Rhodobacter sphaeroides - American Society for Microbiology

JOURNAL OF BACrERIOLOGY, Dec. 1993, p. 7629-7638

Vol. 175, No. 23

0021-9193/93/237629-10$02.00/0 Copyright C 1993, American Society for Microbiology

Transcription Properties of RNA Polymerase Holoenzymes Isolated from the Purple Nonsulfur Bacterium Rhodobacter sphaeroides RUSSELL K. KARLS,1 DING JUN JIN,2 AND TIMOTHY J. DONOHUEI*

Department of Bacteriology, University of Wisconsin-Madison, Madison, Wisconsin 53706,1 and National Cancer Institute, Bethesda, Maryland 208922 Received 20 August 1993/Accepted 5 October 1993

We have been characterizing RNA polymerase holoenzymes from Rhodobacter sphaeroides. RNA polymerase purified from R. sphaeroides transcribed from promoters recognized by Escherichia coli Eif32 or Eif70 holoenzyme. Antisera to E. coli if32 or f70 indicated that related polypeptides of -37 kDa (if37) and 93 kDa (if93), respectively, are present in this preparation. Transcription of Cf32-dependent promoters was observed in a further fractionated R. sphaeroides holoenzyme containing the r37 polypeptide, while a preparation enriched in o93 transcribed Or70-dependent promoters. To demonstrate further that the ir93 polypeptide functions like E. coli C70, we obtained an R. sphaeroides Eir93 holoenzyme capable of transcription from Cf70-dependent promoters by combining ir93 with (i) an Eif37 fraction with diminished ir93 polypeptide content or (ii) E. coli core RNA polymerase. The generation of analogous DNase I footprints on the lacUVS promoter by R. sphaeroides Eor93 and by E. coli Eir70 suggests that the overall structures of these two holoenzymes are similar. However, some differences in promoter specificity between R. sphaeroides Ecf93 and E. coli E{J70 exist because transcription of an R. sphaeroides rRNA promoter was detected only with Ecr93.

The purple nonsulfur photosynthetic bacterium Rhodobacter sphaeroides has the ability to regulate transcription of a large number of genes in response to light, oxygen, and other environmental signals (25-27, 29). Although many genes that are transcriptionally controlled by specific stimuli have been identified, little is known about the cis- and trans-acting elements required for this regulation. In addition, promoter structure in this bacterium is not well understood and little information is available on what forms of RNA polymerase (RNAP) holoenzyme recognize individual promoters. This information is necessary to determine how this facultative organism controls expression of genes responsible for growth under different physiological conditions. Previous analysis of RNAP from the photosynthetic bacteria Rhodospirillum rubrum (43), Rhodobacter capsulatus (17), and R. sphaeroides (24) identified polypeptides comparable in molecular weight to the Escherichia coli

P',

I,

ot, and

u

since the pufQ and pufB control regions do not bear a strong resemblance to any known prokaryotic promoters. To advance the study of transcription in R. sphaeroides further, we analyzed the promoter recognition properties of RNAP holoenzymes. In addition to testing of RNAP activity on an R. sphaeroides template, transcription was assayed with a series of well-characterized E. coli promoters which do not require trans-acting factors in vitro. Our results indicate that R. sphaeroides has separate RNAP holoenzymes (Ea93 and Eu37) that recognize E. coli Ea70- and Eu32-dependent promoters, respectively. In addition, the major sigma and core subunits of the E. coli and R. sphaeroides enzymes are sufficiently compatible to allow promoter recognition by heterologous enzymes containing core from one bacterium and sigma from the second. Despite this overall similarity, there appear to be significant differences in promoter specificity, since a native rRNA operon promoter (rrnB) recognized by R. sphaeroides Ea93 is not transcribed by E. coli Eu70.

70

subunits. In addition, Western blot (immunoblot) analysis of the R. sphaeroides enzyme with E. coli E(J70 antiserum confirmed that these polypeptides are related to E. coli RNAP (24). Unfortunately, only limited information is available on the promoter sequences recognized by these purified enzymes. For example, R. capsulatus RNAP could generate transcripts from pUC19 plasmid DNA, but the specific promoter(s) used by this enzyme has not been reported (17). Specific promoter recognition by R rubrum RNAP was demonstrated by generation of transcripts from intact phage T7 DNA similar to those synthesized by the E. coli enzyme (43). Kansy and Kaplan provided the first in vitro analysis of a host promoter when they showed that R. sphaeroides RNAP was able to generate transcripts from the pufQ and pufB promoter regions (24). However, it is not clear what form of R. sphaeroides RNAP holoenzyme was responsible for synthesizing these products

MATERUILS AND METHODS Enzymes and materials. Restriction enzymes and DNAmodifying enzymes were purchased from Promega Corp., Madison, Wis.; Bethesda Research Laboratories, Gaithersburg, Md.; or New England BioLabs, Beverly, Mass. DNase I was purchased from Worthington Biochemical Corp., Freehold, N.J. E. coli Ea32, Eur70, core RNAP, and the u70 subunit were purified as previously described (16). Radiolabeled nucleotides were obtained from either Dupont NEN, Boston, Mass., or Amersham Life Sciences, Arlington Heights, Ill. Nonradioactive nucleotides, chromatography supplies, and resins were obtained from Pharmacia LKB Biotechnology, Milwaukee, Wis.

Enzyme purification. Wild-type R. sphaeroides 2.4.1 was aerobically at 32°C in 5-liter Pivitsky bottles containing 3.5 liters of Sistrom's minimal medium A (38). Aeration was provided by sparging with 30% oxygen, 69% nitrogen, and 1%

grown * Corresponding author. Electronic mail address: Donohue@ Macc.Wisc.Edu.

7629

7630

KARLS ET AL.

