, Volume 227, Issue 1, pp 151–165

Molecular cloning and characterization of phosphorus starvation responsive genes in common bean (Phaseolus vulgaris L.)


  • Jiang Tian
    • Department of Horticulture and Landscape ArchitecturePurdue University
    • Laboratory of Plant Nutritional Genetics, Root Biology CenterSouth China Agricultural University
  • Perumal Venkatachalam
    • Department of Horticulture and Landscape ArchitecturePurdue University
  • Hong Liao
    • Laboratory of Plant Nutritional Genetics, Root Biology CenterSouth China Agricultural University
    • Laboratory of Plant Nutritional Genetics, Root Biology CenterSouth China Agricultural University
    • Department of Horticulture and Landscape ArchitecturePurdue University
Original Article

DOI: 10.1007/s00425-007-0603-2

Cite this article as:
Tian, J., Venkatachalam, P., Liao, H. et al. Planta (2007) 227: 151. doi:10.1007/s00425-007-0603-2


Common bean (Phaseolus vulgaris L.) is one of the most important food legumes in the world and its production is limited by low phosphate (Pi) availability in many arable soils. To gain better insight into the molecular mechanisms by which common bean adapts to low Pi availability, we generated a suppression subtractive cDNA library to identify genes involved in P starvation responses. Over 240 putative Pi starvation-responsive genes were identified. The identified clones were sequenced and BLASTx/BLASTn analysis revealed an array of 82 genes showing a high degree of sequence homology to known and unknown proteins in the database. Transcript abundance of seven genes representing different functional categories was examined by Northern blot analysis. Six genes were strongly induced/enhanced under Pi deficiency confirming the results of SSH. Full length cDNAs for three genes, representing PvIDS4-like, PvPS2, and PvPT1 were cloned and characterized. The open reading frame (ORF) of PvIDS4-like encodes a 281-amino acid protein, containing a SPX domain. The ORF of PvPS2 gene encodes a 271-amino acid protein coding for a putative phosphatase. The PvPT1 encodes a 531-amino acid protein exhibiting high homology with high affinity Pi transporters. Expression patterns of these three genes in relation to Pi availability were evaluated with two contrasting genotypes (P-inefficient Dor364 and P-efficient G19833). Both Northern and RT-PCR results showed enhanced accumulation of phosphate transporters and phosphatases in P-efficient genotype, implying that in addition to modified root morphology and architecture, increased P transport and phosphatases activity might contribute to efficient Pi acquisition and translocation in G19833 common bean genotype under limited Pi conditions.


Phaseolus vulgaris L.Pi starvationSuppression subtractive hybridization (SSH)Polymerase chain reaction (PCR)Common beanCloningSPX domain





Phaseolus vulgarisIron Deficiency Specific-4 like gene


Phaseolus vulgarisPhosphate Starvation induced-2 gene


Phaseolus vulgarisPi Transporter-1


Suppression subtractive hybridization


Open-reading frame


Base pairs


Reverse transcriptase polymerase chain reaction


Phosphate (Pi) is one of the essential macronutrients, directly or indirectly involved in many important physiological and biochemical processes, including photosynthesis and respiration. Phosphorous and its derivates (i.e. IP3) are also involved in signal transduction pathways. It is well established that low Pi availability in soils is one of the major factors limiting crop production. Pi is easily fixed either by organic compounds or by Fe or Al oxides into forms that are unavailable to plants (Raghothama 1999). However, plants have developed many adaptive strategies such as increasing organic acid exudation, changing root morphological traits and activating phosphate transporters to enhance Pi acquisition (Raghothama 1999; Vance et al. 2003). Common bean (Phaselous vulgaris L.) is one of major food legumes cultivated in Pi deficient soils in the tropics. The origin of this crop from two major gene pools representing Middle America and Andean mountain (Pi efficient) areas has provided excellent genetic resources to study and improve Pi efficiency. Significant genetic diversity in Pi efficiency in common bean has been demonstrated through greenhouse and field experiments (Yan et al. 1995a, b; Beebe et al. 1997).

Two common bean genotypes, G19833 (Pi efficient) and Dor364 (Pi inefficient), as well as their recombinant inbred lines, are widely used to study the morphological, physiological, and genetic mechanisms underlying Pi efficiency. It has been shown that G19833 developed a shallower root system under Pi limited conditions by decreasing basal root growth angle, and forming longer and finer roots (Liao et al. 2001). The shallower root system may have contributed to the superior Pi acquisition ability of G19833 by allocating more roots to shallow soil horizons rich in available Pi, a trait closely associated with some quantitative trait loci (Liao et al. 2004). Along with changes in root architecture and morphology, exudation of more organic acids and protons was observed in G19833 as compared to that in Dor364 at low Pi level. In addition, G19833 also forms more cortical aerenchyma in roots, which is implicated in increased metabolic efficiency of plants under reduced Pi (Shen et al. 2002; Fan et al. 2003). In spite of the detailed analysis of adaptive mechanisms of Pi starvation at both genetic and physiological levels, not much is known about the molecular regulation of Pi starvation responses in common bean. With the advent of macro- and microarray technologies several candidate genes induced by low Pi have been isolated in various plant species including legumes. Sequencing of 2,000 ESTs from Pi-deficient proteoid roots revealed the specific involvement of several genes in carbon and secondary metabolism, Pi scavenging and remobilization, plant hormone metabolism, and signal transduction in white lupin (Uhde-Stone et al. 2003). A total of 13, 245 ESTs representing Pi-starved roots and leaves of Medicago truncatula and about 5, 429 ESTs representing Pi-starved roots of soybean are found in the data base [The Institute for Genomic Research (TIGR),]. There are several reports of differentially expressed genes under Pi deficiency emerging from the analysis of macro and microarrays in Arabidopsis (Hammond et al. 2003; Wu et al. 2003; Mission et al. 2005; Morcuende et al. 2007). In P. vulgaris, a total of 3,165 ESTs belonging to Pi-starved root cDNA library was reported by Ramírez et al. (2005). An in silico approach for the identification of genes involved in adaptation of P. vulgaris and other legumes to P-deficiency has also been reported (Graham et al. 2006).

