Introduction

Phytic acid (myo-inositol-1,2,3,4,5,6-hexakisphosphate, InsP 6), beside being the storage form of phosphorus in seeds, is also a major seed antinutrient. In fact it contributes to reduce zinc and iron bioavailability of in the small intestines of humans and monogastric animals when legumes and grains are eaten, and this can lead to serious alimentary deficiencies, typical of humans living in the developing countries where staple foods are mainly seed-based (Sandberg 2002). Phytic acid also contributes substantially to environmental pollution, since non-ruminant animals are unable to metabolize it, and the undigested phytic acid promotes water eutrophycation. Thus, the negative effects of InsP 6 on nutrition and the environment are the main reason for developing low-InsP 6 crops (Abelson 1999; Raboy 2006).

In the last decade a great deal of interest has emerged in the possible health-beneficial roles of seed dietary phytic acid and several articles have reported the role of InsP 6 in cancer prevention and control of experimental tumour growth, progression, and metastasis. In addition, InsP 6 possesses antioxidant activity and other significant benefits for human health, such as the ability to enhance immune system, prevent pathological calcification and kidney stone formation, lower elevated serum cholesterol, and reduce pathological platelet activity (reviewed by Vucenik and Shamsuddin 2006; Graf et al 1987; Grases et al. 2008; Zhang et al. 2005).

On the light of these data, any manipulation of InsP 6 content in the seed, either directed towards increasing or lowering InsP 6, requires knowledge of the key enzymes involved in its biosynthetic pathway.

A simple two-step de novo myo-inositol (Ins) biosynthetic pathway is shared by all organisms. The Ins ring is provided only through the conversion of glucose 6-phosphate to 1D-myo-inositol 3-phosphate (Ins(3)P 1 ), carried out by D-myo-inositol 3-phosphate synthase (MIPS, EC 5.5.1.4) (Loewus and Murthy 2000). The second step is the hydrolysis of Ins(3)P 1 to Ins and Pi (inorganic P) catalyzed by inositol monophosphatase (IMP, EC 3.1.3.25). This reaction is reversed by the action of myo-inositol kinase (MIK, EC 2.7.1.6.4). The finding of lpa mutants defective in MIPS and MIK activities (Hitz et al. 2002; Shi et al. 2005; Kim et al. 2008) demonstrates the importance of these enzymes for phytic acid synthesis and accumulation in seeds. Conversely, no lpa mutants in IMP function have been described so far; however, Fu et al. (2008) reported that IMP expression is reduced in developing seeds of barley lpa mutants that accumulate Ins, thus suggesting that the control of Ins levels might be important for “feeding” the phytic acid pathway.

Phosphorylation of the of myo-inositol ring may occur either by a “lipid-independent” pathway, entirely independent of inositol lipid synthesis, or by a “lipid-dependent” pathway or via a combination of the two (Stevenson-Paulik et al. 2005). In the “lipid-dependent” pathway myo-inositol is converted to phosphatidylinositol (PtdIns) by a phosphatidylinositol synthase (PtdIS), PtdIns is then sequentially phosphorylated to produce PtdIns(4,5)P 2 . This molecule is the substrate of a PtdIns-specific phospholipase C activity that releases Ins(1,4,5)P 3 , a molecule central to signal transduction (York et al. 1999; Odom et al. 2000; Fujii and York 2005; Seeds et al. 2004). In the “lipid-independent” pathway myo-inositol phosphate is sequentially phosphorylated to lead to the formation of phytic acid (reviewed by Raboy and Bowen 2006). These sequential phosphorylation steps involve at least three type of inositol kinases: inositol 1,4,5-trisphosphate kinase (IPK2, EC 2.7.1.151 or EC 2.1.7.140), inositol 1,3,4-trisphosphate 5/6-kinase (ITPK, EC 2.7.1.134) and inositol 1,3,4,5,6 pentakisphosphate 2-kinase (IPK1, EC 2.7.1.158). The involvement of these inositol kinases in InsP 6 synthesis is supported by the isolation of corresponding lpa mutants (Shi et al. 2003; Stevenson-Paulik et al. 2005).

Our main aim is to lower phytic acid amount in common bean (Phaseolus vulgaris L.) seed and towards this goal we have undertaken a two-strategy approach: the isolation of mutants from an EMS mutagenised population (Campion et al. 2009) and the identification of genes coding for candidate enzymes involved in inositol phosphate metabolism for future targeted mutant identification. Here we report results of this second approach regarding the isolation and molecular characterisation of common bean structural genes important for phytic acid synthesis in the seed. We used bioinformatic tools to identify and map the bean homologues coding for myo-inositol 1-phosphate synthase, inositol monophosphatase, myo-inositol kinase, inositol 1,4,5-tris-phosphate kinase, inositol 1,3,4-triphosphate 5/6-kinase and inositol 1,3,4,5,6 pentakisphosphate 2-kinase. We also present data on the genomic organisation of these genes and on their expression during seed development as well as in other plant tissues.

Materials and methods

Plant material

Common bean plants (cv. Taylor’s Horticultural and breeding line 905, Campion et al. 2009) were grown in a phytotron under a 16/8 h light/dark photoperiod, with temperatures set at 25/18° day/night and relative humidity between 60 and 70%.