carbon dioxide as previously reported (3). Mid- to late-exponential-phase cells (density, 8 x 108 to 1.4 x 109) were harvested by centrifugation (8,000 x g, 10 min) and washed once with Sistrom's minimal medium, and cell pellets were stored at - 80°C. RNAP from R. sphaeroides was purified by the method introduced by Burgess and Jendrisak (4) with the following modifications. Cells (-300 g [wet weight]) were incubated for 1 h at 10°C in 575 ml of grinding buffer supplemented with 1.5 mg of lysozyme per ml to facilitate cell lysis. Following centrifugation (10,000 x g, 45 min) to remove cell debris, polyethyleneimine P was added to 0.2% to precipitate nucleic acids. Proteins nonspecifically bound to nucleic acid were eluted with TGED (10 mM Tris-Cl [pH 7.9], 10% glycerol, 0.1 mM EDTA, 0.1 mM dithiothreitol) containing 0.5 M NaCl. RNAP was eluted from the nucleic acid with TGED supplemented with 1 M NaCl. After ammonium sulfate precipitation (35%), the protein was resuspended in a volume of TGED necessary to make the conductivity equal to that of TGED containing 0.25 M NaCl. The solution was then loaded onto a single-stranded DNA agarose column (28). After being washed with TGED-0.25 M NaCl, RNAP was eluted with steps of TGED-0.4 M NaCl and TGED-1 M NaCl. The column fractions containing RNAP subunits were precipitated with ammonium sulfate (35%), resuspended in 0.5 ml of TGED-0.5 M NaCl, and passed through a 100-ml Sephacryl-300 (S-300) column (1.6 by 100 cm). For some experiments, after S-300 chromatography RNAP was fractionated on Q-Sepharose Fast Flow anion-exchange resin, a resin analogous to that used to separate E. coli Eu70 from core RNAP (21). The enzyme was loaded at 5 mb/h with a peristaltic pump onto a 10-ml column (1 by 20 cm) at a conductivity equal to that of TGED-0.1 M NaCl, washed with 3 column volumes of TGED-0.1 M NaCl, and eluted with a linear gradient of TGED-0.35 to 0.50 M NaCl. Column fractions were pooled on the basis of (i) the protein profile determined by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and (ii) the results of specific transcription assays described in Results. Protein levels were determined with Bio-Rad Assay Solution (Bio-Rad Laboratories, Richmond, Calif.) with bovine serum albumin as the standard. Apparent molecular weights of polypeptides were estimated by comparison to Bethesda Research Laboratories high-range protein molecular weight standards. General transcription assay. The ability of RNAP to incorporate radiolabeled UTP into products was monitored with a modified form of a general transcription assay previously reported (6). The optimized assay for the R. sphaeroides enzyme through S-300 chromatography was as follows. A 0.5-pmol sample of supercoiled template DNA (plasmid pRSF1 [see Table 1; 24]) and 2 [LI of RNAP extract (-2 pmol of protein) were preincubated in 20 RI of 40 mM Tris-Cl (pH 7.9)-50 mM KCl-5 mM MgCl2-0.1 mM dithiothreitol-0.1 mg of bovine serum albumin per ml for 20 min at 30°C. Reactions were started by addition of the four ribonucleotides (each to 0.5 mM) plus 1 ,lI of [3H]UTP (3,000 Ci/mmol). After 3 min, assays were terminated by spotting on DEAE filters (W-DE81; McLeester Research Equipment, Madison, Wis.). After drying, filters were washed to remove unincorporated nucleotides and redried, and the absorbed radioactivity was determined (24). One unit of enzyme activity corresponds to the incorporation of 1 nmol of UTP in 1 min at 30°C. Specific transcription assay. Assays to monitor specific transcription initiation contained plasmid template DNA at 20 nM and 10 to 50 nM RNAP in a 20-pu total volume of

J. BACTERIOL.

transcription buffer (40 mM Tris-Cl [pH 7.9], 35 mM NaCl, 5 mM MgCl2, 0.1 mM dithiothreitol, 0.1 mg of bovine serum albumin per ml). The same preincubation step as in the general transcription assay was used to allow RNAP binding to template DNA. Reactions proceeded for 20 min at 30°C after addition of nucleoside triphosphates (0.5 mM ATP, 0.5 mM CTP, 0.5 mM GTP, and 0.05 mM UTP) and 10 ,uCi of [a-32P]UTP (3,000 Ci/mmol). The assays were terminated by addition of Taqtrack Stop Solution (10 mM NaOH, 95% formamide, 0.05% bromphenol blue, 0.05% xylene cyanole; Promega Corp.), heated to 90°C for 3 min, and analyzed on 6% polyacrylamide-7 M urea-sodium acetate gradient gels (37). Gels were dried prior to autoradiography with Kodak XAR-5 film. When indicated, the levels of specific transcripts were estimated via radioanalytic imaging on an AMBIS system (Ambis, Inc., San Diego, Calif.). Templates used to monitor specific transcription initiation had promoter-containing DNA fragments cloned immediately upstream of the spf or the rrnB T1 and T2 transcription terminators in pUC19spf' (15) or pRLG770 (35), respectively (see Table 1). The transcripts produced (-100 to 300 nucleotides long, depending on the test promoter) were readily detected on denaturing polyacrylamide gels. In most cases, the templates were supercoiled DNA purified by two successive CsCl gradients (31). When linear templates were used, the plasmid was digested at a unique EcoRI site outside the cloned promoter and transcription terminator. Western blot analysis. Protein samples solubilized in the presence of 2-mercaptoethanol for 15 min at 85°C were separated on SDS-12% polyacrylamide gels and transferred to 0.2-,um-thick nitrocellulose for 3.25 h at 0.5 A with a Transphor electrophoresis unit (Hoefer Scientific Instruments, San Francisco, Calif.). To remove nonspecific antibodies, rabbit polyclonal serum against E. coli or (provided by the C. A. Gross laboratory) was preincubated with a cell extract from E. coli CAG 9333 (ArpoH). Mouse monoclonal serum 2G10 against E. coli cu70 (provided by N. Thompson of the R. R. Burgess laboratory) recognizes a domain in conserved region 3 (39). The blots were developed with secondary antibody-alkaline phosphatase conjugates to either anti-rabbit (Boehringer Mannheim, Indianapolis, Ind.) or anti-mouse (Promega Corp.) immunoglobulin G. DNase I protection experiments. EcoRI-BamHI DNA fragments containing the lacUV5 promoter (- 140 to +63 relative to the transcript start) were 3' end labeled on either the top or bottom strand as follows. Plasmid pKO41acUV5 (Table 1) was linearized with EcoRI or BamHI and labeled with T7 DNA polymerase and [a-32P]dATP and ddTlTP for the EcoRI site or with [a-32P]dGTP and ddATP for the BamHI site. The DNA was precipitated, washed with 0.5 M ammonium acetate in 75% ethanol, and dried. The labeled DNA was suspended and digested with BamHI or EcoRI, and the EcoRI-BamHI lacUV5 promoter fragments were purified from 5% nondenaturing polyacrylamide gels as previously described (33). DNase I treatments were modified slightly from those previously reported (33). Approximately 10,000 cpm of endlabeled DNA was incubated with 150 fmol of RNAP (either E. coli Eor70 or R. sphaeroides Eu93; see Results)-0.1 mg of bovine serum albumin per ml-40 mM Tris-acetate (pH 7.9)-30 mM KCl-10 mM MgCl2-1 mM dithiothreitol in a final volume of 25 pA for 15 min at 30°C. DNase I was added to 0.2 pug/ml for 30 s. The reactions were stopped by adding EDTA and sodium acetate to 10 and 200 mM, respectively, along with an equal volume of unbuffered phenol. To each assay, 5 ,ug of calf thymus DNA and 2 volumes of ethanol were added to precipitate DNA. The precipitated DNA was washed with 90%

