Accumulation of chloroplast <i>psbB</i> RNA requires a nuclear factor in <i>Chlamydomonas reinhardtii</i>

(1)

Article

Reference

Accumulation of chloroplast psbB RNA requires a nuclear factor in Chlamydomonas reinhardtii

MONOD KAS, Caroline, GOLDSCHMIDT-CLERMONT, Michel P., ROCHAIX, Jean-David

Abstract

We have isolated and characterized a nuclear mutant, 222E, in Chlamydomonas reinhardtii, which is defective in photosystem II (PSII). Polypeptide P5, the product of psbB, is not produced in this mutant, leading to a destabilization of other PSII components. The mutant specifically fails to accumulate psbB transcripts and displays an altered transcription pattern downstream of psbB. Pulse-labelling experiments suggest that mRNA stability and/or processing are affected by the alteration of a nuclear gene product in this mutant. We show that the C. reinhardtii psbB gene is co-transcribed with a small open reading frame that is highly conserved in location and amino acid sequence in land plants. The 5′ and 3′ termini of the psbB transcript have been mapped to 35 bases upstream of the initiation codon and approximately 600 bases downstream of the stop codon. The 3′ flanking region contains two potential stem-loops, of which the larger (with an estimated free energy of −46 kcal) is near the 3′ terminus of the transcript.

MONOD KAS, Caroline, GOLDSCHMIDT-CLERMONT, Michel P., ROCHAIX, Jean-David.

Accumulation of chloroplast psbB RNA requires a nuclear factor in Chlamydomonas reinhardtii . Molecular and General Genetics , 1992, vol. 231, no. 3, p. 449-459

DOI : 10.1007/BF00292715

Available at:

http://archive-ouverte.unige.ch/unige:138212

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Mol Gen Genet (1992) 231:449-459

MG G

Accumulation of chloroplast psbB RNA requires a nuclear factor in Chlamydomonas reinhardtii

Caroline Monod, Michel Goldschmidt-Clermont, and Jean-David Rochaix

Departments of Molecular Biology and Plant Biology, University of Geneva, 30 Quai Ernest-Ansermet, Ch-1211 Geneva 4, Switzerland Received July 30, 1991

Summary. We have isolated and characterized a nuclear mutant, 222E, in Chlamydomonas reinhardtii, which is defective in photosystem II (PSII). Polypeptide P5, the product ofpsbB, is not produced in this mutant, leading to a destabilization of other PSII components. The mutant specifically fails to accumulate psbB transcripts and displays an altered transcription pattern downstream of psbB. Pulse-labelling experiments suggest that mRNA stability and/or processing are affected by the alteration of a nuclear gene product in this mutant. We show that the C. reinhardtii psbB gene is co-transcribed with a small open reading frame that is highly conserved in location and amino acid sequence in land plants. The 5' and 3' termini of the psbB transcript have been mapped to 35 bases upstream of the initiation codon and approximately 600 bases downstream of the stop codon. The 3' flanking region contains two potential stem-loops, of which the larger (with an estimated free energy of -46 kcal) is near the 3' terminus of the transcript.

Key words: Chloroplast genes - Chloroplast mRNA sta- bility - RNA 3' inverted repeat - Photosynthetic mutant

Introduction

Photosynthesis requires the coordinated regulation of gene products of both the nuclear and chloroplast genomes. The photosystem II (PSII) complex, which per- forms the initial reactions that convert light to chemical energy, is a multimolecular protein-pigment complex.

The core consists of the polypeptides D1, D2, P5, P6, two apocytochromes bss9 and a 4.5 kDa polypeptide, encoded by the chloroplast genes psbA, psbD, psbB, psbC, psbE, psbF and psbI, respectively. Three nuclear encoded polypeptides on the lumen side of the thylakoid membrane are required for oxygen evolution, and sever-

al additional small molecular weight polypeptides encoded by the chloroplast genes psbH to psbN have been found to be associated with PSII, although their exact roles are unknown (Erickson and Rochaix 1991). Ex- pression of the chloroplast genes encoding PSII subunits requires several trans-acting nuclear gene products that act either post-transcriptionally or post-translationally (Jensen et al. 1986; Kuchka et al. 1988, 1989; Rochaix et al. 1989; de Vitry et al. 1989; Sieburth et al. 1991).

Analysis of the mechanism(s) of post-transcriptional regulation has revealed that chloroplast protein-coding transcripts typically contain inverted repeats that can form stem-loop structures. These have been shown in vitro to interact with specific RNA-binding proteins and to affect exonuclease and endonuclease activities (Stern and Gruissem 1987; Stern et al. 1989; Stern and Gru- issem 1989; Adams and Stern 1990; Schuster and Gru- issem 1991 ; Stern et al. 1991). Both the stem-loop structures and the trans-acting factors are required for message processing and stabilization, and they could be in- volved in mediating the differential mRNA stability observed in chloroplasts at different periods of development and under various growth conditions (Mullet and Klein 1987; Gruissem et al. 1988).

We have characterized a Chlamydomonas reinhardtii nuclear mutant that is unable to accumulate transcripts covering not only psbB (encoding P5, the 47 kDa chlorophyll-a apoprotein of PSII) and a co-transcribed open reading frame (ORF), but also sequences further downstream. Northern hybridizations and in vivo pulse-labelling experiments suggest that mRNA degradation and/

or a change in mRNA processing is responsible for the phenotype of the mutant. Taken together these results suggest the existence of a nuclear encoded trans-acting factor responsible for the stabilization and/or processing of transcripts from the psbB region. We also discuss the reliability of pulse-labelling experiments after toluene permeabilization of C. reinhardtii ^cells.

Offprint requests to: J.-D. Rochaix

(3)

Materials and methods

Strains, mutant isolation, and genetic analysis. The C.

reinhardtii wild-type strain 137c, the cell wall-deficient strain cwl 5, and the mutant strains named below, were maintained on TRIS-acetate-phosphate (TAP) agar (Gorman and Levine 1965). The mutant 222E was isolated after mutagenesis of wild-type cells with 5-fluoro- deoxyuridine (Shepherd et al. 1979) and subsequent metronidazole treatment (Schmidt et al. 1977). The PSII deficiency in 222E was detected by screening for high fluorescence (Bennoun and Delepelaire 1982). The double mutant 222E;cw15 was obtained from a cross of 222E(+) with cw15(-). Gametes were obtained by starving cells for nitrogen (Kates and Jones 1964), and tetrad analysis was performed according to standard genetic protocols (Levine and Ebersold 1960; Harris 1989).

