- Email: [email protected]
Positive control of transcription initiation in Escherichia coli
J. Mol. Biol. (1985) 184. l-6
Positive Control of Transcription in Escherichia coli
Initiation
A Base Substitution at the Pribnow Box Renders ompF Expression Independent of a Positive Regulator Tohru Dairi, Kaoru Inokuchi, Takeshi Mizuno and Shoji Mizushima Laboratory of ,Wicrobiology, Faculty of Agriculture Nagoya University, Chikusa-Eu, Nagoya 464, Japan (Received 10 December 1984) Expression of the ompF gene coding for a major outer membrane protein of Escherichia coli is positively regulated by the product of the ompR gene, OmpR. Using an ompF-tet chimera gene, ompF promoter mutants that render the ompF expression independent of the OmpR protein were isolated. In all of the four mutants that were isolated separately, the first, base of the Pribnow box was changed from A to T. The mutant promoter did not require the upstream domain of the -35 region that is required for the OmpR-dependent funct,ioning of the wild-type promoter. It is concluded that the domain upstream from the -35 region plays a role in the positive regulation by the OmpR protein. A statistical survey of the E. coli promoter sequence revealed that almost’ all of the genes that’ do not require an activator protein for their expression possess T at the first position of the Pribnow box, while the position is occupied by other bases in almost all of the positively regulated genes. Based on these facts, the mechanism of positive regulation of the gene expression by an activator protein is discussed.
17 nucleotides downstream from t’he start site. Furthermore, from the results of analysis of the ompF-tet hybrid promoter, it was suggested that the domain responsible for regulation by the OmpR protein is located within the -35 region and the region upstream from it. In this work, in order to study the molecular mechanism of positive regulation of gene expression in detail, we isolated promoter mutants of the ompF gene that functioned independently of the OmpR protein. The conversion of the first base of the Pribnow box from A to T rendered the gene expression OmpR independent. Simultaneously, the requirement for the domain upstream from the - 35 region disappeared. Based on these observations, together with the known promoter sequences, the mechanism of regulation by an activator protein of positively regulated genes is discussed.
1. Introduction Expression of many genes is positively regulated by a specific activator protein in Escherichia coli. The MalT protein is the activator for t’he malPQ, malK and malEFG operons (Hofnung et al., 1971; lXbarbouill6 8: Schwartz, 1979) and the OmpR protein is t’he activator for the ompF and ompC genes (Hall & Silhavy, 1979, 1981aJ). The AcII protein activates the transcription initiated at P,, (R’eichardt, & Kaiser, 1971; Echols & Green, 1971). For expression of these positively regulated genes. a nucleotide domain upstream from the promoters is required (Raibaud et al., 1983; Bedouelle, 1983; Tnokuchi et al., 1984; Ho et al., 1983). Although the domain has been suggested to be involved in the interaction with an activator protein, the molecular mechanism of the positive regulation including the role of the upstream nucleotide domain is unclear. The ompF gene codes for a major outer membrane protein. The gene was cloned (Mutoh et sequence al., 1981) and its total nucleotide determined (Inokuchi et al., 1982). The domain that is responsible for the OmpR-dependent functioning of the ompF promoter has been determined (Inokuchi et nl., 1984). The domain consists of 90 nucleotides upstream from the mRNA start site, including the -35 region and the Pribnow box, and 0022-2X36/85/
1300014%
$oa.oo/o
2. Materials
and Methods
(a) Bacteria and plasmid The following strains derived from E. coli K-12 were used: AB2847 (F- aroB tsx malA supE Z 2: Sato $ Yura, 1981), KY2562 (F- tsx malA ompBlO1 supE I- 1’: Sato & Yura, 1981) and KY2562 recA (a recA mutant of KY2562; a gift from F. Pjara). The plasmid used was pHF129. This plasmid is ampicillin resistant (Ma+) and
1
0
1985 Academic
Press 111~. (London)
Ltd.
T. Dairi
2
et al.
ompF gene
pER322
tef gene v
V
ompF
protem \
I
N termmus
Slgnol
sequence I --
-----t r&1
+I
-113-115-111
-358
I HueJIL
II EcoRI
l-i -35
Ill PB
fill n
V
t122
Cl77
I I
SDATG
H/w1
ret product \
+210t225+249+2L--zII I BgfU
TAP,
SD ATG Hq un
Figure 1. Structure of the ompF-tet junction region and the region upstream from it in pHF129. -35, PB and SD denote the -35 region, Pribnow box and Shine-Dalgarno sequence, respectively. Initiation codons (ATG) and a termination codon (TAA) are indicated. The mRNA start site of the ompF gene is numbered + 1. For the nucleotide sequencing, the EcoRI-BgZII fragment (A) or the HueIII-HhaI fragment (B) was used. The TugI-Hap11 fragment (C) was used for S, nuclease mapping of the mRNA start site with 32P-labeling at the Hap11 end. carries tetracycline-resistant genes @et genes) that are under the control of the ompF promoter (Inokuchi et al., 1984: and see Fig. 1).
used were the TuqI-Hap11 fragments from either pHFld9 or pHF129M (see Fig. 1). S, mapping was performed as described (Inokuchi et al., 1984). The hybridizat,ion temperature was 40°C.
