INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The Clp protease is the second ATP-dependent protease purified from Escherichia coli bacteria (9, 13). Purified Clp protease is composed of an ATP-binding component and a peptide-degrading component (9, 13). The peptide-degrading component of the Clp protease is ClpP (21, 22, 34); the ATP-binding component of the Clp protease may be either the ClpA protein or the ClpX protein (7, 14, 33). The Clp protease can degrade abnormal proteins in vivo (13, 34); it is also responsible for the turnover of specific normal proteins (5, 37). Recognition of specific substrates by the Clp protease is dependent on the particular ATP-binding component of the protease (14, 31). Either ClpA or ClpX can associate with ClpP to form a functional protease, but the ClpAP protease and the ClpXP protease have different substrate specificities (7, 26). The target proteins of the ClpAP protease include a ClpA-LacZ fusion protein and the P1 RepA protein (5, 32); the target proteins of the ClpXP protease include the  O protein and the E. coli RpoS protein (31, 33).

The clpP and clpX genes form an operon that is located at 23 min on the E. coli genetic map (7, 21). In this operon, clpP is the promoter-proximal gene, and clpX is the promoter-distal gene; there is a 125-bp intercistronic region between clpP and clpX (7, 21). The gene upstream of the clpP-clpX operon encodes the trigger factor, and the gene downstream of the clpP-clpX operon encodes the Lon protease (21). The clpX gene is usually transcribed as part of the clpP-clpX operon (7). A promoter for clpX alone, however, appears to be located in the intercistronic region between clpP and clpX; this promoter may be relatively weak (35).

The promoters immediately upstream of the clpP gene, responsible for transcribing the clpP-clpX operon, include two ⁷⁰-dependent promoters (21). The fact that ClpP is an RpoH-dependent heat shock protein (15) suggests that there may also be a heat shock promoter for the clpP gene. By comparing the promoter sequence of the clpP-clpX operon with sequences of known heat shock promoters, a putative heat shock promoter has been identified in the promoter region of the clpP-clpX operon (8). This heat shock promoter is located between the two ⁷⁰-dependent promoters (8).

In E. coli, the cellular levels of certain regulatory proteins are actively regulated by proteases (4, 6). The RpoS protein (16, 30), responsible for induction of stationary-phase-specific genes, is degraded by the ClpXP protease (26, 37). During entry into stationary phase, the cellular level of the RpoS protein increases. The increase in RpoS protein appears to be a consequence of increased translation of rpoS mRNA and also of increased stability of RpoS protein (17, 23, 36). When E. coli is ready to leave stationary phase, the ClpXP protease is required to degrade protein products of the stationary-phase-specific genes (3). In this study, we describe some of the effects of carbon deprivation on transcription of the clpP-clpX operon of E. coli.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and strain construction. The bacterial strains, the transducing phage, and the plasmid used in this study are listed in Table 1. The rpoB mutations in E. coli RFM443 (rpoB8) and in E. coli RFM443 (rpoB3595) were transferred into E. coli E103S/pWPC9 by P1 transduction. P1 transduction was carried out as described by Silhavy et al. (27). Insertion of pWPC9 into bacterial strains was done by electroporation. Electroporation was carried out as described by the manufacturer of the apparatus (Bio-Rad Laboratories).

Reagents and kits. Agarose was purchased from Life Technologies. Acidic phenol, amino acids, and ampicillin were purchased from Fisher Scientific. Diethylene pyrocarbonate (DEPC), MgSO₄, NaCl, and sodium acetate were purchased from Sigma. The nick translation kit was purchased from Promega. RNA size markers were purchased from Promega.

