ABSTRACT
Transcription of the bacteriophage Mu
mom
operon requires transactivation by the phage-encoded C protein. DNase I footprinting showed that in the absence of C,
Escherichia coli
RNA polymerase E
[sigma]
70 (RNAP) binds to the
mom
promoter (P
mom
) region at a site, P2 (from -64 to -11 with respect to the transcription start site), on the top (non-transcribed) strand. This is slightly upstream from, but overlapping P1 (-49 to +16), the functional binding site for rightward
transcription. Host DNA-[
N
6-adenine] methyltransferase (Dam) methylation of three GATCs immediately upstream of the C binding site is
required to prevent binding of the
E.coli
OxyR protein, which represses
mom
transcription in
dam
- strains. OxyR, known to induce DNA bending, is normally in a reduced
conformation
in vivo
, but is converted to an oxidized state under standard
in vitro
conditions. Using DNase I footprinting, we provide evidence supporting the
proposal that the oxidized and reduced forms of OxyR interact differently with
their target DNA sequences
in vitro
. A mutant form, OxyR-C199S, was shown to be able to repress
mom
expression
in vivo
in a
dam
-
host.
In vitro
DNase I footprinting showed that OxyR-C199S protected P
mom
from -104 to -46 on the top strand and produced a protection pattern
characteristic of reduced wild-type OxyR. Prebinding of OxyR-C199S completely blocked RNAP binding to P2 (in the absence of C),
whereas it only slightly decreased binding of C to its target site (-55 to -28, as defined by DNase I footprinting). In contrast, OxyR-C199S strongly inhibited C-activated recruitment of RNAP to P1. These results
indicate that OxyR repression is mediated subsequent to binding by C. Mutations
have been isolated that relieve the dependence on C activation and have the
same transcription start site as the C-activated wild-type promoter. One such mutant,
tin7
,
has a single base change at -14, which changes a T
6
run to T
3
GT
2
. OxyR-C199S partially inhibited RNAP binding to the
tin7
promoter
in vitro
, even though the OxyR and RNAP-P1 binding sites probably do not overlap, and
in vivo
expression of
tin7
was reduced 5- to 10-fold in
dam
- cells. These results suggest that OxyR can repress
tin7
.
Among the bacteriophages, Mu is unique. Besides its remarkable property of functioning as a transposable element (
1
-
4
), Mu employs two unusual strategies to extend its host range. One involves expression of
alternative sets of tail fibers that specify different adsorptive capabilities
on various host cells; this is accomplished by DNA inversion of the phage G
region, which encodes these proteins (
5
,
6
). The other strategy involves an unusual type of DNA modification function,
encoded by the phage
mom
gene. `Momification' is an unusual sequence-specific modification of adenines that protects the phage DNA against a
variety of host-controlled restriction/modification systems (
7
-
9
). However, untimely Mom production is cytotoxic (
10
,
11
). Therefore, in order to establish and maintain the lysogenic pathway, Mu has evolved an
intricate set of transcriptional and translational controls to regulate
mom
expression (
12
).
The
mom
operon is at the rightmost end of the bacteriophage Mu genome and is comprised
of two overlapping genes,
com
and
mom
(
13
); the
com
gene product is a sequence-specific mRNA binding protein that appears to activate
mom
translation by melting a stem-loop structure to expose the translational start signals (
14
,
15
). Transcription of
mom
by
Escherichia coli
RNA polymerase E[sigma]
70
(RNAP) is subject to a complex regulation scheme; it requires transactivation by the phage-encoded C protein (
16
,
17
), as well as function of the host DNA-[
N
6
-adenine] methyltransferase (Dam) (
18
-
21
). C is also required for the transcription of three other late operons, which
are involved in phage morphogenesis and cell lysis during lytic development (
22
,
23
). Of the four late promoters (P
lys
, P
I
, P
P
and P
mom
), P
mom
has the most conserved -35 hexamer sequence, but is at best poorly homologous to the
E.coli
consensus RNAP promoter sequence, TTGACA-X
16-18
-TATAAT (
24
). Analysis of the
mom
promoter sequence reveals that it has identities of 3/6 and 4/6 respectively
with the canonical
E.coli
-35 and -10 hexamers, but the three matches in the -35 hexamer are not at so-called `invariant' positions (
25
). Furthermore, the spacer between the two hexamers is a suboptimal 19 (instead
of 17) bp. Thus, it is no surprise that RNAP cannot by itself initiate
transcription at the
mom
promoter. The precise mechanism by which C activates transcription is currently
unclear. C binds P
mom
from -55 to -28, as defined by DNase I footprinting (
25
,
26
); chemical footprinting with MPE
.
