Molecular Microbiology (1996) 19(2), 231240

Post-transcriptional regulation of CspA expression in

Escherichia coli
Anna Brandi, Paola Pietroni, Claudio O. Gualerzi* and
Cynthia L. Pon
Laboratory of Genetics, Department of Biology,
University of Camerino, I-62032 Camerino (MC), Italy.

of essential functions such as transcription (nusA), translation (infB ), mRNA degradation ( pnp), recombination
(recA), etc. (for a review see Jones and Inouye, 1994).
Among the cold-shock genes are also hns, the gene
encoding the abundant nucleoid protein H-NS (La Teana
et al., 1991), and cspA, the structural gene for the major
cold-shock protein CspA (also named CS7.4) (Goldstein
et al., 1990), which is homologous to a class of eukaryotic
nucleic acid-binding proteins known as Y box (i.e. the
CCAAT motif)-binding proteins (Wolffe et al., 1992). Protein CspA was found to act as a cold-shock transcriptional
enhancer of the expression of at least some cold-shock
genes, such as hns (La Teana et al., 1991) and gyrA
(Jones et al., 1992). In spite of these recent advancements, several important questions remain open concerning: (i) the existence of additional roles of CspA in the coldshock response; (ii) the relationship between CspA and
the other four homologous proteins of the same family
(CspB, CspC, CspD and CspE) which have recently
been detected (Lee et al., 1994; Yamanaka et al., 1994);
(iii) the molecular mechanism by which this protein stimulates transcription of some (or all?) cold-shock genes; and
(iv) the mechanism responsible for turning on and off the
cspA gene during the early stages of the cold-shock
response.
In this article, we present results which may shed light
on the latter point, suggesting that post-transcriptional
events may play a crucial role in determining the massive
expression of the cspA gene that ensues the lowering of
the temperature from 37 C to 10 C. Furthermore, in
accordance with the suggestion that ribosomes may act
as sensors of heat- and cold shock (VanBogelen and
Neidhardt, 1990), our results provide the first direct in
vitro evidence that the cold-shocked ribosome may participate in the selective expression of a cold-shock gene.

Summary
The Escherichia coli cspA gene, encoding the major
cold-shock protein CspA, was deprived of its natural
promoter and placed in an expression vector under
the control of the inducible k PL promoter. After induction of transcription by thermal inactivation of the k ts
repressor, abundant expression of the product (CspA)
was obtained if the cells were subsequently incubated
at 10 C, but poor expression was obtained if the cells
were incubated at 37 C or 30 C. The reason for this
differential temperature-dependent expression was
investigated and it was found that: (i) the CspA content of the cells decreased more rapidly at 37 C compared to 10 C, regardless of whether transcription was
turned off by addition of rifampicin; (ii) both the chemical and functional half-lives of the cspA transcript
were substantially longer at 10 C compared to 37 C;
(iii) S30 extracts as well as 70S ribosomes prepared
from cold-shocked cells translated CspA mRNA (but
not phage MS2 RNA) more efficiently than equivalent
extracts or ribosomes obtained from control cells
grown at 37 C; and (iv) purified CspA stimulated
CspA mRNA translation. Overall, these results indicate that a selective modification of the cold-shocked
translational apparatus favouring translation of CspA
mRNA, and an increased stability of this mRNA at low
temperature, may play an important role in the induction of cspA expression during cold shock.

Introduction
In Escherichia coli, the expression of a set of genes,
known as the cold-shock regulon, becomes specifically
enhanced or induced de novo during the growth lag
which follows the lowering of the temperature from 37 C
to 10 C (Jones et al., 1987). The genes belonging to the
cold-shock regulon encode proteins involved in a variety

Results

Construction of a CspA hyperproducing strain

The manipulations carried out to deprive E. coli cspA of its

natural promoter and place it under the control of phage
PL promoter (pCspA1) and of phage T7 promoter (pCspA2)
are explained in detail in the Experimental procedures. The
relevant restriction sites and the transcription and translation signals of native cspA and of the two constructs

Received 6 April, 1995; revised 8 August, 1995; accepted 30 August,

1995. *For correspondence. Tel. (737) 40749; Fax (737) 636216.

# 1996 Blackwell Science Ltd

232 A. Brandi, P. Pietroni, C. O. Gualerzi and C. L. Pon

Fig. 1. Schematic representation of the
structure of the constructs containing cspA .
A. The cspA gene as found in the E. coli
chromosome showing the 10 and 35
promoter elements, the transcriptional
startpoint (vertical line with undulating arrow)
and terminator (T), the ShineDalgarno
(SD) sequence, the CspA coding sequence
(dotted bar), and some relevant restriction
sites.
B. The expression vector pCspA1 in which
cspA has been placed under the control of
the PL promoter (shaded arrow). The blunt
ligation between filled-in Fsp I and Acc I sites
resulted in the creation of an Nsi I site. The
thin and thick lines represent DNA sequences
originating from the vector and the
chromosome, respectively.
C. The transcription vector pCspA2 in which
cspA has been placed under the control of
the phage T 7 promoter. All symbols are as
indicated above.

are presented in the scheme of Fig. 1.