In spite of these developments, detailed analyses of genes expressed differentially under Pi deficiency in common bean are lacking. This is partly due to lack of microarrays representing the P. vulgaris genome. In the absence of microarrarys, suppression subtractive hybridization (SSH) is an efficient approach to identify genes that are differentially expressed during Pi deficiency. Among various methods available to identify differentially expressed genes, SSH is an efficient technique to isolate novel genes from plants lacking entire genomic sequence information. In this technique differentially expressed genes can be normalized and enriched over 1000-fold in a single round of hybridization (Diatchenko et al. 1996). SSH has successfully been used on various plant species to isolate genes involved in responses to biotic and abiotic stresses (Zhang et al. 2005; Srivastava et al. 2007; Venkatachalam et al. 2007). In the present investigation, an attempt was made to isolate P starvation-regulated genes from a P-efficient common bean genotype (G19833) using SSH analysis. In addition, three genes representing PvIDS4-like (iron deficiency specific gene) PvPS2 (phosphatase) and PvPT1 (high affinity Pi transporter) were isolated, and their expression patterns were analyzed in Pi efficient and inefficient common bean genotypes. This study has led to a comprehensive understanding of gene regulation in P-efficient common bean genotype during Pi starvation and provided a better insight into the molecular mechanisms by which common bean adapts to low P availability.

Materials and methods

Plant materials and growth conditions

The Pi-efficient genotype G19833 is a relatively well-adapted common bean variety to low Pi conditions. Seeds of G19833 were surface sterilized for 1 min in 10% H2O2 before germinating in the dark on germination papers soaked with 1/2 strength modified nutrient solution (Yan et al. 2001). Uniformly germinated, six-day-old seedlings were transplanted into modified nutrient solution supplied with 5 (low P) or 1,000 μmol/L (high P) KH2PO4. Nutrient solution was continuously aerated and replaced every seven days. Shoots and roots were harvested for construction of SSH library on 10th day after transferring plants to hydroponics. To study the effect of phosphate availability and gene expression, uniform seedlings were transplanted into modified nutrient solution supplied with 5, 50, 100 and 250 μmol/L KH2PO4. Seedlings were also transferred to nutrient solution lacking potassium (K), nitrogen (N) or iron (Fe) along with Pi. Shoots and roots were harvested after 10 days of initiation of different nutrient treatments.

Determination of biomass and leaf soluble phosphorus content

Fresh weight of shoots and roots was recorded immediately after harvest and soluble phosphorus content in leaves was determined as described by Murphy and Riley (1963).

Construction of the suppression subtractive hybridization (SSH) library

Total RNA was extracted from the shoots and roots of G19833 subjected to Pi deficiency (Low P) or Pi sufficiency (High P) using Trizol reagent. Poly (A+) RNA was isolated from total RNA using Poly (A+) mRNA isolation kit (Promega, USA). The subtraction library was prepared using PCR-Select cDNA subtraction kit (BD Biosciences, USA) according to the manufacturer’s instructions. Briefly, cDNAs were synthesized separately from 2 μg Poly (A+) RNA, isolated from plants grown at low and high P levels. The cDNAs generated from plants at low P level served as “testers”, while cDNAs generated from plants at high P level were used as “drivers”. cDNA digestion, adaptor ligation, hybridization, and PCR amplification were done according to the protocol described earlier (Srivastava et al. 2007). Subtracted PCR amplified cDNAs were cloned into pCRscript vector using TOPO TA Cloning kit (Invitrogen, USA). The ligated product was transformed into E. coli competent cells (TOPO 10) and the transformed cells were plated onto LB agar medium containing 100 μg/ml kanamycin, and incubated at 37°C overnight. About 1,000 positive clones were obtained, and each clone was picked individually and grown overnight in LB medium with 100 μg/ml kanamycin. The amplified cells, representing subtraction libraries enriched with differentially expressed gene fragments, were stored as glycerol stocks at −80°C.

Screening the SSH library

Differential screening was carried out by dot blot hybridization. About 3 μl of PCR product, amplified from individual clones with T3 and T7 primers, was mixed with 3 μl of 0.6 N NaOH for denaturing, and then 1.5 μl of each was arrayed on to two identical nylon membranes. After air drying, membranes were UV-cross-linked in a Stratalinker (Stratagene, CA). 32P-labeled cDNA probes were synthesized from the tester as well as the driver cDNAs using the random primed labeling kit (Ambion, USA) and they were purified using Sephadex G50 columns. Two sets of membranes were hybridized with 32P labeled driver or tester cDNAs. Hybridization was carried out at 42°C overnight in a solution containing 50% (v/v) formamide, 5 × Denhardt’s, 0.5% (w/v) sodium dodecyl sulfate (SDS), 6 × SSPE and denatured salmon-sperm DNA. The membranes were washed with 2 × SSC, 0.2% (w/v) SDS at 42°C for 10 min, and then with 0.1 × SSC, 0.1% SDS at 42°C for 10 min. Blots were exposed to X-ray film (Fuji photo film, Japan) at −80°C for signal detection.

DNA sequencing and blast analysis

Based on the results of dot blot hybridization, differentially expressed clones were selected and sequenced at the Purdue Genomics Core Facility, Purdue University. The DNA sequences were analyzed using the BlastN and BlastX algorithm ( to identify their putative functions. The sequences of cDNAs identified in this work were deposited (Accession numbers from EG594296 to EG594395) in the GenBank (

Construction of cDNA library

Common bean seedlings of Pi-efficient genotype G19833 were subjected to Pi starvation and total RNA was extracted for cDNA construction according to manufacturer’s instructions (Stratagene, USA). Briefly, cDNAs were synthesized from 5 μg of Poly (A+) RNA and EcoRI adapters were ligated to blunt ended cDNAs. After XhoI digestion and size separation on Spharose CL-2B column, cDNAs were ligated to the EcoRI-XhoI site of the Uni-ZAP XR express vector followed by packaging. The library prepared in lambda phage vector was amplified to a stable stock.