(b) Media and other materials All experiments except for the ompF promoter activity When required, assay were carried out in L-broth. ampicillin and tetracycline were added at concentrations of 50 pg/ml and 15pg/ml, respectively. For solid cultivation, media were supplemented with 1.5% (w/v) agar. Restriction endonucleases, bacteriophage T4 ligase, exonuclease BuZ31, S, nuclease, the Klenow fragment of synthetic EcoRI linker DNA polymerase I, (dG-G-A-A-T-T-C-C) and a dideoxy sequencing kit were obtained from Takara Shuzo Co. (c) DNA technique8 Plasmid DNA was prepared as described by Birnboim & Doly (1979). The conditions used for restriction endonuclease reactions, exonuclease Bul31 digestions and Klenow fragment treatments were those proposed by the manufacturers. Analyses of plasmid DNA fragments were performed by electrophoresis in 08% (w/v) agarose gel or 5oj, (w/v) polyacrylamide gel. Pat1 fragments of bacteriophage 1 DNA were used as molecular weight standards. The buffer for electrophoresis was 50 mw-Tris-borate 1 mM-Na,EDTA. DNA fragments were ($3 8.3) recovered from the gel by electroelution as described (Maniatis et al., 1982). Ligation was performed as described (Inokuchi et al., 1984). DNA sequencing was carried out either by the method of Maxam & Gilbert (1980) or by the dideoxy method (Sanger et al., 1977). In the latter case, isolated restriction fragments were cloned into a single-strand DNA cloning vector, M13mp8. JM103 was used as a host strain. Transformation was carried out by the method of Dagert & Ehrlich (1979). (d) S, n&ease mapping h’. coli KY2562 recA or AB2847 harboring either a recombinant plasmid or the wild-type plasmid was grown Ampicillin (50pg/ml) was to A,,, = 0.8 in L-broth. added to the medium for cells harboring a plasmid. The cells were collected by centrifugation and total RNA was isolated as described (Brosius et al., 1982). 32P-labeled probes were used for S, nuclease mapping. The probes
(e) Assaying of the ompF promoter activity The ompF promoter activity of each recombinant plasmid carrying the ompF-tet gene was examined in terms of the extent of tetracycline resistance. E. coli cells grown in 10 ml of L-broth (A,,, = 0.2) were transformed with each recombinant plasmid (100 ng of DNA) in 0.1 ml of 0.1 M-CaCI,, and 0.05 ml portions were spread on medium A plates (Kawaji et al.. 1979) with or without 15% (w/v) sucrose. The medium also contained ampicillin (50 pg/ml) and tetracycline (O? 15. 25 or 35 pg/ml). The colonies that appeared after 40 h incubation at 37°C were counted. The extent of tetracycline resistance was expressed as the ratio of the number of tetracyclineresistant colonies to that of ampicillin-resistant colonies.
3. Results (a) Isolation of ompF promote,r mutants that function independently of the ompR ge’ne
Plasmid pHF129 was used as a starting material. It carries tet genes that are under the control of the ompF promoter (Inokuchi et al.. 1984). From the sequence shown in Figure 1, it is assumed that the transcription initiated at the ompF promoter continues to the end of the tet genes. In contrast. the translation initiated at the signal sequence for the OmpF protein is assumed to terminate at TAA beyond the ompF-tet junction and to be reinitiated for .the tet genes to render the cells tetracycline resistant. The Shine-Dalgarno sequence of the tet genes remains intact in this plasmid. Since the expression of the ompF gene is positively regulated at the promoter region by the OmpR protein cells harboring pHF129 are tetracycline sensitive and resistant in the ompRand ompR+ backgrounds. respectively (Inokuchi et aE., 1984). Strain KY2562 (ompR-) harboring pHF129 was plated on L-broth containing ampicillin (50 pg/ml)
Positive
Regulation
in E. coli
( a ‘I G
A
T
C
<.iL
-14 cc *TCTTA
tt
A&
-TG~
I
Klenow
fragment
EcoRI
ltnker
fCORI/SOlI
(b)
-I4
4 . I TAGCGAAACGTTAGTTTGAATGGAAAGATGCCTGC: ATCGCTTTGCAATCAAACTTACCTTTCTACGGACGT -35
EcuRI-Sol1 large fragment of pHFl29
+I
EcoRI
so/ I
Bg/D
PB
Figure 2. Nucleotide sequence (hyphens omitted for clarity) of the OWL@ promoter region of the OmpRindependent om$‘-tet gene. (a) The EcoRI-BgZII fragment was prepared from pHF129M (see Fig. 1) and sequenced by the dideoxy method. The antisense strand was used. The positions of + 1 and - 14 are indicated by arrows. (b) The nucleotide sequence of the wild-type and mutant promoters. The base-pair altered in the mutant plasmid is shown in a box. The -35 region ( - 35) and the Pribnow box (PB) are indicated. and tetracycline (15 pg/ml), and four tetracyclineresistant colonies were isolated independently. Plasmids were isolated from these colonies and KY2562 recA was transformed with them. All the transformants were tetracycline resistant, indicating that the mutations responsible for the tetracycline resistance resided in these plasmids. To further confirm that the mutations were within the ompF promoter region, the 320 nucleotide EcoRIBglII fragment that carries the ompF promoter region was exchanged between pHF129 and the mutant, plasmids (see Fig. 1). Replacement of the EcoRI-BgZII of pHF129 by that derived from mutant plasmids rendered the transformed cells t’etracycline resistant in the ompR- background. (b) Determination
of the mutations
The EcoRT-BgZTT fragments were prepared from the mutant plasmids and the nucleotide sequences were determined (Fig. 2). In all the mutants that were obtained independently, the A at the first position of the Pribnow box was altered to T. (c) The upstream region of the ompF required for the OmpR-independent
For the nucleotides
l Llgose
promoter is not expression
ompF promoter function, the upstream from the -35 region
55 are
Figure 3. Deletion mutagenesis in &TO of the OmpRindependent, ompF promoter. pHF129M was used to isolate the upstream deletions. Filled bars represent the ompF promoter region. The approximate location and direction of the ompF promoter (F& is indicated. The bla gene and the tet genes that are fused to the ompF gene are indicated by thin arrows. The EcoRI-Sal1 large fragment of pHF129 IS the same as that of pHF129M. required and this domain is assumed to be involved in the posit’ive regulation by the OmpR protein (Inokuchi et al., 1984). Therefore, it is reasonable to assume that the region is dispensable for a mutant promoter whose function is independent of the OmpR protein. As shown in Figure 3, pHF129M, one of the mutant plasmids, was linearized with EcoRI at t’he position 111 nucleotide upstream from the mRNA start sit’e; and digested with exonuclease Bal31 to produce the ompF promoter region deleted to varying extents. After t,reatment with the Klenow enzyme, the resultant blunt-ended fragments were ligated with EcoRI linker. The ligated products were digest,ed with EcoRI and SalI, and the DNA fragments caarrying t,he ompF promoter with different deletions were ligated with the EcoRI-Sal1 large fragments of pHF129. Plasmids t’hus constructed were transferred to KY2562 recA, and ampicillin-resistant clones were selected. The extent, of deletion was estimated by sizing the fragments generated by EcoRT-HgZTT digestion on 57; polyacrylamide gel electrophoresis. For some of them, the endpoint of the deletion was determined accurately by DNA sequencing. The results are shown in Figure 4. The plasmids carrying these deletions were named pTDA1 through pTDA10. The ten deletions thus selected were examined as
4
T. Dairi
et al.