Carbon starvation of E. coli. A culture of the E. coli strain to be studied was grown overnight with aeration at 30°C in 1× M9 growth medium (25), which contains 0.2% glucose, required amino acids, and vitamins. After overnight growth, the bacterial culture was diluted into fresh 1× M9 growth medium. The optical density at 600 nm (OD₆₀₀) of the diluted bacterial culture was adjusted to about 0.02. The culture was then incubated at 30°C with aeration until its OD₆₀₀ reached 0.2. Half of the bacterial culture was then transferred to a chilled centrifuge tube containing one-half of the original culture volume of ice-cold 40 mM sodium acetate. This bacterial sample, collected in log phase, was kept on ice. The other half of the bacterial culture was collected on a Milli-Q type HA 0.45-µm-pore-size filter membrane. Bacteria on the membrane were washed twice with the same volume of prewarmed 1× M9 wash-starvation medium, which contains required amino acids and vitamins but no glucose. Bacteria on the filter membrane were then resuspended in an equal volume of 1× M9 wash-starvation medium, and the resuspended bacteria were returned to 30°C with aeration. After the desired time of starvation, the culture was transferred to a chilled centrifuge tube containing one-half of the original volume of ice-cold 40 mM sodium acetate. The bacterial samples were then centrifuged in a Sorvall SS34 rotor at 10,000 rpm for 10 min at 4°C. The supernatants were discarded after centrifugation, and the bacterial pellets were stored on dry ice. Then either the bacterial pellets were stored at 70°C, or the RNA was immediately extracted from the frozen bacterial samples.

Extraction of total RNA from E. coli. The procedure for RNA extraction was a modification of the method described by Aiba et al. (1). Lysis buffer (500 µl) and acidic phenol (500 µl) were added to each bacterial pellet. The bacterial pellets were dissolved in lysis buffer together with acidic phenol by vortexing vigorously. The lysed pellets were then transferred to microcentrifuge tubes, where they were incubated for 5 min at 65°C. The organic solvent was separated from the aqueous solution after centrifugation in a microcentrifuge at maximum speed (14,000 rpm) for 4 min at 4°C. The acidic phenol extraction was repeated two more times. The aqueous solution was extracted again with a 1:1 mixture of acidic phenol and chloroform. The aqueous phase from each sample was then mixed with 3 volumes of absolute ethanol, and the RNA was precipitated on dry ice for 30 min. RNA in the solution was pelleted by centrifugation in a microcentrifuge at maximum speed for 15 min at 4°C. RNA pellets were washed twice with 80% ethanol. Each RNA sample was resuspended in 200 µl of DEPC-treated water, and 10 µl of DEPC-treated 5 M sodium chloride was added to each RNA sample. Three volumes of absolute ethanol was added to each sample. The RNA was again precipitated on dry ice for 30 min. RNA in each sample was recovered after centrifugation in a microcentrifuge at 14,000 rpm for 15 min at 4°C. RNA pellets were washed twice with 80% ethanol. DEPC-treated water was used to resuspend the RNA pellets. The concentration and purity of the RNA samples were determined spectrophotometrically.

Production of the template for the radiolabeled probe used in Northern blotting. In most experiments, the template used to produce the radiolabeled probe was amplified from pWPC9 by PCR (2). The forward primer, 5'-GTCATACAGCGGCGAACGAG-3', annealed to the clpP sequence between bp 380 and bp 399. The reverse primer, 5'-CGACCAGACCGTATTCCAC-3', annealed to the clpP sequence between bp 975 and bp 957. The PCR mix included 10 ng of pWPC9, 1 µg of the forward primer, 1 µg of the reverse primer, 5 µl of 10× buffer for Taq polymerase, 0.5 µl of Taq polymerase, and 41.5 µl of H₂O.

Northern blotting. RNA glyoxal-agarose gel electrophoresis was carried out as described by Ausubel (2). RNA was electrotransferred from the agarose gel to a Zeta probe GT membrane as suggested by the manufacturer of the Zeta probe GT membrane (Bio-Rad Laboratories). The radiolabeled probe used in Northern blotting was produced by nick translation. The probe was labeled during nick translation with [-³²P]dCTP. The template used in nick translation was a PCR fragment amplified from the sequence of the clpP gene between bp 380 and bp 975. Hybridization and prehybridization were carried as recommended by the manufacturer of the Zeta probe GT membrane (Bio-Rad Laboratories).