Isolation and cloning of full-length cDNAs for PvIDS-4, PvPS-2 and PvPT1

Based on partial sequences identified in the SSH library, two sets of primers were designed for PCR amplification of PvPS2 (Forward, 5′-GAAAGTTCTTCCTTGGCATCA-3′ & Reverse, 5′-GAAAGTTCTTCCTTGGCATCA-3′) and PvPT1 (Forward, 5′-GAGCCACAGGTATGGTTTGTTC-3′ & Reverse, 5′-GATTGCAAGCTTCAGACTCTG-3′) genes. cDNA library stock was used as template for amplification of these genes in separate reactions, in combination with T7 and T3 sequence primers. A combination of T3 sequence primer as forward and gene specific primer as reverse (5′-CATCCGTGTTGTTTGTTATCTG-3′) was used in PCR amplification of full-length sequence of PvIDS4-like gene. PCR amplified cDNA fragments of three genes were cloned into pGEM-T vector and sequenced. Alignment of amino acid sequences and phylogenetic analysis were conducted with ClustalW program.

Northern blot analysis

Total RNA (10 μg) was size fractionated in 1.2% (w/v) agarose gel and then transferred to nylon membrane by capillary method and UV-cross linked. Gene fragments representing different functional groups were amplified by PCR and purified DNA was separately labeled using random primed labeling kit (Ambion, USA). The labeled probes were purified on Sephadex G50 columns and added into prehybridized RNA blots. Hybridization and washing conditions used were the same as described above. Membranes were exposed to X-ray film (Fuji photo film, Japan) to detect signals at −80°C.

Semiquantitative RT-PCR analysis

Total RNA was isolated as described previously, treated with DNaseI and subsequently used for RT-PCR. Two microgram total RNA of each sample was used for 1rst-strand cDNA synthesis in 20 μl reaction containing 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, 10 mM DTT, 50 mM dNTPs, 200 U MMLV reverse transcriptase (Promega Inc., Madison, WI, USA) and 50 pmol oligonucleotides T15. Reverse transcription was performed at 42°C for 60 min with a final denaturation at 70°C for 15 min. The cDNA (1 μl) was used directly for PCR with 1 U of Taq polymerase (Promega Inc., Madison, WI, USA) in the presence of dNTPs at 250 μM and the appropriate primers. The following primer pairs were used for RT-PCR analysis of PvPT1 (5′ GGTACCACTTCACAGCAA 3′ and 5′ ACTTCAGCTTCAAGCTCC 3′), PvIDS4 like (5′ CTCTAGCAACCAGATCGAG 3′ and 5′ CCACCCATCTGCAGCGGT 3′), PvPS2 (5′ CGACAGCGACAACTGGGT 3′ and 5′ GCCTCCAAAACAGAGAGT 3′), actin (5′GTGCTCGATTCTGGTGATGGTGTG 3′ and 5′ CCACGACCTTGATCTTCATGCTGCT 3′). RT-PCR was done using with the following thermal cycling parameters: 94°C for 3 min followed by 25 cycles of 94°C for 1 min, 58–60°C for 1 min and 72°C for 1 min.


Effect of Pi deficiency on growth and leaf soluble P concentration of common bean genotype G19833

Common bean seedlings exhibited distinct symptoms of Pi starvation (dark green leaves and reduced leaf expansion) after 10 days of transfer to hydroponic solution lacking Pi. Growth of G19833 was significantly inhibited by low Pi availability in the medium (Table 1). The overall fresh weight of G19833 was reduced by 56% and that was accompanied by a significant reduction in leaf soluble phosphorus content under Pi deficiency. However, the root/shoot ratio was increased to 0.66 under low Pi conditions. Pi sufficient and deficient plant materials showing distinct symptoms to Pi deficiency were used for gene expression analysis.
Table 1

Effect of Pi on growth of common bean genotype G19833

Phosphorus concentration


Fresh weight (g plant−1)

Leaf soluble P concentration (mg g−1)

5 μM

0.68 ± 0.02a

9.21 ± 0.16b

0.099 ± 0.003b

1 mM

0.53 ± 0.05b

21.22 ± 2.35a

0.44 ± 0.13a

Seedlings were grown for 10 days in the nutrient solution supplied with 5 μM or 1mM KH2PO4, then harvested for measuring root to shoot ratio, fresh weight and leaf soluble P concentration. Data represent the mean of four replicates with standard error. Different letters in the same column indicate the significant difference at 0.05 level (P < 0.05)

Construction and differential screening of SSH library

In order to understand the molecular mechanisms leading to the adaptation of common bean to low Pi, an SSH library was constructed to identify Pi starvation regulated genes. The SSH library was screened by dot blots to identify Pi-responsive genes for further characterization. PCR products amplified from individual SSH clones were arrayed on two identical sets of nylon membranes, and then hybridized separately with two 32P labeled probes (tester and driver cDNAs). Hybridization signal intensity varied greatly among the clones. Many of them showed strong signal when the membrane was hybridized with the tester cDNA probe (Fig. 1a). Meanwhile, only weak signals were detected when the same membrane was hybridized with the driver cDNA as a probe (Fig. 1b). Based on dot blot results, 240 clones were identified as differentially regulated cDNAs under low P condition and they were selected for sequencing. BLASTx analysis of the cDNA sequences was carried out to identify their putative functions ( Sequences without significant homology with known proteins were further analyzed with BLASTn to predict their putative functions. This analysis revealed that 99 clones represent non-redundant cDNA inserts and details are summarized in Tables 2, 3. Bioinformatic analysis resulted in the identification of several Pi starvation-regulated genes involved in various biological processes.
Fig. 1

Differential screening of the SSH library with dot blot hybridization. An array of 100 clones was hybridized with 32P-labeled tester (a) and driver (b) cDNA probes synthesized from RNA isolated from plants grown at low and high Pi levels, respectively. The spots differing in their intensity between membranes were recognized as either up-regulated (circled by continuous line) or down-regulated (circled be dotted line)

Table 2

Identification of P starvation responsible genes isolated from the SSH library in common bean through analysis of BLASTx

SSH number

Accession number

Size (bp)



Homology protein [species and accession number]

Transporter and channel




Putative cationic amino acid transporter [Arabidopsis thaliana, BAC42395]

4 × e−59




Phosphate transporter [Sesbania rostrata, CAC28218]

4 × e−65




PDR-like ABC transporter [Nicotiana tabacum, CAH39854]

1 × e−27




Mitochrondrial voltage-dependent anion-selective channel [Phaseolus coccineus, AAY82249]

2 × e−53




Carbohydrate transporter/sugar porter/transporter [Arabidopsis thaliana, NP_187911]

1 × e−8




Sec 14p-like lipid-binding domain [Oryza sativa, XP_463507]