Hyperexpression of CspA occurs preferentially at low

temperature

D D

E. coli K12 H1 trp cells were transformed with pCspA1

and CspA hyperexpression experiments were carried
out. After a brief incubation at 42 C to inactivate the ts

repressor, the cells were incubated either at 37 C or at

10 C for increasing lengths of time. As seen from the
Coomassie blue-stained SDSPAGE (Fig. 2), in the cells
transformed with the vector lacking cspA , a low level of
CspA was observed only after 90 and 210 min at 10 C;
this production can be attributed to the cold-shock activation of the chromosomal cspA gene. On the other hand,
abundant expression of CspA was already obtained after

D D

Fig. 2. Hyperexpression of CspA from pCspA1. Two exponentially growing cultures of E. coli K12 H1 trp, one transformed with pCspA1
(cspA+ ) and the other with pPLc2833 (cspA7 ), were shifted from 30 C to 42 C as described in the Experimental procedures. After induction,
half of each culture was incubated at 37 C and the other half at 10 C. Aliquots were withdrawn at the following times, after incubation at the
indicated temperatures, for analysis by SDS(18%) PAGE and staining with Coomassie brilliant blue. Only the relevant portion of the gel
(approximately the lower one-fifth) is presented in the figure. Lanes 13 (after 10, 15 and 20 min); lanes 4 and 7 (after 30 min); lanes 5, 8, 10
and 13 (after 60 min); lanes 6, 9, 11 and 14 (after 90 min); lanes 12 and 15 (after 210 min). Lane S contains two molecular mass markers:
lower band, E. coli IF1 (8.1 kDa); upper band, lysozyme (14.3 kDa). The two lanes indicated by * contain samples which are not related to this
experiment. The identification of CspA in the gel was established using the following criteria: (i) by reference to the molecular weight markers
loaded in lane S, of which only two (IF1 and lysozyme) are visible in the lower portion of the gel shown; (ii) by its absence at 37 C and at
early times after cold shock in control cells harbouring the vector without the insert, and by its appearance in the extracts of the same cells
exposed to long periods of cold shock; and (iii) by immunoblotting with anti-CspA antibodies, a gel run in parallel to that stained.

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

Cold-shock expression of CspA 233

Fig. 3. Stability of CspA and CspA mRNA at 10 C and 37 C.

A. Western blot of CspA present in cells after hyperproduction and incubation at 37 C or 10 C. E. coli K12 H1 trp transformed with pCspA1
were grown in LB as described in the Experimental procedures. After 10 min at 42 C, rifampicin (200 g ml71) was added then one half of the
culture was incubated at 37 C, and the other half was incubated at 10 C. Samples for Western blot analysis were taken after (lane 1) 10 min,
(lane 2) 20 min, (lane 3) 30 min, (lane 4) 60 min, (lane 5) 90 min, and (lane 6) 120 min of incubation at the indicated temperatures, rapidly
mixed with SDS sample buffer, and aliquots corresponding to 2.5 g of total protein, as determined by the Bradford dye-binding method (Bradford, 1976), were subjected to SDS(15%) PAGE. The proteins were electroblotted onto a nitrocellulose membrane, essentially as described
by Sambrook et al. (1989) and the CspA protein was detected with rabbit anti-CspA antiserum using the Blotting Detection Kit for rabbit antibodies (Amersham). Lane C, non-induced control; Lane I, sample taken immediately after incubation at 42 C prior to addition of rifampicin.
Lanes
16, samples taken at the times indicated above.
B. Quantification of the Western blot of (A) by densitometric scanning. The value obtained for the sample I was taken to be 100%.
C. Chemical stability of the CspA mRNA at 10 C and 37 C. Extraction and electrophoresis of the RNA, Northern blotting and hybridization

D D

60 min incubation at 10 C in the cells transformed with the

plasmid containing cspA. In contrast, the amount of CspA
found in the same cells after incubation for 3090 min at
37 C (or 30 C in some experiments) was rather modest.
As, in this experiment, any transcriptional control which
may operate on the natural promoter of cspA is bypassed
in the construct with the promoter, and the transcription
step is identical for all the samples, our results indicate that
the preferential cspA expression at low temperature is due
to a post-transcriptional event.

Differential stability of CspA mRNA and CspA at

10 C and 37 C

In the following experiments, the chemical stability of the

CspA mRNA and of its product, the CspA protein, were
investigated at 10 C and 37 C to determine if a differential rate of decay at the two temperatures could be correlated with the differential efficiency of cspA expression.
For this purpose, cells harbouring pCspA1 were exposed
to 42 C to inactivate the ts repressor and induce cspA
transcription from the PL promoter, and exposed to rifampicin before being divided into two portions which were
then incubated at 37 C and at 10 C, respectively. After dif-

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

234 A. Brandi, P. Pietroni, C. O. Gualerzi and C. L. Pon

Fig. 4. Functional stability of CspA mRNA.

