• Open Access

Spatial and temporal expression of endosperm transfer cell-specific promoters in transgenic rice and barley


* Correspondence (fax +63 8 830 37 102, e-mail sergiy.lopato@ACPFG.com.au)


Two putative endosperm-specific rice genes, OsPR602 and OsPR9a, were identified from database searches. The promoter regions of these genes were isolated, and transcriptional promoter:β-glucuronidase (GUS) fusion constructs were stably transformed into rice and barley. The GUS expression patterns revealed that these promoters were active in early grain development in both rice and barley, and showed strongest expression in endosperm transfer cells during the early stages of grain filling. The GUS expression was similar in both rice and barley, but, in barley, expression was exclusively in the endosperm transfer cells and differed in timing of activation relative to rice. In rice, both promoters showed activity not only in the endosperm transfer cells, but also in the transfer cells of maternal tissue and in several floral tissues shortly before pollination. The expression patterns of OsPR602 and OsPR9a in flowers differed. The similarity of expression in both rice and barley suggests that these promoters may be useful to control transgene expression in the transfer cells of cereal grains with the aim of altering nutrient uptake or enhancing the barrier against pathogens at the boundary between maternal tissue and the developing endosperm. However, the expression during floral development should be considered if the promoters are used in rice.


The cereals rice, wheat, maize and barley are major sources of human food and animal feed, and also provide the raw material for many industries. Some of the limitations of conventional breeding of cereals may be overcome by applying the techniques of genetic transformation or engineering (Vonwettstein, 1993; Sinclair et al., 2004; Shrawat and Lörz, 2006). An important application of genetic engineering is expected to be in the improvement of grain size, quality and yield by the modulation of the levels of expression of particular gene(s), whilst retaining the desirable qualities in the selected cultivar. Negative effects of constitutive transgene expression on plant development may be prevented by the use of tissue- or development-specific promoters (Potenza et al., 2004). Endosperm-specific promoters are of great importance for the development of modified grains with altered grain size, shape and composition, as well as biotic and abiotic stress tolerance.

After cellularization, the endosperm of cereals comprises at least four different tissues which have specialized roles: the endosperm transfer cells (ETCs), embryo surrounding region (ESR), starchy endosperm and aleurone. ETCs are responsible for the uptake of nutrients and act as a barrier against pathogens. The ESR protects the embryo, transfers nutrients to the embryo and acts as a channel for regulatory cross-talk between the endosperm and embryo. Starchy endosperm cells make up the bulk of the endosperm and accumulate storage proteins and starch. The aleurone layer is important as the site of synthesis of enzymes which hydrolyse the storage reserves in the endosperm cells at the time of seed germination (Becraft, 2001; Olsen, 2001, 2004).

Most of the endosperm-specific promoters from monocots described in the literature have been shown to be active in the late phases of grain development, and are expressed mainly in starchy endosperm (Russell and Fromm, 1997; Lamacchia et al., 2001; Qu and Takaiwa, 2004; Su et al., 2004). They show maximal activity at the middle of grain maturation, when cell proliferation and differentiation are mostly complete. These promoters are useful for biotechnological projects attempting to increase starch and protein content or to enrich grain with particular amino acids (Vonwettstein, 1993). Strong endosperm-specific promoters of storage proteins are also widely used for the high-level expression of proteins of non-grain origin, with the aim of large-scale protein production for applications such as the production of pharmaceuticals and in industry (Wang et al., 1994b; Hood and Jilka, 1999; Hood et al., 2002). However, for the manipulation of grain size and shape, it is probable that the expression of transgenes will be needed before or during cellularization and in tissues involved in nutrient transfer to the developing endosperm, such as the ETC layers.

ETCs are in contact with cells of the maternal tissue. These cells transfer efficiently sugars and amino acids from maternal tissue to the endosperm (Wang et al., 1994b; Becraft, 2001; Olsen, 2004). To increase the efficiency of transfer, ETCs develop cell wall ingrowths, which increase the surface area by up to 22-fold (Wang et al., 1994b).

In maize, transfer cells are restricted to the aleurone layer and three to four adjacent layers of starchy endosperm (Thompson et al., 2001; Olsen, 2004). In barley and wheat, transfer cells are formed from maternal nucellar cell projections adjacent to the endosperm cavity and at least one layer of ETCs on the other side of the endosperm cavity (Wang and Fisher, 1994; Wang et al., 1994a,b; Broekaert et al., 1997; Olsen, 2004). In rice, the pericarp has four vascular bundles, three of which are situated on the ventral part and side faces of the grain and are responsible for the supply of nutrients and water for the development of the pericarp. The fourth vascular bundle is situated on the dorsal side of the grain and transports organic solutes and minerals to the endosperm. In the portion of aleurone adjacent to the vascular bundles, there are three to five layers of cells responsible for the transfer of nutrients to the starchy endosperm (Hoshikawa, 1989).

