Expression of the OsPI1 gene, cloned from rice roots using cDNA microarray, rapidly responds to phosphorus status


Author for correspondence: Jun Wasaki Tel: +81 11 706 3845 Fax: +81 11 706 3845 Email:


  • • Plants have developed several methods of adapting to conditions of low phosphorus (P). However, details of the regulation of the gene expression system that responds to P status of plants is unknown. Here, a phosphorus limitation inducible novel gene was isolated and characterized to provide further information of plant adaptation to low P.
  • • Rice plants (Oryza sativa) were grown hydroponically with or without P. A novel gene was isolated by cDNA microarray analysis and designated as OsPI1 (Oryza sativaPhosphate-limitation Inducible Gene 1). mRNA accumulation was examined by Northern blot and quantitative real time PCR.
  • • The OsPI1 gene was rapidly induced by phosphate starvation in both shoots and roots. When phosphate was supplied to phosphate-deficient plants, the OsPI1 transcripts rapidly disappeared. OsPI1 cDNA consisted of 375 bp and contained several small open reading frames (ORFs). The OsPI1 gene shows the same characteristics as the TPSI1/Mt4 family (the phosphate starvation inducible novel gene family).
  • • It is suggested that OsPI1 acts as riboregulator, that is, it binds with other molecules under phosphate starvation and regulates their function.

Oryza sativaphosphate-limitation inducible gene 1


acid phosphatase


open reading frame


expression sequence tag


inorganic phosphate


Plants have developed several methods for adapting to low P conditions, and these methods can be classified into two strategies. The first strategy is the efficient acquisition of P from the rhizosphere. When plants are grown under P deficient conditions, their roots secrete acid phosphatase (APase) and organic acids to liberate inorganic phosphate (Pi) from organic compounds and sparingly soluble Pi compounds, respectively (Gardner et al., 1983; Tadano & Sakai, 1991). Several related genes were isolated and characterized, an example of which is LASAP2 encoding APase secreted from white lupin roots (Wasaki et al., 2000). It has also been reported that gene expression of high affinity Pi transporters increases in the roots of P deficient plants (Mucchal et al., 1996; Mucchal & Raghothama, 1999). The second strategy is the efficient utilization of absorbed P in plant tissues. The bypass pathway contributing to saving Pi pool in plant tissues rather than producing energy has been published elsewhere (Duff et al., 1989, 1991). Production of ribonucleases also increases under P deficiency, and its function is to mobilize Pi from internal RNA pools (Löffler et al., 1992; Green, 1994). Goldstein et al. (1989) proposed that plants also have a ‘pho regulon’ that regulates the responses of P deficiency, such as in yeast (Bergman et al., 1986) and Escherichia coli (Torriani & Ludtke, 1985). However, details of the regulation of the gene expression system that responds to P status of plants is unknown.

Several novel genes, which were induced under low P conditions, were also isolated and characterized. The first paper on phosphate starvation inducible novel genes described TPSI1 cloned from tomato (Liu et al., 1997). Subsequently, other related genes were isolated and characterized: Mt4 from Medicago truncatula (Burleigh & Harrison, 1997, 1998), At4 (Burleigh & Harrison, 1999) and AtIPS1 (Martín et al., 2000) from Arabidopsis thaliana. They were designated as the TPSI1/Mt4 family because of their common characteristics. They are up-regulated by P limitation and rapidly down-regulated by supply of P or colonization of mycorrhizal fungi. This fact indicates the possibility that they have an important role on the strategies for low P adaptation, although their physiological function is not still clarified.

Burleigh & Harrison (1998) conducted a split-root experiment, that is the roots were split and grown with or without Pi, and investigated the Pi concentration and accumulation of Mt4 mRNA. They found that Mt4 accumulation in both –P and +P split-roots remained at the same low level as the +P nonsplit roots. On the other hand, because the Pi concentration of the –P split-root was about half of the +P split-root, they suggested that the signal of P status was transmitted from +P split-root to –P split root. It was also indicated that there was a possibility that long-distance signals from the shoots to the roots were involved in the expression of AtIPS1, which is also a member of TPSI1/Mt4 (Martín et al., 2000). However, it is confirmed that the external status of Pi is affected in the expression of the TPSI1/Mt4 family. Varadarajan et al. (2002) estimated the effects of phosphite (Phi), an analog of Pi, on the expression of P-deficient inducible genes of tomato and Arabidopsis. Phi down-regulated the P deficient inducible genes of Arabidopsis over a short period (Varadarajan et al., 2002). It is appropriate that Phi influences a signal transduction pathway involving Pi in the roots. That is to say, it suggests the existence of a molecular mechanism, which is able to receive not only a long-distance signal, but also the external or internal Pi concentration, and respond during earlier stress. To clarify the molecular mechanism, investigation of expression of early responsive genes is important.