Seeds and pods were collected from bean plants at different developmental stages over a period of several Days After Flowering (DAF). At 4–6 DAF only pods could be collected while after 12 DAF the cotyledons were sufficiently developed to allow the removal of the seed tegument and the embryo. Leaves, roots and epicotyls were sampled at 8 or 18 days After Germination (DAG). All collected samples were immediately frozen in liquid nitrogen and stored at −80°C until use.

Sequence analyses, genes and cDNAs isolation

Genes related to the biosynthesis of phytic acid in P. vulgaris were identified by searching the public database at the NCBI website (http://www.ncbi.nlm.nih.gov) using the tBLASTx program (Altschul et al. 1997). Genes from Zea mays L., Arabidopsis thaliana (L.) Heynh. and Glycine max (L.) Merr. were used as query against plant nucleotide and EST databases. Nucleotide sequences were multialigned using the ClustalW program (http://www.ebi.ac.uk/Tools/clustalw2/index.html). Primers for PCR were designed based on conserved regions found in the resulting alignments (Table IS reports primers and annealing temperatures used for target genes isolation). PCR amplifications were performed using leaf genomic DNA or cDNA from developing cotyledons (12 DAF) or leaf (only for PvMIPSv isolation) (see Table IS for more details). cDNAs were obtained from total RNA using the reverse transcriptase Superscript II (Invitrogen), according to the protocol suggested by the manufacturer. To obtain full-length cDNA sequences, 3′-rapid amplifications of cDNA ends (RACE)-PCR were carried out with the 3′-RACE system (Invitrogen), using a forward-specific primer in combination with AUAP (Invitrogen).

Multiple alignment of deduced amino acid sequences as well as the related neighbor-joining branched tree were done using Clustal W software at http://align.genome.jp/.

All the genomic and cDNA sequences isolated were deposited in the EMBL Nucleotide Sequence Database (accession numbers are reported in Table IS).

DNA and RNA isolation and blot analysis

For Southern hybridization, 5 μg of genomic DNA, extracted from leaf as described by Dellaporta et al. (1983), were digested with EcoRI and HindIII, respectively, resolved on a 0.8% agarose gel and transferred to a nylon membrane (Hybond™ N+, GE Healthcare). Total RNA was isolated from leaves, immature pods or cotyledons as described by van Tunen et al. (1988). Ten microgram of RNA from each sample were resolved on 1.2% (w/v) agarose/formaldehyde/MOPS gel and transferred to a nylon membrane (Hybond™ N+, GE Healthcare).

The isolated cDNA sequences coding for the genes of the phytic acid pathway were labelled with α-[32P]-dCTP using a random primer DNA labelling kit (Fermentas), according to the manufacturer’s instructions, and used for membrane hybridisations as described in Sparvoli et al. (1994). Membranes were washed at 0.1× SSC, 0.5% (w/v) SDS at 65°C.

In silico mapping

Bean genes of phytic acid pathway were mapped onto the soybean genome by “virtual Southern blot” based on similarity searches (Salse et al. 2004) using BLASTn tools with their standard servers’ settings (http://www.phytozome.net/search.php?show=blast&targetType=genome&method=Org_soybean). For each bean gene the corresponding cDNA was used as query (Table 1 column I). The identified genomic region containing the soybean hortologous genes (Table 1 columns II, IV) was then analysed at the web site http://soybase.org/gbrowse/cgi-bin/gbrowse/gmax1.01 to look for anchored bean markers surrounding the soybean hortologous gene. The position of each marker was verified on the common bean reference genetic map of McClean (NDSU) 2007, at the Legume Information System web site using the cMAP software (http://lis.comparative-legumes.org/cgi-bin/cmap/viewer?changeMenu=1). Additional bean markers proximal to the first one were placed on the soybean genome in order to identify the closest one linked to the soybean homolog of each gene. An E value of 10−10 was adopted to claim a significant match between bean marker sequence and the soybean genome. In cases of hits with 10−10 < E value < 10−3 these were also considered to be syntenic with the soybean genome when consistent with the other markers’ synteny results. Syntenous regions were defined based on the number and relative order of markers supporting colinearity with the common bean reference map of McClean (NDSU) 2007.

Table 1 List of the closest bean markers (column V) to the bean genes of the phytic acid pathway (column I)
Full size table

Semi-quantitative RT-PCR and quantitative real-time RT-PCR

For semi-quantitative RT-PCR the pairs of primers K1-574F and K1-495, K3-343F and K3-EndR, K3-3-97F and K3-3-EndR, were used for amplifying PvIPK1, PvITPKα and PvITPKβ genes, respectively (Table IS). For quantitative real-time RT-PCR, primers were designed using the ABI-PRISM Primer Express™ Software V2.0 (Applied Biosystems) (gene specific pairs of primers are reported in Table IIS).

Total RNA, extracted as above, was treated with DNAse I by incubating 10 μg of RNA with 2 U of DNAse I (Roche) at 37°C for 30 min and then heat inactivated for 15 min at 75°C. cDNA was produced from 1 μg of total RNA using the “High Capacity cDNA Archive kit” (Applied Biosystems) according to the manufacturer’s protocol (RT step). Each reaction for semi-quantitative RT-PCR was performed in 20 μl containing 10 μl of 2× GoTaq Master Mix (Promega), 0.4 μM each of forward and reverse primers and 50 ng of cDNA. The PCR amplification program was: step 1: 94°C 5 min; step 2: 94°C 30 s, 55°C 30 s, 67°C 1 min (we used 25 cycles for 18S rRNA and 35 cycles for PvIPK1, PvITPKα and PvITPKβ genes); step 3: 67°C 10 min. Control RNA samples were run without the RT step to confirm the absence of any genomic contamination.