Several types of gene have been found to be expressed only or mainly in ETCs (Thompson et al., 2001; Olsen, 2004). Most encode low-molecular-weight, cysteine-rich proteins with hydrophobic signal peptides. Four types of these proteins have been found in maize basal endosperm transfer layers (BETLs). BETL-1 and BETL-3 show sequence homology to defensin-like proteins; BETL-2 has no homologous sequences; BETL-4 has some homology to the Bowman–Birk family of α-amylase/trypsin inhibitors (Hueros et al., 1995, 1999b). As defensins and low-molecular-weight trypsin inhibitors have been shown to inhibit the growth of fungi and bacteria (Broekaert et al., 1997), BETL proteins may help protect the grain from infection. Defensins can also alter the permeability of fungal plasma membranes, and hence may act as regulators of transport through the plasmalemma (Thompson et al., 2001).

The BETL-1 promoter was isolated and used to direct β-glucuronidase (GUS) gene expression in transgenic maize. GUS activity was found exclusively in BETL cells (Hueros et al., 1999a,b). Expression of BETL proteins is greatly reduced in the maize rgf1 (reducing grain filling1) mutant, which also shows decreased uptake of sugars in endosperm cells at 5–10 days after pollination (DAP) (Maitz et al., 2000). The identification and characterization of homologues of BETL genes in other grasses have not been reported.

One further class of ETC genes, encoding low-molecular-weight, cysteine-rich proteins, was identified in the barley transfer cell domain of the endosperm coenocyte. This cell type gives rise to the cells that differentiate into transfer cells (Doan et al., 1996). The gene was designated Endosperm 1 (END1). The expression activities of the barley gene HvEND1 and its orthologue from wheat were studied using in situ hybridization (Doan et al., 1996; Drea et al., 2005) to show that END1 is expressed in the coenocyte above the nucellar projection during the free-nuclear division stage. After cellularization, END1 transcripts accumulated mainly in the ventral endosperm over the nucellar projection, but, from 8 DAP, a low level of expression was also detected in the modified aleurone and the neighbouring starchy endosperm (Doan et al., 1996).

Several other ETC genes have also been identified and partially characterized. These include ZmTCRR-1 (Muniz et al., 2006), the transcription factor ZmMRP-1 (Gomez et al., 2002) and meg1 (maternally expressed gene1) (Gutierrez-Marcos et al., 2004) from maize.

In this study, the promoters of two rice genes, designated OsPR602 and OsPR9a, were isolated and characterized. They were cloned into the pMDC164 plant transformation vector upstream of the GUS reporter gene, and the activity of the promoters was analysed using stable Agrobacterium-mediated transformation of rice and barley. The spatial and temporal activities of the promoters were studied using whole-mount and histological analysis. It was shown that OsPR602 and OsPR9a promoters lead to GUS expression in ETCs of developing caryopses in both rice and barley. However, the promoters exhibit less specific expression in rice than in barley, as rice shows expression in certain flower tissues shortly before pollination.


Identification and isolation of promoter regions of OsPR602 and OsPR9a

A cDNA library was prepared from the liquid endosperm of wheat at 3–6 DAP (Lopato et al., 2006). Inserts from 100 randomly selected clones were sequenced, and several cDNAs encoding full-length, low-molecular-weight, cysteine-rich proteins were identified. One such clone, designated TaPR60 (Accession Number EU264062), showed protein sequence identity with HvEND1 (Doan et al., 1996) (Figure 1a). Another cDNA, designated TaPR9 (Accession Number EU264058), encoded a short protein with no close homologues in the databases, although the position of some cysteines was similar to the position of cysteines in BETL-3, a defensin-like protein from maize (Hueros et al., 1995) (Figure 1b).

Figure 1.

Multiple alignments of protein sequences of OsPR9a to TaPR9 (a) and OsPR602 to TaPR60 and HvEND1 (b). Identical amino acids are shown in black boxes; similar amino acids are shown in grey boxes.