RHODOBACTER SPHAEROIDES IN VITRO TRANSCRIPTION

VOL. 175, 1993

7631

TABLE 1. Plasmids used in this study Plasmid

pRSF1

pBR328 pUC19spf' pRLG770 pJET40 pDC401 pRKK60 pDC421 pKO500 pRLG593 pKO41acUV5 pKO4 pRKK61 pKWT12 pRLG597 pRKK4 pRME19 pRME1

pUC19 pRKK5 a

Promotera

Reference and/or plasmid construction

24; EcoRI fragment F containing at least one promoter recognized by R. sphaeroides RNAP, cloned into pBR328 7 15 35 E. coli dnaK P1 315-bp HindIII-EcoRV fragment of pDC401 cloned into HindIII-HincII site of pUC19spf'; plasmid constructed by D. Cowing 9 E. coli htpG EcoRV-SmaI htpG fragment from pDC421 cloned into HinclI site of pUC19spf'; plasmid constructed by A. Mitin 150-bp SaullIA htpG fragment cloned into BamHI site of pKO500; plasmid constructed by D. Cowing 10 E. coli lacUV5 35 203-bp HaeIII lacUV5 promoter fragment cloned into filled EcoRI- and BamHI-digested pKO4; plasmid constructed by G. Bellomy 34 E. coli rpoD PHS HindIII-EcoRV rpoD PHS fragment of pKWT12 cloned into pUC19spf'; plasmid constructed by A. Mitin 41 E. coli rrnB P1 35 Tn9O3kan EcoRI-XhoI fragment from pRME19 cloned into EcoRI-XhoI-digested pRLG770 SphI fragment containing Tn9O3kan gene from pRME1 cloned into SphI site of pUC19 12 44 R. sphaeroides rrnB rrnB Pstl-Alul promoter fragment cloned into PstI-HincIl-digested pUC19spf'

Refers to the specified promoter fragment cloned upstream of a transcription

ethanol, dried, resuspended in 5 pd of Taqtrack Stop Solution, heated for 3 min at 90°C, and separated on a 6% polyacrylamide-7 M urea gel alongside an A+G DNA sequencing ladder of the same end-labeled template (32). Gels were dried prior to autoradiography with Kodak XAR-5 film. RNAP reconstitution experiments. To test for stimulation of transcription, -0.6 pmol of core RNAP was mixed with a oneor fivefold molar excess of a subunit and preincubated for 20 min at 30°C with 20 nM plasmid DNA template in 1 x transcription buffer (20-iLl volume). Reactions were completed as described for specific transcription assays. To concentrate the R sphaeroides U93 sample (Q-Sepharose fraction C; see Results), a Centricon 30 (Amicon, Inc., Beverly, Mass.) was used.

terminator.

R. sphaeroides RNAP recognizes both E. coli Er70- and Eo32-dependent promoters. To test for specific transcription initiation, the S-300 RNAP preparation was assayed with promoters known to be dependent on either the Eu70 or Eu32 form of E. coli RNAP with both linear and supercoiled

U,~~~~~~~~~~4 e:

Y

u

Z

z

tu ~~~~~~~~4l

RESULTS Purification of R. sphaeroides RNAP. We purified RNAP by a modified version of the Polymin P method of Burgess and Jendrisak (4) because it yields active RNAP from a variety of bacteria (43). SDS-PAGE analysis of the protein content following each purification step (Fig. 1) showed subunits comparable in size to that of E. coli Eu70. By using this protocol, we obtained -25 mg of protein after S-300 chromatography from -300 g (wet weight) of aerobically grown cells. The S-300 fraction was used to optimize the general transcription assay conditions (described in Materials and Methods) for temperature and salt (MgCl2 and NaCl) and supercoiled plasmid template concentrations (data not shown). Under optimal conditions, the specific activity of the S-300 material with this general transcription assay was -55 U/mg. While this value is --15-fold lower than that of E. coli RNAP at an equivalent point during purification (-800 U/mg; 4), it could reflect differences in the number and strength of promoters on the DNA templates used to measure enzyme specific activity. Indeed, in the specific transcription assays presented below, the R. sphaeroides enzyme showed activity comparable to that of E. coli RNAP on several test promoters.

tu

ffi

.

200 97 68

VI

0

4

a70

43 29 -*

FIG. 1. SDS-PAGE analysis of RNAP purification. Proteins were separated on a 10% polyacrylamide gel and visualized by Coomassie blue staining. The numbers at the left are molecular masses (in kilodaltons) of Bethesda Research Laboratories prestained high-range protein molecular weight (Mol. wt.) standards. The rightmost lane contained 2 jxg of the purified E. coli Eu70 holoenzyme. The remaining lanes contained 25 ,g of protein from the indicated steps of R sphaeroides RNAP purification. Individual protein subunits in the S-300 peak that cross-react with antiserum to E. coli Eu70 (data not shown) are indicated at the right.

7632

KARLS ET AL. TIOM3

E. coll cuvS E. cel RNAP R. phhrdes RNAP

Tempbte (onm

J. BACTIERIOL. kam

Ee Ee

En En + S

L

L S

S

L

S

E. ceol

E. ce

nuB P,

duK P,

Ee Ee EE

L S

L

S

L

S

E. coli Promoter

LSj

L

S

-35

-10

A

Ee Ee

L

lacUV5

CCCAGGCTTTAC& CTTTATGCTTCCGGCTCG TATAATGTGTGGA

Tn9O3kan

AGCCACGTTGTGT CTCAAAATCTCTGATGT

orliv

AGAGTTCTTG&AG TGGTGGCCTAACTACGGC TACaCTAGAAGGA

ttnB P,

TTTCCTCTTGTCA GGCCGGAATAACTCCC

E. coli e' consensus

E. coli Promoter

TTGACA

TACATTGCACAAGA

TATAATGCGCCACCA TATAAT

17 bp

-10

-35

B rtnB P, dnaK

P,

ATTTCCTCTTGTC

AGGCCGGAATAAC

TCTCCCCCTTGAT GACGTGGTTTACGA

TCTATAATGCGCCACCA CCCCATTTAGTAGTCAA

htpG

GCT_GCTTGAA

rpoD Ps

TGCCACCCTTGAA AAACTGTCGATGTGG GACGATATAGCAGATAA

E. coli o'3 consensus

TCTC-CCCTTGAA

ATTATTCTCCCTTGT CCCCATCTCTCCCACATC

13-15 bp

CCCCAT-TA

-35

-10

C rrnB

AATCCGCTTGCGC CCGGGGCCGTCTGCTCC

TAGAAACCGCTTC

rrnA

TTTCCTCTTGCGG GTTTTTTTGCGGTTCCC

TAGATAGCGCCTC

TTTCCTCTTGCGG GTTTTTTTGCGGTTCCC

TAGATAGCGCCTC

E. coli a" consensus

1

2

3

4

S

6

7

S

9

10

11

12

13

14

15

16

17

15

FIG. 2. In vitro transcripts generated from E. coli promoters. R. sphaeroides RNAP from an S-300 column (Fig. 1) or purified E. coli Ea70 and E&12 RNAP holoenzymes were tested for transcription with E. coli promoters cloned on plasmids upstream of transcription terminators. The promoters tested are indicated at the top of grouped lanes. Both supercoiled (S) and EcoRI-linearized (L) versions of these templates were analyzed. Specific transcripts are indicated by arrowheads. The asterisk refers to a transcript made from the oriV promoter present on all plasmids. The enzyme levels used were -1, 0.4, and 0.2 pmol of R. sphaeroides RNAP, E. coli Eu70, and E. coli Eu32, respectively. The conditions for specific transcription assays are described in Materials and Methods.

