34 81 Sigma Factor Switching

wea25324_ch08_196-221.indd Page 197
11/17/10
4:41 PM user-f468
/Volume/204/MHDQ268/wea25324_disk1of1/0073525324/wea25324_pagefile
8.1 Sigma Factor Switching
8.1
Sigma Factor Switching
When a phage infects a bacterium, it usually subverts the
host’s transcription machinery to its own use. In the process, it establishes a time-dependent, or temporal, program
of transcription. In other words, the early phage genes are
transcribed first, then the later genes. By the time phage T4
infection of E. coli reaches its late phase, essentially no
more transcription of host genes takes place—only transcription of phage genes. This massive shift in specificity
would be hard to explain by the operon mechanisms described in Chapter 7. Instead, it is engineered by a fundamental change in the transcription machinery—a change in
RNA polymerase itself.
Another profound change in gene expression occurs
during sporulation in bacteria such as Bacillus subtilis.
Here, genes that are needed in the vegetative phase of
growth are turned off, and other, sporulation-specific
genes are turned on. Again, this switch is accomplished by
changes in RNA polymerase. Bacteria also experience
stresses such as starvation, heat shock, and lack of
nitrogen, and they also respond to these by shifting their
patterns of transcription.
Thus, bacteria respond to changes in their environment by global changes in transcription, and these
changes in transcription are accomplished by changes in
RNA polymerase. Most often, these are changes in the
s-factor.
(a) Early transcription; specificity factor: host σ (
197
)
Early genes
Early transcripts
Early proteins, including gp28 (
)
(b) Middle transcription; specificity factor: gp28 (
)
Middle genes
Middle transcripts
Middle proteins, including gp33 (
) and gp34 (
(c) Late transcription; specificity factor: gp33 (
)
) + gp34 (
)
Phage Infection
What part of RNA polymerase would be the logical
candidate to change the specificity of the enzyme? In
Chapter 6, we learned that s is the key factor in determining specificity of phage T4 DNA transcription in vitro, so
s is the most reasonable answer to our question, and
experiments have confirmed that s is the correct answer.
However, these experiments were not done first with the
E. coli T4 system, but with B. subtilis and its phages,
especially phage SPO1.
SPO1, like T4, has a large DNA genome. It has a temporal program of transcription as follows: In the first 5 min or
so of infection, the early genes are expressed; next, the middle genes turn on (about 5–10 min after infection); from
about the 10-min point until the end of infection, the late
genes switch on. Because the phage has a large number of
genes, it is not surprising that it uses a fairly elaborate
mechanism to control this temporal program. Janice Pero
and her colleagues were the leaders in developing the model
illustrated in Figure 8.1.
The host RNA polymerase holoenzyme handles transcription of early SPO1 genes, which is analogous to the
T4 model, where the earliest genes are transcribed by
the host holoenzyme (Chapter 6). This arrangement is
necessary because the phage does not carry its own
Late genes
Late transcripts
Late proteins
Figure 8.1 Temporal control of transcription in phage SPO1infected B. subtilis. (a) Early transcription is directed by the host
RNA polymerase holoenzyme, including the host s-factor (blue); one
of the early phage proteins is gp28 (green), a new s-factor. (b) Middle
transcription is directed by gp28, in conjunction with the host core
polymerase (red); two middle phage proteins are gp33 and gp34
(purple and yellow, respectively); together, these constitute yet another
s-factor. (c) Late transcription depends on the host core polymerase
plus gp33 and gp34.
RNA polymerase. When the phage first infects the cell,
the host holoenzyme is therefore the only RNA polymerase available. The B. subtilis holoenzyme closely
resembles the E. coli enzyme. Its core consists of two
large (b and b9), two small (a), and one very small (v)
polypeptides; its primary s-factor has a molecular mass
of 43,000 kD, somewhat smaller than E. coli’s primary
s (70,000 kD). In addition, the polymerase includes a
d-subunit with a molecular mass of about 20,000 kD.
wea25324_ch08_196-221.indd Page 198
4:42 PM user-f468
/Volume/204/MHDQ268/wea25324_disk1of1/0073525324/wea25324_pagefile
Chapter 8 / Major Shifts in Bacterial Transcription
β′
β
σ
α
gp28
gp34
δ
gp33
ω
Enzyme B
This subunit helps to prevent binding to nonpromoter
regions, a function performed by the E. coli s-factor but
not by the smaller B. subtilis s-factor.