-100
-80
pTDA4
pTDA6
-60
pTDA8
Figure 4. Nucleotide sequence of the promoter region and the region upstream from it of the deletions derived from the OmpR-independent ompF promoter of pHF129M. The nucleotide sequences of individual deletion mutants were determined by the method of Maxam & Gilbert (1980) .by using the E&III-HhaI fragments (see Fig. 1) with 32P-labeling at the Hoe111 end. The upstream ends of the individual deletions are indicated. The mRh’A start site of the ompF gene is numbered + 1. The - 35 region ( - 35) and Pribnow box (PB) are indicated. Sequence hyphens are omitted for clarity. to the ompF promoter function in terms of tetracycline resistance of the transformants in the ompRbackground (Fig. 5). First of all. the activity of the mutant ompF promoter in the ompR- background was as high as that of the wildtype promoter in the ompR’ background. Secondly, as expected, the OmpR-independent expression of the mut’ant promot,er did not require the domain
HFl29M
\1
I
pTDA’ I
I -100
pTDA2
c I -80
upstream from the -35 region. However. no promoter activity was observed when the deletion reached the -35 region wit,h the exception of pTDA7 that weakly showed t~etracycline resistance. In pTD,47, the first three ba,ses of the -35 region had been delet’ed (Fig. 4). However. as a result of the linkage with the EcoR81 linker a possible -35 region, T-C-C-C-(:-A. was newly formed in the
pT?A3
pT?A4
pTDA5\\/
$1 c 4 I %~Lh
-60 Base-
-40
-35
----L-J -20
PB
+’
pale
Figure 5. Effect of upstream deletions on the OmpR-independent ompF promoter activity of pHFl29M. The upstream ends of individual deletions are indicated by arrows, The mRNA start site is numbered + 1 and the positions of the - 35 region ( - 35) and Pribnow box (PB) are indicated. The ompF promoter activity of the deletions was assayed in the ompR- background (KY2562 recA) and expressed as the ratio of the number of tetracycline-resistant colonies to that of ampicillin-resistant colonies as described in Materials and Methods. Concentrations of tetracycline in the media were (a) 15 pg/ml, (0) 25 pg/ml and (A) 35 pg/ml. The activity of the wild-type ompF promoter in pHF129 was assayed in the ompR+ background (AB2847) in the presence of 15 pg tetracycline (A).
Positive Regulation in E. coli corresponding region. In contrast to this, deletion of the T at the firsi position of the -35 region resulted in the complete loss of the promoter activity (pTDA6). The importance of the T at the first position of the -35 region was proposed by Hawley & McClure (1983) and has been shown in this work. This is discussed further below. The mRNA st’art, site was determined for the OmpR-independent mutant promoter and some of the deletion promoters (pHF129M. pTDA1. pTDA2, pTDA3 and pTDA5) by S, nuclease mapping. In all cases. the start site was the same as cwnpF promoter, indicating that of the wild-type that t’he OmpR-independent expression of these promoters was not due to the formation of an artificial promoter (data not, shown). Th e promoter activity of t,he deletions was examined also in the ompR’ background using AB2847 as a host’ strain. The results were essent~ia,lly the same as those for the ompKbackground.
4. Discussion (a) Role of the OmpR protein in expression of the ompF gene The ompF gene coding for a major outer regulated membrane protein of E. coli is positively by OmpR: a product of the ompR gene. In a previous work (Inokuchi et al., 1984), the 107 basepair fragment was found to function as the ompF promoter. The fragment consists of 90 nucleotides upstream from the mRNA start site including the and the Pribnow box, and 17 - 35 region nucleotides downstream from the start site. Although the sequence at. the -35 region and that at the Pribnow box show low degrees of homology to the respective consensus sequences, the ompF gene is one of the most efficiently expressed genes in E. coli cells in the ompR’ background. It is assumed that the OmpR protein interacts with the domain upstream from the -35 region, which in turn facilitat,es the efficient binding of RNA polymerase to the ompF promoter region. To confirm this hypothesis, promoter mutants that show OmpR-independent expression of the wnpF gene were isolated with the aid of the ompFtut hybrid gene. In all four promoter mutants that were isolated independently, the first base of the Pribnow box was changed from A to T. No ot,her changes were observed. Upon the mutation, the became independent of both oNIp F expression OmpR, and the domain upstream from the -35 region but, on the other hand, the mRNA start site was not changed. These results strongly support the view that) the domain upstream from the -35 region is responsible for the positive regulation by the OmpR protein; most probably, the OmpR protein binds t,o this domain to induce the efficient transcription initiation. Recently, precise analysis of t.hr o,mpC’ gene, a gene coding for another major out.er membrane protein of E. coZi, was carried out in this laboratory (T. Mizuno & S. Mizushima,
5
unpublished results). The ompC gene is also positively regulated by the OmpR protein, and a domain upstream from the -35 region was found to be required for the gene expression. This domain shows a high degree of homology to the corresponding domain of the OmpF gene. (In) Positive regulation
and the Pribnow box
