Rice UDP-Glucose Pyrophosphorylase1 Is Essential for Pollen Callose Deposition and Its Cosuppression Results in a New Type of Thermosensitive Genic Male Sterility^,

Abstract

UDP-glucose pyrophosphorylase (UGPase) catalyzes the reversible production of glucose-1-phosphate and UTP to UDP-glucose and pyrophosphate. The rice (Oryza sativa) genome contains two homologous UGPase genes, Ugp1 and Ugp2. We report a functional characterization of rice Ugp1, which is expressed throughout the plant, with highest expression in florets, especially in pollen during anther development. Ugp1 silencing by RNA interference or cosuppression results in male sterility. Expressing a double-stranded RNA interference construct in Ugp1-RI plants resulted in complete suppression of both Ugp1 and Ugp2, together with various pleiotropic developmental abnormalities, suggesting that UGPase plays critical roles in plant growth and development. More importantly, Ugp1-cosuppressing plants contained unprocessed intron-containing primary transcripts derived from transcription of the overexpression construct. These aberrant transcripts undergo temperature-sensitive splicing in florets, leading to a novel thermosensitive genic male sterility. Pollen mother cells (PMCs) of Ugp1-silenced plants appeared normal before meiosis, but during meiosis, normal callose deposition was disrupted. Consequently, the PMCs began to degenerate at the early meiosis stage, eventually resulting in complete pollen collapse. In addition, the degeneration of the tapetum and middle layer was inhibited. These results demonstrate that rice Ugp1 is required for callose deposition during PMC meiosis and bridges the apoplastic unloading pathway and pollen development.

INTRODUCTION

UDP-glucose pyrophosphorylase (UGPase) is found in all prokaryotic and eukaryotic organisms. It is a key enzyme in carbohydrate metabolism that catalyzes the reversible production of glucose-1-phosphate and UTP to UDP-glucose and pyrophosphate, depending on the metabolic status of the tissue. In photosynthetic source tissues, UGPase is coupled with sucrose phosphate synthase; it converts glucose-1-phosphate to UDP-glucose and is primarily involved in the synthesis of sucrose, the major sugar for photoassimilate export (reviewed in Kleczkowski et al., 2004). In nonphotosynthetic sink tissues, which to some extent depend on imported carbon resources, UGPase is linked to sucrose degradation pathways by converting UDP-glucose produced by sucrose synthase (SuSy) to glucose-1-phosphate for the demand of metabolic processes (Winter and Huber, 2000). In the cytoplasm of cereal seed endosperm, UGPase may also be coupled with cytosolic ADP-glucose pyrophosphorylase, resulting in a direct conversion of UDP-glucose to ADP-glucose for starch synthesis (reviewed in Kleczkowski, 1994).

In addition to its involvement in sucrose and starch metabolism, UGPase also takes part in cell wall biosynthesis. The substrate/product of UGPase, UDP-glucose, acts as a precursor for the synthesis of the carbohydrate moiety of glycolipids, glycoproteins, and cell wall components, including callose, pectin, and cellulose (reviewed in Kleczkowski, 1994; Amor et al., 1995; Gibeaut, 2000; Dong, 2004). UGPase mutant strains of Dictyostelium failed to complete the developmental cycle due to insufficient UDP-glucose levels for cellulose formation (Dimond et al., 1976). When the UgpB gene is disrupted in Dictyostelium discoideum, cells undergo aberrant differentiation and development with decreased spore viability and glycogen levels (Bishop et al., 2002). Also, Ugp is involved in cellulose biosynthesis in Acetobacter xylinum, as confirmed by its complementation of cellulose-negative mutants (Valla et al., 1989). A 10-fold–reduced UGPase activity in the yeast mutant resulted in a slight increase in the calcofluor sensitivity and a decrease in the cell-wall β-glucan content (Daran et al., 1995). Point mutations in the UGP gene lead to cellular UDP-glucose deficiency and decreased glycogen levels (Flores-Diaz et al., 1997).

In plants, Ugp genes are expressed in all tissues, including roots, tubers, leaves, stems, and young seeds (Zrenner et al., 1993; Abe et al., 2002). Pi deficiency, light exposure, and sucrose feeding could upregulate its expression, and upregulation of Ugp by sucrose is mediated via a hexokinase-independent and abscisic acid–insensitive pathway that involves an okadaic acid–responsive protein phosphatase (Ciereszko et al., 2001a, 2001b; Kleczkowski et al., 2004). Many Ugp genes have been cloned from various species of plants, including potato (Solanum tuberosum) (Katsube et al., 1990), barley (Hordeum vulgare) (Eimert et al., 1996), and rice (Oryza sativa) (Abe et al., 2002). Earlier studies concerning the biological function of plant Ugp genes focused on whether it represents a rate-limiting step in carbohydrate metabolism. In stored potato tubers, when UGPase activity was decreased by 30 to 50% by antisense suppression, significant reductions in sugar content were observed (Spychalla et al., 1994; Borovkov et al., 1996). In Arabidopsis thaliana, a decrease of 30% in UGPase activity led to a large decrease in carbohydrate content in the antisense plants, but without any detectable changes in plant growth parameters (Johansson, 2003).

The rice genome contains two homologous UGPase genes, Ugp1 on chromosome 9 (Abe et al., 2002) and Ugp2 on chromosome 2 (Mu, 2002). The rice Ugp1 gene on chromosome 9 was first cloned from immature endosperm. Its mRNA level was high at 10 to 15 d after flowering and then slightly decreased until up to 35 d after flowering (Abe et al., 2002). Remarkably, promoter reporter analysis has suggested that the rice Ugp2 gene on chromosome 2 is expressed only in binucleate pollen (Mu, 2002). Here, we report a detailed functional characterization of rice Ugp1. Ugp1 is expressed throughout the rice plant, with highest expression in florets, especially in pollen cells during anther and pollen development. Silencing of Ugp1 by RNA interference (RNAi) or cosuppression affects callose deposition during pollen wall development and leads to the degeneration of the pollen mother cell (PMC) at the early meiosis stage, resulting in male sterility in rice. These results demonstrate that Ugp1 is essential for PMC meiosis and microspore development in rice. Moreover, Ugp1-cosuppressing plants contained unprocessed intron-containing primary transcripts derived from the transcription of the overexpression construct. These aberrant transcripts undergo temperature-sensitive splicing in florets, resulting in a new type of thermosensitive genic male sterility.

RESULTS

Ugp1 Transcript Is Abundant in Florets

The rice genome contains two UGPase genes: Ugp1 on chromosome 9 (Abe et al., 2002) and Ugp2 on chromosome 2 (Mu, 2002). They share 81 and 88% sequence identity at the nucleotide sequence and the amino acid levels, respectively. We have previously isolated the rice Ugp1 gene (Chen et al., 2007). To understand the role of the gene in rice plant development, we first examined its expression in different rice organs by RNA gel blot analysis. The expression pattern is shown in Figure 1A. Ugp1 was expressed in various organs, including roots, seedling stems, mature stems, seedling leaves, and mature leaves. However, Ugp1 transcripts were most abundant in pooled florets at various stages before flowering (Figure 1A). A duplicate RNA membrane was hybridized with Ugp2 probes under the same high stringency conditions as those used to hybridize Ugp1, and no visible Ugp2 signal was detected (data not shown), indicating that the Ugp mRNA bands shown in Figure 1A consist solely of Ugp1 transcripts and that Ugp2 is expressed at extremely low levels. We therefore performed a more sensitive method, semiquantitative RT-PCR, to estimate the level of Ugp2 transcript. Low but detectable levels of Ugp2 transcript were found in roots, seedling stems, mature stems, seedling leaves, and mature leaves (Figure 1B). Ugp2 transcripts were also present at higher levels in pooled florets at various stages before flowering (Figure 1B). Together, these data indicate that both Ugp1 and Ugp2 are ubiquitously expressed throughout rice development and that Ugp1 is expressed at much higher levels than Ugp2. The presence of greater amounts of Ugp1 transcripts in florets before flowering strongly suggests that it plays a special role in rice flower development.

