Promoter upstream bent DNA activates the transcription of the Clostridium perfringens phospholipase C gene in a low temperature‐dependent manner

Seiichi Katayama, Osamu Matsushita, Chang‐Min Jung, Junzaburo Minami, Akinobu Okabe

Author Affiliations

  1. Seiichi Katayama1,
  2. Osamu Matsushita1,
  3. Chang‐Min Jung1,
  4. Junzaburo Minami1 and
  5. Akinobu Okabe*,1
  1. 1 Department of Microbiology, Faculty of Medicine, Kagawa Medical University, Miki‐cho, Kita‐gun, Kagawa, 761‐0793, Japan
  1. *Corresponding author. E-mail: microbio{at}


The phospholipase C gene (plc) of Clostridium perfringens possesses three phased A‐tracts forming bent DNA upstream of the promoter. An in vitro transcription assay involving C.perfringens RNA polymerase (RNAP) showed that the phased A‐tracts have a stimulatory effect on the plc promoter, and that the effect is proportional to the number of A‐tracts, and more prominent at lower temperature. A gel retardation assay and hydroxyl radical footprinting revealed that the phased A‐tracts facilitate the formation of the RNAP–plc promoter complex through extension of the contact region. The upstream (UP) element of the Escherichia coli rrnB P1 promoter stimulated the downstream promoter activity temperature independently, differing from the phased A‐tracts. When the UP element was placed upstream of the plc promoter, low temperature‐dependent stimulation was observed, although this effect was less prominent than that of the phased A‐tracts. These results suggest that both the phased A‐tracts and UP element cause low temperature‐dependent activation of the plc promoter through a similar mechanism, and that the more efficient low temperature‐dependent activation by the phased A‐tracts may be due to an increase in the bending angle at a lower temperature.


Sequence‐directed DNA curvature, which is also referred to as an intrinsic DNA bend, results when special sequence motifs such as A‐tracts (Crothers et al., 1990; Pérez‐Martín et al., 1994) and GC‐type elements (Bolshoy et al., 1991; Brukner et al., 1994) are repeated in phase with the DNA helical repeat. A characteristic feature of bent DNA is the anomalous gel mobility at low temperature (Wu and Crothers, 1984; Diekmann, 1987; Koo and Crothers, 1988). Bent DNA sequences are preferentially located upstream of promoters (Plaskon and Wartell, 1987; Tanaka et al., 1991; Helmann, 1995). Furthermore, transcriptional regulation by bent DNAs has been demonstrated in a number of cases (McAllister and Achberger, 1989; Ellinger et al., 1994; Pérez‐Martín et al., 1994; Brahms et al., 1995; Kim et al., 1995). Promoter‐proximal and ‐distal bends tend to facilitate the formation of closed and open complexes, respectively, although there is no clear‐cut boundary between their positions (Pérez‐Martín et al., 1994).

AT‐rich sequences were also noted upstream of several strong promoters in Escherichia coli (Galas et al., 1985; Deuschle et al., 1986) and other bacteria (Banner et al., 1983; Graves and Rabinowitz, 1986; McAllister and Achberger, 1988; Frisby and Zuber, 1991). The AT‐rich sequence located upstream of the E.coli rrnB P1 promoter has been shown to stimulate transcription through contact with αCTD, the C‐terminal domain of the α subunit of RNA polymerase (RNAP) (Ross et al., 1993; Blatter et al., 1994; Busby and Ebright, 1994; Jeon et al., 1995). Such an AT‐rich sequence, known as an upstream (UP) element, has also been demonstrated for other promoters (Rao et al., 1994; Fredrick et al., 1995; Négre et al., 1997), and its consensus sequence has been determined (Estrem et al., 1998). A proximal bend and an AT‐rich sequence are often located at similar positions (between −60 and −40 bp relative to the transcription start site) and both show phase‐dependency (McAllister and Achberger, 1989; Newlands et al., 1992). Thus, it is difficult to distinguish the effects on promoter activity of bent DNA and an AT‐rich sequence when they overlap.

Phased A‐tracts appear to share some features with AT‐rich sequences. They are preferentially bound by some DNA‐binding proteins, such as histone‐like protein HU and integration host factor (IHF) (Shimizu et al., 1995). AT‐rich sequences are flexible, adopting a bent conformation (Grosschedl et al., 1994), which facilitates the recognition and binding of HMG‐like domains (Miao and Wang, 1996). Both AT‐rich sequences and phased A‐tracts show supercoiling‐dependent flexibility so as to be localized preferentially in the terminal loops of superhelical domains (Tsen and Levene, 1997). They exhibit characteristic properties, strand opening or transition to a pre‐melting state, under low ionic conditions or upon binding to proteins (Chan et al., 1993; Bowater et al., 1994; Economides et al., 1996; Miao and Wang, 1996). Such similarities have raised controversy as to whether bent DNA is or is not responsible for the functions suggested so far.