The conversion of the first base of the Pribnow box from A t.o T rendered the ompF gene activatorindependent. This suggests that. the T at the first. position of !,he Pribnow box may be important for the expression of a promoter that does not require an activator. From the results of statistical studies on about 106 promoters of E. co& genes, a so-called consensus sequence (T-A-T-A-A-T) was proposed for the Pribnow box (Hawley & McClure, 1983). Among the six bases, T. A and T at the first, second and sixt’h positions, respectively, are highly conserved, indicating the importance of these bases in recognition by RNA polymerase. Consistent with our observation in the present work, the first base was found to be almost exclusively T for promoters that do not require a positive regulator. An exception is the major lipoprotein gene for which A-AT-A-C-T was proposed with 18 base-spacing from the -35 region. However, this would preferably be T-A-A-T-A-C with 17 base-spacing. In contrast to the promoters that do not require a positive regulator, most of those requiring one do not possess T at this position. They are ompC (G-A-G-A-AT: Mizuno et al., 1984). maEPQ ((I-A-A-A-C-T), malK (C-A-T-A-G-T), malEFG (G-A-A-A-C-T: Raibaud et al., 1983; Bedouelle, 1983) and Pa, (A-A-G-T-A-T: Ho et al., 1983). The A and T at the second and sixth positions are conserved in these promoters. Furbhermore, these genes possess an additional domain upstream from the -35 region that is a possible interaction site for et al., 1983; an activator protein (Raibaud Bedouelle. 1983; Inokuchi ct nl.. 1984: Ho et al., 1983). Based on these facts, the following hypothesis can be proposed for positive regulation. (1) The T of the Pribnow box is at the first position indispensable for the functioning of a promoter that does not, require an activator protein; replacement of t.he T by other bases resuhs in the disappearance of the promoter function. (2) However. the defect can be compensated for by t’he presence of both an activator protein and a proper nucleotide domain upstream from the -35 region. (3) The activator protein possibly interacts with the upstream domain to facilitate the binding of RNA polymerase to t’he promoter region. Consistent with this hypothesis, transcription initiation of the malEFG operon becomes independent of the activator protein (MalT) upon conversion by mutation of the G at, the first position of the Pribnow box to T (Bedouelle, 1983). It should be mentioned, however, that the T of the Pribnow box is not t,he only base that
T. Dairi
6
determines positive regulation. In a previous work, we constructed a hybrid promoter in which the - 35 region (T-A-G-C-G-A) and the region upstream from it are from the ompF gene and Pribnow box (T-T-T-A-A-T) from the tet genes (Inokuchi et al., 1984). Functioning of the hybrid promoter was OmpR dependent. When the -35 region was further replaced by that’ from tet genes leaving the upstream region from the ompF gene as it was, the promoter function became OmpR independent (Y. Ozawa et aZ., unpublished results). These results indicate that the -35 region may also have something to do with the positive regulation. In many positively regulated genes. the -35 region is rather atypical, showing a low homology to the consensus sequence (Rosenberg & Court, 1979: Bedouelle et al., 1982: DBbarbouilE & Raibaud, 1983; Ho et al., 1983; Inokuchi et al., 1984). In addition to the regulation by the ompR gene, expression of the ompF gene is affected by the envZ gene (Hall & Silhavy, 1981a,h) and medium osmolarity (van Alphen & Lugtenberg, 1977; Kawaji et al., 1979). The functions of these effecters in the OmpR independent mutant were, however, difficult to study with the ompF-tet system we used in this work. We thank S. Teranishi for excellent secretarial assistance. This work was supported by grants from the Ministry of Education, Science and Culture of Japan and the Special Co-ordination Fund for Science Technolog? Agency of Japan. References Bedouelle, H. (1983). J. Mol. Biol. 170, 861-882. Bedouelle, H., Schmeissner, U.. Hofnung, M. & Rosenberg, M. (1982). J. Mol. Biol. 161. 519-531. Birnboim, H. C. & Doly. J. (1979). Nucl. Acids Res. 7, 1513-1523. Brosius, J., Cate, R. L. & Perlmutter, A. P. (1982). J. Biol. Chem. 257, 9205-9210. Dagert, M. & Ehrlich, S. D. (1979). Gene, 6, 23-28.
et al. D6barbouillB. M. & Raibaud, 0. (1983). .J. Bacterial. 153. 1221-1227. DBbarbouillB, M. & Schwartz, M. (1979). ,J. No/. Riol. 132. 521-534. Echols. H. & Green, I,. (197 1). Proc. ,Vut. Acud. Sri., U.S.A. 68, 2190-2194. Hall. M. X. & Silhavy. T. J. (1979). J. BactevioZ. 140. 843h 847. Hall, M. N. & Silhavy. T. tJ. (19810). .I. Mol. Biol. 146. 23-43. Hall, M. N. & Silhavy, T. .J. (1981b). .J. Mol. Kiol. 151. l-15. Hawley. D. K. & McClure. W. R. (1983). XUCI. &ids Res. 11, 2239-2255. Ho. Y.-S., Wulff, D. I,. & Rosenberg. M. (1983). .Yrtl~r(, (London), 304. 703 708. Hofnung. M.. Schwartz, M. 8: Hatfield. I). (1971). J. No/. Riol. 61. 681-694. Inokuchi. K.. Wutoh. X., ,\llatsuyamn, S. & Ylizushima. S. (1982). ic’ucl. Acids RPS. 10, 6957--6968. Tnokuchi. K.. Furukawa. H.. h’akamura. K. B Mizushima. S. (1984). ,I. Mol. Viol. 178. 653-668. Kawaji. H.. Vizuno, T. & Mizushima. S. (19%). .I. Bnrtrriol. 140. 84S847. Maniatis. ‘I’.. Frit,scah. 15. F. & Sambrook. ,I. (1982). 111 Molec~ular Cloning. pp. 150-I 72. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Maxam, A. M. 8: (Gilbert. W. (1980). In Jlrfhods i)t EnzymoZogy ((:rossman. I,. & Moldavr. E;.. rds), vol. 65. pp. 499-560. Academir Press. Kew York. Mizuno. T.. Chou, Yl .-Y. 8r Inouyr, 11. (1984). t’roc.. .z’n/. Acad. Sri.. [-.A’.,~. 81. 1966Gl970. Mutoh. N., n’agasawa, T. 6 Mizushima. S. (l!tHI). ,J. Hactrriol. 145. 10X5-1090. Raibaud, O., Ddbarbouill@. M. & Schwartz. Al. (I9X:l). J. Mol. Biol. 163, 395-~408. Rricahardt,. L. & Kaiser. A. I). (1971 ). /‘rot. .Vnf. .,IuIc/. Sci., I’.R.d.
68. %I85 Zl89.
Rosenberg. iV. & (‘out?. I). (1979). ;Z,t~r(. fZv11.(:rnet. 13. 319-353. Sanger. F.. Nicklen. S. & (‘oulxon . A. R. (1!177). /‘/YN~..Z;rf. A cud. Sci., U.S.d
74. 5463-5467.
Sato. T. & Yura. T. (1981). ,J. Kacfrriol. 145. kW96. van Alphen. W. & Lugtenberg. K. (1977). .J. b’crctrriol. 131, 62~3.-630.