templates. The data in Fig. 2 demonstrate that R sphaeroides RNAP recognized several cloned o70-dependent promoters (lacUVS, Tn9O3kan, and rrnB P1; lanes 1 to 12) and the plasmid origin of replication inhibitor promoter P4 (oriV, 2) present on all templates (lanes 1 to 12). The synthesis of transcripts similar in size to those made by E. coli E"70 from each of these promoters suggests that transcription termination signals present on these templates are recognized by R sphaeroides

RNAP. DNA sequences for the O °-dependent promoters transcribed by this preparation of R. sphaeroides RNAP are shown in Fig. 3A. All of these promoters have -10 and - 35 hexamers with considerable homology to the a70 consensus and differ from the optimal 17-bp spacing between these elements by at most 1 bp. To determine whether R. sphaeroides RNAP could recognize a o32-dependent promoter, the S-300 material was assayed with the E. coli dnaK P1 promoter, which is known to require Ea32 for transcription (9). R sphaeroides RNAP yielded a dnaK P1 transcript identical to that made by E. coli Eu32 (Fig. 2, lanes 15 to 18). These experiments indicate that R sphaeroides RNAP recognizes both C70- and u32-dependent promoters. In most cases, the yield of transcripts in Fig. 2 showed little dependence on the superhelical nature of the template. The specific activity of this enzyme preparation appears to be comparable to that of pure Eu70 or E.32 when one compares the amount of product generated in Fig. 2 with the molar amounts of enzyme added to these assays.

TTGACA

17 bp

TATAAT

FIG. 3. Compilation of E. coli E&0- and E&2-dependent promotsphaeroides rRNA operon promoter sequences. (A) Eu70-dependent promoters which function as templates for transcription by R. sphaeroides RNAP in vitro. The E. coli or' promoter consensus (22, 23) is shown. Individual bases homologous to the promoter consensus are in boldface. *, promoter sequence inferred from proximity to the deduced 5' ends of in vitro transcripts and in vivo mRNA 5'-end analysis in R. sphaeroides (36). (B) E. coli E(J32_ dependent promoters tested for transcription by R sphaeroides RNAP. The E. coli 32 consensus (8, 9) is also shown. Bases homologous to the E. coli a32 consensus are underlined. (C) Putative R sphaeroides rRNA operon promoter elements with bases identical to the E. coli c70 consensus sequence are in boldface. ers and putative R

R. sphaeroides polypeptides of 93 and 37 kDa are structurally and functionally related to cr70 and r32, respectively. We next sought to separate the Eu70 and Eu32 transcription activities present in the R sphaeroides RNAP preparation. The S-300 material was fractionated on a strong anion-exchange resin (Q-Sepharose Fast Flow). This type of resin was chosen because it had previously been used to separate E. coli Eo70 from a fraction containing both core RNAP and Ec&2 (21).

The SDS-PAGE profile shown in Fig. 4 indicates that fraction C contains the 93-kDa polypeptide without much of core, fraction F contains the 93-kDa polypeptide plus core, and fraction H contains the 37-kDa polypeptide plus core but has a diminished amount of the 93-kDa polypeptide. Figure 5A shows that a monoclonal antibody (2G10) which is specific for a 37-amino-acid epitope in region 3 of E. coli c70 (39) reacts with the 93-kDa polypeptide in fractions C, F, and H. Therefore, in descriptions of all subsequent experiments we will refer to this 93-kDa protein as o93. When a polyclonal antibody specific to E. coli cr32 was used in a similar experiment, it reacted with the 37-kDa polypeptide in fraction H (Fig. SB). In contrast, no related proteins were found in fraction F (Fig. 5B), which lacked detectable levels of the 37-kDa polypeptide (Fig. 4). Therefore, in descriptions of all subsequent experiments, we will refer to this 37-kDa protein as c37. When these fractions were assayed for specific transcription activity, Euc (fraction F) generated the same products as E. coli Eu70 from the lacUVS and oriV promoters (Fig. 6A, lanes 1 and 5) but did not transcribe from the ur32-dependent dnaK

VOL. 175, 1993


A.

Pooled column fractions ,

o o- A B C D E F G H I 4

200

4

97

4

68

4

43

4

29

4

18

a70 -

av

7633

B. Ea70 Eg70 H

F

C

0

F Eg32 H

--

a32 -p

4- 14

FIG. 4. R. sphaeroides RNAP fractionation by Q-Sepharose chromatography. An SDS-12% polyacrylamide gel was used to analyze pooled Q-Sepharose column fractions by silver staining. The first two lanes contained 2.80 and 0.56 jig of protein from the column onput (S-300 RNAP in Fig. 1). Approximately 15 jil of each pooled fraction was loaded onto the gel in the order in which it eluted from the column (A through I). RNAP subunits present in the column onput that are related to E. coli EU70 or u32 are indicated at the left (Western blot analysis data not shown). Migration of Bethesda Research Laboratories prestained high-range protein molecular mass standards (in kilodaltons) is indicated at the right.

P1 promoter (Fig. 6A, lanes 2 and 4). The control lane (Fig. 6A, lane 3) confirmed that the dnaK P1 promoter cannot be transcribed by E. coli Eu70. These results support the notion that core RNAP and u93 present in fraction F transcribe a70-dependent promoters. Further support for this idea comes from transcription from u70-dependent promoters with a reconstituted RNAP holoenzyme. In this experiment, addition of a 1:1 or 5:1 molar ratio of (93 (fraction C) to fraction H (Eu37 with a small amount of u93) increased transcription from the r70-dependent lacUVS and oriV promoters to the levels observed with Eu93 (fraction F; Fig. 7A, lanes 1 to 5; radioanalytic quantitation not shown). These results indicate that R sphaeroides u93 can be added to core RNAP in vitro to yield a functional enzyme that is capable of recognizing C70-dependent promoters. The data in Fig. 6 also show that ECu37 (fraction H) transcribes a32-dependent promoters. While fraction H had a relatively low level of activity with the lacUVS or oriVpromoter (Fig. 6B, lane 1), it had transcription activity comparable to that of E. coli Er32 on three cu32-dependent promoters (dnaK P1, htpG, and rpoD PHS; Fig. 6B, lanes 4 to 9; promoter sequences are shown in Fig. 3B). The control lane shows that the lacUV5 and oriV promoters are not transcribed by E. coli Eu32 (Fig. 6B, lane 2). In summary, these results indicate that distinct forms of R

1 2 3 4 5 6 1 2 3 FIG. 5. Western blot analysis of R sphaeroides RNAP Q-Sepharose column fractions. Panel A was probed with monoclonal antibody 2G10, which is specific for E. coli u70. Samples are E. coli E(J70 (lane 1, 1.92 jig; lane 2, 0.64 jig); Q-Sepharose fractions H (lane 3, 3.80 jig), F (lane 4, 1.92 j,g), and C (lane 5, 0.38 ,ug); and the column onput (lane 6, 1.92 jig). Panel B was probed with polyclonal antiserum to E. coli U32. Samples are Q-Sepharose fractions F (lane 1, 2.31 jig) and H (lane 3, 4.20 jig) and E. ccli E32 (lane 2, 1.65 jig). The presence of lower-molecular-weight products in lane 2 probably reflects minor degradation products present in the E. coli Eu32 sample.