The phospholipase C gene (plc) of Clostridium perfringens has three phased A‐tracts upstream of the plc promoter (Figure 2A). We have previously shown that plc gene transcription is stimulated in the presence of A‐tracts more prominently at lower temperature than at higher temperature (Matsushita et al., 1996). Although this low temperature‐dependent stimulation seems to be related to the role of bent DNA, the promoter upstream region (−66 to −40 with respect to the transcriptional initiation site) is rich in A+T (96.3%). Thus, elucidation of the underlying mechanism would solve the controversy as to the role of bent DNA in the activation of the promoter. In this study, we examined the mechanism by means of the following approaches: (i) analysis of the temperature effect on the in vitro transcriptional activity of the plc promoter with and without phased A‐tracts; (ii) determination of the stability of the RNAP–plc promoter complex by means of a gel shift assay; (iii) analysis of the complex by hydroxyl radical footprinting; and (iv) comparison of the temperature effects on the transcriptional activation between bent DNA and an AT‐rich UP element using the plc and rrnB P1 promoters. This paper shows that bent DNA enhances the formation of the RNAP–plc promoter complex, and that bent DNA of the plc gene, but not the UP element of rrnB P1, has a low temperature‐dependent stimulatory effect on the downstream promoter. Evidence has also been presented that the UP element of rrnB P1 has a low temperature‐dependent stimulatory effect on the plc promoter, although less prominently than the three phased A‐tracts. Based on these results, we discuss the mechanism(s) underlying the similar temperature‐responsive stimulatory effects of the phased A‐tracts and UP element on the plc promoter, and also the possible involvement of the conformational transition of the bent DNA in the additional effect shown by the three phased A‐tracts.

Figure 1.

SDS–PAGE of RNAPs from E.coli and C.perfringens. EcRNAP (lane 1) and CpRNAP (lane 2) were purified as described in Materials and methods. One microgram of each sample was subjected to SDS–PAGE on a 10% polyacrylamide gel. The numbers on the left are the molecular masses (in kDa) of the markers. The positions of the subunits are indicated by α, β, β′ and σ.


Purification of RNAP from C.perfringens and E.coli cells

Clostridium perfringens RNAP (CpRNAP) of high purity was obtained by a modification of the method of Garnier and Cole (1988). The present method yielded 0.25 mg of RNAP holoenzyme/liter of culture, of which the specific activity was 489 U/mg of protein. Analysis by sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS–PAGE) revealed that no other major peptide than the α‐, β‐, β′‐ and σ‐subunits was present, although only contaminating levels of other peptides were still present in the CpRNAP preparation (Figure 1). The identity of the σ‐subunit was verified by its cross‐immunoreactivity with E.coli σ70 on immunoblotting (data not shown). The σ70 E.coli RNAP holoenzyme (EcRNAP) was also purified to homogeneity (Figure 1). Its specific activity was 393 U/mg of protein, this being comparable with that of CpRNAP.

Figure 2.

Nucleotide sequences of the plc promoter and its upstream regions, and physical maps of the DNA fragments used in this study. (A) Nucleotide sequences of the plc promoter and its upstream regions. The −35 and −10 regions are boxed and the transcription start site is indicated as +1. The TG motif in the ‘extended −10 element’ is denoted by closed circles. The three phased A‐tracts are underlined. (B) Physical maps of the plc gene and its fragments: plc, the plc gene cloned into pJIR418 (Matsushita et al., 1996); 3Ap, 2Ap, 1Ap and 0Ap, fragments containing 3, 2, 1 and 0 A‐tracts, respectively, which were used as templates for the in vitro transcription assay; 0Ap′, a fragment containing no A‐tract, which was used as a template for the promoter competition assay; and +UP‐plc and −UP‐plc, fragments with and without the UP element of E.coli rrnB P1 promoter. The filled and shaded blocks represent the A5‐6‐tract of the plc gene and the UP element of the E.coli rrnB gene, respectively. The plc and rrnB P1 promoters (open circles) and the regions derived from the vectors (vertical line) are also indicated. The transcriptional start site is indicated by the arrow. The sizes of the fragments are indicated by the numbers on the right. (C) Physical maps of E.coli rrnB gene‐containing fragments: +UP‐rrnB P1 and −UP‐rrnB P1, fragments with and without the UP element, respectively. The regions derived from the rrnB gene and vectors are indicated by filled boxes and vertical lines, respectively. Abbreviations: AI, AccI; AII, AccII; Ea, EarI; Ec, EcoRI; F, FokI; H, HindIII; P, PvuII.

