Organization of Genes Coding for LH，RC，and Pigment...

Organization of Genes Coding for LH，RC，and Pigment Biosynthetic Enzymes and Cytochromes

The genetic control of photosynthesis and photosynthetic membrane formation has been an important issue in the last few years.Recombinant DNA technology has made possible the isolation of most of the genes associated with the photosynthetic apparatus of the photosynthetic bacteria.The isolated genes have been subjected to a detailed analysis of nucleotide sequence and structure.Therefore，a great deal of information about the structure and function of the photosynthetic apparatus both in R.capsulatus and in R.sphaeroides has been obtained.

The study of photosynthetic genes in R.capsulatus is aided by an indigenous genetransfer agent（Yen and Marrs 1976）and isolation of a conjugative R-prime plasmid，pRPS404，containing R.capsulatus 46 kb DNA，which complements most mutations affecting the differentiation of the photosynthetic apparatus（Marrs 1981）.A procedure for the transposon mutagenesis of pRPS404 in Escherichia coli that allows for mapping of the positions of the transposon insertions was developed in our laboratory（Youvan et al.1982）.In order to improve the stability of the mutants generated in this way，a new transposon was created that had the transposition characteristics of Tn5 and the antibiotic resistant characteristics of Tn7（Zsebo et al.1984）.By screening both the earlier mutants generated by Marrs's laboratory and the transposon mutants generated in this laboratory for enhanced near-infrared fluorescence indicating a defect in the utilization of absorbed photons，two restriction fragments from the R-prime plasmid pRPS404 were found to complement all enhanced-fluorescence mutants defective in LH-Ⅰ and RC genes.The complete nucleotide sequence（8，867 bp）and a deduced polypeptide sequence for these two restriction fragments showed that they encoded 11 proteins from the photosynthetic gene cluster of R.capsulatus（Youvan et al.1984b）.Four structural genes for LH-Ⅰβ（pufB），LH-Ⅰ α（pufA），RC-L（pufL），RC-M（pufM），and two putative genes，C2397 and C2814，were found in the BamHI-C-EcoRI-B restriction fragment，whereas four putative genes-F108，F460，F1025，and F1696-and one structural gene for RC-H（puhA）were found in the BamHI-F fragment（see Figure 2）.The pufB，pufA，pufL，pufM，and C2397 genes are arranged contiguously and comprise a single operon：Anoxygenically induced mRNA hybridizes to this region（Clark et al.1984；Belasco et al.1985）and a specific oligonucleotide（17 mer）complementary to the C2397 gene hybridizes to a 2.6 kb long transcript from the puf operon（Zhu and Hearst 1986）.From the deduced amino acid sequences of the RC subunits it was concluded that both the L and the M subunits are very hydrophobic proteins with 282 and 307 amino acids，respectively.As mentioned earlier，hydropathy plots suggest that the L and M subunits are transmembrane proteins that may cross the membrane five times.On the other hand，the H subunit is a hydrophilic polypeptide（254 amino acids long）with a very hydrophobic amino terminal of 30 amino acids.All five structural and six putative genes possess a ShineDalgarno sequence（A/C）GGAG（A/G）N3-10 ATG that is complementary to the 3'terminal of R.capsulatus 16S rRNA.However，no E.coli-like consensus promoters were found over the entire 9 kb of sequenced DNA.Three potential hairpins have been found between the genes for LH-1 α and RC-L，after the gene for C2397，and between the genes for RC-H and F3981.

The puf and the puh operons are separated by 35 kb and are transcribed in the opposite direction.A large number of genes（at least 17）associated with Bchl and Crt biosynthetic enzymes are located between the puf and the puh operons（see Figure 2）.Studies of the structure and the regulation of genes for pigment biosynthesis are in progress in our laboratory.

