The Plant Cell, Vol. 3, 1289-1303, December 1991 O 1991 American Society of Plant Physiologists

Expression Dynamics of the Tomato rbcS Gene Family during Development Leslie A. Wanner' and Wilhelm Gruissem2 Department of Plant Biology, University of California, Berkeley, California 94720

The tomato rbcS gene family is composed of five genes (rbcS7, 2, 3 4 38, and 3C)that are differentially expressed during tomato development. Nuclear run-on transcription assays and RNA analysis were used to determine the contribution of transcriptional and post-transcriptional regulation to the accumulation of mRNA from the five rbcS genes in tomato seedlings, leaves, and fruit. We found that the qualitative pattern of mRNA accumulation is regulated at the transcriptional level and that, in general, there is a correlation of rates of rbcS transcription with overall rbcS mRNA abundance in fruit and leaves. Although transcriptional control is a primary determinant for rbcS gene expression in tomato, examination of relative transcription rates and mRNA accumulation of each rbcS gene demonstrated that there is also significant post-transcriptional control of rbcS gene expression during organ development. Individual rbcS mRNAs, which have highly conserved coding sequences and differ only in their 5' and 3' untranslated sequences, have different stabilities. We showed that both transcription and stability of individual rbcS mRNAs are altered in different organs and by the developmental program within these organs as well as by exposure to light. Together, the results provide a comprehensive analysis of the extent of transcriptional and post-transcriptional control that operates within the rbcS gene family during plant development.

INTRODUCTION

Ribulose-l,5-bisphosphatecarboxylase (Rubisco) is a critical enzyme that comprises approximately 50% of the soluble protein in leaves and is present in much lower levels in other photosynthetic organs of the plant (Bonner and Varner, 1976; Ellis, 1979). The enzyme catalyzes the initial fixation of atmospheric carbon dioxide into carbohydrates. Rubisco is composed of eight identical large subunits (LSU) and eight identical small subunits (SSU). It has been of considerable interest to understand how accumulation of each of the subunits is regulated because the LSU and the SSU proteins are products of the chloroplast rbcL gene and the nuclear rbcS genes, respectively, and the subunits are both required (and found) in chloroplasts in stoichiometric amounts. Excess amounts of either SSU or LSU are apparently rapidly degraded (Schmidt and Mishkind, 1983; lnamine et al., 1985). LSU is probably made constitutively in most plants, although it accumulates only when SSU is synthesized to form the holoenzyme. Thus, regulation of formation of this enzyme is dependent on the control of rbcS gene expression (Rodermel et al., 1988). All known plant rbcS genes occur in families of four to 13 genes, but there is no known functional difference

' Current address: Biotechnology Center, 1060 Carmack Road, Ohio State University, Columbus, OH 43210. To whom correspondence should be addressed.

between mRNAs or proteins from each of the rbcS genes within plant species. All genes within a family produce proteins of similar or identical sequence that presumably function with equal efficiency. Many rbcS genes are, however, expressed in qualitatively and quantitatively differing patterns in different organs and at different developmental stages (for review, see Manzara and Gruissem, 1988; Dean et al., 1989). Previous reports have shown that regulation of rbcS gene expression is primarily exerted at the level of transcription (Gallagher and Ellis, 1982; Silverthorne and Tobin, 1984; Berry-Lowe and Meagher, 19851, although recent work has also demonstrated control at the level of mRNA stability (e.g., Shirley and Meagher, 1990; Shirley et al., 1990; Silverthorne and Tobin, 1990). The molecular mechanisms of transcriptional regulation, primarily during light induction, have been dissected for selected rbcS genes, and cis-regulatory elements and corresponding frans-acting proteins that have a role in their transcriptional control are currently under investigation (for review, see Gilmartin et al., 1990). No information is available on cis- or trans-acting components that are critical to the differential stability of rbcS mRNAs. Also, most of the previous studies on the transcriptional and post-transcriptional regulation of rbcS gene expression have either focused on overall rbcS mRNA or protein accumulation without differentiating between members of

1290

The Plant Cell

the gene families or have examined only one or a few of the members in a gene family. We are investigating the molecular mechanisms that control transcription, mRNA accumulation, and protein accumulation of all members in the tomato rbcS gene family to understand differential expression of rbcS genes in plants. The tomato rbcS gene family consists of five members, which are arranged into three loci that are located on two chromosomes (Pichersky et al., 1986; Sugita et al., 1987). Three of the five genes (rbcS3A, rbcS3B, and rbcS3C) are linked within a 1O-kb region and encode identical mature SSU proteins, whereas rbcS7 and rbcS2 are located on different chromosomes and encode SSU proteins that differ from each other and from the locus 3 proteins (Sugita et al., 1987). A preliminary analysis with cDNA probes and gene-specific oligonucleotides showed that the accumulation of the tomato rbcS mRNAs is controlled by unique developmental and organ-specific programs, as well as by light and diurna1 cycles (Piechulla and Gruissem, 1987; Sugita and Gruissem, 1987). For example, rbcS7, 2, and 3A mRNAs accumulate in cotyledons in. the absence of light, whereas accumulation of rbcS3B and 3C mRNAs is light dependent. In contrast, all five mRNAs are found at high but different levels in illuminated leaves, and transfer to darkness significantly decreases the leve1 of rbcS7 mRNA as well as of rbcS3B and 3C mRNAs (Sugita and Gruissem, 1987). In photosynthetically active young fruit, only rbcSl and 2 mRNAs are found, and they decrease to undetectable levels during fruit development (Piechulla et al., 1986; Sugita and Gruissem, 1987). The five mRNAs do not accumulate in roots. Consequently, effectors of tomato rbcS gene expression must integrate different developmental, organ-specific, and environmental signals to establish the distinct mRNA accumulation patterns. We used nuclear run-on transcription assays and steady-state mRNA quantification in combination with single-stranded DNA probes to investigate the transcriptional, and post-transcriptional components of differential rbcS mRNA accumulation in response to the developmental program and light induction in tomato. In this paper, we showed that there is a strict correlation between the transcription of individual rbcS genes and the presence of their cognate mRNAs, indicating that transcriptional regulation acts as a primary control within the gene family. In the accompanying paper (Manzara et al., 1991), we showed that the differences in transcriptional activities during development correlate with changes in DNA-protein interactions in the promoter regions of the five genes. We also demonstrated in the report here that differential rbcS mRNA stability contributes to the quantitative changes in mRNA accumulation that occur during light activation and development. Because the DNA coding sequences of the five tomato rbcS genes are highly conserved, differences in rbcS mRNA stability must be controlled by differences in 5' and/or 3' untranslated regions.

RESULTS

Construction Of rbcs Gene-Specific Probes Gene-specific probes longer than the 20-nucleotide oligomers used previously (Sugita and Gruissem, 1987; Sugita et al., 1987) were constructed to assay rbcS transcript levels at high stringency. Protein-coding and intron DNA sequences of tomato rbcS genes are highly conserved, and the 5' untranslated regions are only 8 to 73 nucleotides in length (Figure 1; Sugita et al., 1987). The DNA sequences in the 3' untranslatedregions of the five tomato rbcS genes are longer and divergent (Sugita et al., 1987), therefore allowing construction of gene-specific probes

-

rbcS 1

ioobp p i y A (t904)

tix (-73)

1

I

I D

r r I

S B

D

N

L

II DY

'+I S

u 164bp

-

rbcS 2

200 bp

IlX

(-281

polyA (+2248)

.1 1 1

u

I D

I

R

J N T D B K

K

u i15bp

i R

J.

- , , , , , , 1

Y

I 1 DJ

PNNT

I 1 SD

I

S'tX

UK B

I

4

I1

I

AR

E

D

H

I

2.7 hb

-

rbcS 3A

1OObp p l y A (+877)

trx (-391

c r u 0.4 hb

14kb

pdyA (+e571

I

,,"

i

,,,,,,,,

J P N

DDU

NT

B

K

RJ M

./1

AI

E

1

i3hb

rbcS 3c

,B

-

lOObp p l y A (+058. 1861)

I l X 1-17]

.1

i

,,,,,, PN

NT

B

RJ D JE

J

2;

A 2 775 bp

Figure 1. Structure of the Five Tomato rbcS Genes. Filled boxes represent exons. Transcription start sites (trx) and polyadenylation sites (polyA) are indicated by arrows, with their positions in nucleotides relative to the translation start site given in parentheses. The gene-specific probes from the 3' untranslated regions constructed for this study are indicated by hatched boxes and are located immediately 3' to the last exon of each gene. Restriction sites are indicated by letters: A, Accl; B, Ball; C, Clal; D, Dral; E, EcoRV; G, Bgll; H, Hindlll; J , Hincll; K, Haelll; L, Hhal; M, Hpal; ,N, Nhel; P, Pvull; Q, Pstl; R, EcoRI; S, Sau3A; T, Sphl; U, Sspl; X, Xbal; Y , Ndel.