Temperature effect on the activity of the plc promoter with different numbers of A‐tracts

For the in vitro transcription assay involving the plc promoter with different numbers of A‐tracts, template DNAs containing 3, 2, 1 and 0 A‐tracts were constructed and designated as 3Ap, 2Ap, 1Ap and 0Ap, respectively (Figure 2B). First, we calculated the bending angles of these template DNAs with the published program (Shpigelman et al., 1993), since curvatures predicted by computer analyses are fairly consistent with those determined experimentally from the anomalous electrophoretic mobility. The magnitude of the bending near the promoters 3Ap and 0Ap are predicted to be ∼40° and ∼0°, respectively (Figure 3). Those of 2Ap and 1Ap are predicted to be ∼30° and ∼10°, respectively. Thus, the magnitude of the bending is roughly proportional to the number of A‐tracts.

Figure 3.

Predicted DNA paths of fragments 3Ap and 0Ap used as DNA templates. The two DNA fragments were derived from pCMΔT and pCM0AΔT as described in the Materials and methods and Figure 1. (A), 3Ap; (B), 0Ap. The predicted DNA paths were calculated using the CURVATURE program (Shpigelman et al., 1993).

An in vitro transcriptional assay was performed using 3Ap, 2Ap, 1Ap and 0Ap at 25 and 45°C (Figure 4). In the case of CpRNAP, the stimulatory effect was more prominent with an increase in the number of A‐tracts. Furthermore, the effect was more prominent at 25 than at 45°C. On the other hand, EcRNAP did not show such A‐tract‐ or low temperature‐dependent stimulation. Thus, the effect is determined by the RNAP species, being dependent on the number of A‐tracts and low temperature.

Figure 4.

The effect of the number of A‐tracts on the transcriptional activity of the plc promoter. 3Ap, 2Ap, 1Ap and 0Ap, containing 3, 2, 1 and 0 A‐tracts, respectively, were used as templates for the in vitro transcription assay. The assay was performed with CpRNAP (closed circles) and EcRNAP (open circles) at 25 (A) and 45°C (B). The level of each transcript was determined by measuring the intensity of the band corresponding to each transcript, and expressed as a percentage of that of the transcript from the wild‐type template (3Ap). The means and standard deviations of three determinations are plotted. Some of the error bars are too small to be distinguishable.

Promoter competition study on plc promoters with and without A‐tracts

A promoter competition assay was performed to compare the strengths of the plc promoters with and without three A‐tracts. To distinguish two transcripts, a PvuII–AccI fragment (0Ap′) and a PvuII–AccII fragment (3Ap) were used as the template DNAs (Figure 2B). Equimolar amounts of the 3Ap and 0Ap′ templates were incubated at 25°C under various conditions, and then the levels of their transcripts were determined. The transcript ratio, 3Ap versus 0Ap′, increased as the RNAP concentration decreased, and as the KCl concentration increased (Tables I and II). Thus, the magnitude of the upstream region‐dependent stimulation is greatest when the concentration of RNAP is limiting, because of either lowering of the concentration of RNAP or increasing of the concentration of KCl. This suggests that the 3Ap template can sequester RNAP from the 0Ap′ template.

View this table:
Table 1. Effect of the RNAP concentration on the transcriptional activation by bent DNA in the promoter competition assay
View this table:
Table 2. Effect of the KCl concentration on the transcriptional activation by bent DNA in the promoter competition assay

Affinity of RNAP to the plc promoters with and without A‐tracts

The relative affinities of the C.perfringens and E.coli RNAPs to 3Ap and 0Ap were measured at 25°C by means of a bandshift assay using 32P‐labeled 3Ap and 0Ap (Figure 2B). A discrete bandshifted RNAP–DNA complex increased as the amount of RNAP increased (Figure 5). The amount of CpRNAP required to retard 0Ap by 50% was ∼5‐fold higher than that of 3Ap (Table III), indicating that transcriptional activation occurs mainly, if not wholly, at the level of binding of CpRNAP to the promoter. EcRNAP bound more preferentially and/or tightly to 3Ap than to 0Ap (Table III). Nonetheless, EcRNAP transcribed the former less efficiently than the latter. The reason for this discrepancy cannot be explained at present but it could be related to an inhibitory effect on promoter clearance (see Discussion).