The genes coding for another LH complex and theβ（pucB）and a（pucA）subunits of LH-Ⅱ were isolated using a deoxyoligonucleotide probe and sequenced（Youvan and Ismail 1985）.This completed the sequencing of all seven structural genes coding for the LH and RC polypeptides that bind the pigments and cofactors participating in the primary light reaction of photosynthesis.The pucA and the pucB genes are at least 11 kb away from pufL and pufM，or at least 7 kb away from puhA.Since the organization of theβ and a structural genes for both LH-Ⅰ and LH-Ⅱ is identical，and very high homologies exist between the two β-polypeptides and between the two α-polypeptides，it has been suggested that both complexes arose by gene duplication from a single ancestral LH gene.A highly conserved sequence，Ala-x-x-x-His，is believed to bind Bchl（Youvan and Ismail 1985）.

Hearst and Sauer（1984a）first reported at the Vlth International Congress on Photosynthesis in 1983 that the RC subunits L and M of R.capsulatus share amino acid sequence homology with the QB protein of spinach.Particularly，a strong homology has been found between the L subunit and the QB protein（residues 196-221）.It was proposed that this sequence of amino acids is involved in quinone binding and function（You van et al.1984a and 1984b；Hearst and Sauer 1984a and 1984b）.

The pufM（Williams et al.1983），pufL（Williams et al.1984），puh A（Donohue et al.1986a），pufB，pufA（Kiley et al.1987a），and pucB and pucA（Kiley et al.1987b）in R.sphaeroides have also been isolated and sequenced.The same arrangement of genes coding for LH-Ⅰ（β，α）and RC（L，M）proteins and a striking homology of deduced amino acid sequences of the LH and RC polypeptides were found between R.sphaeroides and R.capsulatus.The considerable sequence homology between the L and M subunits indicates that their genes arose by duplication of a single gene.The RC-L and RC-M polypeptides of R.sphaeroides also share homology with the 32 kd thylakoid membrane protein in spinach（Williams et al.1983 and 1984）and are proposed to span the membrane five times，like the RC-L and RC-M polypeptides in R.capsulatus.One of the histidine residues（His-219）in a hydrophobic segment of the RC-M subunit is a candidate for a Bchl or a quinone binding site（Williams et al.1983）.The genes pufB and pufA are contiguous with pufL and pufM.A region about 170 bp upstream from the 5'end of the structural gene is sufficient for the in vitro expression of LH-Ⅰβ and α polypeptides（Kiley et al.1987a）.In addition，three regions of dyad symmetry upstream from pufB have been found.These inverted repeats may serve as recognition sites for DNA binding proteins in regulating expression of this operon.

Unlike most of the other photosynthetic bacteria，R.rubrum contains a single holochrome antenna complex，designated B880，that also consists ofβ and α subunits.Genes for B880 holochrome have recently been isolated and sequenced（Bérard et al.1986）.The genes for theβ and α polypeptides are contiguous and show high homologies to pufB and pufA in R.capsulatus.They are presumed to be transcribed as a single operon.An additional C-terminal tail of 10 and 13 amino acid residues on the α andβ subunits，respectively，may serve as a means of orienting these polypeptides in the membrane.

The pufL and pufM（Michel et al.1986）and the puhA（Michel et al.1985）genes in R.viridis have been cloned using oligonucleotide hybridization and expression vectors，respectively，and sequenced.The arrangement of pufL and pufM is the same as in R.capsulatus and R.sphaeroides.However，due to a deletion of five bases at the end of pufL in R.viridis，pufL and pufM do not overlap.Therefore，overlap of these two genes is not necessary for 1∶1 stoichiometry of the L and M polypeptides（Williams et al.1984）.Another structural feature different from R.capsulatus and R.sphaeroides is that the gene encoding the RC cytochrome c2，which is tightly bound to the RC and unique in R.viridis，is located at the end of pufM and overlaps pufM by one base pair.In R.capsulatus pufM is followed by an unknown gene C2397（Youvan et al.1984b）.Furthermore，the M subunit from R.viridis has 17 more amino acids than R.capsulatus and 18 more amino acids than R.sphaeroides at its carboxy-terminal.This extension binds the cytochrome subunit（Deisenhofer et al.1985a and 1985b）.

