The Plant Cell, Vol. 3, 1155-1165, November 1991 O 1991 American Society of Plant Physiologists

RESEARCH ARTICLE

Analysis of Rice Actl 5' Region Activity in Transgenic Rice Plants Wanggen Zhang," David McElroy,* and Ray W U " ~ ~ , ' a

Field of Botany, Cornell University, Ithaca, New York 14853 Section of Biochemistry, Molecular, and Cell Biology, Cornell University, Ithaca, New York 14853

The 5' region of the rice actin 1 gene ( A c t l ) has been developed as an efficient regulator of foreign gene expression in transgenic rice plants. To determine the pattern and level of rice Actl 5' region activity, transgenic rice plants containing the Actl 5' region fused to a bacterial 8-glucuronidase (Gus) coding sequence were generated. Two independent clonal lines of transgenic rice plants were analyzed in detail. Quantitative analysis showed that tissue from these transgenic rice plants have a level of GUS protein that represents as much as 3% of total soluble protein. We were able to demonstrate that Actl-Gus gene expression is constitutive throughout the sporophytic and gametophytic tissues of these transgenic rice plants. Plants from one transgenic line were analyzed for the segregation of GUS activity in pollen by in situ histochemicalstaining, and the inheritance and stability of Actl-Gos expression were assayed in subsequently derived progeny plants.

INTRODUCTION

Severa1 groups have reported the introductionand expression of foreign genes in rice (Toriyama et al., 1988; Zhang and Wu, 1988; Zhang et al., 1988; Shimamoto et al., 1989; Hayashimoto et al., 1990). These studies have relied on regulatory elements from the cauliflower mosaic virus (CaMV) 35s gene and/or maize alcohol dehydrogenase 1 gene (Adh 7 ) to direct foreign gene expression. However, it has been reported that these elements have relatively low activity in transformed monocot cells (McElroy et al., 1990b, 1991; Peterhans et al., 1990). Furthermore, in transgenic rice plants it has been shown that the CaMV 35s promoter is not active in all cell types (Battraw and Hall, 1990; Tereda and Shimamoto, 1990) and that the maize Adhl promoter is most active in anaerobically induced root tissue (Zhang and Wu, 1988; Kyozuka et al., 1990, 1991). Plant actin promoters are likely to be active in all tissues because actin is a fundamental component of the plant cell cytoskeleton. Although little has been published concerning plant actin promoters,the activity of a number of animal cytoplasmic actin promoters, including those from Dictyostelium, Drosophila, chicken, mouse, and human, have been well characterized, and the cis- and trans-acting factors responsible for their respective patterns of expression have been delineated both in vitro and in vivo (Fyrberg et

' To whom correspondence should be addressed.

al., 1983; Sanchez et al., 1983; Parker and Topol, 1984; Cohen et al., 1986; Fregien and Davidson, 1986; Nellen et al., 1986; Elder et al., 1988; Frederickson et al., 1989). Many of these animal actin promoters have been found capable of directing high-leve1 foreign gene expression in transgenic animal cells (Guild et al., 1988; Sugiyama et al., 1988; Miyazaki et al., 1989; Liu et al., 1990). It has been reported previously that the 5' region of the rice actin 1 gene (Actl) confers high-leve1expression to a bacterial 0-glucuronidase gene (Gus) in transient assays of transformed rice and maize cells (McElroy et al., 1990b, 1991). Whereas Escherichia coli p-glucuronidase is encoded by the uidA gene, we have denoted this gene as Gus for the purposes of this publication. Although the Actl sequence was previously described as a promoter, this region contains, in addition to 1.64 kb of nontranscribed 5' flanking sequence, severa1Actl transcribed sequences. These sequences include a 5' intron and a portion of the first coding exon of the Actl gene. These sequences have both been shown to contribute to the high level of ActlGus gene expression in transformed rice and maize cells (McElroy et al., 1990b, 1991). Therefore, this Actl sequence will be referred to as the Actl 5' region for the purposes of this report. Studies on rice Actl gene expression to date have either involved the determination of Actl transcript abundance in whole organs (McElroy et al., 199Oa) or the analysis of

1 156

The Plant Cell

Actl 5' region activity in undifferentiatedcells (McElroy et al., 1990b, 1991). To assess the potential of the Actl 5' region for regulating the constitutive expression of foreign genes in whole plants, we set out to determine the level and spatial localization of Actl 5' region activity in transgenic rice. Clonal lines of rice plants expressing an Actl-Gus fusion construct were generated. These were found to have a level of GUS protein that was between 0.5% and 3.2% of total soluble protein (depending upon the tissue sampled). Self- and back-crossing of one clonal transgenic rice line allowed us to determine that an equivalent level of Act 7-Gus gene expression was maintained in subsequent generations. In situ characterization of Act I-Gus gene expression revealed that the Actl 5' region has a virtually constitutive pattern of activity in both the vegetative and reproductive tissues of transgenic rice plants. This ubiquitous pattern of Actl 5' region activity in transgenic rice is quite distinct from that reported for either the CaMV 35s promoter (Battraw and Hall, 1990; Tereda and Shimamoto, 1990) or the maize Adhl promoter (Zhang and Wu, 1988; Kyozuka et ai., 1990, 1991), which each have a more restricted spatial pattern of activity than that observed for the rice Actl 5' region. As such, the rice Actl 5' region represents an efficient constitutive promoter of foreign gene expression in transgenic rice.

RESULTS Regenerationof Rice Plants from Transformed Protoplasts and Screening for Actl-Gus Expression

To investigate rice Actl 5' region activity in vivo, plants were regenerated from rice protoplasts transformed with the Actl-Gus fusion construct pActl-D. The structure, restriction map, and sequence of the Actl 5' transcribed region and Actl-Gus junction region in pActl-D are shown in Figure 1. pActl-D contains 1.64 kb of 5' flanking sequence; a 79-bp, GC-rich, 5' noncoding exon; a 447-bp 5' intron (a corrected version of the intron sequence in McElroy et al., 1990b); and 25 bp from the first coding exon of the Actl gene (including the Actl translation initiation codon), fused to a Gus coding sequence and nopaline synthase (nos)gene 3' noncoding region (Figures 1A and 18). This Actl-Gus fusion encodes a transcript containing two in-frame translation start codons, one each from the Actl and Gus genes, adding 15 amino acids to the N-terminal end of the wild-type GUS protein (Figure 1B), resulting in the production of an ACTl -GUS fusion protein of 72.5 kD in transgenic rice tissue (Y. R. Mawal and R. Wu, unpublished data). For protoplast transformation and plant regeneration, the procedures of Zhang and Wu (1988) were followed.