Figure 1.

Expression Patterns of Ugp1 and Ugp2 in Wild-Type Hejiang 19 Rice Plants.

(A) RNA gel blot analysis of Ugp1 transcript levels in wild-type Hejiang 19 rice plants. Total RNA (20 μg) extracted from wild-type tissues, including roots, seedling stems, mature stems, seedling leaves, mature leaves, and pooled florets at various stages before flowering was probed with full-length Ugp1 cDNA. An ethidium bromide stain of the gel is shown to confirm equal RNA loading.

(B) RT-PCR analysis of Ugp2 expression in wild-type Hejiang 19 rice plants. The amplification of the rice Actin1 gene was used as a control to show that approximately equal amounts of total RNA had been used in the RT-PCR analysis.

Identification of Ugp1-Overexpressing, Ugp1-Cosuppressing, and Ugp1-RNAi Transgenic Rice Plants

To elucidate the role of Ugp1 in rice growth and development, we constitutively increased and suppressed the expression of the Ugp1 gene in transgenic rice. The overexpression construct (Ugp1-OX), antisense construct (Ugp1-AS), and double-stranded RNA interference (dsRNAi) construct (Ugp1-RI) of Ugp1 were developed under the control of the maize (Zea mays) ubiquitin 1 (Ubi1) promoter (Christensen et al., 1992; Christensen and Quail, 1996) (Figure 2A). Transgenic plants were generated by introducing the constructs into the japonica rice variety Hejiang 19 by Agrobacterium tumefaciens–mediated transformation (Hiei et al., 1994). Forty independent Ugp1-OX plants (T0 generation), 24 independent antisense Ugp1-AS T0 plants, and 20 independent Ugp1-RI T0 plants were generated, each containing one to several copies of the transgene, as confirmed by DNA gel blot analysis (data not shown).

Figure 2.

Ugp1 Expression Patterns in Transgenic Rice Plants.

(A) Structures of constructs for rice transformation. The promoter of the maize ubiquitin 1 gene of ∼2 kb comprises, in the 5′ to 3′ direction, the following: a promoter with a transcription start site (+1); two overlapping heat shock elements located at positions −214 and −204 from the transcription start site; an 83-bp leader sequence adjacent to the transcription start site (exon); an intron of ∼1 kb; and a translation start site. The overexpression construct (Ugp1-OX), antisense construct (Ugp1-AS), and dsRNAi construct (Ugp1-RI) of Ugp1 were developed under the control of the Ubi1 promoter and nopaline synthase (Nos) terminator cassette. Arrows represent the primers used to generate the RT-PCR products shown in Figure 7D.

(B) RNA gel blot analysis of Ugp1 transcript levels in transgenic rice. Total RNA extracted from rice plants at the heading stage was probed with full-length Ugp1 cDNA. Arrowheads indicate (1) the unprocessed longer-than-full-length transcript and (2) endogenous Ugp1 mRNA; (3) to (5) indicate the silencing-related RNA degradation intermediates. Loading of equal amounts of RNA was confirmed by ethidium bromide staining.

(C) Protein gel blot analysis of UGPase protein from transgenic plants using an antibody against rice UGPase. Equal loading of proteins in each lane was confirmed by probing a duplicate blot with an anti-ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco) antibody.

(D) UGPase activities in transgenic plants. UGPase activities are shown as means ± se (n = 3).

(E) siRNA analysis in transgenic plants. Seventy-five nanograms of Ugp1 primers (18 and 21 nucleotides) were intermixed or mixed with wild-type RNA and served as size standard and hybridization controls. Closed arrowhead indicates the siRNA probed by the Ugp1 cDNA fragment between +333 and +852 (related to the ATG at +1 bp), and the open arrowhead indicates the siRNA by the 3′-end of the Ugp1 coding region. 5S rRNA was used as the loading control.

Co10, Co14, and Co27, cosuppressing plants of lines 10, 14, and 27, respectively; OX10 and OX14, overexpressing plants of lines 10 and 14, respectively; NS14, null segregant plant of line 14; R19 and R25, RNAi plants of lines 19 and 25, respectively.

Ugp1-OX T0 plants with a single T-DNA insertion were selected for further characterization. T1 lines were germinated and transplanted in the field at Wuhan University on April 28 and May 27, 2004. T1 plants carrying the Ugp1-OX construct were identified by PCR amplification of the transgene. Expression of Ugp1 was examined by RNA gel blot analysis of total RNA extracted from leaves of transgenic plants (Figure 2B). According to the abundance of Ugp1 transcripts in transgenic plants, the PCR-positive transgenic T1 plants of the Ugp1-OX lines segregated into two subpopulations. In one subpopulation, referred to as overexpressing plants, Ugp1 mRNA strongly accumulated (Figure 2B, OX10 and OX14). In the other subpopulation, endogenous Ugp1 expression was completely suppressed; no signal was detected in the regions corresponding to endogenous Ugp1 mRNA (Figure 2B, Co27, Co10, and Co14), suggesting that it was cosuppressed (Napoli et al., 1990; van der Krol et al., 1990). Thus, this subpopulation is referred to as cosuppressing plants. Besides low molecular weight and silencing-related RNA degradation intermediates, longer-than-full-length Ugp1 mRNAs, which contained the unspliced intron in the 5′-untranslated region of the Ubi1 promoter, were also present in the cosuppressing plants (Figure 7A).

Figure 7.

Effects of Temperature on Expression of Ugp1 and Ugp2 in Cosuppressing Plants.

The numbers above the lanes indicate the following anther developmental stages: 1, premeiosis; 2, meiosis; 3, young microspore; 4, later microspore; and 5, pollen mitosis. Co27, cosuppressing plant of line 27; R19, RNAi plant of line 19.

(A) RNA gel blot analysis of Ugp1 transcript levels in leaves of fertility-reverted cosuppressing plants at low temperature. The RNA gel blot was hybridized with the corresponding Ugp1 and subsequently with the Ubi1 intron probes. Arrowheads indicate (1) the unprocessed longer-than-full-length transcript and (2) endogenous Ugp1 mRNA; (3) to (5) indicate the silencing-related RNA degradation intermediates. Loading of equal amounts of RNA was confirmed by ethidium bromide staining.

(B) RNA gel blot analysis of Ugp1 transcript levels in florets of fertility-reverted cosuppressing plants grown at low temperature. The RNA gel blot was hybridized with corresponding Ugp1 probes. Loading of equal amounts of RNA was confirmed by ethidium bromide staining.

(C) Protein gel blot analysis of UGPase protein in cosuppressing plants. Protein samples were extracted from florets and leaves of cosuppressing plants grown at high temperature (male sterile stage) and low temperature (male fertile stage).

(D) RT-PCR analysis of the expression of Ugp2 and correctly spliced Ugp1 from unprocessed primary transcripts derived from transcription of the Ugp1-OX construct (Figure 2A) in florets of cosuppressing plants. Amplification of the rice Actin1 gene was used as a control to show that approximately equal amounts of total RNA had been used in the RT-PCR analysis.

