High mature grain phytase activity in the Triticeae has evolved by duplication followed by neofunctionalization of the purple acid phosphatase phytase (PAPhy) gene

Abstract

The phytase activity in food and feedstuffs is an important nutritional parameter. Members of the Triticeae tribe accumulate purple acid phosphatase phytases (PAPhy) during grain filling. This accumulation elevates mature grain phytase activities (MGPA) up to levels between ~650 FTU/kg for barley and 6000 FTU/kg for rye. This is notably more than other cereals. For instance, rice, maize, and oat have MGPAs below 100 FTU/kg. The cloning and characterization of the PAPhy gene complement from wheat, barley, rye, einkorn, and Aegilops tauschii is reported here. The Triticeae PAPhy genes generally consist of a set of paralogues, PAPhy_a and PAPhy_b, and have been mapped to Triticeae chromosomes 5 and 3, respectively. The promoters share a conserved core but the PAPhy_a promoter have acquired a novel cis-acting regulatory element for expression during grain filling while the PAPhy_b promoter has maintained the archaic function and drives expression during germination. Brachypodium is the only sequenced Poaceae sharing the PAPhy duplication. As for the Triticeae, the duplication is reflected in a high MGPA of ~4200 FTU/kg in Brachypodium. The sequence conservation of the paralogous loci on Brachypodium chromosomes 1 and 2 does not extend beyond the PAPhy gene. The results indicate that a single-gene segmental duplication may have enabled the evolution of high MGPA by creating functional redundancy of the parent PAPhy gene. This implies that similar MGPA levels may be out of reach in breeding programs for some Poaceae, e.g. maize and rice, whereas Triticeae breeders should focus on PAPhy_a.

Introduction

Phytases (myo-inositol hexakisphosphate phosphohydrolase; EC 3.1.3.8 and EC 3.1.3.26 are defined as phosphatases which can initiate the sequential hydrolysis of the major storage form of phosphorus in plant seeds, phytate (InsP₆, myo-inositol 1,2,3,4,5,6-hexakisphosphate). This provides bio-available phosphate, inositol phosphates, and eventually inositol required for a range of cellular activities (Brinch-Pedersen et al., 2002). Expression or activation of phytase and the concomitant hydrolysis of InsP₆ has been observed during germination in many plant species (Duff et al., 1994). The digestive tracts of non-ruminants possess negligible phytase activity and the activity provided by food and feedstuffs is often insufficient for efficient hydrolysis of the InsP₆ content. Unable to be absorbed, InsP₆ is excreted along with chelated cations. InsP₆ is therefore considered the single, most important antinutritional compound for the bioavailability of a range of essential minerals including Zn²⁺, Ca²⁺, and Fe²⁺ (Brinch-Pedersen et al., 2007).

Endogenous plant phytases can only be beneficial to nutrition if they are present at the stage of consumption. This usually means in the mature seeds. Among cereals, the Triticeae tribe grains exhibits a remarkable high mature grain phytase activity (MGPA) (Eeckhout and De Paepe, 1994). Triticeae phytases are have been proven beneficial during bread making and in the digestive tract of pigs and humans (Pointillart et al., 1987; Sandberg and Andersson, 1988; Türk et al., 1996). Indexing MGPA against wheat (Triticum aestivium L.; 100%) places barley (Hordeum vulgare L.) at 49%, rye (Secale cereale L.) at 430% and triticale (× Triticosecale, Wittmack) at 141%. In comparison, the non-Triticeae cereals oat (Avena sativa L.), maize (Zea maize L.) and rice (Oryza sativa L.) index at 3.5, 5.7, and 6.0%, respectively (Eeckhout and De Paepe, 1994; Lucca et al., 2001). Moreover, there is also variation within species, in wheat up to 2-fold differences between cultivars (Ram et al., 2010).

The cereal phytase complement has been found to consist of the multiple inositol polyphosphate phosphatase (MINPP) phytases and the purple acid phosphatase phytases (PAPhy) (Dionisio et al., 2007, 2011). For PAPhy, 12 wheat, barley, maize, and rice cDNAs expressed in the grain have been described. In a phylogenetic analysis including 42 plant purple acid phosphatases, the isolated PAPhy genes grouped in a single clade (PAP type 1). The wheat and barley PAPhy could further be grouped as PAPhy_a isoforms, predominantly expressed during grain filling and as PAPhy_b isoforms, predominantly expressed during germination. The isoforms can also be distinguished by the C terminal but are otherwise of very similar sequence. Biochemical characterization of the recombinant proteins demonstrated similar enzyme kinetic parameters except for a slightly higher pH optimum for the examined PAPhy_a and a different preference for metal cofactors. The PAPhy_a protein accumulates in the protein storage vacuoles of the aleurone cells in close proximity to the major InsP₆ deposits (Dionisio et al., 2011).

The MINPPs are also present in different isoforms and they are also expressed both during grain development and germination (Dionisio et al., 2007). However, purification of phytases from wheat bran and dry and germinating barley grains, in hindsight and based on molecular mass and N-terminal amino acid sequence, revealed the PAPhy as the major contributor to the MGPA in wheat and barley (Nakano et al., 1999; Greiner et al., 2000). This was further substantiated when it was shown that a transformation mediated duplication of the HvPAPhy_a gene in barley doubled the MGPA (Holme et al., 2012).