Fe(II) narrows the boundaries from -53 to -35 (
16
). C binding promotes RNAP binding at a site, P1, functional in rightward
transcription, and precludes RNAP binding at another site, P2, which is
slightly upstream and overlapping P1 (
2
7
,
2
8
; unpublished results). Whether or not C is also required at some later step in
transcription initiation is unknown. Recent
in vitro
studies indicate that RNAP bound at P2 produces a low level of short
transcripts in the leftward direction (
2
8
), adding one more dimension to the complexity of
mom
regulatory control.
Mutations have been isolated that relieve the dependence on C activation,
without altering the transcription start site (
2
7
). One such mutant,
tin7
, has a single base change at -14, which changes a T
6
run (in the top strand) to T
3
GT
2
. Not only does this substitution disrupt any intrinsic bending due to the T
6
run, but it also creates a TG at -15, -14, which has been shown to relieve dependence upon accessory
factors for several other
E.coli
promoters (
2
9
).
Interestingly, Dam methylation of the adenine residues in three closely spaced
GATCs located immediately upstream of the proposed C recognition site is also
required for
mom
transcription (
18
-
21
). This was unexpected because of the cases known where DNA methylation
negatively regulates gene expression in eukaryotes. Later studies showed that Dam methylation blocks binding of another host
protein, OxyR, which represses
mom
transcription in
dam
-
strains (
30
). Recent experiments indicate that OxyR modulates
mom
expression even in
dam
+
cells (unpublished observations). The
E.coli
OxyR protein is a redox-sensitive transcriptional regulator that, under oxidative stress, induces
the expression of a set of antioxidant defense genes. OxyR is 305 amino acids
in length and belongs to the LysR family, whose members share a conserved helix-turn-helix motif involved in DNA binding (
3
1
,
3
2
). OxyR represses its own transcription during growth in the absence of
oxidative stress (
3
1
). Detailed studies indicated that increased expression of OxyR-activated genes by treatment with low doses of hydrogen peroxide (H
2
O
2
) is a consequence of an induced conformational change in the protein and that
the reduced and oxidized forms of OxyR have different DNA binding properties
and protection patterns (
3
3
,
3
4
). During growth in the absence of oxidative stress intracellular OxyR is in the
reduced conformation (
3
3
).
In this paper, we show that both reduced and oxidized forms of OxyR
in vitro
bind P
mom
at a distance further upstream relative to its binding in the
oxyR
promoter (
3
1
), suggesting that the mechanisms of repression of the two promoters might be
different. Our results on
in vitro
DNase I footprinting of P
mom
support the notion (
3
4
) that the reduced and oxidized forms of OxyR make different contacts in DNA
binding. In order to carry out an
in vitro
investigation of P
mom
binding by reduced OxyR, C and RNAP, it was necessary to obviate the
requirement for high dithiothreitol (DTT) concentrations. Therefore, we took advantage of the existence of a mutant, OxyR-C199S, which appears to be `locked' in a reduced conformation (
3
4
,
3
5
). We observed that, like the wild-type OxyR
in vivo
, OxyR-C199S was able to repress
mom
transcription in
dam
-
cells. Furthermore,
in vitro
OxyR-C199S gave a DNase I footprint identical to that of reduced wild-type OxyR. Although OxyR-C199S only slightly reduced C binding to its target site
in vitro
, prebinding of OxyR-C199S prevented both RNAP binding at P2 as well as C-activated RNAP binding at P1. These results indicate that OxyR-mediated repression of P
mom
transcription is mediated subsequent to binding by C.