Expression Dynamics of the rbcS Gene Family

01 rbcS Probe

(SA)

1

(3.0)

2

(3.9)

3A

(4.9)

3B

(4.4)

3C

(4.2) H9| RNA

3

6

GR

1291

B Oh light 6 h light green

111II __LL1 i nii i I 5

1.25

2

3A

3B

3C

rbcS Gene Figure 2. rbcS mRNA Abundances in Etiolated, Greening, and Fully Green Seedling Cotyledons. (A) Replicate RNA slot blots of 5 ^g of total RNA from seedling cotyledons grown in the dark for 7 days and then exposed to light for 0, 1, 3, or 6 hr, and 1.25 ^g of total RNA from seedlings grown for 7 days in a 14-hr light/10-hr dark cycle (GR) were hybridized with rbcS gene-specific probes. Specific activities (SA x 107 Cerenkov cpm/^g template) of the DNA probes are given next to the rbcS probe designations. (B) Quantitative representation of the data in (A). Individual slots were excised from the hybridized filters in (A), and radioactivity was determined by liquid scintillation counting. The cpm displayed are normalized to 5 ^g of total RNA and are representative of the results obtained in four independent experiments.

from these regions. The polymerase chain reaction (PCR) was used to amplify precise regions of maximum hybridizable length, utilizing sequence information from tomato rbcS genomic clones (Sugita et al., 1987; Wanner, 1990), because the DNA sequences in the 3' untranslated regions are AT rich and provide few restriction sites. Sequences of PCR fragments were determined to confirm their identity. The locations of the probes are shown in Figure 1. Polyadenylation sites for the three genes were mapped. Single sites of polyadenylation occur in rbcS1, 2, 3A, and 3B, and confirm the previously established polyadenylation sites in rbcS1 and 2 (McKnight et al., 1986). The polyadenylation sites in rbcS3A and 36 genes are located 156 and 158 nucleotides 3' to the stop codon, respectively, and are within the cloned PCR DNA fragments used in this study. In rbcS3C, two polyadenylation sites are located 137 and 140 nucleotides 3' to the stop codon and are also within the cloned PCR DNA fragment (Figure 1). Although the lengths of hybridizable sequences in the gene-specific DNA fragments vary between 133 and 164 nucleotides, the GC content and the calculated Tm values are very similar (see Table 3 and Methods). The rbcS gene-specific probes hybridized with approximately equal

efficiencies to the cognate DNA fragments from the five genes and did not cross-hybridize (Wanner, 1990). The rbcS gene-specific probes were used to compare the relative rates of transcription in isolated nuclei with the corresponding levels of rbcS mRNAs detected by RNA blot analysis in response to light and to the developmental program in tomato.

Light Affects rfocS Gene Transcription and mRNA Stability in Developing Seedlings Tomato seedlings grown in the dark for 7 days differ in development from seedlings of the same age grown in light/dark cycles. Etiolated seedlings had 5-cm-long white hypocotyls and light yellow cotyledons that were just emerging from the seed coats. Seedlings grown in 14-hr light/10-hr dark cycles had 2.5- to 3-cm anthocyanincontaining hypocotyls, their seed coats were shed, and they had dark green cotyledons with the first pair of true leaves appearing between them. The rbcS gene-specific probes were hybridized to total RNA from cotyledons of 7-day-old seedlings grown in the dark or grown in light/

1292

A

The Plant Cell

Etiolated Cotyledon

O

-— flH?

etiolated seedling cotyledon

I 100

green seedling

o

FR2+

B

Green Cotyledon

FR2-

rD

1+

1-

rD

2+

2-

rD

3A+ 3B+

3A3B-

3C+

3C-

d> N

phytphyt + 0.

KEY

^^m

^^

«•»

o

V)

§

^

1

2

3A

3B

rtocS Gene Figure 3. Nuclear Run-On Transcription Activities in Etiolated and Green Seedling Cotyledons. (A) and (B) Run-on transcription assays were carried out in nuclei from 7-day-old etiolated and green seedling cotyledons. Radiolabeled transcripts were hybridized to filters containing 2 ^g of single-stranded DNA with sequences either complementary to (+) or the same (-) as mRNA, as indicated in the key to the autoradiograms. Different amounts of single-stranded DNA containing sequences complementary to ribosomal RNA (2, 0.2, and 0.02 /jg) were blotted to determine that hybridization of transcripts is linear until saturation is reached. Abbreviations for gene probes shown in the key are: 1, rbcS1; 2, rbcS2; 3A, rbcS3A; 3B, rbcS3B; 3C, rbcS3C; FR2, a cDNA clone for a proline-rich protein expressed at high levels in immature fruit; phyt, an Arabidopsis phytochrome A clone; rD, the ribosomal DNA probe pHA2.3. Total transcript cpm of 1.1 x 107 and 0.9 x 107 were hybridized in (A) and (B), respectively. (C) Quantitative results are plotted as cpm hybridizing to each rbcS probe per 1 x 107 total transcript cpm, as described in Methods. Columns are averages of four (etiolated) or two (green) run-on assays in independent nuclear preparations. Error bars show the range in results between experiments.

dark cycles. Figure 2 shows that the abundance of the five rbcS mRNAs differed significantly between etiolated and green seedlings. In etiolated seedling cotyledons, rbcS1 and 3A mRNAs provided most of the rbcS mRNA. rbcS2 mRNA was present at low but detectable levels, but mRNAs from rbcS3B and 3C were not detected. In contrast, all five rbcS mRNAs were abundant in light-grown seedlings. Etiolated seedlings that had been exposed to light for 6 hr were darker yellow to yellow-green. The rbcS mRNAs that were not detected in etiolated seedlings were readily detected only 3 hr after light exposure (Figure 2). Although exposure to light enhanced the accumulation of all five rbcS mRNAs, light was required for accumulation of rbcS3B and 3C mRNAs. Nuclei were isolated from the same seedlings used in preparation of RNA shown in Figure 2. Nuclear run-on transcripts from etiolated or green seedlings were hybridized to filter-bound, single-stranded DNA probes for each rbcS gene and to DNA probes for ribosomal RNA, phytochrome (phyA), and a proline-rich protein (FR2; Salts

et al., 1991), as shown in Figure 3. FR2 mRNA accumulates to high levels in young tomato fruit and to lower levels in all other tomato organs (Salts et al., 1991), and transcription of this gene was similar in etiolated and green cotyledons. In addition, there was no significant difference in rRNA transcription between etiolated and green seedlings. Accumulation of phytochrome mRNA is reduced by light, and this decrease is regulated at the level of transcription (Lissemore and Quail, 1988). Transcription of the tomato phytochrome gene that cross-hybridizes to the Arabidopsis phyA DNA probe was, as expected, detectable only in etiolated seedling cotyledons. Nuclear run-on transcripts from all five rbcS genes were readily detectable in green seedlings (Figure 3B). In etiolated seedlings, rbcS1, 2, and 3A were actively transcribed, although transcripts from 3A were near the limits of detection. Transcription was not detected from rbcS3B and 3C, suggesting that the transcription of these two genes in seedlings is light dependent. Single-stranded DNA probes of the same sense as rbcS mRNAs [(-)-strand DNA] were used as controls in

ExpressionDynamics of the rbcS Gene Family

hybridizations with run-on transcripts. As shown in Figures 3A and 36, a low level of hybridization was detected with the (-)-strand DNA for FR2, rbcS3A, and 38. Although these hybridization signals were reproducible in independent experiments, it is not known whether this signifies the presence of stable antisense RNA from these genes. Quantification of the hybridization results (Figure 3C) confirmed that rbcS7, 2, and 3A are transcribed at twofold to eightfold higher levels in green seedling cotyledons than in etiolated seedling cotyledons, demonstrating that light enhances rbcS transcription in tomato seedlings. It is likely that light enhancement of fbcS3A transcription is underestimated because (-)-strand DNA hybridization was subtracted as background in the quantitative data in Figure 3C. Light enhancement of transcription is specific to the rbcS genes because significant differences were not detected in the transcription of the ribosomal RNA genes or FR2. Comparison of the rates of transcription among the tomato rbcS genes with their respective mRNA levels provided evidence for post-transcriptionalcontrol. In etiolated seedlings, rbcS3A mRNA was most abundant, but the rbcS7 gene was more highly transcribed than was rbcS3A (Figures 26 and 3C). Therefore, it appears that there are differences in stability among the three rbcS mRNAs that accumulate in etiolated seedlings; rbcS3A mRNA was more stable in the dark than were either rbcS7 or rbcS2 mRNAs. The rbcS7 mRNA level was at least fivefold higher in light-grown seedlings than in etiolated seedlings, but rbcS7 transcription rates differed by only 2.5-fold. A similar discrepancy between transcriptional activity and mRNA abundance was detected for rbcS2. Thus, light had a stabilizing effect on rbcS7 and rbcS2 mRNAs but not rbcS3A mRNA, suggesting that light-dependent post-transcriptionaldifferences in rbcS mRNA stability play a significant role in defining rbcS mRNA levels during tomato seedling development.