Figure 5.

Relative binding affinity assay of CpRNAP and EcRNAP as to 3Ap. The 32P‐labeled 3Ap fragment (5 fmol) was mixed with 30 μl of the reaction buffer containing various amounts of RNAPs. The amounts of EcRNAP in lanes 1–3 were 0, 0.015 and 0.03 U, respectively. Those of CpRNAP in lanes 4–6 were: 0, 0.005 and 0.01 U, respectively. After incubation at 25°C for 15 min, the samples were loaded and run on an 8% polyacrylamide gel at 25°C. The intensities of the band corresponding to the RNAP–DNA complex (bound form, indicated by B) and that corresponding to unbound DNA (free form, indicated by F) were measured as described in the text.

View this table:
Table 3. Gel mobility shift assaying of the 3Ap–RNAP and 0Ap–RNAP complexes

Hydroxyl radical footprinting of the RNAP–plc promoter complex

Complexes between RNAP and 3Ap or 0Ap DNA were formed at 25°C, and then footprints of the resulting binary complexes were determined with hydroxyl radicals. A selection of results is shown in Figure 6. For 0Ap, CpRNAP affords protection to five sets of bases, i.e. from −50 to −48, −40 to −38, −29 to −27, −19 to −17 and −8 to −1, on the top strand (note that these coordinates are with respect to the plc start site). In contrast, hydroxyl radical cleavage of the 3Ap–CpRNAP complex generated one additional set of bases, from −61 to −59, on the top strand. Footprints of the bottom strand revealed that the 3′ end of the protected region was −54 for 0Ap, while it was extended further upstream to −64 for 3Ap (data not shown). These additional protected regions lie within a region containing three phased A‐tracts. The vectors of the bending created by the compression of the minor groove due to the propeller‐twist of an A–T base pair were estimated for this region (from −66 to −36), as described by Katahira et al. (1990). The results indicate that the bending is directed toward one face, and that this direction is in good agreement with that predicted with the CURVATURE program (Shpigelman et al., 1993). When the protected bases are plotted, they reside on the face toward which the bending is directed. Thus, the bent DNA seems to adopt a favorable conformation to facilitate binding to and complex formation with CpRNAP by wrapping around CpRNAP. The patterns of hydroxyl radical generated bands of EcRNAP–3Ap and EcRNAP–0Ap complexes were similar to those of CpRNAP–3Ap and CpRNAP–0Ap complexes, respectively, in the upstream region, although both the downstream protected regions of the EcRNAP complexes were extended to +24, unlike those of the CpRNAP ones (data not shown). This suggests that EcRNAP is in contact with the bent DNA similar to CpRNAP in the promoter upstream region.

Figure 6.

Hydroxyl radical footprinting analysis of the 3Ap–CpRNAP and 0Ap–CpRNAP complexes. (A) Autoradiographs of hydroxyl radical footprinting of the 3Ap–CpRNAP and 0Ap–CpRNAP complexes. The footprint of the bottom strand is not shown. Lanes: 1, 4, 5 and 8, A and G chemical sequencing ladder; 2, 3Ap DNA alone; 3, 3Ap DNA plus CpRNAP; 6, 0Ap alone; and 7, 0Ap DNA plus CpRNAP. The location of the three phased A‐tracts, and the −35 and −10 regions are also shown by filled and dotted boxes, respectively. (B) A schematic representation of the footprints of the 3Ap–CpRNAP and 0Ap–CpRNAP complexes. Open rectangles represent regions of the DNA duplex protected from hydroxyl radical cleavage due to RNAP binding. Note that RNAP binds to the face of the DNA helix opposite to the three phased A‐tracts.

Promoter competition study on the rrnB P1 promoter with and without the UP element

The low temperature‐dependent activation of the wild‐type plc promoter seems to be due to the phased A‐tracts. It is equally possible that the upstream AT‐rich sequence is responsible for this activation. To assess these possibilities, we performed two sets of in vitro promoter competition experiments, one involving EcRNAP, and wild‐type (+UP‐rrnB P1) and mutant (−UP‐rrnB P1) rrnB templates, and the other involving CpRNAP, and 3Ap and 0Applc templates. The transcript ratio of 3Ap versus 0Ap′ increased with a decrease in temperature, indicating that the extent of the stimulation is correlated with that of bending (Table IV). On the other hand, the transcript ratio of +UP‐rrnB P1 versus −UP‐rrnB P1 was nearly constant at all temperatures within the range of 25–45°C, although the effect of UP was maximum at 37°C (Table IV). This suggests that the activation of the plc promoter is due to the bent DNA. However, this may not only reflect the difference in the promoter upstream sequence between the plc and rrnB P1 templates, since they also differ in their downstream sequences including the core promoter sequence. To assess the possible effect of the downstream region, we constructed +UP‐plc, a chimera of the rrnB P1 UP element and the plc promoter (Figure 2), and examined transcriptional activity at various temperatures by means of the in vitro competition assay. The stimulation of the plc promoter by the UP element was greater at lower temperatures than at higher temperatures. The stimulatory effect at 25°C was much lower than that of 3Ap, being comparable with that of 2Ap (Table IV). This indicates that both the rrnB P1 UP element and the three phased A‐tracts cause low temperature‐dependent stimulation of the plc promoter, the stimulatory effect of the latter being more prominent than that of the former.