We also performed protein gel blot analysis to investigate UGPase protein levels in the transgenic plants (Figure 2C). In accordance with the results of the RNA gel blot analysis, very high levels of UGPase protein were detected in the overexpressing plants, whereas UGPase signals in samples from the cosuppressing plants were very weak or undetectable. The UGPase activity correlated well with the amounts of UGPase protein (Figure 2D). The overexpressing plants had almost 260% wild-type UGPase activities, while the cosuppressing plants had only 11% wild-type activities.

RNA gel blot analysis showed that endogenous Ugp1 mRNA expression was not affected in PCR-positive transgene segregants of the Ugp1-AS population (data not shown). These plants exhibited no signs of abnormal growth and development; therefore, they were not studied further.

Ugp1-RI T0 plants were first analyzed by UGPase activity assays. Fourteen out of 20 Ugp1-RI T0 plants showed substantial reductions in leaf UGPase activity, ranging from 5.9 to 27.0% of that in wild-type Hejiang 19 control plants (see Supplemental Figure 1 online). Of these, R19 and R25 lines were selected for further molecular analysis. RNA gel blot analysis showed that endogenous Ugp1 mRNAs in R19 and R25 lines were completely suppressed, as in cosuppressing plants, since no signal was detected in the regions corresponding to endogenous Ugp1 mRNA (Figure 2B). These results indicate that Ugp1 was suppressed in these Ugp1-RI plants by RNAi (Wesley et al., 2001; Baulcombe, 2004).

To further confirm the silencing of Ugp1 in Ugp1-RI and cosuppressing plants, we examined them for the presence of small interfering RNAs (siRNAs), the hallmark of RNAi and cosuppression (Hamilton and Baulcombe, 1999; Baulcombe, 2004). As shown in Figure 2E, Ugp1-derived siRNAs were detected in the Ugp1-RI plants and cosuppressing plants, but not in any wild-type, null segregant, or overexpressing plants examined. In the Ugp1-RI lines, the only siRNAs detected corresponded to the region (333 to 852) designed to produce the inverted repeat hairpin RNA, while in cosuppressing plants, the detected siRNAs corresponded to the 3′-end of the Ugp1 coding region. Both the Ugp1-RI and cosuppressing plants showed no presence of siRNAs corresponding to the 5′-end of the Ugp1 coding region (data not shown). Taken together, these results strongly suggest that Ugp1 was silenced in both the Ugp1-RI and cosuppressing plants.

Silencing of Ugp1 by RNAi or Cosuppression Results in a Male-Sterility Phenotype in Transgenic Plants

None of the T1 plants of Ugp1-OX lines with a single T-DNA insertion showed any obvious phenotypic changes during the vegetative growth stage. However, during the reproductive growth stages, segregation was observed. All null segregants (PCR-negative) had a wild-type phenotype with normal fertility, whereas the PCR-positive transgenic T1 plants of the Ugp1-OX lines segregated into two subpopulations. The overexpressing subpopulation produced round, well-developed pollen grains and had a higher rate of seed-setting than the wild-type plants at maturity. In another cosuppressing subpopulation (Figure 3B), the panicle number was increased and the flowering time was delayed by several days compared with the wild-type Hejiang 19 plants (Table 1). They developed pale, shrunken anthers (Figures 3E and 3G), containing few collapsed and irregularly shaped pollen grains that adhered to each other and did not stain in the presence of KI-I₂ (Figure 3I). Consequently, these cosuppressing plants were completely male-sterile and bore no seeds on their panicles (Table 1).

Figure 3.

Phenotypes of Ugp1-Silenced Plants.

(A) Wild-type plant at maturity stage.

(B) Cosuppressing plant at maturity stage showing no seed set on the panicles.

(C) RNAi plant at maturity stage.

(D) Wild-type spikelets at the heading stage.

(E) Cosuppressing plant spikelets at the heading stage.

(F) Flower and anther morphology of a wild-type plant.

(G) Flower and anther morphology of a cosuppressing plant.

(H) Wild-type pollen grains stained with KI-I₂ solution.

(I) Cosuppressing plant pollen grains stained with KI-I₂ solution.

Co27, cosuppressing plant of line 27; R19, RNAi plant of line 19. Bar = 10 cm.

Table 1.

Morphological Characteristics of Ugp1-Silenced Plants

Line	Plant Height (cm)	Panicle Number	Flowering Time (Day)	Panicle Length (cm)	Grain Number (per Panicle)	Seed-Setting Rate (%)
Wild type	56.8 ± 2.01	24.0 ± 2.66	61.0 ± 0	13.4 ± 0.54	47.5 ± 2.23	75.9 ± 3.73
Co27a	54.8 ± 2.31	31.0 ± 4.02	65.7 ± 0.48	12.7 ± 1.53	48.3 ± 7.07	0
R19b	34.0 ± 6.34	6.0 ± 0.70	75.6 ± 2.07	11.5 ± 0.84	35.1 ± 4.17	0

Data represent means ± se (n = 10).

^aCo27, cosuppressing plant of line 27.

^bR19, RNAi plant of line 19.

Ugp1-RI plants showed pleiotropic developmental abnormalities (Figure 3C), including retarded growth, lower tiller numbers, reduced plant height, and delayed flowering (Table 1). When entering the reproductive stage, these plants also exhibited a male-sterile phenotype. Along with the male-sterility phenotype observed in cosuppressing plants, these results suggest that silencing of rice Ugp1 led to male sterility in the transgenic rice.

Sterility was solely due to the defective male gametes; female fertility in the Ugp1-silenced plants (including Ugp1-RI and cosuppressing plants) was not affected. When the cosuppressing male-sterile plants were cross-pollinated with wild-type plants and other varieties, they produced normal seeds and the F1 plants exhibited normal fertility. When the male-sterile Ugp1-RI T0 plants carrying single T-DNA insertions were crossed with wild-type plants, the F1 plants segregated in a 1:1 fertile:male-sterile ratio, indicating that the male-sterility phenotype in the Ugp1-RI lines can be inherited and is only present in F1 individuals that carry the transgene.

Cosuppression and Male Sterility in Ugp1-OX Plants Occur Only When the Transgene Locus Is Homozygous

As shown in Table 2, T1 individuals from each Ugp1-OX T0 line with a single insertion segregated in a Mendelian fertile:sterile fashion, with a ratio of 3:1, suggesting that cosuppression of Ugp1 and male sterility may be triggered in plants with a homozygous transgene locus. The 3:1 segregation ratio (Table 2) of fertile-to-sterile plants in the T2 progenies derived from T1 fertile transgenic plants indicates that these fertile T1 plants were heterozygous for the transgene locus. By contrast, all T2 progenies of null T1 segregants were fertile (Table 2). These fertility segregation analyses of T2 progenies further support the notion that cosuppression and its resulting male sterility are induced only in homozygotes, a previously documented phenomenon (de Carvalho et al., 1992). In addition, these results demonstrate that the male-sterility phenotype observed in the Ugp1-OX lines is attributable to the presence of the Ugp1 transgene rather than to any pleiotropic effect of rice transformation and tissue culture.

Table 2.

Fertility of Transgenic Plants with a Single Insertion of Ugp1-OX and Their Progenies

Line	Fertile	Sterile	χ2 (3:1)a
Segregation of fertility in Ugp1-OX T1 plantsb
10	9	4	0.2308
14	22	6	0.1905
27	20	3	1.7536
Segregation of fertility in T2 lines derived from heterozygous T1 plantsc
10-11	29	11	0.1333
14-8	96	30	0.0952
27-9	48	13	0.4426
Fertility in T2 lines derived from null segregant T1 plantsc
10-10	70	0
14-1	90	0
27-4	37	0
Fertility in T2 lines derived from homozygous T1 plantsc
27-10	0	98
27-12	0	38
27-19	0	25

^aχ² values test the fit of counts of fertile/sterile progeny of T0 or T1 plants to a 3:1 ratio.