A

1 kb O

Xh,H,E

Xh E

-s-ykSm

S Xb

Sc

Xh

B TATA BOX +1 - 4 1 ~TACCCCCCCCTCTCCTCCTCCCATCCCCCCAACCCT ACCACCACCACCACCACCACCTCCT

SStI

BStEII

26 C c C C C C T C t C T G C C W i A C ~ C ~ c T C c C C C C T C C C C C T C C G C C t C C t C C G ~ ~ a a c c a c c c c ~ 92 cgtcc~ctcctctttctttCtCCgftfttttt:ttcc~ctc~ctcqatctttggccttg~a~t

159 tqgqggcqaqaqqCqgcttC~cqcccaqatcggtqcgcgggaqqqgcg~atctcgc~ct~~c BamHI EqlII 226 t c q q c ~ q c g g c c ~ t c q c q g g q a a t q g g g c t c t c q g a t ~ ~ t g a t c c g c c ~ t ~ t g q

293 gggaqatqatqqggcgtttaaaatttcqccatgctaaacaagatcaggaagaggqqaaaaqgqcact EqlII

360 atggtttatatttttatatatttctqctgctqctc~caggcttagat~qct~ttctttct 4 2 7 tcttttt~gggtagaatttqaatccctcagcatt~tcatcq~agtttttcttttca:qatt~~,

i?% 494 gacaaatqcagcctcqtgcgqagcttttttgtagl

pa

GTAGRRG Id.t Ala GCT AGAC S p AGCC la G GAG l u ASP GXT

B?.lnH&?2L

561 GGG GGA TCC CCG GGT GGT CAG TCC CTT A I S TTA Gly G l y Ser Pro Gly Gly G l n Ser Phe M t Leu

Figure 1. Structure, Restriction Map, and Partia1 Sequence of the Actl-Gus Fusion Vector pActl-D. (A) Restriction map of pActl-D that contains the 5' region of the rice Actl gene (noncoding and coding portions of Actl exons are depicted as open boxes), the Actl 5' intron (open box with dark arrowhead), the Gus gene coding region (filled box), and the 3' noncoding region of the nos gene (vertically stripped box). The ampicillin resistance gene (AmpR)and a bacterial origin of replication of the pBluescript plasmid (Ori) are indicated (horizontally striped boxes). The region of pActl-D whose sequence is shown in (e) is indicated by a black bar above the restriction map of pActl-D. Restriction sites are abbreviated as follows: B, BamHI; E, EcoRI; H, Hindlll; S, Sstl; Sc, Scal; Sm, Smal; Xb, Xbal; Xh, Xhol. The 1.8-kb BamHI-Sstl Gus restriction fragment used as a probe in DNA gel blot analyses is indicated below the pAct7-D restriction map. (6)Nucleotide sequence of the rice Actl 5' region in pActl-D between the Actl TATA box and Gus translation initiation codon. Nucleotides are numbered with the A of the Actl transcription Restriction sites are indicated and initiation site designated as +l. underlined. The different regions of the Actl 5' transcribed sequences are: upper case italic letters, Actl 5' flanking region; upper case letters, Actl exon sequence; lower case letters, Actl 5' intron sequence; slashed lines, exon-intron splice sites; bold letters, Actl (5' most) and Gus (3' most) translation initiation codons. The translation products of the Actl-Gus junction region are shown below their respective codons.

From 2 x 10' rice protoplasts subjected to the transformation protocol, 1036 plantlets, each from individual calli, were regenerated in two independent transformation experiments. Callus from the base of each individual plantlet shoot was assayed for GUS activity (Zhang and Wu, 1988) by staining with 5-bromo-4-chloro-3-indolyl-/3-~-glucuronic acid (X-gluc) (Jeffersonet al., 1987). A total of 96 individual calli showed blue coloration (indicative of GUS activity)

Actin 5' Region Activity in Transgenic Rice

within 1 to 3 days after staining (see Figure 4A). In a secondary screening step, leaf and root tissue from those individual rice plantlets whose shoot-base callus had been determined to be GUS positive were stained for GUS activity. By visual inspection, it was estimated that 44 of the 96 plantlets assayed exhibited Gus expression in root and/or leaf tissue. From these 44 GUS positive plantlets, two were selected for detailed analysis; clonal plant lines produced from the individual calli that had given rise to these GUS positive plants were named T8-1 and T8-2, respectively. T8-1 was chosen because upon visual inspection, it was found to display a high level of GUS activity, whereas T8-2 displayed an intermediate level of GUS activity that was typical of many plants. Eventually, 20 plants from each line were grown up and transferred to the greenhouse. Compared to untransformed rice plants, regenerated plants (both transgenic and nontransgenic) produced relatively few flowers, and those flowers that were produced showed relatively low seed set and seed viability.

DMA Gel Blot Analysis of Transgenic Rice Plants DNA gel blot analyses of leaf genomic DMA was carried out to estimate the number of pAct1-D copies in the transgenic rice lines T8-1 and T8-2. Gel blots using either undigested genomic DNA or genomic DNA digested with restriction enzymes that have one or no sites within pAct1-D were probed with a radiolabeled Gus coding sequence (Figure 1A) to confirm that the Act 1-Gus sequence had integrated into the genome of the transgenic rice plants. The results of this analysis are shown in Figure 2. Undigested leaf genomic DNA from the two putative transgenic lines showed a high molecular weight hybridizing band when probed with the Gus coding region restriction fragment (Figure 2, lanes 5 and 8, respectively). The undigested hybridizing band from each transgenic line was different in mobility from that of undigested pAct 1-D (Figure 2, lane 1), suggesting that the Act1-Gus construct had integrated into the genomes of the putative transgenic rice plants. When the genomic DNA was digested with Ncol, for which there are no restriction sites within pAct1-D, the T8-1 line appeared to have three hybridizing bands, indicating three integration sites, with the upper band estimated to represent two copies of the Act 1-Gus sequence integrated at one locus (Figure 2, lane 6). The T8-2 line appeared to have four hybridizing bands in such Ncol digestions (Figure 2, lane 9). When digested with Hindlll, which cuts once within pAct1-D (Figure 1A), genomic DNA from the T8-1 line clearly showed four hybridizing bands (Figure 2, lane 7), with one band having an apparent mobility equivalent to that of linearized pAct 1-D (Figure 2, lane 2). The T8-2 line also appeared to show four