View this table:
Table 4. Effects of temperature on the transcriptional activation of plc and rrnB P1 promoters by phased A‐tracts and rrnB P1 UP element in the promoter competition assay


Our previous in vivo study showed that plc gene expression is stimulated by lowering of the temperature (Matsushita et al., 1996). The present in vitro study has revealed evidence that bent DNA stimulates the plc promoter activity depending on the number of phased A‐tracts and low temperature. These two factors should increase the bending angle, as rationalized by the junctional model (Koo et al., 1986), and exemplified by the anomalous mobility of bent DNA in the gel matrix and by the circular permutation assay (Wu and Crothers, 1984). Therefore, it is highly possible that bent DNA modulates the plc gene expression by sensing an alteration in temperature. On the contrary, the activation of rrnB P1 promoter by the UP element is temperature‐independent, at least between 25 and 45°C.

The results of the experiment involving the chimeric template indicated that the UP element stimulates the plc promoter in a low temperature‐dependent manner, similarly to the two phased A‐tracts (2Ap). This suggests possible involvement of the downstream region in the low temperature‐dependent stimulation of the plc promoter. A gal P1 promoter, which consists of the −35 region unrelated to the consensus −35 hexamer and the ‘extended −10’ element, i.e. 5′‐TGnTATAAT‐3′ (Barne et al., 1997), has been shown to form an open complex at low temperature. On the contrary, a modified gal P1con promoter containing the consensus −35 hexamer, 5′‐TTGACA, requires a higher temperature to form an open complex (Grimes et al., 1991). Furthermore, the open complex formation of the gal P1 promoter at lower temperature requires the upstream sequence around −50, which is thought to bend or distort the DNA of the intermediate prior to open complex formation (Grimes et al., 1991; Burns et al., 1996). The extended −10 element is also present in the plc promoter (Figure 2A), suggesting that a similar mechanism underlies the open complex formation of the plc promoter. The difference between 3Ap and 0Ap in the transcriptional activity was more than that in the binding affinity to RNAP. This may be due to the difference in the ease with which the two templates form an open complex at low temperature. Whether or not the phased A‐tracts require the extended −10 element to cause the temperature responsive stimulation remains to be determined.

The finding that the UP element exhibited almost the same effect on the plc promoter as the two phased A‐tracts raises the question of whether the temperature responsiveness is specified by DNA bending or the sequence of the upstream region. The major bent DNA has been shown by electrophoretic analysis to be located at ∼100 bases upstream of the transcriptional initiation site, and not just upstream of the −35 hexamer of the rrnB P1 promoter (Gaal et al., 1994). Furthermore, the phased A‐tracts have been shown to activate the downstream promoter through interactions with the RNA polymerase α subunit, similar to UP elements (Aiyar et al., 1998). Thus, one likely explanation is that the phased A‐tracts exhibit a low temperature‐dependent stimulatory effect like the UP element, possibly through interactions with αCTD. However, this does not explain the difference in the extent of the low temperature‐dependent stimulation between the three phased A‐tracts and the UP element: the stimulation range between 25 and 45°C is 30.8–1.9 for the former, while it is 13.5–3.5 for the latter (Table IV). Therefore, it seems also possible that the bending of the upstream region is responsible at least for the additional effect of the three phased A‐tracts. Recently, a UP element consensus sequence for bacterial promoters was determined by the SELEX procedure to be −59 nnAAA (A/T)(A/T)TTTTnnAAAAnnn −38 (Estrem et al., 1998). Thus, there is no doubt that the UP element is a sequence‐specific site for αCTD binding. Considering the results of all the analyses together, we speculate that the promoter upstream phased A‐tracts function simply as UP elements at higher temperature, while they stimulate promoter activity efficiently at lower temperature, possibly by increasing the bending angle. A more precise study involving constructs with various AT‐rich sequences and the plc promoter is necessary to prove our hypothesis.

