Portal de Periódicos da CAPES

Heat Shock RNA Polymerase (Eς32) Is Involved in the Transcription of mlc and Crucial for Induction of the Mlc Regulon by Glucose in Escherichia coli

2001; Elsevier BV; Volume: 276; Issue: 28 Linguagem: Inglês

10.1074/jbc.m101757200

1083-351X

Dongwoo Shin, Sangyong Lim, Yeong‐Jae Seok, Sangryeol Ryu,

Fungal and yeast genetics research

Mlc is a global regulator of carbohydrate metabolism. Recent studies have revealed that Mlc is depressed by protein-protein interaction with enzyme IICBGlc, a glucose-specific permease, which is encoded by ptsG. Themlc gene has been previously known to be transcribed by two promoters, P1(+1) and P2(+13), and have a binding site of its own gene product at +16. However, the mechanism of transcriptional regulation of the gene has not yet been established. In vitrotranscription assays of the mlc gene showed that P2 promoter could be recognized by RNA polymerase containing the heat shock sigma factor ς32 (Eς32) as well as Eς70, while P1 promoter is only recognized by Eς70. The cyclic AMP receptor protein and cyclic AMP complex (CRP·cAMP) increased expression from P2 but showed negative effect on transcription from P1 by Eς70, although it had little effect on transcription from P2 by Eς32 in vitro. Purified Mlc repressed transcription from both promoters, but with different degrees of inhibition. In vivotranscription assays using wild type and mlc strains indicated that the level of mlc expression was modulated less than 2-fold by glucose in the medium with concerted action of CRP·cAMP and Mlc. A dramatic increase in mlc expression was observed upon heat shock or in cells overexpressing ς32, confirming that Eς32 is involved in the expression of mlc. Induction of ptsG P1 andpts P0 transcription by glucose was also dependent on Eς32. These results indicate that Eς32 plays an important role in balancing the relative concentration of Mlc and EIICBGlc in response to availability of glucose in order to maintain inducibility of the Mlc regulon at high growth temperature. Mlc is a global regulator of carbohydrate metabolism. Recent studies have revealed that Mlc is depressed by protein-protein interaction with enzyme IICBGlc, a glucose-specific permease, which is encoded by ptsG. Themlc gene has been previously known to be transcribed by two promoters, P1(+1) and P2(+13), and have a binding site of its own gene product at +16. However, the mechanism of transcriptional regulation of the gene has not yet been established. In vitrotranscription assays of the mlc gene showed that P2 promoter could be recognized by RNA polymerase containing the heat shock sigma factor ς32 (Eς32) as well as Eς70, while P1 promoter is only recognized by Eς70. The cyclic AMP receptor protein and cyclic AMP complex (CRP·cAMP) increased expression from P2 but showed negative effect on transcription from P1 by Eς70, although it had little effect on transcription from P2 by Eς32 in vitro. Purified Mlc repressed transcription from both promoters, but with different degrees of inhibition. In vivotranscription assays using wild type and mlc strains indicated that the level of mlc expression was modulated less than 2-fold by glucose in the medium with concerted action of CRP·cAMP and Mlc. A dramatic increase in mlc expression was observed upon heat shock or in cells overexpressing ς32, confirming that Eς32 is involved in the expression of mlc. Induction of ptsG P1 andpts P0 transcription by glucose was also dependent on Eς32. These results indicate that Eς32 plays an important role in balancing the relative concentration of Mlc and EIICBGlc in response to availability of glucose in order to maintain inducibility of the Mlc regulon at high growth temperature. cyclic AMP receptor protein phosphoenolpyruvate: carbohydrate phosphotransferase system When Mlc is overproduced on a multicopy plasmid inEscherichia coli grown in the presence of glucose, it causes reduction of acetate accumulation and E. coli makes large colonies (1Hosono K. Kakuda H. Ichihara S. Biosci. Biotechnol. Biochem. 1995; 59: 256-261Crossref PubMed Scopus (51) Google Scholar). Mlc has been proposed to be a new global regulator of carbohydrate metabolism (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar, 3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar). It has been reported that Mlc regulatesmanXYZ encoding enzyme II of the mannose PTS (4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar),malT encoding the activator of maltose operon, andmlc itself negatively (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar). Moreover, ptsGencoding enzyme IICB of the glucose PTS (EIICBGlc) and thepts operon encoding general PTS proteins also proved to be repressed by Mlc (3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar, 6Kim S.Y. Nam T.W. Shin D. Koo B.M. Seok Y.J. Ryu S. J. Biol. Chem. 1999; 274: 25398-25402Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar, 8Tanaka Y. Kimata K. Inada T. Tagami H. Aiba H. Genes Cells. 