Protein isolation, polyacrylamide gel electrophoresis, and western analysis. Thylakoid membrane proteins from wild-type and 222E cells were prepared according to Ro- chaix et al. (1988), with the addition of protease inhibi- tors after sonication (see below for the concentrations used). Gradient (7.5%-15%) polyacrylamide-SDS gels were prepared according to a modified protocol of N.

Chua described in Rochaix et al. (1988). After electrophoresis the proteins were electroblotted onto nitrocellulose membranes and reacted with antisera followed by [125I]-protein A.

Pulse-labelling of proteins with [14C]acetate. Cells were prepared as in Mayfield et al. (1987), except that cyclo- heximide was added to 10 gg/ml final concentration.

[14C]acetic acid, sodium salt (200 gCi/ml) was added to each 30 ml culture to a final concentration of 2 gCi/ml.

After shaking the cultures for 10 rain at room temperature, the cell suspension was adjusted to 50 mM sodium acetate and processed according to a protocol by Ben- noun et al. (1986), with the addition of protease inhibi- tors 0.2 mM phenylmethylsulfonyl fluoride, 1 mM ben- zamidine, and 5 mM e-amino caproic acid. The cellular extract was then fractionated by electrophoresis on a polyacrylamide-SDS gel as described above, which after staining and destaining was treated with Enhance according to the manufacturer's instructions (New Eng- land Nuclear) and subjected to fluorography.

DNA sequencing. Nucleotide sequences were determined using the Sequenase Version 2.0 kit manufactured by United States Biochemical.

Primer extension. Primer extension was carried out according to Ausubel et al. (1990), using as the primer the oligonucleotide 5'-CAGGGTCATTGATTACTAC, complementary to psbB coding nucleotides 49-31.

Northern analysis. Total RNA was prepared according to Rochaix et al. (1988) and separated electrophoretically on denaturing formaldehyde agarose gels according to Goldschmidt-Clermont (1986). The gels were blotted

psbA

EH H H E

A

psbD 1.3 1.7 kb H H psbB

,kb

27"11,1,71071,,I A

H P H P HH P H

psbC atpA rbcL psaB

1.6kb ' 1.4 1.5

E A E E H H H E

Fig. 1. Restriction maps of the subclones digested to generate gene- specific fragments. Each gene, including any introns and 5' and 3' flanking regions, is drawn as a black rectangle, with a black circle indicating the 5' end. Unidentified chloroplast sequences are drawn as horizontal black lines. Vector sequences are drawn as zigzag lines. The restriction enzymes used to digest the subclones are as follows: for the psbA subclone (R14 in pBR328), EcoRI (E) and HindIII (H); for the psbD subclone (R3 in pBR328), HinfI (H); for the psbB subclone (Bal0 in pBR313), HincII (H) and PstI (P); for the psbC subclone (R9 in pBR328), EcoRI (E) and AvaI (A); for the atpA rbcL and psaB subclone (R15 in pUC), EcoRI (E) and HindIII (H)

to Amersham Hybond N + membrane by capillary transfer, and the filters were prehybridized, hybridized and washed, according to Amersham protocols with mi- nor modifications.

S1 analysis. Sl analysis was performed according to Au- subel et al. (1990) with some modifications. The probe was a 977 base PstI-EcoRI fragment 3' end-labelled at the EcoRI site and gel purified.

Pulse-labelling after toluene treatment. Wild-type and 222E cells were permeabilized with toluene and pulse- labelled with [32P]UTP according to Guertin and Belle- mare (1979). The pulse was carried out at 30°C for 5 rain, after which the samples were chilled, spun down, resuspended in 0.6 ml of TENS (200 mM TRIS-HCI, pH 8, 10 mM EDTA, 0.5 M NaC1, 0.2% SDS), 0.6 ml phenol:chloroform:isoamyl alcohol, and 0.6 ml glass beads. Samples were extracted, RNA was ethanol pre- cipitated, resuspended in 100 pl 5 mM EDTA treated with diethyl pyrocarbonate, and added as probe to the hybridization solution. The filters were prepared in the following manner: plasmids shown in Fig. 1 were digested and separated electrophoretically on 0.8% agarose, 1 x TBE (0.09 M TRIS-borate, 0.002 M EDTA) gels. The gels were denatured and neutralized, the blots were transferred by capillary action, and the Hybond N + (Amersham) membranes were fixed in 0.4 M NaOH, prehybridized, hybridized and washed, according to the manufacturer's instructions.

(4)

In vivo pulse-labelling. 100 ml cultures of wild-type ;cwl 5 and 222E;cw15 cells were grown in 30 gM PO4 TAP to approximately 2 x 106 cells/ml, then washed and resuspended in 10 ml POg-free TAP. Subsequently each culture was supplemented with 500 gCi of carrier-free [aZP]orthophosphate, shaken for 10 or 3 rain at room temperature, and chilled on ice. Then 10 ml of cold 1 mM PO~ (regular) TAP was added, and the culture was centrifuged. Each sample was then resuspended in 4 ml of TENS (see above), 4 ml phenol:chloroform:

isoamyl alcohol, and 4 ml glass beads. Samples were extracted, re-extracted with chloroform, ethanol precipi- tated, resuspended in 300 pl 5 mM EDTA, and the RNA used as probe in an identical manner to the toluene- pulse experiment, except that the prehybridization and hybridization solutions contained 1% SDS and 1 x sodium phosphate/pyrophosphate solution to reduce background hybridization (25 x sodium phosphate/pyrophosphate is 160 mM Na2HPO+, 110 mM NaH2PO~H20, and 33 mM Na4PzO7HaO, pH 7.2). To further reduce background hybridization, the hybridization solution containing the probe was hybridized for approximately 1 h to a membrane blotted as usual but containing no DNA fragments, and then the hybridization solution containing the probe was transferred to the membrane bearing the DNA fragments. Washes were the same as for the toluene-pulse experiment except that they contained 1% SDS, another measure to reduce background.

Computer search. The homologies between ORF31 in C. reinhardtii and land plants were identified using the program ideas developed by Minoru Kanehisa.

