Microbiology (2009), 155, 115–123
DOI 10.1099/mic.0.023432-0
Role of the Escherichia coli Hfq protein in GcvB
regulation of oppA and dppA mRNAs
Sarah C. Pulvermacher, Lorraine T. Stauffer and George V. Stauffer
Correspondence
Department of Microbiology, University of Iowa, Iowa City, IA 52242, USA
George V. Stauffer
[email protected]
Received 18 August 2008
Revised
17 October 2008
Accepted 22 October 2008
The gcvB gene encodes a small non-translated RNA (referred to as GcvB) that regulates oppA
and dppA, two genes that encode periplasmic binding proteins for the oligopeptide and dipeptide
transport systems. Hfq, an RNA chaperone protein, binds many small RNAs and is required for the
small RNAs to regulate expression of their respective target genes. We showed that repression by
GcvB of dppA : : lacZ and oppA : : phoA translational fusions is dependent upon Hfq. Double
mutations in gcvB and hfq yielded similar expression levels of dppA : : lacZ and oppA : : phoA
compared with gcvB or hfq single mutations, suggesting that GcvB and Hfq repress by the same
mechanism. The effect of Hfq is not through regulation of transcription of gcvB. Hfq is known to
increase the stability of some small RNAs and to facilitate the interactions between small
RNAs and specific mRNAs. In the absence of Hfq, there is a marked decrease in the half-life of
GcvB in cells grown in both Luria–Bertani broth and glucose minimal medium with glycine,
suggesting that part of the role of Hfq is to stabilize GcvB. Overproduction of GcvB in wild-type
Escherichia coli results in superrepression of a dppA : : lacZ fusion, but overproduction of GcvB in
an hfq mutant does not result in significant repression of the dppA : : lacZ fusion. These results
suggest that Hfq also is likely required for GcvB–mRNA pairing.
INTRODUCTION
The Escherichia coli gcvTHP operon encodes the glycine
cleavage enzyme complex that provides one-carbon units
for cellular methylation reactions (Kikuchi, 1973). The
gcvB gene encodes a small, non-translated RNA (sRNA) of
206 nt (referred to as GcvB) that, in turn, regulates several
genes encoding proteins involved in various metabolic
pathways. These proteins include OppA and DppA, the
periplasmic binding proteins for the oligopeptide and
dipeptide transport systems, respectively, as well as SstT,
the primary serine transport protein (Urbanowski et al.,
2000) (S. C. Pulvermacher and others, unpublished
results). The GcvA and GcvR proteins regulate both the
gcvTHP operon and the gcvB gene (Stauffer & Stauffer,
1994; Urbanowski et al., 2000). Both gcvTHP and gcvB are
activated by GcvA in response to glycine and repressed by
GcvA and GcvR together in the absence of glycine; this
repression is enhanced in the presence of purines (Stauffer
& Stauffer, 1994; Urbanowski et al., 2000; Wilson et al.,
1993). Because expression of gcvB follows a pattern of
regulation qualitatively similar to the gcvTHP operon, we
hypothesize that GcvB plays a role in an integrated cellular
response to regulate other genes in conjunction with the
glycine cleavage enzyme system.
Abbreviations: sRNA, small non-translated regulatory RNA; WT, wildtype.
023432 G 2009 SGM
Printed in Great Britain
The E. coli chromosome encodes .80 sRNAs that function
as regulators in different cellular processes (Argaman et al.,
2001; Gottesman, 2004; Hershberg et al., 2003; Wagner &
Flardh, 2002; Wassarman et al., 1999, 2001; Wassarman,
2002). Many of the sRNAs regulate expression of target
genes post-transcriptionally by complementary base pairing with the target mRNAs in trans. A common
characteristic of these sRNAs is the requirement of the
RNA chaperone protein, Hfq (Zhang et al., 2003). Hfq is a
small, abundant RNA binding protein first described in
E. coli as host factor I (HF-I), and was found to be
necessary for replication of the RNA phage Qb (Franze de
Fernandez et al., 1968); further analyses have shown that
Hfq homologues are found in both Gram-positive and
Gram-negative bacteria (Moller et al., 2002; Zhang et al.,
2002). Hfq is a pleiotropic regulator that controls the
expression of proteins by affecting the translation, stability
or polyadenylation of mRNAs (Gottesman, 2004). Hfq
preferentially binds AU-rich sequences in single-stranded
regions, usually close to structured regions of RNA (Lease
& Belfort, 2000; Moller et al., 2002; Schumacher et al.,
2002; Zhang et al., 1998, 2002), and sRNAs in particular are
targets for Hfq binding (Wassarman et al., 2001). In E. coli
Hfq is known to bind the sRNAs OxyS, DsrA, RprA, RyhB,
MicC, GadY and Spot42 (Chen et al., 2004; Masse &
Gottesman, 2002; Moller et al., 2002; Opdyke et al., 2004;
Sledjeski et al., 2001; Vecerek et al., 2003; Wassarman et al.,
2001; Zhang et al., 1998). Binding of Hfq to sRNAs is
115
S. C. Pulvermacher, L. T. Stauffer and G. V. Stauffer
predicted to change the secondary structure of some
sRNAs, thereby affecting the processing or intracellular
stability of the sRNA transcript (Moller et al., 2002; Zhang
et al., 2002). Additionally, binding of Hfq to the sRNA,
target mRNA, or both the sRNA and target mRNA may
function to facilitate RNA–RNA interactions (Moller et al.,
2002; Zhang et al., 2002). Therefore, it has been proposed
that enhancement of RNA–RNA pairing is a general
function of Hfq (Vecerek et al., 2003; Zhang et al., 2002).
Previously we have shown that GcvB regulates both DppA
and OppA in LB (Pulvermacher et al., 2008; Urbanowski et
al., 2000), and microarray analysis has indicated that oppA,
dppA and sstT mRNAs are negatively regulated by Hfq
(Guisbert et al., 2007). Hfq has been reported to bind GcvB
(Zhang et al., 2003), and Hfq has been shown to have a
stabilizing effect on GcvB in Salmonella (Sharma et al.,
2007). Thus, we tested whether Hfq is required for GcvB to
regulate dppA : : lacZ and oppA : : phoA in E. coli, as well as
the effect of Hfq on GcvB stability. Our results demonstrate
that repression of dppA : : lacZ and oppA : : phoA translational fusions is dependent on Hfq. Furthermore, our
results suggest that Hfq functions to stabilize GcvB. These
results suggest that an Hfq–GcvB interaction is necessary
for GcvB-mediated regulation of these genes as well as to
facilitate RNA–RNA pairing of GcvB with its target
mRNAs.