1999; 4: 319-399Crossref Scopus (49) Google Scholar). The Mlc regulon is also under the positive control of the CRP1·cAMP complex. It has been discovered that repression of the Mlc regulon is relieved in cells grown in the media containing glucose or other PTS sugars (7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar). It has been shown that the unphosphorylated EIICBGlc can sequester Mlc from its binding sites by direct protein-protein interaction to induce expression of the Mlc regulon in response to glucose (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar, 10Tanaka Y. Kimata K. Aiba H. EMBO J. 2000; 19: 5344-5352Crossref PubMed Scopus (124) Google Scholar, 11Lee S.J. Boos W. Bouche J.P. Plumbridge J. EMBO J. 2000; 19: 5353-5361Crossref PubMed Scopus (121) Google Scholar).The mlc gene encoding a 44-kDa Mlc protein is located around 35 min of chromosomal locus (1Hosono K. Kakuda H. Ichihara S. Biosci. Biotechnol. Biochem. 1995; 59: 256-261Crossref PubMed Scopus (51) Google Scholar). This gene was also identified as the same allele of the dgsA gene (4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 12Morris P.W. Binkey J.P. Henson J.M. Kuempel P.L. J. Bacteriol. 1985; 163: 785-786Crossref PubMed Google Scholar, 13Roehl R.A. Vinopal R.T. J. Bacteriol. 1980; 142: 120-130Crossref PubMed Google Scholar). Decker et al. (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar) have shown that mlc transcription starts from two promoters called upstream "+1" and downstream "+13" and there exists one Mlc-binding site centered at +16. In addition, a highly conserved CRP-binding site is present within the mlcpromoter. However, the detailed mechanisms of transcriptional regulation of mlc have not yet been reported.The majority of the E. coli promoters are recognized by the RNA polymerase containing the housekeeping sigma factor, ς70 (14Ishihama A. Annu. Rev. Microbiol. 2000; 54: 499-518Crossref PubMed Scopus (446) Google Scholar, 15Helmann J.D. Chamberlin M.J. Annu. Rev. Biochem. 1988; 57: 839-872Crossref PubMed Scopus (713) Google Scholar). Several genes that are necessary to respond to various environmental or nutritional changes require specific recognition by RNA polymerase associated with the alternative sigma factors, ς32 (16Grossman A.D. Erickson J.W. Gross C.A. Cell. 1984; 38: 383-390Abstract Full Text PDF PubMed Scopus (307) Google Scholar), ςE (17Erickson J.W. Gross C.A. Genes Dev. 1989; 3: 1462-1471Crossref PubMed Scopus (291) Google Scholar), ς54 (18Hirschman J. Wong P.K. Sei K. Keener J. Kustu S. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 7525-7529Crossref PubMed Scopus (185) Google Scholar), or ςS (19Mulvey M.R. Loewen P.C. Nucleic Acids Res. 1989; 17: 9979-9991Crossref PubMed Scopus (181) Google Scholar). The heat shock response in E. coli is mediated by Eς32 (20Yura T. Nagami H. Mori H. Annu. Rev. Microbiol. 1993; 47: 321-350Crossref PubMed Scopus (395) Google Scholar) and it is known that expression of at least 26 genes is induced by heat shock in E. coli (21Chuang S.E. Blattner F.R. J. Bacteriol. 1993; 175: 5242-5252Crossref PubMed Google Scholar). Many essential genes in E. coli have multiple promoters including one recognized by Eς32 in order to respond to various environmental conditions (22Charpentier B. Branlant C. J. Bacteriol. 1994; 176: 830-839Crossref PubMed Google Scholar, 23Newland J.T. Gaal T. Mecsas J. Gourse R.L. J. Bacteriol. 1993; 175: 661-668Crossref PubMed Google Scholar, 24Ryu S. Ramsier T.M. Michotey V. Saier Jr., M.H. Garges S. J. Biol. Chem. 1995; 270: 2489-2496Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar). It has been shown that the pts P0 promoter is recognized by Eς32 as well as Eς70 (25Ryu S. Mol. Cells. 1998; 8: 614-617PubMed Google Scholar) as is expected for a system as central to carbohydrate metabolism as the PTS. In this work, we studied the transcriptional regulation of the mlc gene in vitro as well as in vivo and the role of Eς32 in maintaining glucose-dependent induction of the Mlc regulon at high growth temperature.DISCUSSIONIt has been suggested that the mlc gene has two promoters, P1 and P2, which are separated by 12 bases and autoregulated by its product (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar). Here, we report that the transcription of themlc gene is regulated in a highly sophisticated manner and that heat shock ς factor, ς32, is involved in its transcription.It is known that the expression level of several genes encoding transcriptional repressors such as galS (33Weickrt M.J. Adhya S. J. Bacteriol. 1993; 175: 251-258Crossref PubMed Google Scholar),nagC (34Plumbridge J. J. Bacteriol. 1996; 178: 2629-2636Crossref PubMed Google Scholar), purR (35Meng L.M. Kilstrup M. Nygaard P. Eur. J. Biochem. 1990; 187: 373-379Crossref PubMed Scopus (68) Google Scholar), and trpR (36Kelley R.L. Yanofsky C. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 3120-3124Crossref PubMed Scopus (74) Google Scholar) ofE. coli is low and that their expression level is not modulated much in various growth conditions. It seems likely that both CRP·cAMP and Mlc work together in E. coli to maintain the level of Mlc optimum in response to availability of glucose. All genes known to be regulated negatively by Mlc, such as manXYZ,malT, ptsG, and pts, are also regulated positively by CRP·cAMP (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar, 3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar, 6Kim S.Y. Nam T.W. Shin D. Koo B.M. Seok Y.J. Ryu S. J. Biol. Chem. 1999; 274: 25398-25402Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar, 8Tanaka Y. Kimata K. Inada T. Tagami H. Aiba H. Genes Cells. 