The present study addresses the characterization of genes encoding the Triticeae PAPhy and uncovers their evolutionary history towards a high MGPA. Fifteen Triticeae PAPhy genes from the five species wheat, barley, rye, einkorn, and Aegilops tauschii are cloned. It is demonstrated by chromosomal mapping in wheat and barley that the PAPhy genes can be assigned to two paralogous loci encoding the PAPhy_a and PAPhy_b isoforms. In rye, the PAPhy_a locus has undergone an additional duplication. The promoters responsible for the differentiated expression of the two loci were examined and the evolutionary origin of the paralogues is discussed in the context of the sequenced grasses rice, maize, sorghum, and Brachypodium (Brachypodium distachyon (L.) P. Beauv).

Materials and methods

Plant materials used for cloning

Cultivars/accessions for two rye, five Ae. tauschii, three wheat, barley, einkorn, and Brachypodium were used for PAPhy cloning and/or phytase assays (Supplementary Table S1, available at JXB online). Plants were grown in the greenhouse according to Brinch-Pedersen et al. (2000).

In silico sequence analysis

Sequence information was retrieved from the databases of Gramene (www.gramene.org/) and NCBI (www.ncbi.nlm.nih.gov/) as well as the raw 454 wheat sequence reads from CerealsDB (www.cerealsdb.uk.net/Index_Home.html). Alignments and trees were carried out using the CLC-workbench (CLC bio). Promoters were analysed for cis-acting regulatory elements with the PlantCARE webtool (http://bioinformatics. psb.ugent.be/webtools/plantcare/html/). Graphic representations of sequences were drawn with Vector NTI (Invitrogen). Synteny was analysed with the Narcisse webtool (www1.clermont.inra.fr/umr1095/narcisse_cereals/cgi-bin/narcisse15.cgi). Dotblots were drawn with the EMBOSS Polydot webtool (http://emboss.bioinformatics.nl/cgi-bin/emboss/polydot).

Genomic library constructions, screening, and gene cloning

Genomic libraries of wheat cultivar Skagen and rye cultivar Picasso were generated using the Lambda Fix II/XhoI Partial Fill-In Vector Kit (Agilent Technologies) according to the manufacturer’s instructions. The initial libraries were estimated to have 2.8–7.2 and 4.5–11.5-times genome coverage, respectively. The libraries were amplified once on NZY agar at a density of 300 pfu cm^–2.

Screening probes were labeled with 20 μCi ³²P using PCR. Probe one was amplified using the forward primer 5ʹ-CTTGAGCCTGG GACGAAGT-3ʹ and the reverse primer 5ʹ-GAGAAGGACCC GCTCTCC-3ʹ. The primers amplified a 479-bp fragment of the TaPAPhy_b2 cDNA. Probe two was amplified using the forward primer 5ʹ-TGGACCTGCCGGACAC-3ʹ-and the reverse primer 5ʹ-CGAAGGAGCAGGAGTAGCAG-3ʹ, amplifying a 770-bp fragment of the ScPAPhy_b1 gene.

For screening, the amplified libraries were plated out on 240×240mm NZY agar at a density of 600 pfu cm^–2. Plaque lifts were preformed with Hybond N+ membranes (GE Healthcare) followed by alkaline denaturation and UV crosslinking. Prehybridization (2 hours) and hybridization (overnight) were performed in Church’s buffer (0.25M sodium phosphate buffer, pH 7.2, with 7% SDS and 0.17mg ml^–1 salmon sperm DNA) at 65 °C. The membranes were washed twice at 65 °C and once at room temperature with SSPE buffer. X-ray films were exposed with the membranes at –80 °C for 2–5 days and the film was subsequently developed and analysed. λ preparations were performed with the Lambda midi kit (Qiagen), according to the manufacturer’s instructions.

Sequencing of the λ clones facilitated the design of primers in conserved regions 3–400bp upstream of the start codon and 50–250bp downstream of the stop codon of the PAPhy genes (Supplementary Table S2). PCR was performed using the Herculase II polymerase in a reaction mixture with 6% DMSO (Agilent Technologies) according to the manufacturer’s instructions. Amplicons were cloned in the pCR 4Blunt-TOPO vector (Invitrogen) and sequenced. At least two clones from independent PCR reactions were sequenced.

Inverted PCR was used to extend known sequences in the 5ʹ-direction. The templates were prepared by digestion of cereal genomic DNA with restriction endonucleases, BamHI, XhoI, or ScaI followed by inactivation of the enzyme and ligation with T4 DNA ligase in a final DNA concentration of 2ng μl^–1. Templates were generated from barley and einkorn genomic DNA. Inverted PCR consisted of a reaction with a primary primer set (5ʹ-GAATTCCAGATGGGCGGC-3ʹ and 5ʹ-CTGCTCGGGAGCCCAG-3ʹ) followed by a second nested PCR with a secondary primer set (5ʹ-TACAGCCAGCTCTACCCCT-3ʹ and 5ʹ-ACCGCGTGGCCCCTGT-3ʹ). PCR, cloning, and sequencing were preformed as described above.

Chromosomal mapping

Wheat chromosomal mapping was performed using the Chinese Spring nullisomic-tetrasomic lines described by Kimber and Sears (1979). Out of original 42 possible nullisomic-tetrasomic lines, two were missing in the present set of lines, the nullisomic (N) 2A tetrasomic (T) 2B and the N4BT4D lines. Their absence did not compromise the mapping because there are two nullisomic lines for each chromosome.

For chromosomal mapping in barley, the addition lines consisting of individual barley chromosomes (from cv. Betzes) translocated to wheat (cv. Chinese spring) was used (Islam et al., 1981; Islam and Shepherd, 2000). A part of each PAPhy isogene was PCR amplified using a specific set of primers (Supplementary Table S3) and Taq polymerase.