sphaeroides RNAP holoenzyme which specifically recognize a70- and u32-dependent promoters can be partially separated by Q-Sepharose chromatography. The data suggest that Eu93 recognizes ur70-dependent promoters, while Eu37 transcribes u32-dependent promoters. Although this resin appears to separate R. sphaeroides Eu93 and Ecr37 effectively, the relative elution of these two holoenzymes appears to be the opposite of that observed with E. coli RNAP (21). Reconstitution of functional RNAP holoenzyme with the core and sigma subunits from R. sphaeroides and E. coli. Potential structural and functional similarities between the R. sphaeroides and E. coli RNAP holoenzymes were examined further by determining whether active enzymes could be obtained by mixing the core and sigma subunits of these two bacteria. In the first experiment, R. sphaeroides u93 (fraction C) was tested for the ability to stimulate transcription from a70-dependent promoters (lacUW5 and oriV) when added to E. coli core RNAP. Neither R. sphaeroides cr3 nor E. coli core RNAP had detectable transcription activity from either promoter (Fig. 7, lanes 3 and 6). However, when the R. sphaeroides u 93 fraction was added at a 1:1 or 5:1 molar ratio, transcription from both promoters was detected (,Fig. 7A, lanes 7 and 8). In a similar experiment, E. coli U7 was tested for the ability to increase transcription of the same promoters when added to R sphaeroides core RNAP. The source of the R. sphaeroides core was Q-Sepharose fraction H, which has diminished levels of u93. Purified E. coli C70 did not produce detectable transcripts (Fig. 7B, lane 3), and R. sphaeroides fraction H had minimal transcription from these u70-dependent promoters compared with Eau93 (fraction F; Fig. 7B, lanes 6 and 9). However, when a 1:1 or 5:1 molar ratio of E. coli u70 was added to R. sphaeroides fraction H, radioanalytic imaging indicated that transcription from the lacUV5 promoter was stimulated about sixfold (Fig. 7B, lanes 7 and 8; quantitation not shown). Under these conditions, no detectable increase from the oriV promoter was observed when E. coli C70 was mixed with R. sphaeroides fraction H. A possible explanation for this result is that the low level of Ea93 in fraction H is sufficient to saturate the oriV promoter.

7634

KARLS ET AL.

J. BACTERIOL.

A lacuvS dnaK P, E. coli Ea"7 E. coli Eu'3 R. sphaeroides F

+ +

+

+ +

B

+

+

+

lacuv5

E. coli EO*" E. coli R. sphaeroides H

dnaK P,

tpG rpoD P,n

+

EO3"

+

+

+

+

+

+

+

+

+ 4- rpoD P.,

dnaK P, -+

dnaK P,

4-

h:pG

lac uvS

lacuvS -_

oriV * oriv *

1

2

3

4

S

1

2

3

4

S

6

7

8

9

FIG. 6. In vitro transcription of c2-dependent promoters by R sphaeroides Eo3. Supercoiled plasmid templates containing the indicated promoter cloned upstream of transcription termination signals are listed at the top. Control RNAP used in these assays was either purified E. coli Ea7O or Eu32 holoenzyme (0.6 pmol of each). The R sphaeroides RNAP used (-0.6 pmol), was either Q-Sepharose fraction F (A) or H (B). In

vitro-generated transcripts

are

identified alongside each panel.

R. sphaeroides Eo93 and E. col Eo70 produce similar DNase I footprints on the lacUV5 promoter. We further characterized interactions of Eu93 with promoters by analyzing its DNase I footprint on the lacUV5 promoter (Fig. 8). In general, the extent of the protected region on either strand was similar with R sphaeroides Eu93 and E. coli Ec70. For example, the region from -47 to + 17 showed virtually identical protection and enhancement patterns with both holoenzymes on both strands. The only detectable differences appear to be in enhancements at the ends of the protected region. For example, R sphaeroides Ea93 lacks enhancements at + 18 (Fig. 8, lane 7) and +21 to +23 (Fig. 8, lane 3) that are observed with E. coli Ec70. However, the R sphaeroides enzyme shows a pronounced enhancement at -49, which is only weakly visible with E. coli Ea70 (Fig. 8, lane 4 versus lane 3). Therefore, with the resolution afforded by this type of analysis, these data support the notion that R sphaeroides Eu93 recognizes a C70 promoter via interactions similar but not identical to those used by E. coli EuJ7o. In addition, since R sphaeroides Eur93 protects a region of DNA comparable to that protected by E. coli Ea7, it predicts that this holoenzyme has roughly the same size and shape when bound to a promoter.

R. sphaeroides Eo93 initiates transcription from the R. sphaeroides nwB promoter, while E. coli Eo70 does not. When transcription by R sphaeroides RNAP was tested on an R. sphaeroides rRNA promoter (rrnB), a discrete transcript was produced by the partially purified S-300 material (Fig. 9A, lanes 1 and 2). To determine whether the Eu93 present in the S-300 material was responsible for this transcript, Q-Sepharose-fractionated RNAP was assayed on supercoiled templates (Fig. 9B). Eu93 (fraction F) generated the same rrnB transcript seen with the partially purified enzyme (Fig. 9B, lane 1). The limited transcription of R. sphaeroides rrnB from fraction H could reflect the small amount of U93 present or poor transcription of this promoter by Eur37 (Fig. 9B, lane 2). When r93 (fraction C) alone was used, no transcripts were detected (Fig. 9B, lane 3). However, when u93 was added to fraction H at a

1:1

5:1 molar ratio, transcription from R. sphaeroides rrnB stimulated to levels detected for Ec93 (Fig. 9B, lanes 1, 4, and 5; quantitation not shown). These results are consistent with the idea that Eu93 directs transcription from the R. sphaeroides rrnB promoter. Dryden and Kaplan have previously shown that R sphaeroides contains three rrn operons (13). Comparison of the was

or

7635


VOL. 175, 1993

R. sphaeroides F E. coli Ea7 R. sphaeroides H E. coli core R. sphaeroides C

E07' R. sphaeroides F

+ +

+

E. coli

+

+

+

I

5

lac uv5

BE. coli

lacuv5

A

+

+

I15

+ +

+I

+

core

R. sphaeroides H E. coli a7'

+ +

+

1

+

+

15s

s

*

*

1 2 3

45S 6 7 8 9

1

2 3

4 5 6

7

8

9

FIG. 7. In vitro transcription by heterologous RNAP holoenzymes. Plasmid templates contained the lacUV5 promoter cloned upstream of transcription termination signals. The lacU5 transcript is indicated by the arrow alongside each panel. The oriV transcript is denoted by the asterisk. The form of RNAP used in each assay is indicated at the top of each lane. Either a one- or fivefold molar excess of the E. coil a70 subunit (A) or R. sphaeroides fraction C (B) was added, as indicated above the lanes.