Post-Transcriptional Differences in rbcS mRNA Stability during Leaf Development

The above results raised the possibility that both transcriptional and post-transcriptional control mechanisms coordinate the differential expression of the tomato rbcS gene family at other developmental stages as well. For this reason, we examined the relative rbcS run-on transcription rates and rbcS mRNA levels in tomato leaves at two developmental stages. Immature, light green, incompletely expanded leaves were collected from apical regions of the plant. Fully expanded, green leaflets were collected from mature leaves. RNA and nuclei were prepared from portions of the same collection of leaves. Figure 4A shows that the five rbcS mRNAs were found at similar levels in immature and mature leaves. The quantitative data in Figure 4 6 show that rbcS mRNAs were either similar

1293

(rbcS3C) or reduced in mature leaves as compared with immature leaves. Although the relative rates of run-on transcription of ribosomal DNA and the FR2 gene were similar in mature and immature leaves, as shown in Figure 5A, the relative rates of transcription of the five rbcS genes were twofold to threefold higher in mature leaves than in immature leaves. The discrepancy between increased rates of rbcS transcription in mature leaves and the decreased rbcS mRNA abundance suggests that the overall stabilities of rbcS mRNAs decrease during leaf development. In addition to these general changes, the results shown in Figures 4 and 5 reveal changes in the relative stability of individual rbcS mRNAs during leaf development. The approximate changes in relative rates of rbcS transcription and in rbcS mRNA abundance between immature and mature leaves are summarized in Table 1, together with the calculated relative increase in mRNA turnover required to accommodate the differences in mRNA abundance. From the results in Table 1, it appears that the decay rates forrbcS7 and 3A mRNA increase relative to their transcription rates, thus resulting in a lower “steady-state” level of these mRNAs. The increase in rbcS2, 38, and 3C mRNA decay in mature leaves was paralleled by an increase in rates of transcription; thus, the “steady-state” levels of these mRNAs did not decline significantly during leaf maturation. We conclude that both transcriptional and posttranscriptional mechanisms regulate the differential expression of the members of the rbcS gene family during tomato leaf development.

rbcS Gene Expression in Developing Fruit 1s Controlled by Temporal Changes in Transcriptional Activity and mRNA Stability

Previous results have shown that onlyrbcS7 and 2 mRNAs accumulate early during tomato fruit development and that their mRNA levels decrease rapidly during fruit ripening (Sugita and Gruissem, 1987).To understand the molecular basis for this rbcS gene expression pattern, we determined the transcriptional activity of rbcS genes in developing fruit. Total RNA and nuclei were prepared from young fruit (3 to 5 and 5 to 8 mm diameter)and fully expanded, mature green fruit. Fruit of 3 to 5 mm diameter are approximately 3 to 6 days postpollination and are growing primarily by cell division. Fruit of 5 to 8 mm diameter are approximately 7 to 10 days postpollination. During this time, cell division stops and growth continues by cell expansion until the mature green stage is reached 35 to 40 days postpollination. At the mature green stage, fruit are fully enlarged but have no visible color change resultingfrom the degradation of chlorophyll and the concomitant accumulation of carotenoids, which are characteristic phenotypic markers of the ripening process.

1294

The Plant Cell

B rbcS Probe (SA)

i

m

1

(0.5)

If

2

(1.0)

II

3A

(1.0)

II

3B

(1.5)

II

3C

(1.5)

«•

FJ immature leaf • mature

1200n

3A

3B

3C

rbcS Gene Figure 4. rbcS mRNA Accumulation in Immature and Mature Leaves. (A) Slot blots containing 2 Mg of total RNA from immature (I) and mature (M) leaves were hybridized to each rbcS gene-specific DNA probe. Specific activities (SA x 107 Cerenkov cpm/^g template) in this particular experiment are indicated next to the rbcS probe designations. (B) Average cpm hybridizing in slot blots as shown in (A) and similar sets of slot blots/2 ^g total RNA. Hybridized cpm were corrected for the specific activities of the individual rbcS gene probes. Numbers are the averages of separate hybridizations to three independent RNA preparations, with the range in quantification between blots indicated by the error bars.

Figure 6 shows that rbcS1 and rbcS2 mRNAs can be readily detected throughout development of the fruit until the mature green fruit stage. There was a small increase in the relative levels of both mRNAs from 3 to 5 mm diameter to 5 to 8 mm diameter fruit (Figure 6B), which correlates with an increase in chlorophyll content and photosynthetic activity (Piechulla et al., 1987). The significant reduction of rbcS mRNA levels in mature green fruit (Figure 6B), however, is temporally uncoupled from the loss of chlorophyll and photosynthetic activity during the ripening phase (Piechulla et al., 1987). Of the locus 3 rbcS mRNAs, only rbcS3A was found near the limit of detection in early fruit development. These results are consistent with the failure to detect any SSL) isoform 3 protein in young fruit (D. Somers and W. Gruissem, unpublished data). To determine the extent to which the rbcS expression pattern in fruit was controlled at the transcriptional level, the activity of all five genes was assayed by nuclear runon transcription. Figure 7 shows that only rbcS1 and 2 genes were transcribed at significant levels in developing and mature green fruit and that transcription of locus 3

rbcS genes was at or below the level of detection. The hybridization to the (—)-strand DNA probes for rbcSSA and 3B was pronounced in run-on transcripts from developing fruit. Therefore, the average counts per minute hybridizing to the (-)-DNA strand of rbcS3C were used as background levels in the quantification of transcription of rbcS3A and 36. The quantitative data in Figure 7D confirmed that the transcriptional activity of rbcS1 and 2 increases early in fruit development but decreases sixfold to eightfold by the time fruit reach the mature size. The transcriptional activation/inactivation during fruit development was specific to the rbcS genes and not a general property of developing fruit because other genes (e.g., FR2 and FR5) show a different temporal transcription activity profile. We conclude from these results that the unique rbcS gene expression pattern in fruit is established at the transcriptional level. Comparison of the relative rates of transcription of both rbcS1 and 2 with the accumulation of their respective mRNAs indicated that differences in mRNA stability are superimposed on the transcriptional control in fruit, as in seedling and leaves. The approximately threefold increase

Expression Dynamics of the rbcS Gene Family

in transcriptional activity of rbcS1 and 2 was not accompanied by a corresponding increase in their mRNA levels, indicating that rbcS mRNA turnover rates, similar to leaves, increase during fruit development. In addition, transcriptional activity of rbcS2 was only approximately 1.5-fold higher than transcriptional activity of rbcS1, but rbcS2 mRNA increased to an approximately threefold higher level, indicating that rbcS2 mRNA is more stable than rbcS1 mRNA in developing fruit. Additional evidence for increased nbcS2 mRNA stability was also found in mature green fruits, where rbcS1 and rbcS2 transcriptional activities have declined between sixfold and eightfold, but nbcS2 mRNA levels have only declined approximately twofold.