Positive Control of Transcription in Escherichia coli
Initiation
A Base Substitution at the Pribnow Box Renders ompF Expression Independent of a Positive Regulator Tohru Dairi, Kaoru Inokuchi, Takeshi Mizuno and Shoji Mizushima Laboratory of ,Wicrobiology, Faculty of Agriculture Nagoya University, Chikusa-Eu, Nagoya 464, Japan (Received 10 December 1984) Expression of the ompF gene coding for a major outer membrane protein of Escherichia coli is positively regulated by the product of the ompR gene, OmpR. Using an ompF-tet chimera gene, ompF promoter mutants that render the ompF expression independent of the OmpR protein were isolated. In all of the four mutants that were isolated separately, the first, base of the Pribnow box was changed from A to T. The mutant promoter did not require the upstream domain of the -35 region that is required for the OmpR-dependent funct,ioning of the wild-type promoter. It is concluded that the domain upstream from the -35 region plays a role in the positive regulation by the OmpR protein. A statistical survey of the E. coli promoter sequence revealed that almost’ all of the genes that’ do not require an activator protein for their expression possess T at the first position of the Pribnow box, while the position is occupied by other bases in almost all of the positively regulated genes. Based on these facts, the mechanism of positive regulation of the gene expression by an activator protein is discussed.
17 nucleotides downstream from t’he start site. Furthermore, from the results of analysis of the ompF-tet hybrid promoter, it was suggested that the domain responsible for regulation by the OmpR protein is located within the -35 region and the region upstream from it. In this work, in order to study the molecular mechanism of positive regulation of gene expression in detail, we isolated promoter mutants of the ompF gene that functioned independently of the OmpR protein. The conversion of the first base of the Pribnow box from A to T rendered the gene expression OmpR independent. Simultaneously, the requirement for the domain upstream from the - 35 region disappeared. Based on these observations, together with the known promoter sequences, the mechanism of regulation by an activator protein of positively regulated genes is discussed.
1. Introduction Expression of many genes is positively regulated by a specific activator protein in Escherichia coli. The MalT protein is the activator for t’he malPQ, malK and malEFG operons (Hofnung et al., 1971; lXbarbouill6 8: Schwartz, 1979) and the OmpR protein is t’he activator for the ompF and ompC genes (Hall & Silhavy, 1979, 1981aJ). The AcII protein activates the transcription initiated at P,, (R’eichardt, & Kaiser, 1971; Echols & Green, 1971). For expression of these positively regulated genes. a nucleotide domain upstream from the promoters is required (Raibaud et al., 1983; Bedouelle, 1983; Tnokuchi et al., 1984; Ho et al., 1983). Although the domain has been suggested to be involved in the interaction with an activator protein, the molecular mechanism of the positive regulation including the role of the upstream nucleotide domain is unclear. The ompF gene codes for a major outer membrane protein. The gene was cloned (Mutoh et sequence al., 1981) and its total nucleotide determined (Inokuchi et al., 1982). The domain that is responsible for the OmpR-dependent functioning of the ompF promoter has been determined (Inokuchi et nl., 1984). The domain consists of 90 nucleotides upstream from the mRNA start site, including the -35 region and the Pribnow box, and 0022-2X36/85/
1300014%
$oa.oo/o
2. Materials
and Methods
(a) Bacteria and plasmid The following strains derived from E. coli K-12 were used: AB2847 (F- aroB tsx malA supE Z 2: Sato $ Yura, 1981), KY2562 (F- tsx malA ompBlO1 supE I- 1’: Sato & Yura, 1981) and KY2562 recA (a recA mutant of KY2562; a gift from F. Pjara). The plasmid used was pHF129. This plasmid is ampicillin resistant (Ma+) and
1
0
1985 Academic
Press 111~. (London)
Ltd.
T. Dairi
2
et al.
ompF gene
pER322
tef gene v
V
ompF
protem \
I
N termmus
Slgnol
sequence I --
-----t r&1
+I
-113-115-111
-358
I HueJIL
II EcoRI
l-i -35
Ill PB
fill n
V
t122
Cl77
I I
SDATG
H/w1
ret product \
+210t225+249+2L--zII I BgfU
TAP,
SD ATG Hq un
Figure 1. Structure of the ompF-tet junction region and the region upstream from it in pHF129. -35, PB and SD denote the -35 region, Pribnow box and Shine-Dalgarno sequence, respectively. Initiation codons (ATG) and a termination codon (TAA) are indicated. The mRNA start site of the ompF gene is numbered + 1. For the nucleotide sequencing, the EcoRI-BgZII fragment (A) or the HueIII-HhaI fragment (B) was used. The TugI-Hap11 fragment (C) was used for S, nuclease mapping of the mRNA start site with 32P-labeling at the Hap11 end. carries tetracycline-resistant genes @et genes) that are under the control of the ompF promoter (Inokuchi et al., 1984: and see Fig. 1).
used were the TuqI-Hap11 fragments from either pHFld9 or pHF129M (see Fig. 1). S, mapping was performed as described (Inokuchi et al., 1984). The hybridizat,ion temperature was 40°C.