Northern blot hybridization suggested that both TaPR9 and TaPR60 mRNAs were transcribed and accumulated at high levels in the developing grain between 6 and 10 DAP, but were not detectable in any other tissues tested (Figure 2a). Quantitative real-time polymerase chain reaction (PCR) analysis showed that the amount of TaPR9 mRNA started to increase in the grains at 6–7 DAP, reached a maximum at 8–10 DAP and remained present until 16–20 DAP (Figure 2b). The level of TaPR60 transcripts was up to 100-fold higher than that of TaPR9 transcripts. These were first seen at 3 DAP, reached a maximum expression level at 7–8 DAP and were not detected at 18–20 DAP. At 5 DAP, the expression of both TaPR9 and TaPR60 was found mainly in the liquid fraction of the endosperm (Figure 2b). As the objective of the research was to isolate cereal promoters that showed specific expression during early endosperm development, the rice genome sequence was used to identify the promoter regions. The amino acid sequences of TaPR60 and TaPR9 allowed the identification of rice homologues, designated as OsPR602 and OsPR9a, respectively, in expressed sequence tag (EST) databases. ESTs for OsPR602 originated from cDNA libraries prepared from the panicle or pistil a short time after flowering [Institute for Genomic Research (TIGR) libraries #ILF, #68F, #IL6, #IL0, #IKS, #IJM]. The EST for OsPR9a was represented by a singleton in a library prepared from rice immature seed (TIGR library #OS36). Contigs containing full-length coding sequences were found for both OsPR602 and OsPR9a [Accession Numbers CA767165 (EST), EU264061 (gene) and CA760707 (EST), EU264060 (gene), respectively]. Quantitative real-time PCR analysis showed that the amount of OsPR602 mRNA started to increase in the grains at 2–5 DAP, with a very small amount remaining detectable after 11 DAP. A very low copy number of OsPR602 mRNA was also found in pre-anthesis panicles (Figure 2b). OsPR9a mRNA was detected in pre-anthesis panicles and in panicles between 2 and 11 DAP, but was not detectable at 12–18 DAP. No expression of either gene was detected in the other tissues tested (Figure 3). The level of OsPR602 transcripts in panicles was up to several 100-fold higher than that of OsPR9a transcripts.

Figure 2.

Expression of TaPR60 and TaPR9 in different wheat tissues shown by Northern hybridization (a) and quantitative real-time polymerase chain reaction (Q-PCR) (b). Q-PCR was performed using four independent replicates for each tissue. DAP, days after pollination.

Figure 3.

Expression of OsPR9a and OsPR602 in different rice tissues, as shown by quantitative real-time polymerase chain reaction. DAP, days after pollination.

The nucleotide sequences of the two EST contigs were used to identify the translational start site of the genes in the rice genomic database. DNA fragments of 2808 and 1730 bp, upstream of the translational start site of OsPR602 and OsPR9a, respectively, were isolated by PCR using rice (Oryza sativa ssp. japonica cv. Nipponbare) genomic DNA as template. These promoter sequences were then cloned into the plant transformation vector pMDC164 (Curtis and Grossniklaus, 2003) to provide the transcriptional GUS fusion promoter constructs designated as pMDC164-OsPR602 and pMDC164-OsPR9a.

Computer analysis for putative cis-elements in the promoters of OsPR602 and OsPR9a

Computer analysis of the OsPR602 and OsPR9a promoters using place software (http://www.dna.affrc.go.jp/PLACE/signalscan.html) revealed several putative cis-elements which could be involved in endosperm-specific activation. These included: the GCN4-like motif (GLM; 5′-ATGAG/CTCAT-3′), recognized by bZIP proteins of the Opaque2 subfamily (Albani et al., 1997; Wu et al., 1998; Conlan et al., 1999; Onate et al., 1999; Vicente-Carbajosa and Carbonero, 2005); the prolamin box (PB; 5′-TGTAAAG-3′) recognized by one zinc finger (DOF) class of transcription factors (Mena et al., 1998; Lijavetzky et al., 2003; Yanagisawa, 2004); and a motif for the R2R3 subclass of MYB transcription factors (5′-AAC/TA-3′) (Suzuki et al., 1998; Diaz et al., 2002). It has been demonstrated previously that DOF transcriptional factor(s) interact with R2R3MYB (GAMYB) in vivo to activate endosperm-specific promoters (Diaz et al., 2005).

The recently discovered cis-element from the BETL-1 promoter (Barrero et al., 2006), which comprises two -TATCTC- repeats and specifically interacts with ZmMRP-1, was not identified in either the OsPR602 or OsPR9a promoter. Two single -TATCTC- sequences were found at 1189 and 1677 bp upstream of the translational start site in the OsPR602 promoter. However, no confirmation of the role of predicted cis-elements in the activation of the OsPR602 or OsPR9a promoter was undertaken.