^bThe seeds from Ugp1-OX T0 plants of lines 10, 14, and 27 were soaked, sown, and transplanted on April 28, May 5, and May 27, 2004, respectively.

^cThese plants were soaked, sown, and transplanted on April 22, April 28, and May 14, 2005, respectively.

Characterization of Male Gametogenesis in Ugp1-Silenced Plants

To clarify the nature of the male sterility in the Ugp1-silenced plants (including Ugp1-RI plants and cosuppressing plants), anther transverse sections were further examined. Based on the cellular events visible under the light microscope and the previous classification of anther development (Feng et al., 2001; Itoh et al., 2005), we delineated rice anther development into seven stages. Light microscopy observations indicated that the anther development was normal prior to meiosis of PMCs in Ugp1-silenced plants (exemplified by cosuppressing plants). Compared with wild-type anthers (Figure 4A), no differences were observed in the four outer layers (the epidermis, endothecium, middle layer, and tapetum) of the anther or PMCs of the cosuppressing plants before meiosis (Figure 4H). During the process of meiosis, the four outer layers of the anther remained normal (Figures 4I and 4J), but the PMCs of the cosuppressing plants began to degenerate in the locules of the anthers at an early stage of meiosis (Figure 4I). The degeneration continued into the later meiosis stage (Figure 4J). At the end of meiotic cell division, free microspores were released into the anther locules and the middle layers narrowed in the wild-type anthers (Figure 4D). By contrast, in the anthers of cosuppressing plants, the microspores were irregularly shaped and adhered to each other, and the middle layers retained their initial shape (Figure 4K).

Figure 4.

Anther Development in Wild-Type Plants and Cosuppressing Plants.

The micrographs show one of the four lobes in cross sections of wild-type anthers ([A] to [G]) and anthers from cosuppressing plants ([H] to [N]) at the premeiosis stage ([A] and [H]), early meiosis stage ([B] and [I]), meiosis stage ([C] and [J]), young microspore stage ([D] and [K]), later microspore stage ([E] and [L]), pollen mitosis stage ([F] and [M]), and mature pollen stage ([G] and [N]). DMC, degenerated meiocyte; Ep, epidermal cell layer; En, endothecial cell layer; M, middle layer; Tp, tapetum layer; YMsp, young microspore; DMsp, degenerated microspore; LMsp, later microspore; BP, binucleated pollen; MP, mature pollen. Bars = 10 μm.

Subsequently, in the later microspore stage, the microspores developed and the tapetum layers became very dense and thick in the wild-type anthers (Figure 4E). By contrast, in the cosuppressing plants, the microspores were shrunken and collapsed, and the tapetum layers began to manifest premature degeneration as the vacuoles enlarged (Figure 4L). At the pollen mitosis stage, the uninucleate pollen developed into trinucleate pollen through two mitotic divisions. In the wild-type plants, the tapetum layers were completely degenerated and the endothecial cell layers eventually ruptured, releasing the mature pollen grains (Figures 4F and 4G). At the same stage in anthers of cosuppressing plants, the pollen had completely degenerated, leaving only remnants in the anther locules, while the tapetum had become abnormally large and extremely vacuolated. The anthers also retained all four outer layers and did not dehisce (Figures 4M and 4N). These findings suggest that rice Ugp1 is essential for PMCs meiosis and microspore development as early as the early meiosis stage. Moreover, this gene is needed for the degeneration of the tapetum and middle layer.

Callose Deposition Is Disrupted in Pollen of Ugp1-Silenced Plants

Pollen development involves dynamic turnover of cell wall layers and changes in their composition (Polowick and Sawhney, 1992; Owen and Makaroff, 1995). Previous studies have shown that callose, one of the cell wall components, appears to play a vital role in the process of pollen development, since the lack of callose affects proper pollen wall formation (Waterkeyn and Beinfait, 1970; Worrall et al., 1992; Dong, 2004; Dong et al., 2005). We reasoned that the degeneration and irregular shape of the pollen in Ugp1-silenced plants may be due to the loss of the Ugp1 function in cell wall biosynthesis, which is required to maintain the integrity of the cell wall structure (Figures 4I to 4M). To test this possibility, we examined callose contents of the anthers at various developmental stages by aniline blue staining. Before the initiation of meiosis, callose was not detectable in PMCs of either wild-type or Ugp1-silenced plants (exemplified by cosuppressing plants; Figures 5A and 5E). At this stage, PMCs in cosuppressing plants appeared normal with no signs of degeneration (Figure 5E). During the process of meiosis, callose was deposited around the PMCs (seen as bright-yellow fluorescence in Figures 5B, 5C, 5F, and 5G) and the tetrads (Figures 5D and 5H). However, the PMCs and the tetrads of cosuppressing plants produced much less callose than those of wild-type plants and gradually degenerated (Figures 5F, 5G, and 5H). When microspores were released into the anther locule, both the normal microspores of the wild type and the degenerated microspores of cosuppressing plants exhibited weak callose fluorescence (Figures 5I and 5M). Subsequently, no callose signal was detected in the later microspores, binucleated pollens, or mature pollen of the wild-type plants (Figures 5J to 5L). However, strong callose fluorescence appeared in the abnormally large and extremely vacuolated tapetum of cosuppressing plants (Figures 5O and 5P). These results indicate that callose deposition was greatly reduced in the pollen (PMCs and tetrads) of cosuppressing plants during meiosis, and the lack of callose deposition due to cosuppression of Ugp1 affected the pollen wall structure and viability of these plants.

Figure 5.

Callose Deposition in Anthers during Microsporogenesis.

Cross sections of wild-type anthers and anthers from cosuppressing plants were stained with aniline blue to detect callose. Wild-type sections are shown in (A) to (D) and (I) to (L), and other panels show cosuppressing plant sections. Callose deposition is shown as bright-yellow fluorescence (indicated by arrows). Note that callose deposition was greatly reduced in PMCs and tetrads of cosuppressing plants. DMC, degenerated meiocyte; Td, tetrad; YMsp, young microspore; DMsp, degenerated microspore; LMsp, later microspore; BP, binucleated pollen; Tp, tapetum layer; MP, mature pollen. Bars = 10 μm.

(A) and (E) The premeiosis stage.

(B) and (F) The early meiosis stage.

(C) and (G) The later meiosis stage.

(D) and (H) The tetrad stage.

(I) and (M) The young microspore stage.

(J) and (N) The later microspore stage.

(K) and (O) The pollen mitosis stage.

(L) and (P) The mature pollen stage.

UGPase Protein Accumulates in Pollen Cells throughout Pollen Development in Wild-Type Plants but Not in Ugp1-Silenced Plants

To gain further insight into the action of the UGPase protein, we analyzed its expression pattern in anthers at different stages using immunolocalization techniques. In the wild-type plants, the UGPase protein was located predominately in PMCs at the premeiosis stage (Figure 6A), the early meiosis stage (Figure 6B), and the subsequent meiosis stage (Figure 6C) but was not present in other anther tissues, such as the epidermis, endothecium, middle layer, and tapetum. UGPase protein signals were still detectable in microspores, but they gradually declined in strength (Figure 6D). In control sections treated with preimmune serum, only weak background signals were observed (Figure 6E). By contrast, when anthers at the premeiosis stage of Ugp1-silenced plants (exemplified by cosuppressing plants) were analyzed, only background signal levels were detected (Figure 6F). Therefore, these results strongly confirm that the expression of UGPase is associated with PMC meiosis and microspore development.