1999; 4: 319-399Crossref Scopus (49) Google Scholar). In vitrotranscription assay with Eς70 and CRP·cAMP showed two opposite effects on each promoter of mlc, that is the positive effect on P2 and the negative effect on P1 (Fig. 2). This can be explained based on the fact that the CRP-binding site of P2 centered at −71.5 to the transcription start site is more compatible for a functional CRP site (37Kolb A. Busby S. Buc H. Garges H. Adhya S. Annu. Rev. Biochem. 1993; 62: 749-795Crossref PubMed Google Scholar) compared with that of P1 centered at −58.5 to the transcription start site. However, Eς32-directed P2 transcription was insensitive to CRP·cAMP. The P2 promoter of mlc should be a good model system to assess the effect of ς factor on transcription activation by CRP·cAMP because it has been known that CRP·cAMP activates transcription by direct protein-protein interaction with the α-subunit of RNA polymerase (38Busby S. Ebright R.H. J. Mol. Biol. 1999; 293: 199-213Crossref PubMed Scopus (632) Google Scholar).Repression of the P2 transcription when cells were grown in the presence of glucose implies that the action of CRP·cAMP is dominant over the self-repression by Mlc in the regulation of the mlcP2 promoter. The low binding affinity of Mlc to its own promoter that is 10 times weaker than that to ptsG or pts P0 promoters (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar) seems to be a major reason for the low influence of Mlc on regulation of its own gene. In vitro transcription assay revealed that each promoter of mlc has a different sensitivity to Mlc (Fig. 3). When cells were grown in the absence of glucose, a similar level of expression from both P1 and P2 was observed even though the P2 transcription is more sensitive to Mlc repression probably because the intracellular concentration of Mlc is limiting inE. coli (3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar). The condition seems to be similar to thein vitro transcription condition where both CRP·cAMP and a small amount of Mlc were present as shown in lane 6 of Fig.3 A. However, P1 was as active as P2 when the mlcstrain was grown in the absence of glucose. These results imply that the concentration of intracellular CRP·cAMP is lower than that of CRP·cAMP used for in vitro transcription reactions (40 nm) (39Ishizuka H. Hanamura A. Kunimura T. Aiba H. Mol. Microbiol. 1993; 10: 341-350Crossref PubMed Scopus (84) Google Scholar). When Mlc was induced and the concentration of CRP·cAMP was lowered by the addition of glucose in the growth medium, the P1 promoter was activated slightly while the P2 promoter was repressed because the P2 promoter is active only in the presence of CRP·cAMP. This situation is similar to the in vitrotranscription condition where neither CRP·cAMP nor Mlc were present (lane 1 of Fig. 3 A). Therefore, the addition of glucose in the growth medium resulted in the reduction ofmlc expression by about half. These results also agree with the previous report by Decker et al. (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar) that expression ofmlc is reduced by half when cells were grown in the presence of glucose by measuring the β-galactosidase activity of themlc-lacZ fusion. Level of Mlc expression can vary precisely in response to the available sugars but the variation range is less than 2-fold in that the availability of unphosphorylated EIICBGlc may be more critical than the intracellular level of Mlc for induction of the Mlc regulon by glucose as shown in our previous report (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar).Eς32 is involved in the transcription of themlc gene. In vitro transcription assay with Eς32 showed that Eς32 could recognize the P2 promoter of the mlc gene. Transcription of P2 was increased when ς32 was overexpressed (Fig.4 A). Moreover, mlc expression was increased upon heat shock. It is known that the intracellular concentration of ς32 in E. coli increases from 15–20-fold within 5 min then declines to a new steady-state level severalfold higher than the preshift level in response to temperature shift from 30 to 42 °C (20Yura T. Nagami H. Mori H. Annu. Rev. Microbiol. 1993; 47: 321-350Crossref PubMed Scopus (395) Google Scholar, 32Straus D.B. Walter W.A. Gross C.A. Nature. 1987; 329: 348-351Crossref PubMed Scopus (265) Google Scholar). Transcription from P2 recognized by Eς32 was induced transiently to an extraordinary level upon heat shock when cells were grown in the absence of glucose (Fig.4 B, lane 4). The level of mlctranscription was changed parallel to the changes in intracellular concentration of ς32. In addition, P2 transcription was activated upon heat shock even when cells were grown in the presence of glucose. These results imply that the major RNA polymerase which activated the P2 transcription upon heat shock was Eς32because Eς32 was less sensitive to Mlc than Eς70 and the P2 transcription by Eς32 was not dependent on CRP·cAMP as revealed by the in vitrotranscription assay (Fig. 3). It is not clear why the P1 transcription was reduced in the ΔrpoH strain and activated by heat shock or when cells overexpressing ς32 were grown in the presence of glucose even though P1 promoter was not recognized by Eς32 in vitro. Because heat shock should exert pleiotropic effects by regulating transcription of various genes (21Chuang S.E. Blattner F.R. J. Bacteriol. 