Primer extension. The procedure for primer extension was a modification of the method described by Sambrook et al. (25). The primer 5'-TATCTCGTTCGCCGCTGTATGACAT-3' was used in primer extension. This primer annealed to the noncoding strand of the clpP gene between bp 378 and bp 402. The primer was radiolabeled at its 5' end by T4 polynucleotide kinase. The primer labeling reaction included 100 ng of the primer in 16 µl of H₂O, 1 µl of 20-mCi/ml [-³²P]ATP, 2 µl of 10× buffer for T4 polynucleotide kinase, and 1 µl of T4 polynucleotide kinase. The reaction mixture was incubated for 1 h at 37°C. T4 polynucleotide kinase was inactivated by incubating the reaction mixture for 15 min at 65°C. The radiolabeled primer was further diluted in 200 µl of DEPC-treated water and then stored at 20°C.

Measurement of band intensities on autoradiographs. The autoradiographs produced from Northern blots and from primer extension gels were scanned, and the images were analyzed by the gel plotting function of the image analysis software (NIH Image). After the background was subtracted, the integrals of areas under the peaks corresponding to the predominant bands in the autoradiographs were calculated.

Repetition of experiments. Unless stated otherwise, all reported experiments were performed at least three times with essentially similar results.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Transcription of the E. coli clpP-clpX operon during carbon starvation. To study the expression of the clpP-clpX operon, Northern blotting was performed to determine the sizes of the transcripts produced from the clpP-clpX operon in log phase and during carbon starvation. A culture of E. coli E103S bacteria in log phase was subjected to carbon starvation. A sample of bacteria was collected before the culture was starved. At 15 min after the onset of carbon starvation, another bacterial sample was collected. Total RNA from each bacterial sample was purified, and Northern blotting was performed to determine the sizes and the relative levels of the transcripts produced from the clpP-clpX operon. The radiolabeled probe used to detect the transcripts of the clpP-clpX operon was prepared by nick translation of PCR fragments of the clpP sequence between bp 380 and bp 975. This probe hybridizes specifically to the clpP sequence; it was used for all the Northern blots in this study, with one noted exception.

View larger version (52K): [in this window] [in a new window]   FIG. 1.   (A) Northern blotting was performed to determine the sizes and relative levels of the transcripts produced from the clpP-clpX operon in E. coli E103S. RNA samples were prepared from a culture of E103S in log phase (lane L) and after 15 min of carbon starvation (lane S). The probe used in the Northern blotting hybridized specifically to the clpP sequence, not to the clpX sequence. This probe was used in all of the Northern blot experiments in this study. (B) Northern blotting was performed to analyze the clpP-clpX transcripts produced in E103S/pWPC9. Total RNA was extracted from E103S/pWPC9 in log phase (lane L) and after 7, 15, and 20 min of carbon starvation (lanes 7, 15, and 20, respectively).