Figure 6.

Expression of UGPase Protein in Wild-Type Anthers and Cosuppressing Anthers.

UGPase protein was immunolocalized in anthers of wild-type plants ([A] to [E]) and cosuppressing plants (F). The blue staining in the anthers of the wild-type plants at the premeiosis stage (A), early meiosis stage (B), later meiosis stage (C), and microspore stage (D) shows the localization of the UGPase protein in the PMCs and microspores. A consecutive section of (A) treated with preimmune serum as a negative control (E). No UGPase protein was detected in the anthers of cosuppressing plants at the premeiosis stage (F). Msp, microspore. Bars = 10 μm.

Male Sterility of Cosuppressing Plants Is a New Type of Thermosensitive Genic Male Sterility

While the male sterility of Ugp1-RI lines was stable under various environmental conditions, the cosuppressing plants that were completely sterile under natural summer conditions could revert to fertility in the autumn. When the cosuppressing plants headed in September, they showed normal fertility levels with seed setting 75.3% ± 1.47% (n = 10), equivalent to 75.9% of wild-type levels (Table 1). To determine if the male-sterility phenotype in T1 homozygous cosuppressing plants could be inherited by the seeds harvested in autumn, fertility parameters were also investigated in T2 progenies of these plants. As shown in Table 2, under natural summer conditions, the male sterility could be maintained through successive generations. Plants that were completely male-sterile in summer also reverted to fertility in autumn.

The fertility reversion of cosuppressing plants followed a period of consistently low temperatures (≤21°C) and short daylength photoperiod (<12.5 h). These environmental parameters associated with the reversion to fertility in cosuppressing plants resemble those that trigger reversion in photoperiod-sensitive genic male-sterility and thermosensitive genic male-sterility (TGMS) lines, which have been widely used in hybrid rice seed production (Virmani et al., 2003). The photoperiod-sensitive genic male-sterility lines are stably sterile under long-day conditions (>13.75 h) but fertile under short-day conditions (<13 h), regardless of the temperature. By contrast, the TGMS lines are sensitive to temperature, since they retain male sterility at high temperatures (>27°C) but revert to fertility at low temperatures (<24°C) (Virmani et al., 2003). To determine whether the fertility reversion in cosuppressing plants is determined by the photoperiod, T2 progeny of cosuppressing plants were subjected to a short-day treatment (9-h-light/15-h-dark cycles) under natural summer conditions. The results showed that cosuppressing plants remained completely male-sterile under the short-day conditions, suggesting that the fertility transformation in the cosuppressing plants is controlled by temperature, not by the photoperiod. Thus, the male-sterility phenotype in cosuppressing plants is a new type of TGMS.

Temperature-Sensitive Splicing in Florets Is the Basis for Fertility Reversion of TGMS in Cosuppressing Plants

The fertility reversion of Ugp1 cosuppressing plants at low temperature led us to hypothesize that the cosuppression of Ugp1 in these plants might reset. In other words, the silenced gene may be reactivated, a well-documented characteristic of cosuppression (de Carvalho et al., 1992; Hart et al., 1992; Dorlhac de Borne et al., 1994). To test this possibility, we first investigated the expression of Ugp1 in leaves of the fertility-reverted cosuppressing plants grown at low temperature (21°C). All fertility-reverted cosuppressing plants contained RNA degradation intermediates and the aberrant Ubi1 intron-containing longer-than-full-length RNAs, and no signal was detected in the regions corresponding to endogenous Ugp1 mRNA (Figure 7A). The expression of Ugp1 was further analyzed in florets of the fertility-reversed cosuppressing plants at low temperature (21°C). The RNA samples were prepared from florets at different anther developmental stages, according to the previous classification of anther development (Feng et al., 2001; Itoh et al., 2005). The aberrant Ubi1 intron-containing longer-than-full-length RNAs and RNA degradation intermediates were still present in florets of fertility-reversed cosuppressing plants, and endogenous Ugp1 mRNAs were still completely suppressed (Figure 7B). These results indicate that cosuppression of Ugp1 was still maintained in the fertility-reverted cosuppressing plants at low temperature (21°C).

The formation of organs and tissues usually requires the coordinated expression of a complex network of genes, and the function of the Ugp1 gene in anther development may be replaced by other genes in cosuppressing plants at low temperature. This possibility is based on the hypothesis that the expression of UGPase protein in florets is also suppressed. Thus, to test this possibility, we analyzed the expression of UGPase protein in florets at various anther developmental stages by protein gel blot analysis. When cosuppressing plants were in the male-sterile stage under natural summer conditions (high temperature), they did not show any UGPase expression in florets of any of the developmental stages. Significantly, however, when they reverted to fertility in the autumn at low temperature (21°C), they accumulated UGPase protein in their florets at close to wild-type levels (Figure 7C). These findings indicate that fertility reversion in cosuppressing plants is due to the accumulation of UGPase protein in florets at low temperature (21°C). In contrast with the florets, UGPase protein expression remains unchanged in leaves of cosuppressing plants regardless of the temperature (Figure 7C).

In attempts to trace the source of the UGPase proteins that accumulated in florets of fertility-reverted cosuppressing plants at low temperature (21°C), the possible contribution of endogenous Ugp1 mRNA could be excluded since it was completely suppressed (Figure 7B). However, as shown in Figure 1B, Ugp2 is also expressed in pooled florets at various stages before flowering. To determine whether the expression of Ugp2 contributes to the accumulation of UGPase proteins observed in florets of fertility-reverted cosuppressing plants at low temperature, we analyzed Ugp2 expression by RT-PCR of total RNA extracted from florets at different anther developmental stages. PCR primers were designed based on the 5′- and 3′-untranslated regions of Ugp2 to amplify full-length fragments of Ugp2 cDNA. As shown in Figure 7D, Ugp2 was mainly expressed in florets in the pollen mitosis stage, and the expression of Ugp2 was not affected in florets of cosuppressing plants, regardless of whether the temperature was high or low. Thus, the possible contribution of Ugp2 to the accumulation of UGPase proteins in florets of fertility-reverted cosuppressing plants at low temperature could also be excluded.

The aberrant longer-than-full-length RNAs, which contained the unspliced intron in the 5′-untranslated region of the Ubi1 promoter (Figures 7A and 7B), are unprocessed primary transcripts derived from the transcription of the Ugp1-OX construct (Figure 2A). These molecules in cosuppressing plants are destined to be processed for degradation or, alternatively, for processing into mature RNAs for translation, possibly contributing to the accumulation of UGPase proteins in florets of fertility-reverted cosuppressing plants at low temperature. To test the latter possibility, we performed RT-PCR to analyze the correctly spliced Ugp1 mRNAs from longer-than-full-length RNAs in florets of cosuppressing plants. As shown in Figure 7D, RT-PCR amplification of the correctly spliced Ugp1 mRNA sequence between the exon of the Ubi1 promoter and the 3′-untranslated regions of the nopaline synthase (Nos) terminator revealed the predicted 1521-bp product in florets of cosuppressing plants, both at high temperature and low temperature. Florets of cosuppressing plants that developed at low temperature accumulated much more correctly spliced Ugp1 mRNAs from longer-than-full-length RNAs than florets that developed at high temperature (Figure 7D). This finding shows that splicing of Ugp1 mRNA is temperature sensitive. Taken together, the results described above demonstrate that temperature-sensitive splicing of Ugp1 mRNAs from longer-than-full-length RNAs contributes to the accumulation of UGPase proteins in florets of cosuppressing plants, which is the molecular basis of the TGMS fertility reversion in cosuppressing plants.