5 × e−57




ADP, ATP carrier protein precursor [Picea abies, CAC27140]

8 × e−5

Signaling and transcription




Ring finger protein family [Arabidopsis thaliana, BAA97070]

4 × e−37




Type IIB calcium ATPase [Medicago truncatula, AAL17950]

2 × e−142




Ser/Thr protein kinase [Lotus japonicus, BAD95891]

3 × e−10




R2R3-MYB transcription factor-like protein [Oryza sativa, BAD33914]

3 × e−20




Putative splicing factor [Oryza sativa, BAD354671]

3 × e−18




Transcription factor [Arabidopsis thaliana, NP_567030]

3 × e−36




ids-4 like protein [Arabidopsis thaliana, AAM64762]

7 × e−32




ids-4-like protein [Castanea sativa, AAL17697 ]

1 × e−49




ids-4 like protein [Arabidopsis thaliana, AAM64762]

6 × e−12




phosphoprotein phosphatase [Arabidopsis thaliana, NP_567261]

4 × e−8




Contain PHF5 domain [Oryza sativa, XP_474232]

7 × e−29




Protein kinase C inhibitor [Arabidopsis thaliana, CAB88052]

3 × e−25

Carbon metabolism




ribulose 1,5-bisphosphate carboxylase/oxygenase small subunit [Phaseolus vulgaris, AAB97165]

2 × e−31




ribulose 1,5-bisphosphate carboxylase large subunit [Cladrastis sinensis, CAB08876]

1 × e−53




Q(B) polypeptide [Medicago sativa, AAK50366]

6 × e−97




Putative cellulose synthase [Fagus sylvatica, AAZ79659]

1 × e−10




Photosystem II thylakoid membrane protein [Glycine max, AAQ67339]

4 × e−95




glyceraldehydes-3-phosphate dehydrogenase (phosphorylating) [Antirrhinum majus, CAA42103]

2 × e−74




cytosolic phosphoglucomutase [Pisum sativum, CAB60127]

3 × e−35




Chloroplast hypothetical protein [Zea mays, AAR91119]

1 × e−17




Chlorophyll a/b-binding protein [Glycine max, AAA50172]

7 × e−119

Stress and defence




Wound-inducible carboxypeptidase [Lycopersicon esculentum, AAF44708]

4 × e−26




endo-1,3-beta-glucanase [Glycine max, AAR26001]

3 × e−34




Putative senescence-associated protein [Pisum sativum, BAB33421]

2 × e−58




DNAJ heat shock protein [Arabidopsis thaliana, NP_177796]

1 × e−112




PR1 a precursor[Glycine max, AAD33696]

1 × e−6




guanine nucleotide-exchange-like protein [Arabidopsis thaliana, CAB89051]

9 × e−48




glycerophosphodiester phosphodiesterase [Arabidopsis thaliana, NP_198924]

4 × e−58




Cytochrome P450 [Medicago truncatula, AAQ20042]

7 × e−11




glutathione peroxidase 1 [Lotus japonicus, AAP69867]

1 × e−32




coronatine-insensitive 1 [Glycine max, AAZ66745]

4 × e−23




Endochitinase precursor [Phaseolus vulgaris, P06215]

3 × e−75




beta-galactosidase like protein [Arabidopsis thaliana, CAB16852]

2 × e−18




non-cyanogenic beta-glucosidase [Cicer arietinum, CAG14979]

3 × e−22




Mlo-h1-like protein [Arabidopsis thaliana, CAB80753]

7 × e−40

Other metabolism




Unspecific monooxygenase [Nicotiana tabacum, T02995]

4 × e−41




S-adenosylmethionine decarboxylase [Phaseolus lunatus, BAB83763]

6 × e−46




N-methyltransferase [Coffea liberica, AAM18507]

1 × e−18




Ribosomal protein L23a [Arabidopsis thaliana, AAB87692]

3 × e−18




Putative phosphatase [Lycopersicon esculentum, CAD30864]

4 × e−16




Putative phosphatase [Lycopersicon esculentum, CAD30864]

2 × e−3




Putative FK506-binding protein [Oryza sativa, XP_469487]

2 × e−15




Putative acid phosphatase [Lupinus luteus, CAE85073]

4 × e−25




Phosphoric monoester hydrolase [Arabidopsis thaliana, NP_173213]

2 × e−53




O-methyltransferase [Prunus dulcis, CAA11131]

7 × e−10




Inorganic diphosphatase/magnesium ion binding/pyrophosphatase [Arabidopsis thaliana, NP_190930]

6 × e−58




In2-1 protein [Glycine max, AAG34872 ]

5 × e−22




GMPase [Medicago sativa, AAT58365]

1 × e−63




Cu2+ plastocyanin-like [Arabidopsis arenosa, AAR15487]

5 × e−35




Bifunctional nuclease [Zinnia elegans, AAD00694]

1 × e−19




Putative phosphatase [Arabidopsis thaliana, AAU15169]

4 × e−54




cytochrome b6/f-complex subunit IV [Oryza nivara, BAD26809]

1 × e−18




1-deoxyxylulose 5-phosphate synthase [Catharanthus roseus, CAA09804]

9 × e−35





Nodulin like protein[Arabidopsis thaliana, AAC27411]

1 × e−12




Unknown [Arabidopsis thaliana, NP_197515]

2 × e−5




Unknown [Pisum sativum, CAA33557]

9 × e−26




Unknown [Arabidopsis thaliana, AAG51819]

6 × e−47




Unknown [Arabidopsis thaliana, AAM98235]

2 × e−21




Unknown [Arabidopsis thaliana, NP_850445]

9 × e−35




Unknown [Arabidopsis thaliana, NP_171899]

3 × e−3




Unknown [Arabidopsis thaliana, BAC42041]

1 × e−3




Unknown [Arabidopsis thaliana, BAD43699]

4 × e−33




Unknown [Arabidopsis thaliana, AAF82203]

1 × e−28




Unknown [Arabidopsis thaliana, AAU94402]

6 × e−21




Calcium-binding protein [Medicago truncatula, ABE94507]

4 × e−15

All 72 clones with homologies to BLAST x-expected values lower than 1.0 × e−2 are listed

Table 3

Identification of P starvation responsible genes isolated from the SSH library in common bean through analysis of BLASTn