1993; 175: 5242-5252Crossref PubMed Google Scholar), further study on the mechanism of heat shock is needed for a better understanding of these phenomena.To investigate whether the increased level of mlc expression resulting from heat shock can influence Mlc-dependent gene expression, we analyzed changes in ptsG expression by heat shock. The P1 expression of ptsG was increased significantly by heat shock only when cells were grown in the presence of glucose regardless of the presence of Mlc (Fig. 5 A). These results suggest that activation of ptsG P1 by heat shock was not mediated by Mlc or CRP·cAMP even though Mlc repression might be dominant over activation of ptsG P1 by heat shock. We have reported that the unphosphorylated form of EIICBGlcsequesters Mlc from its target promoters upon glucose uptake by direct protein-protein interaction (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar). Therefore, glucose is required to maximize the level of dephosphorylated EIICBGlc necessary to sequester Mlc that is increased by heat shock. However, contrary to the case of mlc P2 transcription in which Eς32plays a major role in its regulation, it is likely that additional factors independent of Eς32 are involved in regulation of the ptsG expression because ptsG P1 expression was increased partially upon heat shock even in theΔrpoH strain grown in the presence of glucose. It means that two separate mechanisms involving Eς32-dependent and Eς32-indendent activation of ptsG P1 may work additively for full activation of ptsG P1 by glucose when cells were heat-shocked. The importance of Eς32 in glucose induction of the Mlc regulon was manifested by the fact that glucose induction of ptsG could not be observed in theΔrpoH strain. The inability of glucose to activate ptsG expression resulted in an insensitivity of thepts P0 promoter to glucose in ΔrpoHstrain. We are trying to elucidate the mechanism of activation ofptsG transcription by heat shock in the absence of Eς32 in order to understand the general role of Eς32 in regulation of genes involving carbohydrate metabolism. When Mlc is overproduced on a multicopy plasmid inEscherichia coli grown in the presence of glucose, it causes reduction of acetate accumulation and E. coli makes large colonies (1Hosono K. Kakuda H. Ichihara S. Biosci. Biotechnol. Biochem. 1995; 59: 256-261Crossref PubMed Scopus (51) Google Scholar). Mlc has been proposed to be a new global regulator of carbohydrate metabolism (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar, 3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar). It has been reported that Mlc regulatesmanXYZ encoding enzyme II of the mannose PTS (4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar),malT encoding the activator of maltose operon, andmlc itself negatively (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar). Moreover, ptsGencoding enzyme IICB of the glucose PTS (EIICBGlc) and thepts operon encoding general PTS proteins also proved to be repressed by Mlc (3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar, 6Kim S.Y. Nam T.W. Shin D. Koo B.M. Seok Y.J. Ryu S. J. Biol. Chem. 1999; 274: 25398-25402Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar, 8Tanaka Y. Kimata K. Inada T. Tagami H. Aiba H. Genes Cells. 1999; 4: 319-399Crossref Scopus (49) Google Scholar). The Mlc regulon is also under the positive control of the CRP1·cAMP complex. It has been discovered that repression of the Mlc regulon is relieved in cells grown in the media containing glucose or other PTS sugars (7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar). It has been shown that the unphosphorylated EIICBGlc can sequester Mlc from its binding sites by direct protein-protein interaction to induce expression of the Mlc regulon in response to glucose (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar, 10Tanaka Y. Kimata K. Aiba H. EMBO J. 2000; 19: 5344-5352Crossref PubMed Scopus (124) Google Scholar, 11Lee S.J. Boos W. Bouche J.P. Plumbridge J. EMBO J. 2000; 19: 5353-5361Crossref PubMed Scopus (121) Google Scholar). The mlc gene encoding a 44-kDa Mlc protein is located around 35 min of chromosomal locus (1Hosono K. Kakuda H. Ichihara S. Biosci. Biotechnol. Biochem. 1995; 59: 256-261Crossref PubMed Scopus (51) Google Scholar). This gene was also identified as the same allele of the dgsA gene (4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 12Morris P.W. Binkey J.P. Henson J.M. Kuempel P.L. J. Bacteriol. 1985; 163: 785-786Crossref PubMed Google Scholar, 13Roehl R.A. Vinopal R.T. J. Bacteriol. 1980; 142: 120-130Crossref PubMed Google Scholar). Decker et al. (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar) have shown that mlc transcription starts from two promoters called upstream "+1" and downstream "+13" and there exists one Mlc-binding site centered at +16. In addition, a highly conserved CRP-binding site is present within the mlcpromoter. However, the detailed mechanisms of transcriptional regulation of mlc have not yet been reported. The majority of the E. coli promoters are recognized by the RNA polymerase containing the housekeeping sigma factor, ς70 (14Ishihama A. Annu. Rev. Microbiol. 