METHODS
Bacterial strains, plasmids and phages. The E. coli strains and
plasmids used in this study are listed in Table 1. Strains GS1148 and
GS1149 were constructed by transduction. P1clr phage grown on the
hfq-1 : : VCMR strain GS081 was used to transduce strains GS162 and
GS1132 to chloramphenicol-resistant (CMR), generating strains
GS1148 and GS1149, respectively. The ldppA : : lacZ and
loppA : : phoA translational fusion phages have been described
previously (Pulvermacher et al., 2008; Urbanowski et al., 2000). The
lgcvB : : lacZ transcriptional fusion phage includes 150 bp upstream of
the gcvB transcription initiation site to bp +50 within gcvB fused to
translationally competent lacZYA genes in phage lgt2 (Urbanowski et
al., 2000). These phages were used to lysogenize various E. coli host
strains as described previously (Urbanowski & Stauffer, 1986). Each
lysogen was tested to ensure that it carried a single copy of the l
chromosome by infection with lcI90c17 (Shimada et al., 1972). All
lysogens were grown at 30 uC since all fusion phages carry the lcI857
mutation, resulting in a temperature-sensitive lcI repressor
(Panasenko et al., 1977).
Plasmid pGS609, carrying the E. coli hfq gene, was constructed as
follows. In a PCR, HindIII primer 1 (59-CCCCCAAGCTTCCGCCTGCTCACC) hybridized to a region beginning 992 bp upstream of
the Hfq translation start site and HindIII primer 2 (59-CCCCCAAGCTTGATGGGGCCGACGGCCTCCGGT) hybridized to a region
beginning 109 bp downstream of the Hfq translation stop codon.
PCR amplification was carried out using Vent DNA polymerase (New
England Biolabs). Cycling conditions were as follows: 1 min 30 s
denaturation at 95 uC, 1 min annealing at 55 uC, 2 min extension at
72 uC for 30 cycles. The 1409 bp PCR-generated HindIII fragment
was cloned into the HindIII site in plasmid pACYC177 (Chang &
Cohen, 1978), and this was verified by DNA sequence analysis at the
DNA Core Facility of the University of Iowa.
Media. The complex medium used was Luria–Bertani broth (LB)
(Miller, 1972). The minimal medium used was the salts of Vogel &
Bonner (1956) supplemented with 0.4 % (w/v) glucose (GM). Agar
was added at 1.5 % (w/v) to make solid media. Supplements were
added at the following concentrations (mg ml21): ampicillin (AP), 50;
chloramphenicol (CM), 20; glycine, 300; inosine, 50%; phenylalanine,
50; rifampicin, 250; thiamine, 1. GM media were always supplemented with phenylalanine and thiamine, since all strains carry the
pheA905 and thi mutations.
Enzyme assays. b-Galactosidase assays were performed on midexponential-phase cells (OD600 ~0.5) as described by Miller (1972)
using the chloroform/SDS lysis procedure. Alkaline phosphatase
assays were performed on mid-exponential-phase cells (OD600 ~0.5)
as described by Brickman & Beckwith (1975). Results are the averages
of two or more assays with each sample done in triplicate.
RNA extraction and Northern blot analysis. For Northern blots, E.
coli strains GS162 [wild-type (WT)], GS1148 (Dhfq), and GS1148
[pGS609] [Dhfq (phfq++)] (Table 1) were grown in 40 ml LB or
GM+glycine (AP was added to the transformant) to the midexponential phase of growth. Rifampicin (250 mg ml21) was added to
Table 1. Strains and plasmids
Strain or plasmid
Strains*
GS081
GS162
GS1132
GS1144
GS1148
GS1149
Plasmids
pGS341
pGS571
pGS594
pGS609
Relevant genotype
Source or reference
MC4100 hfq-1 : : V(CMr)
WT
D(gcvA gcvB) : : VaadA (referred to as DgcvAB)
D(gcvB) : : VCMr (referred to as DgcvB)
hfq-1 : : VCMr (referred to as Dhfq)
D(gcvA gcvB) : : VaadA hfq-1 : : VCMr (referred to as DgcvAB Dhfq)
G. Storz
This laboratory
Urbanowski et al. (2000)
Pulvermacher et al. (2008)
This study
This study
Single-copy vector+gcvA (gcvA+)
Multicopy vector+WT gcvB (gcvB++)
Single-copy vector+WT gcvB (gcvB+)
Multicopy vector+hfq (hfq++)
Jourdan & Stauffer (1998)
This laboratory
This laboratory
This study
*All strains, except GS081, also carry the pheA905 thi araD129 rpsL150 relA1 deoC1 flbB5301 ptsF25 and rbsR mutations.