RpoH protein required for production of the shorter transcript. The ClpP protein is a heat shock protein (15), and the RpoH protein is required for the increase in the level of ClpP during heat shock (15). The RpoH protein is also involved in the induction of various stress-related proteins during entry into stationary phase (10, 11). The role of the RpoH protein in the increase in the level of the shorter transcript during carbon starvation was therefore determined. E. coli CAG15071 and CAG15137 were kindly provided by C. Gross. CAG15071 is an rpoH30::kan strain; CAG15137 is an rpoH⁺ strain otherwise isogenic to CAG15071. RNA was prepared from bacterial samples collected from cultures of CAG15071 and CAG15137 in log phase and after 15 min of carbon deprivation. Northern blotting was performed to determine the expression of the clpP-clpX operon in CAG15071 and CAG15137. The results of Northern blotting are shown in Fig. 2. CAG15137, the rpoH⁺ strain, shows changes in the relative amounts of the longer and shorter clpP-clpX transcripts similar to those seen in E103S during carbon starvation (Fig. 1A). The longer transcript is the predominant transcript produced from the clpP-clpX operon in the rpoH⁺ strain CAG15137 during log phase (Fig. 2, lane wt L); the shorter clpP-clpX transcript is predominant when CAG15137 is subjected to carbon deprivation (Fig. 2, lane wt S). CAG15071, the rpoH::kan strain, produces the longer transcript from the clpP-clpX operon in log phase (Fig. 2, lane rpoH L). CAG15071, however, does not accumulate the shorter transcript during carbon starvation (Fig. 2, lane rpoH S). The RpoH protein is the  factor responsible for initiating transcription of genes involved in the heat shock response (29). The results with the rpoH mutant (Fig. 2) suggest that activation of the ³²-dependent promoter of the clpP-clpX operon is required for the increase in the level of the shorter transcript during carbon starvation.

View larger version (66K): [in this window] [in a new window]   FIG. 2.   Northern blotting was performed to analyze the transcripts produced from the clpP-clpX operon in E. coli strains CAG15137 (wt) and CAG15071 (rpoH) in log phase and during carbon starvation. Total RNA samples were prepared from the cultures in log phase (lanes labeled L) and after 15 min of carbon starvation (lanes labeled S).

clpP-clpX promoter utilization during carbon deprivation. Primer extension was employed to determine which promoters of the clpP-clpX operon are used in log phase and during carbon starvation. Total RNA was extracted from the bacterial samples collected from cultures of E103S/pWPC9 in log phase and after 15 min of carbon starvation. The increase in the level of the shorter clpP-clpX transcript from starved E103S/pWPC9 cells was confirmed by Northern blotting (data not shown). Total RNA prepared from these bacterial samples was used for primer extension. The radiolabeled primer was annealed to the RNA sequence near the 5' end of the clpP coding sequence, and the primers were then extended by reverse transcriptase. The DNA fragments extended from the annealed primer by reverse transcription were resolved in a sequencing gel. A sequencing ladder was generated from pWPC9 and the same primer was used in the primer extension experiments. The transcription initiation sites were determined by comparing the positions of DNA fragments to the sequencing ladder. The results of the primer extension experiments are shown in Fig. 3. Three transcription initiation sites, located at bp 279, bp 306, and bp 344, are evident. The transcription initiation sites for the two ⁷⁰-dependent promoters of the clpP-clpX operon are at bp 279 and bp 344 (21). The 10 region of the ³²-dependent promoter is thought to be located between bp 292 and bp 297 (8). Thus, bp 306 is probably the transcription initiation site of the ³²-dependent promoter of the clpP-clpX operon.

View larger version (59K): [in this window] [in a new window]   FIG. 3.   Activities of the 70-dependent promoters and 32-dependent promoter of the clpP-clpX operon analyzed by primer extension. Total RNA samples were prepared from a culture of E103S/pWPC9 in log phase (lane L) and after 15 min of carbon starvation (lane S). The same primer used for primer extension was also used to generate a sequencing ladder from pWPC9. The sequencing ladder and the DNA fragments from primer extension were resolved in an 8% sequencing gel. The initiation sites of the two 70-dependent promoters are located at bp 279 and bp 344. The initiation site of the 32-dependent promoter is located at bp 306.