A. Comparison of the functional stability of CspA mRNA at 37 C and 10 C. Transcription of CspA mRNA from the PL promoter was induced
at 42 C. After addition of rifampicin (0.25 mg ml71), the cultures were incubated at 37 C (S) or 10 C (T), and the residual capacity of the
CspA mRNA to direct the synthesis of CspA within 5 min pulses with [35S]-methionine was determined by immunoprecipitation followed by
electrophoresis. Further details are given in the Experimental procedures.
B. Comparison of the functional stability of CspA mRNA and that of bulk mRNAs. The experiment was carried out as described above,
except for the fact that the cells pulse-labelled with methionine at 10 C were lysed, and the radioactivity associated with the electrophoretically
separated CspA band was taken as a measure of the functional stability of CspA mRNA (T), while the total radioactivity associated with all
electrophoretically resolved polypeptides (minus CspA) was taken as a measure of the functional stability of the bulk mRNA (k). Further
details are given in the Experimental procedures.

ferent lengths of time, the cells were harvested, ruptured,

and their total proteins were separated by SDSPAGE.
The proteins were then blotted onto membranes which
were developed with anti-CspA antiserum. As seen from
the Western blot presented in Fig. 3A, and from its densitometric quantification (Fig. 3B), following induction of the
promoter at 42 C (lane I), the amount of CspA increases
substantially compared to the sample of non-induced cells
(lane C) and then decreases at different rates depending
upon the temperature of incubation. Under the conditions
of the experiment, the apparent half-life ( 1/ 2 ) of CspA is
approx. 7 min at 37 C and between five and 10 times
longer at 10 C. It should be noticed that, in this
experiment, transcription had been inhibited by addition
of rifampicin, but translation was allowed to continue
so that the estimated rates of CspA disappearance
reflect the rates of both mRNA and protein degradation.
A faster decay of CspA at 37 C compared to 10 C
was also obtained in an experiment similar to that of
Fig. 3(A and B) in cells not exposed to rifampicin (not
shown).
To compare the chemical stability of CspA mRNA at
37 C and 10 C, in the subsequent experiment, transcription of CspA mRNA was induced at 42 C for 12 min, rifampicin (0.2 mg ml71) was added, and the cells were
transferred to 37 C and 10 C. After various times of incubation at these temperatures, the total RNA was extracted,

subjected to electrophoresis, Northern blotting, hybridization with a radioactive cspA probe, and quantification. As
seen from Fig. 3C, the stability of CspA mRNA is by far
greater (over two orders of magnitude) at 10 C than at
37 C, at which 1/ 2 2 min. However, since small mRNA
truncations that could destroy the translational capacity
of the mRNA might have escaped detection in this experiment, the functional stability of CspA mRNA was compared at 37 C and 10 C. Thus, after induction at 42 C,
transcription of CspA mRNA was stopped by addition of
rifampicin (0.25 mg ml71), and the cells were divided in
two portions which were incubated at 37 C and at 10 C.
Aliquots of these cells were subjected to 5 min pulse
chase with [ 35S]-methionine at different times after stoppage of transcription and the radioactivity associated with
CspA synthesized during this period was detected by
immunoprecipitation. The results of this experiment are
presented in Fig. 4A. As seen from Fig. 4, CspA mRNA
continues to be active in directing the synthesis of its product for a substantially longer time at 10 C than at 37 C.
Because stabilization of the mRNA at the lower temperature may simply reflect a general reduction of enzyme
reaction rates and, therefore, be non-specific, the functional stability of the CspA mRNA at 10 C was compared
to that of the bulk cellular mRNAs. As seen from
Fig. 4B, the capacity of CspA mRNA to direct the synthesis of CspA persists longer than the average capacity

t &

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

Cold-shock expression of CspA 235

Fig. 5. In vitro translation of CspA mRNA.

A. SDSPAGE analysis of the radioactive products (visualized using a BioRad GS-250 Molecular Imager) obtained from 10 l of reaction
mixtures prepared as described in the Experimental procedures. The mixtures contained 15 pmol CspA mRNA, [14C]-valine and 3 l, 5 l or
8 l (lanes 13, respectively) of S30 fractions prepared from E. coli MRE600 cells grown at 37 C to A 620 = 0.8 (A), and after 20 min (B), 40 min
(C) and 90 min (D) cold shock at 10 C. The total protein concentrations of the individual S30 fractions were: A = 43 g ml71, B = 30.4 mg ml71,
C= 36.8 mg ml71, D = 40 mg ml71. The lanes labelled N in each block correspond to mixtures each containing 8 l of the corresponding S30
fractions and to which no mRNA was added. Translation was carried out at 37 C for 30 min.
B. In vitro translation of CspA mRNA (k, l) or MS2 RNA (S, T) by S30 fractions prepared from control cells grown at 37 C (k, S) and cells
after 90 min of cold shock (l, T), each adjusted to contain 23 pmol of 70S ribosomes (as determined by sucrose gradient centrifugation
analysis). The experiment was performed as described in the Experimental procedures with the indicated amounts of template and [35S]methionine as labelled precursor. After 30 min incubation at 37 C, 10 l of each reaction mixture were placed onto Whatman 3MM filter discs
which were processed batchwise by the hot trichloroacetic acid procedure.
C. Effect of CspA on the in vitro translation of CspA mRNA. The experiment was carried out as described above, supplementing with CspA
mRNA (5 pmol), the S30 fractions (each corresponding to 20 pmol 70S) prepared from 37 C control (k) and 90 min cold-shocked (l) cells.
CspA protein, purified as described previously (La Teana et al., 1991), was added to each sample to reach the concentrations indicated in the
abscissa. After 30 min incubation at 37 C, the samples were processed as indicated above.
D. Time-course of CspA mRNA (k, l) or MS2 RNA (S, T) translation by 70S ribosomes prepared from S30 fractions of control cells (k, S)
and 40 min cold-shocked cells (l, T). The experiment was performed as described in the Experimental procedures using 10 pmol of CspA
mRNA or 6.9 pmol of MS2 RNA, 20 pmol of ribosomes and [35S]-methionine as labelled precursor. At the indicated times, 10 l aliquots of
each reaction mix were withdrawn and processed as indicated above.