One of the genes transcribed in the early phase of
SPO1 infection is called gene 28. Its product, gp28,
associates with the host core polymerase, displacing the
host s (s43). With this new, phage-encoded polypeptide
in place, the RNA polymerase changes specificity. It
begins transcribing the phage middle genes instead of
the phage early genes and host genes. In other words,
gp28 is a novel s-factor that accomplishes two things: It
diverts the host’s polymerase from transcribing host
genes, and it switches from early to middle phage
transcription.
The switch from middle to late transcription occurs in
much the same way, except that two polypeptides team up
to bind to the polymerase core and change its specificity.
These are gp33 and gp34, the products of two phage middle genes (genes 33 and 34, respectively). These proteins
constitute a s-factor that can replace gp28 and direct the
altered polymerase to transcribe the phage late genes in
preference to the middle genes. Note that the polypeptides
of the host core polymerase remain constant throughout
this process; it is the progressive substitution of s-factors
that changes the specificity of the enzyme and thereby directs the transcription program. Of course, the changes in
transcription specificity also depend on the fact that the
early, middle, and late genes have promoters with different
sequences. That is how they can be recognized by different
s-factors.
One striking aspect of this process is that the different
s-factors vary quite a bit in size. In particular, host s, gp28,
gp33, and gp34 have molecular masses of 43,000, 26,000,
13,000, and 24,000 kD, respectively. Yet they are capable
of associating with the core enzyme and performing a
s-like role. (Of course, gp33 and gp34 must combine forces
to play this role.) In fact, even the E. coli s, with a molecular
mass of 70,000 kDa, can complement the B. subtilis core
in vitro. The core polymerase apparently has a versatile
s-binding site.
How do we know that the s-switching model is valid?
Two lines of evidence, genetic and biochemical, support it.
First, genetic studies have shown that mutations in gene 28
prevent the early-to-middle switch, just as we would predict if the gene 28 product is the s-factor that turns on the
middle genes. Similarly, mutations in either gene 33 or 34
prevent the middle-to-late switch, again in accord with the
model.
Pero and colleagues performed the biochemical studies. First, they purified RNA polymerase from SPO1infected cells. This purifi cation scheme included a
phosphocellulose chromatography step, which separated
three forms of the polymerase. The first of the separated
polymerases, enzyme A, contains the host core polymerase, including d, plus all the phage-encoded factors.
Enzyme C
198
11/17/10
δ σ
Figure 8.2 Subunit compositions of RNA polymerases in SPO1
phage-infected B. subtilis cells. Polymerases were separated by
chromatography and subjected to SDS-PAGE to display their
subunits plus. Enzyme B (first lane) contains the core subunits (b9, b, a,
and v), as well as gp28. Enzyme C (second lane) contains the core
subunits plus gp34 and gp33. The last two lanes contain separated
d- and s-subunits, respectively. (Source: Pero J., R. Tjian, J. Nelson, and R.
Losick. In vitro transcription of a late class of phage SPO1 genes. Nature 257 (18
Sept 1975): f. 1, p. 249 © Macmillan Magazines Ltd.)
The other two polymerases, B and C, were missing d, but
B contained gp28 and C contained gp33 and gp34.
Figure 8.2 presents the subunit compositions of these latter
two enzymes, determined by SDS-PAGE. Without d,
these two enzymes were incapable of specific transcription because they could not distinguish clearly between
promoter and nonpromoter regions of DNA. However,
when Pero and colleagues added d back and assayed
transcription specificity, they found that B was specific
for the delayed early phage genes, and C was specific for
the late genes.
SUMMARY Transcription of phage SPO1 genes in
infected B. subtilis cells proceeds according to a
temporal program in which early genes are transcribed first, then middle genes, and finally late
genes. This switching is directed by a set of phageencoded s-factors that associate with the host core
RNA polymerase and change its specificity of promoter recognition from early to middle to late. The
host s is specific for the phage early genes; the phage
gp28 protein switches the specificity to the middle
genes; and the phage gp33 and gp34 proteins switch
to late specificity.
wea25324_ch08_196-221.indd Page 199
11/17/10
4:42 PM user-f468
/Volume/204/MHDQ268/wea25324_disk1of1/0073525324/wea25324_pagefile
8.1 Sigma Factor Switching
(a)
(b)
Figure 8.3 Two developmental fates of B. subtilis cells.