The band shift experiments revealed that bent DNA stabilizes the CpRNAP–plc promoter complex. The results of the hydroxyl radical footprinting indicated that the presence of bent DNA extends the contact region in the complex to nucleotide position −64. This extension seems to stabilize the complex through the wrapping of the bent DNA around CpRNAP (Nickerson and Achberger, 1995). The downstream end of the contact region differs from that observed for the EcRNAP–plc promoter complex. To focus on the difference in the stable complex formation, the complex was formed under the conditions used for the band shift experiments, which differ from those for the analysis of the sequential steps on transcription initiation. Thus, we cannot explain the difference in the downstream contact regions. Clarification of this will require more detailed biochemical analysis, especially with respect to the nature of CpRNAP and the kinetics of its transcription initiation steps. It should be noted that the plc gene transcription by EcRNAP is repressed in the presence of bent DNA, despite the more efficient binding to EcRNAP observed for 3Ap than for 0Ap. The difference in the effect of the bent DNA between the two RNAPs may be related to the finding by other workers (Ellinger et al., 1994) that facilitated RNA polymerase binding in the presence of A‐tracts stimulates binding‐limited promoters but inhibits the promoter function in which polymerase escape and promoter clearance are rate limiting.

The induction or stimulation of protein synthesis by lowering of the temperature has been reported for some proteins. The low temperature‐dependent expression of CspA, an E.coli major cold‐shock protein, is mainly due to increased mRNA stability at low temperature (Brandi et al., 1996; Goldenberg et al., 1996). The activation of certain promoters in response to low temperature and dehydration in Arabidopsis thaliana results from the interaction of the cis‐acting C‐repeat/dehydration responsive element (DRE) with activator protein CBF1 (Stockinger et al., 1997). The activation of the lambda PL promoter by lowering of the temperature has been suggested to involve an architectural alteration of the promoter region due to the intrinsic properties of the promoter as well as the binding of IHF (Giladi et al., 1995). In contrast to these examples, the temperature‐responsive regulation of plc gene expression only involves the cis‐element. It does not require the co‐evolution of a trans‐acting factor(s) and hence may be one of the most primitive forms of gene regulation.

Many bacterial virulence genes are expressed on a shift from an ambient low temperature to the body temperature upon entering a host (Mekalanos, 1992; Abe et al., 1997; Ohlsen et al., 1997). Why is PLC, a major virulence factor of C.perfringens, downregulated at high temperature? Type A of this organism inhabits the large intestines of humans and other animals, and also lives in soil and water, while the four other types, B through E, live exclusively in the guts of animals. The plc genes from all the types share the promoter upstream phased A‐tracts (Tsutsui et al., 1995). When living symbiotically at 37°C, C.perfringens produces low levels of PLC, a potent phospholipase C, as this would be detrimental to the host. When present in the soil, or following the death of its host, the bacterium must adapt to lower temperatures and adverse conditions, and thus induces PLC which degrades cell membranes or other available lipids to generate a source of carbon and energy for growth.

Materials and methods

Purification of RNAP

The σ70 E.coli RNAP holoenzyme was purchased from Boehringer Mannheim. It was purified further by chromatography on a DNA–cellulose column (denatured DNA; Pharmacia Biotech) according to the method of Lowe et al. (1979). CpRNAP was purified from C.perfringens NCTC 8237 by the method of Garnier and Cole (1988) with a slight modification. Briefly, the cell extract was treated with polyethyleneimine (pH 8.0) at a final concentration of 0.5% to precipitate nucleic acids. Proteins were eluted from the precipitate, and then reprecipitated with ammonium sulfate (65% saturation). This fraction was subjected to gel filtration on a Sephacryl S‐200 column (Pharmacia Biotech), chromatography on a heparin–agarose column (prepacked type II; Sigma Biochemical), FPLC on a MonoQ column (Pharmacia Biotech), and chromatography on a DNA–cellulose column.