m m

of the other mRNAs to direct the synthesis of all the other

electrophoretically resolved polypeptides. Taken together,
the results of Figs. 3C and 4, A and B indicate that both
chemical stability and functional stability of CspA mRNA
are increased at 10 C compared to 37 C, and that the
extent of stabilization of CspA mRNA at 10 C exceeds
that displayed by other mRNAs at the same tempera-

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

ture, and must, therefore, be regarded as specific and

not the result of a reduced rate of hydrolysis at low temperature. Thus, the different rates of CspA mRNA
degradation at high and low temperatures can
account for the preferential expression of cspA during
cold shock, even without invoking a faster degradation of
its product at 37 C.

236 A. Brandi, P. Pietroni, C. O. Gualerzi and C. L. Pon

Preferential translation of CspA mRNA by the

cold-shocked translational apparatus
VanBogelen and Neidhardt (1990) have shown that cells
exposed to low dosage of antibiotics of the H-group
type (i.e. puromycin, streptomycin and kanamycin) and
the C-group type (i.e. chloramphenicol, tetracycline,
erythromycin, spiramycin and fusidic acid) preferentially
express heat-shock and cold-shock proteins, respectively.
The two types of antibiotics differ in their mechanism of
action on the ribosome leaving the A-site occupied or
empty and supposedly giving rise to a heat-shock and a
cold-shock conformation. Furthermore, a translational
bias in favour of cold-shock mRNAs by cold-shocked ribosomes could account for the stabilization of CspA mRNA
at the low temperature reported above, as it is well
known that the rates of mRNA translation and degradation are interconnected, albeit in a rather complex
manner (Petersen, 1993). To test this hypothesis, the
cspA gene was placed under the control of the phage T7
promoter (Fig. 1) and transcribed in vitro with the phage
polymerase; the resulting CspA mRNA was tested in
cell-free extracts prepared from control cells growing at
37 C (Fig. 5A, lanes 13, block A) as well as in extracts
prepared from cells subjected to 20, 40, and 90 min cold
shock at 10 C (Fig. 5A, blocks of lanes B, C, and D,
respectively). As seen from the Fig. 5, all cell extracts
directed the synthesis of mainly one protein (corresponding to CspA) that was not seen in the control extracts to
which no CspA mRNA was added (lanes N of blocks A
D). However, the amount of CspA protein synthesized
was not the same in all cases, but increased with time of
cold shock to which the cells had been subjected, being
maximal with the extracts of cells shocked for 90 min at
10 C (Fig. 5A, lanes of block D). This difference occurred
in spite of the fact that all cell extracts had been prepared
under rigorously identical conditions and normalized for
their protein and ribosome content (as determined by
sucrose-gradient centrifugation). The experiment shown
in Fig. 5A was performed using non-dialysed S30 fractions to prevent the possible loss of some potentially relevant small regulatory molecules. Similar results were
obtained, however, when the experiment was repeated
with dialysed extracts and when the precursor amino
acid was changed ([35S]-methionine was used in place of
[14C]-valine). In this experiment, some cell extracts were
also challenged with another mRNA (MS2 RNA) which,
being a non-cold-shock mRNA, served as a control. As
seen from Fig. 5B, CspA mRNA is translated more efficiently by the S30 fraction prepared from cold-shocked
cells compared to the extract from control cells, while
both extracts translated MS2 RNA with similar efficiency.
As there are indications that CspA may bind to its own
mRNA (Newkirk et al., 1994), we investigated the possibi-

lity that CspA itself might be responsible for the increased

translation of CspA mRNA displayed by cell-free extracts
from cold-shocked cells. As seen from the experiment of
Fig. 5C, addition of purified CspA protein to the S30
extracts, at concentrations comparable to those reached
by this protein during cold shock, was indeed found to
stimulate translation of CspA mRNA while having no
effect on the translation of MS2 RNA (not shown). This
stimulation, however, was seen with cell-free extracts of
both control and cold-shocked cells and, in the presence
of CspA, the control extracts reached a plateau level of
translation substantially lower than that attained by the
cold-shocked extract. Thus, while providing evidence for
a direct effect of CspA protein on translation, these findings indicate that the presence of different concentrations
of CspA does not fully explain the different translational
activity of the cell extracts from control and cold-shocked
cells. To determine whether the translational bias in
favour of the cold-shock mRNA is a property of the coldshocked ribosomes, these were prepared from control
and cold-shocked cells, and compared for their translational capacity in a system supplemented with a postribosomal fraction derived from control cells (i.e. cells
growing at 37 C). As seen from Fig. 5D, both types of
ribosomes translate MS2 RNA at approximately the
same rate and with the same efficiency; the ribosomes
from cold-shocked cells, however, were found to translate
the CspA mRNA noticeably better than those from control
cells, indicating that the ribosomes are, at least in part,
responsible for the difference in the translational efficiency of CspA mRNA seen with the S30 fractions.