GUS constructs

The construct pCLEAN-G185-PAPhy_a (Holme et al., 2012) was modified to serve as a GUS construct. The PAPhy_a coding open reading frame and terminator was replaced by the UidA open reading frame followed by the NOS terminator. The cloning was performed with In-Fusion technology (Zhu et al., 2007), ensuring that the start codon context was preserved.

The vector backbone and promoter of the construct was amplified using the primers Cis to GUS Fw (5ʹ-TCGAGTCG ACGTTCTTGAC-3ʹ) and Cis to GUS Rv (5ʹ-GTTGATGTTGTT GCTTGGCATTG-3ʹ). The UidA and NOS terminator was amplified from a donor construct using the primers GUS Fw with overhang (5ʹ-AGCAACAACATCAACATGTTACGTCCTGTAGAAA CC-3ʹ) and GUS Rv with overhang (5ʹ-GGAACGTCGACTCGACT ATGACCATGATTACGAATTCC-3ʹ). The resulting construct was named pCLEAN-G185-wt-proGUS.

A potential regulatory motif was identified in silico (see Results). In order to evaluate the motif in vivo, a GUS construct where the motif had been obstructed was created. This was done by taking 20bp including the 14-bp motif from the pCLEAN-G185-PAPhy_a construct and process them trough a nucleotide randomizer (http://molbiol.ru/eng/scripts/01_16.html) with settings to preserve the nucleotide ratios of the original sequence. The resulting sequence, 5ʹ-gcatacgaagcatagtacga-3ʹ, was only identical to the original in three nucleotide positions and did not contain any regulatory elements known by PlantCARE. The original 20bp in pCLEAN-G185-wt-proGUS was replaced by the randomized sequence as described by Zhu and co workers (2007). The primers were Kill triad Fw (5ʹ-gcatacgaagcatagtacgaCGTAGGCGTCCAAACTTTG-3ʹ) and Kill triad Rv (5ʹ-tcgtactatgcttcgtatgcCTACAAAATGTAGCTTGA AATTAAAGAG-3ʹ). The resulting construct was pCLEAN-G185- KOtriad-proGUS.

Transient expression

Immmature barley (cv. Golden Promise) endosperms, 15–20 days after pollination, were isolated and placed on 0.25% phytagel pH 5.8 containing (l^–1) 4.4g MS medium (Duchefa) supplemented with 0.15g myo-inositol, 1.5g ammonium nitrate, 1g casein hydrolysate, 0.7g l-proline, 1.5mg 2,4-dichlorophenoxyacetic acid, 30g maltose, and 72.88g mannitol. The endosperms were bombarded in a DuPont PDS 1000 Helium Biolistic Delivery system using 1100-psi rupture discs and 1 μm gold particles coated with plasmid DNA as described by the manufacturer (BioRad). The endosperms were incubated in the dark at 23 °C overnight before histochemical staining as described by (Jefferson et al., 1987).

Phytase assay

Mature cereal grains were assayed for phytase activity essentially as Engelen et al. (1994). Grains were ground in a mortar. Protein was extracted from flour using 10ml of 25mM sodium acetate buffer (pH 5.5) with 0.1M calcium chloride per gram and shaking for 60 minutes. After centrifugation, 30 μl was diluted to 800 μl with extraction buffer containing sodium phytate to a final concentration of 5 μM. Phytate hydrolysis proceeded for 60 minutes at 37 °C. The reaction was stopped with 800 μl of freshly prepared 25g l^–1 ammonium heptamolybdate with 0.25% ammonia/6.9g l^–1 ammonium vanadate with 1% nitric acid/65% nitric acid (15:15:10). The mixture was centrifuged for 5 minutes at 3800 g before absorbance was measured at 415nm. Blanks were made by substituting the protein extract with extraction buffer. The concentration of phosphate was calculated according to a standard curve. All samples were measured in technical duplicates.

Results

Cloning of the PAPhy gene complement of selected Triticeae

Screening of two λ phage libraries lead to the identification of four wheat and two rye unique PAPhy clones. The clones were partially sequenced in order to localize the coding sequence as well as the promoter, introns, and terminator sequences. Wheat TaPAPhy_a2 and rye ScPAPhy_a1 λ clones contained only part of the coding sequences. In addition, a barley PAPhy λ clone isolated previously was analysed in more detail in the present study (Holme et al., 2012).

Introns and exons were annotated after alignment with the known cDNA clones and identification of the in-frame start and stop codons (Fig. 1). One interruption of the coding sequence on the TaPAPhy_a2 λ clone was only 8bp and did not follow the GT–AG rule for introns. PCR screening of nine wheat cultivars (including Chinese spring) showed that the 8-bp interruption was only present in the cultivars Skagen and Bob White (results not shown). Hence, this interruption was annotated as an insertion rather than an intron. All other interruptions of the coding sequences followed the GT–AG rule as expected for typical introns.

Fig. 1.

Schematic representation of the sequenced part of the λ clones and the PCR clones. Green, exons; red, introns; light blue, retrotransposons and other recurring DNA segments; dark blue, TATA boxes, near upstream elements, and insertions (not to scale).