In this study, we investigated the transcriptional alteration of P-deficient rice roots using cDNA microarray. A novel gene was isolated and designated as OsPI1 (Oryza sativaphosphate-limitation inducible gene 1). It was shown that the OsPI1 gene responded to short-term P deficiency and had the same features as the TPSI1/Mt4 family. The role of the OsPI1 gene under P-deficient conditions is also discussed.

Materials and Methods

Plant material

Rice (Oryza sativa L. ssp. japonica cv. Michikogane) seeds were sterilized using 70% ethanol and germinated in 1 mm CaCl2 for 36 h with aeration under dark conditions. The seeds were precultured in 1 mm CaCl2 for 4 d to diminish the level of P in the plant in order to observe the effect of P status in the plant. After preculture, the plants were transferred to nutrient solutions either with or without P. The composition of the nutrient solution was as follows: 0.83 mm NH4NO3, 0 µm (−P solution) or 32 µm (+P solution) NaH2PO4, 0.38 mm KCl, 0.19 mm K2SO4, 0.75 mm CaCl2, 0.82 mm MgSO4, 36 µm Fe (III)-EDTA, 9.1 µm MnSO4, 46 µm H3BO3, 3.1 µm ZnSO4, 0.16 µm CuSO4 and 7.4 nm (NH4)6Mo7O24. The pH was adjusted to 5.2 using 0.1 m NaOH, and plants were grown in a growth chamber at 28°C with 24 h light (c. 130 µmol photons m−2 s−1 at leaf surface and 50% rh).

CDNA microarray analysis Half of the plants grown on the +P solution were transferred to the −P solution after 8 d, and all plants were collected the following day. The solution was changed every 24 h. The treatments were defined as the +P treatment, short-term −P treatment (transferred from +P solution after 8 d), and long-term −P treatment (continuously grown on the −P solution). The roots were washed in deionized water and the f. wt of shoots and roots were measured, then the roots and shoots were frozen immediately using liquid nitrogen. The samples were stored at −80°C.

Investigating the growth, P uptake and OsPI1 mRNA accumulation Rice seedlings were precultured in a nutrient solution containing 32 µm NaH2PO4 and transferred to the +P or −P treatments. The solution was changed every 24 and 12 h during preculture and treatment, respectively. The plants were divided into shoots and roots and collected at 12, 36, 72, 84, 108 and 144 h after transference to the treatment solutions. At 72 h, half the −P plants were transferred to the +P solution (PR treatment) and collected using the same method at 84, 108 and 144 h after the start of the treatment. Shoots and roots were separated and immediately frozen by liquid nitrogen and stored at −80°C for further analysis. For the measurement of P, a frozen sample was lyophilized, ground, then a sample of approx. 50-mg was digested with H2SO4-H2O2. The P concentration in the digested solution was measured using the vanado-molybdate yellow method described by Watanabe et al. (1998). All experiments were replicated three times.

DNA and RNA gel blot analysis

Genomic DNA was extracted from rice leaves using the cetyl trimethyl ammonium bromide (CTAB) method (Murray & Thompson, 1980) and total RNA was extracted using the sodium dodecyl sulfate (SDS)–phenol method (Palmiter, 1974). Seven micrograms of genomic DNA were digested with the restriction enzymes BamHI, EcoRV, HindIII and XbaI, and separated on a 0.8% agarose gel. Twenty micrograms of total RNA were separated on 1.0% agarose gel denatured with formaldehyde, and after electrophoresis the digested DNA and total RNA were transferred to a Hybond™-N+ membrane (Amersham Bioscience, Piscataway, NJ, USA). Probe labeling and hybridization were performed using the Gene Images™ system (Amersham Bioscience) in accordance with the manufacturer's instructions. The blots were hybridized and washed at 65°C under stringent conditions.

Microarray analysis

Total RNA was isolated using the aforementioned method and purified using the CsCl gradient ultracentrifuge method (Sambrook et al., 1989). Forty micrograms of purified RNA were used per array set as targets for microarrays, and labeled with Cy5-dCTP (Amersham Bioscience) using Superscript™ II (Gibco BRL, Rockville, MD, USA). Unlabeled primers and dyes were removed with GFX™ columns (Amersham Bioscience).