Discussion
The available data indicate that the major cold-shock protein CspA acts as transcriptional enhancer of cold-shock
genes such as hns (La Teana et al., 1991) and gyrA
(Jones et al., 1992). Consistent with its role of transcriptional enhancer of the other genes, CspA appears to be
the first cold-shock protein to be synthesized and cspA
the gene responding to the highest level to this particular
type of stress. Thus, the question to be answered is how
cspA is turned on in the early stages of the stress response
(Jones et al., 1987; Goldstein et al., 1990). Addressing this
problem, Inouyes laboratory reported that an as yet
unidentified DNA-binding factor, present in minute amounts
in cold-shocked cells, might be responsible for the induction of cspA expression (Tanabe et al., 1992). Treatment
of the cells with chloramphenicol was also found to
induce cspA expression and the results obtained strengthened the opinion that cold-shock induction of this gene
occurs at the transcriptional level (Jiang et al., 1993;
Jones and Inouye, 1994). In these articles it was reported,
however, that the CspA mRNA is extremely unstable at

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

Cold-shock expression of CspA 237

37 C and that its stabilization, as a result of the chloramphenicol treatment, might contribute to cspA activation.
In our laboratory, we have been studying the mechanism of cspA expression and found that this is indeed
transcriptionally regulated (A. Brandi, C. L. Pon, C. O.
Gualerzi, in preparation). In the course of these studies,
however, we obtained evidence that other regulatory
mechanisms are also operating. In the present article,
we provide evidence supporting a post-transcriptional
mechanism of cspA regulation, showing that synthesis of
CspA occurs much more efficiently at 10 C compared to
37 C (or 30 C) under conditions in which a potential
cold-shock transcriptional control is bypassed by placing
the cspA gene under the control of the PL promoter.
Our data actually suggest the existence of three possible
levels of post-transcriptional regulation of cspA expression: (i) translation of mRNA, (ii) stability of the transcript, and (iii) stability of the product. Of these three
potential mechanisms, translational control is the most
interesting for its novelty and for its mechanism. In fact,
unlike most of the other known cases of translational control (Gold, 1988; McCarthy and Gualerzi, 1990), regulation, in this case, seems to be an intrinsic property of the
translational machinery which is modified during cold
shock in such a way that it will preferentially translate
CspA mRNA. Thus, we find that cell-free (S30) extracts
from cold-shocked cells translate CspA mRNA better
than extracts from control (i.e. non cold-shocked) cells,
prepared in the same way and normalized for their ribosome content, while both extracts translate MS2 RNA
with almost equal efficiency. Furthermore, the difference
between the two cell-free extracts in their capacity to sustain CspA mRNA translation did not depend on the nature
of the radioactive precursor amino acid (methionine or
valine) used to detect synthesis, did not disappear upon
dialysis of the S30 fractions and was retained, to a large
extent, by the ribosomal fraction obtained from them. In
an attempt to find an explanation for this phenomenon,
we found that purified protein CspA can stimulate translation of CspA mRNA. Together with the fact that CspA,
whose 3-D structure is characteristic of a protein interacting with single-stranded nucleic acid, can bind to an oligonucleotide corresponding to the 24 nucleotides of the 5 leader region of its mRNA (Schindelin et al., 1993; Schnuchel et al., 1993; Newkirk et al., 1994; Schindelin et al.,
1994), our present finding introduces the interesting perspective that one of the CspA functions may be a selective stimulation of the translation of its own mRNA, and
possibly of other cold-shock mRNAs. As translation initiation requires that the translation-initiation region of the
mRNA is devoid, as much as possible, of secondary
structure (Gualerzi and Pon, 1990; McCarthy and
Gualerzi, 1990), CspA may act in maintaining an unstructured configuration of the translation-initiation region of a

'