DISCUSSION

In this report, we showed that the developmental, organspecific, and light-dependent differential expression of the rbcS gene family in tomato is regulated both by overlapping and complex transcriptional and post-transcriptional

A

Immature Leaf

Mature Leaf

Table 1. Estimation of the Change in rbcS mRNA Stabilities during Leaf Development

B

rbcS Gene

1 2 3A 3B 3C

A/B

Increased Turnover trxa Mature RNA Mature Required (Decrease trx Immature RNA Immature in mRNA f. ? ) 2.7 2.3 2.6 4.5 2.3

4.9 X

0.55 1.0 0.67 0.92 1.07

2.3 X

3.9 X 4.9 X 2.1 X

"trx, transcription as measured in nuclear run-on assays.

controls. This is particularly striking because even though they show significant differences in their regulation, the five genes are highly conserved at the DNA and protein levels (Sugita and Gruissem, 1987) and the locus 3 genes are clustered within a 10-kb region. Our results demonstrated that the qualitative pattern of rbcS mRNA

t FR2-

B

1295

FR2+

rD

1+

1-

rD

2+

2-

rD

3A+

3A-

3B+

3B-

3C+

3C-

KEY

3A

rbcS

3B

Gene

Figure 5. Nuclear Run-On Transcription Activities in Immature and Mature Leaves. (A) and (B) Labeled transcripts were hybridized to membranes containing 2 ;/g of single-stranded DNA with sequences either complementary to (+) or the same (-) as mRNA, as indicated in the key. The amounts of single-stranded DNA containing sequences complementary to ribosomal RNA were 2. 0.2. and 0.02 /ug. respectively. Abbreviations for gene probes shown in the key are as in Figure 3 (except that the phytochrome cDNA probe was not included). Total transcript cpm of 4.5 x 107 and 1.75 x 107 were hybridized in (A) and (B), respectively. (C) Quantitative results of run-on transcription assays are plotted as cpm hybridizing to each rbcS probe/107 total transcript cpm, as described in Methods. Columns are averages of two (immature leaf) or three (mature leaf) run-on assays in independent nuclear preparations. Error bars show the range of cpm between experiments.

1296

The Plant Cell

rbcS Probe (SA)

1

(3.0)

2

(3.9)

in co

op w

B M 3-5 mm fruit ^ E3 5-8 mm fruit U mature green fruit/

3A (4.9) 3B (4.4) 3C (4.2) 3A

3B

3C

rbcS Gene Figure 6. rbcS mRNA Accumulation in Developing Fruit. (A) Replicate slot blots containing 20 ^g of RNA from 3- to 5-mm (3-5), 5- to 8-mm (5-8), and mature green (MG) fruit total RNA were hybridized to each rbcS gene-specific probe. Specific activities (SA x 107 Cerenkov cpm//ig template) of the DNA probes are indicated next to the rbcS probe designations. (B) Quantification of rbcS DNA probes hybridizing to RNA slot blots. The columns represent averages of separate hybridizations to two independent RNA preparations, with the range in quantifications between RNA blots shown by error bars.

accumulation in tomato cotyledons, leaf, and fruit is strictly correlated with the transcriptional activity of the corresponding genes. Comparison of relative rates of transcription in seedlings (Figure 3C), leaves (Figure 5C), and fruit (Figure 7D) to rbcS mRNA abundance in these organs (Figures 2B, 4B, and 6B, respectively) showed that there is a general correlation between rbcS transcription rates and mRNA accumulation. Both transcription rates and mRNA accumulation from each rbcS gene are 10-fold to 20-fold higher in leaves than at any stage of fruit development. As would be expected, when an rbcS mRNA is not detected in an organ (e.g., rbcS3B and rbcS3C in fruit and etiolated seedlings), it is not transcribed. Thus, our results confirmed that a primary determinant of rbcS mRNA accumulation is the rate of transcription of each gene. In addition, we found that regulation of temporal and organ-specific differences in rbcS mRNA stability is superimposed on the transcriptional control. TherbcS transcription and mRNA accumulation patterns in tomato differ from those in pea and petunia, for which it has been suggested that rbcS gene expression and chloroplast function are tightly coupled (Fluhr and Chua, 1986; Fluhr et al., 1986). In tomato, rbcS transcriptional activities are primarily controlled by developmental and organ-specific programs. Light conditions (or chloroplast

function) can modulate the quantitative expression patterns of all five genes in cotyledons and leaves, but light (or chloroplasts) does not activate the transcription of locus 3 genes in fruit to any significant extent. Transcription of rbcS1 and 2 is down regulated during fruit maturation, prior to differentiation of chloroplasts into photosynthetically inactive chromoplasts, providing evidence that rbcS transcription can be uncoupled from plastid differentiation. In dark-grown seedling cotyledons, rbcS1, 2, and 3A are transcribed and all three SSU isoforms accumulate (D. Somers and W. Gruissem, unpublished data), although chloroplast development has not been initiated. Only rbcS3B and 3C are strictly light dependent for transcriptional activation and coordination with chloroplast development in cotyledons. We conclude, therefore, that rbcS transcription in tomato can be uncoupled from light induction and chloroplast development/differentiation, as well as the presence of photosynthetically active chloroplasts. Since the first report that light activates transcription of pea rbcS genes (Gallagher and Ellis, 1982), most of the work on regulation of rtbcS gene expression in plants has focused on defining sequence elements and protein factors involved in rbcS gene transcription, c/s-acting sequences mediating light-responsive and organ-specific expression of rbcS genes have been identified in several plants,

Expression Dynamics of the rbcS Gene Family

A 3-5 mm Fruit

t

1297

Mature Green Fruit

B 5-8 mm Fruit FR21+ 2+ 3A+ 3B+ 3C+

FR2+

rD

123A3B3C-

rD rD

FR5+ FR5-

KEY

3A

3B

3C

rbcS Gene

Figure 7. Nuclear Run-On Transcription Activities in Developing Fruit. (A) Run-on transcription assays were carried out in nuclei from 3- to 5-mm green fruit. Labeled transcripts were hybridized to filters containing 2 M9 of single-stranded DNA with sequences either complementary to (+) or the same (-) as mRNA, as indicated in the key. Two, 0.2, and 0.02 /^g of single-stranded DNA containing sequences complementary to ribosomal RNA were loaded on the filter. Abbreviations for gene probes shown in the key are as in Figure 5; FR5, a cDNA expressed at high levels in immature fruit. The run-on transcript cpm were 1.3 x 108 in the hybridization solution. (B) Run-on transcription assays were carried out in nuclei from 5- to 8-mm green fruit. Hybridization conditions and DNA probes on the filter were as described in (A). The run-on transcript cpm were 2.1 x 10s in the hybridization solution. (C) Run-on transcription assays were carried out in nuclei from mature green fruit. Hybridization conditions and DNA probes were as described in (A), except that only 2.0 and 0.02 ^g of single-stranded DNA containing sequences complementary to ribosomal RNA were loaded on the membrane. The run-on transcript cpm were 4.4 x 107 in the hybridization solution. (D) Quantitative results are plotted as cpm hybridizing to each rbcS probe per 1 x 107 total transcript cpm, as described in Methods. The column for 3- to 5-mm fruit nuclei represents the average of three independent nuclear run-on assays. Columns for 5- to 8-mm fruit and mature green fruit are derived from single experiments.

including pea, tobacco, Arabidopsis, and tomato (reviewed in Manzara and Gruissem, 1988; Dean et al., 1989; Gilmartin et al., 1990). In the accompanying report, we showed that differences in the organ-specific rbcS gene transcription pattern can be correlated with specific changes in DNA-protein interactions in the promoter regions of all five genes (Manzara et al., 1991). Most strikingly, we found that in tomato roots and ripe fruit, where we were unable to detect transcription of any of the five rbcS genes (data not shown), DNA-protein interactions do not occur at promoter sites that are generally protected in cotyledons, leaves, or young fruit. The differential transcriptional control is also striking for the tightly linked locus 3 genes in tomato, for which the developmental regulation of rbcS3A transcription in dark-grown cotyledons contrasts with the light-dependent transcriptional regulation of the rbcS3B and 3C genes. It is interesting to note that the rbcS3B and 3C genes share several, but not all, conserved

DNA sequence motifs in their promoter regions with the rbcS3A promoter, which may account for their differential regulation (Manzara et al., 1991). The contribution of post-transcriptional regulation to rbcS gene expression in plants has been less extensively examined. Recent reports have indicated that the expression of individual rbcS genes in soybean, petunia, and Lemna gibba is also regulated by changes in the stability of their corresponding mRNAs (Shirley and Meagher, 1990; Shirley et al., 1990; Silverthorne and Tobin, 1990). Our results demonstrated that the quantitative patterns of rbcS mRNA accumulation in tomato are produced by posttranscriptional regulation of mRNA levels from specific members in the gene family. We found that mRNA stability within the tomato rbcS gene family is altered by organspecific and developmental programs. rbcS3A mRNA is more stable than rbcS1 mRNA in etiolated seedlings. The stabilities of rbcS1 and rbcS2 mRNAs, but not rbcS3A