A PCR-based cloning strategy was designed as supplement to the library screenings. Sequence information from the λ clones and known cDNA sequences was used to design primers in conserved regions 3–400bp upstream of the start codon and 50–250bp downstream of the stop codon. In wheat, all genes already known from the library screenings were isolated again (Fig. 1). However, two additional genes, termed TaPAPhy_b1 and TaPAPhy_b2, were identified. In rye, the full coding sequence of ScPAPhy_a1 was obtained and an additional gene, ScPAPhy_a2, was discovered. From barley, HvPAPhy_b1-2 was identified. One PAPhy_a gene and one PAPhy_b gene were amplified from both einkorn and Ae. tauschii. The PAPhy 5ʹ-flanking regions of the PCR clones were extended by inverted PCR and by consulting raw 454 reads of wheat available at CerealsDB. Using these techniques, the 5ʹ-flanking regions were extended from 277bp in TaPAPhy_a2 to a maximum of 713bp in HvPAPhy_b1-2. In summary, three PAPhy_a and three PAPhy_b genes could be isolated from wheat. One PAPhy_a and one PAPhy_b gene were cloned from each of barley, einkorn, and Ae. tauschii. In rye, two PAPhy_a genes and one PAPhy_b genes were isolated. The clones are described in Table 1 with the details of their accession numbers, size, and gene coverage.

Table 1.

Summary of isolated clones, the chromosomal location of their corresponding genes and the extent to which the coding sequences and flanking regions have been sequenced

Species	Clone type	Accession	Gene	Chromosome	5ʹ- end of start codon (bp)	Exons and introns (bp)	3ʹ-end of stop codon (bp)	Extension
Wheat	λ	JF838306	TaPAPhy_a1	5D	2090	2003	721
	PCR	JF838305			309	2003	76
	λ	JF838308	TaPAPhy_a2	5A	2373	1119	–
	PCR	JF838307			324	1975	76
	λ	JF838310	TaPAPhy_a3	5B	3837	1982	703
	PCR	JF838309			261	1982	94
	PCR	JF838311	TaPAPhy_b1	3D	391 (697)	2081	192	306bp by 454 read lcl|GKTGI2C01AVAXX
	PCR	JF838312	TaPAPhy_b2	3B	378 (655)	2072	203	277bp by 454 read lcl|F3ODV4H01C3ZB5
	λ	JF838314	TaPAPhy_b3	3A	1883	2088	209
	PCR	JF838313			361	2087	196
Einkorn	PCR	JF838315	TmPAPhy_a1		316 (627)	1972	76	311bp by inverted PCR
	PCR	JF838316	TmPAPhy_b1		364	2058	197
Aegilops tauschii	PCR	JF838317	AePAPhy_a1		307	2008	76
	PCR	JF838318	AePAPhy_b1		391	2085	193
Barley	λ	FR851293	HvPAPhy_a1	5H	2728	2246	878
	PCR	JF838323	HvPAPhy_b1/2	3H	319 (1032)	2066	202	713bp by inverted PCR
Rye	λ	JF838320	ScPAPhy_a1		1493	149	–
	PCR	JF838319			277	1982	102
	PCR	JF838321	ScPAPhy_a2		504	1968	102
	λ	JF838322	ScPAPhy_b1		2723	2016	698

Gene structure, copy number, and chromosomal mapping

The cloned PAPhy genes contain either four or five introns (Fig. 1). Pairwise alignments of the genomic wheat and barley clones and the corresponding cDNA clones previously cloned and classified as PAPhy_a and PAPhy_b by Dionisio et al. (2011) revealed that the PAPhy_a genes possessed four introns and the PAPhy_b genes possess five introns. The four introns of PAPhy_a were positioned as introns 1, 2, 4, and 5 of PAPhy_b. Barley is known to express one PAPhy_a transcript and two PAPhy_b transcripts (Dionisio et al., 2011). The known PAPhy_a transcript corresponds to the λ clone of HvPAPhy_a1, but only one PAPhy_b gene was successfully cloned by PCR. However, this gene could account for both transcripts by differential splicing at the 3ʹ-border of the third intron, where two alternative AG dinucleotides facilitate the inclusion or omission of a GAG codon. Because of the differential splicing of the barley PAPhy_b transcripts, this gene is referred to as HvPAPhy_b1-2 and the two transcripts as HvPAPhy_b1 and HvPAPhy_b2. The three TaPAPhy_a and TaPAPhy_b genes isolated from Chinese Spring suggest that the two groups are a pair of paralogues with homoeologous loci on each of the three subgenomes A, B, and D. Moreover, only one PAPhy_a gene and one PAPhy_b gene could be isolated from the diploid carriers of the A and D genomes (einkorn and Ae. tauschii, respectively) and from barley (diploid).

The wheat and barley PAPhy_a and PAPhy_b genes were chromosomally mapped to confirm or disprove the assumption of a single pair of paralogues (Supplementary Figs. S1 and S2 and Supplementary Table S2). The three TaPAPhy_a genes were mapped with one gene on each of the three subgenomes A, B, and D. The PAPhy_a genes were located on chromosome 5 in all wheat subgenomes and barley. The wheat TaPAPhy_b genes were distributed in a similar manner, with one locus on each of the three homoeologous subgenomes. In wheat and barley, the genes mapped to chromosome 3 (Table 1).

The PAPhy genes of sequenced grasses

The TaPAPhy_a1 coding sequence was blasted against the sequenced genomes of rice, maize, sorghum, and Brachypodium in the LATESTGP database (Gramene.org). The search sensitivity was set at ‘near exact matches’. Rice, maize, and sorghum were found to posses one PAPhy locus, whereas Brachypodium posses two loci. Contig views showing the TaPAPhy_a1 coding sequence aligned to the chromosomes and predicted genes are shown in figure 2. According to the Narcisse webtool, conserved synteny exists between Brachypodium chromosome 1, barley chromosome 5, maize chromosome 1, rice chromosome 3, and sorghum chromosome 2. This suggests that the PAPhy genes on these chromosomes are orthologues. There was no conserved synteny between Brachypodium chromosomes 1 and 2. Likewise, there was no conserved synteny between barley chromosomes 3 and 5 (results not shown).