2000; 54: 499-518Crossref PubMed Scopus (446) Google Scholar, 15Helmann J.D. Chamberlin M.J. Annu. Rev. Biochem. 1988; 57: 839-872Crossref PubMed Scopus (713) Google Scholar). Several genes that are necessary to respond to various environmental or nutritional changes require specific recognition by RNA polymerase associated with the alternative sigma factors, ς32 (16Grossman A.D. Erickson J.W. Gross C.A. Cell. 1984; 38: 383-390Abstract Full Text PDF PubMed Scopus (307) Google Scholar), ςE (17Erickson J.W. Gross C.A. Genes Dev. 1989; 3: 1462-1471Crossref PubMed Scopus (291) Google Scholar), ς54 (18Hirschman J. Wong P.K. Sei K. Keener J. Kustu S. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 7525-7529Crossref PubMed Scopus (185) Google Scholar), or ςS (19Mulvey M.R. Loewen P.C. Nucleic Acids Res. 1989; 17: 9979-9991Crossref PubMed Scopus (181) Google Scholar). The heat shock response in E. coli is mediated by Eς32 (20Yura T. Nagami H. Mori H. Annu. Rev. Microbiol. 1993; 47: 321-350Crossref PubMed Scopus (395) Google Scholar) and it is known that expression of at least 26 genes is induced by heat shock in E. coli (21Chuang S.E. Blattner F.R. J. Bacteriol. 1993; 175: 5242-5252Crossref PubMed Google Scholar). Many essential genes in E. coli have multiple promoters including one recognized by Eς32 in order to respond to various environmental conditions (22Charpentier B. Branlant C. J. Bacteriol. 1994; 176: 830-839Crossref PubMed Google Scholar, 23Newland J.T. Gaal T. Mecsas J. Gourse R.L. J. Bacteriol. 1993; 175: 661-668Crossref PubMed Google Scholar, 24Ryu S. Ramsier T.M. Michotey V. Saier Jr., M.H. Garges S. J. Biol. Chem. 1995; 270: 2489-2496Abstract Full Text Full Text PDF PubMed Scopus (39) Google Scholar). It has been shown that the pts P0 promoter is recognized by Eς32 as well as Eς70 (25Ryu S. Mol. Cells. 1998; 8: 614-617PubMed Google Scholar) as is expected for a system as central to carbohydrate metabolism as the PTS. In this work, we studied the transcriptional regulation of the mlc gene in vitro as well as in vivo and the role of Eς32 in maintaining glucose-dependent induction of the Mlc regulon at high growth temperature. DISCUSSIONIt has been suggested that the mlc gene has two promoters, P1 and P2, which are separated by 12 bases and autoregulated by its product (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar). Here, we report that the transcription of themlc gene is regulated in a highly sophisticated manner and that heat shock ς factor, ς32, is involved in its transcription.It is known that the expression level of several genes encoding transcriptional repressors such as galS (33Weickrt M.J. Adhya S. J. Bacteriol. 1993; 175: 251-258Crossref PubMed Google Scholar),nagC (34Plumbridge J. J. Bacteriol. 1996; 178: 2629-2636Crossref PubMed Google Scholar), purR (35Meng L.M. Kilstrup M. Nygaard P. Eur. J. Biochem. 1990; 187: 373-379Crossref PubMed Scopus (68) Google Scholar), and trpR (36Kelley R.L. Yanofsky C. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 3120-3124Crossref PubMed Scopus (74) Google Scholar) ofE. coli is low and that their expression level is not modulated much in various growth conditions. It seems likely that both CRP·cAMP and Mlc work together in E. coli to maintain the level of Mlc optimum in response to availability of glucose. All genes known to be regulated negatively by Mlc, such as manXYZ,malT, ptsG, and pts, are also regulated positively by CRP·cAMP (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar, 3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar, 6Kim S.Y. Nam T.W. Shin D. Koo B.M. Seok Y.J. Ryu S. J. Biol. Chem. 1999; 274: 25398-25402Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar, 8Tanaka Y. Kimata K. Inada T. Tagami H. Aiba H. Genes Cells. 1999; 4: 319-399Crossref Scopus (49) Google Scholar). In vitrotranscription assay with Eς70 and CRP·cAMP showed two opposite effects on each promoter of mlc, that is the positive effect on P2 and the negative effect on P1 (Fig. 2). This can be explained based on the fact that the CRP-binding site of P2 centered at −71.5 to the transcription start site is more compatible for a functional CRP site (37Kolb A. Busby S. Buc H. Garges H. Adhya S. Annu. Rev. Biochem. 1993; 62: 749-795Crossref PubMed Google Scholar) compared with that of P1 centered at −58.5 to the transcription start site. However, Eς32-directed P2 transcription was insensitive to CRP·cAMP. The P2 promoter of mlc should be a good model system to assess the effect of ς factor on transcription activation by CRP·cAMP because it has been known that CRP·cAMP activates transcription by direct protein-protein interaction with the α-subunit of RNA polymerase (38Busby S. Ebright R.H. J. Mol. Biol. 1999; 293: 199-213Crossref PubMed Scopus (632) Google Scholar).Repression of the P2 transcription when cells were grown in the presence of glucose implies that the action of CRP·cAMP is dominant over the self-repression by Mlc in the regulation of the mlcP2 promoter. The low binding affinity of Mlc to its own promoter that is 10 times weaker than that to ptsG or pts P0 promoters (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar) seems to be a major reason for the low influence of Mlc on regulation of its own gene. In vitro transcription assay revealed that each promoter of mlc has a different sensitivity to Mlc (Fig. 3). When cells were grown in the absence of glucose, a similar level of expression from both P1 and P2 was observed even though the P2 transcription is more sensitive to Mlc repression probably because the intracellular concentration of Mlc is limiting inE. coli (3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar). The condition seems to be similar to thein vitro transcription condition where both CRP·cAMP and a small amount of Mlc were present as shown in lane 6 of Fig.3 A. However, P1 was as active as P2 when the mlcstrain was grown in the absence of glucose. These results imply that the concentration of intracellular CRP·cAMP is lower than that of CRP·cAMP used for in vitro transcription reactions (40 nm) (39Ishizuka H. Hanamura A. Kunimura T. Aiba H. Mol. Microbiol. 1993; 10: 341-350Crossref PubMed Scopus (84) Google Scholar). When Mlc was induced and the concentration of CRP·cAMP was lowered by the addition of glucose in the growth medium, the P1 promoter was activated slightly while the P2 promoter was repressed because the P2 promoter is active only in the presence of CRP·cAMP. This situation is similar to the in vitrotranscription condition where neither CRP·cAMP nor Mlc were present (lane 1 of Fig. 3 A). Therefore, the addition of glucose in the growth medium resulted in the reduction ofmlc expression by about half. These results also agree with the previous report by Decker et al. (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar) that expression ofmlc is reduced by half when cells were grown in the presence of glucose by measuring the β-galactosidase activity of themlc-lacZ fusion. Level of Mlc expression can vary precisely in response to the available sugars but the variation range is less than 2-fold in that the availability of unphosphorylated EIICBGlc may be more critical than the intracellular level of Mlc for induction of the Mlc regulon by glucose as shown in our previous report (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar).Eς32 is involved in the transcription of themlc gene. In vitro transcription assay with Eς32 showed that Eς32 could recognize the P2 promoter of the mlc gene. Transcription of P2 was increased when ς32 was overexpressed (Fig.4 A). Moreover, mlc expression was increased upon heat shock. It is known that the intracellular concentration of ς32 in E. coli increases from 15–20-fold within 5 min then declines to a new steady-state level severalfold higher than the preshift level in response to temperature shift from 30 to 42 °C (20Yura T. Nagami H. Mori H. Annu. Rev. Microbiol. 1993; 47: 321-350Crossref PubMed Scopus (395) Google Scholar, 32Straus D.B. Walter W.A. Gross C.A. Nature. 1987; 329: 348-351Crossref PubMed Scopus (265) Google Scholar). Transcription from P2 recognized by Eς32 was induced transiently to an extraordinary level upon heat shock when cells were grown in the absence of glucose (Fig.4 B, lane 4). The level of mlctranscription was changed parallel to the changes in intracellular concentration of ς32. In addition, P2 transcription was activated upon heat shock even when cells were grown in the presence of glucose. These results imply that the major RNA polymerase which activated the P2 transcription upon heat shock was Eς32because Eς32 was less sensitive to Mlc than Eς70 and the P2 transcription by Eς32 was not dependent on CRP·cAMP as revealed by the in vitrotranscription assay (Fig. 3). It is not clear why the P1 transcription was reduced in the ΔrpoH strain and activated by heat shock or when cells overexpressing ς32 were grown in the presence of glucose even though P1 promoter was not recognized by Eς32 in vitro. Because heat shock should exert pleiotropic effects by regulating transcription of various genes (21Chuang S.E. Blattner F.R. J. Bacteriol. 1993; 175: 5242-5252Crossref PubMed Google Scholar), further study on the mechanism of heat shock is needed for a better understanding of these phenomena.To investigate whether the increased level of mlc expression resulting from heat shock can influence Mlc-dependent gene expression, we analyzed changes in ptsG expression by heat shock. The P1 expression of ptsG was increased significantly by heat shock only when cells were grown in the presence of glucose regardless of the presence of Mlc (Fig. 5 A). These results suggest that activation of ptsG P1 by heat shock was not mediated by Mlc or CRP·cAMP even though Mlc repression might be dominant over activation of ptsG P1 by heat shock. We have reported that the unphosphorylated form of EIICBGlcsequesters Mlc from its target promoters upon glucose uptake by direct protein-protein interaction (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar). Therefore, glucose is required to maximize the level of dephosphorylated EIICBGlc necessary to sequester Mlc that is increased by heat shock. However, contrary to the case of mlc P2 transcription in which Eς32plays a major role in its regulation, it is likely that additional factors independent of Eς32 are involved in regulation of