1298

The Plant Cell

mRNA, are apparently enhanced in green cotyledons as compared with etiolated cotyledons. Differences in stabilities of rbcSl and rbcS2 mRNAs were also observed during fruit development. The transcription rates for both rbcS7 and rbcS2 increased approximately threefold during early fruit development, but there was no corresponding increase in mRNA levels. Furthermore, the stability of rbcS2 mRNA was increased relative to rbcS7 mRNA during fruit maturation, as indicated by the rapid decline in rbcS2 transcription that was not accompanied by a similar rapid decline in rbcS2 mRNA. These results, together with two related reports (Shirleyet al., 1990; Silverthorne and Tobin, 1990), establish that individual members of the rbcS gene family in plants show different patterns of transcriptional and post-transcriptional regulation. Sequences in the 3’ (and 5 ’ ) untranslated regions of the tomato rbcS genes are not highly conserved, in contrast to the high degree of sequence similarity in the rbcS coding regions (Sugita et al., 1987). It is therefore possible that RNA sequences in the 3‘ and 5’ untranslated regions could regulate the quantitative accumulation of the five tomato rbcS mRNAs. A recent study of the role of 3’ untranslated regions showed that 3’ sequences from different genes can alter the expression of a reporter gene by as much as 60-fold (Ingelbrecht et al., 1989), but in that study no distinction was made between effects on transcriptional activity by 3’ enhancer elements and mRNA stability. No information is available on specific cis-elements or trans-acting molecules that regulate cytoplasmic mRNA stability in higher plants. There are no sequence motifs in the 3’ untranslated regions of cytoplasmic rbcS mRNAs that could form stem-loop or other significant secondary structures, comparable with the inverted repeat sequences in the 3‘ untranslated regions of chloroplast rbcL mRNA and other mRNAs. These chloroplast mRNA 3‘ inverted repeat sequences function as cis-regulatory elements for stability (Stern and Gruissem, 1987; Stern et al., 1991) and interact with specific proteins (Stern et al., 1989). The expression of at least one of the 3‘ inverted repeat RNA binding proteins (28RNP) correlates with the accumulation of chloroplast mRNAs (Schuster and Gruissem, 1991). Like most cytoplasmic mRNAs, rbcS mRNAs in plants have a 3’ poly(A) tail. In animals and yeast, there is evidence for a role of the poly(A) tail in mRNA stability and translation (Jackson and Standart, 1990). Poly(A) tail lengths were not examined for the tomato rbcS mRNAs in this study, but based on the discrete size of tomato rbcS mRNAs on RNA blots, we can exclude substantial length heterogeneity. Poly(A) tail lengths were also found to be similar in two soybean rbcS mRNAs that differ significantly in their stability (Shirley et al., 1990). Alternatively, the 5’ untranslated regions of the tomato rbcS mRNAs could be determinants for mRNA stability. We have shown previously that the sequences in the 5‘ untranslated regions differ in size from 8 to 75 nucleotides

and are not conserved. The length of the 5’ untranslated region could influence the rate of translation initiation of rbcS mRNAs, and this in turn may affect their stability. We have found, however, that there is no significant difference in binding of individualrbcS mRNAs to ribosomes in mature leaves (L. Harper and W. Gruissem, unpublished data), suggesting that ribosome binding does not appear to be dependent on the length of the 5’ untranslated regions. Similarly, no differences in ribosome binding were detected between two soybean rbcS mRNAs that differ eightfold in their relative stabilities (Shirley et al., 1990). It is also possible that the rate of intron splicing affects the differential rbcS mRNA accumulation. Intron-exon boundaries are conserved in all tomato rbcS genes, however, and only rbcS2 differs from the other four tomato rbcS genes in intron number and structure. We conclude, therefore, that the rbcS 3‘ untranslated regions are the most likely sites of regulation of the differential rbcS mRNA stabilities in different organs and at different developmental stages. It is not known whether there is any functional difference between the three different Rubisco SSU protein isoforms encoded by the five tomato rbcS genes. The individual patterns of developmental and organ-specific expression suggest, however, that the multiple rbcS genes in tomato do not serve simply to amplify the expression of their abundant and important protein product in photosynthetic organs. Rather, tomato appears to have evolved precise and multiple mechanisms to coordinate the activity of rbcS genes to maximize expression in different light environments and at different stages in development. These mechanisms operate at both the transcriptional and posttranscriptionallevels and are specific for different members within the rbcS gene family.

METHODS Plant Material All plant material was from the VFNT cherry tomato cultivar LA 1221. Mature leaf material consisted of leaflets from fully ex-

panded, dark green leaves of 6-week-old to 5-month-old greenhouse-grown plants. lmmature leaves were collected as entire light green shoot tips of 0.5 to 1.5 cm in length from plants grown either hydroponically for 3 to 5 weeks, in flats of soil in a greenhouse for 3 weeks, or from 6-week-old to 5-month-old greenhouse-grownplants. Supplemental lightingwas provided to greenhouse-grown plants between September and May. To grow seedlings in the dark, seeds were sterilized for 20 min in 5% bleach and rinsed thoroughly in distilled water. They were then spread one layer thick on three layers of presoaked, sterile Whatman 3MM filter paper in aluminum pans, which were covered with aluminum foil. Seedlings were grown for 7 days at 25OC, after which cotyledons were collected into liquid nitrogen under green safelight. For light induction studies, 7-day-old dark-grown seedlings were exposed to a combination of fluorescent/ incandescent light of 250 pE.m-*.sec-’ for 1, 3, or 6 hr, after

Expression Dynamics of the rbcS Gene Family

which cotyledons were collected into liquid nitrogen. For seedlings grown with a light/dark cycle, sterilized seeds were spread thickly on soil and grown in a greenhouse with a 14-hr light/lO-hr dark photoperiod. Dark green cotyledons from 7-day-old seedlings were collected into liquid nitrogen. Tomato fruits from greenhouse-grown plants were separated from sepals and frozen in liquid nitrogen. Frozen fruits were later separated by size using sieves with hole diameters of 3, 5, and 8 mm for immature stages. Fruits of less than 3, 3 to 5, and 5 to 8 mm diameter were approximately 3,5, and 7 days postpollination, respectively. Mature green fruit stages were of the same average diameter (2.5 to 3 cm) as fully mature red fruits but showed no externa1 or interna1 signs of lycopene accumulation. Pollination efficiency in greenhouse-grown plants was increased by gently shaking branches with mature flowers every 3 days; this also served to produce batches of staged fruit. All plant material was harvested between 1o:OO AM and 1:OO PM directly into liquid nitrogen and stored at -8OOC.

Oligonucleotides Twenty nucleotide synthetic DNA oligomers used as primers in the PCR synthesis of gene-specific probes were isolated from a 20% acrylamide/7 M urea/TBE gel (Maniatis et al., 1982), eluted from the gel into 0.3 M sodium acetate, desalted on Bio-Gel P6DG columns (Bio-Rad, Inc.) in water, dried down, and stored dry at -2OOC until needed. The sequences of the oligonucleotides used to make rbcS gene-specific probes are found in Table 2.

DNA Probes The rbcS gene-specific probes used in this study are described in detail in Wanner (1990). Briefly, DNA fragments corresponding to approximately the first 160 nucleotides 3‘ to the stop codon of

Table 2. Primer Oligonucleotide DNAs for Synthesis of rbcS Gene-Specific DNA Probes ~~~~~~~

OEgonucleotidea

Sequence (5’ to 3’)

1 PCR-1

CCAAACATATAATCAAAATG CATATAACCACCGATTCATC TCATATTAGGAACAAAATTG SCACCAAATCATTATCATAT GTTGTAGAAAAACTAATTGC CATTAATATAGTTCTTGTAA AATGCCAATTTTTAATTATGTAA ACAAGTTTCAGAAGCASATT GTGTATATGTCAACAGTSAG CACCAGAAATCATATGTGAA

1 PCR-2

2 PCR-1 2 PCR-2 3A PCR-1 3A PCR-2 38 PCR-1 38 PCR-2 3C PCR-1 3C PCR-2

‘Oligonucleotides designated X PCR-1 and X PCR-2 (X = 7, 2, 3A, 38, 3C) demarcate the sequence used to synthesize genespecific PCR DNA fragments and correspond to the first 20 nucleotides of the 3’ nontranslatedcoding strand (X PCR-1) and the complement to a 20-nucleotide sequence approximately 160 nucleotides downstream of the last codon fX PCR-2).