Quantitative real-time RT-PCR runs were performed in triplicate in 96-well optical plates using an ABI 7300 Sequence Detection Systems (Applied Biosystems). Each reaction was in 25 μl containing 11.25 μl of 2.5× Real Master MIX/20X SYBR Solution (Eppendorf), 0.5 μM each of forward and reverse primers and 7.5 ng of cDNA. For the 18S rRNA detection (reference gene), 0.1 ng of cDNA were used. The PCR amplification program was: step 1: 50°C 2 min; step 2: 95°C 10 min; step 3 95°C 15 s, 60°C 1 min 95°C repeated for 40 cycles; step 4: 95°C 15 s, 60°C 30 s, 95°C 15 s. Control RNA samples were run without the RT step to confirm the absence of any genomic contamination.

PCR efficiencies of target and reference genes were determined by generating standard curves. For each target gene, relative transcript levels were normalized to the 18S rRNA reference gene level and calculated using the 2−∆∆CT method (Livak and Schmittgen 2001) following indications described in the SDS ABI 7300 User manual.

Results

Identification and characterization of structural genes of the phytic acid biosynthetic pathway

myo-inositol 1-phosphate synthase

In bean two different MIPS protein forms have been reported: one is specifically expressed in developing seeds while the other one is localized in the plastids (Johnson and Wang 1996). In addition, the sequence of a MIPS cDNA (PvMIPSr, GenBank accession U38920) expressed in roots has also been reported (Wang and Johnson 1995). Surprisingly, this sequence is very similar (99% of identity) to the Arabidopsis AtMIPS1 gene (NCBI Reference Sequence NP_195690).

In order to isolate the gene coding for the MIPS form expressed in developing seeds we performed RT-PCR on cDNA derived from developing common bean seeds using pairs of primers designed on the basis of available bean EST sequences and on the partial genomic common bean MIPS (GenBank accession AF282263) present in the GenBank database (www.ncbi.nlm.nih.gov). A full length cDNA of 1,736 bp (called PvMIPSs) was isolated and shown to contain an open reading frame of 1,533 bp encoding for a polypeptide of 510 amino acids. When leaf genomic DNA was used as the template for PCR, two types of MIPS genes were isolated: one is 2,671 bp long, contains nine introns and its transcript was 100% identical to PvMIPSs isolated from developing cotyledons cDNA. The other gene shares a high degree of sequence identity (88%) with PvMIPSs, from which it differed mainly for the length of the third intron (Fig. 1). Analysis of the sequences in the bean EST database, together with expression analyses by real-time RT-PCR (Fig. 2) showed that this second gene is expressed in several plant organs, but is poorly represented in developing seeds. It was therefore referred to as PvMIPSv (MIPS vegetative) to differentiate it from the PvMIPSs (MIPS seed). The phylogenetic tree derived from the multiple alignment of the deduced amino acid sequences of the bean genes together with those derived from soybean and other species’ MIPS genes confirmed that PvMIPSs clusters together with GmMIPS1, which is highly expressed in the cotyledons, while PvMIPSv is in the same branch with GmMIPS3 and GmMIPS4, which are expressed in vegetative tissues and poorly, or not at all, in seed tissues (Fig. 3) (Chappell et al. 2006).

Fig. 1
figure 1

Intron-exon structure of the bean PvMIPSs and PvMIPSv genes. Exons are indicated by boxed numbers, lines represent introns, upward arrows indicates sites for restriction enzymes

Full size image
Fig. 2
figure 2

Quantitative real-time RT-PCR analysis of PvMIPs and PvMIPSv on cDNAs from different bean plant tissues. Each bar represents the mean of three replicates with the indicated standard deviation of the mean. DAF indicates Days After Fertilisation; DAG indicates Days After Germination

Full size image
Fig. 3
figure 3

Phylogenetic tree of MIPS proteins. Accession numbers are: GmMIPS (AAK72098), GmMIPS1 (ABC55419), GmMIPS2 (ABC55420), GmMIPS3 (ABC55421), GmMIPS4 (ABC55422), ZmMIPS (AAG40328), OsMIPS (BAA25729), AtMIPS1 (NP_195690), PvMIPSr (Q41107), PeMIPS (ABF51620), SiMIPS (AAG01148), PvMIPs (CAH68559), PvMIPSv (CAJ15162)

Full size image

The presence of two different MIPS genes as well as their copy number was determined by Southern blot hybridization (Fig. 4a). The hybridisation pattern obtained with the two genes clearly indicated the presence of one copy for each MIPS gene. In fact, comparing the two blots it was evident that the stronger hybridising fragments obtained with the PvMIPSs probe corresponded to the weaker ones recognised by the PvMIPSv probe. Moreover, when genomic DNA was digested with HindIII and hybridised to the PvMIPSs probe, an additional hybridising fragment was detected that corresponded to the 5′ region of the gene which contains a HindIII site which is absent in the PvMIPSv gene (Fig. 1, 4a).