(b) Media and other materials All experiments except for the ompF promoter activity When required, assay were carried out in L-broth. ampicillin and tetracycline were added at concentrations of 50 pg/ml and 15pg/ml, respectively. For solid cultivation, media were supplemented with 1.5% (w/v) agar. Restriction endonucleases, bacteriophage T4 ligase, exonuclease BuZ31, S, nuclease, the Klenow fragment of synthetic EcoRI linker DNA polymerase I, (dG-G-A-A-T-T-C-C) and a dideoxy sequencing kit were obtained from Takara Shuzo Co. (c) DNA technique8 Plasmid DNA was prepared as described by Birnboim & Doly (1979). The conditions used for restriction endonuclease reactions, exonuclease Bul31 digestions and Klenow fragment treatments were those proposed by the manufacturers. Analyses of plasmid DNA fragments were performed by electrophoresis in 08% (w/v) agarose gel or 5oj, (w/v) polyacrylamide gel. Pat1 fragments of bacteriophage 1 DNA were used as molecular weight standards. The buffer for electrophoresis was 50 mw-Tris-borate 1 mM-Na,EDTA. DNA fragments were ($3 8.3) recovered from the gel by electroelution as described (Maniatis et al., 1982). Ligation was performed as described (Inokuchi et al., 1984). DNA sequencing was carried out either by the method of Maxam & Gilbert (1980) or by the dideoxy method (Sanger et al., 1977). In the latter case, isolated restriction fragments were cloned into a single-strand DNA cloning vector, M13mp8. JM103 was used as a host strain. Transformation was carried out by the method of Dagert & Ehrlich (1979). (d) S, n&ease mapping h’. coli KY2562 recA or AB2847 harboring either a recombinant plasmid or the wild-type plasmid was grown Ampicillin (50pg/ml) was to A,,, = 0.8 in L-broth. added to the medium for cells harboring a plasmid. The cells were collected by centrifugation and total RNA was isolated as described (Brosius et al., 1982). 32P-labeled probes were used for S, nuclease mapping. The probes
(e) Assaying of the ompF promoter activity The ompF promoter activity of each recombinant plasmid carrying the ompF-tet gene was examined in terms of the extent of tetracycline resistance. E. coli cells grown in 10 ml of L-broth (A,,, = 0.2) were transformed with each recombinant plasmid (100 ng of DNA) in 0.1 ml of 0.1 M-CaCI,, and 0.05 ml portions were spread on medium A plates (Kawaji et al.. 1979) with or without 15% (w/v) sucrose. The medium also contained ampicillin (50 pg/ml) and tetracycline (O? 15. 25 or 35 pg/ml). The colonies that appeared after 40 h incubation at 37°C were counted. The extent of tetracycline resistance was expressed as the ratio of the number of tetracyclineresistant colonies to that of ampicillin-resistant colonies.
3. Results (a) Isolation of ompF promote,r mutants that function independently of the ompR ge’ne
Plasmid pHF129 was used as a starting material. It carries tet genes that are under the control of the ompF promoter (Inokuchi et al.. 1984). From the sequence shown in Figure 1, it is assumed that the transcription initiated at the ompF promoter continues to the end of the tet genes. In contrast. the translation initiated at the signal sequence for the OmpF protein is assumed to terminate at TAA beyond the ompF-tet junction and to be reinitiated for .the tet genes to render the cells tetracycline resistant. The Shine-Dalgarno sequence of the tet genes remains intact in this plasmid. Since the expression of the ompF gene is positively regulated at the promoter region by the OmpR protein cells harboring pHF129 are tetracycline sensitive and resistant in the ompRand ompR+ backgrounds. respectively (Inokuchi et aE., 1984). Strain KY2562 (ompR-) harboring pHF129 was plated on L-broth containing ampicillin (50 pg/ml)
Positive
Regulation
in E. coli
( a ‘I G
A
T
C
<.iL
-14 cc *TCTTA
tt
A&
-TG~
I
Klenow
fragment
EcoRI
ltnker
fCORI/SOlI
(b)
-I4
4 . I TAGCGAAACGTTAGTTTGAATGGAAAGATGCCTGC: ATCGCTTTGCAATCAAACTTACCTTTCTACGGACGT -35
EcuRI-Sol1 large fragment of pHFl29
+I
EcoRI
so/ I
Bg/D
PB
Figure 2. Nucleotide sequence (hyphens omitted for clarity) of the OWL@ promoter region of the OmpRindependent om$‘-tet gene. (a) The EcoRI-BgZII fragment was prepared from pHF129M (see Fig. 1) and sequenced by the dideoxy method. The antisense strand was used. The positions of + 1 and - 14 are indicated by arrows. (b) The nucleotide sequence of the wild-type and mutant promoters. The base-pair altered in the mutant plasmid is shown in a box. The -35 region ( - 35) and the Pribnow box (PB) are indicated. and tetracycline (15 pg/ml), and four tetracyclineresistant colonies were isolated independently. Plasmids were isolated from these colonies and KY2562 recA was transformed with them. All the transformants were tetracycline resistant, indicating that the mutations responsible for the tetracycline resistance resided in these plasmids. To further confirm that the mutations were within the ompF promoter region, the 320 nucleotide EcoRIBglII fragment that carries the ompF promoter region was exchanged between pHF129 and the mutant, plasmids (see Fig. 1). Replacement of the EcoRI-BgZII of pHF129 by that derived from mutant plasmids rendered the transformed cells t’etracycline resistant in the ompR- background. (b) Determination
of the mutations
The EcoRT-BgZTT fragments were prepared from the mutant plasmids and the nucleotide sequences were determined (Fig. 2). In all the mutants that were obtained independently, the A at the first position of the Pribnow box was altered to T. (c) The upstream region of the ompF required for the OmpR-independent
For the nucleotides
l Llgose
promoter is not expression
ompF promoter function, the upstream from the -35 region
55 are
Figure 3. Deletion mutagenesis in &TO of the OmpRindependent, ompF promoter. pHF129M was used to isolate the upstream deletions. Filled bars represent the ompF promoter region. The approximate location and direction of the ompF promoter (F& is indicated. The bla gene and the tet genes that are fused to the ompF gene are indicated by thin arrows. The EcoRI-Sal1 large fragment of pHF129 IS the same as that of pHF129M. required and this domain is assumed to be involved in the posit’ive regulation by the OmpR protein (Inokuchi et al., 1984). Therefore, it is reasonable to assume that the region is dispensable for a mutant promoter whose function is independent of the OmpR protein. As shown in Figure 3, pHF129M, one of the mutant plasmids, was linearized with EcoRI at t’he position 111 nucleotide upstream from the mRNA start sit’e; and digested with exonuclease Bal31 to produce the ompF promoter region deleted to varying extents. After t,reatment with the Klenow enzyme, the resultant blunt-ended fragments were ligated with EcoRI linker. The ligated products were digest,ed with EcoRI and SalI, and the DNA fragments caarrying t,he ompF promoter with different deletions were ligated with the EcoRI-Sal1 large fragments of pHF129. Plasmids t’hus constructed were transferred to KY2562 recA, and ampicillin-resistant clones were selected. The extent, of deletion was estimated by sizing the fragments generated by EcoRT-HgZTT digestion on 57; polyacrylamide gel electrophoresis. For some of them, the endpoint of the deletion was determined accurately by DNA sequencing. The results are shown in Figure 4. The plasmids carrying these deletions were named pTDA1 through pTDA10. The ten deletions thus selected were examined as
4
T. Dairi
et al.