the ptsG expression because ptsG P1 expression was increased partially upon heat shock even in theΔrpoH strain grown in the presence of glucose. It means that two separate mechanisms involving Eς32-dependent and Eς32-indendent activation of ptsG P1 may work additively for full activation of ptsG P1 by glucose when cells were heat-shocked. The importance of Eς32 in glucose induction of the Mlc regulon was manifested by the fact that glucose induction of ptsG could not be observed in theΔrpoH strain. The inability of glucose to activate ptsG expression resulted in an insensitivity of thepts P0 promoter to glucose in ΔrpoHstrain. We are trying to elucidate the mechanism of activation ofptsG transcription by heat shock in the absence of Eς32 in order to understand the general role of Eς32 in regulation of genes involving carbohydrate metabolism. It has been suggested that the mlc gene has two promoters, P1 and P2, which are separated by 12 bases and autoregulated by its product (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar). Here, we report that the transcription of themlc gene is regulated in a highly sophisticated manner and that heat shock ς factor, ς32, is involved in its transcription. It is known that the expression level of several genes encoding transcriptional repressors such as galS (33Weickrt M.J. Adhya S. J. Bacteriol. 1993; 175: 251-258Crossref PubMed Google Scholar),nagC (34Plumbridge J. J. Bacteriol. 1996; 178: 2629-2636Crossref PubMed Google Scholar), purR (35Meng L.M. Kilstrup M. Nygaard P. Eur. J. Biochem. 1990; 187: 373-379Crossref PubMed Scopus (68) Google Scholar), and trpR (36Kelley R.L. Yanofsky C. Proc. Natl. Acad. Sci. U. S. A. 1982; 79: 3120-3124Crossref PubMed Scopus (74) Google Scholar) ofE. coli is low and that their expression level is not modulated much in various growth conditions. It seems likely that both CRP·cAMP and Mlc work together in E. coli to maintain the level of Mlc optimum in response to availability of glucose. All genes known to be regulated negatively by Mlc, such as manXYZ,malT, ptsG, and pts, are also regulated positively by CRP·cAMP (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar, 3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar, 4Plumbridge J. Mol. Microbiol. 1998; 27: 369-380Crossref PubMed Scopus (102) Google Scholar, 5Plumbridge J. Mol. Microbiol. 1998; 29: 1053-1063Crossref PubMed Scopus (95) Google Scholar, 6Kim S.Y. Nam T.W. Shin D. Koo B.M. Seok Y.J. Ryu S. J. Biol. Chem. 1999; 274: 25398-25402Abstract Full Text Full Text PDF PubMed Scopus (54) Google Scholar, 7Plumbridge J. Mol. Microbiol. 1999; 33: 260-273Crossref PubMed Scopus (74) Google Scholar, 8Tanaka Y. Kimata K. Inada T. Tagami H. Aiba H. Genes Cells. 1999; 4: 319-399Crossref Scopus (49) Google Scholar). In vitrotranscription assay with Eς70 and CRP·cAMP showed two opposite effects on each promoter of mlc, that is the positive effect on P2 and the negative effect on P1 (Fig. 2). This can be explained based on the fact that the CRP-binding site of P2 centered at −71.5 to the transcription start site is more compatible for a functional CRP site (37Kolb A. Busby S. Buc H. Garges H. Adhya S. Annu. Rev. Biochem. 1993; 62: 749-795Crossref PubMed Google Scholar) compared with that of P1 centered at −58.5 to the transcription start site. However, Eς32-directed P2 transcription was insensitive to CRP·cAMP. The P2 promoter of mlc should be a good model system to assess the effect of ς factor on transcription activation by CRP·cAMP because it has been known that CRP·cAMP activates transcription by direct protein-protein interaction with the α-subunit of RNA polymerase (38Busby S. Ebright R.H. J. Mol. Biol. 1999; 293: 199-213Crossref PubMed Scopus (632) Google Scholar). Repression of the P2 transcription when cells were grown in the presence of glucose implies that the action of CRP·cAMP is dominant over the self-repression by Mlc in the regulation of the mlcP2 promoter. The low binding affinity of Mlc to its own promoter that is 10 times weaker than that to ptsG or pts P0 promoters (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar) seems to be a major reason for the low influence of Mlc on regulation of its own gene. In vitro transcription assay revealed that each promoter of mlc has a different sensitivity to Mlc (Fig. 3). When cells were grown in the absence of glucose, a similar level of expression from both P1 and P2 was observed even though the P2 transcription is more sensitive to Mlc repression probably because the intracellular concentration of Mlc is limiting inE. coli (3Kimata K. Inada T. Tagami H. Aiba H. Mol. Microbiol. 1998; 29: 1509-1519Crossref PubMed Scopus (73) Google Scholar). The condition seems to be similar to thein vitro transcription condition where both CRP·cAMP and a small amount of Mlc were present as shown in lane 6 of Fig.3 A. However, P1 was as active as P2 when the mlcstrain was grown in the absence of glucose. These results imply that the concentration of intracellular CRP·cAMP is lower than that of CRP·cAMP used for in vitro transcription reactions (40 nm) (39Ishizuka H. Hanamura A. Kunimura T. Aiba H. Mol. Microbiol. 