Chloroplast rpoA, rpoB, and rpoC genes specify at least three components of a chloroplast DNA-dependent RNA polymerase active in tRNA and mRNA transcription.

The purpose of this study was to determine the relationship between putative chloroplast RNA polymerase subunit genes and known chloroplast transcriptional activities. We have prepared fusion polypeptide genes from fragments of chloroplast DNA homologous to bacterial RNA polymerase subunit genes and expression vectors carrying portions of the anthranilate synthetase gene (trpE). Fusion proteins for chloroplast homologs of the RNA polymerase alpha (rpoA), beta (rpoB), and beta' (rpoC) subunits were obtained from these genes. The fusion polypeptides synthesized by Escherichia coli in vivo were purified and used as antigens for production of rabbit polyclonal anti-RNA polymerase subunit-specific antibodies. The purified antibodies were able to immobilize chloroplast DNA-dependent RNA polymerases from spinach, pea, and Euglena gracilis. In addition, the soluble chloroplast RNA polymerase activity in tRNA and mRNA synthesis was strongly inhibited by these antibodies under conditions which had little effect on transcription by the chloroplast transcriptionally active chromosome that preferentially transcribed rRNA genes (Greenberg, B. M., Narita, J. O., DeLuca-Flaherty, C., Gruissem, W., Rushlow, K. A., and Hallick, R. B. (1984) J. Biol. Chem. 259: 14880-14887). From these data we conclude that the chloroplast genes homologous to bacterial RNA polymerase subunit genes are expressed in vivo and that the protein products specify at least three of the components of the chloroplast RNA polymerase(s) involved in tRNA and mRNA transcription.

minants with RNA polymerases from other eukaryotes, and also share relatedness to each other (12). Another characteristic of nuclear polymerases is that protein factors are required for accurate initiation of RNA synthesis with cloned templates in vitro ( 5 ) .
DNA-dependent RNA synthesis from organellular genes is catalyzed by distinct DNA-dependent RNA polymerases that differ from the nuclear enzymes (1,(8)(9)(10)(11)(12)(13)31). In chloroplasts of Euglena gracilis and higher plants, there are at least two RNA polymerase activities that are each capable of selective transcription of different classes of genes (9). One of these enzymes, which is tightly bound to the chloroplast chromosome, is denoted the chloroplast TAC' (for transcriptionally active chromosome). It is capable of faithfully transcribing the rRNA genes, in analogy to the nuclear class I enzyme (8,9,12). The other activity, denoted the "soluble" activity, may be readily dissociated from the chloroplast DNA with 0.5 M KC1. This soluble enzyme retains the ability to faithfully transcribe, in a template-dependent manner, mRNA (22) and tRNA genes (8,10,11). Aside from the biochemical differences between these two RNA polymerases, putative subunits for the TAC from Euglena (19), spinach ( l ) , and mustard ( 2 ) have been reported that are different in some regard from subunits of the highly purified "soluble" enzyme (16). Collectively, these studies are consistent with the idea that at least two RNA polymerases are involved in the transcription of chloroplast genes.
Recently, several chloroplast DNA sequences with significant homology to bacterial RNA polymerase subunit genes have been reported (20,27,28). Sijben-Muller et al. (28) first reported the sequences of a gene with homology to the gene for the a subunit of Escherichia coli RNA polymerase (denoted rpoA) (Fig. 1). This was followed by the publication of the complete chloroplast DNA sequences of Marchantia (20) and tobacco (27), which also contained an a-like subunit gene. In addition, these plants contain two additional chloroplast genes, possibly contained within a single transcription unit, with homology to the p' (rpoC gene product) and p (rpoB gene product) subunits of E . coli RNA polymerase. Euglena chloroplast DNA also contains both an rpoB2 and rpoC3 gene. The Marchantia (20), spinach: and Euglena3 chloroplast rpoC loci appear to encode two polypeptide genes that have been designated rpoCl and rpoC2. These loci are homologous to the amino-terminal and carboxyl-terminal domains, respec-  (28). B, comparison of the RNA polymerase p subunit of tobacco (middle, 1070 codons) with that of E. coli (lower, 1343 codons) and with RPBZ, the yeast RNA polymerase I1 second largest subunit (upper, 1225 codons). Regions of homology are indicated by the blackened areas. The length of the homology domains and the percent identical residues are indicated. Note the two deletions in the tobacco gene relative to the E. coli gene. Regions of the coding sequence used for the expression cloning are indicated by arrows.
tively, of E. coli rpoC. Based on this DNA sequence analysis, it appears that the homolog of the p' subunit in bacteria may occur as two separate polypeptides in chloroplasts.