(a) B. subtilis vegetative cells dividing and (b) sporulation, with an
endospore developing at the left end, and the mother cell at the right
and surrounding the endospore. (Source: Courtesy Dr. Kenneth Bott.)
Sporulation
We have already seen how phage SPO1 changes the
specificity of its host’s RNA polymerase by replacing its
s-factor. In the following section, we will show that the
same kind of mechanism applies to changes in gene expression in the host itself during the process of sporulation.
B. subtilis can exist indefinitely in the vegetative, or growth,
state, as long as nutrients are available and other conditions are appropriate for growth. But, under starvation or
other adverse conditions, this organism forms endospores—
tough, dormant bodies that can survive for years until
favorable conditions return (Figure 8.3).
Sporulation begins with the formation of a polar septum
between daughter cells. Unlike a vegetative septum that
divides the cell equally, the polar septum forms toward one
end, dividing the cell into two unequal parts. The smaller
part (on the left in Figure 8.3), is the forespore, which develops into a mature endospore. The larger part is the mother
cell, which surrounds the endospore.
Gene expression must change during sporulation; cells
as different in morphology and metabolism as vegetative
and sporulating cells must contain at least some different
gene products. In fact, when B. subtilis cells sporulate, they
activate a whole new set of sporulation-specific genes. The
switch from the vegetative to the sporulating state is accomplished by a complex s-switching scheme that turns
off transcription of some vegetative genes and turns on
sporulation-specific transcription.
As you might anticipate, more than one new s-factor is
involved in sporulation. In fact, several participate: sF, sE, sH,
sC, and sK each play a role, in addition to the vegetative sA.
Each s-factor recognizes a different class of promoter. For
example, the vegetative sA recognizes promoters that are
very similar to the promoters recognized by the major E. coli
s-factor, with a 210 box that looks something like TATAAT
and a 235 box having the consensus sequence TTGACA. By
contrast, the sporulation-specific factors recognize quite different sequences. The sF-factor appears first in the sporulation process, in the forespore. It activates transcription of
about 16 genes, including the genes that encode the other
sporulation-specific s-factors. In particular, it activates
spoIIR, which in turn activates the gene encoding sE in the
mother cell. Together, sF and sE put the forespore and mother
cell, respectively, on an irreversible path to sporulation.
To illustrate the techniques used to demonstrate that
these are authentic s-factors, let us consider some work by
Richard Losick and his colleagues on one of them, sE. First,
they showed that this s-factor confers specificity for a known
sporulation gene. To do this, they used polymerases containing either sE or sA to transcribe a plasmid containing a
piece of B. subtilis DNA in vitro in the presence of labeled
nucleotides. The B. subtilis DNA (Figure 8.4) contained promoters for both vegetative and sporulation genes. The vegetative promoter lay in a restriction fragment 3050 bp long,
and the sporulation promoter was in a 770-bp restriction
fragment. Losick and coworkers then hybridized the labeled
RNA products to Southern blots (Chapter 5) of the template
0.4 kb
Veg
EcoRI
HincII
3050 bp
199
HincII
EcoRI
770 bp
Figure 8.4 Map of part of plasmid p213. This DNA region contains two promoters: a vegetative promoter (Veg) and a sporulation promoter
(0.4 kb). The former is located on a 3050-bp EcoRI–HincII fragment (blue); the latter is on a 770-bp fragment (red). (Source: Adapted from Haldenwang
W.G., N. Lang, and R. Losick, A sporulation-induced sigma-like regulatory protein from B. subtilis. Cell 23:616, 1981.)
wea25324_ch08_196-221.indd Page 200
200
11/17/10
4:42 PM user-f468
/Volume/204/MHDQ268/wea25324_disk1of1/0073525324/wea25324_pagefile
Chapter 8 / Major Shifts in Bacterial Transcription
σB
3050
bp
622
404
+
M σE σC
spoIID
309
242
770
660
1
2
Figure 8.5 Specificities of sA and sE. Losick and colleagues
transcribed plasmid p213 in vitro with RNA polymerase containing sA
(lane 1) or sE (lane 2). Next they hybridized the labeled transcripts to
Southern blots containing EcoRI–HincII fragments of the plasmid.