1993; 10: 341-350Crossref PubMed Scopus (84) Google Scholar). When Mlc was induced and the concentration of CRP·cAMP was lowered by the addition of glucose in the growth medium, the P1 promoter was activated slightly while the P2 promoter was repressed because the P2 promoter is active only in the presence of CRP·cAMP. This situation is similar to the in vitrotranscription condition where neither CRP·cAMP nor Mlc were present (lane 1 of Fig. 3 A). Therefore, the addition of glucose in the growth medium resulted in the reduction ofmlc expression by about half. These results also agree with the previous report by Decker et al. (2Decker K. Plumbridge J. Boos W. Mol. Microbiol. 1998; 27: 381-390Crossref PubMed Scopus (87) Google Scholar) that expression ofmlc is reduced by half when cells were grown in the presence of glucose by measuring the β-galactosidase activity of themlc-lacZ fusion. Level of Mlc expression can vary precisely in response to the available sugars but the variation range is less than 2-fold in that the availability of unphosphorylated EIICBGlc may be more critical than the intracellular level of Mlc for induction of the Mlc regulon by glucose as shown in our previous report (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar). Eς32 is involved in the transcription of themlc gene. In vitro transcription assay with Eς32 showed that Eς32 could recognize the P2 promoter of the mlc gene. Transcription of P2 was increased when ς32 was overexpressed (Fig.4 A). Moreover, mlc expression was increased upon heat shock. It is known that the intracellular concentration of ς32 in E. coli increases from 15–20-fold within 5 min then declines to a new steady-state level severalfold higher than the preshift level in response to temperature shift from 30 to 42 °C (20Yura T. Nagami H. Mori H. Annu. Rev. Microbiol. 1993; 47: 321-350Crossref PubMed Scopus (395) Google Scholar, 32Straus D.B. Walter W.A. Gross C.A. Nature. 1987; 329: 348-351Crossref PubMed Scopus (265) Google Scholar). Transcription from P2 recognized by Eς32 was induced transiently to an extraordinary level upon heat shock when cells were grown in the absence of glucose (Fig.4 B, lane 4). The level of mlctranscription was changed parallel to the changes in intracellular concentration of ς32. In addition, P2 transcription was activated upon heat shock even when cells were grown in the presence of glucose. These results imply that the major RNA polymerase which activated the P2 transcription upon heat shock was Eς32because Eς32 was less sensitive to Mlc than Eς70 and the P2 transcription by Eς32 was not dependent on CRP·cAMP as revealed by the in vitrotranscription assay (Fig. 3). It is not clear why the P1 transcription was reduced in the ΔrpoH strain and activated by heat shock or when cells overexpressing ς32 were grown in the presence of glucose even though P1 promoter was not recognized by Eς32 in vitro. Because heat shock should exert pleiotropic effects by regulating transcription of various genes (21Chuang S.E. Blattner F.R. J. Bacteriol. 1993; 175: 5242-5252Crossref PubMed Google Scholar), further study on the mechanism of heat shock is needed for a better understanding of these phenomena. To investigate whether the increased level of mlc expression resulting from heat shock can influence Mlc-dependent gene expression, we analyzed changes in ptsG expression by heat shock. The P1 expression of ptsG was increased significantly by heat shock only when cells were grown in the presence of glucose regardless of the presence of Mlc (Fig. 5 A). These results suggest that activation of ptsG P1 by heat shock was not mediated by Mlc or CRP·cAMP even though Mlc repression might be dominant over activation of ptsG P1 by heat shock. We have reported that the unphosphorylated form of EIICBGlcsequesters Mlc from its target promoters upon glucose uptake by direct protein-protein interaction (9Nam T.W. Cho S.H. Shin D. Kim J.H. Jeong J.Y. Lee J.H. Roe J.H. Peterkofsky A. Kang S.O. Ryu S. Seok Y.J. EMBO J. 2001; 20: 491-498Crossref PubMed Scopus (105) Google Scholar). Therefore, glucose is required to maximize the level of dephosphorylated EIICBGlc necessary to sequester Mlc that is increased by heat shock. However, contrary to the case of mlc P2 transcription in which Eς32plays a major role in its regulation, it is likely that additional factors independent of Eς32 are involved in regulation of the ptsG expression because ptsG P1 expression was increased partially upon heat shock even in theΔrpoH strain grown in the presence of glucose. It means that two separate mechanisms involving Eς32-dependent and Eς32-indendent activation of ptsG P1 may work additively for full activation of ptsG P1 by glucose when cells were heat-shocked. The importance of Eς32 in glucose induction of the Mlc regulon was manifested by the fact that glucose induction of ptsG could not be observed in theΔrpoH strain. The inability of glucose to activate ptsG expression resulted in an insensitivity of thepts P0 promoter to glucose in ΔrpoHstrain. We are trying to elucidate the mechanism of activation ofptsG transcription by heat shock in the absence of Eς32 in order to understand the general role of Eς32 in regulation of genes involving carbohydrate metabolism. We are grateful to Dr. C. Park for providing KD413 and Dr. T. Yura for providing bacterial strains and a pKV10.