putative promoter sequences for R. sphaeroides rrnB, rrnA, and rrnC with the E. coli cr70 consensus promoter (Fig. 3C) showed that all three share some of the most highly conserved bases in the 35 and 10 hexamers. In addition, all three of these R sphaeroides rrn promoters contain the ideal 17-bp spacing between these elements; thus, it was not surprising to see that partially purified R sphaeroides RNAP also generated specific transcripts of approximately the expected sizes from rrnA and rrnC templates (data not shown). However, it was surprising to find that E. coli Eu70 produced no detectable transcripts with R. sphaeroides rrnB (Fig. 9A, lanes 3 and 4), rrnA, or rrnC (data not shown). Despite the overall similarity between Ec93 and Eu70 suggested by our experiments, this result indicates that Eu70 is defective in transcription from any of the R. sphaeroides rrn promoters under the conditions used in these assays. Possible explanations for this are presented in the Discussion. -

-

DISCUSSION To characterize the RNAP holoenzymes present in R. sphaeroides, we analyzed transcription in vitro by using a series of well-defined promoters. Earlier studies indicated that RNAPs from other photosynthetic bacteria contain core subunits similar in size to E. coli RNAP and that R. sphaeroides RNAP contains a 93-kDa polypeptide that has antigenic determinants in common with E. coli ur70 (17, 24, 43). Our results demonstrate that this 93-kDa protein is responsible for recognition of u70-dependent promoters. We have also dem-

onstrated that R sphaeroides contains an Ea37 holoenzyme which recognizes several E. coli a32-dependent promoters.

Synthesis of identical-size transcripts by enzymes from both bacteria suggests that the R sphaeroides RNAP holoenzymes are able to recognize the transcription terminators (rrnB T1 and T2 or spf ) located downstream of these promoters. The following sections describe the properties of R. sphaeroides Ea93 and E&37 revealed by our experiments. Properties of the R. sphaeroides Er93 holoenzyme. Our results suggest that R sphaeroides contains a functional homolog of E. coli Eur70. A partially purified R sphaeroides RNAP preparation produced identical-size transcripts from several promoters (lacUVS, rrnB P1, oriV, and Tn903kan) known to be recognized by E. coli Eu70. Of these, only Tn9O3kan is also known to be expressed in R sphaeroides (11). Primer extension analysis of in vivo Tn9O3kan mRNA (36) and the size of the Tn9O3kan transcript generated in vitro by both enzymes place the 5' end downstream of a putative promoter with limited homology to a U70 consensus promoter. Although homogeneous R sphaeroides RNAP subunits are not available, several findings strongly suggest that "93 is the sigma factor responsible for recognition of u70-dependent promoters. Monoclonal antibodies against a70 showed that "93 has a structural motif in common with conserved region 3. Addition of u93 to core RNAP from E. coli or R. sphaeroides increased transcription from the lacUVS promoter. Addition of E. coli r70 to R sphaeroides Ea37 stimulated transcription from

7636

J. BAcrRIOL.

KARLS ET AL. Bottom Strand

A.

Top Strand

B.

+

A+G sequence + no RNAP

E. coli Ea7" R. sphaeroides F

+

+

rrnl

rFnB

+

+

R. sphaeroides

R. sphaeroides

+

E. coli Eo70 R. sphacreides RNAP Template form

+

+

+

+

S

L

S

+ L

R. sphaerodes F R. sphaeroldes H R. sphaeroides C

+

+

+ +

+ S 4-

+17

-SCD

-53

1

+22

1

2

3

4

5

6

7

8

FIG. 8. DNase I footprints of the E. coli lacUVS promoter. The 207-bp EcoRI-BamHI lacUVS DNA fragment from plasmid pKO41acUV5 was used as the template for DNase I protection studies. Lanes: 1 to 4, the bottom strand was labeled at the EcoRI site located upstream of the start of transcription (+ 1); 5 to 8, the upper strand was labeled at the BamHI site downstream of the start of transcription initiation. The first lane in each set contained a Maxam-and-Gilbert A+G cleavage pattern. The second lane in each set contained the lacUV5 promoter fragment treated with DNase I. The third lane in each set contained the lacUV5 promoter fragment preincubated with E. coli Eu70 RNAP prior to DNase I treatment. The fourth lane in each set contained the lacUVS promoter fragment preincubated with R sphaeroides Er93 (Q-Sepharose fraction F) prior to DNase I treatment. Protected regions are highlighted by the bars beside lanes 1 and 8. The numbers refer to locations on the DNA sequence relative to the transcription initiation site. Similar amounts of RNAP (150 nmol) were used in each assay. Further details can be found in Materials and Methods.

70-dependent promoters. Although the cr93 and Eu93 fractions both contain a 29-kDa polypeptide, we do not believe that this protein is responsible for recognition of u70-dependent promoters. If this 29-kDa protein is a sigma factor, it does not contain the region 3 epitope that is shared by the major sigma subunits from several genera of bacteria (5). Therefore, when all of the data are considered, r93 is the most probable candidate for an R. sphaeroides C70 homolog.

2

1

4

FIG. 9. In vitro transcription from the R sphaeroides rrnB promoter region. (A) A plasmid containing the R sphaeroides rrnB promoter region cloned upstream of transcription termination signals was tested for transcription by R. sphaeroides RNAP after the S-300 column (Fig. 1) or by purified E. coli EuJ70 holoenzyme with supercoiled (S) and EcoRI-linearized (L) plasmid templates. The arrow denotes the transcript from the rrnB promoter. The asterisk denotes the transcript produced from the plasmid oriV promoter. (B) Reconstitution assays performed on the same promoter template with Q-Sepharose RNAP fractions as described in Materials and Methods.

From our results, we can also infer that some physical features of R sphaeroides Ea93 are similar to those of E. coli Er70. In addition to the region 3 epitope in u70 and u93, the core RNAP subunits from R sphaeroides must contain additional common motifs because they reacted with polyclonal antisera to E. coli RNAP in several studies (data not shown; 24). In addition, the interaction domains of R. sphaeroides core and u93 must be sufficiently similar to their E. coli counterparts to yield transcriptionally competent enzymes when subunits from these two bacteria are mixed. The similarity of the DNase I footprints on the lacUVS promoter made by R sphaeroides Eu93 and E. coli Eu70 suggests that the overall size and shape of these holoenzymes are also comparable. Regardless of the similarities between R sphaeroides Ea93 and E. coli Eu70, a significant functional difference between these enzymes was observed when activity was tested with R sphaeroides rmn promoters. From the size of this transcript generated by Ecr93 and the 5' end of the precursor rRNA species mapped in vivo (14), it appears that the R. sphaeroides rrnB promoter has limited homology to the E. coli u70 consensus ,Fig. 3C). Therefore, we were surprised to find that E. coli Ea70 failed to transcribe all three of the rrn promoters we