Results

Isolation of nuclear mutant 222E with a PSII defect Wild-type C. reinhardtii was mutagenized with 5-fluoro- deoxyuridine (Shepherd et al. 1979) and then exposed to metronidazole in the light (Schmidt et al. 1977), a procedure that selects against photosynthetically active cells. Surviving photosynthetic mutants, recovered on acetate-containing media and grown in the dark, were screened for high fluorescence (Bennoun and Delepelaire 1982), and mutant 222E, whose fluorescence induction kinetics indicated a defect in PSII, was isolated. In C.

reinhardtii, individual zygotes from a cross between mat- ing-type plus (rot+) and mating-type minus (mt-) gametes can be subjected to tetrad analysis. Chloroplast mutations are inherited uniparentally from the mating- type plus parent and are usually transmitted to all four spores. Nuclear mutations in either parent are inherited according to Mendelian rules, and segregate 2 wild-type progeny and 2 mutant progeny. After mating 222E to wild type, tetrad analysis demonstrated that the mutation was nuclear in origin: all tetrads, 10 from a 222E(+)xWT(--) cross and 15 from a 222E(-)x WT(+) cross, showed 2: 2 segregation.

co L~ L~J t-- E3 a ('4

~ O4

~" LL LL

Fig. 2. The 47 kDa chlorophyll a-binding protein (P5) is not synthesized in the mutant 222E. Autoradiogram of total proteins fi'om wild type and mutant 222E fractionated by polyacrylamide-SDS gel electrophoresis. Cells were pulse-labelled for 10rain with [14C]acetate. Included for comparison are the mutant FuD34, which does not synthesize the 43 kDa chlorophyll a-binding protein (P6), and the mutant FuD50, which does not synthesize the beta subunit of ATPase, A polypeptide is visible in the FuD50 lane that comigrates with the beta subunit of ATPase

PSII components absent in the mutant

To establish whether synthesis of a PSII polypeptide is altered in the mutant, wild-type and mutant cells were starved for acetate and then pulse-labelled with [~4C]acetate for 10 min. Total proteins were extracted and separated on a polyacrylamide-SDS gel. Figure 2 shows that the P5 protein (the 47 kDa chlorophyll-a apoprotein) is no longer synthesized in the mutant 222E, while the synthesis of other proteins is apparently not affected.

For comparison, mutant FuD34 lacking the P6 protein (the 43 kDa chlorophyll-a apoprotein) and mutant FuD50 lacking the beta subunit of ATPase are shown (Rochaix etal. 1989; Woessner et al. 1984). Previous work with mutants unable to synthesize one of the PSII core polypeptides has shown that the presence of all core PSII polypeptides is necessary for the stable assembly of the PSII complex in the thylakoid membranes (Bennoun et al. 1986; Erickson et al. 1986; Jensen et al.

1986; Kuchka et al. 1988, 1989; Rochaix and Erickson 1988; Rochaix et al. 1989). Consequently, the absence of one PSII core polypeptide in 222E could result in the loss of other PSII polypeptides. In order to determine the level of the various PSII components in the mutant 222E, polypeptides from membrane fractions of mutant and wild-type cells were separated on a 7.5%-15% polyacrylamide-SDS gel, transferred onto nitrocellulose and probed with antibodies against proteins D2, P5 (the 47 kDa chlorophyll-a apoprotein) and P6 (the 43 kDa

(5)

chlorophyll-a apoprotein). The mutant did not accumulate any of these proteins (data not shown). Thus it would appear that the lack of P5 synthesis in 222E results in destabilization of the PSII complex and concomi- tant loss of other PSII polypeptides.

Mapping of the psbB gene

Previous work has indicated that the psbB gene was located in the RI0, Bal0 region of the chloroplast genome of C. reinhardtii (Rochaix 1981). Initial Northern data revealed that the mutant lacked a transcript from this region. Partial sequencing of this region allowed us to localize psbB precisely: its first 122 and last 57 codons were sequenced along with its 5' and 3' flanking regions.

The first 122 codons show 77% conservation and the last 57 codons show 66% conservation at the amino acid level with the spinach psbB sequence (Morris and Herrmann 1984). In liverwort, corn, spinach and tobacco, the psbB coding region is highly conserved, with over 90% homology at the amino acid level (Fukuzawa et al.

1988; Rock eta1. 1987; Morris and Herrmann 1984;

Shinozaki et al. 1986). Since the distance separating the 5' and 3' ends of the psbB coding region in C. reinhardtii is about 1.5 kb, and since in land plants the psbB coding region consists of 1524 uninterrupted coding nucleotides, it is unlikely that the gene contains introns. To date, the only known psbB gene with introns is in Eug- lena (Keller et al. 1989).

The precise location of the 5' end of the psbB mRNA was mapped by primer extension analysis. Figure 3A shows that reverse transcriptase stopped at a T residue, corresponding to an A residue in the noncoding strand, 35 bases upstream of the ATG initiator codon. This result concurs with data from SI nuclease and RNAse protection experiments (not shown). Figure 3 B depicts the sequence of the 251 nucleotides preceding the transcription initiation site. Eleven bases upstream of the transcription initiation site is the sequence TATAAC, which resembles the prokaryotic -10 consensus sequence (TATAAT), though in prokaryotes the T in tire sixth position is highly conserved (Hawley and McClure 1983). Thirty-five bases upstream of the transcription initiation site is the sequence TAGAAA, which resembles the prokaryotic -35 consensus sequence (TTGA- CA). Chloroplast promoters often resemble prokaryotic promoters (Hanley-Bowdoin and Chua 1987) but since Chlamydomonas chloroplast DNA is A+T-rich, sequence similarity to prokaryotic promoter consensus sequences is not necessarily significant. (Our attempts to cap the psbB primary message with radioactively labelled GTP and guanydyltransferase were not successful.)

No putative Shine-Dalgarno sequence complementary to the 3' terminus of C. reinhardtii chloroplast ribo- somal 16S RNA could be discerned. Shine-Dalgarno sequences have been found at the 5' ends of C. reinhardtii genes encoding the large subunit of ribulose bisphosphate carboxylase (rbcL), D1 (psbA), D2 (psbD) and P6 (psbC) (Dronet al. 1982; Erickson et al. 1984, 1986;

Rochaix et al. 1989). In the absence of a recognizable

A

1 2 3 4 5 6 7 8

B

GTTAACATCTGCCAT ATTTTATTTTTCGGT ATGCTCTGACAGGAA TATGAGTATCAAAAG

5q

622 CTGATTGCAATTCCT 527

404 TTGTGGCTAATCAAT TCTTTGGTGAAAATA TATAAATTAAAATCA TGAAAAAAAAACTAT 150 _J AATATAAATTAATT T AATTTAAAATCTTAA AAAATTTTTTTTAAC ATAGTTAATTAAATT

2o 9

TTTT GATTTT rr; rGrlhX£ }}

eeeQoee

ATTTCTTTTTATTTT TATAACCTTGTAATA

90 A

ATTAAGTAAAAAAAT CAGTAAAAAATTTTT 76

67

Fig. 3A, B. Primer extension analysis and sequence of the psbB 5' flanking region. A Primer extension was performed with a radiolabelled 19-base oligonucleotide defined in Materials and methods.