Activation of the ³²-dependent promoter and production of the shorter transcript. The RpoH protein is required for the increase in the level of the ClpP protein during heat shock (15); furthermore, the results in Fig. 2 suggest that the increase in the level of the shorter transcript during carbon starvation requires the RpoH protein, presumably to activate the ³²-dependent promoter of the clpP-clpX operon. If activation of the ³²-dependent promoter were the only requirement for the production of the shorter transcript, subjecting the bacteria to heat shock should lead to an increase in the level of the shorter transcript. To determine the effect of heat shock on the transcription of the clpP-clpX operon, a bacterial sample was collected from a culture of E103S/pWPC9 in log phase. Additional bacterial samples were collected from cultures of E103S/pWPC9 15 min after the onset of heat shock at 42°C, 15 min after the onset of carbon starvation at 30°C, and 15 min after the onset of the carbon starvation and heat shock at 42°C. RNA was extracted from these bacterial samples, and Northern blotting was performed to determine the effects of the different conditions on the expression of the clpP-clpX operon in E103S/pWPC9. The results of Northern blotting are shown in Fig. 4. The band intensities of the shorter and longer transcripts in each sample on the blot were measured. The combined intensities of the shorter and the longer clpP-clpX transcripts of each sample were designated as 100%. The intensity of the shorter transcript as a percentage of the combined intensities was then determined for each sample. The results of these measurements are shown in Table 3. It appears that heat shock treatment does not change the ratio of the shorter transcript to the longer transcript (Table 3 and Fig. 4, lanes L and H). Similar to the results shown in Fig. 1, carbon starvation treatment leads to accumulation of the shorter transcript in E103S/pWPC9 (Fig. 4, lane S). The level of the shorter transcript as a percentage of the combined clpP-clpX transcripts increases from 13% in log phase and 16% during heat shock to 58% after 15 min of carbon starvation (Table 3). When E103S/pWPC9 bacteria are subjected to carbon starvation and to heat shock simultaneously, the shorter transcript is evident and the longer transcript is absent (Fig. 4, lane HS). Heat shock and/or carbon starvation treatment may transiently increase the total transcription of the clpP-clpX operon, but only carbon starvation leads to a relative increase in the production of the shorter transcript. These results suggest that activation of the ³²-dependent promoter of the clpP-clpX operon is not sufficient to increase the production of the shorter transcript.

View larger version (53K): [in this window] [in a new window]   FIG. 4.   Northern blotting performed to analyze the transcripts produced from the clpP-clpX operon in E103S/pWPC9 under different growth conditions. A total RNA sample was prepared from a culture of E103S/pWPC9 in log phase at 30°C (lane L). Additional RNA samples were prepared from the culture of E103S/pWPC9 after 15 min of heat shock at 42°C (lane H), after 15 min of carbon starvation at 30°C (lane S), and after 15 min of simultaneous heat shock and carbon starvation at 42°C (lane HS).

Shorter transcript is degraded faster than the longer transcript. Two of the various mechanisms that might lead to production of the shorter transcript from the clpP-clpX operon are (i) processive RNA processing, which removes nucleotides progressively from the 3' end of the longer transcript to produce the shorter transcript, and (ii) premature termination of the longer transcript. As shown in Fig. 1, the level of the shorter transcript is unusually higher than the level of the longer transcript after 15 min of carbon starvation at 30°C. Therefore, if the first mechanism were active and the shorter transcript resulted from a progressive processing of the longer transcript, the half-life of the shorter transcript should be greater than or at least equal to the half-life of the longer precursor transcript. It would be unlikely for a progressive RNA-processing mechanism to be responsible for the relative increase in the level of the shorter transcript if this transcript had a shorter half-life than the longer transcript.

View larger version (34K): [in this window] [in a new window]   FIG. 5.   (A) Northern blotting performed to determine the relative degradation rates of the transcripts produced from the clpP-clpX operon in E103S/pWPC9 during carbon deprivation. A culture of E103S/pWPC9 in log phase was subjected to carbon starvation. After 5 min of carbon starvation, rifampin was added to the starved culture to a final concentration of 100 µg/ml. The RNA samples used in Northern blotting were prepared from the culture of E103S/pWPC9 after 3, 5, 7, 9, and 11 min of carbon starvation (lanes 3, 5, 7, 9, and 11, respectively). (B) Relative levels of the shorter transcript and the longer transcript at various time points after rifampin treatment and the onset of carbon starvation were determined. The Northern blot shown in panel A was scanned, and the signal intensities of the shorter transcript and the longer transcript were measured. The signal intensity of the longer transcript after 3 min of carbon starvation was assigned a value of 1. The relative levels of the longer transcript and the shorter transcript in all other samples are shown as ratios of the band intensity to the signal intensity of the longer transcript after 3 min of carbon starvation. Open bars, longer transcript; solid bars, shorter transcript.