1299

each rbcS gene were synthesizedusing PCR. The 100-pL reaction mixtures contained 10 mM Tris-HCI, pH 8.3, 50 mM KCI, 5 mM MgCI2, 0.02°/~ gelatin, 0.2 mM each dATP, dCTP, dGTP, and TTP, approximately 200 pmol each of one pair of oligonucleotide primers specific to an rbcS gene (see Table 2), 50 ng of the appropriate rbcS plasmid subclone (or 200 ng of genomic phage clone 417 for rbcS3B), and 3 units of Taq DNA polymerase (Stratagene). Amplification was obtained by 25 cycles of denaturation at 94O C for 2 min, reannealing at 45OC for 2 min, and extension at 72OC for 2 min. Amplified fragments were purified by electrophoresis in a 3% NuSieve agarose (FMC BioProducts)/l % agarose gel in TAE buffer (Maniatis et al., 1982), 5’ phosphorylated, and blunt-endligated into EcoRV cut and dephosphorylated pBluescript KS+ vector (Stratagene). Clones in both orientations were identified by sequencing phagemids containing inserts. The probe for cytoplasmic rRNA was either pHA2, an 8.7-kb subgenomic DNA clone encompassing the entire rDNA repeat from pea (Jorgenson et al., 1982), or a derivative of pHA2 designated pHA2.3, which contains a 1.2-kb BamHl fragment from the 28s region of the rDNA cistron in pHA2 cloned into the BamHl site of the pBluescript KS+ vector. Cloned genes from an early fruit cDNA library (constructed by Dr. Y.Salts in the laboratory of W. Gruissem) were used to verify that fruit nuclei actively transcribe RNA other than ribosomal RNA. The cDNA clones used were FR2+ and FR2-, which are opposite orientations of a 750bp cDNA for a proline-rich protein (Salts et al., 1991), and FR5+ and FR5-, which are opposite orientations of a 400-bp cDNA for an unknown protein (Y.Salts and W. Gruissem, unpublisheddata). Both cDNAs were cloned with EcoRl linkers into the EcoRl site of the pBluescript KS+ vector. Orientation of these pairs of clones was determined by sequencing from both ends. Although both of these cDNA probes showed maximum hybridization to early fruit RNA, neither was completely fruit specific. FR2 hybridizes most highly to a 1.15-kb RNA on blots of RNA from 8- to 10-mm fruit and at 1O/O to 2% of the maximum fruit leve1 to RNA from leaves (Salts et al., 1991). FR5+ hybridizes to a 1-kb RNA, with hybridization most strongly to RNA from 4- to 6-mm fruit RNA, and at lower levels to root and leaf RNA (Y. Salts and W. Gruissem, unpublished data). Two additional DNA clones were obtained by subcloning the approximately 900-bp Pstl fragment from the Arabidopsis type A phytochrome clone pSP18-8 (courtesy of Dr. Peter Quail, UCBUSDA Plant Gene Expression Center, Albany, CA), in both orientations into the Pstl site of pBluescript KS+. Orientation of the clones was determined by use of an asymmetric Hindlll site within the phytochrome insert sequence.

Mapping of PolyadenylationSites The 3’ ends of each rbcS gene were cloned directly from total leaf RNA by cDNA synthesis using the poly(dT)primer 5’CTCGAGAATTCGCGGCCGCT(,,)3’, followed by amplification of the rbcS sequences using PCR with DNA primers corresponding to the first 20 nucleotides of the 3’ untranslated region of each gene (X-PCR 1 in Table 2) and the complement to the unique sequence of the poly(dT) primer below (5’GCGGCCGCGAATTCTCGAG3‘). Products of the PCR reaction were isolated from 3% NuSieve agarose/l% agarose gel, phosphorylated, and ligated into either EcoRV-digested (rbcS7, 2, 3A, and 38) or Smal-digested (rbcS3C), dephosphorylated

1300

The Plant Cell

pBluescript KS+ vector. DNA was prepared from six independent transformants of each rbcS cDNA clone, and several clones from each isolate were sequenced. The mapped polyadenylation sites are shown in Figure 1.

RNA Preparation, Hybridization, and Quantification Total RNA was prepared from frozen plant material by the method of Prescott and Martin (1987) and quantified by measuring absorbance at 260 nm in triplicate dilutions. For RNA slot blots, total RNA was either diluted into 100 pL of 6 x SSC (1 x SSC is 0.3 M NaC1/0.03 M sodium citrate, pH 7.0) or denatured in 50% formamide/5% formaldehyde/6 x SSC before dilution into 6 x SSC, and then applied to Hybond N membranes under weak vacuum. No differences in hybridization efficiency were detected whether or not the RNA was formamide/formaldehyde-denatured prior to blotting. RNA was fixed to nylon membranes using a Stratalinker (Stratagene) on auto-crosslink. Slot blots were prehybridized in 0.1 mL/cm2 of filter area in 5 x SSC/5x Denhardt's (Maniatis et al., 1982)/1% sodium sarcosine/l0 mm sodium phosphate, pH 7.0/0.1 mg/mL denatured herring sperm DNA for 2 hr at 65°C. Prehybridization solution was replaced with 0.5 volume of the same solution containing 0.5 to 5 x 10' Cerenkov cpm radioactively labeled DNA probe prepared from 1 pg of single-stranded DNA templates as described below. Filters were hybridized for 24 hr at 65OC. After removal of probe solution, filters were washed at room temperature for 5 min in 2 x SSC/O.8% sodium sarcosine, then for 30 min at 65°C in 2 x SSC/O.8% sodium sarcosine, and 45 min at 65°C in 1 x SSC/ 0.8% sodium sarcosine. Following autoradiography, individual hybridized slots were excised, and relative levels of hybridization were determined by scintillation counting. rbcS mRNA hybridization signals were normalizedto hybridizationof the random primerlabeled rDNA probe isolated from pHA2.3 to minimize errors in quantification by measurement of absorbance at 260 nm and to account for the much higher contribution of plastid RNA to total RNA in leaves than in other organs.

Preparation of Radioactively Labeled DNA Probes Radioactively labeled DNA probes to specific rbcS genes were prepared from single-stranded pBluescript DNA containing rbcS inserts and the appropriate oligonucleotide primer, shown in Table 3, using standard conditions. lncorporation of w ~ ~ P - ~ Awas TP monitored using DE81 paper (Maniatis et al., 1982). The labeled DNA was digested with Xhol, which cleaves in the polylinker near the site of rbcS sequence insertion. The labeled rbcS DNA region was isolated by gel electrophoresis, precipitatedwith ethanol, and dissolved in hybridization solution.

Preparation of Nuclei Plant material used for nuclear preparations was collected into liquid nitrogen and stored frozen at -80% until needed. Nuclei were prepared as rapidly as possible in a cold room (4"C), and all solutions and filtrates were kept on ice. Two to 10 g of frozen plant material were powdered in a coffee grinder together with small pieces of dry ice. The frozen powder was suspended by stirring in a beaker containing 10 volumes (w/v) 0.32 M sucrose/ 1O mM NaC1/25 mM 1,4-piperazinediethanesuIfonicacid, pH 6.8/ 0.2 mM EDTA/0.15 mM spermine/0.5 mM spermidine/20 mM 2mercaptoethanol/O.l"/O Triton X-1O0 (H buffer) until thawed. The suspension was filtered through several layers of fine mesh cheesecloth and 300, 100, and 52 mm nylon mesh into a beaker containing one-third volume of H buffer made with 2.2 M sucrose, and mixed. Nuclei were pelleted by centrifugation at 50009 for 10 min. The crude nuclear pellet was washed twice by resuspending in H buffer (not including Triton X-100) and pelleting nuclei, first by centrifuging 5 min at 45009, and then at 40009. The nuclear pellet was then resuspended in 5 mL of 50 mM Tris-HCI, pH 8.0/ 5 mM MgCI2/75 mM KC1/0.2 mM EDTA/25% glycerol/2 mM dithiothreitol (NRB) and centrifuged for 2 min at 35009. The final nuclear pellet was resuspended in a minimum volume of NRB (0.2 to 0.8 mL, depending on the weight of initial plant material and the volume of starch in the sample) and either used directly in runon transcription assays or frozen in liquid nitrogen and stored at

Preparation of Single-Stranded DNA M13K07 helper phage was prepared according to Vieira and Messing (1987), except that XL-1 Blue (Stratagene) or JM101 were used as host bacterial strains. For preparation of singlestranded DNA, Bluescript phagemids were transformed into competent TG2 or XL-1 Blue host cells. Fifty milliliters of 2x yeast tryptone or Luria broth containing 50 bg/mL ampicillin and 4 X 1O8 to 4 x 1010 plaque-forming units of M13K07 were inoculated with a single colony of phagemid and incubated at 37°C with aeration for 2 hr. Kanamycin was added to a final concentration of 70 pg/mL, and phagemids were incubated for 18 hr. Singlestranded DNA was prepared from the supernatant of these cultures by PEG precipitation and phenol extraction (Sambrook et al., 1989), and DNA was quantified by measuring absorbance at 260 nm (1 OD260= 40 @g/mLsingle-stranded DNA). The accuracy of quantification and extent of contamination with single-stranded helper phage were monitored by electrophoresis in 1% agarose/ TBE gels.