DISCUSSION

UGPase Is Essential for Growth and Development in Rice

UGPase is a key enzyme in plant carbohydrate metabolism and cell wall biosynthesis that catalyzes the reversible production of glucose-1-phosphate and UTP to UDP-glucose and pyrophosphate. It has been well established that UGPase is essential for normal cell physiology and development in yeast and other fungi (Dimond et al., 1976; Daran et al., 1995; Bishop et al., 2002). However, reductions of UGPase activity in potato and Arabidopsis by antisense suppression have been shown to have no detectable effect on plant growth and development (Zrenner et al., 1993; Spychalla et al., 1994; Borovkov et al., 1996; Johansson, 2003). In this study, although silencing of Ugp1 by cosuppression also had no visible phenotypic effects on rice during vegetative growth, when Ugp1 was silenced by expressing a dsRNAi construct, pleiotropic developmental abnormalities (including retarded growth, lower tiller numbers, reduced plant height, and delayed flowering) occurred (Table 1, Figure 3C), providing direct evidence that UGPase plays critical roles in plant growth and development.

The failure to detect growth and developmental defects in previous antisense studies in potato and Arabidopsis was very likely due to low suppression efficacy and the presence of other Ugp genes. The phenotypic differences in vegetative growth between Ugp1-RI plants and cosuppressing plants may also result from the different silencing efficacy. Cosuppression of Ugp1 has no effect on the expression of Ugp2; however, both Ugp1 and Ugp2 are completely suppressed in Ugp1-RI plants (Figure 7D). Generally, the genetic requirements for S-PTGS (cosuppression) and IR-PTGS (dsRNAi) are different, and dsRNA is the branch point between the S-PTGS and IR-PTGS pathways. AGO1, SGS2/SDE1, SGS3, and SDE3 proteins, which are required for the formation and/or replication of dsRNA in S-PTGS, are dispensable for IR-PTGS (Beclin et al., 2002). The high efficiency of IR-PTGS is due to the folding of transgene-derived RNAs into long, stable dsRNA structures (Wesley et al., 2001). Thus, dsRNA is constitutively produced and the endogenous Ugp mRNAs, including Ugp1 and Ugp2, are degraded globally in Ugp1-RI plants (Wesley et al., 2001; Baulcombe, 2004), resulting in severe vegetative growth defects that are absent in cosuppressing plants.

Rice Ugp1 Is Required for Callose Deposition during Pollen Development

Ugp1-silenced plants (including Ugp1-RI plants and cosuppressing plants) are male-sterile but female-fertile. During meiosis, normal callose deposition was disrupted in the PMCs of Ugp1-silenced plants (Figures 5F to 5H), resulting in their degeneration at the early stage of meiosis. The results demonstrated here provide direct evidence that rice Ugp1 is essential for callose deposition during pollen development. Consequently, only the remnants of pollen remain in the locules of Ugp1-silenced plants' anthers. The anthers of Ugp1-silenced plants also show other developmental defects. While the middle layer and the tapetum degenerate in the wild-type anthers, this process does not occur in Ugp1-silenced plants' anthers.

Multiple roles for the callose wall have been proposed over the last 40 years. Prior to meiosis, the sporogenous initials divide to form PMCs surrounded by a simple primary wall with a distinct middle lamella. The wall doubles in width, and newly formed fibrillar wall materials are deposited in it. At the onset of meiosis, the wall microfibrils disappear and a secondary callose wall is deposited between the primary wall and plasma membrane of the PMCs (Polowick and Sawhney, 1992; Owen and Makaroff, 1995). The callose layer is believed to prevent the cohesion and fusion of the PMCs (Waterkeyn, 1962), and it may function as a molecular filter that protects the developing pollen cells from premature swelling and bursting under the influence of the surrounding diploid tissues (Heslop-Harrison and Mackenzie, 1967). In addition, the callose wall can provide a template or mold for formation of the exine wall of pollen grains (Waterkeyn and Beinfait, 1970; Worrall et al., 1992; Dong, 2004; Dong et al., 2005). The Arabidopsis callose synthase mutant cals5 lacks normal callose deposition in meiocytes, tetrads, microspores, and mature pollen. Consequently, the pollen exine wall is not formed properly and the mutant exhibits male sterility (Dong et al., 2005). However, cals5 mutants undergo apparently normal meiosis. Similarly, normal callose deposition was disrupted in PMCs of Ugp1-silenced plants in our study, but this led to degeneration of the PMCs at the early stage of meiosis. Our results demonstrate that callose deposition is also essential for both the integrity of PMC cell walls and male meiosis, in addition to the other important roles of callose deposition in the process of microsporogenesis (Tucker et al., 2001; Dong et al., 2005).

In plants, UDP-glucose can potentially be synthesized by two different pathways. One pathway involves UGPase, while the second involves SuSy, which catalyzes the reversible hydrolysis of sucrose to fructose and UDP-glucose. A membrane-associated form of SuSy has been suggested to channel carbon directly from sucrose to cellulose and/or callose in developing cotton fibers (Amor et al., 1995). However, the developing pollen represents a symplastically isolated sink with no plasmodesmatal connections to the surrounding sporophytic tissue (Clément and Audran, 1995). Thus, an unloading pathway via apoplastic cleavage of sucrose by extracellular invertase and uptake of the resulting hexose monomers by monosaccharide transporter is mandatory (Eschrich, 1980). It has been reported that invertase is the dominant sucrose-cleaving enzyme in anthers (Dorion et al., 1996; Sheoran and Saini, 1996; Ranwala and Miller, 1998; Koonjul et al., 2005). In wheat (Triticum aestivum) and rice, the arrest of pollen development and consequent male sterility by meiotic-stage water stress is preceded by a specific decline in anther invertase activity, resulting from selective transcriptional downregulation of anther invertase genes (Dorion et al., 1996; Sheoran and Saini, 1996; Koonjul et al., 2005). Tissue-specific antisense inhibition of the extracellular invertase gene Nin88 under the control of its own promoter in tobacco (Nicotiana tabacum) reduces apoplastic invertase activity, leading to the developmental arrest of pollen at the unicellular microspore stage, thus causing male sterility (Goetz et al., 2001). Furthermore, cold-induced repression of the rice anther-specific cell wall invertase gene Os INV4 is correlated with the significant reduction in anther cell wall–bound invertase activity, premature callose degradation, and eventual pollen sterility (Oliver et al., 2005; Mamun et al., 2006). Taken together, these results (and the identification of the pollen-specific monosaccharide transporter; Ylstra et al., 1998; Truernit et al., 1999) prove that the apoplastic unloading pathway plays a vital role in pollen development. In this context, rice Ugp1 could bridge the apoplastic unloading pathway and pollen development because it is specifically expressed in pollen cells during anther and pollen development (Figure 6). Briefly, in the anther apoplastic unloading pathways, sucrose is irreversibly hydrolyzed by extracellular invertase. The resulting hexose monomers are taken up by pollen cells through pollen-specific monosaccharide transporters (Ylstra et al., 1998; Truernit et al., 1999) and then metabolized to glucose-1-phosphate, which is converted to UDP-glucose by UGPase for pollen cell wall biosynthesis.