-100
-80
pTDA4
pTDA6
-60
pTDA8
Figure 4. Nucleotide sequence of the promoter region and the region upstream from it of the deletions derived from the OmpR-independent ompF promoter of pHF129M. The nucleotide sequences of individual deletion mutants were determined by the method of Maxam & Gilbert (1980) .by using the E&III-HhaI fragments (see Fig. 1) with 32P-labeling at the Hoe111 end. The upstream ends of the individual deletions are indicated. The mRh’A start site of the ompF gene is numbered + 1. The - 35 region ( - 35) and Pribnow box (PB) are indicated. Sequence hyphens are omitted for clarity. to the ompF promoter function in terms of tetracycline resistance of the transformants in the ompRbackground (Fig. 5). First of all. the activity of the mutant ompF promoter in the ompR- background was as high as that of the wildtype promoter in the ompR’ background. Secondly, as expected, the OmpR-independent expression of the mut’ant promot,er did not require the domain
HFl29M
\1
I
pTDA’ I
I -100
pTDA2
c I -80
upstream from the -35 region. However. no promoter activity was observed when the deletion reached the -35 region wit,h the exception of pTDA7 that weakly showed t~etracycline resistance. In pTD,47, the first three ba,ses of the -35 region had been delet’ed (Fig. 4). However. as a result of the linkage with the EcoR81 linker a possible -35 region, T-C-C-C-(:-A. was newly formed in the
pT?A3
pT?A4
pTDA5\\/
$1 c 4 I %~Lh
-60 Base-
-40
-35
----L-J -20
PB
+’
pale
Figure 5. Effect of upstream deletions on the OmpR-independent ompF promoter activity of pHFl29M. The upstream ends of individual deletions are indicated by arrows, The mRNA start site is numbered + 1 and the positions of the - 35 region ( - 35) and Pribnow box (PB) are indicated. The ompF promoter activity of the deletions was assayed in the ompR- background (KY2562 recA) and expressed as the ratio of the number of tetracycline-resistant colonies to that of ampicillin-resistant colonies as described in Materials and Methods. Concentrations of tetracycline in the media were (a) 15 pg/ml, (0) 25 pg/ml and (A) 35 pg/ml. The activity of the wild-type ompF promoter in pHF129 was assayed in the ompR+ background (AB2847) in the presence of 15 pg tetracycline (A).
Positive Regulation in E. coli corresponding region. In contrast to this, deletion of the T at the firsi position of the -35 region resulted in the complete loss of the promoter activity (pTDA6). The importance of the T at the first position of the -35 region was proposed by Hawley & McClure (1983) and has been shown in this work. This is discussed further below. The mRNA st’art, site was determined for the OmpR-independent mutant promoter and some of the deletion promoters (pHF129M. pTDA1. pTDA2, pTDA3 and pTDA5) by S, nuclease mapping. In all cases. the start site was the same as cwnpF promoter, indicating that of the wild-type that t’he OmpR-independent expression of these promoters was not due to the formation of an artificial promoter (data not, shown). Th e promoter activity of t,he deletions was examined also in the ompR’ background using AB2847 as a host’ strain. The results were essent~ia,lly the same as those for the ompKbackground.
4. Discussion (a) Role of the OmpR protein in expression of the ompF gene The ompF gene coding for a major outer regulated membrane protein of E. coli is positively by OmpR: a product of the ompR gene. In a previous work (Inokuchi et al., 1984), the 107 basepair fragment was found to function as the ompF promoter. The fragment consists of 90 nucleotides upstream from the mRNA start site including the and the Pribnow box, and 17 - 35 region nucleotides downstream from the start site. Although the sequence at. the -35 region and that at the Pribnow box show low degrees of homology to the respective consensus sequences, the ompF gene is one of the most efficiently expressed genes in E. coli cells in the ompR’ background. It is assumed that the OmpR protein interacts with the domain upstream from the -35 region, which in turn facilitat,es the efficient binding of RNA polymerase to the ompF promoter region. To confirm this hypothesis, promoter mutants that show OmpR-independent expression of the wnpF gene were isolated with the aid of the ompFtut hybrid gene. In all four promoter mutants that were isolated independently, the first base of the Pribnow box was changed from A to T. No ot,her changes were observed. Upon the mutation, the became independent of both oNIp F expression OmpR, and the domain upstream from the -35 region but, on the other hand, the mRNA start site was not changed. These results strongly support the view that) the domain upstream from the -35 region is responsible for the positive regulation by the OmpR protein; most probably, the OmpR protein binds t,o this domain to induce the efficient transcription initiation. Recently, precise analysis of t.hr o,mpC’ gene, a gene coding for another major out.er membrane protein of E. coZi, was carried out in this laboratory (T. Mizuno & S. Mizushima,
5
unpublished results). The ompC gene is also positively regulated by the OmpR protein, and a domain upstream from the -35 region was found to be required for the gene expression. This domain shows a high degree of homology to the corresponding domain of the OmpF gene. (In) Positive regulation
and the Pribnow box