Rice cDNA microarrays were prepared by the microarray project in Japan (Kishimoto et al., 2002). The arrays contained 8987 cDNA clones from several organs of rice (Oryza sativa L. ssp. japonica cv. Nipponbare) and were spotted on two slide glasses. Half the cDNAs were spotted on each slide as duplicates and hybridization was performed at 60°C for 4 h using a humidified chamber and an ExpressHyb™ solution (Clontech, Palo Alto, CA, USA). After hybridization, the glasses were washed and dried. The signal intensity of each spot was measured using an array scanner (FLA8000; Fuji Film, Tokyo, Japan) and the scanned signals were analyzed by software from Fuji Film (Array Gauge).

The values measured from each spot were standardized to remove influences among slide glasses caused by unevenness of hybridization and a difference of labeling efficiency. The distribution of the logarithmic value for each spot was similar to normal distribution and the deviation values from all signal intensities were calculated. Each spotted field, consisting of 384 spots, was regarded as a population for normalization. Changes of the expression mass were estimated by the differences of the deviation values between treatments. The normalization and statistical analysis were performed using Microsoft Excel 2001 (Microsoft, Redmond, WA, USA).

Investigation of mRNA accumulation using quantitative real time PCR

Total RNA was isolated using the aforementioned method. The RNA was treated with DNase (RT grade, Nippon Gene, Tokyo, Japan) to digest the contaminated genomic DNA, then reverse transcribed using the 1st Strand cDNA Synthesis Kit for RT-PCR (AMV) (Roche Diagnostics, Basel, Switzerland). The first strand cDNAs were used as templates for quantitative real time PCR using a LightCycler™ System (Roche Diagnostics), and the TaqStart™ antibody (Clontech) was used for the repression of unspecific amplification. The RAc1 gene was selected as a control gene because it was reported that its gene expression was relatively stable among actin isoforms (McElroy et al., 1990). Calibration curves were drawn with two replications based on the specific plasmids. The primer sets and the reaction steps for amplifying OsPI1 and RAc1 fragments were as follows; for amplification of OsPI1 transcripts, opiLC-S primer (sense, 5′- TCC TCT CTA CCC CCA ACA ATG G-3′) and opiLC-A primer (antisense, 5′-TGG AGA AGG AAG ACC TGC CAA A-3′) were used. LightCycler™ was run at 95°C for 0 s, at 65°C for 4 s, at 72°C for 8 s and 83°C for 0 s for 45 cycles with 20°C s−1 of temperate transition rate. For amplification of RAC1 transcripts, ra1LC-S primer (sense, 5′-CTT CAT AGG AAT GGA AGC TGC GGG TA-3′) and ra1LC-A primer (antisense, 5′-CGA CCA CCT TGA TCT TCA TGC TGC TA-3′) were used. LightCycler™ was run at 95°C for 0 s, at 68°C for 5 s, at 72°C for 8 s and 81°C for 0 s for 40 cycles with 20°C s−1 of temperate transition rate.

The amounts of transcripts were shown as relative amounts against the value of the +P roots (Total RNA basis). The amount of OsPI1 transcripts was also shown as the RAc1 basis value, which was divided into the value of RAc1 transcripts (RAc1 basis).

Cloning and sequencing analysis of OsPI1 gene

An EST clone (Accession No. C19881), including partial OsPI1 cDNA, was provided from a DNA bank (National Institute of Agrobiological Sciences, Tsukuba, Japan). The flanking region on the 3′ and 5′ ends were cloned using, respectively, a 3′- and a 5′-Full RACE Core Set (Takara Bio, Otsu, Japan) in accordance with the manufacturer's instructions. cDNA was synthesized from RNA of P-deficient rice roots and used as a template for RACE-PCR. Amplified fragments were subcloned into a pGEM®-T vector (Promega, Madison, WI, USA).

DNA sequencing was performed by the dideoxy chain-terminal method (Sanger et al., 1977) using a Thermo Sequenase™ Cycle Sequencing Kit with 7-deaza-dGTP instead of dGTP (Amersham Bioscience), and the sequence was determined by a DNA sequencer 4000LS (Li-Cor, Lincoln, NE, USA). Homology searches were performed with the blast program (Altschul et al., 1990; Gish & States, 1993) using several databases through the National Center for Biotechnology Information (NCBI). Alignment of the amino acid sequences was performed using the ClustalW program (Thompson et al., 1994) through DDBJ (DNA databank of Japan). Genetyx-Mac software (Genetyx, Tokyo, Japan) was used to find stem–loop structures in the promoter sequence of OsPI1.