Sample Digest

r p/5c/1-D-]

1157

p C -j

|—T8-1 —j

j—T8-2— (

U

H

U

H

U

N

H

U

N

H

1

2

3

4

5

6

7

8

9

10

kb 25.0-

7.4-

1.8Lane

Figure 2. DNA Gel Blot Analysis of Rice Plants Transformed with pAct1-D. Copy number determination of pAct1-D was made using DNA isolated from the leaves of T8-1 and T8-2 transgenic rice plants and from an untransformed control (C) plant, digested with Ncol (N) or Hindlll (H). Digested and undigested (U) leaf genomic DNA was gel blotted and hybridized with a 32P-labeled BamHI-Sstl Gus restriction fragment from pAct1-D. Lanes 1 and 2 contain 5 genomic equivalents of uncut and Hindlll-digested pAct1-D, respectively.

hybridizing bands in such a Hindlll digestion (Figure 2, lane 10). However, the pattern of mobility of the Hindlll hybridizing bands in T8-2 was different from that of both T8-1 and linearized pActl-D. No Gus hybridizing bands were seen in leaf genomic DNA from untransformed control plants either undigested (Figure 2, lane 3) or digested with Hindlll (Figure 2, lane 4).

Quantitative Analysis of GUS Activity in Transgenic Rice Plants Fluorometric and immunoblot analyses were carried out to determine the level of Gus expression at different developmental stages in the two transgenic plant lines. GUS activity in total soluble root or leaf protein was determined fluorometrically by following the conversion of the GUS substrate 4-methylumbelliferyl-/;J-D-glucuronide (MUG) to 4-methylumbelliferone (MU). The results of this analysis are shown in Table 1. The T8-1 and T8-2 transgenic rice lines showed GUS activities that were at least 280- and 340-fold higher in root and leaf tissue, respectively, than that observed for untransformed control plants. For both lines, it was found that GUS specific activity was generally higher in leaf tissue than it was in root material of

1158

The Plant Cell

Table 1. Fluorometric and Immunological Quantification of Gus Expression in Transgenic Rice Plants Transformed with the Act 1 -Gus Fusion Construct pAct1-D

Plant Type8 Untransformed Control

T8-1 Line (R0)

Growth Stage"

Protein Source

NA 4 11 11 NA 3 3 11 11 11 11 11 11 14 14

callus 4th leaf 11th leaf root callus 3rd leaf root 3rd leaf 4th leaf 6th leaf 8th leaf 10th leaf 11th leaf 11th leaf 12th leaf 13th leaf 14th leaf

14 14

T8-1 R, Plants R,-1 R,-2 R,-3 R,-4 R,-5

T8-1 R 2 Plants R2-1 R2-2 R2-3 R2-4 R2-5

T8-2 Line (R0)

GUS Specific Activity (nmol/ mg/min)

Estimated % Total Soluble Protein as GUSC

0.2 0.2 0.3 0.2 413

ND ND 0.0

408 208 131 145 119 124 106 286 134 166 127 263

ND ND 2.7 ND 2.2 ND 1.4 1.3 1.9 ND 1.2 1.4 2.0 3.2

4 4 4 4 4

4th 4th 4th 4th 4th

leaf leaf leaf leaf leaf

134 117 125 110 0.5

ND ND ND ND ND

4 4 4 4 4 8 8

4th 4th 4th 4th 4th 8th root

leaf leaf leaf leaf leaf leaf

148 181 158 2 0.2 104 56

ND ND ND ND ND 0.5 ND

° Ro generation plants were regenerated from protoplasts, R, generation plants are the progeny of the R0 generation, and R 2 generation plants are the progeny of the R! generation. b Rice growth stages are defined in relation to the number of emerged leaves from the shoot, and this definition is not applicable (NA) to rice callus tissue. c The level of GUS protein as a percentage of total soluble protein was calculated from immunoblot experiments using known GUS protein standards. The percentage GUS protein estimates were not determined (ND) for some protein sources.

comparable age. It was also observed that GUS specific activity remained at a relatively high level during all stages of leaf development examined. Total soluble leaf protein from the transgenic plants was subjected to immunoblot hybridization. The results of this analysis are shown in Figure 3. Wild-type control plants (Figure 3, lane 7) or control plants that had been

transformed and regenerated but showed no GUS activity contained no detectable protein that cross-reacted with GUS-specific antiserum (Figure 3, lane 8). Protein from the youngest leaf of a T8-2 plant at the eight-leaf stage (Figure 3, lane 9) showed a band that reacted with GUS-specific antiserum to a level equivalent to 0.2 ^g of standard GUS protein, as determined by densitometry scanning. As this result came from a sample containing 40 ^g of protein, we estimate GUS to represent 0.5% of total soluble protein in this tissue. Similar calculations were made for many of the leaf protein samples examined, and the results are shown in Table 1. The results were consistent with those determined by fluorometric assay and indicate that leaves from these transgenic rice plants produce high levels of GUS enzyme, ranging from 0.5% to 3.2% of total soluble protein, depending upon the plant and tissue examined.

Histochemical Localization ol Act1 5' Region Activity in Transgenic Rice Plants To determine the cellular pattern of rice Act1 5' region activity, T8-1 and T8-2 transgenic rice plants, as well as RT generation progeny plants from the T8-1 line, were sectioned and subjected to histochemical staining (Jefferson et al., 1986) to localize Act1 -Gus gene expression. The results of this histochemical analysis are shown

Plant Sampled

CU CTT8-2 |————————T8 1————————.