Fig. 4
figure 4

Southern blot analysis of genes of the phytic acid pathway. Genomic DNA was digested with the restriction enzymes EcoRI or HindIII. Filters were hybridised to a PvMIPSs or PvMIPSv, b PvIMP, c PvMIK, d PvIPK2, e PvITPKα, f PvITPKβ, g PvIPK1. On the left side of each blot are reported molecular weights in kb

Full size image

Inositol monophosphatase (PvIMP)

The bean cDNA coding for IMP (PvIMP) was isolated by RT-PCR from developing seed cDNA using pairs of primers designed on the basis of P. vulgaris ESTs (GenBank accessions CV532534, CV538392 and FD787042) homologous to LeIMP1 (GenBank accession AY227666) reported in tomato. The isolated PvIMP cDNA is 900 bp long and codes for a polypeptide of 272 amino acids. The amino acid sequence coded by PvIMP was aligned to the sequences of other IMP proteins from tomato, barley, Arabidopsis, rice and ice plant. Amino acid residues reported by Bone et al. (1992) to be involved in metal or sulphate binding (highlighted residues in Fig. 1S) as well as motifs A (WxI/VDPI/LDxTxxF/Y/IxH/K) and B (WDxxA/GA/GxA/I/LI/L/VA/L/VxxxGA/G) described by Neuwald et al. (1991) (boxed residues in Fig. 1S) were highly conserved among all the sequences, thus suggesting that PvIMP is endowed with inositol monophosphatase activity.

On the basis of the alignment between the PvIMP cDNA and the LeIMP1 genomic sequence, we designed other pairs of primers in order to isolate the corresponding entire bean genomic sequence. The genomic sequence of PvIMP was 5,862 bp long and was organised with a 12-exons-11-introns structure. From the comparison of the intron-exon structure of PvIMP with IMP genes from other species we observed that dicot species share the 12-exons-11-introns structure, while monocots have a 10-exons-9-introns organisation, in which monocot exon 1 and exon 3 are equivalent to dicot exon 2 and exons 4 and 5, respectively (Fig. 5).

Fig. 5
figure 5

Comparison of bean PvIMP exon-intron structure with that reported for other IMP genes. Exons are indicated by boxed numbers, lines represent introns, upward arrows indicates sites for restriction enzymes. Accession numbers are: HvIMP (DQ145527), OsIMP (AK071149), AtIMP (AY035150), LeIMP1 (AY227666), PvIMP (FN357287)

Full size image

The number of PvIMP genes present in the bean genome was estimated by Southern blot analysis (Fig. 4b). Digestion of leaf genomic DNA with EcoRI and HindIII showed the presence of few hybridising fragments. Sites for these two restriction enzymes are present in the PvIMP sequence (three for EcoRI and one for HindIII) and the length of the major EcoRI hybridising fragment (about 5,000 bp) together with the number of hybridising fragments obtained with both enzymes is consistent with the hypothesis of the presence of a single copy of the IMP gene in bean (Fig. 4b, 5).

Myo-inositol kinase

Myo-inositol kinase (MIK) phosphorylates myo-inositol to produce Ins(3)P. The involvement of this activity in the production of phytic acid has been shown by Shi et al. (2005) who demonstrated that in the maize lpa3 the MIK gene is disrupted by a Mu transposon insertion. In the lpa3 mutant the decrease in phytic acid is accompanied by an increase in myo-inositol, while myo-inositol phosphate intermediates are not present at significant levels. These characteristics suggest that MIK is not just a salvage enzyme for myo-inositol recycling, but is an important biosynthetic step of phytic acid pathway (Shi et al. 2005).

Since it was not possible to identify any putative MIK sequence in the bean EST database, we used the maize MIK (GenBank accession AY772410) as a query to search genomic and ESTs plant databases to find other putative MIK sequences in order to build a multiple sequence alignment for primer design. The putative MIK sequences used were those of soybean (GenBank accession CS236223), Lotus japonica (GenBank accession BW619491), Medicago truncatula (GenBank accession EV255698), Cicer arietinum (CK148972) and those already reported in rice, maize and Arabidopsis by Shi et al. (2005). The RT-PCR was conducted on developing seed cDNA and the amplification product (715 bp) was sequenced. We were able to obtain only a portion of the PvMIK sequence corresponding to a polypeptide of 238 amino acids that share an overall amino acid identity of 83, 58, 57 and 56% with MIK proteins of soybean, rice, Arabidopsis and maize, respectively.

In maize MIK is coded by a single-copy gene containing only one intron, and also in Arabidopsis a single MIK sequence with only one intron is present. The PvMIK cDNA corresponds to the region coded by exon 1 both in Arabidopsis and maize, therefore, assuming a similar gene organisation in bean, we can affirm that no EcoRI nor HindIII restriction sites are present in the PvMIK gene, since these enzymes do not cut the cDNA. From Southern blot analysis also the bean genome appears to contain a single copy of the MIK gene, in fact a single hybridising fragment is present in genomic DNA digested with EcoRI or HindIII (Fig. 4c). Weaker hybridising fragments are also recognised by the PvMIK probe, however, we hypothesise that they correspond to pseudogenes or distantly related genes, since the blot was washed in highly stringent conditions (0.1xSSC at 65° for 1 h).