Primer extension was performed to determine the transcription start point of the OsPI1 gene according to the DNA sequencer's instruction. A primer set (op1GE-S; 5′-CCA GAT CAT ATG CCA GTG TCT C-3′, op1GE-A; 5′-TGT CGA GAT GGA GAA GGA AGA C-3′) was designed from a sequence of the PAC clone of the rice genome (Accession No. AP003231) containing the complete OsPI1 gene. PCR was performed using the primer set and 10 mg rice genomic DNA as a template. The amplified fragment was subcloned into pGEM®-T vector and used as a template for cycle sequencing of OsPI1 fragments. To sequence and reverse transcribe the OsPI1 specific fragments, an IRD800-labeled primer was designated as op1PE-IRD (5′-IRD800-AGA CCT GCC AAA GTT GTT TAG TTG C-3′) and was obtained from Aloka (Tokyo, Japan). ReverTra Ace® (Toyobo, Osaka, Japan) was used as reverse transcriptase for the reaction of primer extension. Five micrograms of Poly(A) + RNA were isolated from P-deficient rice shoots and used as templete for primer extension experiments.


Screening a novel gene induced under –P conditions using cDNA microarray analysis

Scatter plots were used to compare the relative amount of transcripts spotted on the cDNA microarray between the complete nutrient (+P) and P-deficient nutrient (−P) treatments (Fig. 1). Most scatters were found near the 1 : 1 line, indicating that their expressions were not changed by the −P treatment. The gene spotted on element no. 3526 was most obviously induced by the −P treatment for 9 d (Fig. 1, arrow). A blast search against the sequences of the EST clones (Accession No. C19881 and AU162218) indicated that there are no homologous genes registered on the databases. The novel gene was designated as OsPI1 (Oryza sativaphosphate-limitation inducible gene 1).

Figure 1.

Scatter plots of relative amounts of transcripts analyzed by cDNA microarray. (a), (b) The comparison between +P (X axes) and short-term and long-term −P treatments (Y axes), respectively. The deviation value of signal intensity of each spot is shown. The arrows indicate the OsPI1 (element no. 3526).

To estimate the reproducibility of cDNA microarray analysis, the relative amounts of mRNA for OsPI1 (Element no. 3526) and an actin isozyme (Element no. 5820) analyzed by cDNA microarray were compared with the result of quantitative RT-PCR analysis (Table 1). The actin isozyme, designated as RAc1, has a relatively stable expression under several conditions (McElroy et al., 1990). It was shown by quantitative analysis that mRNA of RAc1 in roots grown under −P for 9 d was about eightfold larger than in the +P roots, while in the −P shoots the amount decreased by about half when compared with +P (Table 1). In both shoots and roots, a notable increase of the amount of OsPI1 mRNA in the −P treatment for 9 d was shown (Table 1). The decrease of OsPI1 accumulation in −P shoots for 24 h was not significantly different.

Table 1.  Relative accumulation of RAcI and OsPI1 mRNAs in leaves and roots. Data of microarray analysis is shown by deviation value. Results of quantitative real time PCR are shown as relative amounts against the value of the +P roots (Total RNA basis). The amount of OsPI1 transcripts was also shown as the RAc1 basis value, which was divided into the value of RAc1 transcripts (RAc1 basis). n.d., not determined
Microarray (Deviation value)Real Time PCR (total RNA basis)Microarray (Deviation value)Real Time PCR (total RNA basis)Real Time PCR (RAc1 basis)
Leaves+Pnd2.97nd  3.56 1.23
−P, 24 hnd2.49nd  2.27 0.923
−P, 9 dnd1.58nd 70.845.1
Roots+P581.040  1.0 1.0
−P, 24 h612.449 21 8.6
−P, 9 d638.06349061

When compared with the results of microarray analysis and individual PCR analysis, the inclines of the expression patterns were similar (Table 1). The difference of the mRNA amount was estimated to be smaller in the microarray analysis than in the RT-PCR analysis and unavoidable cross-hybridization during microarray analysis is believed to have caused the underestimation.