Cytochrome c2 is an electron carrier that is common to the photosynthetic and respiratory machinery of R.capsulatus（Bartsch 1978）.The structural gene for cytochrome c2（cycA）of R.capsulatus has been cloned and sequenced（Daldal et al.1986）.Comparison of the known amino acid sequence of the purified cytochrome c2 and the nucleotide sequence corresponding to the N-terminal part of the cycA product indicates the presence of a putative 21-amino-acid signal sequence.Cytochrome c2 may be synthesized as a precursor that is processed during its secretion to the periplasm.Studies of the insertion and insertion-deletion mutations have led Daldal et al.to conclude that cytochrome c2 is not essential for photosynthetic growth.Donohue et al.（1986b）have analyzed cycA in R.sphaeroides and have also found a 21-amino-acid signal sequence.Small（740 nucleotide）and large（920 nucleotide）cycA-specific mRNA species are found in various growth conditions.In contrast to the result of R.capsulatus（Daldal et al.1986），using cartridge insertion coupled with partial deletions of the cycA gene，they have found that cytochrome c2 obligatory for photosynthetic growth of R.sphaeroides.

Recently the genes encoding the three main subunits of the cytochrome bc1 complex in a strain of R.sphaeroides have been cloned（Gabellini et al.1985）.This strain was identified later by Davidson and Daldal（1987）as most likely a strain of R.capsulatus.The genes are arranged in the order（5'）fbcF，fbcB，and fbcC，encoding the FeS protein，cytochrome b，and cytochrome c1，respectively.The fbc genes are coordinately transcribed in one polycistronic mRNA and constitutively expressed under both photosynthetic and respiratory growth conditions.The nucleotide sequence of 3，874 bp of cloned chromosomal DNA，including the three structural genes fbcF，fbcB，and fbcC，has been determined（Gabellini and Sebald 1986）.The fbcF gene codes for 191 amino acids of the FeS protein，starting from a GTG codon that is preceded by an adenine and guanine-rich sequence complementary to the 3'end of the 16S ribosomal subunit of R.sphaeroides.The fbcF gene is followed by the fbcB gene with a 12 nucleotide space that includes a Shine-Dalgarno-like sequence.The fbcB gene codes for 437 residues of the cytochrome b subunit.The fbcC gene starts 20 bp after the fbcB gene and is preceded by a Shine-Dalgarno-like sequence.The fbcC gene codes for a 280-residue preapocytochrome c1 that has a transient leader sequence（21 amino acids）.S-l nuclease protection analysis indicates that the transcription of the fbc operon starts approximately 240 bp upstream from the start codon of the fbcF gene and terminates 120 bp downstream from the stop codon of the fbcC gene.The deduced amino acid sequences，homologies with other similar proteins，and hydropathy plots indicate the following structural features：① FeS protein，carrying a high-potential 2Fe-2S cluster，has a large hydrophilic domain and is exposed to the aqueous phase；② cytochrome b has nine or ten possible membrane-spanning regions，and four conserved histidine residues are postulated as ligands of the two heme groups；③ cytochrome c1 is synthesized as a precursor form including a transient leader sequence of 21 amino acids.The typical sequence that binds the heme covalently is located near the N-terminal corresponding to Cys-55，Cys-88，and His-50.(https://www.daowen.com)

Similar results of a detailed analysis of the fbc operon in R.capsulatus were recently reported by Davidson et al.（1987）.They designated the fbc operon as the photosynthetic electron transport（pet）operon.A possible promoter region was found to be located upstream from the first petA structural gene coding for Rieske FeS protein.It is concluded that unlike cytochrome c2（Daldal et al.1986），a functional bc1complex is essential for photosynthetic growth.