In the absence of C, OxyR-C199S partially inhibited RNAP binding to the
tin7
promoter
in vitro
, even though the OxyR and RNAP P1 binding sites probably do not overlap (5'- and 3'-boundaries defined by DNase I protection extend
further than the actual contacts), and
in vivo
expression of
tin7
was reduced in
dam
-
cells. These results suggest that OxyR may exert repression of
tin7
through its ability to bend DNA. Finally, the mechanism involved in the
regulation of
mom
transcription initiation is discussed.
Escherichia coli
K12 strains TA4484
oxyR
[Delta]
3
(pMC7, pGSO68) (
3
5
,
3
6
) and GSO9 were kindly provided by Dr G. Storz; GSO9 carries an
oxyR
::
kan
insertion, with the
kan
promoter in the same direction as the
oxyR
promoter. Plasmid pGSO68, a pKK177-3 derivative, contains a mutant
oxyR-C199S
gene with a modified Shine-Dalgarno sequence and under control of the
tac
promoter (
3
5
).
Escherichia coli
DH5[alpha]F' [Phi]
80dlacZ
[Delta]
M15
[Delta](
lacZYA-argF
)
U169 endA1 recA1 hsdR17(r
K
-
m
K
+
) deoR thi-1 supE44
[lambda]
-
gyrA96 relA1
, carrying the wild-type OxyR overproducing plasmid pJL
momR
[Delta]
15
, was a generous gift from R. Kahmann (
30
).
Escherichia coli
JM83
ara
[Delta](
pro-lac
)
rpsL thi
[Phi]
80dlacZ
[Delta]
M15
was from Bethesda Research Laboratories (BRL).
Escherichia coli
GM1853
dam3 dcm6
[Delta](
pro-lac
)
thi
and GM2972
thr1 leuB6 thi1 supE44 lacY1 proA2 galK2 ara14 xyl5 mtl1 rpsL31 his4 tsx33 dam13
::
Tn9 mutH34
were from M. G. Marinus (
3
7
).
Plasmid pLW4 (~8.8 kb) has been described previously (
2
7
); it contains the P
mom
region (from -136 to +79) and a 5' portion of the proximal
com
gene fused in-frame to the
E.coli lacZ
gene. A Com-LacZ fusion protein with [beta]-galactosidase activity is produced by pLW4, but only if
transactivated by C. Plasmid pLW4-
tin7
differs from pLW4 in that it has a T -> G transversion at -14 within a run of six thymines on the top strand (
2
7
); it produces [beta]-galactosidase activity constitutively.
Escherichia coli
TA4484
oxyR
[Delta]
3
(pMC7, pGSO68) was grown in LB + ampicillin (100 [mu]g/ml) in order to promote segregational loss of plasmid pMC7, which confers
tetracycline resistance (
3
6
). Plasmid DNA was prepared from 1.5 ml of overnight cells by the alkaline lysis
method (
3
8
) and used to transform competent
E.coli
DH5[alpha]F' cells. An Amp
r
Tet
s
transformant was characterized by restriction nuclease digestion and confirmed
to carry only pGSO68. Plasmid DNA was cleaved with
Hin
dIII and the single-stranded tails filled-in with the
E.coli
Klenow fragment (PolIK). After digestion with
Eco
RI, the ~1.5 kb fragment containing the
oxyR-C199S
gene was recovered from a low melting temperature agarose gel and then cloned
into the
Eco
RI and
Sca
I sites of pACYC184 (
3
9
) under the
cat
gene
promoter. This construct (~5.3 kb) was designated pROR184.
Plasmid pMLC322 was created by replacing the ~1.2 kb
Sca
I-
Bsu
36I fragment from pMLF-2 (
17
) with the corresponding ~2.1 kb fragment from pTLG1(+) (
2
8
). The resulting Amp
r
plasmid (~11.2 kb) carries the pBR322 replication origin, which makes it compatible
with pACYC184 derivatives; it also contains the Mu
C
gene (under the
bla
promoter), as well as a
com
-
lacZ
translational fusion under control of P
mom
(
2
7
).
Plasmid pMLC322[Delta]
mom
was constructed by removal of a 220 bp
Eco
RI-
Bam
HI
mom
promoter fragment, followed by fill-in with PolIK and blunt-end ligation.