Table 3. Lengths and Theoretical Melting Temperatures (T,) the rbcS Gene-Specific PCR DNA Fragments 3' End Clonea

Lb

(G + CP/o

TmC

rbcS2 +/rbcS3A rbcS3A rbcS3B +/rbcS3C

162 164 134 157 134 137

29.0 30.5 32.8 28.7 28.3 28.5

77.2 77.8 77.8 76.9 76.0 76.2

+

+,

of

a clone from which the single-stranded DNA produced will be the antisense strand; -, a clone from which the single-stranded DNA produced will be the sense strand. L, number of base pairs in the homologous sequence. T, (in degrees centigrade), 69.3 + 0.41 (G C)% - 650/L. rbcS3C has two polyadenylation sites 3 nucleotides apart; T, for both DNA fragments is shown.

a

+

Expression Dynamics of the rbcS Gene Family

-8OOC. No differences in run-on transcriptional activity were detected between nuclei prepared and used immediately and nuclei stored frozen for as long as severa1 months.

1301

(Gallagher and Ellis, 1982; Marzluff and Huang, 1985; Walling et al., 1986).

Run-On Transcript Hybridizationand Quantification Quantificationof DNA in Nuclei Duplicate 1O-pL aliquots of resuspended nuclei were added to 290 pL of 1O mM Tris-HCI, pH 8.0/1 mM EDTA. A 100-pL aliquot of 0.2 M Tris-HCI, pH 8.0/2 M NaCI/100 mM EDTA/2% sodium sarcosine and 2 pL of 10 mg/mL heat-treated RNase A were added, and nuclei were incubated at room temperature for 60 min. Ten microliters of 1O mg/mL proteinase K were added for a further incubationat 37°C for 1 hr. The digest was then extracted once with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1), and a second time with ch1oroform:isoamyl alcohol (241), and DNA was precipitated by addition of 2 volumes of 95% ethanol. The DNA pellet was washed with 70% ethanol, dried briefly, and dissolved in 0.5mL of 10 mM Tris-HCI, pH 8.0/ 10 mM EDTA. DNA concentrations were calculated from the absorbance measured at 260 nm.

Nuclear Run-On Transcription Assay Typically, the DNA concentration in nuclear preparations was 0.5 to 1.0 mg/mL. To reinitiate transcription, 2 mL of RNasin (Promega, 40 units/mL), and 250 pCi (for leaf) or l mCi (for fruit) a32P-UTP(3000Ci/mmol) dissolved in one-quarter final volume 5 x transcription mix were added to 200 pL of nuclei. The 5 x transcription mix contained 250 mM Tris-HCI, pH 8.0/375mM KC1/25 mM MgCIt/l mM EDTA/5 mM dithiothreitol/5 mM ATP/ 2.5 mM GTP/2.5 mM CTP/O.O5 mM UTP. Transcriptionreactions were incubated at 25OC for 20 min. To terminate the transcription reaction, CaCI2,proteinase K,and RNase-freeDNase I (Bethesda Research Laboratories, 2 pg/250 pL) were added to a final concentration of 2 to 3 mm, 0.02 pg/pL, and 40 ng/mL, respectively, and the mixture was incubated at 37°C for 30 min. An equal volume of 0.1 M Tris-HCI, pH 8.0/0.05 M EDTA/1 M NaCI/ 2% sodium sarcosine, and 0.25 mg/mL proteinase K was added and incubation was continued for 30 min at 37OC. The nuclear lysate was extracted once with an equal volume of pheno1:chloroform:isoamylalcohol (25:24:1) and once with chlorof0rm:isoamyl alcohol(24:l). Two micrograms of RNA as carrier were added, and transcripts were precipitated by addition of 2 volumes of 95% ethanol. The pellet was washed with 70% ethanol and dried. Transcripts, dissolved in 100 mL of 10 mM NaCI/10 mM Tris-HCI, pH 8.0/1 mM EDTA, were chromatographed on a Sephadex G-50 spun column (Maniatis et al., 1982) to remove unincorporated (u-~~P-UTP. lncorporation of U-~~P-UTP into run-on transcripts was estimated from duplicate aliquots of the aqueous phase after phenol/chloroform extraction by binding to DE81 filter discs (Maniatis et al., 1982). Typical incorporation was 2 to 1O x 1O5 cpm/pg of DNA. The rate of UTP incorporation was approximately 0.1 pmol/pg of DNA/min, which is comparable with the rate of 0.1 to 0.2 pmol of GTP incorporated/pg of DNA/min in mammalian cell nuclei reported by Marzluff and Huang (1985). The tomato nuclear run-on transcription system has similar monovalent and divalent cation optima and sensitivity to inhibition by a-amanitin (Wanner, 1990), as previously established run-on transcription systems from many tissues, including other plants

Membrane filters containing rbcS DNA probes were prepared by slot-blotting 2 pg of sense and antisense single-stranded DNA for each rbcS transcript onto Hybond N in a volume of 100 pL of 6 x SSC. Slots were rinsed with an additional 200 pL of 6 x SSC, and DNA was cross-linked using a Stratalinker. Blots were prehybridized in 0.1 mL of hybridization solution/cm2of filter area for 6 hr at 65°C. Hybridization solution was 5 x SSC (Maniatis et al., 1982)/5 x Denhardt's (Maniatis et al., 1982)/1YO sodium sarcosine/lO mm sodium or potassium phosphate, pH 7/0.1 mg/mL sheared, denatured salmon or herring sperm DNA. Prehybridization solution was replaced with hybridization solution containing run-on transcripts, and hybridization was continued at 65°C for 24 hr. Hybridized filters were rinsed in 2 x SSC/0.So/~sodium sarcosine at room temperature for 5 min, followed by washes in 2 x SSC/0.8% sodium sarcosine at 65°C for 30 min, and 1 x SSC/O.~O/O sodium sarcosine at 65OC for 45 min. Quantitative results of run-on transcription assays were obtained by scintillation counting of individual hybridized slots. The hybridization to the (+)DNA strand rbcS probes was corrected for hybridization to the (-)DNA strand rbcS probes, except in fruit, where significant hybridization to the (-)DNA strands of rbcS3A and 38 was seen. Transcription data are expressed as cpm hybridizing to rbcS or other gene probes per 107transcript cpm to adjust for differences in transcript specific activities between experiments. This method of quantifying run-on transcripts resulted in reproducible numbers of transcript cpm from a particular organ. The variability between experiments was generally less than 10%.

ACKNOWLEDGMENTS Synthetic oligonucleotides were kindly provided by Dr. Gerard Zurawski (DNAX Research lnstitute of Molecular and Cellular Biology, Palo Alto, CA) and Dr. N. Kent Peters (Ohio State Biotechnology Center, Columbus).We thank Dr. Thianda Manzara for critical reading of the manuscript. This researchwas supported by National Science Foundation (NSF) Grant DMB-8905202 to W.G. and by the NSF Center for Plant Development,Universityof California, Berkeley.

Received September 26, 1991; accepted October 29, 1991.

REFERENCES

Berry-Lowe, S.L., and Meagher, R.B. (1985). Transcriptional regulation of a gene encoding the small subunit of ribulose-l,5bisphosphate carboxylase in soybean tissue is linked to the phytochrome response. MOI.Cell. Biol. 5, 1910-1917.