rRNA from both R.sphaeroides（Marrs and Kaplan 1970）and R.capsulatus（Yu et al.1982）consists of 23S，16S，and 5S species.The 23S rRNA，however，is not stable during isolation and is cleaved into 16S and 14S rRNAs probably by a specific endonuclease（Yu et al.1982；Klug et al.1984；Zhu and Kaplan 1985）.The rRNA genes of R.capsulatus have been cloned（Yu et al.1982）.At least seven rRNA operons are present in the R.capsulatus chromosome.The arrangement of 16S，23S，and 5S rRNAs is the same as that in E.coli.However，the spacer（0.91 kb）between the 16S and the 23S rRNA genes is longer than that（0.44 kb）in E.coli.

Although ribulose-1，5-bisphosphate carboxylase（RuBPCase）is not a membrane protein，a great deal of attention has been paid to it because it is the most abundant protein in the cytoplasm and it plays a key role in carbon dioxide fixation in the Calvin cycle in photosynthetic bacteria，algae，and plants.It is of particular interest that two forms of RuBPCase are present in R.sphaeroides and R.capsulatus（Gibson and Tabita 1977a and 1977b）.The form I RuBPCase resembles that found in plants and most other bacteria.It is a large oligomeric protein composed of eight large（52 kDa）and eight small（11 kDa）subunits（Gibson and Tabita 1977b）.The form Ⅱ enzyme is a hexamer of only large subunits（52 kDa）and does not have a small subunit.Recently，the gene for the form ⅡRuBPCase（rbcL）was isolated from R.rubrum（Somerville and Somerville 1984）and R.sphaeroides（Fornani and Kaplan 1983；Quivey and Tabita 1984；Muller et al.1985）.The nucleotide sequence of the rbcL gene from R.rubrum has been reported（Nargang et al.1984）.No consensus prokaryotic promoter sequence was found in the rbcL gene from R.rubrum.It is likely that the promoters for highly regulated genes such as those involved in carbon dioxide fixation and the photosynthetic apparatus are quite different from typical bacterial promoters（Quivey and Tabita 1984）.A region of dyad symmetry 5'to the coding region of the R.rubrum RuBPCase gene similar to the proposed attenuator regions was found，indicating that additional proteins may be involved in transcription（Nargang et al.1984）.

One of the consequences of the knowledge of structure and organization of nucleotide sequences of the photosynthetic genes in purple photosynthetic bacteria is the ability to mutate the genes that are of interest by site directed mutagenesis and to examine the effect of the resulting mutation on the function of the photosynthetic apparatus（Youvan and Ismail 1985）.Another consequence that will be described here in more detail is to facilitate the elucidation of structure and function of the photosynthetic apparatus in plants and algae as well as in purple bacteria.Proteins can be studied based on sequence homologies that may shed some light on functionally important structural features（Youvan et al.1984a and 1984b；Hearst and Sauer 1984a and 1984b；Hearst 1986）.Amino acid sequence comparison between photosynthetic membrane proteins of purple bacteria and those of blue-green algae and higher oxygen evolvers remains an extraordinarily valuable and insightful exercise.A simple generalization is that there is strong sequence homology between proteins of identical function in all bluegreen algae and higher plants.The purple bacteria are very different but have small regions of homology with plant proteins.Such homologous regions provide special insight regarding the essential functional sites in the corresponding proteins.