Relative levels of the shorter transcript and the longer transcript are affected by E. coli rpoB mutations. If the production of the shorter RNA transcript from the clpP-clpX operon were due to premature termination of the longer transcript, the relative levels of the shorter and longer transcripts might be affected by changes in transcription termination efficiency. Therefore, the effects of rpoB mutations which alter transcription termination efficiency on the relative levels of the shorter and longer clpP-clpX transcripts were examined.

View larger version (48K): [in this window] [in a new window]   FIG. 6.   Northern blotting performed to analyze the sizes and levels of the clpP-clpX transcripts. Cultures of E103S/pWPC9 (wt), E103S rpoB3595/pWPC9 (B3595), and E103S rpoB8/pWPC9 (B8) were grown logarithmically and then subjected to carbon starvation. The RNA samples used in the Northern blot were prepared from the cultures during logarithmic growth (lanes L) and after 15 min of carbon starvation (lanes S).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Two different sizes of transcripts are produced from the clpP-clpX operon in E. coli. In log phase, the level of the longer transcript is higher than the level of the shorter transcript. During carbon starvation, the level of the shorter transcript increases significantly, whereas the level of the longer transcript decreases (Fig. 1). The accumulation of the shorter transcript during carbon starvation is dependent on the RpoH protein (Fig. 2). The RpoH protein is probably required to increase the activity of the ³²-dependent promoter of the clpP-clpX operon during carbon starvation (Fig. 3). Since the increase in the in vivo level of the ClpP protein following heat shock of an E. coli culture requires the RpoH protein (15), it seems that the ³²-dependent promoter of the clpP-clpX operon is also activated during heat shock. Nevertheless, activation of the ³²-dependent promoter by heat shock does not increase the relative level of the shorter transcript (Fig. 4). It appears that transcription initiated at the ³²-dependent promoter can produce either the shorter or the longer clpP-clpX transcript depending on factors associated with bacterial growth conditions.

Several observations lend support to the hypothesis that the shorter transcript is produced by premature termination of the longer clpP-clpX transcript. First, the shorter transcript is the most abundant clpP-clpX transcript shortly after the onset of carbon deprivation (Fig. 1). Second, the half-life of the shorter transcript is significantly less than that of the longer transcript during carbon starvation (Fig. 5). If the shorter transcript were produced by a processive degradation of the longer transcript, it would be unlikely for the level of the less-stable shorter transcript to be equal to or greater than that of the less-abundant longer transcript. Third, if the sorter transcript were produced by endonucleolytic cleavage of the longer transcript, a clpX probe might have detected a promoter-distal fragment of the longer transcript. A clpX-specific probe, however, failed to detect any transcript other than occasional trace amounts of the longer transcript by 15 min after the onset of carbon deprivation. Fourth, the E. coli rpoB3595 mutant with lower transcription termination efficiency (28) has a lower relative level of the shorter transcript during carbon starvation than does the isogenic wild-type strain (Fig. 6). The E. coli strain carrying the rpoB8 mutation, which increases the termination efficiency of RNA polymerase (12), produces elevated amounts of the shorter transcript in log phase and in carbon-deprived cells. Thus, the relative level of the shorter transcript is altered by changes in the transcription termination efficiency of RNA polymerase. These results suggest that the shorter transcript may be produced by premature termination of the longer clpP-clpX transcript.

