Background

Phosphomannomutase (PMM, EC 5.4.2.8), catalyzing the interconversion between mannose-6-phosphate and mannose-1-phosphate, is an essential and conserved enzyme in eukaryotic organisms [1, 2]. Mannose-1-phosphate is necessary for synthesizing the vital cellular metabolite GDP-mannose, which plays a crucial role in the formation of polysaccharide chains required for the glycosylation of protein and lipid molecules [3, 4]. In higher plants, GDP-mannose also acts as an important precursor for the biosynthesis of the key antioxidant ascorbic acid (AsA) [5, 6], and the mannose containing polysaccharides are essential for the development of functional cell walls [2, 4]. From available information, the copy number of functional PMM in diploid eukaryotic species varies from one (i.e., in Saccharomyces cerevisiae and Arabidopsis thaliana) to two (e.g., in human). PMM proteins from different eukaryotic species are highly similar in primary structure, with more than 50% amino acid sequence identities found among the PMMs from S. cerevisiae, Arabidopsis and human [6]. The crystal structure and catalytic mechanism of mammalian PMMs are well understood [7], and there is also a wealth of information on human PMM2, whose point mutations lead to a genetically inherited disease CDG-Ia [8–11].

Two recent molecular genetic studies in Arabidopsis show that defect in PMM function reduces AsA biosynthesis and protein glycosylation, causing enhanced susceptibility to oxidation stress and plant death at high growth temperature (28°C) [6, 12]. The requirement of a functional PMM protein for Arabidopsis to grow at 28°C is consistent with the pioneering finding of temperature sensitive PMM mutations that arrest S. cerevisiae growth at 37°C [1], suggesting that PMM may be generally important for the temperature adaptability of eukaryotes.

Compared to the above progress, little is known about PMM gene and function in crop plants, although broad temperature adaptability is vitally important for crops to achieve high yield potential under diverse environmental conditions. To improve our understanding on PMM in crop plants, we have embarked on a systematic molecular and biochemical analysis of PMM genes in bread wheat (T. aestivum, 2n = 6x = 42, AABBDD), a polyploid crop species cultivated in both temperate and tropical regions and with harvest possible under a wide range of temperature conditions [13, 14]. Bread wheat belongs to the tribe Triticeae, which is composed of approximately 350 species, and contains both diploid (such as barley and rye) and polyploid (e.g., bread wheat) crop plants [15]. The diploid genomes in Triticeae species are homoeologous, and may be combined to form polyploids through natural and artificial hybridizations [16]. Bread wheat was evolved through natural hybridization between tetraploid wheat (T. turgidum, 2n = 4x = 28, AABB) and Aegilops tauschii (2n = 2x = 14, DD) about 8000 years ago [16–18]. T. turgidum was formed about 0.5 million years ago (MYA) through spontaneous hybridization between an unidentified Triticeae species containing the B genome and the wild einkorn wheat T. urartu (2n = 2x = 14, AA) [19–22]. The A, B and D genomes diverged from each other around 2.5 to 4.5 MYA, whereas the divergence of wheat and barley occurred around 11 MYA [19–22]. Although bread wheat and related Triticeae species generally contain large and complex genomes, the advent of structural and functional genomics research in model cereals (rice, Brachypodium distachyon) has made it possible to conduct relatively detailed molecular genetic and biochemical studies of important genes both within Triticeae and among rice, Brachypodium and Triticeae species. The genomic resources of rice and B. distachyon (e.g., molecular markers, annotated genomic information) have both been proved to be very useful in the investigations of important Triticeae genes, such as Ph1 and Lr34/Yr18 in common wheat and Ppd-H1 and Cly1 in barley [23–26].

From the information presented above, the main objectives of this work were to 1) determine PMM genes and loci in bread wheat and related Triticeae species, and 2) investigate the evolution of Triticeae PMM genes and the main active PMM isozymes in bread wheat tissues, with the aid of the genomic knowledge of rice and B. distachyon.