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

class of mRNAs, thus favouring their interaction with the

ribosome.
However, since CspA was found to stimulate translation
of its mRNA in extracts from both control and cold-shocked
cells, and because the translational activity of the control
extracts cannot be increased by CspA to the same level
as that of cold-shocked cell extracts, the higher activity
of the latter must have some additional cause. In this connection, our experiments have shown that the ribosome
itself might acquire an increased capacity to translate
CspA mRNA after cold shock. Taken together, our results
are compatible with and support the theory suggested by
VanBogelen and Neidhardt (1990), that the ribosome
might play a direct role in the control of the expression of
cold- (and heat-) shock genes.
Concerning the stability of the transcript, it has been
shown here that after being transcribed from the PL promoter, the CspA mRNA decays both chemically and functionally more rapidly at 37 C than at 10 C. The
stabilization of CspA mRNA at 10 C seems to be specific
since the functional half-life of this mRNA at this temperature is longer than that of the bulk mRNAs. The CspA
mRNA studied in this paper differs somewhat from the natural cspA transcript in its 5 -untranslated region in that it
lacks approximately half of the 159 nucleotides preceding
the initiation triplet, and this difference could obviously
influence its stability.
Nevertheless, results qualitatively similar to ours were
also obtained with the natural transcript by Goldenberg
and coworkers (1996 see the accompanying paper)
who have also concluded that the stability determinants
of CspA mRNA are located at the 3 -end of this mRNA.
Thus, taken together, these results indicate that a differential rate of mRNA degradation may play an important role
in the control of cspA expression. It remains to be seen,
however, if, and to what extent, the differential stability of
the mRNA results from the aforementioned different rate
of CspA mRNA translation, because it is known that translational efficiency of an mRNA may affect its stability (for a
review see Petersen, 1993). Several clues (our unpublished results) seem to indicate, however, that the two
phenomena co-exist and co-operate in controlling the
cold-shock activation of CspA synthesis.
The third post-transcriptional level of regulation suggested by our work is a possible differential stability of
the CspA protein at 37 C and 10 C. While such an hypothesis is appealing, also in light of the fact that it would
establish a more complete similarity between cold-shock
and heat-shock response (it is known that protein 32 is
rapidly degraded at normal temperature and becomes
stabilized during heat shock (for a review see Yura et al.,
1993)), it is premature to draw conclusions on this point
from our data. In fact, the higher rate of CspA degradation
at 37 C was observed following the mild heat shock

l
8

'

'

238 A. Brandi, P. Pietroni, C. O. Gualerzi and C. L. Pon

required to inactivate the ts repressor, and it is known

that there are proteases among the heat-shock genes
(Yura et al., 1993); the lower degradation rate, on the
other hand, was observed after cold shock, and it is
known that, under these conditions, the expression of
heat-shock genes is reduced (Taura et al., 1989). Additional experiments to clarify this point are obviously
necessary.
In conclusion, the present data, together with that presented in the accompanying paper by Goldenberg et al.
(1996), give a strong indication that mRNA translation
and stability play an important role in the cold-shock activation of cspA. These results do not, however, rule out
the possibility that transcriptional regulation may also
play an important regulatory role in this process.

Experimental procedures

The 1.2 kb Pst I Eco RI DNA fragment containing cspA was

phage 9F6 (Khohara et al., 1987) and
isolated from
cloned into pTZ19R. Subsequently, a 320 bp Mnl I Eco RI
fragment (cspA deprived of its promoter) was subcloned
into the polylinker (Sma I Eco RI) of pTZ18R, then excised
with Eco RI and Bam HI, and the cspA gene placed under
PL promoter in the expression vector
the control of the
pPLc2833 (Remaut et al., 1981) digested with Bam HI and
Hin dIII (ligation of Bam HI followed by Klenow fill-in and
blunt-end ligation). To complete the construction of pCspA1,
the cloned gene was used to walk along the chromosome to
search for the terminator, which was then inserted as the
230 bp Bst XI Fsp I fragment into the Bst XI Acc I-digested
pPLc clone (via ligation of Bst XI and filled-in Fsp I and Acc I
sites; see Fig. 1). The sequence of the reconstructed gene
in pCspA1 was verified by determination of the sequence
(Sanger et al., 1977).
The Eco RI Nsi I fragment excised from pCspA1 was
placed into the polylinker of pTZ18R (digested with Eco RI
and Pst I) to give rise to pCspA2, in which cspA is under the
control of the T7 promoter.

Expression of cspA

D D

E. coli K12 H1 trp transformed with pCspA1, was grown at

30 C in M9 medium (Sambrook et al., 1989) supplemented
with yeast extract (0.01%), tryptophan (50 g ml71), all other
amino acids (30 g ml71), ampicillin (50 g ml71) or in Luria
Bertani (LB) medium containing ampicillin (50 g ml71).
When the absorbance at 620 nm reached 0.6, the culture
was transferred to 42 C for 20 min and then incubated at
either 37 C or 10 C to allow expression of cpsA.

Determination of the chemical stability of mRNA

D D

E. coli K12 H1 trp transformed with pCspA1, was grown at

30 C in LB containing ampicillin (50 g ml71) to A 620 = 0.6,
then shifted to 42 C. After 12 min incubation, rifampicin was

Determination of the functional stability of mRNA

D D
8

E. coli K12 H1 trp cells transformed with pCspA1, were

grown at 30 C in supplemented M9 medium (see above).
After reaching A 500 = 1.0, the culture was transferred to
42 C for 4 min and rifampicin added to a final concentration
of 0.25 mg ml71. The culture was divided into two equal portions, which were incubated at 37 C and 10 C. At the times
indicated in Fig. 4, 1.5 ml aliquots of each culture were
removed, labelled with 6 Ci of [ 35S]-methionine (1000 Ci
mmole71 ) for 5 min and chased with non-radioactive methionine (final concentration 30 mM). The cells were collected by
centrifugation, washed in 0.9% NaCl, and their protein content subjected to electrophoretic analysis preceded, or not,
by immunoprecipitation. For the electrophoretic analysis in
the absence of immunoprecipitation, the cells were resuspended in Laemmli sample buffer, heated at 100 C for
5 min, and the supernatant obtained after centrifugation was
loaded onto SDS(18%) PAGE. For immunoprecipitation,
which was carried out essentially as described by Sambrook
et al. (1989), the cells were resuspended in 200 l PBS containing 1% Nonidet P-40 (Sigma), 10 mM DTT and 100 g
ml71 PMSF, disrupted by sonication, and 40 l aliquots of
the supernatant obtained by centrifugation were incubated
with Protein A Sepharose (Sigma) pre-coated with rabbit polyclonal anti-CspA. After 150 min at room temperature, unbound
protein was removed by washing with PBS containing 0.05%
Tween 20, and 50 l Laemmli sample buffer was added. After
heating at 80 C for 10 min, the samples were subjected to
electrophoresis as described above. After electrophoresis,
the gels were dried and the radioactivity in the bands was
quantified using a BioRad GS-250 Molecular Imager.