Generation of transgenic plants

pMDC164-OsPR602 and pMDC164-OsPR9a constructs were transformed into Agrobacterium tumefaciens strain AGL1, and the presence of plasmid in selected colonies was confirmed by PCR using specific primers. Transformed Agrobacterium was subsequently used to introduce constructs into rice and barley. The integration of promoter:GUS fusions in transgenic plants was confirmed by either PCR (for rice) or Southern blot hybridization (for barley) (Table 1). Southern blot hybridization gave additional data showing the number of inserts in transgenic lines of barley. All GUS-expressing lines showed the same spatial and temporal patterns of expression after GUS staining. However, the level of expression differed.

Table 1.  Information about T0 transgenic lines
 Transgenic line
OsPR602:GUS in transgenic riceOsPR602:GUS in transgenic barleyOsPR9a:GUS in transgenic riceOsPR9a:GUS in transgenic barley
  1. GUS, β-glucuronidase.

Number of putative transgenic lines selected on antibiotic166119
Number of transgenic lines confirmed by polymerase chain reaction and Southern blot hybridization166119
Number of transgene insertionsN/ALines 1, 2, 4 and 5 have single-copy insertions. Line 3 has two copies. Line 6 contains three copiesN/ALines 1, 4, 6, 7 and 9 have single-copy insertions. Line 2 has three copies. Lines 5 and 8 contain two copies. Line 3 has seven copies
Number of transgenic lines with GUS expression6: lines 3, 5, 8, 9, 12 and 16; line 8 has the strongest GUS expression6: lines 1–6; line 1 has the strongest GUS expression5: lines 2, 3, 8, 9 and 113: lines 2, 5 and 8; line 2 has the strongest GUS expression

Sixteen rice lines transformed with pMDC164-OsPR602 were selected on hygromycin, and the integration of the transgene was confirmed by PCR in all lines. However, GUS activity was detected in only six lines. Six from six selected barley plants transformed with pMDC164-OsPR602 were positive in staining for GUS activity. Lines 1, 2, 4 and 5 had single-copy insertions, whereas two copies of the transgene were found in line 3 and at least three copies were integrated in line 6. GUS analyses were performed with all six T0 lines using untransformed wild-type barley and/or barley transformed with empty vector plasmid as the negative control. No differences were found between wild-type plants and plants transformed with empty vector. The profile of OsPR602 promoter activity was identical in all six independent T0 lines. The highest level of GUS activity was detected in line 1.

Integration of pMDC164-OsPR9a was confirmed by PCR for 11 transgenic rice lines. However, only five T0 lines exhibited GUS activity. The pattern of GUS expression was the same in all five lines. Nine transgenic barley lines with the same construct were selected on hygromycin, and integration of the transgene was confirmed in all selected lines by Southern blot hybridization (Table 1). Lines 1, 4, 6, 7 and 9 contained single-copy insertions. Two to three copies were integrated in lines 2, 5 and 8; line 3 contained more than five copies of the transgene in the genome. However, only three transgenic lines (2, 5 and 8) showed GUS activity. The strongest GUS expression was found in line 2.

The numbers of transgenic lines, transgene copy number and relative strength of expression are summarized in Table 1. All T0 lines, and T1 progeny derived from at least two independently transformed T0 transgenic lines, were used for analysis.

Spatial and temporal control of GUS expression by the OsPR602 promoter in transgenic rice and barley

In the transgenic rice lines transformed with pMDC164-OsPR602, GUS activity was detected in the pistils and some flower tissues at anthesis. In flowers, it was found in the rachilla and vascular bundles of the lemma and palea, but was not detected in the anther and ovary (Figure 4a, A2). No GUS expression was observed in the leaf blade, sheath, culm, auricle, ligule, root or rachis. In the T1 caryopsis, the OsPR602 promoter was found to be switched on at 7 DAP. No GUS staining was detected in any grain tissues before 7 DAP (data not shown). In the grain collected at 15 DAP, GUS activity was clearly observed towards the dorsal side of the grain (Figures 3 and 4), in the aleurone transfer cell layers and the adjacent starchy cells. The aleurone transfer cell layers retained GUS activity until 24 DAP. No GUS activity was detected in the grain of control plants at any stage (Figure 4a, A3).

Figure 4.