Effects of P status on OsPI1 expression in rice

Rice plants were grown hydroponically with or without P, and dry matter production, total P and accumulation of OsPI1 mRNA were investigated (Fig. 2). There were no differences in d. wts of shoots and roots among the treatments (Fig. 2a,b), and total P concentration in the −P plants decreased continuously (Fig. 2c,d). When P was supplied to the −P plant (PR treatment), the P concentration of the shoots and roots recovered, respectively, by 80% and 83% at 12 h after supply and by 100% and 92% at 72 h after supply (Fig. 2c,d).

Figure 2.

Effects of P limitation and resupplying on growth, P concentration and OsPI1 expression. Open and closed circles indicate −P and +P treatment, respectively. Triangle symbols indicate P resupply treatments. Dry matter production of leaves and roots are shown in (a) and (b), respectively. P concentration of leaves and roots are shown in (c) and (d), respectively. (e) The results of Northern blot analysis. rRNA was also shown as control of electrophoresis. The symbols (+, − and R) indicate +P, −P and PR treatment. P was resupplied at 72 h after emergency in PR treatment.

Fig. 2(e) shows the result of a Northern blot analysis of OsPI1, the transcripts of which increased precisely in shoots and roots grown under the −P conditions. In the −P treatment, the transcripts increased continuously until the end of the treatment (Fig. 2e). OsPI1 mRNA was not detected in the +P shoots and roots (Fig. 2e). When P was supplied to the −P plants, the OsPI1 transcripts rapidly disappeared (Fig. 2e). The amount of transcripts in the −P roots increased at 72 h and later doubled that of the −P in the shoots (Fig. 2e). The major signals were detected between 0.5 and 0.6 kb (Fig. 2e), and a minor band was observed at about 0.3 kb in the −P roots treated for 84 and 144 h (Fig. 2e).

The effect of P depletion over several hours on OsPI1 expression in rice roots was investigated using quantitative RT-PCR (Table 2). A change of the amount of transcripts was not significant after 1 h, but after 4 h the OsPI1 transcripts significantly increased in the −P treatment. This was calculated based on both the total RNA and RAc1 transcript amounts (Table 2).

Table 2.  Change of relative accumulation of OsPI1 mRNA in roots during short-term P limitation
 Relative mRNA accumulation
Total RNA basisRAc1 basis
  1. The data are shown as relative amounts against the value of + P roots. *0.001 < P < 0.01, **P < 0.001.

+P    1 ± 0.040   1 ± 0.105
−P, 1 h0.843 ± 0.2861.16 ± 0.276
−P, 4 h 2.44 ± 0.032**3.32 ± 0.105*

Cloning and characterization of OsPI1 cDNA

An EST clone (Accession No. C19881) encoding OsPI1 cDNA was provided from a gene bank and sequenced by ourselves. The 3′- and 5′-flanking sequences of OsPI1 cDNA were determined by the 3′- and 5′-RACE methods, respectively (Fig. 3). It was estimated that OsPI1 cDNA, excluding the poly A, was 375 bp in length (Fig. 3). There was a large difference in molecular mass of transcripts estimated by the Northern blot analysis, which detected major bands between 0.5 and 0.6 kb (Fig. 2e). Generally, the poly A tail on the 3′ end of mRNA was 20–250 bases (Sugiura & Takeda, 2000), and it was assumed that OsPI1 mRNA added a longer poly A tail during processing. Compared with the genomic sequence, the OsPI1 gene has no introns (data not shown). A base on the 5′ end of cDNA (G) was different from the genomic sequence (T) (Fig. 3) and an error of PCR amplification was believed responsible. However, we obtained the same results over several replications (data not shown) and post-transcriptional modifications of the 5′ end may be needed.

Figure 3.

cDNA sequence and deduced amino acid sequences of OsPI1. Lower case letters in 5′-end are different with the genomic sequence (T). Bold letters indicates the longest and considerable open reading frames (ORFs) deduced from three frames. Double underlined bases indicate the transcription starting points estimated by primer extension (Fig. 4a,b). Underlined bases indicate the consensus region among the TPSI1/Mt4 family (Fig. 5).