Fig. 2.

Contiq views of the PAPhy loci in sequenced grasses. Boxes, exons; lines, introns; bright red, significant alignments of the TaPAPhy_a1 coding sequence to the genomes; dark red, predicted genes.

All-against-all dotblots were generated using 20 and 200 Kb chromosome sequences centred around the start codon of the PAPhy genes in the sequenced grasses (Fig. 3). The best results were obtained with word size 11 for the 20-kb blots and word size 14 for the 200-kb blots. The sequences were orientated according to the PAPhy genes. The genes were clearly recognizable as a diagonal line from the centre and upwards to the right in all 20-kb blots. The synteny extended beyond the PAPhy gene in 20-kb blots of Brachypodium chromosome 1 vs. rice and in rice vs. sorghum. In 200-kb blots, the PAPhy genes were only just visible as an enlarged dot in the centre. Nearby regions with more extensive identity were visible as diagonals or tracks of enlarged dots in the blots of Brachypodium chromosome 1 vs. rice and sorghum, sorghum vs. rice as well as sorghum vs. maize.

Fig. 3.

All-against-all dotblots of 20 and 200kb sequences centred around the PAPhy start codon.

Taken together, this confirms the hypothesis that the Brachypodium chromosome 1 gene and the single-copy PAPhy genes of rice, maize, and sorghum are orthologues. Dotblot confirmation regarding inclusion of the gene on Triticeae chromosome 5 in this group is pending because current sequence data is insufficient.

The Brachypodium chromosome 2 segments did not form clear diagonals or tracks beyond the PAPhy gene with any other sequence, although enlarged dots were present. A multiple alignment of the two Brachypodium 20-kb sequences and the predicted coding sequences of the two PAPhy genes was generated. The genomic sequences aligned from the start codon until 43/52bp before the stop codons and they were 53.5% identical in this region. Three smaller regions of identity (72bp/97.22%, 51bp/86.27%, and 30bp/83.33%) were identified downstream of the PAPhy genes. The first region was found only 100bp downstream of the PAPhy gene on chromosome 1 but 5051bp downstream of the gene on chromosome 2. All introns of the Brachypodium chromosome 1 gene were also present in the gene on chromosome 2. This gene on the other hand had two additional introns, numbers 2 and 6.

The TaPAPhy_a1 and TaPAPhy_b1 genomic and coding sequences were added to the alignment. The two introns specific to the gene on Brachypodium chromosome 2 were found in both wheat genes and all wheat introns were also found on the Brachypodium chromosome 2 gene. The Brachypodium coding sequences were aligned to each other and were found to be 78.1% identical.

The Triticeae PAPhy core promoter and untranslated regions

The longest available sequences representing the individual Triticeae PAPhy genes were aligned and adjusted to include the beginning of the coding sequence and the 5ʹ-flanking region (Supplementary Fig. S3). One in-frame ATG codon was conserved and aligned in the beginning of the predicted coding sequences of all 15 PAPhy genes. The four bases preceding the ATG agrees with the consensus start motif found for monocots G(A/C)(G/A)GC(A/C/G)(G/A)(C/A)(G/C)AUGGCG (Nakagawa et al., 2008). This codon was therefore assigned as the start codon.

All 15 PAPhy genes were represented in the alignment up to position –601bp (relative to the start codon). CAP-trapper-generated cDNA clones of TaPAPhy_a2 and TaPAPhy_b2 are available (Genbank AK331144.1 and AK331704.1) (Kawaura et al., 2009). The CAP trapper method ensures full transcript length clones (with an additional 5ʹ-guanidine) so the clones facilitated the identification of conserved motives with positions and sequences corresponding to the TATA box and the near upstream element. Most genomic clones contained two adjacent TATA boxes and one contained two near upstream elements (Fig. 1 and Supplementary Fig. S3).

Transcription is expected to begin 25–39bp downstream of the TATA box (Joshi, 1987). Genomic vs. cDNA pairwise alignment demonstrated that the full transcript length TaPAPhy_a2 cDNA clone starts 35bp downstream of the TATA box closest to the start codon (TATA box 1). The context of this site is conserved and it fits well with the ‘cap signal’ (T/G)C(A/T)(G/T/C)(T/C/A)(C/T)(T/C/G)(T/C) reported by (Bucher, 1990). The adjacent alignment position is a better match to the cap signal for the PAPhy_b genes. The two positions were assigned as the transcriptional start sites when TATA box 1 is used (Supplementary Fig. S3).

The TaPAPhy_b2 clone on the other hand have been transcribed via the second TATA box (TATA box 2) since it begins 38bp downstream at alignment position –206. This position is only covered by the Triticum and Aegilops PAPhy_b genes, which appear to have an insertion here when they are compared to the PAPhy_a genes and to the barley and rye PAPhy_b sequences. The transcriptional start site of genes without the insertion is expected to be closer to the start codon because of the strict positional constraints of the TATA box and the transcriptional start site. A good candidate is alignment position –186. This residue is conserved in all PAPhy genes and its consensus context CCAGACGA (position –186 underlined) fits well with the cap signal.

TATA box 1 aligns across all PAPhy genes at positions –115 to –121bp and it is surrounded by a 121-bp block of high conservation, even between PAPhy_a and PAPhy_b (Supplementary Fig. S3). The position of TATA box 2 was also conserved, although it is missing from TaPAPhy_b1 and AePAPhy_b1. Moreover, as seen for TATA box 1, box 2 is surrounded by a segment of significant conservation across all PAPhy genes. However, for TATA box 2, the conserved region is much shorter (58bp) and less conserved (Supplementary Fig. S3). The distance between the TATA boxes is 90 +/– 28bp in all promoters except ScPAPhy_a2.