Growth Stage

11

8

8

3

11

11

It

11 14 14

14

Leal Sampled

11

8

8

3

3

6

8

10

13

14

7

8

9

10

11

12

13

14 15

17

18

Lane

1

2

5

6

11 12 16

14

97 —

Figure 3. Immunoblot Analysis of Rice Plants Transformed with the/4cf7-Gus Fusion Construct pAct1-D. Immunoblot analysis was conducted using GUS-specific antiserum and soluble protein extracts from control plants that were either untransformed (CU, lane 7), transformed but showing no histochemically detectable GUS activity (CT, lane 8), or from transgenic rice plants containing the Act1-Gus fusion construct pActl-D (lanes 9 to 18). Standard protein preparations were: lane 1, prestained protein molecular mass standards; lanes 2 to 6, 2.5 ng, 1.0 Mg, 0.5 ^g, 0.25 ^g, and 0.05 ^g, respectively, of purified GUS protein; lanes 9 to 18, plant protein samples with plant name, growth stage, and leaf assayed, as indicated above each lane. Forty micrograms of total soluble protein were loaded into each lane. Molecular mass marker units in kD are shown to the left of the immunoblot.

Actin 5’ RegionActivity in Transgenic Rice

in Figure 4. In general, it was found that vegetative material from all transgenic rice lines stained rapidly and displayed the same pattern of Gus expression, whereas material from untransformed control plants displayed no staining. Five minutes of X-gluc staining was sufficient to detect blue coloration in root hair cells (Figure 46). Although we observed strong GUS staining in the root tip and zone of root elongation (data not shown), we found the most intense GUS staining around the site of secondary root formation (Figures 4C and 4D). Staining of transverse root sections revealed strong GUS activity in the cells of the phloem and epidermis, whereas weaker staining was observed in the cells of the root cortex. In contrast, the cells of the root xylem, endodermis, and exodermis showed relatively little GUS staining (Figure 4E). When transverse sections of stem material were stained for GUS activity, little coloration was observed in cortical cells, whereas intense staining was seen in the cells of the vascular phloem and in adventitious root primordia (Figures 4F and 4G). Only brief staining with X-gluc was required to obtain a strong staining reaction in leaf trichomes (Figure 4H) and leaf epidermal guard cells, whereas weaker reactivity was observed in unspecialized leaf epidermal cells (Figures 41 and 4J). Longer incubation of leaf tissue revealed GUS staining in phloem tissue, bulliform cells, and leaf mesophyll cells. However, little or no staining was observed in the xylem cells of the leaf vascular tissue (Figure 4K). Shoot apex longitudinal sections from plants at the eight-leaf stage revealed that the shoot meristem, leaf intercalary meristem, and axillary bud all show strong staining, whereas pith cells show little GUS activity (Figures 4L and

4M). Actl-Gus expression was analyzed during panicle development in T8-1 plants. GUS activity was detected at all developmental stages and in all tissues, including the ovary, stigma, stamen (Figure 4N), and pollen (Figure 40), during panicle floret formation. GUS staining was observed throughout R1 generation transgenic rice seeds, including the embryo and endosperm (Figures 4P and 4Q).

1 159

Analysis of Act7-Gus Gene Expression in R, and R2 Generation Progeny Plants Individual plants from the T8-1 transgenic line were grown in the greenhouse and plants that flowered were used for back-crossingto untransformed cv Taipei 309 rice plants. In most cases, seed set was relatively low, endosperm failed to develop normally, and embryo rescue on N6 basic media was employed to ensure seed survival and progeny plant production. To determine if Actl-Gus genes can be transmitted to the subsequent generation, we conducted gel blot analysis using DNA from the leaves of a T8-1 progeny plant, denoted as T8-1, R,-1 (or T8-1-1) that showed high GUS activity (Table 1). The results of this analysis are shown in Figure 5. It appears from Hindlll digests of genomic DNA that the T8-1-1 plant (Figure 5, lane 8) had inherited all four integrated copies of the Actl-Gus sequences that were present in the progenitor T8-1 lhe (Figure 5, lane 5). To determine if the high level of Gus expression was stably maintained in the subsequent genehtions, quantitative fluorometric assays were carried out on leaf material from individual progeny plants. The results of this analysis are shown in Table 1. This analysis revealed that most of the progeny plants expressed a level of GUS specific activity that was equal to that of the T8-1 parental lhe. The low level of GUS activity observed in some individual progeny plants is presumed to be due to a failure to obtain any functional Actl-Gus genes after back-crossing to the untransformed parental line. Morphologically, the R1 generation progeny plants showed no obvious difference from control plants in their vegetative growth and development and did not display any of the abnormalities observed in the Ro generation. At their reproductive stage of development, they displayed the same low seed set and low seed viability that had been found in the Ro generation transgenic rice plants, and embryo rescue was again employed to produce R2 generation transgenic rice plants. However, R2 generation seeds produced by back-crossing to untransformed rice plants developed and germinated normally (data not shown).

Segregation Analysis of Act 7-Gus Gene Expression in Transgenic Rice Plant Pollen

DlSCUSSlON

The pollen of the T8-1 plants was collected at different stages of floret development and stained for GUS activity. Mature pollen from such plants showed differential staining, with Gus-expressing pollen showing intense coloration within 10 min after staining (Figure 40). To characterize the segregation of the Actl-Gus-containing loci in the T8-1 line, over 1O00 pollen grains were histochemically stained for GUS activity. Statistical analysis of their blue to white staining ratio, shown in Table 2, revealed that their distribution followed a 7:l pattern.

We have described the regeneration of rice plants from protoplasts transformed with the Actl-Gus fusion construct pActl-D (Figure 1). Histochemical screening for GUS activity showed that approximately 4% (44/1036) of these regenerated rice plants displayed Actl-Gus gene expression at the seedling stage of development. Visual inspection of the growth of regenerated rice plants, both transgenic and nontransgenic, indicated that they displayed morphological abnormalities. It has been well documented

scu

Figure 4. Histochemical Localization of Act1-Gus Expression in Transgenic Rice Plants.