Spatial and temporal β-glucuronidase (GUS) expression in rice (a) and barley (b) directed by the OsPR602 promoter. GUS activity in transgenic rice detected in flower (A2) and towards the dorsal side of the grain at 15 days after pollination (DAP) (A4), but not in control flower (A1) and grain (A3). Bars, 1 mm. Histochemical GUS assay in a transverse section of a T1 caryopsis at 9 DAP (A5, A6) and 13 DAP (A7, A8); longitudinal section of rice caryopsis at 15 DAP (A9, A10). Bars: A5, A7 and A9, 800 µm; A6, A8 and A10, 100 µm. GUS activity in barley detected in hand-cut longitudinal (B2) and transverse (B4) sections of transgenic lines, but not in wild-type (B1, B3) caryopsis at 10 DAP. Histochemical GUS assay counter-stained with safranin in a 10 µm thick transverse section of transgenic barley caryopsis at 10 DAP (B5), longitudinal section of wild-type caryopsis at 10 (B7), 16 (B9), 26 (B11) and 30 DAP (B13), and transgenic barley caryopsis at 10 (B6), 16 (B8), 23 (B10) and 30 DAP (B12). Arrows show GUS-stained tissues. AL, aleurone; DVB, dorsal vascular bundle; ETC, endosperm transfer cell; mp, maternal pericarp; NE, nucellar epidermis; NP, nucellar projection; OVT, ovular vascular trace; PS, pigment strand; Ra, rachilla; SEC, starchy endosperm cells; VB, vascular bundle; VT, vascular tissue. Bars, 1 mm.

Following the GUS assay, the samples of rice grain were embedded in paraffin wax, sectioned and counter-stained with safranin orange to achieve high contrast and resolution of cell morphology. The micrographs A5–A10 in Figure 4a demonstrate the distribution of GUS expression during early grain filling. The ovular vascular traces formed a large tissue mass down the dorsal side of the grain, which serves as a path linking the maternal vascular tissue to the inner side of the nucellus. In the grain at 9 DAP, GUS was strongly expressed in the ovular vascular cells, the adjacent pigment strand and testa, the three to four layers of aleurone transfer cells and the adjacent starchy cells (Figure 4a, A5 and A6). GUS activity gradually declined and could not be detected in the ovular vascular trace or pigment strand at 13 DAP, but remained in other transfer tissues (Figure 4a, A7 and A8). Figure 4a (A9 and A10) illustrates GUS activity in the posterior pole of a caryopsis at 18 DAP.

The whole-mount GUS analyses of grain from transgenic barley plants transformed with pMDC164-PR602 revealed a spatial pattern similar to that in rice grain; GUS staining could be seen as two parallel lines in the ventral groove zone of transgenic T1 caryopses. The activity of the promoter was detected in the ETCs and a few layers of adjacent starchy endosperm cells (Figure 4b, B2 and B4). No GUS activity was observed in control grain (Figure 4b, B1 and B3). No GUS activity was found in transgenic caryopses before 10 DAP. GUS expression in ETCs of barley was observed throughout grain maturation. It started at 10 DAP and finished at 30 DAP (Figure 4b, B6 and B12). In one of the six barley lines (line 5), GUS expression was weaker than in the other lines, but the OsPR602 promoter was still active at 35 DAP and GUS expression was observed in two to three layers of transfer cells.

A transverse section (B5) through the middle of the barley caryopsis (line 1) at 10 DAP is shown in Figure 4b. The identities of the aleurone layers at the endosperm adaxial axis and modified aleurone cells near the abaxial side were already established at this stage, and could be distinguished from the central starchy endosperm cells. The GUS expression is clearly seen in the transfer cell layer and adjacent starchy endosperm.

The spatial activation patterns of the OsPR602 promoter in barley matched well with the data obtained for this promoter in rice, except that the ventral pigment strand and the vascular bundles down the chalaza pad in barley did not show GUS expression. GUS activity was not detected in the flowers or any vegetative tissues of the transgenic barley plants.

The temporal expression patterns of OsPR602 in rice differed slightly from those in barley. In rice, GUS expression started 3 days earlier and stopped 6 days earlier than in barley.

Spatial and temporal control of GUS expression by the OsPR9a promoter in transgenic rice and barley

The whole-mount GUS staining of a T1 caryopsis of rice at 26 DAP is shown in Figure 5a (A2 and A3). OsPR9a activity was present in the rachilla, anthers and mature pollen shortly before pollination (Figure 5a, A1). No activity in the anthers and mature pollen was observed for the OsPR602 promoter in either rice or barley. No activity of the OsPR9a promoter was detected in the vascular tissues of the lemma and palea or in other plant tissues. The strongest OsPR9a promoter activity in rice was found in aleurone transfer cell layers between 5 and 26 DAP. After 26 DAP, the activity diminished, and no GUS activity was detected later than 35 DAP.

Figure 5.