Lanes 1-4, sequencing reactions (A, G, C and T) using the same primer as for the extension; 5, radiolabelled MspI digested pBR322; 6, primer extension control with 25 gg of yeast tRNA;

7, primer extension with 25 pg wild-type total RNA; 8, primer extension with 25 ~tg 222E total RNA. The primer extension product is indicated with an arrow. B DNA sequence of the 286 nucleotides upstream of the psbB initiation codon (boxed). The 5' end of the psbB mRNA is marked by an arrowhead. A sequence resembling the prokaryotic -10 region is underlined, and a sequence resembling the prokaryotic -35 region is underlined with dashes.

A palindromic sequence is indicated by dots

(6)

Shine-Dalgarno sequence, sequences 5' and 3' to the psbB initiation codon may play a role in translation initiation (Bonham-Smith and Bourque 1989).

psbB is co-transcribed with a small open reading frame To determine the 3' end of the psbB transcript, S1 nuclease analysis was performed using as a probe the 977 base PstI-EcoRI fragment containing the last 51 coding bases of psbB. Figure 4 shows three major bands 580, 600 and 660 bases in size, that correspond to 3' flanking regions of approximately 530, 550 and 610 bases. The DNA sequence of the PstI-EcoRI fragment containing the 3' nontranslated region is shown in Fig. 5A, also indicating that the 3' terminus is located near a potential stem-loop structure (stem-loop 1 in Fig. 5A and B) with an estimated free energy value of -46 kcal (Tinoco et al.

1973). Also shown is a smaller potential stem-loop structure (stem-loop 2) with an estimated free energy value of -25 kcal, 86 bases downstream of the stop codon.

In land plants, the psbB mRNA is part of a polycistronic transcription unit that consists of psbB, an ORF (33 to 38 codons), psbH, petB and petD; downstream of the ORF and transcribed from the other strand is psbN. Interestingly in C. reinhardtii, the psbB gene is also co-transcribed with a small similar ORF located upstream of stem-loop 1. Comparison of this ORF in C. reinhardtii and land plants (Fig. 6) illustrates its high degree of conservation. In contrast to land plants, in C. reinhardtii psbN is not found downstream of this ORF, and the petB and petD genes are unlinked to psbB (Harris et al. 1987).

123 Q

910- i,~

659J_ O

655

521 -

403- I

281- I

257- 0 226- o

136- I 100-

WT 222E

L 5 6 7 8 9 10 11

Steady-state RNA accumulation for psbB and flanking regions

To determine the steady-state amount of transcripts that accumulate for psbB and neighbouring regions in the wild type and in the mutant, Northern analysis was performed using as probes various fragments of DNA that comprise a 7.3 kb portion of the chloroplast genome, containing psbB and its flanking regions (Fig. 7). Filters probed with DNA fragments that cover the psbB coding region (probes e and f), show two bands in the wild type: probe f (containing 5' nontranslated and coding sequences) and probe e (containing only coding sequences) detect 2.3 and 1.7 kb bands. Probe d which contains the last 51 coding nucleotides and the 3' flanking region, detects only the 2.3 kb band. The 2.3 and the 1.7 kb bands are absent in the mutant. Two filters probed with DNA sequences directly down- and upstream of psbB (probes c and g) show no difference between the wild type and mutant. Although probe c detects a 1.7 kb band, it is not the 1.7 kb band that is missing in the mutant, but rather the 1.7 kb band detected by probe a and present in both wild type and mutant. Interestingly, a probe corresponding to sequences further downstream of psbB (probe b), hybridizes not only to four bands present in the wild type

Fig. 4. $1 analysis of the psbB 3' flanking region. Lane 1, radiolabelled AluI digested pBR322; 2, the 977 base PstI-EcoRI probe containing the last 51 coding bases of psbB and 3' end-labelled at the EcoRI site; 3, the probe without RNA but SI nuclease digested; 4, control reaction with 50 Ixg yeast tRNA; 5, reaction with 6 gg wild-type RNA; 6, 12.5 gg wild-type RNA; 7, 25 gg wild-type RNA; 8, 50 Ixg wild-type RNA; 9, 12.5 gg 222E RNA;

10, 25 gg 222E RNA; 11, 50 gg 222E RNA

and missing in the mutant (approximately 0.9, 0.8, 0.5 and 0.4 kb in size), but also to a novel 1.1 kb band in the mutant, as well as to a very small RNA (at the bottom of the filter). The 1.1 kb band is also detected with probe a, a 4.4 kb EcoRI fragment that spans probes b, c and d. If one assumes that the psbB mRNA in C. reinhardtii consists of 1524 coding bases as it does in vascular plants, 35 bases corresponding to the 5' nontranslated region and approximately 600 bases corresponding to the 3' flanking region, then the full-length psbB mRNA is about 2.1-2.2 kb in size. The band

(7)

A

.1 psbB. .so

TTCGGTAAATA~AAAAAA~iTGGTGATACiT~TTCTCTAcGTGAAG~ATT~TAAGTTTTTCTTTTCTGGCAGTTGG~AGA FGKYKKLGDTSSLREAF---

• ^.lO0 ^- ^. ^. ^. ^. ¹⁵⁰ ^.