SA

Actin 5' Region Activity in Transgenic Rice

Table 2. Determination of Segregation Ratio of Gus-Expressing Loci in Pollen from T8-1 Transgenic Rice by xZ Analysis' No. Pollen No. GUS No. GUS x 2 Stainedb Positive Negative (7:l) Observed Expected

1049 1049

916 917.25

133 131.25

ProbabilityC

0.015 0.9 0.000 1.0

'The x 2 test assesses whether a series of observed frequencies deviates from what a theory predicts to an extent that renders the theory implausible. Mature pollen were isolated from anthers on the same day as flower opening, and GUS activity in individual pollen was determined 10 min after staining with X-gluc. Probability that the sample of pollen investigated comes from a population that has the predicted 7:l ratio of GUS positive:GUS negative pollen.

that tissue culture can cause a wide variety of genetic changes that result in altered phenotypes in plants derived from protoplasts (Larkin et al., 1989). Such abnormal phenotypes have been reported in transgenic rice plants transformed with a hygromycin (antibiotic) resistance gene (Hayashimoto et al., 1990) and transgenic maize plants transformed with a phosphinothricin (herbicide) resistance gene (Gordon-Kamm et al., 1990). lntegration of pAct7-D sequences in the transgenic rice lines was assayed by digesting genomic DNA with restriction enzymes that have either one (Hindlll) or no (Ncol) restriction sites within the sequence of pAct7-D (Figure 2). To produce the observed linear restriction fragments in the DNA gel blot, these enzymes would have to be cut in the rice genomic DNA flanking the integrated A c t l -Gus

1161

construct. The observed hybridization pattern of integrated pAct7-D is different from that observed for nonintegrated pAct7-D plasmid. From such DNA gel blots, it was estimated that the transgenic rice plant lines T8-1 and T8-2 each contain four integrated copies of the pAct7-D construct, although in the case of the T8-1 line integration of two copies of the pAct7-D sequence may have been at a single locus. The T8-1 parental line is believed to have three independently segregating Act l-Gus-containing loci based on analysis of DNA gel blots (Figure 2) and pollen segregation ratios (Table 2). Therefore, our finding that the T8-1 progeny plant that we chose to analyze has the same pattern of restriction fragments as its T8-1 parental line (Figure 5) has the probability of 1 out of 8. Preliminary analysis of progeny produced from these transgenic rice lines revealed that their Act7-Gus sequences are transmitted to subsequent R I and R2 generations. A full analysis of the segregation of integrated Actl-Gus genes will have to await the production of larger numbers of R, generation

plants. Quantitative fluorometric analysis of GUS specific activity in leaf and root material from the transgenic rice plants revealed that Actl-Gus expression was constitutive at the tissue level. The range of GUS specific activity in the transgenic rice lines T8-1 and T8-2 was 56 to 208 nmol/ mg root protein/min and 104 to 408 nmol/mg leaf proteinl min. Furthermore, this level of GUS specific activity was stably maintained in progeny plants produced from the T8-1 transgenic rice line. lmmunoblot analysis of Gus expression showed that the level of GUS protein in the tissue of plants from the T8-1 transgenic rice line can be as high as 3.2% of total soluble protein. This suggests that the accumulation of protein

Figure 4. (continued). Handcut thin sections or whole tissue preparations were stained for GUS activity for 5 to 30 min (except where stated). (A) Embryogenic callus with emerging shoot (S), and root (R), T8-1, 2-hr staining, bright field, magnification x15. (e) Root hair cells, T8-1-1 (or Rl-l), bright field, magnification X16. (C) and (D) Longitudinal sections of lateral root initiation regions, T8-1, bright field, magnification ~ 6 6 SR, . secondary root; SRC, secondary root conjunction; VB, vascular bundle. (E) Primary root transverse section, R1-l, bright field, magnification x132. C, cortical cells; EP, epidermal cells; EX, exodermal cells; P, phloem; X, xylem. (F) Stem transverse section, T8-1, dark field, magnification X66. ARP, adventitious root primordium; C, cortical cells; EP, epidermal cells; VB, vascular bundle. (G) Stem transverse section, T8-2, 24-hr staining; dark field, magnification ~ 1 3 2C, . cortical cells; P, phloem; X, xylem. (H) Leaf surface trichomes (T), T8-1, bright field, magnification x33. (I)Leaf epidermal guard cells (G), T8-2, bright field, magnification x66. (J) Leaf midrib transverse section, T8-1, dark field, magnification x66. G, guard cell; MR, leaf midrib. (K) Leaf blade transverse section, Rl-l, bright field, magnification X66. BC, bulliform cells; BS, bundle sheath cells; MC, mesophyll cells; P, phloem; SC, sclerenchyma cells; X, xylem. (L) Leaf-shootjunction longitudinal section, T8-2, dark field, magnification x66. AB, axillary bud; SL, subtending leaf. (M) Shoot apex longitudinal section, T8-1, dark field, magnification x33. LIM, leaf intercalary meristem; PT, pith cells. (N) A mature floret, R1-l, bright field, magnification x3. AN, anther; FI, filament; LE, lemma; OV, ovary; ST, stigma. (O) Mature pollen, bright field, magnification x33. (P) Seed section, R1-l, bright field, magnificationx5. EM, embryo; EN, endosperm cells. (a)Embryo section, R,-1, bright field, magnificationx12. CO, coleoptile; RA, radicle; SA, shoot apex; SCU, scutellum.

1162

The Plant Cell

Sample Digest

pAcM-D

C

|—T8-1 —|

p T8-1-1 -|

H

H

U

U

BS

H

BS

Lane

BS

8

9

t

kb

7.4-

H

Mi:

1

2

3

4

5

6

7

Figure S. DNA Gel Blot Analysis of an R, Generation Transgenic Rice Plant. Copy number determination of pAct1-D was made using DNA isolated from the leaves of T8-1 and T8-1-1 (or R,-1) transgenic rice plants, and from untransformed control (C) plants, digested with Hindlll (H) or BamHI and Sstl (BS). Digested and undigested (U) leaf genomic DNA was electrophoresed, blotted, and hybridized with a 32P-labeled restriction fragment containing the Gus coding region. Lanes 1 and 2 contain 5 genomic equivalents of Hindlll- or BamHI- and Sstl-cut pAct1-D. Molecular size markers in kb are shown to the left of the autoradiogram.