Inositol 1,4,5-tris-phosphate kinase/inositol polyphosphate multikinase

Inositol 1,4,5-tris-phosphate kinase, or more appropriately inositol polyphosphate multikinase, is a dual-specificity InsP 3 /InsP 4 6-/3-kinase that sequentially generates InsP 5 from InsP 3 (Stevenson-Paulik et al. 2002; Shears 2004; Xia and Yang 2005). The involvement of this enzyme activity in phytic acid biosynthesis has been demonstrated by Stevenson-Paulik et al. (2005) who showed that the loss-of-function obtained by T-DNA insertion in the Arabidopsis AtIpk2β gene resulted in an accumulation of intermediate inositol phosphorylated forms and in a seed phytic acid reduction of about 35%.

To look for bean IPK2 homologues, the AtIpk2β gene of Arabidopsis was used as query in a tBLASTx search against the soybean EST database. A contig corresponding to the GmIPK2 was found and aligned with the AtIpk2β gene to allow the designing of PCR primers in conserved regions. By RT-PCR on cDNA from developing seeds we isolated a sequence of 1,116 bp, named PvIPK2, coding for a polypeptide of 264 amino acids. The alignment of this polypeptide to other plant IPK2 proteins showed a degree of identity ranging from 72% with the soybean Ipk2 to 48% with the rice sequence (OsIPK2) (Fig. 2S). The PvIPK2 protein contains the PxxxDxKxG conserved InsP 3 binding domain, which is a catalytic site for phosphate transfer from ATP to the inositol ring (Bertsch et al. 2000). The motif [L/M] [I/V]D[F/L][A/G] [H/K] which is considered a putative ATP/Mg2+ binding site is also present, together with the “SSLL” domain necessary for enzymatic activity (Saiardi et al. 2001).

The PvIPK2 gene was intronless as shown by PCR on genomic DNA. Indeed, Arabidopsis and rice IPK2 are also intronless genes. Southern blot analysis showed only one strong hybridising fragment when genomic DNA was digested with restriction enzymes EcoRI and HindIII. The latter yields a hybridising band of about 3,400 bp, suggesting that PvIPK2 is a single copy gene (Fig. 4d).

Inositol 1,3,4-triphosphate 5/6-kinase

The family of inositol 1,3,4-triphosphate 5/6-kinases belongs to a larger family of ATP-grasp fold proteins. These also show some functional and structural similarity with inositol polyphosphate multikinases, which include inositol hexakisphosphate kinase and inositol 1,4,5-triphosphate 3-kinase. In plants this type of activity is coded by at least three different types of genes (Suzuki et al. 2007; Sweetman et al. 2007; Stiles et al. 2008), among which only one has been demonstrated to be necessary for phytic acid synthesis in the seed (Shi et al. 2003). We used this last sequence, together with the Arabidopsis homologous gene, At5g16760, as a query against the bean EST database. Using the tBLASTx program we found one EST showing 66 and 68% similarity to the maize and Arabidopsis sequences, respectively. Primers for PCR were based on this bean EST and allowed the isolation of a PCR fragment of 1,017 bp coding for a polypeptide of 317 amino acids, both from seed cDNA and leaf genomic DNA. The isolated gene, named PvITPKα, did not show the presence of any intron and, most likely, is a single copy gene, since from Southern blot analysis with two different restriction enzymes only one hybridising band is detected in high stringency conditions (Fig. 4e).

Although a role in seed phytic acid synthesis has been proven only for the maize lpa2 gene, other genes coding for inositol 1,3,4-triphosphate 5/6-kinases have been isolated and shown to be expressed in developing seeds (Suzuki et al. 2007; Sweetman et al. 2007; Stiles et al. 2008). Therefore, we decided to look for other members of this family of proteins. The derived amino acid sequences of soybean, rice and Arabidopsis ITPKs, together with those from potato and barley, were multiple aligned and the resulting phylogenetic tree confirmed the existence of three types of genes that grouped into three branches (Fig. 6). Soybean genes, belonging to β-branch, as well as Arabidopsis and rice genes belonging to γ-branch, were used as query in a tBLASTx search against bean ESTs. Matching ESTs were found only for β-type ITPK sequences and used to design primers for RT-PCR on seed cDNA. An amplification product of 658 bp was isolated and shown to correspond to a partial cDNA, named PvITPKβ, of which the derived amino acid sequence was 96% identical to that coded by the GmITPK3 gene. We also performed PCR amplification on leaf genomic DNA and obtained a sequence of 1,903 bp containing a portion of the PvITPKβ gene (Fig. 7). A comparison of the genomic sequences of β-type sequences from Arabidopsis and barley shows they have 10 exons and 9 introns. The alignment of the bean PvITPKβ genomic fragment with these genes indicates that PvITPKβ covers a region that spans exon 5 to exon 10 (Fig. 8). All the β-type genes have corresponding exons of similar length, while the length of the corresponding introns is more variable, even among genes from the same species (rice).