A weak，but significant，sequence homology exists between the L and M subunits of the RC of R.capsulatus and the 32 kDa thylakoid membrane protein（also known as D1，QB，herbicide-binding protein）of the RC of photosystem Ⅱ in higher plants and algae（Zurawski et al.1982；Youvan et al.1984b；Hearst and Sauer 1984b；Hearst 1986）.Hearst and Sauer（1984b）proposed a model suggesting that a highly conserved sequence of amino acids between the L and M subunits of the RC of R.capsulatus and 32 kDa thylakoid membrane protein of spinach and tobacco is involved in quinone binding and function.Figure 3（A）reveals that a sequence PFHMLG----F------AMHG-LV-S is common to both the L subunit of R.capsulatus and the 32 kDa（Dl）proteins of spinach，Nicotiana，Chlamydomonas，and Euglena and starts at the one hundred seventy-second amino acid from the N-terminal of the L subunit（282 amino acids）and at the one hundred ninety-sixth amino acid from the N-terminal of the 32 kDa（Dl）protein（353 amino acids）.Later another photosystem Ⅱ RC protein，the D2 protein，was also found to possess sequence homologies to the Dl protein and the L and M subunits（Rochaix et al.1984；Alt et al.1984；Holschuhk et al.1984）.A weaker homology between the R.capsulatus M subunit（from the one hundred ninety-eighth amino acid of a 306-amino-acid molecule）and the pea and Chlamydomonas D2 protein also reveals a conserved sequence PFHM----G-L-A-LLC---GATV（see Figure 3（B））.The data strongly indicate that the L and M subunits and the 32 kDa protein arose from a common precursor.The fact that selection pressure preserved such a precise amino acid sequence over three billion years（Olson 1981）suggests that this region is functionally very important.Subsequently，the crystal structure of the RC from R.viridis implicated the conserved histidine residues in forming the site for the non-heme Fe2+（Deisenhofer et al.1985a and 1985b）.The discovery of sequence homologies between the L and Dl proteins and between the M and D2 proteins suggests that the Dl and D2 proteins are involved in photochemical activity，a hypothesis in conflict with the old concept that two other larger polypeptides of 47—51 kDa and 43—45 kDa are the core of the photosystem Ⅱ.

This hypothesis is further supported by the finding that both the L subunits and Dl proteins bind the photo affinity-labeled herbicide azido atrazine（Pfister et al.1981；DeVitry and Diner 1984）.Through the use of a tritium-labeled quinone alkylating reagent，Worland et al.（1987）have demonstrated the affinity labeling of a 38 kDa protein that they suggest may be D2 in spinach photosystem Ⅱ complexes.In addition，the x-ray structure analysis has shown that both L and M subunits are needed to establish the primary electron donor，which is made up of two nearly parallel Bchl moleucles（the special pair）and the electronaccepting quinone-iron com-plex.All the evidence supports the proposal of Hearst（1986）and Michel and Deisenhofer（1986）that the Dl and D2 proteins form the core of the photosystem Ⅱ RC and the L and M subunits form the RC core of purple photosynthetic bacteria in similar ways.The L subunit is assumed to be the equivalent of Dl，and M to be the equivalent of D2.

Figure 3（a）provides a sequence alignment that clearly accentuates the additional sequences in the higher plants that are not found in the corresponding protein in the purple bacteria.The region of particular interest to us at the present time is the carboxy terminal of the Dl or herbicide-binding protein.This region is highly conserved among the oxygen evolvers and might be associated with the oxygen-evolving manganese center.Evidence for such a hypothesis comes from several sources.First，Metz et al.（1987）have identified a Dl mutant that is incapable of oxygen evolution.Second，Dl in higher plants is rapidly turned over because of photolability，while the L subunit in purple bacteria is quite stable（Kyle et al.1984；Ohad et al.1984；Bishop 1987）.This photolability might well be the consequence of the presence of the oxygen-evolving site.Finally，the amino acid composition of this carboxy terminal region is unusual，containing three acidic residues，four asparagines，two arginines，two methionines，one phenylalanine，and one proline in the twenty-one residues from position 322 to position 342 on the alignment in Figure 3（a），as indicated by the bracket.This suggests a function for this region that must operate on the periplasmic side of the photosynthetic membrane，where oxygen evolution is known to occur，if the analogy with the purple bacteria RC structure of Michel and Deisenhofer（1986）is valid.

Other regions of this protein that are believed to be in the periplasm and that may also be involved with the site of oxygen evolution include amino acids 56 to 76 and 92 to 108，the aspartic acid at position 170，and the amino acids from 308 to 321，as marked with brackets in Figure 3（a）.