from agronomically important genes, whose expression is under the control of the Act1 5' region, could potentially represent a significant fraction of total soluble protein. In transgenic tobacco, direct correlations have been reported between the level of an insect resistance gene product and the degree of tolerance to insect attack (Milder et al., 1987). Use of the/4cf7 5' region should prove beneficial in similar types of experiments involving transgenic rice material. In situ histochemical localization of Act1-Gus gene expression in transgenic rice plants revealed that the Act 1 5' region is active in most sporophytic cell types, as well as in gametophytic pollen tissue. This staining pattern probably reflects a ubiquitous requirement for cytoskeletal components (of which actin is one) in all plant cell types (Staiger and Schliwa, 1987; Seagull, 1989). The most intense GUS staining was found in those cell types that are known to be either cytoplasmically dense, metabolically active, or undergoing rapid division and elongation. However, although fluorometric and immunoblot analysis allowed us to quantify the organ-specific level of Gus expression in the transgenic rice plants, a similar qualitative comparison at the cellular level using histochemical staining is not feasible. This is due to the indeterminable effects

of cell size, cytoplasmic density, and metabolic activity, as well as GUS substrate penetrability, on the observed differential patterns of histochemical stain intensity (Jefferson et al., 1987). Therefore, differences between cell types in their apparent level of Act 1-Gus expression (as determined by in situ histochemical staining for GUS activity) may not necessarily reflect differences in GUS specific activity or in the level of Act1 5' region activity. Using the results presented here, it is possible to compare the pattern of Act1 5' region activity with what is known about the localization of CaMV 35S and maize Adh1 promoter activity in the vegetative and reproductive tissues of transgenic rice plants. Tereda and Shimamoto (1990) reported that CaMV 35S promoter activity in the vegetative organs of transgenic rice is mainly localized in and around the vascular tissue. Unlike the activity of the Act1 5' region, these authors detected no 35S promoter activity in the floral palea, stigma, or stamen tissue in the transgenic rice plants they examined. A more extensive histochemical analysis by Battraw and Hall (1990) revealed that the CaMV 35S promoter also functions in the leaf epidermal and mesophyll tissue, as well as in the root cortical cells of transgenic rice plants. However, unlike the rice,Acf7 5' region, little or no CaMV 35S promoter activity was detected in leaf trichome, bulliform, or bundle sheath cells or in root epidermal cells. Kyozuka et al. (1991) have reported that the maize Adh1 promoter is constitutively expressed in the root cap and meristem, stamen, pollen, and seeds of transgenic rice plants with increased promoter activity being induced by anaerobiosis in transgenic root tissue. However, unlike the Act1 5' region, they detected little or no maize Adh1 promoter activity in the shoot, leaf, ovary, or stigma tissue of transgenic rice plants. Therefore, the rice Act 1 5' region appears to display a more constitutive pattern of activity in transgenic rice plants than either the CaMV 35S or maize Adh1 promoter. It has previously been found that transcribed sequences from the Act 1 5' region, including the Act1 5' intron and the sequence context of the Act1 translation initiation codon, contribute to the high level of Actl-Gus gene expression in transient assays of transformed rice and maize cells (McElroy et al., 1991). Therefore, both the pattern and high level of GUS enzyme activity observed in rice plants expressing the Act 1-Gus fusion gene from pAct 1-D may well be influenced by post-transcriptional and translational processes associated with the transcribed sequences of \\r\eAct1 5' region. Whether this is the case or not, we found that higher levels of foreign gene expression in transgenic maize tissue can be obtained using rice >4cf7-based vectors than were previously obtained with maize Adh1-based vectors (McElroy et al. 1991; A.D. Blowers, D. McElroy, R. Wu, and E.D. Earle, unpublished data). Furthermore, high levels of Act1 5' region activity have also been obtained in transient assays of Act1-Gus gene expression in cells from a number of other monocot

Actin 5' Region Activity in Transgenic Rice

species, including wheat, oats, barley, and sorghum (A.D. Blowers, B. Jenes, J. Cao, D. McElroy, and R. Wu, unpublished data). Using antisera to divergent classes of soybean actin isoforms, McLean et al. (1990) showed that one class has a generalized pattern of expression in soybean root tissue, whereas another class has a more tissue-specific pattern of expression. However, each class of soybean actin isoform is coded for by at least two genes in the soybean genome (Hightower and Meagher, 1985). Therefore, the direct relationship between actin gene expression and the abundance of a specific actin isoform in soybean root tissue could not be determined. By transforming rice with an Actl-Gus fusion construct, we have been able to determine the expression pattern of a single member of the rice actin gene family. This type of study opens the way toward understanding the cellular mechanisms that act to regulate the expression of individual members of plant actin gene families. Such an understanding will eventually contribute toward an elucidation of actin function in higher plants. METHODS Rice Plant Regeneration from Transformed Protoplasts and GUS Activity Screening

1163

at 42OC in a solution containing 50% formamide and 5 x SSC (1 x SSC is 0.15 M NaCI, 0.015 M sodium citrate). Filters were washed three times at 6OoC for 20 min in 0.2 x SSC and 0.1% SDS. Quantitative Analysis of Actl-Gus Expression in Transgenic Rice Plants Fluorometric analysis of GUS activity in transgenic rice plant tissue was performed as described above for rice protoplasts using an SLM-Aminco spectrofluorometer (SLM Instruments, Urbana, IL). For immunoblot analysis, total soluble protein (40 pg) was incubated with 2 volumes of treatment buffer (Laemmli, 1970) and boiled for 1O min before being subjected to electrophoresis through a 12.5% polyacrylamide gel. Polyacrylamide gels were electroblotted to Immobilon-P Transfer Membrane (Millipore) as previously described (Le Gendre and Matsudaira, 1988). Membranes were incubated in a 300-fold diluted solution of rabbit GUS IgG (Clontech, Palo Alto, CA), washed, and incubated with a 3000-fold diluted goat anti-rabbit IgG-horseradish peroxidase (HRP) conjugate (Clontech). Bound peroxidase was detected using an HRP Substrate Kit following manufacturer's procedures (Bio-Rad). Protein molecular weights were determined by cornparison with prestained standards (Bethesda Research Laboratories). Filters were scanned using a Quick-Scan densitometer (Helena Laboratories, Beaumont, TX), and the results were compared to those obtained from known amounts of purified GUS protein (Clontech). In Situ Histochemical Staining for GUS Activity in Transgenic