The chloroplast rpoA, rpoB, rpoCI, and rpoC2 genes are a11 transcribed. The spinach chloroplast rpoA gene is expressed as determined by Northern hybridization and in vitro translation (28) of chloroplast RNA. Transcripts of the tobacco genes were also reported (27). The Euglena rpoB-rpoC(1-2) locus lies within the Bum-Sal fragment designated BS-4 of the chloroplast DNA.3*4 Transcripts from this locus were previously quantified in this laboratory (3, 4). Therefore, we conclude that the chloroplast RNA polymerase loci are expressed and most likely encode components of the chloroplast transcriptional apparatus. In contrast, there is some indirect evidence that chIoroplast RNA polymerases are nuclear-encoded (6, 15).
In the present study, we have begun to analyze the relationship between the protein products of these putative chloroplast DNA-encoded RNA polymerase genes and the known transcriptional activities from chloroplasts (Fig. 2). In this paper we report that these genes do indeed specify at least three components of at least one of the chloroplast RNA polymerases.

MATERIALS AND METHODS AND RESULTS AND DISCUSSION'
Chloroplast RNA Polymerase Gene Structure and Fusion Protein Expression-Defined segments of the spinach rpoA, tobacco rpoB, and Euglena rpoC loci were cloned as transla-Portions of this paper (including Materials and Methods, part of "Results and Discussion," and Fig. 1s) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press. The tobacco chloroplast rpoB gene (Fig. 1B) has four major domains of homology with the bacterial ,t? subunit gene and five domains with that of the yeast RNA polymerase I1 second largest subunit (28). The region of the chloroplast rpoB that was cloned for expression is an internal BamHI fragment spanning codons 573-935. This region is 42% homologous to the bacterial rpoB gene. This chloroplast gene contains two large deletions relative to the bacterial gene of approximately 100 codons each. Interestingly, one of these deletions (codons 976-1080 in the E. coli gene) occurs in a region of tandem repeat of the E. coli rpoB amino acid sequence. It corresponds to codons 965-1083 in the chloroplast gene. An E. coli RNA polymerase mutant with deletions in this region of the rpoB gene has been isolated which has an altered promoter specificity relative to the wild type (7).
The third fusion gene construct involved the E. gracilis rpoC locus. This gene lies proximal to a tRNA gene cluster (21). The carboxyl-terminal DNA sequence has been previously reported, but the rpoC gene was not identified6 (21). The entire gene has subsequently been seq~enced.~ The region of the gene that was used for the expression cloning begins a t an internal Hind111 site and continues to the carboxyl terminus. As shown in Fig. lS, this region is 39% homologous in derived amino acid sequence to the bacterial gene from condons 1110 to 1363 and 38-42% homologous to the carboxyl terminus of different plant rpoC2 loci.
Expression of Chloroplast RNA Polymersae Fusion Protein in E. coli-The segments of the chloroplast rpoA, rpoB, and rpoC genes that had been cloned as translation fusion polypeptide genes by joining with the E. coli trpE gene were each In the original published DNA sequence an omitted base resulted in a frameshift error. The original sequence beginning at base 310 read 5"CTTGTGCTGC. The correct sequence is 5"CTTGTAG-CTGC (C. Radebaugh and R. B. Hallick, unpublished observations). expressed in uiuo upon induction of host cells with indoleacrylic acid. The proteins synthesized in response to induction were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. For each fusion gene construct, a major new polypeptide product was synthesized following induction. The fusion polypeptides were found to be the major protein components in the insoluble protein pellets from lysed cells (Fig.   3). Lanes I, 2, and 3 are solubilized polypeptides derived from the insoluble pellets of cells expressing the rpoC, rpoB, and rpoA fusion genes, respectively. The yield for rpoA and rpoB was 100 mg of fusion polypeptide/liter of culture, with that of rpoC being approximately 4-fold lower. The reason for the lowered expression of the Euglena rpoC fusion gene is not known. We have observed the yield to also depend on the host E. coli strain, with HBlOl giving much lower expression than RR1. The fusion polypeptides were purified by preparative electrophoresis and used as antigens in the preparation of anti-RNA polymerase subunit antisera (see "Materials and Methods" in the Miniprint Supplement' for experimental details).