In conclusion, the Triticeae PAPhy genes have a conserved core promoter usually consisting of two TATA boxes approximately 90bp apart and adjacent transcriptional start sites. The PAPhy_b 5ʹ-UTR is 60–70bp longer than the PAPhy_a 5ʹ-UTR when transcripts are derived from the same TATA box. Pairwise alignment of the Brachypodium promoters (1000bp 5ʹ-flanking regions) showed that they do not possess a conserved core promoter (results not shown).

The differentiated Triticeae PAPhy_a and PAPhy_b promoters

There was no significant conservation between paralogous promoters upstream of the core promoter. However, orthologous promoters were still highly conserved until alignment position –650 and –910 for PAPhy_a and PAPhy_b, respectively (Supplementary Fig. S3). Blasting of the 5ʹ- and 3ʹ-flanking regions against the NCBI nucleotide collection revealed no known genes. However, λ clones of TaPAPhy_a2, TaPAPhy_a3, TaPAPhy_b3 and ScPAPhy_a1 contained regions which returned several matches typically annotated as transposons (Table 2). Moreover, comparing the ScPAPhy_a2 to other PAPhy_a genes reveals a 219-bp insertion between the two TATA boxes (Fig. 1 and Supplementary Fig. S3). The size and presence of flanking short direct repeats suggest the insertion is a short interspersed element.

Table 2.

Matches to the 5’-flanking regions of the λ clones

λ clone	Match position (relative to the start codon)	Maximum identity (%)	Annotation at match	Match accession	Origin of match
TaPAPhy_a2	–507 to –2360	86	None, long terminal repeat retrotransposons at similar matches	AY494981.1	T. turgidum
TaPAPhy_a3	–800 to –3837	95	Long terminal repeat	FN564428.1	Wheat (3B)
TaPAPhy_b3	–714 to –1883	96	None, flanked by repeat regions	FN564434	Wheat (3B)
ScPAPhy_a1	–698 to –1447	82	Retrotransposon, gypsy	FN564428	Wheat (3B)

The breakdown of intra-orthologue conservation coincides with the occurrence of transposon-like sequences (Table 2 and Supplementary Fig. S3). In the further investigation of the PAPhy promoters, it was assumed that cis-acting regulatory elements responsible for the differentiated expression pattern of PAPhy_a and PAPhy_b are found upstream of the core promoter and downstream of the intra-orthologue unconserved region. These assumptions defined othologue-specific promoter sequences of 103–157bp and 121–259bp for PAPhy_a and PAPhy_b, respectively (Supplementary Fig. S3, red and blue boxes). These sequences were screened for cis-acting regulatory elements.

Five motifs were common in the PAPhy_a promoters (Table 3). The RY element was the only omnipresent element, but all the elements were present in at least half of the promoters and many cases of absence were only single-nucleotide sequences that could reflect variations of the element. The observations were drawn to scale using the TaPAPhy_a2 5ʹ-flanking region as template (Fig. 4A).

Table 3.

cis-Acting regulatory elements in PAPhy promoters

The occurrence is given as the number of orthologous promoters with the element compared with the number of sequences represented at the particular position in the alignment.

	Motif	Sequence	Description in PlantCARE	Origin according to PlantCARE	Occurrence
PAPhy_a	TGATC	TGACT	cis-Acting regulatory element involved in the MeJA-responsiveness	Barley	4 of 6
	GCN4	TGAGTCA	cis-Regulatory element involved in endosperm expression	Rice	5 of 8
	Skn1	GTCAT	cis-Acting regulatory element required for endosperm expression	Rice	7 of 8
	RY-element	CATGCATG	cis-Acting regulatory element involved in seed-specific regulation	Sunflower	8 of 8
	G-box	CACGTT	cis-Acting regulatory element involved in light responsiveness	Maize	4 of 8
PAPhy_b	ABRE	CACGTG	cis-Acting element involved in the abscisic acid responsiveness	Arabidopsis thaliana	5 of 7
	GARE	AAACAGA	Gibberellin-responsive element	Brassica oleracea	5 of 5
	TGACG	TGACG	cis-Acting regulatory element involved in the MeJA-responsiveness	Barley	6 of 7

Fig. 4.

Schematic representation of the regulatory elements in the (A) PAPhy_a and (B) PAPhy_b –300 and –376bp 5ʹ-flanking regions.

The overlapping motifs of GCN4 and Skn-1 and the preceding base are identical to the odd base C-box ATGAGTCAT reported by De Pater et al. (1994). Hence, the two elements and the RY element can be interpreted as two elements overlapping in three nucleotides and spanning a total of 14bp. The significance of the combined odd base C-box and RY element was examined by transient expression of GUS driven by the barley PAPhy_a promoter in its native form (construct pCLEAN-G185-wt-proGUS) and with the two elements randomized (construct pCLEAN-G185-KO-proGUS). The pCLEAN-G185-wt-proGUS construct yielded several blue dots (Fig. 5) whereas the pCLEAN-G185-KO-proGUS did not yield any blue dots (results not shown). The two Brachypodium promoters (1000bp 5ʹ-flanking regions) were also screened for the double element but it was not present.

Fig. 5.

Barley endosperms bombarded with pCLEAN-G185-wt-proGUS and stained in GUS buffer.