VOL. 175, 1993


tested (data not shown). The failure of the R. sphaeroides rrn promoters to function in E. coli (14) suggests that our results do not simply reflect the use of suboptimal in vitro assay conditions. Preliminary experiments suggest that this lack of R sphaeroides rrn transcription by Eu70 is due to functional differences in promoter recognition by the housekeeping sigma factors of these bacteria. For example, reconstitution experiments have occasionally yielded a holoenzyme competent at transcribing R. sphaeroides rrnB when R sphaeroides cr93 was added to E. coli core RNAP. In contrast, R. sphaeroides rrnB transcription was not observed when E. coli c70 was mixed with R. sphaeroides core RNAP. Unfortunately, a final conclusion on this point awaits further experiments since reproducible stimulation of R. sphaeroides rrnB transcription by a mixture of ub93 and E. coli core has been difficult to achieve. One possible explanation for the failure of E. coli Eu70 to transcribe R. sphaeroides rrn promoters is that Eu70 and Eu93 recognize slightly different consensus sequences. If this were true, nonconsensus bases present in the R. sphaeroides rrn promoters (Fig. 3C) could prevent Eu70 from forming a stable promoter complex under our assa conditions. Indeed, there are data which suggest that Ec7 analogs in bacteria with higher guanine-plus-cytosine contents in their DNAs recognize slightly different promoter elements. For example, in members of the family Rhodospirillaceae (1, 29, 30, 42) and in pseudomonads (18, 19), putative promoters with limited homology to the cu70 consensus have been detected upstream of transcript 5' ends. In addition, a compilation of a large number of streptomycete promoters (40) has led to a consensus (T T G A C Pu 16 to 18 bp T A g Pu Pu T) which differs somewhat from that recognized by E. coli Eur70 (Fig. 3C). An alternative explanation for the failure of Eu70 to transcribe the R. sphaeroides rrn promoters is that their transcription is not directed by the analogous form of holoenzyme in their natural host. While our data strongly suggest that Eu93 transcribes R. sphaeroides rrnB, until homogenous protein preparations are obtained, it remains to be proven that another sigma factor copurifying with ur93 is not responsible for the transcription of these rrn promoters. It is also possible that multiple sigma factors transcribe this promoter in R sphaeroides, since E. coli rrnB P1 is known to be transcribed by both Ecr70 and Eur32 (33). Properties of R. sphaeroides Eo37. Our results also suggest that R. sphaeroides has a holoenzyme related to E. coli Eu32. The separation of an activity capable of transcribing several E. coli c3 -dependent promoters indicated that this holoenzyme is distinct from Eu93. In addition, Western blot analysis revealed that samples which transcribe u32-dependent promoters contain a 37-kDa protein that reacts with antiserum against E. coli RpoH. While the reaction of this antibody to the 37-kDa polypeptide was not as strong as that to E. coli c32, the antibody should be highly specific for u32 since an extract from an E. coli rpoH null mutant was used to remove nonspecific antibodies. If this 37-kDa polypeptide is the functional equivalent of E. coli (r32, it suggests that R. sphaeroides also uses a specific cr factor to control heat shock protein synthesis (20). Attempts to clone this putative R. sphaeroides c32 homolog and analyze its physiological functions are in progress. In summary, our results have demonstrated that R. sphaeroides contains unique RNAP holoenzymes which recognize c 70- or a32-dependent promoters from well-studied enteric bacteria. Both of these R. sphaeroides holoenzymes were capable of faithfully initiating and terminating transcription from a variety of test promoters which do not require the presence of trans-acting factors. In addition to providing investigators with important information regarding the me-

7637

chanics of transcription and the elements of promoter structure in R. sphaeroides, these purified holoenzymes will provide the reagents necessary for analysis of how specific gene products regulate transcription of individual promoters in response to changing environmental conditions. ACKNOWLEDGMENTS This work was supported by NIH grant GM37509 to T.J.D. Predoctoral NIH training grant GM07215 supported R.K.K. These experiments were initiated while D.J.J. was supported by NIH grant AI19635 to C. A. Gross. We thank C. A. Gross, R. Gourse, W. Ross, and P. Kiley for critical reading of the manuscript. We are grateful to Anni Mitin for construction of plasmids pRKK60 and pRKK61, which were used in this work. We also thank Yan-ning Zhou, Janet Newlands, and members of C. A. Gross's laboratory for technical advice and assistance. REFERENCES 1. Armstrong, G. A., M. Alberti, F. Leach, and J. E. Hearst. 1989. Nucleotide sequence, organization, and nature of the protein products of the carotenoid biosynthesis gene cluster of Rhodobacter capsulatus. Mol. Gen. Genet. 216:254-268. 2. Balbas, P., X. Soberon, E. Merino, M. Zurita, H. Lomeli, F. Valle, N. Flores, and F. Bolivar. 1986. The plasmid vector pBR322 and its special-purpose derivatives-a review. Gene 50:3-40. 3. Brandner, J. P., A. G. McEwan, S. Kaplan, and T. J. Donohue. 1989. Expression of the Rhodobacter sphaeroides cytochrome c2 structural gene. J. Bacteriol. 171:360-368. 4. Burgess, R. R., and J. J. Jendrisak 1975. A procedure for the rapid, large-scale purification of Escherichia coli DNA-dependent RNA polymerase involving polymin P precipitation and DNAcellulose chromatography. Biochemistry 14:4634-4638. 5. Burgess, R. R., N. E. Thompson, S. A. Lesley, D. R. Gentry, D. A. Hager, and M. A. Brow. 1991. Use of monoclonal antibodies in studying the structure and function of E. coli RNA polymerase sigma factors, session 1, paper 1, p. 3-15. Gene Expression, Proceedings of the Second International Mochida Memorial Symposium. The Mochida Memorial Foundation, Tokyo. 6. Chen, C.-Y. A., G. R. Galluppi, and J. P. Richardson. 1986. Transcription termination at X tRl is mediated by interaction of rho with specific single-stranded domains near the 3' end of cro mRNA. Cell 46:1023-1028. 7. Covarrubias, L., L. Cervantes, A. Covarrubias, X. Soberon, I. Vichido, A. Blanco, Y. M. Kupersztoch-Portnoy, and F. Bolivar. 1981. Construction and characterization of new cloning vehicles. V. Mobilization and coding properties of pBR322 and several deletion derivatives including pBR327 and pBR328. Gene 13:2535. 8. Cowing, D. W. 1988. Ph.D. thesis. University of Wisconsin, Madison. 9. Cowing, D. W., J. C. A. Bardwell, E. A. Craig, C. Woolford, R. W. Hendrix, and C. A. Gross. 1985. Consensus sequence for Esche-

10.

11.

12.