As shown in Figure 8.4, this plasmid has a vegetative promoter in a
3050-bp EcoRI–HincII fragment, and a sporulation promoter in a 770-bp
fragment. Thus, transcripts of the vegetative gene hybridized to the
3050-bp fragment, while transcripts of the sporulation gene hybridized
to the 770-bp fragment. The autoradiograph in the figure shows that
the sA-enzyme transcribed only the vegetative gene, but the sEenzyme transcribed both the vegetative and sporulation genes.
(Source: Haldenwang W.G., N. Lang, and R. Losick, A sporulation-induced sigmalike regulatory protein from B. subtilis. Cell 23 (Feb 1981), f. 4, p. 618. Reprinted by
permission of Elsevier Science.)
DNA. This procedure revealed the specificities of the
s-factors: If the vegetative gene was transcribed in vitro, the
resulting labeled RNA would hybridize to a 3050-bp band
on the Southern blot of the template DNA. On the other
hand, if the sporulation gene was transcribed in vitro, the
labeled RNA product would hybridize to the 770-bp band.
Figure 8.5 shows that when the polymerase contained sA,
the transcript hybridized only to the vegetative band (3050 bp).
By contrast, when the polymerase contained sE, the transcript hybridized to both vegetative and sporulation bands
(3050 and 770 bp). Apparently sE has some ability to recognize vegetative promoters; however, its main affinity seems
to be for sporulation promoters—at least those of the type
contained in the 770-bp DNA fragment.
The nature of the sporulation gene contained in the 770-bp
fragment was not known, so Abraham Sonenshein and
colleagues set out to show that sE could transcribe a wellcharacterized sporulation gene. They chose the spoIID gene,
which was known to be required for sporulation and had
been cloned. They used polymerases containing three different s-factors, sB, sC, and sE, to transcribe a truncated fragment of the gene so as to produce a run-off transcript
(Chapter 5). Previous S1 mapping with RNA made in vivo
had identified the natural transcription start site. Because the
truncation in the spoIID gene occurred 700 bp downstream
of this start site, transcription from the correct start site in
vitro produced a 700-nt run-off transcript. As Figure 8.6
160
147
110
90
76
Figure 8.6 Specificity of sE determined by run-off transcription
from the spoIID promoter. Sonenshein and associates prepared a
restriction fragment containing the spoIID promoter and transcribed it
in vitro with B. subtilis core RNA polymerase plus sE (middle lane) or
sB plus sC (right lane). Lane M contained marker DNA fragments
whose sizes in base pairs are indicated at left. The arrow at the right
indicates the position of the expected run-off transcript from the
spoIID promoter (about 700 nt). Only the enzyme containing sE made
this transcript. (Source: Rong S., M.S. Rosenkrantz, and A.L. Sononshein,
Transcriptional control of the Bacillus subtilis spoIID gene. Journal of Bacteriology
165, no. 3 (1986) f. 7, p. 777, by permission of American Society for Microbiology.)
shows, only sE could produce this transcript; neither of the
other s-factors could direct the RNA polymerase to recognize the spoIID promoter. A similar experiment showed that
sA could not recognize this promoter, either.
Losick and his colleagues established that sE is itself the
product of a sporulation gene, originally called spoIIG.
Predictably, mutations in this gene block sporulation at an
early stage. Without a s-factor to recognize sporulation
genes such as spoIID, these genes cannot be expressed, and
therefore sporulation cannot occur.
SUMMARY When the bacterium B. subtilis sporulates, a whole new set of sporulation-specific genes is
turned on, and many, but not all, vegetative genes
are turned off. This switch takes place largely at the
transcription level. It is accomplished by several new
s-factors that displace the vegetative s-factor from
the core RNA polymerase and direct transcription of
sporulation genes instead of vegetative genes. Each
s-factor has its own preferred promoter sequence.
wea25324_ch08_196-221.indd Page 201
11/17/10
4:42 PM user-f468
/Volume/204/MHDQ268/wea25324_disk1of1/0073525324/wea25324_pagefile
201
8.1 Sigma Factor Switching
–35 P1
Genes with Multiple Promoters
The story of B. subtilis sporulation is an appropriate introduction to our next topic, multiple promoters, because
sporulation genes provided some of the first examples of
this phenomenon. Some genes must be expressed during
two or more phases of sporulation, when different s-factors
predominate. Therefore, these genes have multiple promoters recognized by the different s-factors.