Three cis-acting regulatory elements were common in the PAPhy_b promoters (Table 3). The observations were drawn to scale using the TaPAPhy_b3 5ʹ-flanking region as template (Fig. 4B).

Phylogeny

A multiple alignment of the genomic sequences of all 15 Triticeae PAPhy genes was created in CLC workbench using the settings gap open cost 10, gap extension cost 1 and end gap cost ‘as any other’. Sequences from the PCR clones were used when available. Alternatively, sequences from λ clones were shortened to lengths similar to the PCR clones and used for the alignment. The initial alignment was adjusted to include only the section between the first T of TATA box 1 and the T in the near upstream element.

A UGPMA tree was constructed from this alignment (Fig. 6A). The PAPhy_a and PAPhy_b genes form two distinct groups across the species. The PAPhy genes from einkorn (A genome) and Ae. tauschii (D genome) group most closely with the wheat genes located on the A and D subgenomes, respectively. In both the PAPhy_a and PAPhy_b groups, the einkorn and wheat A subgenome PAPhy genes are more diverged than Ae. tauschii and wheat D subgenome PAPhy genes. The divergence between the two PAPhy_a genes in rye is comparable to the divergence between einkorn and the wheat PAPhy genes of the A subgenome. Rye and barley genes fall outside a close Triticum/Aegilops grouping for both the PAPhy_a and PAPhy_b groups.

Fig. 6.

UGPMA trees of: (A) the genomic Triticeae PAPhy sequences from TATA box 1 to the first near upstream elements; (B) full-length coding sequences from PAPhy genes isolated in this study and sequenced grasses; (C) coding sequences from PAPhy genes isolated in this study and sequenced grasses trimmed to equal lengths. Bootstrap values are shown at branch nodes.

Arguably, the tree is heavily affected by the additional intron present in the PAPhy_b genes. In order to investigate whether the tree would be different without the introns, a multiple alignment of the coding sequences was generated (Fig. 6B). The same settings were used but additional coding sequences of the PAPhy genes from sequenced grasses were included in order to put the PAPhy_a and PAPhy_b groups into a broader context. When using this approach, the rye PAPhy_a genes group with their Triticum counterparts. However, apart from this, the Triticeae PAPhy genes group as before. ScPAPhy_a1, ScPAPhy_a2, HvPAPhy_a1, and TaPAPhy_a3 coding sequences are 27bp shorter than other PAPhy_a genes due to a polymorphism introducing a stop codon. When the coding sequences are shortened to equal length, they group almost as seen for the genomic clones (Fig. 6C). Moreover, the Brachypodium PAPhy located on chromosome 2 groups with the PAPhy_b genes. The Brachypodium PAPhy located on chromosome 1 is the most diverged of all genes included. The rice, sorghum, and maize PAPhy sequences fall outside the two groups formed by the Triticeae PAPhy genes and the maize and sorghum PAPhy genes group closely together.

Phytase activity in Triticeae and Brachypodium

The MGPA of the plant material used for cloning and Brachypodium accessions was measured. Additional cultivars/accessions of each species were measured to provide biological replication. The samples were greenhouse grown as described or primary seedbank material in the case of AC 1548, AE 1068, AE 938, GRA 787 and GRA 788. The values differed significantly between the materials. The highest MGPA (6019 FTU/kg) was measured in Ae. tauschii accession NGB90403 and the lowest (736 FTU/kg) was measured in wheat cultivar Skagen. The MGPA of Brachypodium was in the same range as the Triticeae.

Discussion

The PAPhy_ a and PAPhy_b paralogues

Chromosome mapping demonstrated that the PAPhy genes of wheat and barley consists of a set of paralogous genes located on chromosomes 3 and 5 in each of the homoeologous genomes. This is in good agreement with a previous division of PAPhy proteins and transcripts into two isoforms, PAPhy_a and PAPhy_b (Dionisio et al., 2011). The gene on chromosome 5 generates the transcripts termed PAPhy_a and the gene on chromosome 3 generates the PAPhy_b transcripts. It is reasonable to assume a similar chromosomal localization in the diploid carriers of the A and D genomes, i.e. einkorn and Ae. tauschii, respectively.

Rye possesses as the only investigated member of the Triticeae a duplicated PAPhy_a gene. The exact nature of this duplication remains to be investigated but the phylogenetic trees indicate that the two ScPAPhy_a genes are as divergent as the PAPhy genes of wheat subgenome A and those of einkorn. This is notably more divergent than the PAPhy genes of wheat subgenome D and Ae. tauschii. The wheat A subgenome and einkorn diverged between 0.5 and 1 million years ago whereas the hexaploidization which separated the wheat D subgenome from Ae. tauschii occurred only 8000 years ago (Huang et al., 2002). Divergence of the Secale and Triticum genera occurred around 7 million years ago and rye was domesticated 10,000–12,000 years ago (Salamini et al., 2002; Huang et al., 2009). This suggests that the rye ScPAPhy_a duplication occurred long after speciation but prior to domestication. It is also possible that two different wild Secale species each have contributed a locus to domesticated rye.