Primer extension was used to precisely determine the starting point of the transcription of OsPI1 (Fig. 4). The signal detected in the uppermost region coincided with the transcription initiation point estimated by 5′-RACE (Figs 2 and 3). The primer was extended to two obvious signals (Fig. 4a,b), with one-base shorten weak signals (Fig. 4a,b). These were extended even at 60°C, which was stringent annealing of the primer; therefore, we cannot exclude the possibility that both signals originated from OsPI1 mRNAs. It seems that the transcript corresponded to the signal B, and B′ in Fig. 4 was correlated to the weak bands detected at about 0.3 kb in the Northern blot analysis (Fig. 2e). It was considered that the shorter mRNA was produced by alternative splicing or originated from the secondary transcription initiation site.

Figure 4.

Primer extension analysis. Arrows indicates the detected parts of the extended primer. Left, center, and right panels are the results of sample and primer annealed at 42, 50 and 60°C, respectively. A and B indicate the strong signal, and A′ and B′ indicate the weak signal observed in one base downward of A and B, respectively.

OsPI1 cDNA had no similarities with any known plant genes (data not shown), whereas it was similar to two ESTs: BG836358 of maize and AL504365 of barley (Fig. 5). Despite a huge number of dicotyledon ESTs on the database, it was interesting to note that OsPI1-like ESTs were found only among monocotyledons (URL: index.html).

Figure 5.

The sequences conserved among TPSI1/Mt4 family, OsPI1 and OsPI1-like ESTs. Asterisks indicate the coincided bases among all genes. Dots indicate the parts mismatched with the sequence of OsPI1.

P starvation inducible novel genes that have been isolated are as follows: TPSI1 of tomato (Liu et al., 1997), Mt4 of Medicago truncatula (Burleigh & Harrison (1997), 1998), At4 of Arabidopsis thaliana (Burleigh & Harrison, 1999) and AtIPS1 (Martín et al., 2000). When their sequences were aligned with OsPI1 and OsPI1-like ESTs, they had two conserved groups consisting of eight bases each (Fig. 5). There were no other similarities in their sequences.

Fig. 6 shows the result of the genomic Southern blot analysis. Single bands were detected in all lanes, which coincided with the estimated sizes calculated from a PAC clone AP003231 containing OsPI1. Therefore, because the PAC clone was on chromosome 1, this indicates that OsPI1 is a single copy gene that is also located on chromosome 1.

Figure 6.

Genomic Southern blot analysis. Genomic DNA of rice was digested by BamHI, EcoRV, HindIII and XbaI (these are shown in the top of lanes), and used for genomic Southern blot analysis of OsPI1. The estimated length of fragments digested by BamHI, EcoRV, HindIII and XbaI are 27,942, 13,767, 1,321 and 8,613 bp, respectively.


The TPSI1/Mt4 family has many common characteristics. These are as follows: their expression is only induced by P starvation in both leaves and roots; they are rapidly down-regulated by fertilization with Pi or colonization with arbuscular mycorrhizal fungi; they share only a consensus sequence that consist of approx. 20 bases; the cDNAs are less than 1,000 bp in length and contain only short nonconserved open reading frames (ORFs); and they are single copy genes without introns. All these features were observed in the OsPI1 gene and therefore we concluded that the OsPI1 gene belongs to the TPSI1/Mt4 family.

As none of the consensus sequences of the TPSI1/Mt4 family consist of a coordinate ORF, Martín et al. (2000) suggested the possibility that the TPSI1/Mt4 family acts as riboregulators. OsPI1 and OsPI1-like ESTs (BG836358 and AL504345; data not shown) also encode only small ORFs, and none of the consensus sequences are coordinated. This fact supports the suggestion that the TPSI1/Mt4 family, including OsPI1, does not encode a polypeptide, but plays a role as a riboregulator (regulatory RNA). It is possible that to act as regulatory factors, the consensus sequence will bind with some molecules, such as RNA, genomic DNA or proteins that are involved in quick responses to P deficiencies.

Few papers about plant riboregulators have been published. In the case of E. coli, DsrA and OxyS form three stem–loop structures bound with complementary bases of other mRNAs, and regulate the translation of a transcriptional silencer and a transcriptional activator, respectively (Altuvia et al., 1998; Lease et al., 1998). Because it has been predicted that OsPI1 mRNA had no stem–loop structure (data not shown), OsPI1 cannot have a direct relationship with the riboregulators of E. coli. Further studies are necessary to clarify the molecular function of OsPI1.