Binding of Chloroplast RNA Polymerases with the Fusion Protein Antibodies-The ability of RNA polymerases from different chloroplast sources and also from E. coli to be immobilized by the fusion protein antibodies is shown in Fig.  4. In this experiment, the binding of filter-bound immobilized antibodies to solvent-accessible epitopes of the enzyme in solution was first accomplished. Then, the ability of antibodybound enzyme to synthesize RNA in situ was used as a "reporter" activity to assess the binding reaction. As a control, the in uitro transcription products were all shown to be  Adsorption of these antibodies against an E. coli lysate removes the determinants against trpE, but some cross-reactivity against the bacterial subunits is retained. The antibody to the bacterial enzyme also binds the chloroplast RNA polymerases from each source. This is expected since the chloroplast and bacterial genes are also homologous ( Fig. 1s and Supplemental "Results and Discussion"). In contrast, the preimmune antibody (lane 4 ) shows little, if any, binding of the enzymes. Collectively, these data are a demonstration of the relationship between the potential protein products encoded by the chloroplast rpoA, rpoB, and rpoC genes and the RNA polymerase enzymatic activity of the "soluble" enzymes from chloroplasts of Euglena, pea, and spinach.
Inhibition of Chloroplast Transcription by Fusion Protein Antibodies-The transcription of the intron-containing chloroplast valine tRNA gene of spinach chloroplast DNA by the soluble chloroplast RNA polymerase activity has previously been described (11). As a second test of the relationship between the potential rpoA-, rpoB-, and rpoC-gene products and the known chloroplast transcriptional activities, selective inhibition of chloroplast tRNA transcription by the highly purified fusion protein antibodies was assayed. Transcription of a spinach pre-trnV template by the spinach-"soluble" enzyme is sensitive to the addition of each of the fusion protein antibodies (Fig. 5A). No inhibition is observed when preimmune antibody is added. The inhibition of pre-tRNA""' transcription is not a result of ribonuclease contamination, since the incubation of antibody with radiolabeled transcripts following RNA synthesis did not result in a decrease in the intensity of the transcript band (Fig. 5B). The antibodies purified via DEAE-Affi-Gel Blue appear free of detectable RNase. The observed immunoinhibition of transcription by these antibodies is consistent with the results on in situ RNA spinach chloroplast TAC is shown in Fig. 6A. The RNA produced by the spinach chloroplast TAC in panel A shows hybridization to a cloned tobacco rRNA gene as well as to Sari-9 and BarnHI-7 fragments of tobacco chloroplast DNA.
The latter fragments of chloroplast DNA contain genes specifying tRNAs andpsbE polypeptide, and a cluster of ribosomal proteins, respectively. Thus, the spinach chloroplast TAC is similar to that of mustard in that it produces in uitro transcripts from protein-coding loci and rDNA genes (23). In contrast, the Euglena TAC makes only rDNA, as seen in the Euglena TAC RNA product.
The effect of antibody concentration on both Euglena and spinach TAC RNA synthesis was determined (Fig. 7). Preincubation of antibody with the TAC has little if any effect on transcription from the rDNA genes. The same amount of any of the antibodies completely inhibited tRNA transcription by the soluble RNA polymerases (Fig. 5). In addition, an attempt was made to bind the TAC RNA polymerases to the cellulose nitrate-bound antibodies. No in situ RNA synthesis could be obtained. From these results, it may be concluded that the anti-rpoA, rpoB, and rpoC antibodies can discriminate between the in uitro RNA synthesis reactions of the soluble uersus the TAC RNA polymerases. Whether this discrimination is due to a different subunit structure for the TAC uersus the "soluble" polymerase, or merely reflects the fact that the epitopes for the TAC subunits are inaccessible to the antibodies, is unknown. However, the inaccessibility is apparently not due to the DNA component of the TAC inhibiting the binding of the antibodies. We find that in a Euglena "lysed chloroplast" run-on transcription system, where in uiuo initiated transcripts are extended in uitro, mRNA transcription presumably from DNA-bound RNA polymerase is still sensitive to the anti-rpoA, B, and C a n t i b~d i e s .~ We have exhaustively attempted to determine the size of the subunits of the soluble and TAC RNA polymerase that immuoreact with the antibodies. Unfortunately, the signals obtained to date with the fusion peptide antibodies as probes for Western blots of chloroplast RNA polymerase preparations have been very weak and inconclusive.
In summary, the purified antibodies were able to immobilize chloroplast DNA-dependent RNA polymerases from spinach, 'A. Sun and R. B. Hallick, unpublished observations. pea, and E. gracilis as determined by an antibody-linked in situ RNA polymerase assay. The tRNA-specific RNA polymerases from spinach, pea, and Euglena were also strongly inhibited by these antibodies. By contrast, the same antibody treatments had little effect on RNA synthesis by the transcriptionally active chromosome RNA polymerases from both Euglena and spinach chloroplasts. From these data we conclude that the chloroplast genes homologous to the E. coli rpoA, rpoB, and rpoC are expressed in organello and that the protein products specify at least three of the components of the soluble RNA polymerase. The identification of genes for subunits of the TAC RNA polymerase activity was not possible with the available antibodies.