Phage P1
vir
GM2972
dam
13::Tn
9
was used to transduce
E.coli
JM83, essentially by the method of Miller (
40
). The
dam
-
phenotype of chloramphenicol-resistant transductants was screened by
Mbo
I digestion of plasmids transformed into and isolated from these strains and by
poor growth phenotype on 2-aminopurine (
4
1
). JM83
dam
13::Tn
9
was subsequently used to generate JM83
dam
13::Tn
9
[Delta]
oxyR
::
kan
by P1
vir
GSO9[Delta]
oxyR
::
kan
transduction. The [Delta]
oxyR
genotype was confirmed by the loss of OxyR repression, as evidenced by increased
expression of [beta]-galactosidase activity from pMLC322 in JM83
dam
13::Tn
9
[Delta]
oxyR
::
kan
relative to the JM83
dam
13::Tn
9
parent.
For qualitative screening, cells were streaked on MacConkey lactose plates (Difco Laboratories) supplemented with appropriate antibiotics and grown
for 12-16 h at 37oC. For quantitative liquid culture assays, overnight cells grown in
LB plus appropriate antibiotics were diluted 1:100 into fresh medium and
allowed to grow to log phase at 37oC. Aliquots were assayed in triplicate as described by Miller (
40
).
C protein was previously purified as described (
25
,
26
). Wild-type OxyR protein was overproduced from the heat-inducible phage [lambda] P
L
-P
R
tandem promoters in pJL
momR
[Delta]
15
as described (
30
). OxyR-C199S protein was overexpressed from the
tac
fusion promoter in pGSO68 (
3
5
) by IPTG induction. After sonication, the wild-type and mutant OxyR forms were purified by the method of Bölker and Kahmann (
30
). To further purify and concentrate the proteins, we diluted the final
fractions to [NaCl] = 0.1 M and applied them to 2 ml SP Sepharose ion-exchange columns and eluted at 0.5 M NaCl. Protein concentration was
determined according to the BioRad protein microassay using BSA as standard and
adjusted to 1 mg/ml. Both proteins were at least 90% pure as estimated by 15%
SDS-PAGE and silver staining.
Unmethylated supercoiled plasmid DNA (pLW4 or pLW4-
tin7
) was isolated from a
dam
-
host (GM1853) by the alkaline lysis method and purified by CsCl/EtBr gradient
ultracentrifugation (
3
8
). One microgram of plasmid DNA (~0.18 pmol) was incubated in a total volume of 20 [mu]l binding buffer (25 mM Tris-HCl, pH 7.9, 30 mM KCl, 6 mM MgCl
2
, 50 [mu]g/ml BSA, 4% glycerol and 1 mM DTT) at room temperature (22oC) with various combinations of proteins added in different orders, as
described in the figures. One microliter of 0.2 U/[mu]l DNase I (Pharmacia Biotech) was added and, after 45 s, the digestion was
terminated by addition of 20 [mu]l stop buffer (0.1 M Tris-HCl, pH 7.5, 25 mM EDTA and 0.5% SDS). The sample was then brought to
125 [mu]l with H
2
O and extracted successively with phenol/chloroform/isoamyl alcohol (25:24:1)
and chloroform/isoamyl alcohol (24:1) and then precipitated in 0.2 M NaCl plus
2 vol 95% ethanol in the presence of carrier yeast tRNA.
Primer extension was performed essentially as previously described (
4
2
). Briefly, DNase I-treated DNA was suspended in 31 [mu]l H
2
O and incubated with 5 [mu]l (3 * 10
5
c.p.m.)
32
P-end-labeled
lac
primer (NEB), which was complementary to the top strand of P
mom
. Following alkali denaturation with 4 [mu]l 0.1 M NaOH at 80oC for 2 min, the DNA was placed on ice. The primer was annealed at 50oC for 3 min after adding 5 [mu]l TMD buffer (0.5 M Tris-HCl, pH 7.2, 0.1 M MgCl
2
and 2 mM DTT). The annealed primer was extended for 10 min at 45oC following addition of 5 [mu]l of all four dNTPs (5 mM each) and 1 [mu]l 1 U/[mu]l PolIK. The reactions were terminated with 5 [mu]l 0.1 M EDTA and precipitated with 110 [mu]l 95% ethanol. After centrifugation, pellets were
suspended in 10 [mu]l 0.5* loading dye and applied to a 6% polyacrylamide-urea denaturing gel; sequencing reactions with untreated DNA
were run in parallel lanes.