Analysis of β-glucuronidase (GUS) expression in rice (a) and barley (b) directed by the OsPR9a promoter. GUS activity detected in flowers at anthesis (A1), and in uncut (A2) and hand-cut (A3) caryopsis of rice at 26 days after pollination (DAP). Hand-cut transverse and longitudinal sections of wild-type caryopsis at 18 and 10 DAP (B1, B2) and transgenic barley caryopsis at 20 DAP (B3, B4). Histochemical GUS assay and safranin counter-staining in transverse section of wild-type caryopsis at 20 DAP (B5), and in transverse (B6, B8) and longitudinal (B7) sections of transgenic barley caryopsis at 20 DAP. Arrows show GUS-stained tissues. An, anthers; ETC, endosperm transfer cell; Ra, rachilla. Bars, 1 mm (except B8, 0.25 mm).

In the transgenic T1 barley caryopsis, the OsPR9a promoter was active in the ETC layers and the adjacent starchy endosperm cells (Figure 5b, B3, B4, B6 and B7). This resembles the pattern observed for the OsPR602 promoter in barley. The promoter was active at 9 DAP and activity was detectable until 35 DAP. No GUS staining was found in the pedicels and anthers of transgenic barley plants before or at anthesis. No GUS activity was detected in vegetative plant tissues.

Although transgenic lines were generated with different levels of GUS expression, on the whole, the OsPR9a promoter was weaker than the OsPR602 promoter.


The manipulation of grain size, shape and composition depends on access to seed-specific promoters, in particular, promoters that can drive transgene expression during early endosperm development. An important group of promoters are those that are active in the ETC. Several ETC-specific genes have been described in the literature (Hueros et al., 1995, 1999b; Doan et al., 1996; Gomez et al., 2002; Gutierrez-Marcos et al., 2004; Muniz et al., 2006), but the cloning and analysis of only one ETC-specific promoter, from BETL-1 of maize, has been reported (Hueros et al., 1999a). The activity of the BETL-1 promoter was tested in maize, but not in other plant species. A number of sequences were identified encoding cysteine-rich proteins similar to the products of the END1 and BETL genes in mRNA from the liquid fraction of the syncytial endosperm of wheat. Northern blots and quantitative PCR confirmed that transcription of two of the cloned genes, designated TaPR60 and TaPR9, is specifically activated in grain at the end of the cellularization phase, and that the genes are expressed until the end of grain maturation (Figure 2). The cloning of wheat promoters is still a slow process because of the absence of a genome sequence. In contrast, rice genome databases are available, but ETC-specific genes from rice have not been reported. Therefore, ETC-specific promoters from rice were identified and characterized in rice and barley to evaluate their broad applicability in cereal biotechnology. Protein sequences of TaPR60 and TaPR9 were used to identify their rice homologues, OsPR602 and OsPR9a, respectively, in rice EST databases. The alignment of the protein sequences derived from the rice ESTs with proteins from other plants is shown in Figure 1. Quantitative PCR analysis of OsPR602 and OsPR9a expression revealed their specific expression in panicles before pollination and at 11 DAP. This is very similar to the data obtained for wheat homologues, except that, in rice, expression was detected in pre-anthesis panicles (flowers) and expression in both panicles and developing grain strongly decreased after 11 DAP.

The nucleotide sequences of the coding regions of OsPR602 and OsPR9a were subsequently used to identify the translation start sites of their genes. Regulatory sequences containing promoters and 5′-untranslated regions (5′-UTRs) were cloned, and the activity of the promoters was analysed using the stable transformation of rice and barley plants with promoter:GUS gene fusion constructs.

Surprisingly, the OsPR602 promoter in rice was active not only in ETCs, but also in transfer cells of maternal tissue. Indeed, the spatial pattern was similar, but not identical, to the previously described expression pattern of END1 in barley and wheat, as shown by in situ hybridization analyses (Doan et al., 1996; Drea et al., 2005). In rice, the pericarp has four vascular bundles, one on each of the ventral and dorsal sides, and one on each side face. The vascular bundle on the dorsal side is the widest. It contains many conducting elements and extends to the upper end of the grain (Hoshikawa, 1989; Krishnan and Dayanandan, 2003). The activity of OsPR602 was found in the dorsal vascular bundle, but was not detected in the other three narrow vascular bundles of the grain. In the early stages (7–12 DAP) of rice endosperm differentiation, the OsPR602 promoter was active in the ovular vascular trace, the adjacent pigment strand, the adjacent nucellar epidermis, the three to four layers of ETCs and the adjacent starchy endosperm cells. Although the activity of the OsPR602 promoter was not detected in the remaining three narrow vascular bundles of the pericarp, it was observed in vascular tissues of the lemma and palea shortly before anthesis (Figure 4a, A2). The lemma has five swollen vertical ridges with a vascular bundle running through each. The palea has three vascular bundles – one at the centre of the dorsal surface and one on both sides (Hoshikawa, 1989). GUS activity was found in all eight vascular bundles at anthesis, but disappeared at the early stages of grain development. The OsPR9a promoter showed rather different patterns of GUS expression in rice flowers. It was active in the rachilla, anthers and the mature pollen shortly before pollination, but showed no activity in the ovary or vascular tissue of the lemma and palea (Figure 5a, A1). GUS expression was detected in aleurone transfer cells and adjacent layers of starchy cells, but was not found in transfer cells of maternal tissues of grain (data not shown).