Brachypodium is the closest relative of the Triticeae which has been fully sequenced and assembled. The Brachypodium set of PAPhy paralogues is located on chromosomes 1 and 2. These chromosomes share orthologous linkage blocks with the Triticeae chromosomes 5 and 3, respectively (International Brachypodium Initiative, 2010). The presence of orthologous linkage blocks on chromosomes carrying the paralogous PAPhy genes in Brachypodium and in Triticeae suggests that they originate from the same gene duplication event. This is supported by the high degree of shared intron/exon structure and the close phylogenetic relationship between the PAPhy_b genes and the Brachypodium chromosome 2 gene. On the other hand, the gene on Brachypodium chromosome 1 is the most distant member of the clade. However, accelerated divergence can be caused by the functional redundancy introduced by the gene duplication itself (Zhang, 2003). This is supported by the dotblot analysis which show that the Brachypodium chromosome 1 PAPhy is the orthologue of the PAPhy genes in rice, maize, and sorghum. The dotblot and alignment generated from the two Brachypodium sequences demonstrated that the duplication is of the same approximate size as the gene itself. Small duplications may result from retroposition but this would usually result in the loss of introns and certainly the loss of the core promoter (Zhang, 2003). The core promoter is preserved in the Triticeae and introns are preserved in both Brachypodium and Triticeae. The duplication must therefore be of the segmental type.

Other sequenced grasses (rice, maize, and sorghum) contain only one PAPhy gene. This raises the question of gene duplication vs. gene loss. All the Triticeae PAPhy genes and the Brachypodium chromosome 2 gene were phylogenetically closer to each other than to the PAPhy genes of maize, rice, and sorghum. The duplication is therefore younger that the divergence of the species. Further, it is older than the divergence of Brachypodium and the Triticeae since they share it. This indicates duplication between 54 and 32 million years ago (International Brachypodium Initiative, 2010). The most ancient gene must be that, which is shared with all the examined grasses, i.e. the PAPhy_a/Brachypodium chromosome 1 gene.

The divergent promoters of PAPhy_a and PAPhy_b

Analysis of the 5ʹ-flanking region of eight Triticeae PAPhy_a and seven Triticeae PAPhy_b genes revealed a conserved core promoter with two TATA boxes. The two full-length transcripts demonstrate that both TATA boxes are active albeit in different genes. The high conservation of both TATA boxes and their associated transcriptional start sites suggest that both TATA boxes are generally active.

There was no significant conservation between PAPhy_a and PAPhy_b upstream of the core promoter, but orthologues showed conservation for an additional 150–200bp. The upstream sequences of the paralogous promoters of PAPhy_a and PAPhy_b contained cis-acting regulatory elements consistent with the differentiated expression reported previously (Dionisio et al., 2011). Thus, PAPhy_b contained the giberellic acid response element typical for hydrolases expressed during germination in response to gibberellic acid (Ritchie and Gilroy, 1998). For PAPhy_a, the promoters contain the overlapping motifs of the odd base C box and the RY-element. The same motif can also be described as a triad consisting of the GCN4 motif, the skn1 motif, and the RY-element. These motives have been described in promoters driving expression of storage protein genes during seed formation and grain filling. For instance, the RY element has been found and proven functional in promoters of dicot storage proteins such as legumins or conglycinin (Bäumlein et al., 1992; Fujiwara and Beachy, 1994). Similarly the odd base C box has been found in lectin promoters, hordeins and rice glutelins (Müller and Knudsen, 1993; De Pater et al., 1994; Takaiwa et al., 1996). This study’s reporter gene experiment shows that this double motif or triad is critical for the expression of PAPhy_a during grain filling in barley.

Phytase activity in the wider Triticeae tribe and Brachypodium

The present data agrees with previously reported values for wheat, barley, and rye (Eeckhout and De Paepe, 1994; Viveros et al., 2000; Steiner et al., 2007). A collection encompassing einkorn, Ae. tauschii and Brachypodium accessions was also assayed (Fig. 7). As far as is known, these are the first published MGPA values for the three species. The MGPA measurements of einkorn and Ae. tauschii provide further evidence of the association of high MGPA with the Triticeae tribe but the Brachypodium data demonstrate that the trait is not restricted to the Triticeae. Rather, it is associated with the duplication of the PAPhy gene which is shared by Brachypodium and the Triticeae. However, it is not a simple matter of gene dosage, since the trait is conferred by a specialized locus in the Triticeae (PAPhy_a). The PAPhy_a locus is the parent locus so the changed expression pattern cannot result from the duplication as such, for instance because of a new chromosomal context of the gene. The expression pattern of the Brachypodium PAPhy genes remain to be investigated but it is clear that phytase is expressed during grain filling even though neither gene share the features of the Triticeae PAPhy_a promoter. This is not unexpected, since numerous promoters can provide expression during grain filling. The observation suggests that evolution of high MGPA has happened independently in Brachypodium and the Triticeae after the gene duplication. The neofunctionalization of the gene is best explained by positive selection in light of the observations above. Neofunctionalization appears to have been countered by stronger purifying selection in grasses with a single PAPhy gene, presumably because phytase activity during germination is important for the successful lifecycle of the plant. Relaxation of the purifying selection by the gene duplication was therefore critical for the evolution of high MGPA in the Triticeae and Brachypodium. This indicates that the room for improvement of the MGPA by breeding is much more restricted in cereals with a single PAPhy than in cereals with two or more PAPhy genes.

Fig. 7.

Phytase activities of mature seeds.

Supplementary material

Supplementary data are available at JXB online.

Supplementary Fig. S1. Chromosomal mapping of the TaPAPhy genes in Chinese Spring nullisomic-tetrasomic wheat lines

Supplementary Fig. S2. Chromosomal mapping of the TaPAPhy genes in wheat barley translocation lines

Supplementary Fig. S3. Alignment of the Triticeae PAPhy 5ʹ-flanking region and first 20bp of the coding sequence

Supplementary Table S1. Summary of plant material used for cloning purposes and MGPA assays

Supplementary Table S2. Primers for amplifying PAPhy genes and flanking regions

Supplementary Table S3. Primers used for PCR-based chromosomal mapping