Protoplastsfrom rice (Oryza safiva cv Taipei 309) cell suspension cultures were isolated and cultured as previously described (Abdullah et al., 1986). Polyethylene glycol-mediated protoplast transformation with pAct7-D was carried out as previously described (Zhang and Wu, 1988), with the omission of the heat shock treatment. A nurse cell culture procedure was employed for protoplast maintenance and culture using an 0.8-pm millipore filter (Millipore, Bedford, MA) after the gene transfer step (Cao et al., 1991). Cultures were incubated in darkness at 26OC with filters being wetted after 1 week with simplified KPR medium (Kao, 1977; Zhang and Wu, 1988) until calli developed. Rice plant regeneration from calli in differentiationmedium was as described by Zhang and Wu (1988). Putative transgenic plantlets were identified by histochemically screening calli from the shoot-root junction region (Zhang and Wu, 1988) for GUS activity using X-gluc (Jefferson et al., 1987). Putative transgenic plants (and regenerated plantlets that had gone through the transformation and regeneration procedure without the addition of foreign DNA) were transferred to the greenhouse (10 hr/28"C day to 14 hr/ 22OC night cycle) for subsequent growth and development. Analysis of Transgenic Rice Plant DNA Total leaf genomic DNA from putative transgenic (and untransformed control) plants was isolated by the method of Zhao et al. (1989). Ten micrograms of genomic DNA was digested with 30 units of restriction enzyme (Bethesda Research Laboratories) and gel blotted by the procedure of Kao et al. (1984). The 3'P-dATPlabeled Gus probe was generated by random hexamer labeling to give 10' cpm/pg DNA. Overnight hybridization was carried out

Rice Tissue Whole tissues and handcut thin sections from transgenic and untransformed control plant tissue were stained by incubation with X-gluc as previously described (Jefferson et al., 1987; Battraw and Hall, 1990). Pollen was isolated by gentle squeezing of immature or mature anthers. Stained material was examined microscopically using bright-field and dark-field optics. Photomicrographs were taken using an Olympus IMT-2 microscope and Kodak Ektachrome 160 tungsten film. Production of R, and R2 Generation Plants For back-crossingexperiments, anthers from open florets of wildtype rice (cv Taipei 309) were collected, and pollen from the anthers was applied to the stigmatic surface of transgenic rice plants from which anthers had been removed 1 day before flowering. Developing embryos were excised from the resulting seeds 20 days after manual pollination and cultured on N6 medium lacking plant growth regulators. After 3 weeks, the rescued embryos had developed into seedlings. At the three leaf growth stage, the seedlings were transferred to soil and placed in a greenhouse (10 hr/28OC day to 14 hr/22'C night cycle) for subsequent growth and development. ACKNOWLEDGMENTS The authors wish to thank Drs. Maureen Hanson, Alan D. Blowers, Karen Kindle, and Randy Wayne for critical reading of the manuscript, and Dr. Yogesh R. Mawal and Varman T. Samual for

1164

The Plant Cell

technical assistance. This research was supported by Grants RF86058 Allocation No. 60 and RF 90031 Allocation No. 126 from the Rockefeller Foundation, and GM29179 from the National lnstitutes of Health. W.Z. was supported by a predoctoral fellowship from the Rockefeller Foundation. D.M. was supported by a SERC/NATO studentship, a Fulbright Fellowship, scholarships from the British Universities North America Club and the St. Andrew’s Society of Washington, DC, a Cornell University East Asia Program Hu Shih Memorial Fellowship, and predoctoral fellowships from the Cornell University Biotechnology Program and the Cornell Plant Science Center. The Cornell Plant Science Center is a unit of the United States Department of Agriculture, Department of Energy, and National Science Foundation Plant Science Centers Program and a unit of the Cornell Biotechnology Program, which is sponsored by the New York State Science and Technology Foundation, a consortium of industries, and the U.S. Army Research Office.

Received June 27,1991; accepted September 23, 1991

REFERENCES

Abdullah, R., Cocking, E.C., and Thompson, J.A. (1986). EffiCient plant regeneration from rice protoplasts through somatic embryogenesis. Bio/Technology 4,1087-1 090. Battraw, M.J., and Hall, T.C. (1990). Histochemical analysis of CaMV 35s promoter-p-glucuronidasegene expression in transgenic rice plants. Plant. MOI. Biol. 15, 527-538. Cao, J., Zhang, W., McElroy, D., and Wu, R. (1991). Assessment of rice genetic transformation techniques. In Rice Biotechnology, G.S. Khush and G. Toenniessen, eds, (Wallingford, U.K.: C. A. B. International), in press. Cohen, S.M., Knecht, D., Lodish, H.F., and Loomis, W.F. (1986). DNA sequences required for expressionof a Dictyosteliumactin gene. EMBO J. 5,3361-3366. Elder, P.K., Frech, C.L., Subramaniam, M., Schmidt, L.J., and Getr, M.J. (1988). Evidence that the functional p-actin gene is single copy in most mice and is associated with 5’ sequences capable of conferring serum- and cycloheximide-dependent regulation. MOI.Cell. Biol. 8, 480-485. Frederickson, R.M., Micheau, M.R., Iwamoto, A., and Miyamoto, N.G. (1989). 5’ flanking and first intron sequence of the human p-actin gene required for efficient promoter activity. Nucl. Acids Res. 17, 253-270. Fregien, N., and Davidson, N. (1986). Activating elements in the promoter region of the chicken p-actin gene. Gene 48, 1-1 1. Fyrberg, E.A., Mahaffey, J.W., Bond, B.J., and Davidson, N. (1983). Transcription of the six Drosophila actin genes accumulate in a stage and tissue specific manner. Cell33, 115-1 23. Gordon-Kamm, W.J., Spencer, T.M., Mangano, M.L., Adams, T.R., Daines, R.J., Start, W.G., O’Brien, J.V., Chambers, S.A., Adams, W.R., Jr., Willetts, N.G., Rice, T.B., Mackey, C.J., Krueger, R.W., Kausch, A.P., and Lemaux, P.G. (1990). Transformation of maize cells and regeneration of fertile transgenic plants. Plant Cell 2, 603-618.