SSH number

Accession number

Size (bp)


Homology nucleotide [accession number]


Putative function




Arabidopsis thaliana protein transport protein [AY128850]

7 × e−10





Phaseolus vulgaris chitinase mRNA, complete cds [M13968]

1 × e−11

Stress and defence




Vitis vinifera putative transcription factor mRNA [AF281656]

2 × e−8

Signaling and transcription




Platanus acerifolia partial mRNA for translation initiation factor [AM260514]

9 × e−10

Signaling and transcription




Medicago truncatula clone mth2-31k24, complete sequence [AC134521]

1 × e−8





Medicago truncatula Mt4 genomic sequence [AF055921]

7 × e−8





Medicago truncatula Mt4 genomic sequence [AF055921

6 × e−7





Hevea brasiliensis isolate SSH27 mRNA sequence [DQ306756]

3 × e−7





Vicia sativa 25S/18S rRNA intergenic spacer DNA [X61082]

2 × e−12





Cicer arietinum partial mRNA for hypothetical protein, clone Can141 [AJ404639]

5 × e−13





No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology




No significant homology


No significant homology

Clones without high homologies to the reported proteins based on BLASTx analysis were analyzed through BLASTn. Clones which do not show significant homologies with known sequences (expected values higher than 1.0 × e−5) are presented as those with no significant homology

Functional classification of differentially expressed genes

Differentially expressed genes identified by SSH were classified into five functional categories: (1) transporter and channel, (2) signaling and transcription, (3) carbon metabolism, (4) stress and defense, and (5) other metabolism (Tables 2, 3; Fig. 2). Some of the clones were similar to proteins with unknown function or without significant homology to known proteins. The largest number of genes with known function was classified under the category of other metabolism, followed by stress and defense. We also identified 8 clones as transporters/channels and 14 clones involved in signaling/transcription. Of the identified clones, 18 showed homology to genes with unknown functions and 17 clones had no significant homology to the reported genes and proteins in the database.
Fig. 2

Functional classification of Pi starvation induced genes in common bean. Differentially expressed genes were classified into seven groups according to their putative function generated by BLASTx and BLASTn analysis

Expression analysis of differentially induced genes under Pi deficiency

Based on the SSH results, the gene transcripts which were up-regulated in common bean under low Pi conditions were further examined by Northern blot analysis. Seven genes belonging to different functional groups, including phosphate transporter (PvPT, #2), Mt4/TPSI-like (#78), Phosphate starvation-induced 2 (PvPS2, #48), Cytosolicphosphoglumutase (PGM, #26), Glycerophosphoryl diester phosphodiesterase (GPD, #35), Iron deficiency specific 4 (PvIDS4-like, #16) and one with no significant homology genes (NSH, #99) were selected for RNA blot analysis. Transcripts of six genes out of the seven selected were strongly induced or enhanced by Pi starvation in the root tissues except NSH (Fig. 3). The level of mRNA accumulation of five genes other than phosphate transporter increased in shoots under low Pi condition, while the expression of PvPT1 and NSH genes was detected only in roots. This study showed that genes involved in carbon metabolism, Pi acquisition, transport, scavenging and remobilization are coordinately regulated in common bean under low Pi condition, resulting in enhanced P efficiency of G19833 genotype.
Fig. 3

Expression of Pi starvation induced genes in common bean. Total RNA extracted from shoot and root tissues of seedlings grown for 10 days in the nutrient solution supplemented with 5 μM (LP) or 1mM (HP) KH2PO4 was subjected to Northern blot analysis. Seven candidate genes namely Pi Transporter (PvPT1, #EG594339), Mt4/TPSI1-like (#EG594373), Phosphate Starvation-induced2 (PvPS2, #EG594341), Cytosolic Phospho Gluco Mutase (PGM, #EG594318), Glycerophosphoryl diester Phospho Diesterase (GPD, #EG594327), Iron Deficiency Specific 4 (PvIDS4-like, #EG594308) and one of the No Significant Homology genes (NSH, #EG594299) were selected to prepare 32P labeled probes. Ethidium bromide-stained rRNAs indicate uniform loading of samples

Isolation and characterization of full-length cDNAs for PvIDS4-like, PvPS2 and PvPT1 gene

To further characterize Pi starvation regulated genes in common bean, three full length cDNAs coding for PvIDS4-like, PvPS2 and PvPT1 proteins were cloned through PCR amplification. The molecular features of these three genes are described below.

The PvIDS4-like gene (Accession No. EF191350) contains a 846-bp open reading frame, encoding a 281 amino acid protein (Supplement Fig. 1). It belongs to the group of proteins containing the SPX domain, which is named after the proteins SYG1, PHO81, xenotropic and polytropic retrovirus receptor1 (XPR1). The putative protein has theoretical isoelectric point (pI) of 6.2 and a molecular weight of 32.4 kDa. Homology search using the BLASTx algorithm revealed that the deduced amino acid sequence of PvIDS4-like protein has substantial homology to IDS4-like protein in Arabidopsis thaliana (AAM64762, 63%), unknown protein in Arabidopsis thaliana (AAN38706, 63%), IDS4-like protein in Oryza sativa (BAD35467, 60%), which also contain a SPX domain. The alignment of amino acid sequences of these proteins with PHO81 (CAA97261, Saccharomyces cerevisiae) is presented in Fig. 4. Analysis revealed the presence of 49 amino acids and conserved Lys residues among the members (Fig. 4).
Fig. 4

Alignment of the deduced amino acid sequence of PvIDS4-like gene with other proteins containing the SPX domain. AAM64762 and AAN38706 represent IDS4-like proteins in Arabidopsis thaliana; BAD35467is an IDS4-like protein in Oryza sativa; CAA97261 (PHO81) is a SPX domain containing protein in Saccharomyces cerevisiae. Boxed region represent conserved sequences in IDS4-like proteins. Identical residues are indicated with asterisks, and conserved residues are shown by double dots