Genomic Southern blot analysis (Fig. 6) and genomic sequence on database revealed that OsPI1 is a single copy gene located on chromosome 1, which contains two interesting quantitative trait loci (QTL) series. One series of QTL is believed to contribute to increasing tiller number and dry weight under low P conditions (Ni et al., 1998). The second series contributes to an increase of APase activity under low P conditions (Hu et al., 2001). A QTL contributing APase activity was mapped between E7-M5-6 and E4-M1-8, and this region is suggested to contain the OsPI1 gene, which was located 143 cm in chromosome 1 (data not shown). Thus, we suggest that the QTL is involved in the function of OsPI1 on APase activity, and further study is warranted.

OsPI1 was rapidly up-regulated in shoots and roots after depletion of the Pi supply, and was down-regulated within 12 h of Pi fertilization (Fig. 2e). As previously mentioned, an earlier response on P status is commonly observed among the TPSI1/Mt4 family (Liu et al., 1997; Burleigh & Harrison, 1998; Martín et al., 2000). A large difference in total P concentration was not observed between when OsPI1 expression was induced and when it decreased, in either the shoots or the roots (Fig. 2c–e). This suggests that the expression of OsPI1 is not solely regulated by the change of total P concentration in each organ.

We have already shown that the secretion of acid phosphatase (APase) is inducible without a decrease of the P concentration in roots of the lupin, which has an extremely high ability to secrete APase (Wasaki et al., 1999). On the other hand, the APase activity and mRNA accumulation drastically increase when the P concentration of the lupin plants decreases (Ozawa et al., 1995; Wasaki et al., 2003). These results indicate that the response to P stress consisted of two steps (Wasaki et al., 2003), which agrees with the response of other P-deficient inducible genes such as the TPSI1/Mt4 family. It is believed that both internal and external Pi concentration under slight P stress and total P concentration under severe P stress are important in the responses.

A P-deficient responsible transcription factor designated as PHR1 was isolated from Arabidopsis thaliana (Rubio et al., 2001) and shown to have the ability to bind to the promoter region of AtIPS1, a member of TPSI1/Mt4 family (Rubio et al., 2001). The PHR1 binding sequence GNATATNC was also found in the promoter region of OsPI1 (Table 3), other TPSI1/Mt4 members, type 5 APase, Pi transporters and ribonuclease (Rubio et al., 2001). Mukatira et al. (2001) mentioned that the PHO element (CACGTG/C; Oshima, 1997), the helix-loop-helix element (CAT(/G)A(/C)TG; Blackwell & Weintraub, 1990) and the NIT2 element (TATCA(/T)A(/T); Fu & Marzluf, 1990) are common in the TPSI1/Mt4 family as cis-regulatory regions and in the AtPT2 encoding Pi transporter isoform of Arabidopsis. The Pi responsible leucine zipper protein binding domain in the VspB gene, which encodes a vegetative storage protein and has APase activity, was determined and designated as vsp BoxII (Tang et al., 2001). The promoter region of OsPI1 also contains these sequences (Table 3) and it is possible that they act as keys for responses to P stress. Further studies on the promoter analysis will help clarify the detailed mechanisms of responses to P deficiency at the molecular level.

Table 3.  Putative cis-acting elements contained in the promoter region of OsPI1
Name of the ElementSequencePositionReference
PHR1-binding siteGNATATNC −192(Rubio et al., 2001)
Helix-loop-helix elementCAT(/G)A(/C)TG −458(Blackwell & Weintraub, 1990)
vspB boxII elementATTAATT−1318(Tang et al., 2001)
NIT2 elementTATCA(/T)A(/T)−1324(Fu & Marzluf, 1990)
PHO elementCACGTG/C−1483(Oshima, 1997)


This research was funded in part by the Core Research for Evolutionary Science and Technology (CREST) of the Japan Science and Technology Corporation (JST) and the rice genome project MA-2111 of the Ministry of Agriculture, Forestry and Fisheries (MAFF). We thank Dr Junshi Yazaki, Mr Katsumi Sakata, Dr Takuji Sasaki, Dr Naoki Kishimoto, and Dr Shoshi Kikuchi in National Institute of Agrobiological Sciences (NIAS), and Ms. Fumiko Fujii, Ms. Kanako Shimbo, and Ms. Kimiko Yamamoto in Society for Techno-innovation of Agriculture, Forestry and Fisheries for their support of cDNA microarray analysis. We are grateful to DNA bank of NIAS for providing EST clone (Accession No. C19881). We thank Dr Hisanori Bando for the usage of facilities in the Laboratory of Molecular Entomology, Graduate School of Agriculture, Hokkaido University.