GAATGTGTTTGTCT~TTACCGAGTATAAAACCACTCTG~AAGTACAACTGCCACTGT~GTCGCCTTACGT~GTTTAG~T

.200 2

ACqACGTAAGGCGA~GTC~TTTAGGTATTT;~%~AGTTTTACTTACAGAACiTA~TTATA~A~GGAGGG~TCCACA~CAGCA

2 _.250 _, _, _.300

AAG~TGTA~AAGTTTAAAAAAG~A~TC~C~T~GGGACiGCTTAG~AAAT~AC~TAAT~ACT~TT~AA~C~TTAAC

• . .350 . • ,*,o**,o,*,,,,,**.o,,,,, .400

ACTTTACTATGATTCTTT•TATAGCcATACACTTGAAATTCTT•AAG•ATTTTAAAACA•TTTTAAAAATGTTTTAGTGC

• • .450

AATG~T~A~TT~AAGGTAAT~AATAGAiA~TTTG~iTG~AG~TAAAGTATG~A~ATAA~AAACT~AA~TCAAA

.Boo . ORF 31. .s50

AAAATCA~TACATTTTACTTTTTATGGAAGCTTTAGTATATACTTTCiJATTAGTTGGGACTITAGGTATTATTTTCT MEALVYTFLLVGTLGIIF

.600

T~AATTTi~T~AG~GA~T~A~GTATGATTAAATAA~AG~A~TTAATTAGAA~TTAAA~AA~A~ATTT~AAG~TA F S I F F R D P P R M I K--- I

.650 .700

AGCCTACAGAAGC~TAA~T~CAAATATAT~ATTTAAATTcTATTAAATACTA~CC~TTAAGAGJ~A~%%A~ACTT~A q

I .750 .800

TGGACCTA~AGA~TGGTGG~GCAT~ATG~GAG~AATC~ATGTAGCAAG;~ATGCATAATACCGTI~TAATA~T~TTTA

. 8 5 0

AT~GTAAAGCAAGTAAATT~AAA~AATAiAAAT~TATTiA~AG~AcA~TAA~TTA~CG~TAACGAT~CT~AACAG~C

.900 .950

~AGATGAGAATAACTCAATGTTGTTTAGTACAACGGCAN~AGGTC~AT~tu%GACTCACTT~CTAATTATATATATTT~A~

AAATTTGTTTATCTCTG

B

_T^T^A^G_T

T T

G A G.C T.A 150 - G.C

C.G - 165 640 -

A.T T.A T•A 145 - C.G

C.G - 170 635 -

G.C C.G T.A 140 - G.C

/ \ 630 -

TGCCACTGTG GTCCTTTAGG

stem-loop 2

6 2 5 -

620 -

615 -

/\

C A

A G

T A

C A C.G G.C A.T A.T - 655 T.A T.A G.C A.T A.T - 660 G.C T.A T.A A.T - 665 T.A T.A A.T T.A A.T A.T - 670 T.A A.T A.T A.T T.A - 675 T.A T.A A.T A.T G.C - 680 A.T T.A

/ \

AGCATTTAAT TTAAATACTA

stem-loop I

Fig. 5A, B. Sequence and potential secondary structures of the psbB 3' flanking region. A DNA sequence of the 977 base PstI- EcoRI fragment that contains the last 51 coding bases of psbB and its 3' flanking region. Stop codons are indicated by three dashes. Two potential stem-loops are underlined with arrows. The larger and smaller stem-loops are labelled 1 and 2, respectively.

Upstream of stem-loop 1 is an open reading frame (ORF) with 31 codons. The three 3' termini of the psbB mRNA detected by S1 analysis are at nucleotides 580, 600 and 660 +/-50 nucleotides.

A palindromic sequence is indicated by dots. B Base-pairing of the potential stem-loop structures is indicated. Stem-loops 2 and 1 have an estimated free energy of -26 kcal and of -46 kcal, respectively

liverwort MEALVYTFLLUGTLGIIFFAIFFREPPKVPSKGKK C.reinhardtii :::::::::::::::::::::::::::::::

primrose spinach c o r n rice tobacco wheat barley rye

ORF 35 ORF 31 ::::::::::::::::::::::::::::::::::::: ORF 37 ::::::::::::::::::::::::::::::::: ORF 33 ::::::::::::::::::::::::::::::::: ORF 33 ::::::::::::::::::::::::::::::::::: ORF 35 :::::::::::::::::::::::::::::::::: ORF 34 :::::::::::::::::::::::::::::::::::::: ORF 38 :::::::::::::::::::::::::::::::::::::: ORF 38 :::::::::::::::::::::::::::::::::::::: ORF 38 Fig. 6. Comparison of the ORF found downstream ofpsbB among different species. The amino acid sequences were derived from the nucleotide sequences referenced in the text. Colons (:) indicate the residues that are identical to those of ORF 35 in liverwort

marked 2.3 kb in Fig. 7 could correspond to this full- length psbB mRNA containing the entire 3' flanking region. The band marked 1.7 kb with probes e and f in Fig. 7 could correspond to a psbB mRNA with a shorter 3' flanking region• Recall that a second potential stem- loop structure exists, 86 bases downstream of the stop codon (stem-loop 2), and perhaps this marks a second and minor 3' end of the psbB mRNA. Processing of the full-length transcript may be necessary for the translation of the downstream ORF discussed above. We know from RNAse protection experiments (not shown) with strand-specific probes corresponding to probe f, that the 1.7 kb band is not an anti-strand transcript.

Transcript stability ^ofpsbBdiffers in the mutant

The absence of steady-state psbB mRNA in the mutant could be due to either a transcriptional defect or rapid mRNA degradation. To distinguish between these two possibilities, wild-type and mutant cells were permeabilized with toluene and pulse-labelled for 5 rain with [32p]UTP (Guertin and Bellemare 1979). The radioac- tive RNA was then extracted and used to probe DNA fragments specific for various genes encoding photosynthetic functions: psbA (encoding D1), psbD (encoding D2), psbB (encoding P5), psbC (encoding P6), rbcL ^(encoding the large subunit of ribulose bisphosphate carboxylase), psaB (encoding one of the large subunits of the photosystem I reaction centre) and atpA (encoding the alpha subunit of ATPase). These DNA fragments were generated by digesting subclones containing fragments of C. reinhardtii chloroplast DNA (the restriction maps of these subclones are depicted in Fig. i). The results of the toluene experiment are presented in Fig. 8, where each lane corresponds to one of the subclones digested with the appropriate enzymes. The entire psbB coding region and its 3' flanking region are contained in a 2.7 kb fragment. No difference in transcriptional activity between mutant and wild type for either psbB or any other gene represented on the filters is detectable.

Toluene, however, probably drastically alters the physiological state of the cells, and therefore as an independent approach, in vivo RNA pulse-labelling experiments were performed. To facilitate lysis, cell wall-deficient, double mutant cells (222E ;cwl 5) were used. These

(8)

EcoRI 5

2.3 1.7 0.9 I,I

0.5

I ' _t

I I

ORF psbB 5'

=~==

P EcoRI D H

I o# I 'i' I o# 1

^1.2 ^P^I

wt mutant wt m wt m wt m wt m wt

a b c d e

Fig. 7. Northern analysis of psbB and flanking regions. Filters with wild-type (wt) and mutant 222E (m) total RNA were each hybridized to a different probe, designated by the lowercase letter beneath each filter. The vertical dashed line above each filter indicates the chloroplast DNA fragment that comprises the probe, except for the filter marked a, which was probed with the 4.4 kb EcoRI fragment that includes probes b, c and d. Other restriction enzyme sites indicated are ScaI (S), PstI (P), DdeI (D) and HincII

m wt m

g

(H). Panels b to f show exposures of 6 to 18 h, whereas panels a and g show exposures of 12 to 14 days. On the original autoradiogram corresponding to panel g, very faint 1.9 and 1.5 kb bands are visible in both wild-type and mutant lanes. Indicated at the top of the figure are the psbB coding region drawn as a long solid line, the 3' flanking region drawn as a dashed line, and the ORF within the 3' flanking region drawn as a short solid line. Transcrip- tion is from right to left