Results

Isolation and characterization of PMMgenes

Two approaches (bioinformatic analysis, molecular cloning) were used to identify and isolate the PMM genes in B. distachyon, bread wheat and related Triticeae species (barley, T. urartu, T. turgidum, Ae. tauschii). In total, 17 distinct PMM coding sequences were isolated (Table 1). A single PMM gene was found for both rice (OsPMM) and B. distachyon (BdPMM). The cDNA of OsPMM was reported previously [6], whereas that of BdPMM was isolated in this work using RT-PCR (Table 1).

Table 1 PMM genes from B. distachyon, bread wheat and related Triticeae species, and rice a
Full size table

Six different PMM genes were isolated from each of the two bread wheat varieties, Xiaoyan 54 and Chinese Spring (CS). T. urartu, Ae. tauschii and barley were each found to contain two distinct PMM genes, while four different PMM genes were cloned from T. turgidum. Based on the data on exon and intron pattern, nucleotide and deduced protein identities, chromosomal location, and phylogenetic clustering analysis of genomic coding sequences (see below), the A, B and D genomes were each found to contribute two members to the six PMM genes in bread wheat, which were thus named as TaPMM-A1, B1, D1, A2, B2 and D2, respectively (Table 1). The PMM-1 genes in bread wheat and its diploid and tetraploid progenitor species and barley were homoeologous. Likewisely, homoeologous relationship existed among the PMM-2 genes from these species.

The intron and exon patterns of the PMM genes from rice, B. distachyon, and wheat and relatives were highly similar. Except for OsPMM and BdPMM, both of which had 11 exons owing to the split of the first exon into two parts, the remaining genes all had 10 exons and 9 introns (Figure 1). Apart from PMM-A2 homoeologs and HvPMM-2, the remaining Triticeae PMM genes cloned in this work all contained intact ORF. The coding region of HvPMM-2 was disrupted by premature stop codon based on analyzing the PCR products amplified from two barley varieties (Betzes, Zaoshou 3) (Table 1, Additional file 1).

Figure 1
figure 1

Comparisons of the exon (box) and intron (line) patterns of the PMM genes from bread wheat ( TaPMM-A1 , B1 , D1 , A2 , B2 and D2 ), barley ( HvPMM-1 and 2 ), B. distachyon ( BdPMM ), rice ( OsPMM ), and Arabidopsis ( AtPMM ). The exon/intron structures of the compared PMM genes are highly similar except that the first exon in BdPMM and OsPMM (underlined with dashed line) is split into two parts. The positions to which the intron flanking primers (IFf, IFr) bind are indicated by arrows. The chromosomal loci containing the genomic coding sequences of AtPMM and OsPMM are At2g45790 and Os04g0682300, respectively.

Full size image

Consistent with our earlier study [6], the proteins deduced from the PMM genes of common wheat and progenitor species, barley and B. distachyon were more than 50% identical to the PMMs of human and S. cerevisiae (Additional file 2). The identities among the grass PMMs were generally above 90%, with BdPMM showing higher identity to its homologs from bread wheat and related Triticeae species than to OsPMM. The identities among the deduced PMM-1 proteins of bread wheat and its diploid and tetraploid ancestors were generally 98% or higher, with complete identity found among the deduced TaPMM-A1, TuPMM-A1, TtPMM-A1 and TaPMM-B1 proteins. The identities among the deduced PMM-B2 and D2 proteins of bread wheat and its diploid and tetraploid ancestors were generally 97%. By contrast, the identities between the deduced PMM-1 and 2 proteins were comparatively lower, being 95% between B1 and B2 proteins, and 96% between D1 and D2 proteins. The deduced PMM-B1 (or B2) proteins of tetraploid and bread wheats were 99% identical. Interestingly, TaPMM-B2 and TtPMM-B2 displayed unusual amino acid substitutions. The first one, involving the substitution of a broadly conserved glycine residue by alanine, was shared between TaPMM-B2 and TtPMM-B2 (Additional file 3). The second one was unique to TaPMM-B2, involving the replacement of a highly conserved arginine residue by cysteine. The four structural and active site motifs (motifs I to IV), identified by analyzing the crystal structure of HsPMM-1 [7], were generally conserved among the compared plant PMMs (Additional file 3).