The activity of the OsPR602 promoter in rice seeds was similar to that of the barley asi promoter in developing transgenic barley seeds, for which gfp expression was observed specifically in the pericarp, vascular tissue, nucellar projection cells and ETCs (Furtado et al., 2003). However, in barley, both the OsPR602 and OsPR9a promoters are active only in the ETCs and adjacent starchy endosperm cells. GUS expression could not be detected in either transfer cells of maternal tissue or in flowers before pollination. A transcription factor ZmMRP-1, which can specifically activate promoters in ETCs, has been identified in maize (Gomez et al., 2002). The expression pattern of GUS for OsPR602 and OsPR9a in flowers is clearly different from the situation in maize, and could mean that the transcription regulator from rice has a broader specificity than its maize orthologue. However, it is more probable that the OsPR602 and OsPR9a promoters are activated by two different transcription factors. If this is the case, it is probable that only the cis-element responsible for expression in ETCs is conserved between the two species. In maize, ZmMRP-1 can activate the BETL-1 promoter in vivo and interact in vitro with the specific cis-element -TATCTCTATCTC- from the promoter (Barrero et al., 2006). However, this element was not found in the promoters of OsPR602 and OsPR9a.

Quantitative PCR data for the expression of OsPR602 and OsPR9a genes, data obtained for their orthologues from wheat, as well as data published for END1 and BETL (Hueros et al., 1999b), indicate that transcriptional activation of the promoters of these genes takes place at the middle of endosperm cellularization at 3–6 DAP, reaches a maximum activity at the end of endosperm cellularization at 8–10 DAP, and the promoter becomes inactive close to the middle of endosperm maturation at 20 DAP. A 3–5-day difference was found in the activation of the same promoter in rice and barley. This can be partially explained by the observation that endosperm cellularization is complete at 6 DAP in rice, wheat and maize, whereas, in barley, it ends at 8 DAP (Olsen, 2001). If it is assumed that the transcriptional activation of OsPR602 and OsPR9a starts at the middle of the cellularization phase of endosperm development, this could explain the temporal difference in the initiation time of activation of both promoters in barley relative to rice. It is interesting that a gradual increase in GUS activity from the time of promoter activation was not seen. Strong GUS activity appeared suddenly at 9–10 DAP.

Northern hybridization identified the weak expression of END1 in the barley endosperm coenocyte at 5–7 DAP, with strong expression from 9 DAP until 30 DAP (Doan et al., 1996). In contrast, mRNA levels of BETL genes in maize grain deceased at 15–20 DAP (Hueros et al., 1999b). The quantitative PCR data obtained for the orthologous wheat genes matched the maize BETL result. GUS expression in ETCs of barley was observed during the whole phase of grain maturation, long after the expected end of the transcriptional activity of OsPR602 and OsPR9a promoters. Hence, the detection of GUS at 30–35 DAP is most probably a result of the stability of mRNA of the transgene and GUS protein in ETC cells, rather than sustained expression. In most cases, strong expression of GUS was observed until 20 DAP and it then slowly diminished, suggesting an end of transcriptional activity followed by slow degradation of mRNA and protein. This implies that the OsPR602 and OsPR9a promoters can be used to express transgenes until approximately 20 DAP, but thereafter the amount of protein produced may be dependent on mRNA and protein degradation.

The activity of the promoters was not quantified, as their specific expression is restricted to a relatively small number of grain cells, making quantification problematic. The strength of GUS activity is also dependent on the number of copies and position of transgene insertions in the genome. Nevertheless, the OsPR602 promoter appears to be several-fold stronger than OsPR9a in both rice and barley.