116
Microbiology 155
GcvB and Hfq regulation of oppA and dppA mRNAs
each sample, and 5 ml cells was taken immediately (0 min) and at 1,
2, 3, 5 and 10 min after the addition of rifampicin; 5 ml GS162 (WT)
cells was also taken immediately (0 min) and at 2, 10, 20 and 30 min
after the addition of rifampicin. In addition, GS162 (WT) cells were
grown in LB, GM+glycine, GM or GM+inosine medium to the
mid-exponential phase of growth, and strains GS162 (WT), GS1144
(DgcvB), GS1144 [pGS594] [DgcvB (pgcvB+)] and GS1148 (Dhfq) were
grown in LB to the mid-exponential phase of growth. Lastly, GS162
(WT) cells were also grown in LB to exponential phase (OD600 0.19
and 0.47), early stationary phase (OD600 1.0), mid-stationary phase
(OD600 2.0 and 2.9), and late stationary phase (OD600 4.2), or in
GM+glycine medium to exponential phase (OD600 0.1, 0.25 and 0.7),
early stationary phase (OD600 1.4), mid-stationary phase (OD600 2.3)
and late stationary phase (OD600 4.6). All cells were added to one-fifth
volume of stop solution [95 % (v/v) ethanol/5 % (v/v) acidic phenol]
and centrifuged for 5 min at 2000 g, and the bacterial pellets were
frozen at 280 uC. RNA was extracted by phenol extraction (Ledeboer
et al., 2006), and the RNA concentration measured by reading the
absorbance of a 1 ml sample using a NanoDrop ND-1000 spectrophotometer. RNA was checked for degradation by running 10 mg of
each sample on a 1.2 % (w/v) agarose gel and staining with ethidium
bromide. Either 8 % denaturing polyacrylamide gels or 1.2 %
formaldehyde agarose gels were run, typically with 10 mg total RNA,
and transferred to a positively charged nylon membrane (Roche). For
the Northern blot comparing GS162 (WT) strains grown in LB,
GM+glycine, GM or GM+inosine, 5 mg total RNA from the LB
sample and 10 mg total RNA from the GM+glycine, GM and
GM+inosine samples were run on an 8 % denaturing polyacrylamide
gel. For the Northern blot comparing GS162 (WT), GS1144 (DgcvB),
GS1144 [pGS594] [DgcvB (pgcvB+)] and GS1148 (Dhfq) strains grown
in LB, 5 mg total RNA was loaded on an 8 % denaturing
polyacrylamide gel. The blots were hybridized with a DNA probe
specific to bases +1 to +134 of the E. coli gcvB gene that was DIGlabelled using the PCR-DIG Probe Synthesis kit (Roche).
Hybridization was performed at 42 uC as described by Engler-Blum
et al. (1993), and the membrane was either exposed to film overnight
or imaged using the Fujifilm LAS-1000 camera and Intelligent Dark
Box. This camera has a dynamic range of signal detection for
chemiluminescence of over four orders of magnitude. The membrane
was subsequently stripped using 0.1 % (w/v) SDS/26 SSC (sodium
chloride sodium citrate buffer), heated to ~95 uC and reprobed using
a DIG-labelled DNA probe specific for 5S RNA from bases +2 to
+112. Quantification of RNA bands was performed using Image
Gauge 3.12 software. The GcvB value was normalized to the 5S signal
from the same lane of the same blot; the resulting ratio was used for
further comparisons. The half-life for GcvB was determined by taking
the GcvB/5S ratio from time point zero and dividing by two. This
value was then compared with the other GcvB/5S ratio values for each
time point of that respective strain on the same blot. This analysis was
performed for each sample in Figs 3 and 6.
DNA manipulation. The procedures for restriction enzyme digestion,
etc. were as described in Sambrook et al. (1989). Plasmid DNA
purification was performed using Qiagen Miniprep kits.
Materials. Restriction enzymes and other DNA modifying enzymes
were from New England Biolabs.
RESULTS AND DISCUSSION
Hfq is required for GcvB repression of
dppA : : lacZ and oppA : : phoA
Hfq is an RNA binding protein with structural similarities
to eukaryotic Sm proteins (Moller et al., 2002; Zhang et al.,
http://mic.sgmjournals.org
2002). Many sRNAs that regulate by RNA–RNA base
pairing require Hfq (Sledjeski et al., 2001; Zhang et al.,
1998), and thus far there is a correlation between sRNAs
that act by means of RNA–RNA base pairing and the
requirement for Hfq (Masse & Gottesman, 2002). Recently,
it has been shown that GcvB interacts with Hfq in E. coli
(Zhang et al., 2003). To confirm a role for Hfq in
regulation by GcvB, we tested whether Hfq affects
dppA : : lacZ and oppA : : phoA expression. WT, DgcvAB,
Dhfq and DgcvAB Dhfq strains lysogenized with either
ldppA : : lacZ or loppA : : phoA were grown in LB to midexponential phase and assayed for either b-galactosidase or
alkaline phosphatase activity. The DgcvAB strain was used
instead of the DgcvB strain because the DgcvAB strain
carries a spectinomycin-resistance cassette, while DgcvB
carries a CM-resistance cassette. The spectinomycin
cassette in DgcvAB allowed for the transduction of the
hfq-1 : : CMr allele into the DgcvAB strain to generate the
DgcvAB Dhfq double mutant. Both DgcvAB ldppA : : lacZ
and DgcvB ldppA : : lacZ produced similar b-galactosidase
activity levels and DgcvAB loppA : : phoA and DgcvB
loppA : : phoA produced similar alkaline phosphatase
activity, suggesting that both strains are similarly regulated
(data not shown). Expression of dppA : : lacZ in the WT
lysogen was 10-fold lower than expression in the DgcvAB
lysogen (Table 2, lines 1 and 2). We observed a similar fold
difference in ldppA : : lacZ b-galactosidase activity between
the WT and the DgcvB lysogens (data not shown). The
deletion of hfq in the Dhfq ldppA : : lacZ lysogen resulted in
constitutive b-galactosidase expression (Table 2, line 3).
Expression of dppA : : lacZ in the DgcvAB Dhfq double
mutant was slightly elevated compared with either single
mutant (Table 2, compare lines 2 and 3 with line 4).
Although the levels in the double mutant are not additive,
suggesting that GcvB and Hfq are part of the same
regulatory mechanism, the higher levels in the double
mutant are consistent (.10 independent assays). It is
possible that an additional regulatory factor is involved in
Hfq-mediated repression of dppA : : lacZ, or that Hfq
regulation of dppA : : lacZ is not entirely GcvB-dependent.
Expression of oppA : : phoA in the WT lysogen was fivefold
lower than in the DgcvAB lysogen (Table 2, lines 8 and 9).