The PvPS2 gene (Accession No. EF472460) has a 816-bp open reading frame sufficient to code for 271-amino acid protein with a theoretical isoelectric point (pI) of 5.4 and a molecular weight of 30.7 KDa (Supplement Fig. 2). Homology search using the BLASTx algorithm revealed that the deduced amino acid sequence of PvPS2 is similar to proteins such as LePS2 in Lycopersicon esculentum (CAD30866, 65%), putative acid phosphatase and phosphatase in Arabidopsis thaliana (AAM63155, 61%; AAU15169, 65%) and putative phosphatase in Oryza sativa (NP_916687, 61%). The alignment of amino acid sequences of these proteins is illustrated in Fig. 5. Two distinct conserved motifs were observed in these amino sequences, including “FDFDXTI” and “YXGDGXXDXCP”.
Fig. 5

Alignment of the deduced amino acid sequence of PvPS2 with other phosphatase proteins. AAU15169, putative acid phosphatase in Arabidopsis thaliana; AAM63155, putative acid phosphatase in Arabidopsis thaliana; CAD30863, putative phosphatase in Lycopersicon esculentum; NP_694744, phosphatase in Mus musculus; NP_916687, putative phosphatase protein in Oryza sativa. PvPS2 (this work). Identical residues are indicated with asterisks, and well conserved residues are suggested with double dots

The cDNA, PvPT1 (Accession No. EF472461) has 1595-bp with an open reading frame sufficient to code for a 531-amino acid protein (Supplement Fig. 3). The putative protein has theoretical isoelectric point (pI) of 9.0 and a molecular weight of 58 KDa. PvPT1 protein was predicted to be an integral membrane protein containing 12 membrane spanning domains and most similar to the high affinity Pi transporters in plants. Multiple alignments of Pi transporter protein sequences are presented in Fig. 6. Results revealed that the deduced amino acid of PvPT1 share significant homology to other high affinity Pi transporter proteins, including AtPht1:4 (78%) in Arabidopsis thaliana and MtPT2 (84%) in Medicago truncatula.
Fig. 6

Alignment of the deduced amino acid sequence of PvPT1 with other Pi transporter proteins. Phosphate transporters and corresponding plant species are: AtPht1_4, Arabidopsis thaliana; MtPT2 and 4, Medicago truncatula; PvPT1 (this work), Phaseolus vulgaris. Identical residues are indicated with asterisks, and well conserved residues are suggested with double dots

PvPS2, PvPT1 and PvIDS4-like genes are regulated by Pi starvation

Two common bean genotypes, G19833 (Pi efficient) and Dor364 (Pi inefficient), were subjected to starvation of nitrogen, potassium, iron, and phosphate. Total RNA isolated from these nutrition starved plants was used for Northern analysis (Fig. 7). Transcript abundance of three genes was noticed in both roots (Fig. 7a) and shoots (Fig. 7b) of two genotypes under Pi starvation. There was no induction in the gene expression in plants subjected to other nutrient starvation (Fig. 7a, b). This suggests that these genes are specifically induced in response to Pi deficiency.
Fig. 7

Expression pattern of PvPS2, PvIDS4-like and PvPT1 genes in two common bean genotypes (Dor364 and G19833) under nutrient deficiency conditions. Total RNA isolated from a roots and b shoots of seedlings grown for 10 days in hydroponic solution supplemented with P or lacking one of the essential nutrients was used for Northern blot analysis. Ethidium bromide-stained rRNAs indicate uniform loading of the samples

Expression of PvIDS4-likePvPS2, and PvPT1 genes is influenced by Pi availability in two common bean genotypes

Transcripts abundance of PvIDS4-likePvPS2, and PvPT1 genes was observed in two common bean genotypes grown under varying Pi concentrations. Expression of these three genes in both roots and shoots was dependent on Pi availability in the medium (Figs. 8, 9). Both Northern and RT-PCR data showed that expression was greatly suppressed in roots when Pi concentration was 100 μM or higher. Transcripts of the three genes increased with decreased concentration of Pi, reaching a maximum at 5 μM Pi in the medium (Figs. 8 and 9). Major differences in the expression of PvPT1 and PvIDS4-like genes in shoot were observed between the genotypes G19833 and Dor364 (Fig. 9b). However, PvPT1 transcript was not detected in shoots (data not shown). These data suggested that the Pi efficient G19833 responds to Pi starvation more rapidly than Pi inefficient genotype Dor364.
Fig. 8

Expression pattern of PvPS2, PvIDS4-like and PvPT1 genes in two common bean genotypes (Dor364 and G19833) related to P availability. Total RNA isolated from a roots and b shoots of seedlings grown for 10 days in hydroponic solution supplemented with different Pi concentrations (5, 50,100 and 250 μM) was used for Northern blot analysis. Ethidium bromide-stained rRNAs indicate uniform loading of the samples
Fig. 9

RT-PCR analysis of PvPS2, PvIDS4-like and PvPT1 genes in two common bean genotypes (Dor364 and G19833) related to P availability. Total RNA isolated from a roots and b shoots of seedlings grown for 10 days in hydroponic solution supplemented with different Pi concentrations (5, 50,100 and 250 μM) was used for RT-PCR analysis. Actin gene was used a control in PCR amplification


Identification of genes associated with Pi deficiency is essential to understand not only the molecular mechanisms governing Pi efficiency but also for marker assisted selection (MAS) breeding and genetic improvement of common bean. Phosphate starvation had a profound impact on common bean growth as indicated by dark green leaves and reduced leaf expansion. Changes in plant morphology and accumulation of anthocyanin are documented to be associated with changes in gene expression (Mission et al. 2005; Raghothama and Karthikeyan 2005; Jain et al. 2007). In this study, G19833, a common bean genotype with superior Pi efficiency, was used to construct SSH library for identification of low Pi regulated genes. Transcriptional regulation of a subset of Pi starvation induced genes in two contrasting genotypes of common bean was evaluated by Northern blot and RT-PCR analysis. Some salient features of these genes are discussed in the following sections.