The conversion of the first base of the Pribnow box from A t.o T rendered the ompF gene activatorindependent. This suggests that. the T at the first. position of !,he Pribnow box may be important for the expression of a promoter that does not require an activator. From the results of statistical studies on about 106 promoters of E. co& genes, a so-called consensus sequence (T-A-T-A-A-T) was proposed for the Pribnow box (Hawley & McClure, 1983). Among the six bases, T. A and T at the first, second and sixt’h positions, respectively, are highly conserved, indicating the importance of these bases in recognition by RNA polymerase. Consistent with our observation in the present work, the first base was found to be almost exclusively T for promoters that do not require a positive regulator. An exception is the major lipoprotein gene for which A-AT-A-C-T was proposed with 18 base-spacing from the -35 region. However, this would preferably be T-A-A-T-A-C with 17 base-spacing. In contrast to the promoters that do not require a positive regulator, most of those requiring one do not possess T at this position. They are ompC (G-A-G-A-AT: Mizuno et al., 1984). maEPQ ((I-A-A-A-C-T), malK (C-A-T-A-G-T), malEFG (G-A-A-A-C-T: Raibaud et al., 1983; Bedouelle, 1983) and Pa, (A-A-G-T-A-T: Ho et al., 1983). The A and T at the second and sixth positions are conserved in these promoters. Furbhermore, these genes possess an additional domain upstream from the -35 region that is a possible interaction site for et al., 1983; an activator protein (Raibaud Bedouelle. 1983; Inokuchi ct nl.. 1984: Ho et al., 1983). Based on these facts, the following hypothesis can be proposed for positive regulation. (1) The T of the Pribnow box is at the first position indispensable for the functioning of a promoter that does not, require an activator protein; replacement of t.he T by other bases resuhs in the disappearance of the promoter function. (2) However. the defect can be compensated for by t’he presence of both an activator protein and a proper nucleotide domain upstream from the -35 region. (3) The activator protein possibly interacts with the upstream domain to facilitate the binding of RNA polymerase to t’he promoter region. Consistent with this hypothesis, transcription initiation of the malEFG operon becomes independent of the activator protein (MalT) upon conversion by mutation of the G at, the first position of the Pribnow box to T (Bedouelle, 1983). It should be mentioned, however, that the T of the Pribnow box is not t,he only base that
T. Dairi
6
determines positive regulation. In a previous work, we constructed a hybrid promoter in which the - 35 region (T-A-G-C-G-A) and the region upstream from it are from the ompF gene and Pribnow box (T-T-T-A-A-T) from the tet genes (Inokuchi et al., 1984). Functioning of the hybrid promoter was OmpR dependent. When the -35 region was further replaced by that’ from tet genes leaving the upstream region from the ompF gene as it was, the promoter function became OmpR independent (Y. Ozawa et aZ., unpublished results). These results indicate that the -35 region may also have something to do with the positive regulation. In many positively regulated genes. the -35 region is rather atypical, showing a low homology to the consensus sequence (Rosenberg & Court, 1979: Bedouelle et al., 1982: DBbarbouilE & Raibaud, 1983; Ho et al., 1983; Inokuchi et al., 1984). In addition to the regulation by the ompR gene, expression of the ompF gene is affected by the envZ gene (Hall & Silhavy, 1981a,h) and medium osmolarity (van Alphen & Lugtenberg, 1977; Kawaji et al., 1979). The functions of these effecters in the OmpR independent mutant were, however, difficult to study with the ompF-tet system we used in this work. We thank S. Teranishi for excellent secretarial assistance. This work was supported by grants from the Ministry of Education, Science and Culture of Japan and the Special Co-ordination Fund for Science Technolog? Agency of Japan. References Bedouelle, H. (1983). J. Mol. Biol. 170, 861-882. Bedouelle, H., Schmeissner, U.. Hofnung, M. & Rosenberg, M. (1982). J. Mol. Biol. 161. 519-531. Birnboim, H. C. & Doly. J. (1979). Nucl. Acids Res. 7, 1513-1523. Brosius, J., Cate, R. L. & Perlmutter, A. P. (1982). J. Biol. Chem. 257, 9205-9210. Dagert, M. & Ehrlich, S. D. (1979). Gene, 6, 23-28.
et al. D6barbouillB. M. & Raibaud, 0. (1983). .J. Bacterial. 153. 1221-1227. DBbarbouillB, M. & Schwartz, M. (1979). ,J. No/. Riol. 132. 521-534. Echols. H. & Green, I,. (197 1). Proc. ,Vut. Acud. Sri., U.S.A. 68, 2190-2194. Hall. M. X. & Silhavy. T. J. (1979). J. BactevioZ. 140. 843h 847. Hall, M. N. & Silhavy. T. tJ. (19810). .I. Mol. Biol. 146. 23-43. Hall, M. N. & Silhavy, T. .J. (1981b). .J. Mol. Kiol. 151. l-15. Hawley. D. K. & McClure. W. R. (1983). XUCI. &ids Res. 11, 2239-2255. Ho. Y.-S., Wulff, D. I,. & Rosenberg. M. (1983). .Yrtl~r(, (London), 304. 703 708. Hofnung. M.. Schwartz, M. 8: Hatfield. I). (1971). J. No/. Riol. 61. 681-694. Inokuchi. K.. Wutoh. X., ,\llatsuyamn, S. & Ylizushima. S. (1982). ic’ucl. Acids RPS. 10, 6957--6968. Tnokuchi. K.. Furukawa. H.. h’akamura. K. B Mizushima. S. (1984). ,I. Mol. Viol. 178. 653-668. Kawaji. H.. Vizuno, T. & Mizushima. S. (19%). .I. Bnrtrriol. 140. 84S847. Maniatis. ‘I’.. Frit,scah. 15. F. & Sambrook. ,I. (1982). 111 Molec~ular Cloning. pp. 150-I 72. Cold Spring Harbor Laboratory Press, Cold Spring Harbor. Maxam, A. M. 8: (Gilbert. W. (1980). In Jlrfhods i)t EnzymoZogy ((:rossman. I,. & Moldavr. E;.. rds), vol. 65. pp. 499-560. Academir Press. Kew York. Mizuno. T.. Chou, Yl .-Y. 8r Inouyr, 11. (1984). t’roc.. .z’n/. Acad. Sri.. [-.A’.,~. 81. 1966Gl970. Mutoh. N., n’agasawa, T. 6 Mizushima. S. (l!tHI). ,J. Hactrriol. 145. 10X5-1090. Raibaud, O., Ddbarbouill@. M. & Schwartz. Al. (I9X:l). J. Mol. Biol. 163, 395-~408. Rricahardt,. L. & Kaiser. A. I). (1971 ). /‘rot. .Vnf. .,IuIc/. Sci., I’.R.d.
68. %I85 Zl89.
Rosenberg. iV. & (‘out?. I). (1979). ;Z,t~r(. fZv11.(:rnet. 13. 319-353. Sanger. F.. Nicklen. S. & (‘oulxon . A. R. (1!177). /‘/YN~..Z;rf. A cud. Sci., U.S.d
74. 5463-5467.
Sato. T. & Yura. T. (1981). ,J. Kacfrriol. 145. kW96. van Alphen. W. & Lugtenberg. K. (1977). .J. b’crctrriol. 131, 62~3.-630.