Again, the loss of hfq in the Dhfq loppA : : phoA lysogen
resulted in constitutive alkaline phosphatase expression
(Table 2, line 10). Expression of oppA : : phoA in the
DgcvAB Dhfq double mutant was not significantly different
from that in either single mutant (Table 2, compare lines 9
and 10 with line 11). Additionally, in Salmonella, elevated
OppA protein levels have been observed in both DgcvB and
Dhfq strains with no additional pronounced increase in
OppA protein levels in a DgcvB Dhfq double deletion
(Sharma et al., 2007). The results suggest that GcvB and
Hfq are part of the same regulatory mechanism for
oppA : : phoA regulation. As a control, the DgcvAB Dhfq
double mutant lysogenized with either ldppA : : lacZ or
loppA : : phoA was transformed with a single-copy plasmid
carrying gcvA and assayed for either b-galactosidase or
alkaline phosphatase activity. Because both gcvA and gcvB
117
S. C. Pulvermacher, L. T. Stauffer and G. V. Stauffer
Table 2. Effects of Hfq on dppA : : lacZ, gcvB : : lacZ and oppA : : phoA expression
Line
Lysogen*
Relevant genotype
1
2
3
4
5
6
7
GS162ldppA : : lacZ
GS1132ldppA : : lacZ
GS1148ldppA : : lacZ
GS1149ldppA : : lacZ
GS1149ldppA : : lacZ [pGS341]
GS162lgcvB : : lacZ
GS1148lgcvB : : lacZ
WT
DgcvAB
Dhfq
DgcvAB Dhfq
DgcvAB Dhfq [pgcvA+]
WT
Dhfq
8
9
10
11
12
GS162loppA : : phoA
GS1132loppA : : phoA
GS1148loppA : : phoA
GS1149loppA : : phoA
GS1149loppA : : phoA [pGS341]
DgcvAB
Dhfq
DgcvAB Dhfq
DgcvAB Dhfq [pgcvA+]
WT
b-Galactosidase activity
62±12
604±3
994±1
1155±139
1383±96
276±23
358±7
Alkaline phosphatase activity
16±1
77±1
78±4
71±1
77±10
*Lysogens were grown in LB, LB+CM or LB+CM+AP to OD600 ~0.5, harvested, and assayed for either bgalactosidase activity (Miller, 1972) or alkaline phosphatase activity (Brickman & Beckwith, 1975).
are deleted in strain GS1149, we wanted to be certain that
the loss of GcvA is not responsible for the observed altered
regulation. Expression of dppA : : lacZ and oppA : : phoA in
the DgcvAB Dhfq double mutant transformed with plasmid
pgcvA+ did not show an appreciable difference in bgalactosidase or alkaline phosphatase activity compared
with either untransformed lysogen (Table 2, compare lines
4 and 5 and lines 11 and 12). These results are consistent
with Hfq and GcvB functioning in the same regulatory
mechanism to regulate both dppA : : lacZ and oppA : : phoA
expression.
Hfq does not regulate gcvB transcription
It is possible that the hfq gene is required for gcvB
expression. Thus, the failure to observe repression of
dppA : : lacZ and oppA : : phoA in the Dhfq strain may be due
to decreased GcvB levels. Therefore, the WT strain was
lysogenized with the lgcvB : : lacZ phage and transduced to
CMr with P1 phage grown on GS081 and was designated
GS1148 lgcvB : : lacZ. The WT lgcvB : : lacZ and Dhfq
lgcvB : : lacZ lysogens were grown in LB to mid-exponential
phase and assayed for b-galactosidase activity. The deletion
of hfq had no significant effect on gcvB : : lacZ transcription
(Table 2, compare lines 6 and 7). Since Hfq does not
control gcvB transcription, the results suggest that Hfq
likely either plays a direct role in GcvB-mediated repression
of dppA : : lacZ and oppA : : phoA or functions in increasing
the stability of GcvB.
Hfq likely affects GcvB–mRNA pairing
Hfq, a general RNA chaperone, has been shown to directly
stimulate RNA–RNA pairing (Moller et al., 2002; Zhang et
al., 2002) and to stabilize sRNAs (Masse & Gottesman,
2002; Moller et al., 2002; Zhang et al., 2003). We identified
118
a region in the gcvB sequence complementary to both the
dppA and oppA mRNAs near their respective ribosome
binding sites (Pulvermacher et al., 2008). For dppA, the
complementary region is upstream of the translation start
site and for oppA the complementary region overlaps the
translation start site. This suggests that GcvB–mRNA
pairing is a mechanism for GcvB-dependent regulation of
dppA and oppA mRNA. Previous mutation analysis of gcvB,
dppA and oppA suggests that RNA–RNA pairing is likely
part of the mechanism that GcvB uses to control dppA and
oppA expression (Pulvermacher et al., 2008). In addition,
in Salmonella, GcvB has been shown to base pair with dppA
and oppA mRNAs (Sharma et al., 2007). If the role of Hfq
is to facilitate GcvB–mRNA pairing, we predicted that
overexpression of GcvB in a WT strain would increase
repression of dppA : : lacZ. However, overexpression of
GcvB in a Dhfq mutant should have little effect on
dppA : : lacZ expression if Hfq is required for the GcvB–
dppA mRNA interaction. Overexpression of GcvB from the
multicopy plasmid pGS571 carrying WT gcvB (pgcvB++)
in the WT ldppA : : lacZ lysogen resulted in an additional
8.5-fold repression of dppA : : lacZ compared with the WT
lysogen (Table 3, compare lines 1 and 3). However,
overexpression of GcvB in the Dhfq ldppA : : lacZ
[pgcvB++] lysogen only resulted in a 1.5-fold repression
of dppA : : lacZ compared with the Dhfq ldppA : : lacZ
lysogen (Table 3, compare lines 4 and 5). Thus, Hfq is
likely needed to facilitate pairing between GcvB and dppA
mRNA, and the results are consistent with GcvB and Hfq
being part of the same regulatory mechanism. It is possible
that the failure of GcvB to repress dppA : : lacZ in Dhfq
ldppA : : lacZ [pgcvB++] is due in part to decreased stability
of the GcvB RNA. Although Hfq does partially stabilize
GcvB in WT E. coli (see below), reduced GcvB levels seem
an unlikely explanation for loss of repression in Dhfq
ldppA : : lacZ [pgcvB++] for the following reasons.