One of the sporulation genes with two promoters is
spoVG, which is transcribed by both EsB and EsE (holoenzymes bearing sB or sE, respectively). Losick and colleagues
achieved a partial separation of these holoenzymes by
DNA-cellulose chromatography of RNA polymerases from
sporulating cells. Then they performed run-off transcription
of a cloned, truncated spoVG gene with fractions from the
peak of polymerase activity. The fractions on the leading
edge of the peak produced primarily a 110-nt run-off transcript. On the other hand, fractions from the trailing edge of
the peak made predominately a 120-nt run-off transcript.
The fraction in the middle made both run-off transcripts.
These workers succeeded in completely separating the
two polymerase activities, using another round of DNAcellulose chromatography. One set of fractions, containing
sE, synthesized only the 110-nt run-off transcript. Furthermore, the ability to make this transcript paralleled the content of sE in the enzyme preparation, suggesting that sE
was responsible for this transcription activity. To reinforce
the point, Losick’s group purified sE using gel electrophoresis, combined it with core polymerase, and showed that it
made only the 110-nt run-off transcript (Figure 8.7). This
same experiment also established that sB plus core polymerase made only the 120-nt run-off transcript.
These experiments demonstrated that the spoVG gene
can be transcribed by both EsB and EsE, and that these two
σB
σE
σB
+
σE
120 nt
P1
110 nt
P2
1
2
B
3
E
Figure 8.7 Specificities of s and s . Losick and colleagues
purified the s-factors sB and sE by gel electrophoresis and tested
them with core polymerase using a run-off transcription assay. Lane 1,
containing sE, caused initiation selectively at the downstream
promoter (P2). Lane 2, containing sB, caused initiation selectively at
the upstream promoter (P1). Lane 3, containing both s-factors, caused
initiation at both promoters. (Source: Adapted from Johnson W.C., C.P.
Moran, Jr., and R. Losick, Two RNA polymerase sigma factors from Bacillus subtilis
discriminate between overlapping promoters for a developmentally regulated gene.
Nature 302 (28 Apr 1983), f. 4, p. 803. © Macmillan Magazines Ltd.)
–10 P1
P1(σ B )
AGGATT•••AAATC•••GGAATTGAT•••TAATGCTTTTATA
–35 P2
–10 P2
P2 σ E
Figure 8.8 Overlapping promoters in B. subtilis spoVG. P1
denotes the upstream promoter, recognized by sB; the start of
transcription and 210 and 235 boxes for this promoter are indicated
in red above the sequence. P2 denotes the downstream promoter,
recognized by sE, the start of transcription and 210 and 235 boxes
for this promoter are indicated in blue below the sequence.
enzymes have transcription start sites that lie 10 bp apart,
as shown in Figure 8.8. Knowing the locations of these start
sites, we can count the appropriate number of base pairs
upstream and find the 210 and 235 boxes of the promoters recognized by each of these s-factors (also shown in
Figure 8.8). Comparing many 210 and 235 boxes recognized by the same s-factor allowed the identification of
consensus sequences, such as those reported in Chapter 6.
SUMMARY Some prokaryotic genes must be tran-
scribed under conditions where two different
s-factors are active. These genes are equipped with
two different promoters, each recognized by one of
the two s-factors. This ensures their expression no
matter which factor is present and allows for differential control under different conditions.
Other s Switches
When cells experience an increase in temperature, or a
variety of other environmental insults, they mount a defense called the heat shock response to minimize damage.
They start producing proteins called molecular chaperones
that bind to proteins partially unfolded by heating and
help them fold properly again. They also produce proteases that degrade proteins that are so badly unfolded that
they cannot be refolded, even with the help of chaperones.
Collectively, the genes encoding the proteins that help
cells survive heat shock are called heat shock genes.
Almost immediately after E. coli cells are heated from
their normal growth temperature (378C) to a higher temperature (428C), normal transcription ceases, or at least decreases,
and the synthesis of 17 new, heat shock transcripts begins.
These transcripts encode the molecular chaperones and proteases that help the cell survive heat shock. This shift in transcription requires the product of the rpoH gene, which
encodes a s-factor with a molecular mass of 32 kD. Hence
this factor is called s32, but it is also known as sH, where
the H stands for heat shock. In 1984, Grossman and coworkers demonstrated that sH really is a s-factor. They did
this by combining sH with core polymerase and showing
that this mixture could transcribe a variety of heat shock
genes in vitro from their natural transcription start sites.