A 333‐bp PvuII–AccII fragment containing the plc promoter with the three A‐tracts was prepared from plasmid pCMΔT (Matsushita et al., 1996), this fragment being designated as 3Ap (Figure 2B). The fragments with two, one and no A‐tracts, which were designated 2Ap, 1Ap and 0Ap (Figure 2B), respectively, were prepared from plasmids pCM2AΔT, pCM1AΔT and pCM0AΔT (Matsushita et al., 1996), respectively. A 282‐bp PvuII–AccI fragment containing the promoter without the A‐tracts was prepared from pCM0AΔT, and was designated as 0Ap′ (Figure 2B). This was used as the template in the promoter competition experiments. For hydroxyl radical footprinting, 32P‐labeled DNA fragments were prepared as follows. A 305‐bp HindIII–FokI fragment containing the three A‐tracts was labeled at the 5′ end with [γ‐32P]dATP (4500 Ci/mmol, ICN Biochemicals) using a Megalabel™ kit (Takara Shuzo) according to the protocol recommended by the supplier. The 5′ end of the 32P‐labeled antisense strand was removed by cutting the labeled fragment with AccII. The resulting 242‐bp HindIII–AccII fragment was used as the 3Ap template DNA for hydroxyl radical footprinting (Figure 2B). A 254‐bp 32P‐labeled 0Ap fragment was likewise prepared except that a 423‐bp EarI–DraI fragment from pCM0AΔT was labeled with 32P, followed by cutting with AccII (Figure 2B).

DNA fragments containing the rrnB P1 promoter with and without the UP element were constructed as follows. A 113‐bp PCR product, which contained the rrnB P1 promoter region spanning bases −60 to +53, was generated using E.coli DH5α DNA, 5′‐CAGAAAATTATTTTAAATTTC, 5′‐CTCAGGAGAACCCCGCTTGAC‐3′, as the template, and forward and reverse primers, respectively. This fragment was cloned into the EcoRV site of the pT7Blue T‐Vector (Novagen), the resultant plasmid being designated as pUP8. The nucleotide sequence of the insert in pUP8 was confirmed to be identical with that reported for the rrnB gene (Brosius et al., 1981). To place the terminator downstream of the insert, a 179‐bp EcoRI fragment containing rrnBT1 was obtained from pKK232–8 (Pharmacia Biotech). The fragment was filled in at both ends, followed by ligation into the NdeI site of pUP8, which was blunt‐ended and dephosphorylated. The resulting plasmid was designated as pUP9. A 466‐bp PvuII–HindIII fragment of pUP9 was used as the rrnB P1 promoter containing the UP element (+UP‐rrnB P1; Figure 2C). A 116‐bp PCR product, which contained the promoter region of the rrnB P1 promoter spanning bases −39 to +77, was generated using E.coli DH5α DNA and two primers (5′‐CTCTTGTCAGGCCGGAATAAC‐3′ and 5′‐ATTTTTGCTTTCTCTGCCGGA‐3′). This product was then cloned into the pT7Blue T‐Vector as described above. The resultant plasmid, named pUP10, was digested with NdeI, blunt‐ended, and then dephosphorylated. The 179‐bp fragment of pKK232‐8 containing rrnBT1 was filled in and ligated into the blunt‐ended NdeI site of pUP10, the resultant plasmid being designated as pUP11. A 463‐bp PvuII–HindIII fragment of pUP11 was used as the rrnB P1 promoter lacking the UP element (−UP‐rrnB P1; Figure 2C).

DNA fragments containing the plc promoter with and without the UP element were constructed as follows. A 230‐bp PCR product, which contained the UP element spanning bases −60 to −38 and the plc gene spanning bases −37 to +170, was generated using pCMΔT, and 5′‐CAGAAAATTATTTTAAATTTCCTTAGTGAGCTTATGGTAATTATATGG‐3′ and 5′‐ATCAATCTTTCCATCCCAAGC‐3′, as the template, and forward and reverse primers, respectively. This fragment was cloned into the EcoRV site of the pT7Blue T‐Vector, the resultant plasmid being designated as pUP12. A 283‐bp PvuII–AccII fragment was prepared from pUP12 and used as a chimeric promoter consisting of the UP element and plc promoter (+UP‐plc; Figure 2B). A 301‐bp PvuII–AccI fragment for use as a chimeric plc promoter lacking the UP element (−UP‐plc; Figure 2B) was prepared in the same manner except that 5′‐TTAGTGAGCTTATGGTAATTA was used as the forward primer.

DNA manipulations and sequencing

Restriction endonucleases were purchased from Takara Shuzo, Toyobo, and New England Biolabs. The Klenow enzyme and DNA ligation kit were products of Takara Shuzo. All recombinant DNA procedures were carried out as described by Sambrook et al. (1989). Nucleotide sequencing was performed as described previously (Matsushita et al., 1998).