Microbiology 155
GcvB and Hfq regulation of oppA and dppA mRNAs
Table 3. Effects of overexpression of gcvB on dppA : : lacZ expression in an hfq mutant
Line
1
2
3
4
5
Lysogen*
Relevant genotype
b-Galactosidase activity
GS162ldppA : : lacZ
GS1144ldppA : : lacZ
GS162ldppA : : lacZ [pGS571]
GS1148ldppA : : lacZ
GS1148ldppA : : lacZ [pGS571]
WT
DgcvB
DgcvB [pgcvB++]
Dhfq
Dhfq [pgcvB++]
51±1
604±42
6±1
848±69
583±20
Generation time (min)
46±4
47±4
53±7
63±8
81±8
*The lysogens and transformants were grown in LB, LB+CM or LB+CM+AP to OD600 ~0.5 and assayed for b-galactosidase activity (Miller,
1972).
Northern blot analysis of GcvB expression from WT
[pgcvB++], DgcvB [pgcvB++] and Dhfq [pgcvB++] strains
grown to mid-exponential phase in LB showed roughly
equivalent amounts of GcvB produced by the three strains
(Fig. 1). Additionally, high levels of GcvB are known to
inhibit cell growth (Urbanowski et al., 2000). The presence
of extra GcvB in Dhfq ldppA : : lacZ [pgcvB++] increased
the generation time from 63 min without the plasmid to
81 min in the transformed strain (Table 3, compare lines 4
and 5). This suggests that GcvB was expressed in Dhfq
ldppA : : lacZ [pgcvB++] and at levels high enough to begin
to inhibit cell growth. These results suggest that the
presence of GcvB in a Dhfq strain is unable to restore GcvB
repression of the dppA : : lacZ fusion.
Hfq binding is to increase the stability of sRNAs by
preventing degradation by RNase E (Masse & Gottesman,
2002; Moll et al., 2003; Moller et al., 2002; Zhang et al.,
2003). It has been reported previously that GcvB is a target
for Hfq binding in E. coli (Zhang et al., 2003), and GcvB
has several single-stranded AU tracts that could serve as
Hfq binding sites (Fig. 2). Thus, we used Northern blots to
determine the effect of a Dhfq allele on GcvB stability in E.
coli grown in LB. WT, Dhfq or complemented Dhfq
[phfq++] cultures were grown to mid-exponential phase
in LB, and rifampicin was added to terminate transcription. GcvB RNA stability was monitored over 10 min (see
Methods). For strains grown in LB, GcvB RNA was stable
through 10 min in the WT and the complemented Dhfq
Effects of Hfq on GcvB stability in LB
Hfq binds RNA with short single-stranded AU-rich
sequences preceded or followed by stem–loop structures
(Brescia et al., 2003; Moller et al., 2002; Zhang et al., 2002).
These sequences are often sites for RNase E degradation
(Moll et al., 2003). One effect and therefore likely role of
Fig. 1. Northern blot analysis of GcvB expression in the WT
[pgcvB++] [GS162 (pGS571)], DgcvB [pgcvB++] [GS144
(pGS571)] and Dhfq [pgcvB++] [GS1148 (pGS571)] strains. All
strains were grown in LB+AP to OD600 ~0.5. Total cell RNA was
isolated from each strain, and Northern blot analysis was
performed as described in Methods using 5 mg total RNA. Each
membrane was first probed with a DIG-labelled DNA probe for
GcvB. Blots were then stripped and reprobed using a DIGlabelled DNA probe for 5S RNA. Band intensities were quantified
as described in Methods.
http://mic.sgmjournals.org
Fig. 2. Proposed secondary structure of E. coli GcvB using the
mfold prediction program (Zuker, 2003). Predicted stem–loops are
lettered A–F and are predicted to be conserved among E. coli,
Shigella dysenteriae, Yersinia pestis, Haemophilus influenzae and
Vibrio cholerae bacteria. Possible Hfq binding regions are
indicated by arrows.
119
S. C. Pulvermacher, L. T. Stauffer and G. V. Stauffer
Fig. 3. Northern blot analysis of GcvB stability from (a) WT
(GS162), Dhfq (GS1148) and Dhfq [phfq++] [GS1148
(pGS609)] strains and (b) the WT (GS162) strain. All strains
were grown in LB to OD600 ~0.5, and rifampicin (250 mg ml”1)
was added to stop transcription. Aliquots of cells were taken (a) at
0, 1, 2, 3, 5 and 10 min after the addition of rifampicin or (b) at 0,
2, 10, 20 and 30 min after the addition of rifampicin. Total cell RNA
was isolated from each strain at each time point, Northern blot
analysis was performed as described in Methods, and each
membrane was probed sequentially with DIG-labelled DNA probes
specific for GcvB or 5S RNA. The strain from which the RNA was
isolated is indicated to the left of each blot and the probe used is
indicated after the strain name. The time (in minutes) after the
addition of rifampicin is labelled across the top of the blots in (a)
and (b). Band intensities were quantified as described in Methods.
[phfq++] strains, while the half-life of GcvB from the Dhfq
strain was determined to be ~5 min (Fig. 3a). To better
determine the half-life of GcvB in WT cells grown in LB,
another Northern blot was run looking at RNA isolated
from WT cells up to 30 min after the addition of rifampicin.
We determined the half-life of GcvB in WT cells grown in LB
to be ~15 min (Fig. 3b). These data suggest that Hfq does
affect the stability of GcvB in LB-grown cells. Our results
showing that GcvB RNA is stable in E. coli WT cells grown in
LB differ from those for Salmonella GcvB, which has a
reported half-life of 2–3 min in cells grown in rich media
(Sharma et al., 2007). Although a number of genes regulated
by GcvB in E. coli (dppA, oppA and livJ) are also regulated by
GcvB in Salmonella typhimurium (Sharma et al., 2007), it is
possible that other regulatory responses occur in E. coli that
demand higher levels of GcvB and a more stable RNA. A
comparison of the E. coli and S. typhiumurium gcvB
sequences shows ~5 % difference in the DNA sequences
(Pulvermacher et al., 2008), and the mfold program (Zuker,
2003) predicts slightly different secondary structures. Thus,
a difference in half-lives of the two GcvB RNAs is not
unexpected.
120
Analysis using three different gcvB–lacZ transcriptional
fusions indicates that two sizes of GcvB are produced
(Urbanowski et al., 2000). These fusions all have a
common 59 end but contain three different fusion points
downstream with translationally competent lacZYA. These
fusion points occur preceding both predicted terminator
sites t1 and t2, between terminators t1 and t2, or after
terminator t2 (Urbanowski et al., 2000). Results from these
three transcriptional fusions showed that nearly 90 % of the
b-galactosidase activity observed with the fusion occurring
prior to both terminators t1 and t2 is lost when the fusion
point follows terminator t1, indicating that the major site
of transcriptional termination is the t1 site. Analysis also
found that terminator t2 is also functional, but less GcvB of
the 206 nt size is predicted to be produced compared with
the 130 nt size. Despite these in vivo transcriptional fusion
data indicating that two sizes of GcvB are produced
(Urbanowski et al., 2000), we only detected the 206 nt
GcvB RNA by Northern blot analysis. Similarly, GcvB in
Salmonella also appears to have two potential termination
sites, but only one band is detected by Northern analysis,
corresponding to a ~200 nt full-length GcvB sRNA
(Sharma et al., 2007). Whether the 130 nt RNA in E. coli
is too unstable to be detected by Northern blot analysis or
whether the terminator 1 site is not functional is currently
unclear.
Expression of GcvB in different strains grown in
LB
We also determined how much GcvB RNA is produced in
various strains grown in LB. A Northern blot was
performed using WT, DgcvB, DgcvB [pgcvB+] and Dhfq
strains grown in LB to mid-exponential phase. We
observed the highest levels of GcvB RNA in either the
WT or the complemented DgcvB [pgcvB+] strain (Fig. 4,
lanes 1 and 3). There was approximately threefold less
Fig. 4. Northern blot comparison of GcvB RNA levels from WT
(GS162), DgcvB (GS1144), DgcvB [pgcvB+] [GS1144
(pGS594)] and Dhfq (GS1148) strains grown in LB. An 8 %
denaturing polyacrylamide gel was run with 5 mg total RNA from
each strain. The strain from which the RNA was isolated is
indicated at the top of the blot. Blots were probed with DIGlabelled DNA probes specific for GcvB or 5S RNA. Band
intensities were quantified as described in Methods.
Microbiology 155
GcvB and Hfq regulation of oppA and dppA mRNAs
GcvB in the Dhfq strain compared with the WT and the
complemented DgcvB [pgcvB+] strains, confirming that a
function of Hfq is to stabilize GcvB RNA (Fig. 4, compare
lane 4 with lanes 1 and 3). As expected, no GcvB RNA was
detected from the DgcvB strain (Fig. 4, lane 2). The ability
to detect significant levels of GcvB in the Dhfq strain, and
the observation that there is a loss of regulation of both
dppA : : lacZ and oppA : : phoA in the Dhfq strain (Table 2)
support the hypothesis that Hfq has a function in addition
to stabilizing GcvB. Previous results (Pulvermacher et al.,
2008), as well as studies in Salmonella (Sharma et al., 2007),
suggest that this additional function is to facilitate GcvB–
target mRNA pairing.
Expression of GcvB in WT cells grown in different
media
Earlier analysis of the GcvB sRNA in Salmonella has found
that GcvB is detectable only in LB and not in minimal
media (Sharma et al., 2007). To determine whether GcvB is
expressed similarly in E. coli we performed a Northern blot
comparing GcvB expression among WT cells in the midexponential phase of growth grown in LB, GM,
GM+glycine and GM+inosine media. WT cells grown
in LB produced the most GcvB with approximately
fourfold less GcvB detected in GM+glycine-grown cells
(Fig. 5, compare lanes 1 and 2). We did not observe GcvB
when WT cells were grown in GM or GM+inosine (Fig. 5,
lanes 3 and 4). This result is not entirely surprising, since
gcvB is activated by the GcvA protein, and GcvA is unable
to fully activate gcvB expression in GM and GM+inosine
media (Urbanowski et al., 2000). This inability to detect
GcvB RNA in GM and GM+inosine media may explain
why GcvB was not detected when WT Salmonella was
grown in minimal media lacking glycine (Sharma et al.,
Fig. 5. Northern blot comparison of GcvB RNA levels from
different media. An 8 % denaturing polyacrylamide gel was run with
the following total RNA: lane 1, 5 mg total RNA isolated from WT
(GS162) cells grown in LB; lanes 2–4, 10 mg total RNA isolated
from WT (GS162) cells grown in GM+glycine, GM and
GM+inosine, respectively. The strain from which the RNA was
isolated is shown at the top of the blot, and the respective medium
in which each strain was grown is given below the blot. The blot
was probed using DIG-labelled DNA probes specific for GcvB or
5S RNA. Band intensities were quantified as described in
Methods.
http://mic.sgmjournals.org
2007). We hypothesize that GcvB regulates other genes in
defined media in conjunction with the glycine cleavage
system. If true, these target genes likely respond to lower
levels of GcvB than dppA and oppA.
Effects of Hfq on GcvB stability in GM+glycine
Negative regulation of dppA : : lacZ and oppA : : phoA occurs
in LB, but not in GM+glycine-grown cells (Urbanowski et
al., 2000). However, we observed approximately eightfold
higher b-galactosidase levels from a WT lgcvB : : lacZ
lysogen grown in GM+glycine compared with GM and
~25-fold higher levels in GM+glycine compared with
GM+inosine (Urbanowski et al., 2000), suggesting that
GcvB is made and has regulatory targets other than dppA
and oppA in GM+glycine. We have shown that GcvB is
detectable in GM+glycine media. Therefore, we analysed
the stability of GcvB from cells grown in GM+glycine to
determine whether Hfq also functions to stabilize GcvB in
defined media. When WT, Dhfq and complemented Dhfq
[phfq++] strains were grown in GM+glycine, we determined the half-life of GcvB RNA to be ~3 min in WT and
complemented Dhfq [phfq++] cells, while the half-life of
GcvB in Dhfq was less than 1 min (Fig. 6). Thus, the halflife of GcvB in WT and Dhfq cells is much lower in
GM+glycine compared with that of the same strains
grown in LB. This result likely explains why we did not
observe negative regulation of dppA : : lacZ or oppA : : phoA
in GM+glycine, as the GcvB levels may be too low to allow
normal regulation. However, it is interesting that we were
Fig. 6. Northern blot analysis of GcvB stability from WT (GS162),
Dhfq (GS1148) and Dhfq [phfq++] [GS1148 (pGS609)] strains.
All strains were grown in GM+glycine to OD600 ~0.5, and
rifampicin (250 mg ml”1) was added to stop transcription. Aliquots
of cells were taken at 0, 1, 2, 3, 5 and 10 min after the addition of
rifampicin. Total cell RNA was isolated from each strain at each
time point, Northern blot analysis was performed as described in
Methods, and each membrane was probed with a DIG-labelled
DNA probe specific for either GcvB or 5S RNA. The strain from
which the RNA was isolated is indicated at the left along with the
respective probe used. The time (in minutes) after the addition of
rifampicin is indicated across the top. Band intensities were
quantified as described in Methods.
121
S. C. Pulvermacher, L. T. Stauffer and G. V. Stauffer
able to show that GcvB is produced in both WT and Dhfq
cells grown in GM+glycine. Based on our results, Hfq
functions to stabilize GcvB in cells grown in both LB and
GM+glycine.
Timing of GcvB synthesis
We determined when GcvB is expressed in cells to help
determine potential regulatory targets of GcvB. WT cells
were grown in either LB or GM+glycine medium, RNA
was harvested from various stages of growth, and Northern
blots were performed. In LB-grown cells, GcvB RNA was
detected from early exponential phase through early
stationary phase (Fig. 7a, lanes 1–5), with the highest
levels detected at the mid-exponential phase (Fig. 7a, lane
2). Salmonella GcvB is also detected only from cells grown
to exponential phase in LB (Sharma et al., 2007). When
WT E. coli cells were grown in GM+glycine, a similar
GcvB expression pattern was observed, with GcvB RNA
detected from early exponential phase through early
stationary phase (Fig. 7b, lanes 1–4). No GcvB was detected
under either growth condition when RNA was isolated
from cells that had entered stationary phase (Fig. 7a, lane 6;
Fig. 7b, lanes 5 and 6). The results suggest that GcvB
regulates targets only during the exponential and early
stationary phases of growth. Additionally, the level of Hfq
has been shown to decrease gradually during the transition
from exponential phase to stationary phase (Ali Azam et al.,
1999). These findings suggest that Hfq is maintained at a
higher level during the exponential phase because it plays a
role in regulation of targets during this growth stage. It will
be interesting to determine whether GcvB has additional
regulatory targets in the exponential phase in LB- or
GM+glycine-grown cells and whether all GcvB regulatory
targets are dependent upon Hfq for regulation.
ACKNOWLEDGEMENTS
We thank Dr G. Storz (National Institutes of Health, Bethesda, MD)
for strain GS081. This work was supported by Public Health Service
Grant GM069506 from the National Institute of General Medical
Sciences.
REFERENCES
Ali Azam, T., Iwata, A., Nishimura, A., Ueda, S. & Ishihama, A. (1999).
Growth phase-dependent variation in protein composition of the
Escherichia coli nucleoid. J Bacteriol 181, 6361–6370.
Argaman, L., Hershberg, R., Vogel, J., Bejerano, G., Wagner, E. G.,
Margalit, H. & Altuvia, S. (2001). Novel small RNA-encoding genes in
the intergenic regions of Escherichia coli. Curr Biol 11, 941–950.
Brescia, C. C., Mikulecky, P. J., Feig, A. L. & Sledjeski, D. D. (2003).
Identification of the Hfq-binding site on DsrA RNA: Hfq binds
without altering DsrA secondary structure. RNA 9, 33–43.
Brickman, E. & Beckwith, J. (1975). Analysis of the regulation of
Escherichia coli alkaline phosphatase synthesis using deletions and Q80
transducing phages. J Mol Biol 96, 307–316.
Chang, A. C. & Cohen, S. N. (1978). Construction and characteriza-
tion of amplifiable multicopy DNA cloning vehicles derived from the
P15A cryptic miniplasmid. J Bacteriol 134, 1141–1156.
Chen, S., Zhang, A., Blyn, L. B. & Storz, G. (2004). MicC, a second
small-RNA regulator of Omp protein expression in Escherichia coli. J
Bacteriol 186, 6689–6697.
Engler-Blum, G., Meier, M., Frank, J. & Muller, G. A. (1993).
Reduction of background problems in nonradioactive northern and
Southern blot analyses enables higher sensitivity than 32P-based
hybridizations. Anal Biochem 210, 235–244.
Franze de Fernandez, M. T., Eoyang, L. & August, J. T. (1968). Factor
fraction required for the synthesis of bacteriophage Qb-RNA. Nature
219, 588–590.
Gottesman, S. (2004). The small RNA regulators of Escherichia coli:
roles and mechanisms. Annu Rev Microbiol 58, 303–328.
Guisbert, E., Rhodius, V. A., Ahuja, N., Witkin, E. & Gross, C. A.
(2007). Hfq modulates the sE-mediated envelope stress response and
the s32-mediated cytoplasmic stress response in Escherichia coli.
J Bacteriol 189, 1963–1973.
Hershberg, R., Altuvia, S. & Margalit, H. (2003). A survey of small
Fig. 7. Determination of when GcvB is produced during growth of
WT cells. WT (GS162) cells were grown in either LB (a) or
GM+glycine (b) to the respective OD600 listed above each blot.
Aliquots of cells were taken at each OD600 listed, and total cell
RNA was isolated. Northern blot analysis was performed as
described in Methods, and blots were probed with a DIG-labelled
DNA probe specific for GcvB or 5S RNA.
122
RNA-encoding genes in Escherichia coli. Nucleic Acids Res 31, 1813–
1820.
Jourdan, A. D. & Stauffer, G. V. (1998). Mutational analysis of the
transcriptional regulator GcvA: amino acids important for activation,
repression, and DNA binding. J Bacteriol 180, 4865–4871.
Kikuchi, G. (1973). The glycine cleavage system: composition,
reaction mechanism, and physiological significance. Mol Cell
Biochem 1, 169–187.