Genetic manipulations of cspA

added to a final concentration of 0.2 mg ml71 and divided

into two equal portions, one of which was incubated at 37 C
and the other at 10 C. At the indicated times, 10 ml aliquots
were withdrawn and total RNA was isolated as previously
described (La Teana et al., 1989). Aliquots (6 g) of each
sample were separated on a 1.2% denaturing agarose gel
containing formaldehyde (Sambrook et al., 1989) and subsequently blotted onto Hybond N (Amersham). The blot was
probed with the Eco RI Nsi I fragment ( 32P-labelled using
the random-primer reaction) derived from pCspA1 using the
hybridization conditions described in La Teana et al. (1991).
The blot was quantified by densitometric scanning of the autoradiograph in the Imaging Densitometer Model GS-670 of
BioRad.

Preparation of CspA mRNA

CspA mRNA was prepared by in vitro run-off transcription of
pCspA2 with T7 RNA polymerase. The reaction mixture contained in 0.5 ml: 40 mM Tris-HCl pH 8.1, 3.75 mM of the four
nucleotide triphosphates, 22 mM MgCl2 , 10 mM spermidine,
5 mM dithiothreitol, 50 g BSA, 60 U RNAsin (Amersham),
10 pmol pCspA2 linearized with Hin dIII and 2000 U T7 RNA
polymerase. After incubation at 37 C for 4 h, the mRNA was
purified by precipitation with LiCl followed by two cycles of
ethanol precipitation.

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

Cold-shock expression of CspA 239

Preparation of S30 fractions, ribosomes and in vitro
translation

E. coli MRE600 was grown in LB at 37 C to an A 620 of 0.8.

The culture was then divided into four portions: one portion
was harvested immediately, and the other portions were
incubated at 10 C then harvested after 20 min, 40 min and
90 min, respectively. The cells were washed with 0.9% NaCl
and stored at 80 C. Cell-free extracts (S30 fractions) were
prepared essentially as described by La Teana et al. (1991),
with the exceptions that the extracts were prepared in
10 mM Tris-HCl pH 7.1, containing 10 mM Mg acetate,
60 mM NH4Cl, and 6 mM 2-mercaptoethanol, and the preincubation step was omitted. To obtain 70S ribosomes and the
post-ribosomal supernatant, the S30 fraction, after adjustment of the NH4Cl concentration to 250 mM, was centrifuged
at 240 000 g for 4 h in the Beckman SW60 rotor. The supernatant was dispensed in small aliquots and stored at 80 C.
The ribosomal pellets were resuspended in the above buffer
containing 60 mM NH4Cl and stored at 80 C in small aliquots. Protein synthesis in vitro was carried out in 50 l reaction mixtures containing: 10 mM Tris-HCl pH 7.7, 11.5 mM Mg
acetate, 60 mM NH4Cl, 2 mM ATP, 0.4 mM GTP, 10.5 mM
PEP, 1.2 g pyruvate kinase, 2 mM dithiothreitol, 0.12 mM
citrovorum, 40 g tRNA (from E. coli MRE600), and 0.2 mM
of all non-radioactive amino acids except for either 0.02 mM
[35S]-methionine or 0.035 mM [14C]-valine. The reaction mixtures also contained CspA mRNA or MS2 RNA as template,
and either S30 fractions or 70S ribosomes and S100 postribosomal supernatant, in the amounts indicated in each
experiment.

8
7 8

7 8

7 8
m

Acknowledgements
This work was supported by grants from the Italian Consiglio
Nazionale delle Ricerche (CNR), PF Ingegneria Genetica and
EC Human Capital and Mobility Programme to C.O.G., as well
as MURST and CNR grants to C.L.P.

References
Bradford, M.M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal Biochem 72: 248
254.
Gold, L. (1988) Post-transcriptional regulatory mechanisms
in Escherichia coli. Annu Rev Biochem 57: 199233.
Goldenberg, D., Azar, I., and Oppenheim, A.B. (1996)
Differential mRNA stability regulates the expression of
the cspA gene in the cold-shock response of Escherichia
coli. Mol Microbiol 19: 241248.
Goldstein, J., Pollitt, N.S., and Inouye, M. (1990) Major cold
shock protein of Escherichia coli. Proc Natl Acad Sci USA
87: 283287.
Gualerzi, C.O., and Pon, C.L. (1990) Initiation of mRNA
translation in prokaryotes. Biochemistry 29: 58815889.
Jiang, W., Jones, P., and Inouye, M. (1993) Chloramphenicol
induces the transcription of the major cold shock gene of
Escherichia coli, cspA. J Bacteriol 175: 58245828.