WT RNA

<[ O n~ U .-J

222E RNA

C) rn (.9 ..a .(3 .(3 -13 .(3

Q. 13- Q. Q. L

2.7

1.7 1.4 I.I Q9

Fig. 8. In vitro pulse-labelling of toluene-permeabilized cells. Wild- type and 222E cells were treated with toluene and pulse-labelled for 5 min with [32P]UTP. Wild-type RNA and 222E RNA were extracted and used to probe membranes with DNA fragments specific for photosynthetic genes. Lanes are psbA, psbD, psbB, psbC, and rbcL. Each lane contains a fragment specific for the gene after which it is named, except for the rbcL lane which contains in addition psaB and atpA DNA fragments. (Refer to Materials and methods for restriction maps.) Size markers at the right margin of the figure are in kb

cells were pulse-labelled with [32p]phosphate for 10 or 3 min and compared with wild-type;cw15 cells. RNA was extracted and used to probe filters similar to the ones described above. The results of such a 10 min pulse are shown at the top of Fig. 9. The striking difference between the mutant and wild type is that less labelled RNA hybridized to the 2.7 kb psbB fragment in the mutant; otherwise the two filters at the top of Fig. 9 are alike. The weaker signal hybridizing to psbB sequences in the in vivo pulse-labelled mutant cells indicates that either the psbB message is degraded more rapidly, or that transcription of psbB is less efficient in the mutant as compared with the wild type. The results of the 3 rain pulse, shown at the bottom of Fig. 9, are similar to those of the 10 min pulse except that the 2.7 kb psbB band hybridizes more labelled RNA from the mutant, suggesting that less psbB RNA was degraded in 3 than in 10 rain. The filters shown at the top of Fig. 9 represent the most dramatic difference we observed between the 2.7 kb band in the mutant and the wild type; in other 10 rain pulse experiments, the 222E RNA gave a strong- er signal with the 2.7 kbpsbB fragment than in the experiment shown in Fig. 9. This psbB signal with the 222E RNA however, was always weaker than the psbB signal with the wild-type RNA. This differs from the result obtained with toluene-treated cells, where the 222E and wild-type RNA hybridized to similar extents.

One can appreciate additional differences occurring between the hybridization signals of wild-type RNA from toluene-treated cells, and wild-type (cw15) RNA

(9)

WT RNA 222E RNA

.,~ Q m <,9 j '~ D rn U ,.~

..13 ..13 .13 ..13 u ,.13 .13 .13 13 u i/I tO ffl i/'1 ..~ ~ if) 03 q) ,13 13. 0,. (3- O_ ~ 13. 13. Q. 13.

I

--4

4- 27

1.7 1.4 I.I 0.9

i

m

--4

÷ 2.7

-- 12 1.4 --I.I

0.9

psbB

SH P H P HH P H

Fig. 9. In vivo pulse-labelling of cells. Wild-type;cw15 and 222E ;cwl 5 cells were pulse-labelled with [32p]phosphate for 10 rain (top of figure) and 3 rain (bottom of figure). RNA was extracted and used to probe membranes identical to those shown in Fig. 8.

In the top right photograph, the 2.7 kb band is visible upon overex- posure. A restriction map ofpsbB is shown at the bottom; HincII (H), PstI (P), SeaI (S); the ScaI site is included for comparison with Fig. 7 (ScaI was not used in the digest); transcription is from

right to left

from in vivo labelled cells, by comparing Figs. 8 and 9 and referring to the restriction maps in Fig. 1. In the psbA lane, RNA from toluene-treated cells (unlike RNA from in vivo pulse-labelled ceils) hybridizes to a 1.1 kb band containing sequences upstream of psbA. In the psbD lane, RNA from toluene-treated cells (unlike RNA from in vivo pulse-labelled cells) hybridizes better to the 1.7 kb band containing 10% of the psbD coding region and mostly vector sequences, than to the 1.3 kb band containing the 5' nontranslated region and 90% of the psbD coding region. In the psbB lane, RNA from toluene-treated cells hybridizes to sequences up- and downstream ofpsbB, to which RNA from in vivo pulse-

labelled cells either does not hybridize or hybridizes more faintly. In the rbcL lane, RNA from toluene-treated cells hybridizes better to the 1.5 kb band containing the 3' coding sequences of rbcL and psaB, than to the 1.4 kb band containing the 5' nontranslated regions and 5' coding regions of rbcL and atpA; RNA from in vivo pulse-labelled cells hybridizes to both of these bands equally well.

Discussion

Part of the psbB operon in conserved & C. reinhardtii A portion of the psbB 3' flanking region shares consider- able sequence homology with the ORF found downstream ofpsbB in liverwort, primrose, spinach, tobacco, rice, corn, wheat, barley and rye (Fig. 6) (Fukuzawa et al. 1988; Offermann-Steinhard and Herrmann 1990;

Morris and Herrmann 1984; Shinozaki et al. 1986; Hir- atsuka et al. 1989; Rock et al. 1987; Hird et al. 199l;

Reverdatto etal. 1989; Bukharov et al. 1988). So far, this is the only operon of protein coding genes that is conserved in the chloroplast genomes of C. reinhardtii and higher plants. This highly conserved ORF ranges in size from 31 codons in C. reinhardtii to 38 codons in wheat, barley and rye, and practically all amino acid differences occur near the C-terminus. Through evolution a negatively charged glutamic acid as the second residue, a central hydrophobic region, and several charged residues near the C-terminus have been maintained in this ORF. These observations and the high degree of conservation suggest the existence of a small, functional, membrane-spanning polypeptide, whose expression may be co-regulated with that ofpsbB. ^{We note} that this ORF is not found downstream of psbB ^{in the} cyanobacterium Synechocystis 6803 (Vermaas etal.

1987).