Microbiology 155
GcvB and Hfq regulation of oppA and dppA mRNAs
Lease, R. A. & Belfort, M. (2000). A trans-acting RNA as a control
Sledjeski, D. D., Whitman, C. & Zhang, A. (2001). Hfq is necessary for
switch in Escherichia coli: DsrA modulates function by forming
alternative structures. Proc Natl Acad Sci U S A 97, 9919–9924.
regulation by the untranslated RNA DsrA. J Bacteriol 183, 1997–2005.
Stauffer, L. T. & Stauffer, G. V. (1994). Characterization of the gcv
Ledeboer, N. A., Frye, J. G., McClelland, M. & Jones, B. D. (2006).
control region from Escherichia coli. J Bacteriol 176, 6159–6164.
Salmonella enterica serovar Typhimurium requires the Lpf, Pef, and
Tafi fimbriae for biofilm formation on HEp-2 tissue culture cells and
chicken intestinal epithelium. Infect Immun 74, 3156–3169.
Urbanowski, M. L. & Stauffer, G. V. (1986). Autoregulation by tandem
promoters of the Salmonella typhimurium LT2 metJ gene. J Bacteriol
165, 740–745.
Masse, E. & Gottesman, S. (2002). A small RNA regulates the
Urbanowski, M. L., Stauffer, L. T. & Stauffer, G. V. (2000). The gcvB
expression of genes involved in iron metabolism in Escherichia coli.
Proc Natl Acad Sci U S A 99, 4620–4625.
Miller, J. (1972). A Short Course in Bacterial Genetics. Cold Spring
Harbor, NY: Cold Spring Harbor Laboratory.
Moll, I., Afonyushkin, T., Vytvytska, O., Kaberdin, V. R. & Blasi, U.
(2003). Coincident Hfq binding and RNase E cleavage sites on mRNA
and small regulatory RNAs. RNA 9, 1308–1314.
Moller, T., Franch, T., Hojrup, P., Keene, D. R., Bachinger, H. P.,
Brennan, R. G. & Valentin-Hansen, P. (2002). Hfq: a bacterial Sm-like
protein that mediates RNA–RNA interaction. Mol Cell 9, 23–30.
Opdyke, J. A., Kang, J. G. & Storz, G. (2004). GadY, a small-RNA
regulator of acid response genes in Escherichia coli. J Bacteriol 186,
6698–6705.
Panasenko, S. M., Cameron, J. R., Davis, R. W. & Lehman, I. R.
(1977). Five hundredfold overproduction of DNA ligase after
gene encodes a small untranslated RNA involved in expression of the
dipeptide and oligopeptide transport systems in Escherichia coli. Mol
Microbiol 37, 856–868.
Vecerek, B., Moll, I., Afonyushkin, T., Kaberdin, V. & Blasi, U. (2003).
Interaction of the RNA chaperone Hfq with mRNAs: direct and
indirect roles of Hfq in iron metabolism of Escherichia coli. Mol
Microbiol 50, 897–909.
Vogel, H. J. & Bonner, D. M. (1956). Acetylornithinase of Escherichia
coli: partial purification and some properties. J Biol Chem 218, 97–
106.
Wagner, E. G. & Flardh, K. (2002). Antisense RNAs everywhere?
Trends Genet 18, 223–226.
Wassarman, K. M. (2002). Small RNAs in bacteria: diverse regulators
of gene expression in response to environmental changes. Cell 109,
141–144.
induction of a hybrid lambda lysogen constructed in vitro. Science
196, 188–189.
Wassarman, K. M., Zhang, A. & Storz, G. (1999). Small RNAs in
Pulvermacher, S. C., Stauffer, L. T. & Stauffer, G. V. (2008). The role
Wassarman, K. M., Repoila, F., Rosenow, C., Storz, G. & Gottesman, S.
(2001). Identification of novel small RNAs using comparative genomics
of the small regulatory RNA GcvB in GcvB/mRNA posttranscriptional
regulation of oppA and dppA in Escherichia coli. FEMS Microbiol Lett
281, 42–50.
Sambrook, J., Fritsch, E. & Maniatis, T. (1989). Molecular Cloning: a
Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring
Harbor Laboratory.
Escherichia coli. Trends Microbiol 7, 37–45.
and microarrays. Genes Dev 15, 1637–1651.
Wilson, R. L., Stauffer, L. T. & Stauffer, G. V. (1993). Roles of the GcvA
and PurR proteins in negative regulation of the Escherichia coli glycine
cleavage enzyme system. J Bacteriol 175, 5129–5134.
Zhang, A., Altuvia, S., Tiwari, A., Argaman, L., Hengge-Aronis, R. &
Storz, G. (1998). The OxyS regulatory RNA represses rpoS translation
Schumacher, M. A., Pearson, R. F., Moller, T., Valentin-Hansen, P. &
Brennan, R. G. (2002). Structures of the pleiotropic translational
and binds the Hfq (HF-I) protein. EMBO J 17, 6061–6068.
regulator Hfq and an Hfq-RNA complex: a bacterial Sm-like protein.
EMBO J 21, 3546–3556.
Zhang, A., Wassarman, K. M., Ortega, J., Steven, A. C. & Storz, G.
(2002). The Sm-like Hfq protein increases OxyS RNA interaction with
Sharma, C. M., Darfeuille, F., Plantinga, T. H. & Vogel, J. (2007). A
target mRNAs. Mol Cell 9, 11–22.
small RNA regulates multiple ABC transporter mRNAs by targeting
C/A-rich elements inside and upstream of ribosome-binding sites.
Genes Dev 21, 2804–2817.
Zhang, A., Wassarman, K. M., Rosenow, C., Tjaden, B. C., Storz, G. &
Gottesman, S. (2003). Global analysis of small RNA and mRNA
Shimada, K., Weisberg, R. A. & Gottesman, M. E. (1972). Prophage
Zuker, M. (2003). Mfold web server for nucleic acid folding and
lambda at unusual chromosomal locations. I. Location of the
secondary attachment sites and the properties of the lysogens. J Mol
Biol 63, 483–503.
hybridization prediction. Nucleic Acids Res 31, 3406–3415.
http://mic.sgmjournals.org
targets of Hfq. Mol Microbiol 50, 1111–1124.
Edited by: J.-H. Roe
123