Carbon metabolism

Inorganic phosphate plays a very important role in plant metabolism, including photosynthesis and respiration (Duff et al. 1989). Several bypass enzymes such as phosphoglucomutase (PGM) are activated by Pi starvation and they allow carbon flux through glycolytic cycle despite a reduction in Pi and adenylate levels (Plaxton and Carswell 1999). One of the genes (#26) identified in this study has homology to PGM. Northern blot analysis showed that the expression of PGM was strongly induced in both shoot and root tissues (Fig. 3), indicating that alternative pathway of glycolytic carbon flow is activated during Pi starvation in common bean. Induced expression of PGM gene is likely to enhance carbon sources for the tricarboxylic acid (TCA) cycle, which impacts organic acid exudation in common bean. Interestingly, a phosphorylating glyceraldehyde-3-phosphate dehydrogenase (GAPDH, #25) was also identified in the SSH library. Induction of this gene is consistent with the earlier finding that phosphorylating GAPDH was among the Pi-responsive genes in white lupin roots (Uhde-Stone et al. 2003). Higher level of GAPDH expression in the common bean genotype (G19833) may be in part due to activation of enzymes involved in carbon metabolism during Pi deficiency. Induction of these genes in common bean genotype G19833 used in this study is known to secrete more citrate, tartrate and acetate during Pi starvation (Shen et al. 2002).

Under Pi starvation, the rate of photosynthesis is limited in many crops including common bean (Kondracka and Rychter 1997). This may be due in part to the repression of a set of genes (e.g. ferredoxin NADP + reductase), that are involved in photosynthesis during Pi starvation (Hammond et al. 2003; Wu et al. 2003; Misson et al. 2005). Induction of two genes with homology to 1,5-biophosphate carboxylase/oxygenase (Rubisco) small and large subunits suggests that inhibition of photosynthesis in common bean under low Pi condition might not be due to lack of this isoform of Rubisco. This result differs from an earlier observation that reduced photosynthetic rate in some plants during Pi starvation could be due to decreases in regeneration of ribulose 1,5-biophosphate (RuBP) pool size, and not due to reduction of Rubisco activity (Pieters et al. 2001). However, since Rubisco is encoded by a family of genes the physiological significance of induction of two members in this family needs further investigation. There is a likely hood that some members of the family may be up- or down-regulated in response to environmental stresses.


Phosphate starvation leads to significant changes in the transport of inorganic Pi and organic phosphate compounds. This was reflected by changes in the expression of eight genes coding for transporters/channels (Tables 2, 3). One of the genes (#2) showed highest homology with the high affinity Pi transporter in Sesbania rostrata (Table 2). This was further confirmed by Northern analysis showing that the putative Pi transporter gene was preferentially induced in roots under Pi deficiency conditions in common bean (Fig. 3). Root specific expression of this gene suggests its potential role in Pi acquisition, unlike some other high affinity transporters, which are involved in both Pi acquisition and mobilization in roots and shoots (Karthikeyan et al. 2002; Harrison et al. 2002). Two cDNAs (#1 and 73) encoding peptides with homology to both amino acid transporter and protein transporter were also identified (Tables 2, 3) as Pi starvation-induced genes in common bean. It has been demonstrated that genes encoding vacuolar protein protease are up-regulated under Pi limited conditions in plants, implying that proteins might be transported into vacuole via a protein transporter, and degraded by proteases into amino acids, which are then transported out of vacuole by amino acid transporters (Wu et al. 2003). In addition, one gene (#3) with homology to PDR (pleiotropic drug resistant)-like ABC transporter in tomato was also induced under low Pi conditions (Table 2). The PDR-like ABC transporter in tomato is regulated by both methyl jasmonate and iron deficiency and its induction during Pi starvation may be a consequence of cross talk between different nutrient stresses. This is not particularly surprising considering nutrients deficiency signaling appears to overlap with one another (Wang et al. 2002).

Other metabolism

Under low Pi condition, several metabolic pathways associated with enhancing and recycling the available pools of phosphate are triggered by increased activation of phosphatases, RNases and apyrases (Yun and Kaeppler 2001; Kock et al. 2006; Thomas et al. 1999). In this study, five cDNA clones (#47, 48, 50, 51 and 58) exhibiting homology with phosphatase, including putative phosphatase, putative acid phosphatase (APase) and phosphoric monoesterase hydrase were induced. Two of them (#47 and 48) showed high degree of amino acid homology with a putative phosphatase LePS2 in tomato (Baldwin et al. 2001). LePS2 belongs to a class of protein phosphatases and three members of this family have been identified in tomato (Baldwin et al. 2001; Stenzel et al. 2003). Northern results showed that transcripts of PvPS2 (#48) were strongly induced in both roots and shoots (Figs. 3, 7, 8). Expression of these genes is suggestive of coordinated regulation of biochemical processes associated with enhancement of Pi availability during the nutrient starvation. In addition, some cDNAs with homology to bifunctional nuclease (#57), phosphodiesterase (#35) were also isolated in the present study (Table 2). It is postulated that coordinated induction of a group of genes, encoding bifunctional nuclease, phosphodiesterase and acid phosphatase could enhance degradation of extra cellular or internal nucleic acid substrates leading to efficient reutilization of Pi under Pi limited conditions (Abel et al. 2002). We also identified a S-adenosylmethionine decarboxylase (SAMDC, #44), a N-methyltransferase (NMT, #45) and O-methyltransferase (OMT, #52), which are involved in secondary metabolite synthesis in plants. SAMDC is an important enzyme in polyamine biosynthesis, and it plays a role in abiotic stress (Hu et al. 2005).

Stress/defense associated genes induced under Pi deficiency

Induction of many stress related genes during Pi deficiency was observed in the earlier microarray analysis in Arabidopsis, white lupin and rice (Hammond et al. 2003; Uhde- Stone et al. 2003; Wu et al. 2003; Misson et al. 2005; Wasaki et al. 2006). In this study, 15 genes induced during Pi starvation were identified as defense and stress related (Fig. 2). Several of them were related to plant defense, including pathogenesis-related protein 1 (PR1 protein, #33), endochitinase precursor (#39), coronatine insensitive 1 (#38) and Mlo-h1 (#42) (Hammond et al. 2003). Induction of PR genes under Pi deficiency is an indication of crosstalk between Pi deficiency and biotic stresses. In tomato, one of Pi starvation induced phosphatases (LePS2) is also induced by pathogen infection (Stenzel et al. 2003). A few other genes related to oxidative stress (e.g., cytochrome P450, glutathione peroxidase 1) were also isolated. The over production of reactive oxygen species under Pi deficiency conditions in common bean is well understood (Juszczuk et al. 2001). Interestingly, both dot blot and Northern blot results showed that a homolog of glycerolphosphodieaster phosphodiesterase (GPD, #35) was strongly induced by low Pi. GPD is a senescence-associated protein and that may play a role in mobilization of Pi from senescing tissue to actively growing young tissues.