Transcription of two closely spaced regions is affected in mutant 222E

Northern data (Fig. 7) show in the 222E mutant the loss ofpsbB transcripts (2.3 and 1.7 kb), the loss of transcripts downstream of psbB (e.g. 0.9 and 0.5 kb), and the appearance of a novel transcript downstream ofpsbB (1.1 kb). Concurrently however, in the 1.6 kb ScaI-PstI region, in between psbB and the chloroplast region encoding the downstream novel and missing transcripts, no significant difference between wild-type and mutant transcript levels is apparent. In agreement with this finding is the result from the in vivo phosphate pulse experiments (Fig. 9), which shows that the hybridization signals for the 1.4 kb HincII-PstI fragment, about 200 bases smaller but otherwise identical to the 1.6 kb ScaI-PstI fragment, are identical in the mutant and wild type, dem- onstrating that transcription in this region is unchanged in the mutant. Further downstream in the 1.9 EcoRI- ScaI region, differences between the wild type and the mutant reappear. That transcripts from two non-contig-

(10)

uous chloroplast regions are affected, suggests either that the nuclear mutation is pleiotropic and affects two independent transcription units, or that the nuclear mutation affects a polycistronic transcription unit, the processing of which produces mature transcripts with different stabilities: the psbB coding mRNA is rapidly degraded, mRNA corresponding to the 1.6 kb ScaI-PstI region appears unchanged, and mRNA corresponding to the 1.9 kb. EcoRI-ScaI region is processed or stabi- lized differently than in the wild type since a novel band larger than the wild-type bands accumulates.

Concurring with the Northern data suggesting that this nuclear mutation affects one or several transcripts specifically, are the results from the in vivo phosphate pulse experiments (Fig. 9) showing that only psbB message stability seems affected: psbA, psbD, psbC, rbcL, atpA and psaB transcript levels are the same in the wild type and the mutant. Analysis of other C. rein- hardtii mutants demonstrates that a nuclear gene product can specifically affect the message stability of a par- ticular chloroplast transcript. Kuchka et al. (1989) characterized a nuclear mutant, nac2-26, that specifically alters the stability of the psbD message. Jensen et al.

(1986) have reported on a nuclear mutant (GE2.10) that specifically affects psbB transcript level, and Sieburth et al. (1991) have further characterized this mutant, the phenotype of which resembles that of our mutant, in that psbB mRNA is unstable and the same transcripts downstream of psbB are missing. In 222E however we note the appearance of a novel transcript downstream of psbB. The two mutations (222E and GE2.10) may be allelic and cause subtly different phenotypes, or they may occur in two different loci and produce overlapping phenotypes; the latter is more plausible since the GE2.10 mutation is reported to affect D1 synthesis which is not the case for 222E. The processing and stabilization of chloroplast transcripts is probably multifactorial, and consequently two different mutations could impede the same process and yield similar phenotypes.

The primary defect of the 222E mutation is probably at the RNA stability level

Two types of pulse-labelling experiments were performed in order to determine whether psbB message degradation or inefficient transcription is the cause of the absence of psbB mRNA in the mutant. Experiments with toluene-treated cells (Fig. 8) show that the mutant trans- cribes psbB as well as the wild type, suggesting that the mutant phenotype is due to RNA instability. However, in the in vivo pulse-labelling experiments (Fig. 9) the signal obtained with the mutant RNA (222E;cw15) for the 2.7 kb band representing psbB is diminished compared with the signal with the wild-type RNA (WT ;cwl 5), suggesting that psbB RNA degradation or inefficient transcription is taking place in the mutant.

A further indication that message stability is in fact affected in this mutant derives from the comparison between the 10 rain and the 3 rain in vivo phosphate pulses (top and bottom of Fig. 9, respectively): in the sample obtained after the shorter labelling period more radioac-

tive mutant RNA hybridizes to the 2.7 kb band, imply- ing less degradation. The results from the experiment with toluene-treated cells (Fig. 8) suggest that psbB is transcribed normally in the mutant, but that RNA degradation or processing can no longer take place after toluene treatment.

Other artefactual results have been observed in C.

reinhardtii cells after toluene treatment. Kuchka et al.

(1989) were unable to measure RNA half-life in toluene- treated cells and believe toluene to impede normal RNA degradation. Sieburth et al. (1991) have reported on the presence of an RNase specific for the 5' region of the psbC transcript, or the absence of factors normally pro- tecting this transcript from 5' digestion, after toluene treatment. Concerning the differences seen between the hybridization signals of wild-type RNA from pulse-labelled toluene-treated cells, and wild-type (cwl 5) RNA from in vivo labelled cells (Figs. 8 and 9), we surmise that either abnormal transcription or loss of normal RNA degradation and/or processing are responsible.

However, an alternative explanation, other than the side-effects of toluene treatment, exists as to why the psbB RNA level appears normal in the mutant in a pulse experiment after toluene treatment, and diminished in the mutant in an in vivo pulse experiment. It is possible that the nuclear factor assumed to be responsible for psbB mRNA stabilization has a mutation in its signal sequence, and consequently the factor is only able to enter the chloroplast and stabilize its target after toluene treatment has permeabilized the chloroplast membrane.

In the absence of toluene, as in an in vivo pulse experiment, the nuclear factor with its defective signal sequence would be unable to gain entry into the chloroplast and consequently the psbB RNA will be degraded more rapidly.

A nuclear encoded factor may stabilize the psbB mRNA by binding to its 3' inverted repeat

The 3' flanking region of psbB (Figs. 4 and 5) contains two potential stem-loop structures. The smaller stem- loop 2 (located 86 bases downstream of the stop codon) has a potential free energy of -25 kcal, and the larger stem-loop i (located near the 3' end of the approximately 600 base long 3' flanking region) has a potential free energy of -46 kcal. Most chloroplast transcription units contain 3' inverted repeats, and these potential stem- loop structures are thought to function in post-transcriptional processes responsible for the differential RNA stability observed in chloroplasts (Deng and Gruissem 1987). It has been shown in vitro that 3' inverted repeats protect RNA molecules against an exonuclease activity (Stern and Gruissem 1987), and that chloroplast mRNAs interact with specific proteins, sometimes in a gene-specific manner (Stern et al. 1989; Nickelsen and Link 1989, 1991). Subsequently it has been demonstrated that an endonuclease activity cleaves the 3' inverted repeat of spinach psbA in vitro (Adams and Stern 1990), and that a spinach 28 kDa RNA-binding protein is necessary for correct in vitro 3' end processing of several

(11)

chloroplast mRNAs (Schuster and Gruissem 1991).

Moreover, it has now been shown in vivo that the C.

reinhardtii atpB mRNA requires its 3' inverted repeat in order to accumulate normally in the cell (Stern et al.