Temperature-Sensitive Splicing in Florets Is the Basis for Fertility Reversion of TGMS in Cosuppressing Plants

Cosuppression was originally discovered as the reciprocal and coordinated silencing of transgenes and homologous host genes (Napoli et al., 1990; van der Krol et al., 1990). It is more pronounced in homozygous than in corresponding hemizygous plants (de Carvalho et al., 1992; Hart et al., 1992; Dorlhac de Borne et al., 1994). The frequency of cosuppression positively correlates with the promoter strength of the transgene (Que et al., 1997). In addition, transgene transcription is required for cosuppression (Dorlhac de Borne et al., 1994; Vaucheret et al., 1997). In this study, the Ugp1 transgene was under the control of the strong, constitutive maize Ubi1 promoter, and cosuppression was triggered exclusively when plants were homozygous for the transgene locus. Both of these findings support the cosuppression threshold hypothesis (Lindbo et al., 1993). Significantly, in addition to low molecular weight and silencing-related RNA degradation intermediates, aberrant Ubi1 intron-containing longer-than-full-length Ugp1 mRNAs, which are unprocessed primary transcripts derived from transcription of the Ugp1-OX construct (Figure 2A), were present in the cosuppressing plants (Figure 7A). This can be attributed to the characteristics of the Ubi1 promoter that drives the Ugp1 expression. The Ubi1 promoter of maize contains two overlapping sequences similar to the consensus heat shock element found in heat-inducible genes and an intron (Figure 2A). The heat shock elements of the promoter region enhance the expression of the downstream gene in response to elevated temperatures, and the intron is inefficiently processed following a brief, severe heat shock (Christensen et al., 1992).

Recently, it has been shown that low temperature inhibits virus- and transgene-induced RNA silencing by controlling siRNA generation, leading to enhanced virus susceptibility and loss of silencing-mediated transgenic phenotypes. By contrast, RNA silencing is activated and the amount of siRNAs gradually increases with rising temperature (Szittya et al., 2003). However, our results indicate that cosuppression of Ugp1 was not affected in fertility-reverted cosuppressing plants at low temperature (Figures 7A and 7B). Significantly, temperature strongly affected the splicing of Ugp1 mRNAs from longer-than-full-length RNAs. Florets of cosuppressing plants grown at low temperature accumulated much more correctly spliced Ugp1 mRNAs from longer-than-full-length RNAs than those grown at high temperature (Figure 7D), which contributed to the accumulation of sufficient levels of UGPase protein in their florets and, thus, their reversion to fertility. Our results demonstrate that the TGMS phenotype of the cosuppressing plants is attributable to temperature-sensitive splicing, which may provide clues regarding other unknown, molecular TGMS mechanisms (Kurata et al., 2005).

Several types of temperature-sensitive mutations affecting splicing have been characterized. These mutations affect trans-acting proteins required for pre-mRNA splicing due to the production of a defective protein product (Weidenhammer et al., 1996; Arenas and Abelson, 1997). There are few reported cases of temperature-sensitive mutations that have a cis effect on the splicing of a specific transcript (Hughes et al., 1996; Sablowski and Meyerowitz, 1998; Gemignani et al., 2002). The temperature-sensitive mutant ap3-1 carries a missense mutation near a 5′-splice site in exon 5. At the nonpermissive temperature, exon 5 is frequently skipped by the splicing machinery. The resulting inactive AP3De5 causes ap3-1 to develop sepaloid organs and carpelloid organs instead of stamens. At the permissive temperature, the correctly spliced ap3-1 mRNA (exon 5–containing mRNA) accumulates, and the presence of sufficient levels of the active missense mutation protein AP3–Met-153 allows ap3-1 to produce nearly wild-type flowers (Sablowski and Meyerowitz, 1998). Similarly to ap3-1 mutant, the temperature-sensitive splicing of Ugp1 from transgene transcripts in cosuppressing plants could also be attributable to the temperature sensitivity of a component of the splicing machinery. More interestingly, the temperature-sensitive splicing of Ugp1 in cosuppressing plants provides a novel example of cis element (heat shock element) based temperature-sensitive splicing, which may involve the interactions of heat shock elements, heat shock transcription factors, and heat shock proteins (Yost and Lindquist, 1986, 1991).

METHODS

Plant Materials

The wild-type rice (Oryza sativa) strain used in this study was a japonica variety, Hejiang 19. The transgenic plants and wild-type plants were grown in an experimental field at Wuhan University Institute of Genetics (Wuhan, China) (latitude 30°34' North; longitude 114°17' East). All experimental materials were transplanted in the field at a spacing of 16.7 cm between plants within each row and 26.7 cm between rows. The plants were tended under the routine management regime.

Plasmid Constructs and Plant Transformation

All constructs were made using standard procedures (Sambrook et al., 1989). Enzymatic reactions were performed using the conditions and buffers recommended by the suppliers. Plasmid pAHC17 (Christensen and Quail, 1996) was partially digested with HindIII and EcoRI. The maize (Zea mays) ubiquitin 1 (Ubi1) promoter and Nos terminator cassette were inserted into the HindIII and EcoRI sites of pCAMBIA 1301, resulting in plasmid pCU. The coding region of rice Ugp1 was amplified using primers 5′-ATTGGATCCATGGCGGTCGCCGCCGACGTG-3′ and 5′-CCGGGATCCTCAAAGATCCTCCGGACCGTTG-3′ and ligated to the BamHI site of pCU to make the overexpression construct (Ugp1-OX) and antisense construct (Ugp1-AS), respectively (Figure 2A). The Ugp1-OX and Ugp1-AS constructs were confirmed by sequencing.

To make a dsRNAi construct, a 520-bp fragment between +333 and +852 (related to the ATG at +1 bp) was amplified from the Ugp1 coding sequence using the following primers: 5′-CCTCTCGAGCCTTATTGTGATTCAAATTGAG -3′ (XhoI), 5′-CTTGAATTCATTCACATGCTCATC-3′ (EcoRI), 5′-GTTTCTAGACCTTATTGTGATTC-3′ (XbaI), and 5′-CTTAAGCTTATTCACATGCTCATCAGGGAC-3′ (HindIII). The PCR fragment generated using the XhoI-EcoRI pair of primers (to form the sense fragment) was cloned first into pGEM-T (Promega), and then the XhoI-EcoRI–digested Ugp1 fragment was cloned into the same restriction enzyme sites present in the pKANNIBAL vector (Wesley et al., 2001), thus creating pKAN-Ugp1XE. The PCR fragment generated using the XbaI-HindIII pair of primers (to form the antisense fragment) was also cloned first into pGEM-T vector, and then the XbaI-HindIII–digested Ugp1 fragment was cloned into the same sites in pKAN-Ugp1XE to create pKAN-(Ugp1)₂. The XbaI-XhoI dsRNAi fragment contained the two Ugp1 sequences in opposing orientations separated by an 800-bp Pdk intron sequence. This dsRNAi fragment was amplified using the following primers, with the added BamHI site underlined, 5′-ATTGGATCCATTTGGAGAGGACACGCTCGAG-3′ and 5′-CTCGGATCCCTCATTAAAGCAGGACTCTAG-3′. It was then ligated to the BamHI site of pCU to form the dsRNAi construct (Ugp1-RI, Figure 2A). Overexpression, antisense, and dsRNAi constructs were introduced into Agrobacterium tumefaciens (strain EHA105) by electroporation. The Agrobacterium-mediated transformation was performed using vigorously growing calli derived from mature embryos of Hejiang 19 following a standard procedure (Hiei et al., 1994). T0 transgenic plants were grown in a greenhouse under 14-h-light/10-h-dark cycles at 28°C.

