Orientation of functional activating regions in the
Escherichia coli
CRP protein during transcription activation at class II promoters
Orientation of functional activating regions in the Escherichia coli CRP protein during transcription activation at class II promoters
Roy M.
Williams
1
,
Virgil A.
Rhodius
,
Andrew I.
Bell
,
Annie
Kolb
1
and
Stephen J.W.
Busby*
School of Biochemistry, University of Birmingham, PO Box 363,
Birmingham
B15 2TT,
UK
and
1
Département de Biologie Moléculaire, Institut Pasteur, 25 Rue du Dr Roux,
Paris
15,
France
Received November 15, 1995;
Revised and Accepted January 12, 1996
ABSTRACT
At class II CRP-dependent promoters the DNA site for CRP overlaps the DNA site for RNA
polymerase, covering the -35 region. Transcription activation at class II CRP-dependent promoters requires a contact between an activating region
in the upstream subunit of the bound CRP dimer and a contact site in the C-terminal domain of the
[alpha]
-subunit of RNA polymerase. Transcription activation is suppressed by amino acid substitutions in the
activating region, but activation can be restored by second site substitutions
at K52 or E96. These substitutions identify two separate regions on the surface
of CRP that appear to be able to interact with RNA polymerase specifically at
class II promoters. Using the method of `oriented heterodimers' we show that
these alternative activating regions are functional in the downstream subunit
of the bound CRP dimer.
INTRODUCTION
The
Escherichia coli
cyclic AMP receptor protein (CRP) is a well-characterized transcription activator triggered by cAMP and is responsible
for the induction of more than 100 genes in
E.coli
(reviewed in
1
). CRP is a dimer of two identical subunits and at target promoters it binds to
22 bp sequences. The crystal structure of the CRP-DNA complex has been determined; the complex is 2-fold symmetrical, with one subunit binding to one half of the 22 bp
target and the second subunit binding to the other half (
2
).
The many promoters where CRP alone is sufficient for transcription activation
can be categorized according to the position of the CRP binding site (
3
,
4
). At class I promoters CRP binds upstream of RNA polymerase, the paradigm being
the
E.coli lac
promoter, where the CRP binding site is centred between 61 and 62 bp upstream
from the transcription start site (-61.5). In contrast, at class II promoters the CRP binding site overlaps
the -35 region and is centred at or near -41.5. Transcription activation at both classes of promoter involves a
direct contact with RNA polymerase via a surface-exposed loop including CRP residues T158 and H159. This region (dubbed
activating region 1) makes direct contact with a specific contact site in the C-terminal part of the RNA polymerase [alpha]-subunit (
4
-
9
).
Although CRP is a dimer with an activating region 1 in each subunit, experiments with `oriented heterodimers' have shown that only one
activating region 1 is required for transcription activation (
10
,
11
). At class I promoters,
where the bound CRP is located upstream of bound RNA polymerase, a functional
activating region 1 is essential in the downstream CRP subunit, but not in the
upstream subunit. In contrast, at class II promoters activating region 1 is
functional in the upstream subunit of the bound CRP dimer, but not the
downstream subunit (
11
). This is consistent with results from footprinting experiments on open
complexes at class II promoters containing both CRP and RNA polymerase. These
results suggest that RNA polymerase contacts the promoter DNA both upstream and
downstream of the CRP dimer. Thus activating region 1 in the upstream subunit
of the CRP dimer would be able to contact the C-terminal part of the RNA polymerase [alpha]-subunit, which is located just upstream of CRP (
12
-
15
).
Our recent genetic studies have suggested that other regions of CRP might also
make contacts with RNA polymerase in open complexes at class II promoters.