Determination of PMMcopy number and chromosomal locations

To investigate the copy number of PMM genes, we made use of intron length polymorphisms and designed a pair of intron flanking (IF) primers (Additional file 4), which enabled the amplification of PCR fragments containing introns III to V (Figure 1). By carefully analyzing the products amplified with fluorescently labeled IF primers, six distinct fragments were found for bread wheat (varieties Xiaoyan 54 and CS), and their size equaled to that calculated from the six cloned TaPMM genomic sequences in all cases (Additional file 5). Using the same strategy, two PMM fragments were found for T. urartu (accessions DV877 and IE29-1), Ae. tauschii (accessions AS67 and AS91) and barley (Betzes), respectively, and for all three species, the size of the amplified fragments corresponded exactly to the one deduced from the cloned PMM members (Additional file 5). Four fragments were amplified from T. turgidum (Langdon, LDN), whose size again showed specific correspondence to that derived from each of the four cloned TtPMM members (Additional file 5). Finally, a single fragment was amplified from B. distachyon (Bd21) and its size agreed with the one calculated from the cloned BdPMM genomic sequence (Additional file 5).

The copy number of PMM genes in an additional set of materials (containing 12 diploid and 9 tetraploid Triticeae species) was also investigated by fragment analysis using the IF primers (see above). The results showed that the diploid Triticeae species carrying the AmAm, CC, MM, NN, RR, SS, SbSb, SlSl, SsSs, SshSsh, StSt or UU genomes generally had two different PMM fragments, whereas the tetraploid Triticeae species containing the AAGG, DDCC, DDMM, DDNN or UUMM genomes generally possessed four PMM specific fragments (Additional file 6).

The fluorescently labeled IF primers were further employed for investigating the chromosomal locations of wheat and barley PMM members by PCR amplifications. The fragments derived from PMM-A1, B1 and D1 of bread wheat were not amplified from the NT lines lacking 2A, 2B and 2D chromosomes, respectively (Additional file 7), indicating that TaPMM-1 homoeologs were located on group 2 chromosomes. This finding was confirmed by the fragment analysis data obtained with LDN substitution lines (Additional file 7). The fragments derived from PMM-A2 and D2 members of bread wheat were not amplified from the NT lines lacking 4A and 4D chromosomes, respectively (Additional file 7). Moreover, the fragment from TtPMM-B2 was absent in the LDN substitution line lacking the 4B chromosome (Additional file 7). Together, these results indicated that TaPMM-2 homoeologs were located on group 4 chromosomes. Using a similar strategy, the two barley PMM genes were mapped to the 2H and 4H chromosomes, respectively (Additional file 8), which are homoeologous to wheat groups 2 and 4 chromosomes, respectively [27, 28].

The finding of six PMM genes located on groups 2 and 4 chromosomes in bread wheat by this work is consistent with the data generated in the wheat expressed sequence tag (EST) mapping project http://wheat.pw.usda.gov/NSF/progressmapping.html. The EST BE489906, a 421 bp cDNA exhibiting higher than 95% nucleotide sequence identity to the 5' half of the mRNAs of the six bread wheat PMM genes, has been used as a probe to hybridize with the genomic DNA samples of CS and associated NT lines digested with the restriction enzyme EcoRI. The hybridization yields 11 positive bands, which are mapped to groups 2 or 4 chromosomes. Our bioinformatic analysis suggests the presence of three to five EcoRI sites in the genomic ORF of bread wheat PMM genes, with the last EcoRI site located in the 8th exon (Additional file 9). Because the first seven exons give rise to the 5' half of PMM mRNAs upon post transcriptional processing, the DNA blot hybridization with BE489906 probe is expected to show about 11 to 12 hybridizing bands from six TaPMM gene members. The hybridization with another PMM EST probe (BE591860) yields approximately seven positive bands, which are mapped mainly to groups 2 or 4 chromosomes. Consequently, it is suggested that the number of homoeologous PMM genes in bread wheat is likely to be around six. However, we can not exclude the existence of additional and divergent PMM copies or fragments in bread wheat genome, because they may not be amplified by our PCR strategy (owing to mutations in the primer binding sites), or be distinguishable from one or more of the six completely sequenced copies during fragment analysis (because of the lack of difference in intron length).