13. 14. 15.

richia coli heat shock genes. Proc. Natl. Acad. Sci. USA 82:26792683. Cowing, D. W., and C. A. Gross. 1989. Interaction of Escherichia coli RNA polymerase holoenzyme containing u32 with heat shock promoters. DNase I footprinting and methylation protection. J. Mol. Biol. 210:513-520. Donohue, T. J., and S. Kaplan. 1991. Genetic techniques in Rhodospirillaceae. Methods Enzymol. 204:459-485. Donohue, T. J., A. C. McEwan, S. Van Doren, A. R. Crofts, and S. Kaplan. 1988. Phenotypic and genetic characterization of cytochrome c2 deficient mutants of Rhodobacter sphaeroides. Biochemistry 27:1918-1925. Dryden, S. C., and S. Kaplan. 1990. Localization and structural analysis of the ribosomal RNA operons of Rhodobacter sphaeroides. Nucleic Acids Res. 18:7267-7277. Dryden, S. C., and S. Kaplan. 1993. Identification of cis-acting regulatory regions upstream of the rRNA operons of Rhodobacter sphaeroides. J. Bacteriol. 175:6392-6402. Erickson, J. W., and C. A. Gross. 1989. Identification of the cE subunit of Escherichia coli RNA polymerase: a second alternate cr

7638

16.

17.

18.

19. 20. 21.

22. 23. 24.

25.

26. 27.

28. 29.

KARLS ET AL.

factor involved in high-temperature gene expression. Genes Dev. 3:1462-1471. Erickson, J. W., V. Vaughn, W. A. Walter, F. C. Neidhardt, and C. A. Gross. 1987. Regulation of the promoters and transcripts of rpoH, the Escherichia coli heat shock regulatory gene. Genes Dev. 1:419-432. Forrest, M. E., and J. T. Beatty. 1987. Purification of Rhodobacter capsulatus RNA polymerase and its use for in vitro transcription. Fed. Eur. Biochem. Soc. 212:28-34. Gao, J., and G. N. Gussin. 1991. RNA polymerases from Pseudomonas aeruginosa and Pseudomonas syringae respond to Escherichia coli activator proteins. J. Bacteriol. 173:394-397. Gao, J., and G. N. Gussin. 1991. Activation of the trpBA promoter of Pseudomonas aeruginosa by TrpI protein in vitro. J. Bacteriol. 173:3763-3769. Grossman, A. D., J. W. Erickson, and C. A. Gross. 1984. The htpR gene product of E. coli is a sigma factor for heat-shock promoters. Cell 38:383-390. Hager, D. A., D. J. Jin, and R. R. Burgess. 1990. Use of mono Q high-resolution ion-exchange chromatography to obtain highly pure and active Escherichia coli RNA polymerase. Biochemistry 29:7890-7894. Harley, C. B., and R. P. Reynolds. 1987. Analysis of Escherichia coli promoter sequences. Nucleic Acids Res. 15:2343-2361. Hawley, D. K., and W. R. McClure. 1983. Compilation and analysis of Escherichia coli promoter DNA sequences. Nucleic Acids Res. 11:2237-2255. Kansy, J., and S. Kaplan. 1989. Purification, characterization, and transcriptional analysis of RNA polymerases from Rhodobacter sphaeroides cells grown chemoheterotrophically and photoheterotrophically. J. Biol. Chem. 264:13751-13759. Kaplan, S., and T. J. Donohue. 1993. Genetic analysis of photosynthetic membrane biogenesis in Rhodobacter sphaeroides, p. 101-131. In J. Deisenhofer and J. R. Norris (ed.), The photosynthetic reaction center, vol. 1. Academic Press, Inc., Orlando, Fla. Kiley, P. J., and S. Kaplan. 1988. Molecular genetics of photosynthetic membrane biosynthesis in Rhodobacter sphaeroides. Microbiol. Rev. 52:50-69. Lee, J. K., P. J. Kiley, and S. Kaplan. 1989. Posttranscriptional control of puc operon expression of B800-850 light-harvesting complex formation in Rhodobacter sphaeroides. J. Bacteriol. 171: 3391-3405. Lowe, P. A., D. A. Hager, and R. R. Burgess. 1979. Purification and properties of the cr subunit of Escherichia coli DNA-dependent RNA polymerase. Biochemistry 18:1344-1352. Ma, D., D. N. Cook, D. A. O'Brien, and J. E. Hearst. 1993. Analysis of the promoter and regulatory sequences of an oxygen-regulated bcd operon in Rhodobacter capsulatus by site-directed mutagene-

J. BAC7ERIOL.

sis. J. Bacteriol. 175:2037-2045. 30. MacGregor, B. J., and T. J. Donohue. 1991. Evidence for two promoters for the cytochrome c2 gene (cycA) of Rhodobacter sphaeroides. J. Bacteriol. 173:3949-3957. 31. Maquat, L. E., and W. S. Reznikoff. 1978. In vitro analysis of the Escherichia coli RNA polymerase interaction with wild-type and mutant lactose promoters. J. Mol. Biol. 125:467-490. 32. Maxam, A. M., and W. Gilbert. 1980. Sequencing end-labeled DNA with base-specific chemical cleavages. Methods Enzymol. 65:499-560. 33. Newlands, J. T., W. Ross, K. K. Gosink, and R. L. Gourse. 1991. Factor-independent activation of E. coli rRNA transcription. Characterization of complexes of rrnB P1 promoters containing or lacking the upstream activator region with E. coli RNA polymerase. J. Mol. Biol. 220:569-583. 34. Pouwels, P. H., B. E. Enger-Valk, and W. J. Brammar. 1985. Cloning vectors, p. I-c-i-I. Elsevier Science Publishers BV, New York. 35. Ross, W., J. F. Thompson, J. T. Newlands, and R. L. Gourse. 1990. E. coli FIS protein activates ribosomal RNA transcription in vitro and in vivo. EMBO J. 9:3733-3742. 36. Schilke, B., and T. J. Donohue. Unpublished data. 37. Sheen, J.-Y., and B. Seed. 1988. Electrolyte gradient gels for DNA sequencing. BioTechniques 6:942-943. 38. Sistrom, W. R 1960. A requirement for sodium in the growth of Rhodopseudomonas sphaeroides. J. Gen. Microbiol. 22:778-785. 39. Strickland, M. S., N. E. Thompson, and R. R. Burgess. 1988. Structure and function of the u-70 subunit of Escherichia coli RNA polymerase. Monoclonal antibodies: localization of epitopes by peptide mapping and effects on transcription. Biochemistry 27: 5755-5762. 40. Strohl, W. R 1992. Compilation and analysis of DNA sequences associated with apparent streptomycete promoters. Nucleic Acids Res. 20:961-974. 41. Taylor, W. E., D. B. Straus, A. D. Grossman, Z. F. Burton, C. A. Gross, and R R. Burgess. 1984. Transcription from a heatinducible promoter causes heat shock regulation of the sigma subunit of E. coli RNA polymerase. Cell 38:371-381. 42. Wellington, C. A., and J. T. Beatty. 1989. Promoter mapping and nucleotide sequence of the bchC bacteriochlorophyll biosynthesis gene from Rhodobacter capsulatus. Gene 83:251-261. 43. Wiggs, J. L., J. W. Bush, and M. J. Chamberlin. 1979. Utilization of promoter and terminator sites on bacteriophage T7 DNA by RNA polymerases from a variety of bacterial orders. Cell 16:97109. 44. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene 33:103-119.