1991). Taken together, these results suggest that nuclear factors bind at and around the 3' inverted repeats of chloroplast mRNAs and stabilize them. As transcription of most chloroplast genes is thought to be constitutive (Gruissem 1989) this provides a post-transcriptional mechanism for the nuclear control of chloroplast gene expression. In the context of this model, we propose that in C. reinhardtii the stability ofpsbB mRNA probably requires at least one nuclear factor. In vivo experiments with chimaeric genes introduced by transforma- tion could help determine whether the large potential stem-loop structure is required for the putative nuclear factors to convey stability onto the psbB mRNA. Thus a component of the multifactorial process that leads to differential mRNA degradation and stabilization could be identified.

The psbB product is required for the assembly of a stable PSII complex

As a result of the nuclear mutation in 222E, the chloroplast psbB mRNA no longer accumulates and consequently the P5 protein is no longer synthesized. This mutation does not affect the transcription or the mRNA stability ofpsbA, psbD or psbC, some of the other genes that constitute the PSII complex, and furthermore, labelling with [14C]acetate shows that only the P5 protein is absent in the mutant (Fig. 2). The lack of P5 however, is sufficient to destabilize the PSII complex, since West- ern analysis demonstrates that P6 and D2, though synthesized, are not stably integrated into the thylakoid membranes (at most 10% of the wild-type level). Our mutant shows that in C. reinhardtii the lack of P5 alone can destabilize the PSII complex. This is in contrast to the cyanobacterium Synechocystis, where interruption of the psbB gene by a kanamycin resistance cartridge pre- vents P5 synthesis, but the absence of P5 does not pre- clude the insertion of P6 into the thylakoid membranes (Vermaas et al. 1986). This suggests that the require- ments for stable assembly of the PSII complex vary among diverse photosynthetic species.

Acknowledgements. We wish to thank C. Alff-Steinberger for help in finding the homologies between ORF 31 in C. reinhardtii and land plants, and Dr. E. Kfis for valuable advice and thoughtful comments. We also thank O. Jenni, F. Bujard-Ebener, and F.

Veuthey for help in preparing figures. This work was supported by the Swiss National Fund for Scientific Research (31-26345.89).

References

Adams CC, Stern DB (1990) Control of mRNA stability in chloroplasts by 3' inverted repeats : effects of stem and loop mutations on degradation of psbA mRNA in vitro. Nucleic Acids Res 18:6003-6010

Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl K (eds) (1990) Current protocols in molecular

biology. Greene Publishing Associates and Wiley-Interscience, New York

Bennoun P, Delepelaire P (1982) Isolation of photosynthesis mutants in Chlamydomonas. In: Edelman M, Hallick RB, Chua NH (eds) Methods in chloroplast molecular biology. Elsevier Biomedical Press, Amsterdam, pp 25-38

Bennoun P, Spierer-Herz M, Erickson J, Girard-Baseou J, Pierre Y, Delosme M, Rochaix J-D (1986) Characterization of photosystem II mutants of Chlamydomonas reinhardtii lacking the psbA gene. Plant Mol Biol 6:151-160

Bonham-Smith PC, Bourque DP (1989) Translation of chloroplast- encoded mRNA: potential initiation and termination signals.

Nucleic Acids Res 17: 2057-2080

Bukharov AA, Kolosov VL, Zolotarev AS (1988) Nucleotide sequence of rye chloroplast DNA fragment encoding psbB and psbH genes. Nucleic Acids Res 16:8737

Deng XW, Gruissem W (1987) Control of plastid gene expression during development: the limited role of transcriptional regulation. Cell 49 : 379-387

de Vitry C, Olive J, Drapier D, Recouvreur M, Wollman FA (1989) Posttranslational events leading to the assembly of photosystem II protein complex: a study using photosynthesis mutants from Chlamydomonas reinhardtii. J Cell Biol 109:991-1006

Dron M, Rahire M, Rochaix J-D 0982) Sequence of the chloroplast DNA region of Chlamydomonas reinhardtii containing the gene of the large subunit of ribulose bisphosphate carboxylase and parts of its flanking genes. J Mol Biol 162:775-793 Erickson JM, Rochaix J-D (1991) The molecular biology of photo-

system II. In: Barber J (ed) Topics in photosynthesis, vol 10.

Elsevier Science Publishers, Amsterdam

Erickson JM, Rahire M, Roehaix J-D (1984) Chlamydomonas rein- hardtii gene for the 32000 mol. wt. protein of photosystem II contains four large introns and is located entirely within the chloroplast inverted repeat. EMBO J 3:2753-2762 Erickson JM, Rahire M, Malnofi P, Girard-Bascou J, Pierre Y,

Bennoun P, Rochaix J-D (1986) Lack of the D2 protein in a Chlamydomonas reinhardtii psbD mutant affects photosystem II stability and D1 expression. EMBO J 5:1745-1754

Fukuzawa H, Kohchi T, Sano T, Shirai H, Umesono K, Inokuchi H, Ozeki H, Ohyama K (1988) Structure and organization of Marchantia polymorpha chloroplast genome III. Gene organization of the large single copy region from rbcL to trnI(CAU).

J Mol Biol 203:333-351

Goldschmidt-Clermont M (1986) The two genes for the small subunit of RuBP carboxylase/oxygenase are closely linked in Chla- mydomonas reinhardtii. Plant Mol Biol 6:13-21

Gorman DS, Levine RP (1965) Cytochrome f and plastocyanin:

their sequence in the photosynthetic electron transport chain of Chlamydomonas reinhardtii. Proc Natl Acad Sci USA 54:1665-1669

Gruissem W (1989) Chloroplast gene expression: how plants turn their plastids on. Cell 56:161-170

Gruissem W, Barkan A, Deng XW, Stern D (1988) Transcriptional and post-transcriptional control of plastid mRNA levels in higher plants. Trends Genet 4:258-263

Guertin M, Bellemare G (1979) Synthesis of chloroplast ribonucleic acid in Chlamydomonas reinhardtii toluene-treated cells. Eur J Biochem 96:125-129

Harris EH (1989) The Chlamydomonas sourcebook. A comprehen- sive guide to biology and laboratory use. Academic Press, San Diego, USA

Harris EH, Boynton JE, Gillham NW (1987) Chlamydomonas rein- hardtii. In: O'Brien SJ (ed) Genetic maps 1987. A compilation of linkage and restriction maps of genetically studied organ- isms, vol 4. Cold Spring Harbor Laboratory Press, New York Hanley-Bowdoin L, Chua NH (1987) Chloroplast promoters.

Trends Biochem Sci 12:67-70

Hawley DK, McClure WR (1983) Compilation and analysis of Escherichia coli promoter DNA sequences. Nucleic Acids Res

11:2237-2255