Fig. 6
figure 6

Phylogenetic tree of ITPK proteins. The three subgroups α, β and γ are indicated. Accession numbers are: GmITPK-1 (ABU93831), GmITPK-2 (ABU93832), GmITPK-3 (ABU93833), GmITPK-4 (ABU93834), AtITPK-1 (NP_197178), AtITPK-2 (NP_849342), AtITPK-3 (ABJ98567), AtITPK-4 (NP_850407), HvITPK (CAL49035), OsITPK-1 (BAF25933), OsITPK-2 (BAF11371), OsITPK-3 (BAF13047), OsITPK-4 (Q84Y01), OsITPK-5 (BAF27322), OsITPK-6 (BAF25603), StITPK (CAM12754), ZmITPK (NP_001105901), PvITPKα (CAP09175), PvITPKβ (CAY10403)

Full size image
Fig. 7
figure 7

Comparison of bean PvITPKβ exon-intron structure with that reported for other β-type ITPK genes. Exons are indicated by boxed numbers, lines represent introns, upward arrows indicates sites for restriction enzymes. Dashed lines indicate regions not sequenced. Accession numbers are: AtITPK2 (NM_179011), OsITPK1 (NC_008403), OsITPK3 (NC_008396), PvITPKβ (FN357285)

Full size image
Fig. 8
figure 8

Comparison of bean PvIPK1 exon-intron structure with that reported for other IPK1 genes. Exons are indicated by boxed numbers, lines represent introns, upward arrows indicates sites for restriction enzymes. Dashed lines indicate regions not sequenced. Accession numbers are: AtIPK1 (NP_568613), OsIPK1 (NP_001054147), ZmIPK1 (NP_001106063), PvIPK1 (CAM33431)

Full size image

Southern blot analysis on genomic DNA digested with EcoRI or HindIII enzymes indicates that PvITPKα is a single copy gene, while we cannot exclude the presence of two copies for PvITPKβ (Fig. 4f).

Inositol 1,3,4,5,6 pentakisphosphate 2-kinase

The AtIpK1 gene of Arabidopsis, reported to be necessary for the production of phytic acid (Stevenson-Paulik et al. 2005), was used as query against the plant gene database. In fact, we could not identify any bean EST coding for inositol 1,3,4,5,6 pentakisphosphate 2-kinase. Full-length cDNAs and genomic DNAs with high similarity (66–82%) to the Arabidopsis AtIpk1 sequence were found in soybean, rice and maize. In addition, highly homologous ESTs of soybean, barley, and sorghum were also identified. Multiple sequence alignment of the above identified IPK1 sequences revealed several motifs defining conserved regions present in all IPK1 members. Then, several pairs of PCR primers were designed along these conserved regions and used for PCR amplification on leaf genomic DNA. A 2,258 bp genomic sequence, named PvIPK1, containing an open reading frame of 352 amino acids (from exon 1 to exon 6) was isolated (Fig. 8). The analysis of the exon-intron structure of the IPK1 genomic clones of Arabidopsis, maize, rice and bean indicate that it is conserved among these species, although the length of corresponding introns may be quite different (Fig. 8). The deduced amino acid sequence of the PvIPK1 gene was aligned with those of soybean, Arabidopsis, maize and rice (Fig. 3S) and shown to have a similarity ranging from 83% with the soybean protein to 67% with that of rice. The PvIPK1 protein also contains the conserved motif boxes A, C, E, F and G described by Verbsky et al. (2002), Sweetman et al. (2006) (boxes A-E), and Sun et al. (2007) (boxes F, G).

In order to understand the genomic organization of the PvIPK1 gene a Southern blot analysis was performed. Under stringent washing conditions (0.1xSSC) one strong and one weaker hybridising fragments were detected using the PvIPK1 cDNA as probe on genomic DNA digested with EcoRI and HindIII (Fig. 4g). In the PvIPK1 genomic sequence HindIII cuts twice and the expected hybridization fragments of about 1,900 and 500 bp were detected. Since the PvIPK1 restriction pattern resulting from Southern hybridisation results is coherent with the size of expected HindIII and EcoRI fragments, the results indicate the presence of only one copy of the PvIPK1 gene in the genome of common bean.

In silico mapping of genes of the phytic acid pathway

Taking advantage of the high similarity and synteny between soybean and common bean genomes (Choi et al. 2004) we used an in silico approach to position each gene reported in this paper on bean genetic maps. Soybean has a polyploid genome (2n = 40) that is predicted to have undergone duplication after divergence from common bean (2n = 22). As expected, we found two hortologous soybean genes for each bean cDNA, the only exceptions being PvMIPSs and PvIPK2 genes for which only one soybean homolog could be identified (Table 1; Fig. 4S). In most cases little rearrangements in the order of bean markers on the corresponding soybean chromosomes were observed (Fig. 4Sa–h).

The results of the comparative mapping analyses allowed the positioning of PvMIPSs and PvMIPSv on different soybean chromosomes: Gm18 and the two syntenic Gm05 and Gm08, respectively. The subsequent analysis based on adjacent bean anchored markers indicated that the syntenic regions containing these genes are located on bean chromosomes Pv1 and Pv2, respectively (Fig. 4Sa–b). The closest bean marker was at about 210 Kbp in the case of the soybean MIPSs gene (Glyma18g02210) and at 28,136 and 32,399 bp in the case of soybean MIPSv genes Glyma05g31450 and Glyma08g14670, respectively.

Distances between the closest bean markers and the other soybean genes were in the range of those observed for MIPS genes. The only exception was observed for PvITPKα, in fact, although the genomic regions surrounding the soybean hortologous were syntenic to the bean chromosome Pv3, the closest bean markers we could find were g686 and g2371 distant 599,924 bp from Glyma05g02250 and 993,019 bp from Glyma17g09680, respectively (Fig. 4Sc). Conversely, in the case of PvIMP, which maps on bean chromosome Pv6 (Fig. 4Sd), the marker g2435 was only 1,131 and 1,139 bp from Glyma09g01380 and Glyma15g12230, respectively.