Using random mutagenesis with bacterial transposon Tn
5
, Bölker and Kahmann (
30
) identified OxyR as the repressor of
mom
transcription in
dam
-
cells. They also showed that (oxidized) OxyR protected a region of P
mom
in vitro
against MPE-Fe(II) cleavage from -92 to -50 on the top strand. While both oxidized and reduced forms
bound to the
oxyR
promoter, they made strikingly different contacts with the DNA (
3
4
). In order to investigate how the two redox states of OxyR might affect its binding patterns on
unmethylated P
mom
, we carried out a DNase I footprinting analysis using supercoiled plasmid pLW4
DNA as the substrate.
As shown in Figure
1
A, oxidized OxyR (1 mM DTT) gave a DNase I footprint from -94 to -46 on the top strand. These boundary values are in good agreement
with MPE-Fe(II) cleavage data above (
30
), although DNase I footprinting generally gives broader protection compared with chemical agents. Under reducing
conditions (200 mM DTT) this protection was extended ~10 bp upstream and DNase I hypersensitive sites at -75T and -74A became apparent (Fig.
1
B). This is consistent with the notion that oxidized OxyR binds to four
successive major grooves on one face of the DNA helix, while the reduced OxyR
binds to two pairs of adjacent major grooves separated by one helical turn (
3
4
). The presence of DNase I hypersensitive sites suggests that reduced OxyR can
bend P
mom
DNA, as it does the
oxyR
promoter.
It was proposed that substitution of serine for cysteine at position 199 locks
OxyR in the reduced conformation, e.g. OxyR-C199S gave a DNase I protection pattern at the
oxyR
promoter (even in 1 mM DTT) characteristic of reduced wild-type OxyR (
3
4
,
4
3
). Therefore, we extended our analysis to the mutant OxyR-C199S protein. As seen in Figure
1
B and C, the DNase I footprint of OxyR-C199S was identical to that of reduced wild-type OxyR, consistent with its being locked in the reduced
conformation. Thus, OxyR-C199S could be useful for
in vitro
studies (see below) because it obviated the need for high DTT concentrations,
which might provoke structural or functional changes in C or RNAP.
Intracellular wild-type OxyR is in a reduced conformation during normal bacterial growth,
i.e. when oxidative stress is not applied (
3
3
). To test whether OxyR-C199S can repress
mom
transcription
in vivo
, we constructed a JM83
dam
-
[Delta]
oxyR
derivative (see Materials and Methods). This strain was transformed with
plasmid pMLC322, a pBR322 derivative containing a
com
-
lacZ
translational fusion under control of the
mom
promoter, as well as the
C
gene under the
bla
promoter. A second plasmid, vector pACYC184 (Fig.
2
, strain B) or pROR184 (which has the
oxyR-C199S
gene under the
cat
promoter in vector pACYC184) (strain C), was then introduced by transformation.
As a negative control, JM83
dam
-
[Delta]
oxyR
was transformed with pMLC322[Delta]
mom
(containing a deletion of the
mom
promoter fragment) and vector pACYC184 (strain A). After 12-16 h growth on MacConkey lactose plates at 37oC, strain B gave dark red colonies, while strain C colonies were
white or slightly pink. This indicates that
com
-
lacZ
expression was strongly reduced in the presence of OxyR-C199S. Quantitative [beta]-galactosidase activity assays were also carried out in liquid culture. As shown on Figure
2
, OxyR-C199S-producing cells exhibited a 13-fold reduction in P
mom
-directed production of the Com-LacZ fusion. This inhibition was much lower when a higher copy
number plasmid was used in place of pMLC322 (data not shown), suggesting that
OxyR-C199S may be limiting in this system. Nevertheless, we can conclude that
OxyR-C199S repressed transcription of P
mom
in
dam
-
cells.