This work suggests that the OsPR602 and OsPR9a promoters are suitable for ETC-specific expression of genes of interest in barley. As wheat and barley are closely related species, similar expression patterns can be expected. However, for the modulation of nutrient uptake in rice, the OsPR602 promoter would be preferable for transgene expression, as the OsPR9a promoter is also active in non-target organs, e.g. anthers and pollen.

These promoters offer the potential to improve the grain quality by modifying the quality and quantity of nutrient uptake through the ETC. Improvements in disease resistance may also be gained by enhancing the efficacy of this tissue as a barrier to pathogen movement from maternal tissue to the developing endosperm.

Experimental procedures

Promoter cloning and plasmid construction

Promoters and 5′-UTRs of OsPR602 and OsPR9a were amplified from rice cv. Nipponbare genomic DNA using the primers shown in Table 2. The proof-reading DNA polymerase Pfx (Invitrogen, Mulgrave, VIC, Australia) was used to minimize PCR-induced mutations and to produce blunt-end fragments for directional cloning into the pENTR-D-TOPO vector (Invitrogen). The cloned inserts were sequenced and subcloned into the pMDC164 vector (Curtis and Grossniklaus, 2003) using recombination cloning. The resulting constructs were designated as pMDC164-PR602 and pMDC164-PR9a, and were transformed into Agrobacterium tumefaciens strain AGL1 by electroporation. The presence of plasmids in selected clones was confirmed by PCR using promoter-specific primers (Table 2).

Table 2.  A list of primer sequences used in reverse transcriptase-polymerase chain reaction (RT-PCR), quantitative polymerase chain reaction (Q-PCR) and promoter cloning
GeneForward primerReverse primer

Southern blot hybridization and quantitative PCR analysis

Transgene integration into rice plants was confirmed by PCR using GUS-specific primers (Table 2) with genomic DNA from selected transgenic lines. Transgene integration in barley plants was confirmed by Southern blot hybridization. Genomic DNA from selected barley lines was digested with EcoRV, which cuts the T-DNA region once upstream of the selectable marker gene. The Southern blot was probed with the coding sequence of the hygromycin phosphotransferase selectable marker gene.

Quantitative PCR was carried out according to Burton et al. (2004) using the primer combinations shown in Table 2.

Rice and barley transformation

The constructs pMDC164-PR602 and pMDC164-PR9a were transformed into rice and barley using Agrobacterium-mediated transformation and the method developed by Tingay et al. (1997) and modified by Matthews et al. (2001). Rice (Oryza sativa L. ssp. Japonica cv. Nipponbare) and barley (Hordeum vulgare L. cv. Golden Promise) were used as donor plants. To prevent the generation of multiple plants derived from the same transformation event, individual regenerants were selected from independently transformed and cultured pieces of callus.

The promoter activity was tested in T0 and T1 plants (T1 and T2 grain).

Histochemical and histological GUS assays

GUS activity in transgenic barley plants was analysed by histochemical staining using the chromogenic substrate 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid (X-Gluc) (Bio Vectra, Oxford, CT, USA), as described by Hull and Devic (1995). Different plant organs, whole grain and grain sections of different ages were immersed in a 1 mm X-Gluc solution in 100 mm sodium phosphate, pH 7.0, 10 mm sodium ethylenediaminetetraacetate, 2 mm FeK3(CN)6, 2 mm K4Fe(CN)6 and 0.1% Triton X-100. After vacuum infiltration at ~88 043 Pa for 20 min, the samples were incubated at 37 °C until satisfactory staining was observed. Tissues were incubated in 20%, 35% and 50% ethanol, fixed in FAA (50% ethanol, 5% acetic acid and 4% formaldehyde) and cleared in 70% ethanol. The whole-mount grains were then observed under a dissecting microscope, and photographs were taken using a Leica digital camera (Leica, North Ryde, NSW, Australia). The grain samples were further dehydrated in 80% and 90% ethanol, and then incubated in 95% ethanol with 0.05% of the counter-stain safranin orange for contrast. For longitudinal and transverse sectioning, the tissues were embedded in paraffin wax, sectioned at 9–12 µm, deparaffinized and mounted in DPX mountant (Fluka Biochemika, Buchs, Switzerland), as described in Weigel and Glazebrook (2002). The specimens were observed under a compound microscope and photographed.


We thank Professor U. Grossniklaus for providing us with a collection of pMDC vectors, Dr R. Burton for cDNA prepared from rice tissues, and Ursula Langridge for assistance with growing plants in the glasshouse.

This work was supported by the Australian Grain Research and Development Corporation (grant no. UA00083 to P.L.), International Postgraduate Research Scholarship (IPRS) and Adelaide University Scholarship (AUS).