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240

Jones, P.G., and Inouye, M. (1994) The cold-shock response

a hot topic. Mol Microbiol 11: 811818.
Jones, P.G., VanBogelen, R.A., and Neidhardt, F.C. (1987)
Induction of proteins in response to low temperature in
Escherichia coli. J Bacteriol 169: 20922095.
Jones, P.G., Krah, R., Tafuri, S.R., and Wolffe, A.P. (1992)
DNA gyrase, CS7.4, and the cold shock response in
Escherichia coli. J Bacteriol 174: 57985802.
Kohara, Y., Akiyama, K., and Isono, K. (1987) The physical
map of the whole E. coli chromosome: application of a new
strategy for rapid analysis and sorting of a large genomic
library. Cell 50: 495508.
La Teana, A., Falconi, M., Scarlato, V., Lammi, M., and Pon,
C.L. (1989) Characterization of the structural genes for the
DNA-binding protein H-NS in Enterobacteriaceae. FEBS
Lett 244: 3438.
La Teana, A., Brandi, A., Falconi, M., Spurio, R., Pon, C.L.,
and Gualerzi, C.O. (1991) Identification of a cold shock
transcriptional enhancer of the Escherichia coli gene
encoding nucleoid protein H-NS. Proc Natl Acad Sci USA
88: 1090710911.
Lee, S.J., Xie, A., Jiang, W., Etchegaray, J.P., Jones, P.G.,
and Inouye, M. (1994) Family of the major cold-shock
protein, CspA (CS7.4), of Escherichia coli, whose
members show a high sequence similarity with the
eukaryotic Y-box binding proteins. Mol Microbiol 11:
833839.
McCarthy, J.E.G., and Gualerzi, C. (1990) Translational
control of prokaryotic gene expression. Trends Genet 6:
7885.
Newkirk, K., Feng, W., Jiang, W., Tejero, R., Emerson, S.D.,
Inouye, M., and Montelione, G.T. (1994) Solution NMR
structure of the major cold shock protein (CspA) from
Escherichia coli : Identification of a binding epitope for
DNA. Proc Natl Acad Sci USA 91: 51145118.
Petersen, C. (1993) Translation and mRNA stability in bacteria:
a complex relationship. In Control of Messenger RNA
Stability. Belasco, J., and Brawerman, G. (eds). San
Diego: Academic Press, pp. 117145.
Remaut, E., Stanssens, P., and Fiers, W. (1981) Plasmid
vectors for high efficiency expression controlled by the PL
promoter of coliphage lambda. Gene 15: 8193.
Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989)
Molecular Cloning: A Laboratory Manual. 2nd edn. Cold
Spring Harbor, New York: Cold Spring Harbor Laboratory
Press.
Sanger, F., Nicklen, S., and Coulson, A.R. (1977) DNA
sequencing with chain-terminating inhibitors. Proc Natl
Acad Sci USA 73: 54635467.
Schindelin, H., Marahiel, M.A., and Heinemann, U. (1993)
Universal nucleic acid-binding domain revealed by crystal
structure of the B. subtilis major cold-shock protein. Nature
364: 164168.
Schindelin, H., Jiang, W., Inouye, M., and Heinemann, U.
(1994) Crystal structure of CspA, the major cold shock
protein of Escherichia coli. Proc Natl Acad Sci USA 91:
51195123.
Schnuchel, A., Wiltscheck, R., Czisch, M., Herrier, M.,
Willimsky, G., Graumann, P., Marahiel, M.A., and Holak,
T.A. (1993) Structure in solution of the major cold-shock
protein from Bacillus subtilis. Nature 364: 164171.

240 A. Brandi, P. Pietroni, C. O. Gualerzi and C. L. Pon

Tanabe, H., Goldstein, J., Yang, M., and Inouye, M. (1992)
Identification of the promoter region of the Escherichia coli
major cold shock gene, cspA. J Bacteriol 174: 38673873.
Taura, T., Kusukawa, N., Yura, T., and Ito, K. (1989) Transient
shut-off of Escherichia coli heat shock protein synthesis
upon temperature shift-down. Biochem Biophys Res
Commun 163: 438443.

VanBogelen, R.A., and Neidhardt, F.C. (1990) Ribosomes as

sensors of heat and cold shock in Escherichia coli. Proc
Natl Acad Sci USA 87: 55895593.
Wolffe, A.P., Tafuri, S., Ranjan, M., and Familari, M. (1992)
The Y-box factors: a family of nucleic acid binding proteins
conserved from Escherichia coli to man. New Biol 4: 290
298.
Yamanaka, K., Mitani, T., Ogura, T., Niki, H., and Hiraga, S.
(1994) Cloning, sequencing, and characterization of multicopy suppressors of a mukB mutation in Escherichia coli.
Mol Microbiol 13: 301312.
Yura, T., Nagai, H., and Mori, H. (1993) Regulation of the
heat-shock response in bacteria. Annu Rev Microbiol 47:
321350.

# 1996 Blackwell Science Ltd, Molecular Microbiology, 19, 231240