This suggestion is consistent with an unpublished experiment in which we used E. coli strains K37 (rho⁺ nusA⁺), K7314 (rho), and K7906 (rho nusA), obtained from D. Friedman. When starved for carbon, strain K37 produced a typical short transcript band when examined with the clpP probe. The rho strain, K7314, however, failed to produce a distinct short transcript band; rather, Northern blots of RNA from starved K7314 cells showed an exceedingly diffuse collection of clpP-clpX transcripts. The rho nusA strain, K7906, again showed a clear short transcript in response to carbon deprivation. Therefore, it seems that the rho and the nusA factors, known to be involved in transcription termination, may play a role in the production of the shorter clpP-clpX transcript in response to starvation for carbon.

The band produced by the shorter transcript of the rpoB8 mutant in log phase is slightly diffuse (Fig. 6, lane rpoB8 L), indicating that the sizes of the short transcript produced by the rpoB8 mutant are somewhat diverse. If the shorter transcript were produced by endonucleolytic cleavage of the longer transcript, the location of the cleavage site preferred by the RNase would not likely be changed in the rpoB8 mutant. On the other hand, the locations of transcription termination sites may spread across a length of RNA, especially when termination is dependent on the Rho factor (18, 19). The increased termination efficiency of RNA polymerase caused by the rpoB8 mutation might increase transcription termination at locations downstream of the termination sites normally used by wild-type RNA polymerase. Transcription termination at the terminator sites not used by the transcription apparatus in the wild-type E. coli strain may thus lead to the heterogeneity of the shorter transcript shown on the Northern blots as the diffuse band representing the shorter transcripts of the rpoB8 mutant in logarithmic growth.

Comparing the results from E. coli strains with and without plasmid pWPC9, it is apparent that the longer clpP-clpX transcript is still present in strains with pWPC9 15 min after the onset of carbon starvation, whereas strains without pWPC9 have very low, usually undetectable levels of the longer transcript after the same period of carbon starvation. The shorter clpP-clpX transcript, usually absent or present only at a very low level in strains without pWPC9 in logarithmic growth, is also consistently evident in logarithmically growing strains with pWPC9. These differences seen on Northern blots of strains with and without pWPC9 may be due to higher levels of clpP-clpX transcription in strains with pWPC9. The resulting stronger Northern blot signals may enable the detection of low levels of the clpP-clpX transcripts by autoradiography. It is, however, also possible that the increased copy number of the clpP-clpX operon in strains with pWPC9 may saturate the mechanism which controls the sizes of the transcripts produced from the clpP-clpX operon, leading to the differences in clpP-clpX transcription observed between strains with and without pWPC9.

The RpoS protein plays a central role in controlling the set of genes which are induced during stationary phase (16, 17, 20). The level of RpoS is affected by protein stability (17, 36). RpoS is degraded rapidly in E. coli during exponential phase, but degradation of RpoS is reduced in stationary phase. Degradation of the RpoS protein is dependent on both the RssB protein (24) and the ClpXP protease (26, 37). Loss of ClpXP protease activity extends the half-life of RpoS more than 15-fold; thus, the ClpXP protease appears to be responsible for most of the turnover of RpoS. Shortly after the onset of carbon starvation, the level of the shorter clpP-clpX transcript (which encodes the ClpP protein but not the ClpX protein) increases substantially and the level of the longer transcript decreases (Fig. 1). Assuming that the efficiency of translation of the sorter transcript is similar to that of the longer transcript, the ratio of the ClpP protein to the ClpX protein will increase due to the increased percentage of the shorter transcript in carbon-deprived cells. The level of the clpA transcript, on the other hand, increases after the onset of carbon starvation (data not shown). Since ClpP can form functional proteases with either ClpA or ClpX, during carbon starvation ClpP would be more likely to complex with ClpA than with ClpX. The presumed decrease in the level of the ClpXP protease associated with greater production of the shorter clpP-clpX transcript and decreased production of the longer transcript during carbon starvation may contribute to the stabilization of the RpoS protein after the onset of carbon starvation.