Molecular Analyses

Total RNA was isolated using TRIzol reagent (Gibco BRL) according to the manufacturer's protocol from various rice tissues: roots, seedling stems, mature stems, seedling leaves, mature leaves, and florets. The florets at different anther developmental stages were classified in the following categories according to floret length (Feng et al., 2001; Itoh et al., 2005): premeiosis, meiosis, young microspore, later microspore, and pollen mitosis stages when the floret length was 1 to 2, 2 to 3, 3 to 4, 4 to 5, and 5 to 7 mm, respectively. RNA concentrations were estimated based on absorbance at 260 and 280 nm. For RNA gel blot analysis, total RNA was separated on a denaturing 1.5% formaldehyde agarose gel (Sambrook et al., 1989). Loading of equal amounts of RNA was confirmed by ethidium bromide staining. The RNA was transferred to Hybond N⁺ membrane (Amersham Pharmacia Biotech). The probe was labeled with [α-³²P]dCTP using the Prime-a-Gene labeling system (Promega), and the membranes were hybridized for at least 10 h at 65°C with the labeled probe. Blots were washed for 15 min at 65°C in 1× SSC and 0.2% SDS and subsequently for 15 min at 65°C in 0.5× SSC and 0.1% SDS. The membranes were then exposed to x-ray film for autoradiography.

siRNA detection was performed as previously described (Lechtenberg et al., 2003). Electrophoresis was performed in 3% agarose gels in the presence of formaldehyde. Blots of siRNAs were prehybridized at 58°C for 2 h, and hybridizations were performed at 42°C for 16 h. The blots were washed for 3× 10 min at 37°C in 2× SSC/0.5% SDS and then in 2× SSC/0.2% SDS at 42°C for 10 min. The Ugp1 primers (18 and 21 nucleotides) were used as size standard and hybridization controls.

For RT-PCR analysis, 2 μg of total RNA, extracted as described above, was treated with RQI RNase-Free DNase I (Promega) and then reverse-transcribed in a total volume of 20 μL using the Superscript III first-strand cDNA synthesis system for RT-PCR and oligo(dT)₂₀ primer according to the manufacturer's instructions (Invitrogen). The 20-μL first-strand cDNA product was diluted to 100 μL final volume with TE buffer, and 1 μL of the resulting dilutions was used as template in a 25-μL reaction mixture with the KOD Plus DNA polymerase (TOYOBO) using gene-specific primers. Two primers, 5′-TCATCAGATCAGCGTGAAGC-3′ and 5′-GCCCACTCACAAGGAGAAAA-3′, based on the 5′- and 3′-untranslated regions of rice Ugp2, were used to amplify the full-length fragments of Ugp2 cDNA. For the analysis of spliced Ugp1 from unprocessed primary transcripts derived from the Ubi1 promoter, primer pair Ubi1 exon-F (5′-AACCAGATCTCCCCCAAATC-3′) and Nos-R (5′-AAGACCGGCAACAGGATTC-3′), based on the 83-bp leader sequence adjacent to the transcription start site (exon) of the Ubi1 promoter and 3′-untranslated regions of Nos terminator (Figure 2A), were used. For amplification of the rice Actin1 gene, primer pair Actin1-F (5′-GAACTGGTATGGTCAAGGCTG-3′) and Actin1-R (5′-TCCAACACAATACCTGTGGT-3′) were used. The number of cycles used for amplification with each primer pair was adjusted to ensure that the amplification was in the linear range. The PCR products were sequenced to ensure that they were derived from the targeted genes.

UGPase Enzyme Activity Assay

Protein extraction was performed as previously described (Ciereszko et al., 2001a). Rice leaves and florets at various developmental stages were ground to powder with liquid nitrogen and homogenized in extraction buffer with a sample-to-buffer ratio of 1:5 (w/v). The homogenate was centrifuged for 15 min at 20,000g at 4°C, and the resulting supernatant was used to assay UGPase activity (Sowokinos et al., 1993). Reaction mixtures (1 mL, pH 8.5) contained 80 pmol of glycylglycine, 1 pmol of UDP-Glc, 5 pmol of MgC1₂, 1 unit each of phosphoglucose mutase and Glc-6-P dehydrogenase, 20 pmol of Cys, 0.02 pmol of Glc-1,6-diP, 0.6 pmol of NADP, and 20 μL of protein extract. The reactions were initiated by the addition of 2.5 pmol of inorganic pyrophosphate. The formation of NADPH (340 nm) was recorded continuously at 30°C until the reaction rate was no longer linear. One unit of activity in the pyrophosphorolysis direction was defined as the amount of enzyme that catalyzes the formation of 1 μmol of Glc-1-P/min.

Protein Gel Blot Analysis

Eight microliters of sample extract was mixed with 2 μL of 5× SDS-PAGE sample buffer (Sambrook et al., 1989). Samples were heated for 3 min at 95°C and loaded on 12% SDS-polyacrylamide gels. After electrophoresis, proteins were electroblotted to nitrocellulose membranes. Membranes were blocked in 5% nonfat dry milk in TBS-T (20 mM Tris-HCl, pH 7.6, 137 mM NaCl, and 0.1% Tween 20) overnight at 4°C and then washed three times in TBS-T for 8 min each. The membranes were incubated with the primary UGPase polyclonal antibody (1:2500) diluted with 1% BSA in TBS-T for 1 h at room temperature (Chen et al., 2007). Next, the membranes were washed in TBS-T three times for 8 min each. Membranes were then incubated for 1 h with goat anti-rabbit IgG (1:2500 dilution in TBS-T and 1% BSA) conjugated to alkaline phosphatase, followed by three washes for 8 min each in TBS-T, and detected using the 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium (BCIP/NBT) solution as the substrate.

Histological Analysis

Bright-field photographs of individual flowers were taken using an Olympus SZX12 dissecting microscope equipped with CCD. Rice anthers at various developmental stages were dissected and vacuum infiltrated with 4% paraformaldehyde in PBS, pH 7.0, for 30 min, and then the fixative was renewed with fresh paraformaldehyde solution and left to fix at 4°C overnight. The samples were washed in PBS, dehydrated in a graded ethanol series, and embedded in Paraplast Plus (Sigma-Aldrich). Microtome sections (10 μm thick) were stained with 0.2% hematoxylin to stain the cells and 0.1% aniline blue to stain the callose walls. The slides were inspected and photographed using an Olympus BX51 microscope.

For immunocytochemical analysis, the sections were washed six times, after dewaxing and rehydration, in PBST (PBS and 0.05% Tween 20). After blocking for 30 min in 0.5% BSA and 10% normal goat serum in PBST at room temperature, the slides were incubated overnight with a 1:1000 dilution of the primary UGPase polyclonal antibody in PBST and 0.5% BSA at 4°C. The slides were then washed six times in PBST and incubated for 30 min at 37°C with a 1:500 dilution of the secondary antibody (alkaline phosphatase–conjugated goat anti-rabbit antibody). Subsequently, the slides were washed six times in PBST and developed with BCIP/NBT in TNM (10 mM Tris-HCl, pH 9.5, 100 mM NaCl, and 5mM MgCl₂). Sections were examined and photographed using a cooled CCD on an Olympus BX51 microscope.

Fertility Investigation

The anthers of the wild-type and transgenic plants were collected from spikelets just before anthesis. Pollen grains from the broken anthers were suspended in a 1% iodine–potassium iodide solution (KI-I₂). The stained pollen was examined and photographed under a microscope (Olympus BX51). To investigate seed setting, panicles were bagged before heading to prevent cross-pollination. The spikelets and grains from each of the panicles were counted.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers DQ395328 (Ugp1) and AF249880 (Ugp2).

Supplemental Data

The following material is available in the online version of this article.

• Supplemental Figure 1. UGPase Activity Assay in Ugp1-RI T0 Plants.