Initially, gel retardation assays were used to examine whether OxyR-C199S affected the binding of C to P
mom
. Increasing amounts of C were added to mixtures of OxyR-C199S preincubated with an unmethylated
32
P-end-labeled DNA fragment containing P
mom
. We observed that formation of the OxyR-C199S-DNA binary complex was gradually replaced by a supershifted band,
indicative of an OxyR-C199S-DNA-C ternary complex (data not shown). However, significant
amounts of radioactivity were always seen between the two bands, suggesting
that the ternary complex dissociated during electrophoresis.
DNA supercoiling can play a role in the transcription initiation process,
including binding of RNAP or regulatory proteins to promoter sequences (
4
4
). We have previously observed that
mom
transcription from linearized templates is reduced compared with transcription from supercoiled minicircle DNA (
26
,
2
8
). Therefore, we undertook a DNase I footprinting analysis with supercoiled plasmid DNA. To
determine the effect of prebound OxyR-C199S on C binding, supercoiled P
mom
DNA was preincubated with a saturating amount of OxyR-C199S; densitometric measurements showed that >95% of the P
mom
DNA in each lane was protected. Various concentrations of C were then added and
each sample was run on a sequencing gel alongside a reaction that contained the
same amount of C but without OxyR-C199S (Fig.
3
). In the absence of OxyR-C199S, the C footprint was barely detectable at 350 nM (in dimeric equivalents). However, when the C concentration was increased 4-fold, the target sequence was completely protected in both the
absence (lane 9) and presence (lane 10) of OxyR-C199S, showing that OxyR-C199S only weakly inhibited C binding. Furthermore, this pattern is consistent with a strong cooperativity in C binding (S. Hattman, X. Song, T.L. Cabot and W. Sun, submitted for
publication). The order of protein additions was reversed in order to examine
whether OxyR-C199S could influence C prebound to P
mom
(Fig.
3
, lanes 11-18). As expected, OxyR-C199S added to the preformed C-DNA complex reduced C binding to about the same extent as
when it was incubated with the DNA prior to C addition.
Clearly, the weak
in vitro
reduction of C binding due to OxyR-C199S does not suffice to account for
mom
repression seen
in vivo
, since transcription of the
mom
gene in a
dam
-
strain is at least 20-fold lower than in a
dam
+
strain (
18
). This prompted us to address what effect OxyR-C199S exerts on
E.coli
RNAP binding (under conditions where C binding is not affected). Plasmid pLW4 DNA was incubated with or without OxyR-C199S for 20 min prior to the addition of C; the protein concentrations used
produced complete protection of both their respective target sequences,
regardless of the presence of the other protein (Fig.
3
, lanes 9 and 10). After binding of C, varying amounts of RNAP were added and
DNase I digestions were carried out 20 min later. As seen in Figure
4
(lanes 4 and 6), in the absence of OxyR-C199S, a C-dependent RNAP-protected region from -49 to +16 (corresponding to the P1 site) was clearly
evident at concentrations of 60 nM RNAP or higher. This protection was strongly
diminished when OxyR-C199S was present (lanes 5 and 7). These results indicate that OxyR
repression inhibited C-activated RNAP binding subsequent to binding by C.
In the work described here, we show that OxyR gave distinctly different DNase I
footprints on an unmethylated
mom
promoter under oxidizing (1 mM DTT) versus reducing (200 mM DTT) conditions
in vitro
and that OxyR-C199S behaved like wild-type OxyR in its reduced state (
33
). Because
in vivo
experiments showed that OxyR-C199S acted as the
mom
repressor in a
dam
-
host, we were able to use OxyR-C199S
in vitro
at 1 mM DTT in place of wild-type OxyR at 200 mM DTT. Thus, by DNase I footprinting, we demonstrated
that while OxyR-C199S has only a weak inhibitory effect on C binding, it blocked C-activated RNAP binding at the functional P1 site of P
mom
, but the molecular mechanism remains to be elucidated. The DNase I footprint of
OxyR-C199S extends 10 bp into the region protected by C (Fig.
6
), making it possible that the two proteins contact each other. The slight
decrease in DNA affinity for C can be explained by either steric hindrance or
DNA structure distortion (e.g. bending). Studies with the partially C-independent
tin7-mom
promoter indicate that OxyR-C199S inhibited RNAP binding directly, as well as indirectly through its
influence on C-activation.
Significantly, binding of C to the wild-type
mom
promoter gave rise to a pronounced DNase I hypersensitive site in the spacer
region at -19T, well downstream of the C footprint (Fig.