DNA band shift studies

Two DNA fragments, 3Ap and 0Ap, were labeled at their 5′ ends with [γ‐32P]ATP (4500 Ci/mmol, ICN Biochemicals) using a Megalabel kit. These labeled fragments (5 fmol) were incubated with various amounts of RNAP in 30 μl of buffer consisting of 12 mM HEPES (pH 8.0), 4 mM Tris–HCl (pH 8.0), 60 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 12% (v/v) glycerol, 0.3 mg/ml of bovine serum albumin (BSA), and 0.033 mg/ml of poly(dI–dC) (Pharmacia Biotech) at 25°C for 15 min. The reaction mixture was loaded and run on an 8% polyacrylamide gel at 25°C. The temperature during the electrophoresis was kept constant by means of an apparatus with a device for temperature regulation (Hoefer SE650, Pharmacia Biotech). The band intensity was visualized by autoradiography. The relative amounts of the unbound and bound forms of DNA were determined using a BAS1000 Bio‐Imaging Analyzer (Fuji Photo Film) and Image GaugeTM ver 3.0 software on a Power Macintosh computer. The bound form of DNA was determined by subtracting the amount of the unbound form of DNA in the presence of RNAP from that in the absence of RNAP. The amount of RNAP required to bind to one‐half of the DNA fragment was determined from a calibration curve, which was made using three different concentrations of RNAP at each temperature.

In vitro transcription assay

The in vitro transcription assay was performed as described elsewhere (Matsushita et al., 1996), except that 0.2 pmol of template DNA was incubated for 10 min with 0.1 and 0.01 U of RNAP at 25 and 45°C, respectively. In promoter competition assays, two different templates were transcribed in the same mixture to compare their promoter activities. After incubation, the transcripts were applied to a 6% polyacrylamide gel containing 7 M urea and then electrophoresed. The relative amounts of the transcripts were determined with a BAS1000 Bio‐Imaging Analyzer as described above. The relative ratios of the transcripts were subsequently determined with correction for the number of UMP residues.

Hydroxyl radical footprinting

DNA–RNAP complexes were formed by incubating 32P‐labeled DNAs (30 ng) with CpRNAP (0.2 and 1 U for the 3Ap and 0Ap templates, respectively) under the same conditions as described for the DNA band shift studies. To remove glycerol, 30 μl of the incubation mixture was placed on a Millipore VS filter membrane (0.025 μm; Millipore), and then dialyzed against 30 ml of 8 mM Tris–HCl (pH 7.9) for 1 h at 25°C (Schickor and Heumann, 1994). The protein–DNA complexes were digested with hydroxyl radicals as described by Schickor and Heumann (1994). After digestion for 2 min at 25°C, the reaction was quenched by adding 14 μl of 0.1 M thiourea‐20 mM EDTA (pH 8.0) to 51 μl of the reaction mixture. The samples were subsequently precipitated with ethanol, resuspended in an appropriate volume of formamide gel loading buffer and then run on a 6% sequencing gel. The bands were visualized and analyzed as described above.

Enzyme assay, protein determination, SDS–PAGE and immunoblotting

The enzymatic activity of RNAP was determined using [3H]dUTP (35 Ci/mmol, ICN Biochemicals) and poly(dA–dT) DNA (Pharmacia Biotech) as the labeled nucleotide and template DNA, respectively, as described by Pich and Bahl (1991). One unit of RNAP activity caused the incorporation of 1 nmol of UMP into RNA in 10 min at 37°C. Protein concentrations were determined by the method of Bradford (1976) using the Bio‐Rad protein assay reagent (Bio–Rad Laboratories) with BSA as a standard. SDS–PAGE was performed on a 10% polyacrylamide gel, and the gel was stained with Coomassie Brilliant Blue R as described previously (Jin et al., 1996). For immunoblot analysis of the σ‐subunit of CpRNAP, the purified CpRNAP was electrophoresed on an SDS–10% polyacrylamide gel, and then electroblotted onto a nitrocellulose membrane (Advantec Toyo). Antigens were visualized using antiserum against E.coli σ70, which was donated by C.A.Gross (Department of Bacteriology, University of Wisconsin, Madison, WI), and alkaline phosphatase‐conjugated goat anti‐rabbit immunoglobulin (Promega).


We wish to thank Carol A.Gross (University of California, San Francisco, CA) for providing the antiserum against the σ‐subunit of E.coli RNA polymerase. We also thank Y.Yamasoto (Division of Science and Technology for Intelligence, The Graduate School of Natural Science and Technology, Okayama University, Okayama, Japan) for the invaluable assistance in the computer analysis of the DNA structure. We are also indebted to E.N.Trifonov (Department of Structural Biology, Weizmann Institute of Science, Rehovot, Israel) for providing us with the computer program. This work was supported by a Grant‐in‐Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan.