possible to answer some of these questions
and, more generally, to assess the importance of GerE repres-
sion of sigK during sporulation and germination, by deleting
the GerE-binding site in the sigK promoter.

The cotD gene, like many other genes in the ox regulon, is
positively regulated by GerE. For several genes, GerE has been
shown to bind upstream of the promoter ~35 region and stim-
ulate transcription by 0" RNAP in vitro (3, 4). GerE can in-
crease catD transcription in vitro as shown previously (4). but
we discovered that a higher concentration of GerE causes re.
pression (Fig. 4). DNase I footprinting revealed that GerE
protects not only DNA upstream of the -35 region in the catD
promoter, but the protection extends downstream through the
-35 region (Fig. 5). Binding that extends through the -35
region has not been observed in other promoters activated by
GerE (3, 4). We reasoned that GerE binding in the catD -35
region might repress transcription, and we showed that trun-
cation of catD promoter DNA at -44 prevented activation, but
allowed repression, by GerE (Fig. 6). These in vitro results
suggest the model that a rising level of GerE in sporulating
cells first activates catD transcription by binding upstream of
the -35 region and then represses transcription by binding to
a second site just downstream. It should be straightforward to
test the requirement for the upstream GerE-binding site for
activation in vivo, by making the appropriate 5' deletion (e.g. to
-44) and measuring expression of a fusion to a reporter gene.
However. testing the role of the downstream GerE-binding site
in repression in vivo will be more difficult. As noted above, one
must be careful to distinguish between direct repressive effects
of GerE and indirect effects due to GerE repression of sigK.

Fig. 7 illustrates the regulatory interactions between the
four mother cell-specific transcription factors (circled) and our
model for regulation of catD transcription at different times
during sporulation. Early in sporulation. of: and SpoIIID are
active in the mother cell (Fig. 7A). At this time, sigK is tran-
scribed by 0" RNAP, with SpoIIID serving as an essential
activator (6, 13). As we have shown for GerE (Figs. 5 and 6). it
was shown previously that SpoIIID binds in the -35 region of
the catD promoter and represses transcription in vitro (6). Fig.
7B shows the regulatory interactions in the mother cell slightly
later, at about 4 h into the sporulation process, when the
primary product of the sigK gene, pro-o“, is processed to active
0" in response to a signal from the forespore (22, 29). a" RNAP
initiates a negative feedback loop that inhibits transcription of
the sigE gene encoding 01:, which in turn lowers expression of
spoIIID (8. 12). As the levels of as RNAP and SpoIIID fall, cotD
transcription would no longer be repressed by SpoIIID. At the
same time. a" RNAP transcribes gerE (4, 30), and. according to
our model, the GerE produced initially activates cotD tran-
scription and represses sigK transcription. For a period, all four
mother cell-specific transcription factors probably affect tran-
scription of sigK, because SpoIIID activates both a: RNAP and
a" RNAP to transcribe sigK (6, l3). and GerE represses sigK
transcription (4) (Figs. 1—3). The positive autoregulatory loop
created by o‘K RNAP transcription of sigK is kept in check by
two negative feedback loops (Fig. 78). One inhibits transcrip-
tion of sigE and therefore inhibits production of o2 and SpoIIID
(8). The other leads to synthesis of GerE, which represses sigK
transcription directly (4) (Figs. 1-3). As the level of GerE rises
later in sporulation, GerE may also repress cotD (Fig. 7C).
Presumably, the complex regulatory interactions depicted in
Fig. 7 ensure that the four transcription factors. as well as

ugE' spolIID :ng catD

'

 

Fla. 7. Regulatory Interactions between mother cell-specific
transcription factors and a model for regulation oleotD tran-
scription at different times during sporulation. Dashed arrows
show gene (italicized) to product (proteins are circled) relationships.
Solid arrows represent positive regulation of transcription. Lines with
a barred end represent negative regulation of transcription. A. B, and C
represent early, intermediate. and late stages of sporulation, respec-
tively. as explained in the text.

structural proteins like CotD under their control, are made at
optimal levels for formation of the spore cortex and coat.

Our mapping of GerE-binding sites in the sigK and catD
promoters and our in vitro transcription experiments with catD
promoter fragments differing at the 5' end provide the first
insight into how GerE acts as a repressor. In both the sigK and
cotD promoters, GerE binds within the region likely to be
bound by 0'" RNAP. although the footprints of o" RNAP on
these promoters have not yet been determined. The positions of
GerE binding in these promoters suggest that GerE may re-
press transcription by interfering with the initial binding of o"
RNAP or a subsequent step in transcription initiation. It may
be possible to distinguish between these possibilities by meas-
uring the ability of 0‘K RNAP to bind to these promoters in the
presence or absence of GerE. In the case of cotD. it is unlikely
that GerE inhibits elongation of transcripts or inhibits tran-
scription by any other mechanism that does not depend on
sequence-specific DNA binding, because GerE stimulated cotC
transcription in the same reaction mixtures in which it inhib-
ited cotD transcription (Figs. 4 and 6). These same in vitro
transcription experiments. with cotD promoter fragments dif-
fering at the 5' ends, suggest that at a low GerE concentration,
binding to a site upstream of the -35 region activates tran-
scription, and at a higher GerE concentration, binding to a
second site just downstream represses transcription. In the
DNase I footprint experiment shown in Fig. 5. there is no

60

shill—w

 

B. subtilis GerE Protein Is a Transcriptional Repressar

indication that GerE binds with higher afi'mity to the upstream
site (a perfect match to the GerE binding consensus sequence)
than the downstream site (a 7 out of 10 match to the consen-
sus). Although we cannot rule out a slight difference in affini-
ties. another possible explanation is that a" RNAP competes
with GerE for binding to the downstream site but not the
upstream site.

GerE appeared to bind with higher affinity to the sigK pro-
moter region than to the catD promoter region (compare Figs. 1
and 5) or many other GerE-binding sites mapped previously (3,
4). Within the region of the sigK promoter protected by GerE
from DNase I digestion are two sequences in inverted orienta-
tion that overlap by 4 bases and match the consensus sequence
for GerE binding perfectly or in 7 out of 10 positions. An
identical arrangement of sequences matching the GerE consen-
sus was observed previously just upstream of the -36 region in
the cotYZ promoter, where GerE appeared to bind with high
sfinity and activated transcription (3). It is unknown whether
more than one molecule of GerE at a time binds the sigK and
cotYZ sites. Also unknown is whether GerE is monomeric in
solution. GerE is predicted to possess a helix-turn-helix DNA-
binding motif similar to that in several proteins who” three-
dimensional structure has been determined (31, 32). Some of
these well characterized proteins are dimeric and recognize
inverted repeats in the DNA-binding sites (33). The consensus
GerE-binding sequence is not palindromic, so perhaps GerE
can bind as a monomer. This does not exclude the possibility of
interaction between monomers at sites that exhibit palin-
dromic character, such as the sigK and cotYZ sites.

GerE appears to be similar to SpoIIID in terms of its DNA
sequence recognition characteristics. The SpoIIID-binding site
consensus sequence (WWRRACAR-Y) is of similar length and
degeneracy as that recognized by GerE (RWWTRGGY-YY) and
also is not palindromic (5, 6). Although SpoIIID appears to be
monomeric in solution,2 many strong SpoIIID-binding sites
exhibit a second good match to the consensus in inverted ori-
entation relative to the best match (6). Mutational analysis will
be required to assess the contribution of each consensus match
to binding of GerE and SpoIIID at sites with palindromic char-
acter and to determine whether two monomers interact (ag.
bind cooperatively) at these sites.

In summary, we have investigated the biochemical basis for
negative regulation by the GerE protein. It appears to act like
a classical repressor, binding in promoter regions at sites that
overlap the position of RNAP binding. GerE repression at the
sigK promoter lowers a" production during sporulation about
2—fold, potentially regulating the expression of many genes in
the a‘K regulon. In addition, GerE binds in the promoter regions

 

’ B. Zhang and L. Kroos. unpublished work.

FPFP!‘

8327

of certain genes in the a" regulon and directly represses or
activates transcription by a" RNAP. In the case of catD. it is
likely that GerE binds to a site upstream of the promoter ~36
region and first activates transcription. then, as its level rises
in sporulating cells, GerE binds to a site slightly farther down-
stream and represses transcription.

We thank B. Kunkel. R Losick. P. Zuber. and C.
Moran for providing bacterial strains, phages, and plasmids. We thank
S. Lu for providing anti-pro-o" antibodies and B. Zhang for helpful
discussions. We are grateful to J. Kagurri for comments on the
manuscript.

‘-L l .l

‘_“

REFERENCES

Stragier. P.. and Losick, R. (1996) Anna. Rev. Genet. 90. 297-341
Kroos. L. flung. 8.. Ichikswmld" andeY. T. N. (1999)Mol. Microbiol.31.
in press
Zhangp.J.. Ichikawa.1-1..l-1alberg.R..Kroos.L.sndAronson.A.l.(1994)
J. Mol. Biol. 240. 406-415
Zbeng.L..Halberg,R.Roels.S.. lchikawmllJtroos. L..sndLosick.R.(1992)
' J. Mol. Biol. 229. 1037-1050
Zhang. 8.. Daniel, R.. Errington. J.. and Kroos. L. (1997) J. Bacterial. 179.
972-975
Halberg. R. and Kroos. L. (1994) J. Md. Biol. 243. 425-436
Zheng. L., and Insick. R. (1990) J. Mol. Biol. 213. 646-660
Zhang. 8.. and Kroos. L. (1997) J. Bacteriol. 179. 6138-8144
Ranks]. 8.. Kroos. L, Path. 5.. Youngman. P., and Losick. R. (1989) Genes
Dev. 3. 1736-1744
Stevens. C. M. and Errington. J. (1990) Mel. Microbial. 4. 543-552
Tatti. K. M.. Jana. C. R.. and Moran. C. P., Jr. (1991) J. Bacteriol. 118.
7828-7833
Halberg. R. and Kroos. L. (1992) J. Mal. Biol. 228, 840-849
13. Kunkel. 8.. Sandman. R.. Panzer. 3., Youngman. P., and Losick. R. (1988) J.
Bacteriol. 170. 35134622
14. Mair. A. (1981) J. Bacteriol. 148. 1108-1118
16. Kroos. L., Kunkel. 8.. and Losick. R. (1989) Science 943. 526-529
16. Sambrook. J.. Fritsch. E. P., and Maniatis. ‘1‘. (1989) Molecular Cloning: A
Laboratory Manual, 2nd Ed. Cold Spring Harbor Laboratory. Cold Spring
Harbor, NY
17. Kenney. T. J.. and Moran. C. P. (1991) J. Bacteriol. 178. 32824290
18. Zubsr. P., and Losick. R. (1987) J. Bacteriol. 199. 2223-2230
19. Errington. J.. and Mandelstam. J. (1986) J. Gen. Microbiol. 182. 2987-2976
20. Her-wood. C. R.. and Cutting. 8. M. (1990) Molecular Biological Methods for
Bacillus. John Wiley & Sons Ltd. Chichester. UK
21. Miller. J. (1972) Experiments Ill Molecular Genetics. pp. 352-356. Cold Spring
Harbor Laboratory. Cold Spring Harbor. NY
22. Lu. 8.. Halberg. R.. and Kroos. L. (1990)Proc. Natl. Acad. Sci. U.S.A. 97.
9722-9726
23. Bradford. M. M. (1976)AnaL Biochem. 72. 248-254
24. Carter. H. L., and Moran. C. P. (1986) Proc. Natl. Acad. Sci. U.S.A. II.
9438—9442
25. Sandman. K. Kroos. L., Cutting. 8.. Youngman. P.. and Losick. R. (1988)
J. Hal. Biol. 200. 461-473
26. Daniel. R. A.. and Errington. J. (1993) J. Mal. Bid. 292. 468-483
27. Henriques. A. 0.. Bryan. E. It. Beall. B. W.. and Moran. C. P. (1997)
J. Bacteriol. 179. 389-398
28. Henriques. A. 0.. Beall. B. W.. and Moran. C. P. (1997) J. Bacterial. 119.
1887-1897
29. Cutting. 8.. Oke. V.. Driks,A..Losick.R.. Lu.8.. andKroos. L(1990)C¢ll.2.
239-250
30. Cutting. S. Purser. 8.. andLosick.R(1989)J. Mal. sum. 393-404
31. Holland. S. K., Cutting. 8.. andMandelstarn. J. (1987)J. Gen. Microbid. 183.
2381-2391
32. Kshn. D.. and Ditts. G. (1991) Mol. Microbiol. 5. 987-997
33. Pabo. C. 0., and Ssuer. R. '1‘. (1992)Annu. Rev. Biochem. 41. 1063-1096
34. Roels. 8.. and Losick. R. (1995) J. Bacteriol. 177. 6263-6275

2‘? $99.“?

Hen-
’6

61

CHAPTER V

Combined Action of Two Transcription Factors Regulates Genes Encoding Spore Coat

Proteins of Bacillus subtilis

62

'nrs Jounmu. or Blaimicru. Grammy
O2m0by1bAmencanSocietyforBiochemandMobmlarBiolcgana

Vol. 275. No. 18. issue of May 5. pp. 13849-13855, 2000
Printed in USA

Combined Action of Two Transcription Factors Regulates Genes
Encoding Spore Coat Proteins of Bacillus subtilis“

Hiroshi Ichikawa and Lee Kroost

Received for publication, February 7. 2000

From the Department afBiachemistry. Michigan State University, East Lansing, Michigan 48824

During sporulation of Bacillus subtilis. spore coat pro-
teins encoded by cat genes are expressed in the mother
cell and deposited on the forespore. Transcription of the
cotB, cotC, and cotX genes by tr" RNA polymerase is
activated by a small. DNA-binding protein called GerE.
The promoter region of each of these genes has two
GerE binding sites. 6' deletions that eliminated the more
upstream GerE site decreased expression of lacZ fused
to cotB and cotX by approximately 80% and 60%, respec-
tively but had no eflect on cotC-incl expression. The
cotC-lacz fusion was expressed later during sporulation
than the other two fusions. Primer extension analysis
confirmed thatcothRNAincreasesfirstduringspor-b
nlation, followed by cotX and cotC mRNAs over a 2oh
period. In vitro transcription experiments suggest that
the diflerential pattern of cat gene expression results
from the combined action of GerE and another tran-
scription factor. SpoIIID. A low concentration of GerE
activated cotB transcription by a“ RNA polymerase,
whereas a higher concentration was needed to activate
transcription of cat! or cotC. SpoIIID at low concentra-
tion repressed cotC transcription, whereas a higher con-
centration only partially repressed cotX transcription
and had little eflect on cotB transcription. DNase I foot-
printing showed that SpoIIID binds strongly to two sites
in the cotC promoter region, binds weakly to one site in
the call promoter, and does not bind specifically to
cotB. We propose that late in sporulation the rising level
of GerE and the falling level of SpoIIID, together with
the position and affinity of binding sites for these tran-
scription factors in cat gene promoters, dictates the tim-
ing and level of spare coat protein synthesis. ensuring
optimal assembly of the protein shell on the forespore
surface.

 

Upon starvation, the Gram-positive bacterium Bacillus sub.
tilis initiates a sporulation process involving a series of mor-
phological changes (1). The rod-shaped cell undergoes asym-
metrical division into two compartments, a larger mother cell
and a smaller forespore. Different sets of genes are expressed
from the genome in each compartment. As sporulation pro-
ceeds. the forespore is engulfed by the mother cell, forming a
free protoplast surrounded by a double membrane inside the
mother cell. Cell wall-like material called cortex is then depos-
ited between the forespore membranes. Transcription of cat

 

‘ This research was supported by National Institutes of Health Grant
GM43585 and by the Michigan Agricultural Experiment Station. The
costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked “adver-
tissment" in accordance with 18 [1.8.0. Section 1734 solely to indicate
this fact.

3 To whom correspondence should be addressed: Dept. of Biochemis-
try, Michigan State University, East Lansing, MI 48824. Tel.: 517-355-
9728; Fax: 617-353—9334; E-mail: kroosOpilot.msu.edu.

Mmhssalsbleoelnsthttpd/wwwjbaoq

63

genes, which encode spore coat proteins. occurs in the mother
cell. The coat proteins assemble on the surface of the forespore.
forming a tough shell that protects the spare from environmen-
tal insults after it is released by lysis of the mother cell. When
nutrients become available again, the spare germinates, pro-
ducing a cell that resumes growth and division.

The sporulation process of B. subtilis has been studied as a
model to understand the relationship between developmental
morphogenesis and gene regulation (2). A central feature of
sporulation gene regulation is the synthesis and activation of
four compartment-specific 0‘ subunits of RNA polymerase
(RNAP).’aFanda°directRNAPtotranscribegenesinthe
forespore. 0E and a“ direct transcription in the mother cell.
The four 0 factors form a regulatory cascade in which the
activation of each a depends upon the activity of the prior 0 in
the order a’, a”, a0. and finally a'K (3). Activation of each of the
latter three a factors appears to be coupled to a morphological
step in development and involves signaling between the two
compartments.

In the mother cell. two accessory transcription factors,
SpoIIID and GerE. modulate RNAP activity at specific promoto
era (2). GerE is an 85-kDa protein that binds to DNA se-
quences resembling RWWTRGGY-YY (R is purine, W is A or
T. and Y is pyrimidine) and activates transcription of many cat
genes by a“ (4, 5). GerE can also act as a repressor (6). Like-
wise. SpoIIID is a 10.8-kDa protein that binds to DNA se-
quences resembling WWRRACAR-Y and activates or represses
transcription of many different genes (7. 8).

We have investigated transcriptional regulation of the cotB,
cotC. and cotX genes. These genes were known to be tran-
scribed by aK RNAP, with activation by GerE (4. 5). Two GerE
binding sites had been mapped in the promoter region of each
gene (Fig. 1) (4, 5). Here we report that 6' deletions that
eliminated the more upstream GerE site reduced expression of
cotB and call! but not cotC. Interestingly, we found that the
three genes are differentially expressed during development,
suggesting an additional level of regulation. SpoIIID appears to
provide the additional control, based on the results of in vitro
nanscription and DNase I footprinting experiments presented
here and based on how the SpoIIID level has been shown to
change during sporulation (9, 10). This is the first study to
correlate differential transcription of cat genes with the com-
bined action of GerE and SpoIIID. The discovery of such com-
plex regulation of cat gene expression leads us to speculate that
synthesis of cost proteins is finely tuned to ensure optimal
assembly of the spare coat.

MATERIALS AND METHODS

Construction ofaat-lacZ Fusion Strains—DNA fragments containing
the cotB. cotC, or cat)! promoter region flanked by EcoRI and HindIII
restriction sites at the upstream and downstream ends. respectively,

 

‘ The abbreviations used are: RNAP. RNA polymerase; PCR. polym-
erase chain reaction; R, purine; W, A. or T; Y, pyrimidine.

13849

13850

were synthesized by the polymerase chain reaction (PCR) and direc-
tionally subcloned into EcoRI-I-IindlII-digested pTKlac (11). The tem-
plates of the PCR were pBDl36 (4), lel (4). and p.IZ22 (5), respec-
tively. for cotB. cotC. and cotX. Plasmids containing the cotB promoter
region from -85 to +37 (pl-113) or from -60 to +37 (le4) were const-
ructed using the upstream primer 5'-GGGAATI‘CGCGTGAAAATGG-
GTAT-3' or 5'-GGGAATI‘CAAGCGACAATI‘AGGCTo3' for plil3 and
p814. respectively, and the downstream primer 5‘-GCGAAGQTTAA -
TCCTCCTAGTCA-3' (the restriction site in the primer is underlined).
Plasmids containing the cotC promoter region from - 153 to +13 (pl-[16)
or from —79 to + 13 (pill?) were constructed using the upstream primer
5'-CGG_AAm1‘GTAGGATAAATCGTl‘-3' or 5'-CGG¢5TTCTCTATC-
ATl'l‘GGACAG-S' for p316 and le'l. respectively. and the down-
stream primer 5'-CGGAAQC'1'I‘I'I‘ATI‘1'1'I‘ACTACG 3'. Plasmids cont-
aining the cot)! promoter region from -69 to + 10 (leS) or from -54 to
+10 (pHI9) were constructed using the upstream primer 5'-CGQMT_-
EAAAAAATAGGGTI'CTT-3‘ or 5'-CGQAA1ETCATCAGGATATAT-
GA-3' for pHIB and le9, respectively. and the downstream primer
5'CGGWACTGTTAT-3'. These plasmids were linear-
ised by digestion with Bsol and transformed into B. subtilis 23307. in
which marker replacement-type recombination created an SPB-special-
ised transducing phage containing the lad fusion as described previ-
ously (12). A phage lysate was prepared by heat induction and used to
transduce B. subtilis SG38 (spo’ trpC2) and 522.2 (gerE38 trpC2) (13)
with selection of lysogens resistant to chloramphenicol (5 udml) on LB
agar as described previously (14).

Measurement of B—Galactosidase Activity—Sporulation was induced
by nutrient exhaustion in Difco sporulation medium at 37 'C as de-
scribed previously (14). Samples (1 ml) were collected at hourly inter-
vals during sporulation. cells were pelleted. and pellets were stored at
-20 ’C before the assay. The specific activity of B—galactosidass was
determined by the method of Miller (15). using o-nitrophenol-B-D-galac-
topyranoside as the substrate. One unit of enzyme hydrolyzes 1 pmol of
substrate/minlunit of initial cell absorbance at 595 nm.

Primer Extension Analysis—At hourly intervals from 3 to 7 h after
the onset of sporulation. cells were harvested by centrifugation
(11.950 X g for 10 min). and RNA was prepared as described previously
(16) except the RNA was resuspended in 100 pl of water that had been
treated with 0.1% (v/v) diethylpyrocarbonate. The RNA was treated
with DNase I to remove contaminating chromosomal DNA. Primer
extension motions were performed as described previously (17. 18).
The cotB and cotC primers were those designated as Pr2 previously
(19). The cotX primer we used was also called Pr2 previously (5). Aher
the reaction. the extension products were subjected to electrophoresis in
a 5% polyacrylamide gel containing 8 I urea. and transcripts were
detected by autoradiography. The signal intensities were quantified
using a Storm 820 Phosphorlmager (Molecular Dynamics).

In Vitro Transcription—o" RNAP was partially purified from gerE
mutant cells as described previously (20). The enzyme was comparable
in protein composition and in cotD- and sigK-transcribing activities
with fraction 24 shown in Pig. 2 of Kroos et al. (20). GerE was gel-
purifed fiom Escherichia coli engineered to overproduce the protein as
described previously (4). SpoIIID was gel-purified from fractions of
partially purified a“ RNAP as described previously (20). Transcription
reactions (45 ul) were performed as described previously (21). except
that RNAP was allowed to bind to the DNA template for 10 min at 37 'C
before the addition of nucleotides (the labeled nucleotide was
[or-”PICTP). Heparin (6 pg) was added 2 min after the addition of
nuchotides to prevent reinitiation. After the reactions were stopped by
adding 40 pl of stop buffer (100 mu Tris-HCl. pH 8.0. 50 mil EDTA. 200
pg of yeast tRNA/ml). each reaction mixture was incubated with 250 pl
of ethanol at -70 ‘C for 1 h to allow precipitation of transcripts. Pre-
cipitates were pelleted at 12,000 X g for 15 min and resuspended in 10
pl of loading dye (80% formamide. 10 mil EDTA, 1 mg/ml xylene cyanol
FF and 1 mg/ml bromphenol blue). The resuspension was placed at
100 °C for 3 min, then subjected to electrophoresis in a 5% polyacryl-
amide gel containing 8 M urea. and transcripts were detected by auto-
radiography. The signal intensities were quantified using a Storm 820
Phosphorlmager (Molecular Dynamics). To prepare a DNA template
containing the cotX promoter. DNA between -97 and +82 was ampli-
fied by the PCR using pJZ22 (5) as the template. upstream primer
5'~CGGAAAAACGATAACAATI‘AG—3' and downstream primer 5'-CA-
TCTAACGGATGGTCACAGTCAG-a'.

DNase I Footprinting—DNA fragments labeled at only one end were
prepared as follows. For analysis of the cotC promoter region, DNA
between -143 and +30 was amplified by PCR using lel (4) as the
template: upstream primer. 5'-ATCG'I'I'I‘GGGCCGATGAAAATC-3';
downstream primer. 5'-CCCATATATACTCCTCCTI'I‘A‘I‘T.3'. To gen-

B. subtilis SpoIIID and GerE Regulate Cot Genes

I cotB

as ~41;

 

 

-85 -60
4345 -68.5 I' ""
+ ‘4 4 _‘ 001C
-153 -79
.605 l
;—‘; ‘__ cotX
'69 '54 .375

Pro. 1. Ger-B binding sites in the cotB. cotC. and cotX
regions. Large arrows indicate the position of the transcriptional start
site. Small arrows indicate the orientation of sequences that match the
consensus sequence for GerE binding, and numbers above or below
these arrows indicate the position of the center of the sequence. Num-
bers,below vertical lines indicate the end points of 5' deletions used in
this study.

crate a probe for each DNA strand. two separate reactions were per-
formed containing one or the other of the PCR primers labeled at the 5'
end by treatment with T4 polynucleotide kinase and [ra’PlATP and
purified by passage through a MicroSpin G-25 Column (Amersham
Pharmacia Biotech). DNA probes for analysis of the cotX promoter
region were prepared as described previously (5). Labeled DNA frag-
ments were incubated with different amounts of gel-purified SpoIIID
and then mildly digested with DNase 1 according to method 2 of Zheng
et aL (4). except 0.4 units of DNase l was used. and a 7ofold (wlw) excess
of poly(dA-dT) or poly(dI-dC) as compared with cotC or cotX probe,
respectively, was added as competitor. After DNase l treatment. the
partially digested DNAs were electrophoresed in a 7% polyacrylamide
gel containing 8 M urea alongside a sequencing ladder generated with
T7 Sequenase V 2.0 (Amersham Pharmacia Biotech) and the appropri-
ate primer for cotC or by chemical cleavage of the appropriate and-
labeled DNA for cotX.

RESULTS

Role of Upstream GerE Binding Sites in cot Gene map.
tion—The cotB. cotC, and cod! promoter regions each have two
GerE binding sites (Fig. 1) (4. 5). To test the importance of the
more upstream GerE site in transcription of each gene. we
fused promoter DNA fragments containing different amounts
of upstream sequence to lacZ. These fusions were recombined
into the lysogenic phage SP8, and each phage was transduced
into wild-type and gerE mutant B. subtilis, where the phage
integrated into the chromosome at the attachment site. Trans-
ductants were induced to sporulate by nutrient exhaustion.
and B-galactosidase activities were measured. Fig. 2 shows
that deletion of the more upstream GerE site reduced cotB-lacZ
and cotX-lacZ expression by approximately 80% and 60%, re-
spectively. Deletion of the more upstream site centered at
- 134.5 had no effect on cotC-lacZ expression. All the fusions
failed to be expressed in the gerE mutant (Fig. 2 and data not
shown). These results demonstrate that the more upstream
GerE site is not important for cotC transcription and suggest
that the more upstream sites contribute greatly to GerE tran-
scriptional activation of cotB and cotX.

cot Genes Exhibit Different Patterns of Expression—The data
in Fig. 2 suggest that the cotB. cotX. and cotC promoters are
regulated differently. Expression of cotB-lacZ rose sharply be-
tween 4 and 5 h into sporulation and reached a maximum at
6 h. cotX-lacZ expression also increased between hours 4 and 5
but continued to rise until hour 7. Expression of cotC-lacZ
began later. between hours 5 and 6, and rose until hour 7.

To further examine the apparent difference in the pattern of
cat gene expression, we measured the appearance of cotB. cotX,
and cotC mRNAs during sporualtion of wild-type cells using
primer extension analysis. Fig. 3A shows a representative re.

64

B. subtilis SpoIIID and GerE Regulate Cot Genes

a
:

 

 

 

 

 

A I'll
so
w v
JO
Azo
5 I0
5 c
.15
B e
530*
j;
i”
g 10 .
'5 ..
9 v
C in.
20
ml '
I T "

 

 

 

I s s e 7 8
Hours alter the onset of sporulation

FIG. 2. Expression of cob lacZ fusion. cotB (A). cotX (B) and cotC
(C) promoter regions with (Cl) or without ( O ) the more upstream Ger E
site (5' end points are indicated in Fig.1) were fused to lacZ. and
fl-galoctosidaoe activity during sporulation of wild-type 8038 was

meoIqured escribed under 'Materiale ondM e.'thod.I . ex-
pression of fufions containing the u om GerE site was meas-
ured during sporulation of gerE mutant 522.2 (0). Points on the graph
are averages for isolates of each type, and error bars show 1 SD. of the
data.

Rclaliw mRNA level

 

3 4 5 6 7
Hunt! .II'Ier Ihe onset of unrulatinn

FIG. 3 Levels of cotB. cotC, and cg“ mR‘J'VA during sporula-
tion. RNA was preporedfm
number of hours alter the onset of Iporulotion in Difco “sporulation
mediumwAcothotC andcotXmRNA ooydetectedb rexten
sion analysis. B. primer extension signals for cotB I‘d) ESC (0). and
cotX (0)" "l

foreachmRNA. Poinunnth- gr-ph '
Aprepared from two different culture‘s. and error bars show 1

SD. of the data.

 

 

suit from an experiment in which primers for all three mRNAs
were mixed with RNA. and primer extension was done simul-
taneously. Similar results were obtained when primer exten-

 

13851
A
B
0 [00 200 300
Amount of GerE (pmol)
FIG. 4. Efleet of GerE on cotB. cotC. and ptlon in

pmol (lane 8) of gel-purifiedGe GerE added
RNAP. DNA templates were I 478-boae pair Poul! fragment of p301”
(HO-base cotB transcript), a 338-base pair Halli-Eco!!! fragment of
transcri ‘ PCngeneroted
. orunf -ofi' transcripts
sizes. as judged fromthe migrationn of end-labeled DNA
for cotB ([3). :dtC ( 0). and cotX (0) were quantified and normalized to
the maximum signalf or each transcri on the graph are over-
ageaof normalised Iignols from two experiments, and error bars III
1 SD. of the dots.

 

sion was done separately for each gene (data not shown). cotB
mRNA was detectable at 4 h into sporulation, and its level rose
sharply at hour 5. Reproducibly, cotB mRNA was undetectable
at hour 6. then reappeared at hour 7, indicating that synthesis
and/or stability of this mRNA is mreg'ulated by an unknown
YmRNA was barely detect-
able at hour 4 rose to its maximum level at hour 5. and fell to
a barely detectable level at hour 6. cotC mRNA was present at
a low level at hour 3. The enzyme responsible for this low level
of cotC transcription is unknown, but it could be «3 RNAP
because a”: and 0" 'ze similar sequences in cognate
promoters (22, 23). The cotC mRNA level increased at hour 5
and continued to rise until hour 7. Fig. 3B shows quantification
of the experiment shown in Fig. 3A plus one independent
experiment. The average level at different times is plotted
relative to the maximum level for each mRNA to illustrate the
different patterns of mRN A accumulation. These results to-
gether with the tool expression data (Fig. 2) suggest that cotB
transcription is induced slightly earlier than that of cotX,
whereas full induction of cotC transcription lags behind that of
cotX

 

A Lower Concentration of GerE Activates cotB Transcription
than cotX or cotC Transcription—To investigate how different
patterns of col gene expression might be established, we per-
formed in vitro transcription with UK RNAP and different
amounts of GerE Fig 4A shown a r
which an equimolar mixture of cotB. cotX, and cotC DNA tem-
plates was transcribed by RNAP (partially purified from

65

13862

gerE mutant cells) in the presence of increasing GerE. In this in
vitro systsm.allthreegeneswere transcribedintheabsenceof
GerE. whereas expression of tool fused to these genes was not
observed in gerE mutant cells (Fig. 2). The addition of GerE
activated transcription of all three genes in vitro, as expected
(4. 5). but interestingly, a lower concentration of GerE was
sufficient to activate cotB transcription, whereas a higher con-
centration was needed to activate cotX or cotC transcription
(Fig. 4A). The experiment was repeated, and transcript signals
from both experiments were quantified and normalized to the
maximum signal obtained for each template. Fig. 4B shows
that. on average, cotB transcription was activated about 3-fold
and 0.6 an GerE was required for half-maximal activation. The
activation profiles for cotX and cotC were very similar. Both
genes were activated more than 10-fold. and half-maximal ac-
tivation required 4 Int GerE. These results suggest that earlier
expression of cotB during sporulation (Figs. 2 and 3) may result
from a lower threshold for activation by GerE. since the level of
GerE is believed to increase as o“ RNAP becomes active (24).
The results do not explain the apparent differential expression
ofcotXandcotC(Figs. 2 and3), suggestingthere mightbean
additional level of control.

SpoIIID Is a Potent Repressor of cotC Transcription—We
discovered that extracts of sporulating wild-type cells contain a
protein that binds to the cotC promoter region (data not
shown). The kinetics of appearance of this binding activity and
its absence from extracts of spoIIID mutant cells suggested
that the protein is SpoIIID. The SpoIIID protein had been
shown previously to inhibit transcription of the cotD gene in
vitro and to bind in the ~35 region of the catD promoter (7. 20).
We hypothesized that SpoIIID might contribute to the difi'er-
ential regulation of co: gene expression we had observed (Figs.
2 and 3). This hypothesis is difficult to test in viva, because
SpoIIID is required for production of UK RNAP (7, 25), which
transcribes the cot genes (4, 5). To test whether SpoIIID afl'ects
cot gene transcription in vitro, we modified the experiment
shown in Fig. 4. Different amounts of SpoIIID were incubated
with a mixture of DNA templates before the addition of o"
RNAP and a fixed amount of GerE. In this set of experiments,
equimolar gerE template was included as a control because we
knew that SpoIIID has very little effect on its transcription
(26). Fig. 5A shows a representative result, and Fig. 6B sum-
marizes quantification of two experiments. SpoIIID repressed
cotC transcription about locfold, with 50% repression occurring
at 0.2 In(. cotX transcription was repressed about 2-fold at the
highest SpoIIID concentration tested (approximately 1 all).
SpoIIID had very little effect on transcription of cotB or gerE.
These results provide a plausible explanation for the lag be-
tween cotX and cotC expression (Figs. 2 and 3). The level of
SpoIIID decreases as active «1" RNAP accumulates in the
mother cell (9. 10). Our in vitro transcription results suggest
thatcotXwouldberelIosedfromSpoIIIDrspr-essionfirst.
followedbycothhentheSpoIlIDconcentrationreachesa
much lower level.

SpoIIIDBindstoSpecificSitainthecotCandcotXPmmoter
Regions—The inhibitory effect of SpoIIID on cotC and cat)!
transcription suggested that SpoIIID might bind to spcific
DNA sequences in the promoter regions of these genes. To
examine specific binding by SpoIIID, we performed DNase I
footprinting experiments. SpoIIID protected two regions of
cotC promoter DNA from digestion with DNase I. The protec-
tion spanned positions ~43 to ~29 and positions ~77 to ~56 on
the transcribed strand (Fig. 6A). On the nontranscribed strand.
protection spanned positions ~40 to ~22 and positions ~75 to
~63 (Fig. 68). Protection was observed at the lowest concen-
tration of SpoIIID tested, indicating that SpoIIID binds with

66

B. subtilis SpoIIID and GerE Regulate Cot Genes

123456

A
...-m

COtC- fig "...
cotB- '.

  

Relative transcript level

 

0" I V W

0 20 40 60
Amount ofSpolllD (pmol)

no.5. WofSpolnDoncotB,cotC.andce¢XtI-anseription
in other. A. DNA templates (0.06 pmol each) were incubated without
(loss I)orwitb4(lon¢2).8(lone3). 15(lone4).30(lone6).or60pmol
(lane 6) of gel-purified SpoIIID and transcribed with partially purified
on RNAP (0.2 rag) and 200 pmol of gel-purified GerE. DNA templates
werethssameasinFig.4pluso480-bosepair8coRl-Puull ofpSC146
(382-bsse gerE transcript). The positions of run-off transcripts of the
expected sixes. as judged from the migration of end-labeled DNA frag-
ments of Hopi-digested pBR322. are indicated. B. transcript Iignals for
cotB (Cl). cotC (O ), cotX (O), and gerE (A) were quantified and normal-
ised to the maximum signal for each transcript Points on the graph are
averogssofnormalisedsignalsfromtwoexperiments,and¢rrorbars
show 1 SD. of the data.

relativelyhighamnitytothesesitesascomparedwithother
SpoIIID binding sites mapped previously (7. 8). Fig. BC shows
the sequence of the cotC promoter in the two regions protected
by SpoIIID. Within each protected region is a sequence that
matches the consensus sequence for SpoIIID binding (Fig. 60).
'lhese results may explain why SpoIIID is a potent repressor of
cotC transcription (Fig. 5). The upstream SpoIIID binding site
centered at position ~67.5 (Fig. 6C) overlaps the critical GerE
site centered at position ~68.5 (fig. 1). The downstream
SpoIIID binding site centered at ~36.5 overlaps the ~35 region
of the cotC promoter. which may be important for recognition
by a" RNAP.

SpoIIID also bound specifically to a site in the cot]! promoter.
Protection from DNase I digestion spanned positions ~27 to
~11 on the transcribed strand (Fig. 7A) and at least positions
~23 to ~15 on the nontranscribed strand (Fig. 78). Fig. 70
shows the sequence of the cotX promoter in the region protected
by SpoIIID. Within this region is a sequence that matches the
consensus sequence for SpoIIID binding in 7 of 9 positions (Fig.
7D). SpoIIID binds with relatively low affinity to this site in the
cotX promoter (Fig. 7, A and B) as compared with the two sites
in the cotC promoter (Fig. 6. A and B). which may explain why
SpoIIID was a less potent repressor of cat)! transcription than
cotC transcription (Fig. 5). '

B. subtilis SpoIIID and GerE Regulate Cot Genes 13853

A 3123456 B 123456

-n ;rrr:r._ i;;§5:;,
. .

-a
~ee—~s .4 |;‘.“

2 “ JIIIII
. Ills. '
4 ' 2 2 e - 9 ' ‘
Iiir I ' .IESQQI
transcribed strand nontranscribed strand
0

.u. memwwmfimm -l6

~41
WAR- Y CONSENSUS

FIG 6. SpoIIID footprints In the cotC promoter region. Rodi 0-
active DNA fragments separately end- labeled on the transcribed (A) or
nontranscribed (B) strand were incubated Inse patera reactions with a
carrier prote in( (bovine serum albumin. 310 pmol) only (lane 1) or with

4(lons 2), 8 (lane 3),15 (lane 4). 30 (lane 5.) or 60 pmol (lane 6) of
gel-purified SpoIIID in additicnto the carrier protein and then sub-
jectedto toDNase I footprinting' In a total volume of 45 ul. Filled boars
indicate the region protected from DNase I digestion by SpoIIID. Ar-
rowheads denote the boundaries of protection and numbers to the let!
refer to positions relative to the transcriptional start site. as dedu
from sequencing ladders gewnerated ith T7 Sequen aseV 2. 0 (Amer-
sham Phannscia Biotech) and the appropriate primer. mmumk: indi-

tethe position of sites rendered yperseneitive to DNase I digestion
by “SpoIIID binding. C, positions of SpoIIID footprints in the cotC
pmmr sate-mi
of the cotC promoter region is shown (4). N ucleotIdss In the- region
that match the consensus for recognition onby (In indicates A or
C) are shown as ace capital letters. Overllning and underlining

Indicate regions on the nontranscn bed and transcribed strands, respec-
tively, protected by Spo IIID from DNase I digestion. The dashed lines
indicate regions ofuncertain protection due to a lack ofDNase I diges-
tion in these regions. Numbers refer to positions relative to the tran-
scriptional start site D, nucleotide sequences within the SpoIIID-pro-
tected regions of the cotC promoter are aligned“: with ulthe consensus
sequence for SpoIIID binding. theco
shownas ital lcuc mend numbers refer to positions ulrelative to the
transcriptional start site

DISCUSSION

Our results strongly support the idea that the combined
action of GerE and SpoIIID produces differential patterns of cot
gene expression during B. subtilis sporulation. Previously, cotB
and cotC had been thought to be coordinately regulated by the
appearance of GerE (4, 19). However, expression of a cotB-Incl
fusion begins to increase at least 1 h earlier than expression of
a cotC-lacZ fusion (Fig. 2), and cotB mRNA reaches its maxi-
mum level at least 2 h earlier than cotC mRNA (Fig. 3). The
earlier expression of cotB during sporulation may result. in
from a lower threshold for activation by GerE (Fig. 4). but
in addition, SpoIIID was shown to be a potent repressor ofcotC
transcription (Fig. 5). The repressive effect of SpoIIID on cotC
transcription in vitro appears to be due to the presence of two
relatively high affinity SpoIIID binding sites in the cotC pro-
moter region that overlap binding sites for GerE and 0K RNAP
(Fig. 6). Therefore. we propose that SpoIIID represses cotC
transcription during sporulation. contributing to the observed

lag between cotB and cotC expression.

A 12345 B
3vIg-I
-27 d.--

n::'. 2:].-4' *
5.12338 -1 1'3""

transcribed strand
strand

-35 agtcaaaatsscaggctcgctcétltaat .7

-l2 ATGAgCgAGC -2|
HWRRACAR-Y CONSENSUS
FIG 7. SpoIIID fiootprints n the cotX promoter region. Radio-
active DNA fragments separately end~labeled on the transcribed (A) or
notranscribed(B trsndwero Incu inse ratereactionswitha
carrier protein (bovine serum rumalbumin. 310 pmol) only (lane 1) or with
5 “pmol (lane 2).10 pmol (lane 3). 20 pmol (lane 4), 40 pmol (lane 5), or

and then subjected toD Nase I footprinting' In a total volume of 45" It].

See the Fig. 6 legend for explanation of led

her: to the left, and sets ks. C, position of the SpoIIID footprint' In "the
promoter region. The nucleotide sequence of the non transcribed

strand of the cotX promoter region is shown (39). Overliru'ng and un-

derlining indicate regions on the nontranscribed

strands. respectively. protected by SpoIIID from DNase I digestion. The

dashed lines indicate regions of uncertaI nprotection due to a lack of

DNase I digestion in these Numbers refer to positions relative

WW
consensus sequence for SpoIIID binding. mMatches to the consensus
sequence are shown asca capital letters, and numbers refer to positions
relative to the transcriptional start site

Consideration of our results with the cotX promoter provides
additional support for the proposal that SpoIIID delays full
expression ofcotC during sporulation. The pattern of cotX-lacz
expression and cotX mRNA accumulation was more similar to
that of cotB than cotC (Figs. 2 and 3), yet cotX and cotC
transcription in vitro exhibited similar dependence on the con-
centration of GerE (Fig. 4), providing no explanation for the
observed differential expression of cotX and cotC in viva. This
difference can be plausibly explained by our finding that
SpoIIID is a more potent repressor of cotC transcription in vitro
than of cotX (Fig. 5). SpoIIID appears to be a weak repressor of
cotX because it binds with relatively low affinity to a site in the
promoter that overlaps the binding site for UK RNAP (Fig. 7). If
SpoIIID does repress transcription from the cotX promoter
during sporulation. this repression would be tobe
relieved earlier than repression of cotC as the level of SpoIIID
decreases In the mother cell (9,10).

Differential timing of cotB and cotC expression was over-
looked previously due to hybridization of a primer that was
thought to be cotC-specific with cgcAB mRNA (4, 19, 23).
Hence, the primer extension analysis reported previously
shows that cotB and cgeAB transcripts appear with similar

timing during sporulation (19) and does not conflict with our
finding that cotC expression lags behind that of cotB (Figs. 2
and 3). Expression of a cotCJacZ fusion was shown previously
to be induced about 1 h later during sporulation than expres-
sion of cotD-lacZ (27). Interestingly, the difference in time of

67

Fh~

13354 B. subtilis SpoIIID and GerE Regulate. Cot Genes

  

sly! spomo/ sigK gerE cotBa cotC

Pro. 8. A model showing how the combined action of SpoIIID
and Our! may regulate cot genes in the context of interactions
between mother cell-specific transcription factors. Dashed or-
roses show gene (italicisedHo-product (proteins are circled) relation~
ships. Solid arrows represent positive regulation of transcription. Lines
with s boned end represent negative regulation of transcription. The
bardistinguishescotDandcotX.whichareproposedtobeweakly
repressed by SpoIIID, from cotB and cgeAB, which are not repressed by
SpoIIID. and from cotC, which is expressed later because it is strongly
repressedbySpolllDfindicstedbythethicklimwithaborredsnd).

   

 

induction disappeamd when the genes were artificially induced
by production of a“ during growth (27). Under these conditions,
SpoIIID would not be present. Therefore, we propose that
SpoIIID is responsible for the observed delay during sporula-
tion in cotC expression as compared with that ofcotD. A pre-
diction of this hypothesis is that SpoIIID is a more potent
repressor of cotC transcription than of catD transcription.
SpoIIID was shown previously to bind with relatively high
amnity to a site spanning the -35 region of the cotD promoter
and repress transcription in vitro (7); however. the efl'ect of
SpoIIID on corD and cotC transcription in vitro has not been
compared directly.

Fig. 8 illustrates how the combined action of SpoIIID and
GerE may produce differential regulation of co: gene transcrip-
tion in the context of known regulatory interactions with the
two mother cell-specific 0 factors, rr' and o“. «rt RNAP tren-
scribes the spoIIID gene (28-32). As SpoIIID accumulates, it
activates transcription of sigK by a” RNAP (7. 25). The primary
product of the sigK gene. pro—0" (not shown in Fig. 8), is
processed to a“ in an activation step coupled to a signal from
the forespore (33—35). a" RNAP transcribes gerE (4, 24). As
GerE and a" RNAP accumulate, cotB (4, 19), cgeAB (23), and
other genes begin to be transcribed. The other genes include
cod) (4, 6, 7, 20) and cotX (5. 36). but we propose that SpoIIID
limits transcription of these genes (boxed in Fig. 8) and pre-
vents transcription of cotC for about 1 h. Repression by SpoIIID
is relieved as its level declines due to degradation of the protein
and due to a negative feedback loop initiated by a" RNAP that
inhibits transcription of sigE and other early sporulation
genes, thus inhibiting further production of 0" and SpoIIID (9,
10, 37). The falling levels of as and SpoIIID and the rising
levels of o" and GerE together with the fact that both SpoIIID
and GerE can act as activators or repressors of transcription (4,
6—8) make it possible to regulate the timing and level of indi-
vidual cot gene transcription in a variety of ways.

Our 5' deletion analysis of cot promoters gives further in-
sight into the function of GerE as an activator of transcription.
Deletions designed to eliminate the more upstream GerE bind-
ing site in the 00:8 and cotX promoters greatly reduced expres-
sion of lacZ fusions (Fig. 2), strongly suggesting that GerE
binds to sites centered at -73.5 and -60.5 in the cotB and cat)!
promoters. respectively, and contributes to transcriptional ac-
tivation of these promoters during sporulation. On the other
hand, the finding that elimination of the more upstream GerE
sites in these promoters did not abolish GerE-dependent ex-
pressionsuggeststhatthemoredownstreamGerEsitesare

suficient for weak transcriptional activation. The more down.
stream GerE site in the cotB promoter has the sequence 6'-
AATI'AGGCTATT-3' (4). which matches perfectly the consen-
sus sequence for GerE binding (5). ibis sequence is centered at
-47.5 (4), which seems to be a preferred position for binding in
promoters activated by GerE, since the cotVWX, cotYZ, and
cotD promoters also have a sequence matching the consensus
centered at -47.5 or -46.5, to which GerE appears to bind,
activating transcription (5, 6). The more downstream site in the
cotX promoter has the sequence 5'-GAC'I‘GAGTCATA-3',
which matches in 7 of 10 positions in the consensus sequence
for GerE binding (5). This sequence is centered at -37.6 and is
in the opposite orientation relative to the direction of cell
transcription as compared with the site centered at -4"l.5 in
the cotB promoter. Assuming that GerE binds in a particular
orientation to sequences similar to its nonpalindromic consen-
sus sequence, our results suggest that GerE can activate tran-
scription when bound in opposite orientations to sites centered
at -47.5 and —37.5 (figs. 1 and 2). Our results also suggest
that GerE can activate transcription when bound to a site
centered as far upstream as -73.5 in the cotB promoter or
one-half turn of the DNA helix downstream at -68.5 in the
cotC promoter. Hence, GerE may be less stringent than, for
example. the E. coli catabolite gene activator protein with
respecttothepositionfromwhichitcanactivatetranscriptien
(38).Thisideacanbemstedfurtherbycreatingsinglebasepair
changes that eliminate GerE binding to individual sites and/or
bysystematicallyvaryingthepositionofaGerEbindingsitein
a promoter region.

The 5' deletion we created that eliminates the more up-
streamGerE bindingsiteinthecotXpromotermayprovetobe
useful for investigating the mechanism of GerE transcriptional
activation at this promoter. A recent study suggests that GerE
may interact with a" at the cotX promoter and facilitate the
initial binding of 0" RNAP to the promoter (36). Certain amino
acid substitutions in a“ reduced expression of a cothlocZ fu-
sion but not expression of a gerE-lacZ fusion, which also de-
pends upon 0" RNAP but not on GerE. The authors speculated
that GerE bound to the more downstream site centered at
-37.5 contacts 0", enhancing binding of a" RNAP to the cotX
promoter. To explain the observation that the substitutions in
0" did not eliminate cotX-tool expression, the authors pro-
posed that GerE bound to the more upstream site centered at
-60.5 makes a different contact with o" RNAP (eg. with the
C-terminal domain of the a subunit). This model predicts that
expression of the cotX-Incl fusion we made, lacking the GerE
site centered at -60.5, would be abolished in the mutants with
aminoacidsubstitufiomintheJregionthoughttointeract
with GerE.

OurresultsshedmorelightonhowSpoIIIDcanfunctionas
a transcn'ptional repressor. SpoIIID footprints in the bofA,
cotD, and spoVD promoter regions have been published previ-
ously (7, 8). In each case, SpoIIID binds to sites centered at + 1
and/or -36, presumably preventing RNAP from binding to the
promoter or hindering a subsequent step in transcription ini-
tiation. In the cotC promoter region, SpoIIID binds to a site
centered at -67.5 (Fig. 6). which presumably prevents GerE
from binding to its site centered at -68.6 (4), and SpoIIID
binds to a site centered at -36.5 (Fig. 6), which presumably
interferes with RNAP binding or a subsequent step in initia-
tion. Likewise. SpoIIID binding to a site contend at -16.5 in
the cotX promoter (Fig. 7) probably interferes with RNAP
function.

Whyarecertaincotgenessubjecttodualregulationby
SpoIIID and GerE? One possibility is that fine-tuning of cat
gene expression allows optimal levels of Cot proteins to be

68

-_ 1____4_

B. subtilis SpoIIID and GerE Regulate Cot Genes

synthesized at the proper times for assembly into the spore
cost. Our results suggest that expression of cotC is delayed
relative to expression of other cot genes. We plan to test
whether the delay in cotC expression is important by engineer-
ing cells to produce CotC earlier and measuring spore resist-
ance properties. Fine-tuning of cat gene expression may also
allow the spore coat to be suitably tailored in response to
environmental conditions. Expression of a cotC-lacZ transla-
tional fusion was shown to depend strongly on whether sporu-
lation was induced by sudden or gradual nutritional shift-down
(19). Whether this regulation involves one of the previously
known cotC transcription factors (Le. GerE or a" RNAP), the
newly discovered cotC repressor reported here (is. SpoIIID). or
some other mechanism remains to be elucidated.

‘L [J a

_, We thank R Losick, P. Zuber. A. Aronson. J.
Errington, and C. Moran for providing bacterial strains, phages. and
plasmids.

REFERENCES

1. Stragier, P., and lesick. R. (1996) Annu. Rev. Genet. 30. 297-341

2. Kroos. L, Zhang. 8.. Ichikawa. K., and Yu. Y.-'l'. N. (1999) Mol. Microbial. 81.
1235-1294

3. Losick. R.. and Stragier. P. (1992) Nature 856, 601-604

4. Zheng. L. Halberg. R.. Roels. 8.. lchikswa, R.. Kroos. L. and lesick. R. (1992)

J. Mol. Biol. 226. 1037-1050

5. Zhang. J.. lchikswa. 11.. Halberg, ll. Kroos. L. and Aronson. A. l. (1994) J.

Mol. Biol. 240. 405-415

6. Ichikawa. H.. Halberg, R.. and Kroos. L (1999) J. Biol. Chem. 274. 8322-8327

7. Halberg. R. and Kroos. L (1994) J. Mol. Biol. 243. 425-436

8. Zhang. 8.. Daniel. R.. Errington. J.. and Kroos. L (1997) J. Bacteriol. 179,
972-975

9. Zhang. 8.. and Kroos, L (1997) J. Bacterial. 179. 6138-6144

0. Halberg. R.. and Kroos. L. (1992) J. Mol. Biol. 236. 840-849

1. Kenney. T. J.. and Moran. C. P. (1991) J. Bacterial. 173. 3282—3290

1
1

13855

12. Zubar, P., and Losick. R. (1987) J. Bacteriol. I“. 2223-2230

13.
14.

8 3 fig

69

$3.83 8.3.3 8.13833

Errington. J.. and Mandelstam. J. (1986) J. Gen. Microbiol. 132. 2967-2976
Harwood. C. R.. and Cutting. S. M. (1990) Molecular Biological Methods for
Bacillus. John Wiley & Sons. Inc.. Chrchester. England

. Miller. J. (1972) Experiments in Molecular Genetics, pp. 352455, Cold Spring

Harbor Laboratory. Cold Spring Harbor. NY

. lgo, M. M.. and Losick. R (1986) J. Mol. Biol. 191, 615-624
. l-‘isseha. M. Gloudemans. K., Gill. R, and Kroos. L. (1996) J. Bacterial. 178.

2539 -2550

. Brandner. J.. and Kroos, L. (1998) J. Bacteriol. 1”; 1995-2004

Zhang. L., and Losick. R. (1990) J. Mol. Biol. 312. 645—660
Kroos. L, Kunkel. 8.. and Losick. R. (1989) Science 943. 526-529

. Carter, 1!. L. and Moran, C. P. (1986) Proc. Natl. Acad. Sci. U. 8.4 It.

9438-9442

Roels. S.. Driks. A., and Losick. R (1992) J. Bacteriol. 174. 575-585

Roels. S., and Losick. R. (1995) J. Bacterial. 177. 6263-6275

Cutting. S.. Panser, S.. and Losick, R (1989) J. Mol. Biol. 201. 393-404

Kunkel. 8.. Sandman. K., Panzer. 8.. Youngman. P., and Losick. R. (1988) J.
Bacteriol. 170. 3513—3522

Halberg. R.. Oke. V.. and Kroos. L (1995) J. Bacterial. 177. 1888-1891

Oke. V.. and lesick, R. (1993) J. Bacteriol. 175. 7341-7347

Kunkel. 8.. Kroos. L. Path, 11.. Youngman. P.. and Losick, R. (1989) Gauss
Dev. 8, 1735-1744

Stevens. C. M.. and Errington. J. (1990) Mol. Microbiol. 4, 543-552

Tstti, K. IL. Jones. C. R.. and Moran. C. P. (1991)J. Bacteriol 178. 7828-7833

Jones. C. R.. and Moran. C. P. (1992) Proc. Natl. Acad. Sci. U. 8.4. C.
1968-1962

Jones. C. R.. Tstti. K. K., and Moran. C. P. (1992)J. Bacteriol. 174. 6815—6621

Cutting. 8.. Oke. V.. Driks. A., losick. R.. Lu. 8.. and Kroos, L (1990) Cell 62.
239-250

Cutting. 8.. Driks. A., Schmidt. 8.. Kunkel. 3.. and lasick. I. (1991) Genes
Dev. 5, 456-466

Lu. 8.. Halberg. R.. and Kroos. L (1990) Proc. Natl. Acad. Sci. U. 8.4 I7.
9722-9726

Wade, IL H.. Schyns. 0., Opdyke. J. A., and Moran. C. P. (1999) J. Bacteriol.
181. 4365-4373

. Zhang, 8.. Struffi, P., and Kroos. L. (1999) J. Bacterial. 181, 4431-4088

Zhou. Y., Martel. T. J.. and Ebright. R. H. (1994) J. Mol. Biol. 248. 603-610

. Zhang. J.. l-‘its‘lames. P. C.. and Aronson. A. l. (1993) J. Bacteriol. 178,

3757-3766

CHAPTER VI

Conclusions

70

The goal of my research was to understand the role of GerE and SpoIIID in gene
regulation during Bacillus subtilis sporulation. Experiments were performed to
characterize GerE as a DNA-binding protein, to determine the regulatory role of GerE in
the a cascade during sporulation, and to study coordinate gene regulation by GerE and
SpoIIID.

Genetic studies and the primary structure of GerE suggested that it is a small
DNA-binding protein. This was tested directly by gel-purifying GerE from an E. coli
strain that was engineered to overcxpress gerE. GerE bound to the cotB and cotC
promoter regions in a sequence-specific manner and activated transcription in vitro. GerE
also inhibited transcription of sigK and COM in vitro. The ability of GerE to activate
transcription was further demonstrated in transcription of the cotVWXYZ gene cluster.
GerE bound to and activated transcription from specific sequences in the cot VWX, cotX,
and cotYZ promoters. The GerE binding site consensus sequence was established as
RWWTRGGY--YY( R means purine, W means A or T, and Y means pyrimidine).

Many transcriptional activators that bind upstream of the -35 region of promoters
are unable to activate transcription by RNA polymerase that lacks the C-tcrminal domain
of the a subunit. GerE binding sites are located upstream of the -35 region of the cotB,
cotC, and cotX promoters, but a SpoIIID binding site is located within the -35 region of
the sigK promoter. In vitro transcription experiments with heterologous RNA
polymerases consisting of B. subtilis 0K and E. coli core RNA polymerase with or
without the C-terminal domain of the or subunit suggested that GerE interact with aCTD
to activate transcription of cotB, cotC, and cotX, but SpoIIID does not use this mechanism
to activate sigK transcription (Appendix A).

GerE inhibited sigK transcription in vitro; however, when the sigK promoter

71

 

region was fused to the E. coli IacZ gene, there was no indication that GerE inhibited
sigK-lacZ expression in vivo. This paradox was solved by mapping a GerE binding site
near the sigK transcriptional start site, because the result showed that the GerE binding
site was not included in the sigK-lacZ construct. A new sigK-lacZ fiJsion that included
the GerE binding site exhibited B-galactosidase activity during sporulation that was two
fold higher in gerE mutant cells than in wild-type cells. This showed that oK RNA
polymerase transcription of gerE initiates a negative feedback loop in which GerE
represses sigK transcription. Another example of repression by GerE was demonstrated |

with the cotD gene. Two GerE binding sites centered at -47.5 and -25.5 were mapped in

 

the cotD promoter by DNase I footprinting. In vitro transcription assays indicated that
GerE activates cotD transcription when present at a low concentration, but represses
transcription by binding the site centered at -25.5 when present at high concentration.
The existence of these negative effects by GerE clearly indicates that gene expression at
the later stages of sporulation is still tightly regulated.

The relationship between GerE activation of transcription and promoter structure
was investigated by performing deletion studies. There are two GerE binding sites in each
of several promoters, including cotB, cotC, and cotX. GerE binding sites in the cotB
promoter are centered at —73.5 and -47.5. A GerE binding site centered at or very near -
47.5 is also found in the catD, cotYZ, and cat VWX promoters. Upon deletion of the GerE
binding site centered at -73.5 in the cotB promoter, activity was reduced by
approximately 80 %, demonstrating that the site centered at -47.5 is sufficient for some
activation by GerE, but that the site centered at -73.5 contributes greatly to activation.
GerE binding sites in the cotX promoter region are centered at -60.5 and -36.5. The GerE

binding site centered at -36.5 is interesting because its orientation is the opposite of the

72

others. GerE was able to activate transcription from the cotX promoter containing only
the GerE binding site centered at -3 6.5 about 40 % as well as from the native promoter,
suggesting that GerE can activate transcription when bound to the promoter in either
orientation. There is evidence that GerE activates cotX transcription by two different
mechanisms, one acting through the C-terminal domain of the a subunit of RNA
polymerase (aCTD) and the other acting through 0". Studies of E. coli DNA-binding
proteins suggest that binding upstream of the promoter -35 region may activate
transcription through interaction with aCTD, whereas binding within the —35 region may
activate through interaction with the a subunit. Therefore, it is plausible that GerE bound
to the sites centered at -60.5 and -3 6.5 may activate cotX transcription through interaction
with the «cm and 0‘, respectively. Deletion of a GerE binding site centered at -134.5
did not affect cotC promoter activity, indicating that GerE binding to a site centered at -
68.5 is sufficient to activate cotC transcription. ‘

In the preceding study, there appeared to be differences in the pattern of lacZ
expression from fiision to the cotB, cotC and cotX promoters during sporulation. Indeed
primer extension analyses showed that cotB mRNA accumulates slightly earlier during
sporulation than that of cotX or cotC, whereas the cotC mRNA level reaches its
maximum later than that of cotB or cotX. A possible explanation for these differences
was sought using in vitro transcription assays. An equimolar mixture of cotB, cotC, and
cotX DNA templates was transcribed by (JK RNA plymerase in the presence of different
amounts of GerE. A smaller amount of GerE activated cotB transcription than for cotC or
cotX. Likewise, the effect of SpoIIID on transcription of cotB, cotC, and cotX was tested
in mixed-template in vitro transcription assays with a fixed amount of GerE and different

amounts of SpoIIID. A very small amount of SpoIIID strongly repressed cotC

73

 

transcription, a larger amount of SpoIIID moderately repressed cotX transcription, and
SpoIIID had little or no effect on cotB transcription. Taken together, these results suggest
the following model for the differences in expression patterns of cotB, cotC, and cotX.

As oK RNA polymerase becomes active, gerE transcription begins and the level of
SpoIIID starts to decrease. GerE accumulates to a low level and activates cotB
transcription. For cotX and cotC to be transcribed, GerE must accumulate to a higher
level. As the level of SpoIIID decreases, its repressive effect on cotX is lost, and finally
as its level becomes very low, cotC is fully transcribed.

An important contribution of my thesis research is the finding that regulation of
both the sigK gene encoding 0K and cot genes in the 0K regulon are tightly controlled by
GerE and SpoIIID late in sporulation. Tight regulation of cat gene expression may be
important for proper assembly of the spore coat or for modifying coat composition in
response to environmental cues. These hypotheses can be the basis for fiirther research

on cot gene regulation.

74

APPENDIX A

Requirement of the C-Terminal Region of the a Subunit of RNA Polymerase for

Transcriptional Activation by GerE but not SpoIIID of Bacillus subtilis

 

ABSTRACT

GerE and SpoIIID activate transcription by 0" RNA polymerase during Bacillus
subtilis sporulation. GerE activated transcription in vitro of cotB, cotC, and cotX by
heterologous RNA polymerase reconstituted from E. coli core subunits and B. subtilis 0".
Likewise, SpoIIID activated transcription of sigKby the heterologous enzyme. When E.
coli core containing a with a C-terminal truncation was used in similar experiments,
activation by GerE at the three cot promoters decreased by 30 to 50%, but activation by
SpoIIID at the sigK promoter remained the same. These results suggest that GerE
interacts with the C-terminal domain of the a subunit of RNA polymerase; however,
GerE also activates cot gene transcription by another mechanism, and SpoIIID activation

of sigK transcription depends completely on another mechanism.

76

INTRODUCTION

Activation of transcription is the primary regulatory strategy used during
prokaryotic developmental processes. Sporulation of the gram-positive bacterium
Bacillus subtilis is controlled by a cascade of RNA polymerase (RNAP) 0 subunits
(Kroos et al. 1999). In addition, sequence-specific DNA-binding proteins activate
transcription by RNAP containing sporulation-specific 0 factors. For example, GerE and
SpoIIID activate transcription by UK RNAP in the mother-cell compartment of
sporulating B. subtilis (Halberg and Kroos 1994; Zheng et al. 1992; Zhang et al. 1994).
Here, we investigate the mechanism of transcriptional activation by GerE and SpoIIID.

GerE recognizes the consensus DNA sequence RWWTRGGY-YY (R means
purine, W means A or T, and Y means pyrimidine) and activates transcription of many
genes whose products are spore coat proteins (Zheng et al. 1992; Zhang et al. 1994).
Here, we focus on GerE-activated transcription from the cotB, cotC, and cotX promoters.
There are two GerE binding sites in each of these promoters (Zheng et al. 1992; Zhang et
al. 1994). Deletion of the upstream site centered at -73.5 bp in the cotB promoter
resulted in a partial reduction of cotB-lacZ expression, suggesting that both this site and
one centered at -47.5 bp participate in transcriptional activation (Ichikawa and Kroos
2000). In contrast, deletion of the upstream GerE binding site centered at -134.5 bp in the
cotC promoter did not affect cotC-lacZ expression, suggesting that GerE binding to a site
centered at -68.5 bp is sufficient for full activation of cotC transcription (Ichikawa and
Kroos 2000). In the cotX promoter, the upstream site centered at -60.5 bp is in the same
orientation as the GerE binding sites mentioned above, but the downstream site centered
at -36.5 bp is in the opposite orientation. The downstream site alone is sufficient for

approximately 40% as much cotX-lacZ expression as when both sites are present

77

(Ichikawa and Kroos 2000).

SpoIIID activates transcription of the sigK gene (encoding a“) by 05- or 0"-
containing RNAP (Kroos et al. 1989). There is a match to the consensus sequence for
SpoIIID binding, WWRRACAR-Y, centered at -27.5 bp in the sigK promoter, and
SpoIIID has been shown previously to bind to this site (Halberg and Kroos 1994).

All four core subunits of bacterial RNAP can be targets for physical interaction
with transcriptional activator proteins (Ebright 1985; Niu et al. 1996; Miller et al. 1997;
Lee and Hoover 1995). The C-tenninal domain of the a subunit (aCTD) is most often
the target for activators that bind upstream of the promoter -35 region. Many activators
that bind upstream of the promoter -35 region fail to activate transcription when aCTD is
absent. Other activators do not require ocCTD for transcriptional activation. In many
cases, these proteins bind within the promoter -35 region, and interact with 0 or the N-
terminal domain of a (aNTD) (Niu et a1. 1996).

Because GerE binds upstream of the cotB, cotC, and cotX promoter -35 regions,
we hypothesized that GerE might interact with aCTD to activate transcription of these
genes. Conversely, since SpoIIID binds within the sigK promoter -35 region, aCTD
might be dispensable for activated transcription of sigK. The primary structures of
RNAP a and B subunits of B. subtilis and E. coli have high similarity (N iu et al. 1996;
Boor et al. 1995). Here, we show that heterologous RNAP reconstituted from E. coli
core and B. subtilis oK is responsive to transcriptional activation by GerE and SpoIIID.
We used this system to demonstrate that GerE activation of cotB, cotC, and cotX

transcription depends, in part, on aCTD, but SpoIIID activation of sigK is independent of

aCTD.

78

MATERIALS AND METHODS

DNA templates and proteins

To facilitate preparation of template DNA for in vitro transcription, pHI-ll was
constructed by subcloning a HindIII-EcoRI fragment of pHI-3 between the HindIII and
EcoRI sites of pUC 18. pHI-3 was constructed by subcloning a fragment of the cotB
promoter region, -86 to +37, franked with a EcoRI and HindIII site at the upstream and
downstream end of pTKlac (Kenney and Moran 1991), respectively. The DNA fragment
was prepared by PCR using pBD136 (Zheng et al. 1992) as the template. The upstream
primer was S’GGGAATTCGCGTGAAAATGGGTATB’ (the underlined portion
corresponds to an EcoRI site). The downstream primer was
5’GCGAAGQTTAATTCCTCCTAGTCA3’ (the underlined portion corresponds to a
HindIII site). pHI-12 was constructed by subcloning an EcoRI-HindIII fragment of pHI-
7 between the EcoRI and HindIII sites of pUCl9 (Y anish-Perron et al. 1985). pHI-7 was
constructed by subcloning a fragment of the cotC promoter region, -81 to +14, franked
with EcoRI and HindIII restriction sites at the upstream and downstream ends,
respectively, of pTKlac. The DNA fragment was prepared by PCR using (Zheng et al.
1992) as the template. The upstream primer was
5’CGGAATTCTCTATCATI‘TGGACAG3’ (the underlined portion corresponds to an
EcoRI site). The downstream primer was 5’CGGAAGC'I'I'I’I‘ATI'I‘T'I‘ACTACG3’ (the
underlined portion corresponds to a HindIII site). The DNA template containing the cotX
promoter region from -98 to +82 was prepared by PCR using pJZZZ as the template. The
upstream and downstream primers were 5’CGGAAAAACGATAACAATI‘AGC3’ and
5’CGGTI‘TGCATCAGAACATGT3’, respectively. 0" and SpoIIID were gel purified as

described previously (Kroos et al. 1989). GerE was gel purified as described previously

79

(Zheng et al. 1992). E. coli core RNA polymerase was prepared as described previously
(F ujita and Ishihama 1996).
In vitro transcription assay

E. coli core RNA polymerase (0.1 pmol) with or without intact aCTD was
incubated with oK(1 pmol ) on ice for 10 min. to reconstitute holoenzyme. Transcription
reactions (45111) were performed as described previously (Carter and Moran 1986), except
that RNA polymerase was allowed to bind to the DNA template for 10 min. at 37°C
before the addition of nucleotides (the labeled nucleotide was [a-32P]CTP). Heparin
(6ug) was added 2 min. afier the addition of nucleotides to prevent reinitiation. After
reactions were stopped, 20m of each reaction mixture was subjected to electrophoresis in
a 5% polyacrylamide gel containing 8 M urea, and transcripts were detected by
autoradiography. The signal intensities were quantified using a Storm 820 Phosphor

Imager (Molecular Dynamics).

8O

RESULTS

aCTD plays a role in GerE activation of cotB, cotC, and cotX transcription.

To examine the requirement for aCTD in activation of cat gene transcription by
GerE, we reconstituted heterologous RNAPs consisting of B subtilis 0K and E. coli core
containing full-length a (Ecwtoc oK RNAP) or a truncated after amino acid 257
(EcaaCTD oK RNAP). As a control to measure the relative activities of the reconstituted
RNAPs, each preparation was used to transcribe gerE, a gene with a strong oK-dependent
promoter that can be transcribed in the absence of GerE (Zheng et al. 1992). In the
experiment shown in Figure l, EcAaCTD oK RNAP was 68% as active as Ecwta oK
RNAP at transcribing gerE. Since the gerE promoter has no apparent sequence (UP
element) with which aCTD would be expected to interact directly, we attributed the
difference in gerE transcriptional activity to different efficiencies of reconstituting active
holoenzyme. In the absence of GerE, cot gene transcription by the heterologous RNAPs
was so weak that we could not quantitate the signals (Figure 1). In the presence of GerE,
the signals were markedly stronger for Ecwta oK RNAP than for EcAaCTD oK RNAP.
These signals were quantified. The signals obtained with Ecwta oK RNAP were
multiplied by 0.68, to take into account the higher activity of this preparation of
reconstituted RNAP compared with EcAocCTD oK RNAP, as judged by gerE
transcription. Comparing these adjusted Ecwtoc oK RN AP signals with the EcAaCTD oK
RNAP signals, it appeared that the absence of aCTD reduced GerE-activated cot gene
transcription by 20-50%. The experiment shown in Figure 1 was repeated with different
preparations of reconstituted heterologous RNAPs, and similar results were obtained.

Figure 2 shows that the absence of aCTD reduced cotB, cotC, and cotX transcription, on

81

average by 45%, 29%, and 37%, respectively. These results suggest that GerE interacts
with the aCTD of heterologous RN AP, increasing cot gene transcription. GerE
activation of cot genes also involves another mechanism, because EcAaCTD oK RNAP
was stimulated by GerE, although we could not measure the fold activation because the
signals in the absence of GerE were too weak to quantitate (data not shown).
aCTD is not required for activation of sigK transcription by SpoIIID.

The method described above was used to examine the requirement of aCTD for

SpoIIID activation of sigK transcription. Figure 3 shows the results of one experiment.

 

In this experiment, the reconstituted EcAozCTD oK RNAP preparation was 43% as active
as Ecwta 0K RNAP at transcribing gerE. In the absence of SpoIIID, sigK transcription by
the heterologous RNAPs was barely detectable (Figure 3). In the presence of GerE, the
signal was stronger for Ecwta 0K RNAP than for EcAa (JK RNAP; however, this
difference could be completely accounted for by the higher activity of Ecwta 0K RNAP,
as judged by gerE transcription. Similar results were obtained when the experiment was
repeated with different preparations of reconstituted heterologous RNAPs, as summarized
in Figure 2. Thus, the absence of aCTD appeared to have no effect on the ability of

SpoIIID to activate sigK transcription.

82

DISCUSSION
Our results suggest that aCTD is involved in transcriptional activation by GerE
when the cotB, cotC, and cotX genes are transcribed by heterologous RNAP consisting of
E. coli core subunits and B. subtilis 0". These cot genes have GerE binding sites located
upstream of the promoter -35 regions, which indicated they might be analogous to class I
bacterial promoters. One of the most studied class I promoters is the E. coli lac promoter
which has 3 CAP (catabolite gene activator protein) binding site centered at -6l.5.

Genetic and biochemical evidence support the idea that transcriptional activation by CAP

 

bound at this site involves direct physical interaction with aCTD, increasing the affinity L
of RNAP for the promoter (Dove and Hochschild 1998). When this CAP site was moved
to different upstream positions, CAP activated transcription from sites centered at -72.5, -
82.5, and -92.5, suggesting flexibility in the distance between RNAP and CAP, but
rigidity with respect to the face of the DNA helix (Zhou et al. 1994). Each of the three
cot promoters we tested has at least one GerE binding site expected to position GerE on
the same helix face with respect to RNAP (i. e., GerE sites centered at -68.5, -47.5, and -
36.5 bp in the cotC, cotB, and cotX promoters, respectively). However, GerE bound at
the sites centered at -47.5 bp and -36.5 bp in the cotB and cotX promoters, respectively,
might be too close to the transcriptional start site to interact with aCTD. GerE bound at
the other sites in the cotB (centered at -73.5 bp) and cotX (centered at -60.5 bp) promoters
would not be on the same face of the DNA helix as GerE bound at the site centered at -
68.5 bp in the cotC promoter. These considerations suggest that GerE may be more
flexible than CAP in terms of distance and/or helix face requirements for interaction with

aCTD. We cannot rule out the possibility that a direct interaction between aCTD and

83

promoter DNA contributed to the stronger cot gene transcription with Ecwta oK RNAP
than EcAaCTD oK RN AP in the presence of GerE, because we could not reliably
measure the very low levels of basal transcription of the cot genes by the reconstituted

heterologous RNAPs in the absence of GerE. Like the gerE promoter that we used to

gauge the activity of the reconstituted RNAP preparations, the cot promoters we tested
have no sequence (UP element) expected to interact directly with aCTD. Therefore, it is
likely that Ecwta oK RNAP transcribed the cot genes more strongly than EcAa 0" RNAP
because GerE interacts with aCTD. Further experiments are needed to test this model.
Regulatory protein p4 of Bacillus subtilis phage (D29 activates transcription from the viral
late A3 promoter by interacting with aCTD (Monsalve et al. 1996). Interestingly, p4 was
incapable of interacting with E. coli aCTD to activate transcription. Our results suggest
that GerE can interact with E. coli aCTD at cot gene promoters, but it is possible that
GerE interacts more strongly with B. subtilis aCTD at these promoters.

GerE and SpoIIID activated transcription by the heterologous RNAP by one or
more mechanisms that do not involve aCTD. Deletion of «cm still permitted
considerable GerE activation of cat gene transcription, and it did not diminish SpoIIID
activation of sigK transcription.

The sigK promoter, in which the SpoIIID binding site is centered at -27.5 bp, may
be a class II bacterial promoter. The E. coli galPI promoter is a well-studied class 11
promoter with a CAP binding site centered at -41.5 bp. With CAP positioned so close to
the promoter -35 region, aCTD is thought to be located upstream of CAP in the DNA-
protein complex. Activation involves direct interaction between CAP and a surface on

aNTD (N in et al. 1996). Another example of a transcriptional activator that does not

84

require aCTD is bacteriophage ch protein. cI binds to a site centered at -42 bp in the
APRM promoter and activates transcription by interacting directly with the 0 subunit of

E. coli RNAP (Li et al. 1994). SpoIIID bound at the sigK promoter, and GerE bound to
the downstream site in the cotB or cotX promoter, might interact directly with aNTD or
UK to activate transcription. Recently, it was demonstrated that cotX transcription
activation by GerE was reduced with amino acid substitutions in 0", suggesting a direct
interaction between GerE and 0K. GerE bound to the site centered at -68.5 bp in the cotC
promoter may be too far upstream to make these contacts, yet GerE activated
transcription of this promoter by heterologous RNAP lacking aCTD, so perhaps yet
another mechanism is employed.

In summary, we have shown that heterologous RNAP reconstituted from E. coli
core subunits and B. subtilis 0" can be used to investigate the mechanisms of
transcriptional activation by GerE and SpoIIID. aCTD appears to play a role in GerE
activation of cat genes, but not in SpoIIID activation of sigK. Thus, one mechanism used
by many bacterial activators can, in part, account for activation by GerE, but much more
work will be required to determine whether the other mechanisms employed by GerE and

SpoIIID follow other paradigms or are novel.

85

Figure 1

In vitro transcription of cotB, cotC, cotX and gerE by heterologous RNA polymerase
with or without intact aCTD.

DNA template (0.1 pmole) was transcribed with 0.1 pmole each of reconstituted Ecwta
oK RNAP (lanes 1 and 2) or EcAaCTD 0K RNAP (lanes 3 and 4). GerE (25 pmole for
reactions with cotB template and 400 pmole for reactions with cotC and cotX templates)
was added immediately after 0" RNAP (lanes 2 and 4). DNA templates were pSC146
linearized with HindIII (204-base gerE transcript), a Pqu restriction fragment of pHI-ll
(13 l-base cotB transcript), an EcoRI- PvuII restriction fragment of pHI-12 (l97-base
cotC transcript), and a 180 bp PCR fragment prepared as described in Materials and

Methods (82-base cotX transcript).

86

+orCTD —orCTD
GerE

gerE
cotB

cotC

 

87

Figure 2

Effect of aCTD deletion on transcriptional activation.

In vitro transcription signals produced by heterologous RNA polymerase with or without
intact aCTD in the presence of GerE were quntified from the experiments shown in
Figure 1 and Figure 3, and from a second set of experiments. Signals by Ecwta aK RNAP
were adjusted as described in Results to take into account the higher activity of this
reconstituted RNAP compared with EcAaCTD 0" RNAP. The graph shows the average
ratio of the signal produced by EcAocCTD oK RNAP signal, expressed as a percentage.

Error bars indicate 1 standard deviation.

88

 

 

 

m m m m m 0

1|.

833:3 :0 E05 mo “outm—

 

89

Figure 3

In vitro transcription of sigK and gerE by heterologous RNA polymerase with or
without aCTD.

DNA template (0.1 pmole) was transcribed with reconstituted Ecwta oK RNAP (lanes 1
and 2) or EcAaCTD oK RNAP (lanes 3 and 4). SpoIIID (400 pmole) was added
immediately after oK RNAP (lanes 2 and 4). DNA templates were pSC146 linearized
with HindIII (204-base gerE transcript) and pBKl6 linearized with Xbal ( l70-base sigK

transcript).

90

+0tCTD —orCTD

SpoIIID - + - +
12 3,4

gerE flN/A H N/A

0 i:_. '~"- _ " E , A Ti" ‘ (:4
51 K ~“* in;
g 3...,» g: ’ 4

 

 

91

BIBLIOGRAPHY

92

BIBLIOGRAPHY

Alper, S., L. Duncan, et al. (1994). “An adenosine nucleotide switch controlling the
activity of a cell type-specific transcription factor in B. subtilis.” Q11 77(2): 195-205.

Arigoni, F ., L. Duncan, et al. (1996). “SpoIIE governs the phosphorylation state of a
protein regulating transcription factor sigma F during sporulation in Bacillus subtilis.”

Proc Natl Acad Sci U S A 93(8): 3238-42.

Arigoni, F., A. M. Guerout-Fleury, et al. (1999). “The SpoIIE phosphatase, the
sporulation septum and the establishment of forespore-specific transcription in Bacillus
subtilis: a reassessment.” Mol Microbiol 31(5): 1407-15.

Bagyan, 1., J. Hobot, et al. (1996). “A compartmentalized regulator of developmental
gene expression in Bacillus subtilis”; Bacteriol 178(15): 4500-7.

Baldus, J. M., B. D. Green, et al. (1994). “Phosphorylation of Bacillus subtilis
transcription factor SpoOA stimulates transcription from the spoIIG promoter by
enhancing binding to weak 0A boxes.” J Bacteriol 176(2): 296-306.

Barak, 1., P. Prepiak, et al. (1998). “MinCD proteins control the septation process during
sporulation of Bacillus subtilis.” J Bacteriol 180(20): 5327-33.

Beall, B. and J. Lutkenhaus (1989). “Nucleotide sequence and insertional inactivation of
a Bacillus subtilis gene that affects cell division, sporulation, and temperature
sensitivity.” J Bacteriol 171(12): 6821-34.

Bird, T. H., J. K. Grimsley, et al. (1996). “The Bacillus subtilis response regulator SpoOA
stimulates transcription of the spoIIG operon through modification of RNA polymerase
promoter complexes.” 1 Mol Biol 256(3): 436-48.

Boor, K. J ., M. L. Duncan, et al. (1995). “Genetic and transcriptional organization of the
region encoding the beta subunit of Bacillus subtilis RNA polymerase.” J Biol Chem
270(35): 20329-36.

Bouche, J. P. and S. Pichoff (1998). “On the birth and fate of bacterial division sites.”
Mg] Microbiol 29(1): 19-26.

Bramucci, M. G., B. D. Green, et al. (1995). “Identification of a Bacillus subtilis spoOH
allele that is necessary for suppression of the sporulation-defective phenotype of a SpoOA
mutation”! Bacteriol 177(6): 1630-3.

Cabrera-Hemandez, A. and P. Setlow (2000). “Analysis of the regulation and fimction of
five genes encoding small, acid-soluble spore proteins of Bacillus subtilis.” Gene 248(1-
2): 169-181.

 

93

Carter, H. L. d. and C. P. Moran, Jr. (1986). “New RNA polymerase sigma factor under
spoO control in Bacillus subtilis.” Proc Natl Acad Sci U S A 83(24): 9438-42.

Cutting, S., V. Oke, et al. (1990). “A forespore checkpoint for mother cell gene
expression during development in B. subtilis.” @ 62(2): 239-50.

Diederich, B., J. F. Wilkinson, et al. (1994). “Role of interactions between SpoIIAA and
SpoIIAB in regulating cell-specific transcription factor sigma F of Bacillus subtilis.”
Genes Dev 8(21): 2653-63.

Dove, S. L. and A. Hochschild (1998). “Conversion of the omega subunit of Escherichia

coli RNA polymerase into a transcriptional activator or an activation target.” Genes Dev
12(5): 745-54.

Driks, A. (1999). “Bacillus subtilis spore coat.” Microbiol Mol Biol Rev 63(1): 1-20.

Driks, A. and R. Losick (1991). “Compartmentalized expression of a gene under the
control of sporulation transcription factor sigma E in Bacillus subtilis.” rog Natl Acad
Sci U S A 88(22): 9934-8.

Duncan, L., S. Alper, et al. (1995). “Activation of cell-specific transcription by a serine
phosphatase at the site of asymmetric division [see comments].” Science 270(5236): 641-
4.

Duncan, L., S. Alper, et al. (1996). “SpoIIAA governs the release of the cell-type specific
transcription factor sigma F from its anti-sigma factor SpoIIAB.” J Mol Biol 260(2): 147-
64.

Ebright, R. H. (1985). “Use of "loss-of-contact" substitutions to identify residues
involved in an amino aicd-base pair contact: effect of substitution of Glnl 8 of Lac
repressor by Gly, Ser, and Leu.” !. Biomol. Struct. Dmam. 3: 281-297.

Errington, J. (1993). “Bacillus subtilis sporulation: regulation of gene expression and
control of morphogenesis.” Microbiol Rev 57(1): 1-33.

Feucht, A., T. Magnin, et al. (1996). “Bifunctional protein required for asymmetric cell
division and cell-specific transcription in Bacillus subtilis.” Genes Dev 10(7): 794-803.

Frandsen, N. and P. Stragier (1995). “Identification and characterization of the Bacillus
subtilis spoIIP locus.” ! Bacteriol 177(3): 716-22.

Fujita, N. and A. Ishihama (1996). “Reconstitution of RNA polymerase.” Methods
Enmol 273: 121-30.

 

Garsin, D. A., L. Duncan, et al. (1998). “The kinase activity of the antisigma factor
SpoIIAB is required for activation as well as inhibition of transcription factor sigmaF
during sporulation in Bacillus subtilis.” J Mol Biol 284(3): 569-78.

94

Gholamhoseinian, A. and P. J. Piggot (1989). “Timing of spoII gene expression relative
to septum formation during sporulation of Bacillus subtilis.” J Bacteriol 171(10): 5747-9.

Gomez, M., S. Cutting, et al. (1995). “Transcription of spoIVB is the only role of sigma
G that is essential for pro-sigma K processing during spore formation in Bacillus
subtilis.” J Bacteriol 177(16): 4825-7.

Gomez, M. and S. M. Cutting (1996). “Expression of the Bacillus subtilis spoIVB gene is
under dual sigma F/sigma G control.” Microbiology 142(Pt 12): 3453-7.

Green, D. H. and S. M. Cutting (2000). “Membrane topology of the Bacillus subtilis pro-
sigma(K) processing complex.” ! Bacteriol 182(2): 278-85.

Grossman, A. D. (1995). “Genetic networks controlling the initiation of sporulation and
the development of genetic competence in Bacillus subtilis.” Annu Rev Genet 29: 477-
508.

Halberg, R. and L. Kroos (1994). “Sporulation regulatory protein SpoIIID from Bacillus
subtilis activates and represses transcription by both mother-cell-specific forms of RNA
polymerase.” J Mol Biol 243(3): 425-36.

Haldenwang, W. G. (1995). “The sigma factors of Bacillus subtilis.” Microbiol Rev
59(1): 1-30.

Harry, E. J ., K. Pogliano, et al. (1995). “Use of immunofluorescence to visualize cell-
specific gene expression during sporulation in Bacillus subtilis.” J Bacteriol 177(12):
3386-93.

Hicks, K. A. and A. D. Grossman (1996). “Altering the level and regulation of the major
sigma subunit of RNA polymerase affects gene expression and development in Bacillus
subtilis.” Mol Microbiol 20(1): 201- 12.

Hofmeister, A. (1998). “Activation of the proprotein transcription factor pro-sigmaE is
associated with its progression through three patterns of subcellular localization during
sporulation in Bacillus subtilis.” ! Bacteriol 180(9): 2426-33.

Hofmeister, A. E., A. Londono-Vallejo, et al. (1995). “Extracellular signal protein
triggering the proteolytic activation of a developmental transcription factor in B. subtilis.’
Cell 83(2): 219-26.

9

Ichikawa, H. and L. Kroos (2000). “Combined action of two transcription factors
regulates genes encoding spore coat proteins of bacillus subtilis [In Process Citation].” !
Biol Chem 275(18): 13849-55.

Ju, J. and W. G. Haldenwang (1999). “The &quot;pro&quot; sequence of the sporulation-

specific sigma transcription factor sigma(E) directs it to the mother cell side of the
sporulation septum.” J Bacteriol 181(19): 6171-5.

95

Ju, J ., T. Luo, et al. (1997). “Bacillus subtilis Pro-sigmaE fusion protein localizes to the
forespore septum and fails to be processed when synthesized in the forespore.” !
Bagterigl 179(15): 4888-93.

Ju, J ., T. Luo, et al. (1998). “F orespore expression and processing of the SigE
transcription factor in wild-type and mutant Bacillus subtilis.” J Bacteriol 180(7): 1673-
81.

Ju, J ., T. Mitchell, et al. (1999). “Sigma factor displacement from RNA polymerase
during Bacillus subtilis sporulation.” J Bacteriol 181(16): 4969-77.

Karow, M. L., P. Glaser, et al. (1995). “Identification of a gene, spoIIR, that links the
activation of sigma E to the transcriptional activity of sigma F during sporulation in
Bacillus subtilis.” Proc Natl Acad Sci U S A 92(6): 2012-6.

Kellner, E. M., A. Decatur, et al. (1996). “T wo-stage regulation of an anti-sigma factor
determines developmental fate during bacterial endospore formation.” Mol Microbiol
21(5): 913-24.

Kenney, T. J. and C. P. Moran (1991). “Genetic evidence for interaction of 0A with two
promoters in Bacillus subtilis.” J. Bacteriol. 173: 3282-3290.

Khvorova, A., L. Zhang, et a1. (1998). “The SpoIIE locus is involved in the SpoOA-
dependent switch in the location of FtsZ rings in Bacillus subtilis.” J Bacteriol 180(5):
1256-60.

King, N., O. Dreesen, et al. (1999). “Septation, dephosphorylation, and the activation of
sigmaF during sporulation in Bacillus subtilis.” Genes Dev 13(9): 1156-67.

Kirchman, P. A., H. DeGrazia, et al. (1993). “F orespore-specific disappearance of the
sigma-factor antagonist SpoIIAB: implications for its role in determination of cell fate in
Bacillus subtilis.” Mol Microbiol 8(4): 663-71.

Krogh, S., M. O'Reilly, et al. (1996). “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.” Microbiology 142(Pt 8): 2031-40.

Kroos, L. (1991). “Gene regulation in the mother-cell compartment of sporulating
Bacillus subtilis.” Sem. Dev. Biol. 2(1): 63-71.

Kroos, L., B. Kunkel, et al. (1989). “Switch protein alters specificity of RNA polymerase
containing a compartment-specific sigma factor.” Science 243(4890): 526-9.

 

Kroos, L., B. Zhang, et al. (1999). “Control of sigma factor activity during Bacillus
subtilis sporulation.” Mol Microbiol 31(5): 1285-94.

Kunkel, B., L. Kroos, et al. (1989). “Temporal and spatial control of the mother-cell

96

regulatory gene spoIIID of Bacillus subtilis.” Genes Dev 3(11): 1735-44.

Kunkel, B., R. Losick, et al. (1990). “The Bacillus subtilis gene for the development
transcription factor sigma K is generated by excision of a dispensable DNA element
containing a sporulation recombinase gene.” Genes Dev 4(4): 525-35.

Lee, J. H. and T. R. Hoover (1995). “Protein crosslinking studies suggest that Rhizobium
meliloti C4-dicarboxylic acid transport protein D, a sigma 54-dependent transcriptional
activator, interacts with sigma 54 and the beta subunit of RNA polymerase.” Proc Natl
Acad Sci U S A 92(21): 9702-6.

Levin, P. A. and R. Losick (1994). “Characterization of a cell division gene from Bacillus
subtilis that is required for vegetative and sporulation septum formation [published
erratum appears in J Bacteriol 1994 Jun;176(l l):3442].”! Bacteriol 176(5): 1451-9.

Levin, P. A. and R. Losick (1996). “Transcription factor SpoOA switches the localization
of the cell division protein FtsZ from a medial to a bipolar pattern in Bacillus subtilis.”

Genes Dev 10(4): 478-88.

Levin, P. A., R. Losick, et a1. (1997). “Localization of the sporulation protein SpoIIE in
Bacillus subtilis is dependent upon the cell division protein FtsZ.” M91 Microbigl 25(5):
839-46.

Lewis, P. J., S. R. Partridge, et al. (1994). “Sigma factors, asymmetry, and the
determination of cell fate in Bacillus subtilis.” roc Natl Acad Sci U S A 91(9): 3849-53.

Li, M., H. Moyle, et al. (1994). “Target of the transcriptional activation function of phage
lambda cI protein.” Science 263(5143): 75-7.

 

Londono-Vallejo, J. A. (1997). “Mutational analysis of the early forespore/mother-cell
signalling pathway in Bacillus subtilis.” Microbiology 143(Pt 8): 2753-61.

Londono-Vallejo, J. A. and P. Stragier (1995). “Cell-cell signaling pathway activating a
developmental transcription factor in Bacillus subtilis.” Genes Dev 9(4): 503-8.

Lopez-Diaz, I., S. Clarke, et al. (1986). “spoIID operon of Bacillus subtilis: cloning and
sequence.” ! Gen Microbiol 132(Pt 2): 341-54.

Lu, 8., R. Halberg, et al. (1990). “Processing of the mother-cell sigma factor, sigma K,
may depend on events occurring in the forespore during Bacillus subtilis development.”
Proc Natl Acad Sci U S A 87(24): 9722-6.

Magnin, T ., M. Lord, et al. (1997). “Contribution of partner switching and SpoIIAA
cycling to regulation of sigmaF activity in sporulating Bacillus subtilis”! Bagteriol
179(12): 3922-7.

Miller, A., D. Wood, et al. (1997). “RNA polymerase beta’ subunit: a target of DNA

97

binding-independent activation [see comments].” Science 275(5306): 1655-7.

Niu, W., Y. Kim, et al. (1996). “Transcription activation at class II CAP-dependent
promoters: two interactions between CAP and RNA polymerase.” C_e_1_l 87(6): 1123-34.

Oke, V., M. Shchepetov, et al. (1997). “SpoIVB has two distinct functions during spore
formation in Bacillus subtilis.” Mol Microbiol 23(2): 223-30.

Partridge, S. R. and J. Errington (1993). “The importance of morphological events and
intercellular interactions in the regulation of prespore-specific gene expression during
sporulation in Bacillus subtilis.” Mol Microbiol 8(5): 945-55.

Perez, A. R., A. Abanes-De Mello, et al. (2000). “SpoIIB localizes to active sites of septal
biogenesis and spatially regulates septal thinning during engulfment in bacillus subtilis.”
! Bacteriol 182(4): 1096-108.

Pogliano, K., A. E. Hofmeister, et al. (1997). “Disappearance of the sigma E transcription
factor from the forespore and the SpoIIE phosphatase from the mother cell contributes to
establishment of cell-specific gene expression during sporulation in Bacillus subtilis”!
Bacteriol 179(10): 3331-41.

Resnekov, O. (1999). “Role of the sporulation protein BofA in regulating activation of
the Bacillus subtilis developmental transcription factor sigmaK.”! Baeteriol 181(17):
5384-8.

Resnekov, 0., A. Driks, et al. (1995). “Identification and characterization of sporulation
gene spoVS from Bacillus subtilis”! Bacteriol 177(19): 5628-35.

Resnekov, O. and R. Losick (1998). “Negative regulation of the proteolytic activation of
a developmental transcription factor in Bacillus subtilis.” roc Natl Acad Sei U S A
95(6): 3162-7.

Rudner, D. Z., P. Fawcett, et al. (1999). “A family of membrane-embedded
metalloproteases involved in regulated proteolysis of membrane-associated transcription

factors.” Pree Natl Aces! Sei U S A 96(26): 14765-70.

Sato, T., K. Harada, et al. (1994). “Expression of the Bacillus subtilis spoIVCA gene,
which encodes a site-specific recombinase, depends on the spoIIGB product.” J Bacteriel
176(3): 935-7.

Setlow, B., N. Magill, et al. (1991). “Condensation of the forespore nucleoid early in
sporulation of Bacillus species.” J Bacteriol 173(19): 6270-8.

Sharp, M. D. and K. Pogliano (1999). “An in vivo membrane fusion assay implicates

SpoIIIE in the final stages of engulfinent during Bacillus subtilis sporulation.” Pree Natl
Acad Sci U S A 96(25): 14553-8.

98

Sharpe, M. E. and J. Errington (1995). “Postseptational chromosome partitioning in
bacteria.” Proc Natl Acad Sci U S A 92(19): 8630-4.

Sharpe, M. E. and J. Errington (1996). “The Bacillus subtilis soj-spoOJ locus is required
for a centromere-like function involved in prespore chromosome partitioning.” Mel
Mierobiol 21(3): 501-9.

Smith, K. and P. Youngman (1993). “Evidence that the spoIIM gene of Bacillus subtilis
is transcribed by RNA polymerase associated with sigma E.” J Bacteriol 175(11): 3618-
27.

Stragier, P., C. Bonamy, et al. (1988). “Processing of a sporulation sigma factor in
Bacillus subtilis: how morphological structure could control gene expression.” Cell 52(5):
697-704.

Stragier, P. and R. Losick (1996). “Molecular genetics of sporulation in Bacillus subtilis.”
Anne Rev Genet 30: 297-41.

Sun, D. X., R. M. Cabrera-Martinez, et al. (1991). “Control of transcription of the
Bacillus subtilis spoIIIG gene, which codes for the forespore-specific transcription factor
sigma G.”! Bacteriol 173(9): 2977-84.

Sun, D. X., P. Stragier, et al. (1989). “Identification of a new sigma-factor involved in
compartmentalized gene expression during sporulation of Bacillus subtilis.” Genes Dev
3(2): 141-9.

Voelker, U., A. Voelker, et al. (1995). “Separate mechanisms activate sigma B of
Bacillus subtilis in response to environmental and metabolic stresses.” J Bacteriol
177(13): 3771-80.

Webb, C. D., A. Decatur, et al. (1995). “Use of green fluorescent protein for visualization
of cell-specific gene expression and subcellular protein localization during sporulation in
Bacillus subtilis.” J Baeteriel 177(20): 5906-11.

Wu, L. J. and J. Errington (1994). “Bacillus subtilis spoIIIE protein required for DNA
segregation during asymmetric cell division.” Science 264(5158): 572-5.

 

Wu, L. J ., A. Feucht, et al. (1998). “Prespore-specific gene expression in Bacillus subtilis
is driven by sequestration of SpoIIE phosphatase to the prespore side of the asymmetric
septum.” Genes Dev 12(9): 1371-80.

Wu, L. J ., P. J. Lewis, et al. (1995). “A conjugation-like mechanism for prespore
chromosome partitioning during sporulation in Bacillus subtilis.” Genes Dev 9(11):
13 16-26.

Yanish-Perron, C., J. Vieira, et al. (1985). “Improved M13 phage cloning vectors and
host strains: nucleotide sequences of the M13mp18 and pUCl9 vectors.” Gene 33: 103-

 

99

119.

Yanouri, A., R. A. Daniel, et al. (1993). “Cloning and sequencing of the cell division
gene ppr, which encodes penicillin-binding protein 2B in Bacillus subtilis.” ! Baeteriol
175(23): 7604-16.

Zhang, B., R. A. Daniel, et al. (1997). “Bacillus subtilis SpoIIID protein binds to two
sites in the spoVD promoter and represses transcription by sigmaE RNA polymerase.” !
Baeteriol 179(3): 972-5.

Zhang, B. and L. Kroos (1997). “A feedback loop regulates the switch from one sigma
factor to the next in the cascade controlling Bacillus subtilis mother cell gene
expression.” ! Bacteriel 179(19): 6138-44.

Zhang, B., P. Struffi, et al. (1999). “sigmaK can negatively regulate sigE expression by
two different mechanisms during sporulation of Bacillus subtilis.” J Baeteriel 181(13):
4081-8.

Zhang, J ., H. Ichikawa, et al. (1994). “Regulation of the transcription of a cluster of
Bacillus subtilis spore coat genes”! Mol Biol 240(5): 405-15.

Zheng, L., R. Halberg, et al. (1992). “Sporulation regulatory protein GerE from Bacillus
subtilis binds to and can activate or repress transcription from promoters for mother-cell-
specific genes”! Mol Biol 226(4): 1037-50.

Zheng, L. B. and R. Losick (1990). “Cascade regulation of spore coat gene expression in
Bacillus subtilis”! Mol Biol 212(4): 645-60.

Zhou, Y., P. S. Pendergrast, et al. (1994). “The functional subunit of a dimeric
transcription activator protein depends on promoter architecture.” Embo J 13(19): 4549-
57.

 

100