Mapping of PvIPK2 gene required the use other bean maps together with that of McClean (NDSU) 2007. In fact, the bean marker Bng070, that was 326,417 bp from the soybean Glyma12g36970 and mapped on J linkage group by Vallejos et al. (1992), was not present in the McClean (NDSU) 2007 map (Fig. 4Se). However, using the cMAP software it was possible to align the J linkage group to Pv11 of McClean (NDSU) 2007 map and this allowed the finding of the g188 marker that was much closer (26,966 bp) than the Bng070 to the soybean Glyma12g36970 gene.

Expression of genes involved in phytic acid biosynthesis

In bean seeds phytic acid concentration begins to increase after 16 days after fertilisation (DAF) and reaches a plateau after 26 DAF, while InsP pools (InsP 3 to InsP 5 ) peak mainly at 21 DAF (Coehlo et al. 2005), therefore the expression of the genes involved in the phytic acid pathway was monitored at different stages of seed development, ranging from 4 DAF up to 24 DAF. We performed the analyses either by northern blot or, in the case of those genes that were expressed at low levels, by semi-quantitative RT-PCR (Fig. 9).

Fig. 9
figure 9

Expression analysis of genes of the phytic acid pathway during bean seed development. Panel a: total RNA before blotting; panels b–e: northern analysis using as probes PvMIPS, PvIMP, PvMIK and PvIPK2, respectively. Panels f–i: semiquantitative RT-PCR of PvITPKα, PvITPKβ, PvIPK1, and 18SrRNA, respectively. M indicates molecular weights in kb; DAF indicates Days After Fertilisation

Full size image

Maximal expression of PvMIPS was observed at the earliest development stages analyzed, i.e. pods at 4–6 DAF, in which the seed is still almost undetectable and the embryo is at the globular stage. As seed development progressed, PvMIPS transcripts decreased to nearly undetectable levels (cotyledons at 14 DAF). Since the two PvMIPS genes are highly homologous, in order to establish which one is specifically expressed in the seed we performed a quantitative real-time RT-PCR analysis, using gene specific primers, on the same samples used for northern analysis (pods at 6 DAF and cotyledons at 12 and 14 DAF) and in other plant tissues: epicotyls and young leaves taken from 8 days after germination (DAG) seedlings, leaves, stems and roots from 18 DAG seedlings and flowers (Fig. 2). The results obtained demonstrated that PvMIPSs is highly expressed in cotyledons and at much lower levels in flowers as well as in all the other tissues analysed. By contrast, PvMIPSv expression is higher in all seedling tissues with the exception of roots where its level is similar to that of PvMIPSs.

A clearly regulated expression is also found for PvIMP and PvIPK2 genes during seed development, while the other genes of the pathway do not appear to be significantly regulated during seed development. Their expression was maintained at very similar levels along all the stages analysed, although a very weak reduction can be observed during the latest developmental stages (20–24 DAF) for the PvITPKβ gene, while a very faint peak of expression is detectable at 12–14 DAF for the PvITPKα gene.

We also decided to investigate the expression of PvITPKα, PvIPK2 and PvIPK1 in different plant tissues (Fig. 10). The PvIPK1 gene is expressed at very low levels in all tissues analysed, a result very similar to that observed by Suzuki et al. (2007) in rice. A similar expression was found for the PvITPKα gene. On the contrary, PvIPK2 was highly expressed in developing cotyledons (indeed this was the only gene for which expression could be easily detected by Northern analysis), thus confirming its key role for phytic acid synthesis in this tissue. In the other tissues analysed, PvIPK2 expression was between 4- and 16-fold lower than that found in cotyledons: the lowest expression was in epicotyls and leaves, while the highest was in flowers.

Fig. 10
figure 10

Quantitative real-time RT-PCR analysis of PvIPK1, PvIPK2, PvITPKα, on cDNAs from different bean plant tissues. Each bar represents the mean of three replicates with the indicated standard deviation of the mean

Full size image

Discussion and conclusions

In this paper we report the isolation, molecular characterisation and in silico mapping of a full set of structural genes involved in phytic acid synthesis from a single food crop species, P. vulgaris.

Mutants impaired in phytic acid accumulation have been reported in several species for almost all the genes described here, however, a direct connection to phytic acid accumulation is still lacking for IMP and ITPKβ. Inositol monophosphatase catalyzes the dephosphorylation of Ins(3)P to produce free myo-inositol. No lpa mutant with a defective IMP gene has been reported yet, however, a decrease in IMP-1 expression has been shown in four lpa barley mutants (lpa1-1, lpa2-1, lpa3-1 and M955), in which the level of reduction appeared fairly proportional to both phytic acid decrease and myo-inositol increase (Fu et al. 2008). However, the role of such type of gene in phytic acid synthesis is still unclear, in fact the protein it codes may be a l-Galactose 1-P phosphatase important to ascorbic acid synthesis, or may be a dual-function as both an l-Gal 1-P phosphatase/IMPase (Laing et al. 2004; Torabinejad et al. 2009). Plant genomes also contain other distantly related “IMPase-like” proteins, more related to prokaryotic IMPases, which very recently have been shown to have IMPase activity (Torabinejad et al. 2009). Therefore, we cannot exclude that these IMPase-like proteins may be more important to inositol supply for phytic acid synthesis then the previously reported IMPase, that was used in this study for PvIMP gene identification. However, our finding that PvIMP is developmentally regulated during seed maturation together with the fact that lpa mutants show a decrease in the expression of this gene proportional to both phytic acid decrease and myo-inositol increase (Fu et al. 2008), is a good indication supporting a role of this gene in phytic acid biosynthesis. In addition, we have observed a lower expression of PvIMP in developing seeds of bean lpa-280-10 mutant compared to wt (manuscript in preparation), thus confirming what reported in other lpa mutants.