Guild, B.C., Finer, M.H., Housman, D.E., and Mulligan, R.C. (1988). Development of retrovirus vectors useful for expressing genes in cultured murine embryonal cells and hematopoietic cells in vivo. J. Virol. 62, 3795-3801. Hayashimoto, A., Li, Z., and Murai, N. (1990). A polyethylene glycol-mediated protoplast transformation system for production of fertile transgenic rice plants. Plant Physiol. 93, 857-863. Hightower, R.C., and Meagher, R.B. (1985). Divergence and differential expression of soybean actin genes. EMBO J. 4, 1-8. Hilder, V.A., Gatehouse, A.M.R., Sheerman, S.E., Barker, R.F., and Boulter, D. (1987). A nove1mechanism of insect resistance engineered into tobacco. Nature 330, 160-1 63. Jefferson, R.A., Burgess, S.M., and Hirsh, D. (1986). p-glucuronidase from Eschefichia coli as a gene-fusion marker. Proc. Natl. Acad. Sci. USA 83,8447-8451. Jefferson, R.A., Kavanagh, T.A., and Bevan, M.W. (1987). GUS fusions: p-Glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6,3901-3907. Kao, K.N. (1977). Chromosomal behavior in somatic hybrids of soybean-nicotiana glauca. MOI.Gen. Genet. 150, 225-230. Kao, T.H., Moon, E., and Wu, R. (1984). Cytochrome oxidase subunit II gene of rice has an insertion sequence within the intron. Nucl. Acids Res. 12, 7305-7315. Kyozuka, J., Izawa, T., Nakajima, M., and Shimamoto, K. (1990). Effect of the promoter and the first intron of maize Adhl on foreign gene expression in rice. Maydica 35,353-357. Kyoruka, J., Fujimoto, H., Izawa, T., and Shimamoto, K. (1991). Anaerobic induction and tissue-specific expression cf maize Adhl promoter in transgenic rice plants and their progeny. MOI. Gen. Genet. 228,40-48. Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Larkin, P.J., Banks, P.M., Bhati, R., Brettell, I.S., Davies, P.A., Ryan, S.A., Scowcroft, W.R., Spindler, L.H., and Tanner, G.J. (1989). From somatic variation to variant plants: Mechanisms and applications. Genome 31, 705-71 5. Le Gendre, N., and Matsudaira, P. (1988). Direct protein microsequencing from immobilon@-Ptransfer membrane. Biotechniques 6,154-1 59. Liu, Z.,Moav, B., Faras, A.J., Guise, K.S., Kapuscinski, A.R., and Hackett, P.B. (1990). Development of expression vectors for transgenic fish. Bio/Technology 8, 1268-1 272. McElroy, D., Rothenberg, M., Reece, K.S., and WU, R. (1990a). Characterization of the rice (Oryza sativa) actin gene family. Plant MOI.Biol. 15, 257-268. McElroy,)D., Zhang, W., Cao, J., and Wu, R. (1990b). lsolation of an efficient actin promoter for use in rice transformation. Plant Cell 2, 163-1 71. McElroy, D., Blowers, A., Jenes, B., and Wu, R. (1991). Construction of expression vectors based on the rice actin 1 (Act 7) 5’ region for use in monocot transformation. MOI. Gen. Genet., in press. McLean, B.G., Eubanks, S., and Meagher, R.B. (1990). Tissuespecific expression of divergent actins in soybean root. Plant Cell 2, 335-344.

Actin 5’ Region Activity in Transgenic Rice

Miyazaki, J., Takaki, S., Araki, K., Tashiro, F., Tominaga, A., Takatsu, K., and Yamamura, K. (1989). Expression vector systems based on the chicken 0-actin promoter directs efficient production of interleukind. Gene 79, 269-277. Nellen, W., Silan, C., Saur, U., and Firtel, R.A. (1986). Regulatory sequences in the promoter of the Dicfyosfelium actin 6 gene. EMBO J. 5,3367-3372. Parker, C.S., and Topol, J. (1984). A Drosophila RNA polymerase II transcription factor contains a promoter-region-specificDNAbinding activity. Cell 36, 357-369. Peterhans, A., Datta, S.K., Datta, K., Goodall, G.J., Potrykus, I., and Paszkowski, J. (1990). Recognition efficiency of Dicotyledoneae-specific promoter and RNA processing signals in rice. MOI.Gen. Genet. 222, 361-368. Sanchez, F., Tobin, S.L., Rdest, U., Zulauf, E., and McCarthy, B.J. (1983). Two Drosophila actin genes in detail: Gene structure, protein structure and transcription during development. J. MOI.Biol. 163, 533-551. Seagull, R.W. (1989). The plant cytoskeleton. Crit. Rev. Plant Sci. 8, 131-167. Shimamoto, K., Terada, R., Irawa, T., and Fujimoto, H. (1989). Fertile transgenic rice plants regenerated from transformed protoplasts. Nature 338,274-276.

1165

Staiger, C.J., and Schliwa, M. (1987). Actin localization and function in higher plants. Protoplasma 141, 1-12. Sugiyama, H., Niwa, H., Makino, K., and Kakunga, T. (1988). Strong transcriptional promoter in the 5’ upstream region of the human 0-actin gene. Gene 26, 135-1 39. Tereda, R., and Shimamoto, K. (1990). Expression of CaMV 35s-GUS gene in transgenic rice plants. MOI.Gen. Genet. 220, 389-392. Toriyama, K., Arimoto, Y., Uchimiya, H., and Hinata, K. (1988). Transgenic rice plants after direct gene transfer into protoplasts. Bio/Technology 6, 1072-1 074. Zhang, H.M., Yang, H., Rech, B.L., Golds, T.J., Davis, A.S., Mulligan, B.J., Cocking, E.C., and Davey, M.R. (1988). Transgenic rice plants produced by electroporation-mediated plasmid uptake into protoplasts. Plant Cell Rep. 7, 379-384. Zhang, W., and Wu, R. (1988). Efficient regeneration of transgenic rice plants from protoplasts and correctly regulated expression of foreign genes in the plants. Theor. Appl. Genet. 76, 835-840. Zhao, X., Wu, T., Xie, Y., and Wu, R. (1989). Genome-specific repetitive sequences in the genus Oryza. Theor. Appl. Genet. 78,201-209.