Duplication of PMM in diploid Triticeaespecies

The foregoing experiments indicated that there existed two PMM genes (PMM-1, PMM-2) in diploid Triticeae species, and a single PMM gene in either B. distachyon or rice. These findings raised an important question on how and when the two PMM gene members in diploid Triticeae species were evolved. To investigate this question, phylogenetic analysis was conducted using the genomic DNA sequences of PMM genes. Two distinct clades, one composed of Triticeae PMM-2 homoeologs and the other of the PMM-1 homoeologs from common wheat and progenitors, were consistently found in the phylogenetic trees constructed using different tree building methods (Figure 2). HvPMM-1, BdPMM and OsPMM fell outside of the two clades, with HvPMM-1 showing a closer association with both clades than BdPMM and OsPMM (Figure 2). These observations, plus the established synteny between barley chromosome 2 (harboring HvPMM-1) and rice chromosome 4 (on which OsPMM resides) [27–29], indicated that the Triticeae PMM-1 genes were most likely the orthologs of BdPMM and OsPMM, and that the Triticeae PMM-2 member may be derived from PMM-1 through duplication. Trees with topology highly similar to that displayed in Figure 2 were also obtained when the deduced amino acid sequences of the PMM genes were used for phylogenetic analysis, but they could not reveal the phylogenetic relationships of PMM-A2 homoeologs to other PMM genes because TaPMM-A2, TtPMM-A2 and TuPMM-A2 could not yield intact proteins upon conceptual translation (Table 1).

Figure 2
figure 2

Phylogenetic analysis of PMM genes. The PMM sequences from bread wheat (TaPMM-A1, B1, D1, A2, B2, and D2), T. turgidum (TtPMM-A1, B1, A2, and B2), T. urartu (TuPMM-A1 and A2), Ae. tauschii (AetPMM-D1 and D2), barley (HvPMM-1 and 2), B. distachyon (BdPMM), rice (OsPMM, chromosomal locus Os04g0682300), and Arabidopsis (AtPMM, chromosomal locus At2g45790) were used for this analysis. The filled dot and asterisk indicate the two clades composed of Triticeae PMM-1 and 2 homoeologs, respectively. The tree shown was constructed based on an alignment of the nucleotide sequences of PMM genes by neighbor joining method (with P distance and complete deletion options). Highly similar results were obtained by alternative tree building programs (UPGMA, minimum evolution) (data not shown). The bootstrap values were each estimated using 500 replications. AtPMM was used as an outgroup control.

Full size image

To find the approximate timing at which the duplication of Triticeae PMM occurred, we first calculated the number of substitutions per synonymous site for the grass PMM genes characterized in this work. The values of synonymous substitutions calculated with PMM-1 genes were generally lower compared to those obtained with PMM-2 members (Table 2). This prompted us to calculate the molecular clocks for grass PMM-1 and 2 genes, respectively. Based on the averaged synonymous substitutions between OsPMM and wheat PMM-1 or 2 genes (Table 2), and assuming that rice and wheat diverged about 60 MYA [22], the average nucleotide substitution rate for PMM-1 genes was estimated to be 3.7 × 10-9 per site per year, whereas that for PMM-2 genes was 4.2 × 10-9 per site per year. BdPMM and wheat PMM-1 genes diverged about 40.2 ± 6.6 MYA (Table 3). The divergence time between wheat PMM-1 and 2 genes was estimated to be 18.9 ± 4.0 or 16.6 ± 3.6 MYA, depending on whether the clock rates of PMM-1 or 2 genes were used for the calculation (Table 3).