Working with a semi-synthetic CRP-dependent promoter with the DNA site for CRP centred at -41.5,
CC(-41.5),
we showed that the H159L substitution in activating region 1 suppressed
activation, but that substitutions at K52 or E96 could reverse the suppression
and reactivate CRP (
5
,
16
,
17
). Several lines of evidence led us to conclude that the substitutions at K52 or E96 unmask or improve alternative activating regions,
rather than affecting activating region 1 directly. First, both K52 and E96 are located on the opposite face of the CRP monomer to
activating region 1. Second, CRP activity could be increased by different
substitutions at K52 or E96 and these substitutions could compensate for
different defects in activating region 1 (R.Williams and V.Rhodius, unpublished
results;
5
). Remarkably, these substitutions affected activity only at class II promoters,
such as
CC(-41.5)
, suggesting that they defined class-specific activating regions.
In this work we have used the method of `oriented heterodimers' (
10
) to show that the activating regions unmasked by the K52N and E96G
substitutions are functional in the downstream, but not the upstream, subunit of the CRP dimer. Together with the observation
that activating region 1 functions via the upstream subunit at class II
promoters, this shows that RNA polymerase can make contact with both CRP
subunits when the dimer is bound around -41.
MATERIALS AND METHODS
The
E.coli
host strain used here was the [Delta]
crp
derivative of the [Delta]
lac
strain M182, as in our previous work (
5
,
16
). Recombinant plasmids and the different promoter fragments discussed here are
listed in Table
1
. Standard methods for DNA manipulation and characterization were used
throughout. Strains were propagated on MacConkey agar plates or in Lennox broth
containing 25 [mu]g/ml tetracycline, 35 [mu]g/ml kanamycin or 80 [mu]g/ml ampicillin, as appropriate.
The
CX(-41.5)
,
XC(-41.5)
and
CC(-61.5)
promoters were derived from the starting
CC(-41.5)
promoter as shown in Figure
1
. All promoters were carried on
Eco
RI-
Hin
dIII fragments that could be cloned into pBR322 or M13mp19 for genetic
manipulation or into the low copy number
lac
expression vector pRW50 for measurement of promoter strength
in vivo
. [beta]-Galactosidase activity was determined by the Miller protocol,
exactly as in our previous work (
5
,
16
).
Table
2
lists the
crp
derivatives discussed in this study and their origin. Fragments carrying
crp
could be shuttled between plasmids pDCRP and pDW300. Note that these plasmids
are compatible and carry different antibiotic resistance determinants,
permitting the simultaneous introduction of two different types of CRP into the
[Delta]
crp
host. Fragments carrying different derivatives of
crp
were also shuttled into M13mp19 and the recombinants used for site-directed mutagenesis using the Amersham Oligonucleotide-directed Kit (catalogue no. RPN1523). After mutagenesis the complete
sequence of the new
crp
derivative was checked prior to recloning in pDCRP or pDW300.
For
in vitro
studies CRP H159L and CRP E181V H159L K52N were purified by the Ghosaini method
(
20
). Subunit exchange reactions to generate heterodimers and abortive initiation
experiments to assay the activity of resulting heterodimers at
CX(-41.5)
and
XC(-41.5)
were performed as described by Zhou
et al.
(
10
,
11
).
.
Transcription activation by CRP at
CC(-41.5)
and
CC(-61.5)
pDCRP derivative
[beta]-Galactosidase expression in M182[Delta]
crp
cells
CC(-41.5)
cloned
CC(-61.5)
cloned
in pRW50
lac
vector
in pRW50
lac
vector
pDU9
5
5
pDCRP
550
550
pDCRP H159L
40
20
pDCRP H159L K52N
1900
20
pDCRP H159L E96G
400
20
pDCRP H159L K52N E96G
4000
25
pDCRP K52N
4000
400
pDCRP E96G
2500
600
pDCRP K52N E96G
6100
400
[beta]-Galactosidase levels were measured in extracts of exponentially
growing cells and are expressed in standard Miller units. Cells were grown in
Lennox broth supplemented with 25 [mu]g/ml tetracycline and 80 [mu]g/ml ampicillin. Each assay was performed at least five times on
independent clones.