Our work advances efforts to analyse the complete set of genes and functions important to phytic acid synthesis in seeds. In fact, although the phytic acid pathway has been studied in some detail in several species, i.e. rice, barley, maize, Arabidopsis and soybean, each of these species lacks a detailed description of some specific genes of the pathway. For example, in maize no data are available for IMP and IPK2 genes, while in rice, for which all the genes of the pathway have been identified, data are lacking for MIK gene copy number and MIK, IMP and IPK2 expression during seed development. Here, we demonstrate that beside MIPS also IMP and IPK2 expression are regulated during seed development, underlining the key role of these genes in phytic acid biosynthesis, a result that is supported by the finding that these genes are down-regulated in the bean lpa-280-10 mutant (manuscript in preparation). Moreover, we provide for the first time a clear picture of the expression during seed development of the genes involved in phytic acid synthesis, including those such as MIK, IMP and IPK2, for which this information was lacking.

MIPS enzyme is necessary for myo-inositol production and is the first and rate-limiting one in the synthesis of all inositol-containing compounds. We show that a specific MIPS gene (PvMIPSs) is specifically expressed during seed development and its highest levels of expression are detected at very early stages (from 4 to 12 DAF). This very early MIPS expression in seed development has been reported also in other species like rice (Yoshida et al. 1999) and soybean (Hegeman et al. 2001; Chiera and Grabau 2007). Chiera and Grabau (2007) studied MIPS immunolocalisation during soybean seed development and showed that at very early stages GmMIPS is expressed in maternal tissues and then in embryo, thus suggesting a key role in supplying inositol for early embryo development. Of course MIPS should also plays a role in phytic acid synthesis and indeed highest levels of PvMIPSs expression precede synthesis and accumulation of inositol phosphates (InsP 3 to InsP 5 ) which synthesis begins around 12 DAF and peaks at 21 DAF (Coehlo et al. 2005).

We found that also PvIMP and PvIPK2 expression is regulated during seed development, however, they have different timing of expression: while PvMIPSs is expressed at very early stages (4–12 DAF), PvIMP expression declines slightly later (14–16 DAF) and finally PvIPK2 expression peaks between 12–14 DAF and slowly decreases during the later developmental stages. This type of regulation suggests a different and sequential time of action for the enzymatic activities coded by these genes and is in agreement with the timing of synthesis and accumulation of phytic acid and InsPs (InsP 3 to InsP 5 ) reported by Coehlo et al. (2005) for a number of bean varieties.

In rice, OsMIPS and OsIPK1 genes are developmentally regulated and their expression is coordinated (Suzuki et al. 2007). The other two genes analysed, OsITP5/6k-4 and OsITP5/6k-6 (belonging to the ITPKα and ITPKγ subgroups, respectively), show opposite trends for regulation: OsITP5/6k-4 expression increases after the very early stages of seed development and then remains quite constant, while that of OsITP5/6k-6 decreases. However, the authors also showed that these two genes are expressed in different seed tissues, aleurone and embryo, respectively, and suggested that the timing of expression reflects the timing of phytate accumulation in these tissues. Although we could not isolate an ITPKγ gene in bean and such a type of gene has not been reported in soybean (Stiles et al. 2008), a species very closely related to common bean, we cannot exclude that such type of gene could be present in the bean genome.

In any case, the expression of both PvITPKα and PvITPKβ does not resemble that observed for these two genes in rice, while it closely resembles that described by Stiles et al. (2008) for soybean ITPKα and ITPKβ type genes and by Josefsen et al. (2007) for a rice and barley ITPKα gene which is constitutively expressed in all tissues examined as well as during seed development.

Our findings also show that, although two copies of MIPS gene are present in common bean, they have tissue specific pattern of expression and are located in different genomic regions, being mapped on chromosomes Pv1 (PvMIPSs) and Pv2 (PvMIPSv). In addition, we provide evidence that all the other genes of the phytic acid biosynthetic pathway are single copy genes, a characteristic that minimises the risk of functional redundancy that could mask or reduce the effect of the mutation on phytic acid accumulation in seeds. This gene copy number is confirmed by the comparative mapping analysis: in fact, for each bean gene we never found more than two soybean hortologous. The comparative mapping analysis also allowed the finding of bean molecular markers closely associated to the hortologous soybean genes: distances were ranging from about 1 Kbp (between g2435 and the soybean IMP gene) up to 600–1,000 Kbp (between g686–g2371 and the soybean ITPKα gene). These markers may represent a useful tool for bean breeders, for example in QTL analysis of important agronomic traits.

Finally, bean TILLING populations are under development and will be soon ready for mutant identification by the reverse genetic approach (Gepts et al. 2008). We believe that our data will be useful for identification of bean lpa mutants using TILLING and/or ECOTILLING.