1302

The Plant Cell

Bonner, J., and Varner, J.E. (1976).Plant Biochemistry. 3rd ed. (London: Academic Press, Inc.). Dean, C., Pichersky, E., and Dunsmuir, P. (1989).Structure, evolution, and regulation of rbcS genes in higher plants. Annu. Rev. Plant Physiol. Plant MOI.Biol. 40,41 5-439. Ellis, R.J. (1979).The most abundant protein in the world. Trends Biochem. Sci. 11, 241-244. Fluhr, R., and Chua, N.-H. (1986).Developmental regulation of two genes encoding ribulose-l,5-bisphosphate carboxylase small subunit in pea and transgenic petunia plants: Phytochrome response and blue-light induction. Proc. Natl. Acad. Sci. USA 83,2358-2362. Fluhr, R., Moses, P., Morelli, G., Coruzzi, G., and Chua, N.-H. (1986).Expression dynamics of the pea rbcS multigene family and organ distribution of the transcripts. EMBO J. 5,

2063-2071. Gallagher, T.F., and Ellis, R.J. (1 982).Light-stimulatedtranscription of genes for two chloroplast polypeptides in isolated pea leaf nuclei. EMBO J. 2,1493-1498. Gilmartin, P.M., Sarokin, L., Memelink, J., and Chua, N.-H. (1990).Molecular light switches for plant genes. Plant Cell 2,

369-378. Inamine, G., Nash, E., Weissbach, H., and Brot, N. (1985).Light regulation of the synthesis of the large subunit of ribulose-l,5bisphosphate carboxylase in peas: Evidence for translational control. Proc. Natl. Acad. Sci. USA 82,5690-5694. Ingelbrecht, I.L.W., Herman, L.M.F., Dekeyser, R.A., Van Montagu, M.C., and Depicker, A.G. (1989).Different 3’ end regions strongly influence the leve1 of gene expression in plant cells. Plant Cell 1, 671-680. Jackson, R.J., and Standart, N. (1990).Do the poly(A) tail and 3’ untranslated region control mRNA translation? Cell 62,

15-24. Jorgenson, R.A., Cuellar, R.E., and Thompson, W.F. (1 982). Modes and tempos in the evolution of nuclear-encoded ribosoma1 RNA genes in legumes. Carnegie Inst. Wash. Year Book

81,98-101. Lissemore, J.L., and Quail, P.H. (1988). Rapid transcriptional regulation by phytochrome of the genes for phytochrome and chlorophyll a/b-binding protein in Avena safiva. MOI. Cell. Biol.

8,4840-4850. Maniatis, T., Fritsch, E.F., and Sambrook, J. (1 982). Molecular Cloning: A Laboratory Manual (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory). Manzara,T., and Gruissem, W. (1988).Organization and expression of the genes encoding ribulose-l,5-bisphosphatecarboxylase in higher plants. Photosynth. Res. 16,11 7-1 39. Manzara, T., Carrasco, P., and Gruissem, W. (1991).Developmental and organ-specific changes in promoter DNA-protein interactions in the tomato rbcS gene family. Plant Cell 3,

1305-1 31 6. Marzluff, W.F., and Huang, R.C.C. (1985).Transcription of RNA in isolated nuclei. In Transcription and Translation: A Practical Approach, B.D. Hames and S.J. Higgins, eds (Oxford: IRL Press), pp. 89-131. McKnight, T.D., Alexander, D.C., Babcock, M.S., and Simpson, R.B. (1986).Nucleotide sequence and molecular evolution of

two tomato genes encoding the small subunit of ribulose-13bisphosphate carboxylase. Gene 48,23-32. Pichersky, E., Bernatzky, R., Tanksley, S.D., and Cashmore, A.R. (1986).Evidence for selection as a mechanism in the concerted evolution of Lycopersiconesculentum(tomato) genes encoding the small subunit of ribulose-l,5-bisphosphatecarboxylase/oxygenase. Proc. Natl. Acad. Sci. USA 83,

3880-3884. Piechulla, E., and Gruissem, W. (1987).Diurna1mRNA fluctuations of nuclear and plastid genes in developing tomato fruits. EMBO J. 6,3593-3599. Piechulla, B., Pichersky, E., Cashmore, A.R., and Gruissem, W. (1986).Expression of nuclear and plastid genes for photosynthesis-specific proteins during tomato fruit development and ripening. Plant MOI.Biol. 5, 373-384. Piechulla, B., Glick, R.E., Bahl, H., Melis, A., and Gruissem, W. (1987).Changes in photosynthetic capacity and photosynthetic protein pattern during tomato fruit ripening. Plant Physiol. 84,

911-917. Prescott, A., and Martin, C. (1987).A rapid method for the quantitative assessment of levels of specific mRNAs in plants. Plant MOI.Biol. Rep. 4,219-224. Rodermel, S.R., Abbott, M.S., and Bogorad, L. (1988).Nuclearorganelle interactions: Nuclear antisense gene inhibits ribulose bisphosphate carboxylase enzyme levels in transformed tobacco plants. Cell 55, 673-681. Salts, Y., Wachs, R., Gruissem, W., and Barg, R. (1991).Sequence coding for a nove1 proline-rich protein preferentially expressed in young tomato fruit. Plant MOI. Biol. 17,149-150. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989).Molecular Cloning: A Laboratory Manual, 2nd ed. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory). Schmidt, G.W., and Mishkind, M.L. (1983).Rapid degradation of unassembledribulose-l,5-bisphosphatecarboxylase small subunits in chloroplasts. Proc. Natl. Acad. Sci. USA 80,

2632-2636. Schuster, G., and Gruissem, W. (1991).Chloroplast mRNA 3‘ end processing requires a nuclear-encoded RNA-binding protein. EM60 J. 10,1493-1 502. Shirley, B.W., and Meagher, R.B. (1990).A potential role for RNA turnover in the light regulation of plant gene expression: Ribulose-1,dbisphosphate carboxylase small subunit in soybean. Nucl. Acids Res. 18,3377-3385. Shirley, B.W., Ham, D.P., Senecoff, J.F., Berry-Lowe, S.L., Zurfluh, L.L., Shah, D.M., and Meagher, R.B. (1990).Comparison of the expression of two highly homologous members of the soybean ribulose-1,Bbisphosphate carboxylase Small subunit gene family. Plant MOI. Biol. 14,909-925. Silverthorne, J., and Tobin, E.M. (1984).Demonstration of transcriptional regulation of specific genes by phytochrome action. Proc. Natl. Acad. Sci. USA 81, 1 1 12-1 1 1 6. Silverthorne, J., and Tobin, E.M. (1990). Post-transcriptional regulation of organ-specific expression of individual rbcS mRNAs in Lemna gibba. Plant Cell 2, 1 1 81-1 190. Stern, D.B., and Gruissem, W. (1987).Control of plastid gene expression: 3’ inverted repeats act as mRNA processing and stabilizing elements, but do not terminate transcnption. Cell51,

1145-1 157.

Expression Dynamics of the rbcS Gene Family

Stern, D.B., Jones, H., and Gruissem, W. (1989). Function of plastid mRNA 3’ inverted repeats: RNA stabilization and genespecific protein binding. J. Biol. Chem.264, 18742-1 8750. Stern, D.B., Radwanski, E.R., and Kindle, K.L. (1991). A 3’ stem/loop structure of the Chlamydomonas chloroplast atpB gene regulates mRNA accumulation in vivo. Plant Cell 3, 285-297. Sugita, M., and Gruissem, W. (1987). Developmental, organspecific, and light-dependentexpression of the tomato ribulose1,&bisphosphate carboxylase small subunit gene family. Proc. Natl. Acad. Sci. USA 84, 7104-7108. Sugita, M., Manzara, T., Pichersky, E., Cashmore, AR., and Gruissem, W. (1987). Genomic organization, sequence analysis

1303

and expression of all five genes encoding the small subunit of ribulose-l,5-bisposphatecarboxylase/oxygenase from tomato. MOI.Gen. Genet. 209, 247-256.

Vieira, J., and Messing, J. (1987). Production of single-stranded plasmid DNA. Methods Enzymol. 153, 3-1 1. Walling, L., Drews, G.N., and Goldberg, R.B. (1986). Transcriptional and post-transcriptionalregulation of soybean seed protein mRNA levels. Proc. Natl. Acad. Sci. USA 83, 2123-2127. Wanner, L. (1990). Regulation of Tissue-Specific Expression of the Five Genes for the Small Subunit of Ribulose-l,5-Bisphosphate Carboxylase in Tomato, PhD Thesis (Berkeley: University of California).