In this study, 14 genes known to be involved in transcription and signaling, including MYB transcription factor (#11), type II B calcium ATPase (#9), Ser/Thr protein kinase (#10) and PvIDS4-like protein (#14, 15, 16) were found to be regulated by Pi starvation (Table 2). Expression of similar genes under Pi deficiency has been reported in Arabidopsis thaliana, white lupin, rice and tomato (Muchhal et al. 1996; Uhde-Stone et al. 2003; Wu et al. 2003; Hammond et al. 2003; Misson et al. 2005). Identification of these Pi starvation-responsive genes, which are involved in transcription and signaling, is indicative of complex regulatory changes occurring in common bean during Pi starvation. Expression of MYB transcription factors and PvIDS4-like protein overlap with early and late Pi starvation in Arabidopsis thaliana (Hammond et al. 2003; Misson et al. 2005), indicating that they play important roles at different stages of Pi deficiency in plants.

Genes with unknown function

Eighteen cDNAs representing Pi starvation-induced genes exhibit high similarities to GenBank entries whose functions have not been clearly defined. Two cDNAs (#78 and 79) showed high homology to genes belonging to the Mt4/TPSI1 family (Table 3). Northern analysis indicated that the transcript of Mt4/TPSI1-like gene (#78) was induced in both shoot and root tissues by Pi starvation, like other members of the family in tomato, rice, Medicago truncatula and Arabidopsis thaliana (Liu et al. 1997; Burleigh and Harrison 1998; Hou et al. 2005; Shin et al. 2006). Members of this family were among the first to be identified as Pi starvation-inducible genes. However, the functions of these genes are not completely understood but their transcripts accumulate to high levels in Pi-starved tissues (Liu et al. 1997; Burleigh and Harrison 1998; Shin et al. 2006). Transcripts from the Mt4/TPSI1 genes contain a series of short overlapping ORFs that are conserved among the family members. A recent study provided evidence for the involvement of At4 in Pi allocation between the shoots and roots, possibly by influencing Pi retranslocation to the roots. Further it is stated that the At4 gene is not only induced during Pi-starvation, but the transcript levels may be regulated at posttranscriptional level by the activity of a micro RNA (Shin et al. 2006). About 17 other cDNAs representing novel transcripts that do not share any homology with the reported sequences in databases were identified as up-regulated genes under low Pi conditions.

PvPS2, PvPT1 and PvIDS4-like genes are strongly induced under phosphate starvation in Pi efficient genotype

The PvPS2 gene was induced under Pi deficient condition and its protein contains a putative phosphatase domain. BLASTx algorithm revealed that the deduced amino acid of PvPS2 protein showed significant homology to other putative phosphatase proteins, including LePS2 protein in tomato (AAG40473, 65%), which is a novel acid phosphatase (APase). Northern analysis results showed that the PvPS2 is strongly induced by low Pi (Figs. 3, 7, 8) but not by other nutrient deficiencies. This suggested that PvPS2 induction is specific to phosphate starvation. Our unpublished data on LePS2 suggests that members of this family act as phosphoprotein phosphatases (K. G. Raghothama, personal communication). This confirms the notion that some of the phosphatases may have functions other than releasing Pi during phosphate deficiency (Yan et al. 2001). Though PvPS2 gene is induced in Pi starved common bean plants, the exact role of PvPS2 protein during Pi starvation needs to be analyzed.

Sequence alignment analysis showed that PvIDS4-like gene has high degree of nucleotide sequence homology with IDS4-like genes in other plants. IDS4-like proteins are represented by a family of 4 members in Arabidopsis thaliana (Wang et al. 2004). Functions of IDS4-like protein have not been clearly elucidated. However, some Pi starvation-responsive genes encoding proteins with the SPX domain have been illustrated to play important roles in Pi sensing, transport and Pi homeostasis in yeast and Arabidopsis thaliana (Wykoff and O’Shea 2001; Hamburger et al. 2002; Wang et al. 2004). It has recently been revealed that expression of one gene (At5g20150), with high homology with PvIDS4-like gene, is rapidly induced and persists as long as the plants were exposed Pi deficiency, indicating the gene is involved in both early and late responses of plants to Pi starvation (Misson et al. 2005). In our study, expression of PvIDS4-like gene was specific to Pi starvation, and it was not affected by Fe deficiency, unlike the IDS4-like (gi:285635) gene, which was identified in Fe-deficient barley roots. This suggests that different members of IDS4 like gene family containing SPX domain may have specific roles in response to different nutrient deficiencies. Further characterization of the IDS4 like homolog identified in this study will shed light on functional diversity of this gene family.

The high affinity Pi transporter homolog PvPT1 encodes a 531-amino acid membrane protein. Northern and RT-PCR results showed that PvPT1 was preferentially and specifically induced in roots under Pi deficiency conditions in two common bean genotypes (Figs. 7, 8, 9). This gene is likely to be involved in Pi acquisition during Pi deficiency similar to other high affinity Pi transporters analyzed in various plant species (Muchhal et al. 1996; Liu et al. 1998; Karthikeyan et al. 2002; Harrison et al. 2002). Of the two genotypes tested, accumulation of PvPT1 transcripts was higher in the Pi-efficient genotype compared to inefficient one. Earlier studies demonstrated that altered root traits, including more root hair and root length, increasing adventitious roots and shallower root development of basal roots were noticed in G19833 genotype under Pi deficient conditions (Yan et al. 2004; Beebe et al. 2006). In addition to modified root morphology and root architecture, Northern and RT-PCR data illustrated that enhanced expression of PvPS2 and PvPT1 collectively contribute to the acquisition and translocation of Pi under limited P environment in the Pi efficient genotype. This study highlights the complex molecular changes occurring in Pi efficient common bean genotype during Pi starvation, concomitant with morphological and physiological changes reported earlier.


This research is in part supported by funds from the Collaborative Crop Research Program of the McKnight Foundation to K. G. Raghothama and X. L. Yan and the Key Basic Research Special Funds of China to X. L. Yan (2005CB120902).

Supplementary material

425_2007_603_MOESM1_ESM.doc (44 kb)
Supplement Figures (DOC 44 kb)

Copyright information

© Springer-Verlag 2007