3
), suggesting some C-induced DNA conformational change. In this regard, we propose that C
binding to P
mom
may modify the spacer to a three-dimensional structure productive for RNAP placement and that
tin7
DNA, by eliminating a misoriented DNA bend embedded in the spacer and also by
creating-15T, -14G, compensates to some degree for the divergence of the -10 and -35 sequences from the consensus
E.coli
hexamers. Moreover, the
tin7
- and the wild-type
mom
promoters
have a similar affinity for RNAP in the presence of C (at both 60 and 120 nM
RNAP, as determined from Figs
4
and
5
B), suggesting that the
tin7
mutation and C binding might produce similar functional topologies of the
promoter. It should be noted that the
tin7
and the wild-type promoters have the same transcription start site. It is not known
whether OxyR-C199S makes contact with RNAP at the
tin7
promoter and, if so, whether this interaction accounts for the direct
inhibitory effect on RNAP binding. Considering the fact that the DNase I
footprints of OxyR-C199S and RNAP overlap by at most 4 bp (Fig.
6
) and the marked conformational alteration caused by OxyR-C199S binding, it is more likely that decreased RNAP binding occurs as a
result of a DNA conformational change.
In the absence of C, RNAP binds to the P2 site slightly upstream of and
overlapping P1, protecting a region against DNase I cleavage from -64 to -11 (
2
7
). Both OxyR-C199S and wild-type OxyR completely block P2 binding, presumably by steric
exclusion, and they can also displace RNAP already bound at P2 (data not shown). The same steric hindrance mechanism is probably operative in self-repression of
oxyR
transcription, because OxyR binds to its own promoter in a region spanning the -10 hexamer to the +1 site (
3
1
,
3
3
). All findings taken together, we favor the notion that OxyR acts as a
repressor primarily by blocking C activation of RNAP binding at P1. Whether the
redox states of OxyR are relevant to repression of P
mom
transcription awaits further study. OxyR is proposed to bind on one face of its
target DNA sequences as a dimer of dimers (
3
4
). Oxidation of this protein rearranges its contacts with the upstream half of
the
oxyR
site, while contacts with the downstream half-site remain unchanged. Based on the close similarity of its footprints on
P
mom
and P
oxyR
(under reducing or oxidizing conditions), it appears that OxyR interacts with
the two promoters in essentially the same fashion. C binds to P
mom
just 3' of and overlapping the downstream OxyR half-site, suggesting the likelihood that C-OxyR interactions, if any, are not affected by OxyR redox
state. On the other hand, whether and how the redox-dependent OxyR conformational change and the DNA bend directed by reduced,
but not oxidized, OxyR affect
mom
transcription deserve further investigation.
Intriguingly, the -35 elements of activatable promoters often deviate from consensus,
leading to the hypothesis that one role of the activators in these cases is to
provide a contact point for RNAP (
4
4
). In the
mom
system, the DNase I footprints of C and RNAP overlap extensively around the -35 region (Fig.
6
), suggesting that RNAP may be recruited to the promoter by interaction with C
bound at its target site. In agreement with this view, RNAP binding to
the
tin7
promoter is further stimulated by C. C-induced DNA curvature and protein-protein contacts are potentially two aspects of C function. The
fact that the four Mu late promoters, despite their different DNA sequences, have the same requirement for C activation supports the notion that direct C-RNAP communication plays a critical part in RNAP promoter recognition and transcription initiation. With this consideration, an efficient
mechanism of OxyR repression would be to force C out of register with respect
to RNAP, either indirectly by DNA bending or by direct interaction with C (or
both), so that specific C-RNAP contact cannot be achieved. It will be worthwhile to look for C
mutations that can (partially) circumvent the repression effect by OxyR (or OxyR-C199S). If we can then obtain suppressor mutants of OxyR that can restore repression, support for direct protein-protein interaction will be strengthened and we may also be able to
define the contact domains.
We are grateful to Irene Kline for purification of the C protein. Special thanks
are due to Dr Gisela Storz for providing strains, suggestions and continued
interest in this work, which was supported by a PHS grant GM29227 to S.H.
REFERENCES
Return