Table 2 Relative synonymous substitutions of PMM genes from rice, B. distachyon, bread wheat, and barley a
Full size table
Table 3 Divergence time estimates for PMM genes
Full size table

Analysis of evolutionary rate and potential positive selection

The cDNA sequence of 11 PMM genes (including seven PMM-1 and four PMM-2 members from bread wheat and its progenitor species) were aligned (Additional file 10), and the resultant alignment was used to construct a NJ tree (Additional file 10). The non-synonymous/synonymous substitution rate ratios (ω = dN/dS) were calculated for analyzing evolution rate and potential positive selection on amino acid sites by appropriate models (Table 4). The analysis with one-ratio model (M0) produced an averaged ω value of 0.15, indicating that purifying selection dominated the evolution of the examined PMM gene members (Table 4). The analysis with two-ratio model revealed that the mean ω value of PMM-2 genes was larger than that of PMM-1 members, though this difference was not statistically significant by the likelihood-ratio test (LRT) comparing one-ratio with two-ratio models (Table 4). The analysis with six site specific models, M1a, M2a, M3 (k = 2), M3 (k = 3), M7 and M8, suggested that more than 90% of codon sites were under strong purifying selection (with p ranging from 0.91 to 0.99, Table 4). Two candidate sites (10 and 96) were found to be under positive selection based on naïve empirical Bayes (NEB) posterior probability, one of which (site 10) was significant (Table 4, Additional file 10). However, the positive selection on site 10 was not significant according to Bayes empirical Bayes (BEB) posterior probability, and neither was it supported by LRT (Table 4).

Table 4 Maximum-likelihood analysis of PMM-1 and 2 genes and detection of codon sites under positive selection
Full size table

Transcriptional patterns of six TaPMMgenes

Semiquantitative RT-PCR experiments, using member specific primers, were conducted to compare the transcriptional patterns of six PMM genes in Xiaoyan 54. In general, the transcripts of the six PMM members were detectable in both vegetative (root, culm, leaf) and reproductive (immature spikes) organs, with relatively higher transcript levels detected in the seedling and flag leaves (Figure 3). The transcript levels of TaPMM-1 members (A1, B1, D1) were generally and relatively higher than those of TaPMM-2 members (A2, B2, D2) (Figure 3). The transcriptional profiles of the three TaPMM-1 genes were basically similar, except that the transcript level of D1 in the immature spikes was much lower relative to those of A1 or B1 (Figure 3). Among the three TaPMM-2 members, B2 and D2 showed a highly similar transcriptional profile (Figure 3). The transcript levels of A2 in the roots, culms and immature spikes were substantially lower than those of B2 and D2 in the same set of organs (Figure 3).

Figure 3
figure 3

Comparisons of the transcript levels of six TaPMM genes ( A1 , B1 , D1 , A2 , B2 , and D2 ) in Xiaoyan 54 organs collected at the seedling and bolting stages. The total RNA samples, extracted from the roots (R), seedling leaves (L), culms (C), flag leaves (FL), and immature spikes (IS), were converted into cDNAs through reverse transcription, followed by semiquantitative RT-PCR analysis with gene specific primers. The amplification of tubulin transcripts served as an internal control (for normalizing the cDNA levels of different reverse transcription samples and for checking the kinetics of PCR amplification). The numbers of amplification cycles for TaPMM and tubulin genes are shown on the right side of the graph. The data depicted are representative of three independent experiments.

Full size image

Biochemical activity of TaPMM proteins

First, yeast complementation assays were carried out to test if the proteins expressed from TaPMM-A1, B1, D1, B2 or D2 may complement the temperature sensitivity of the sec53-6 mutant of S. cerevisiae. TaPMM-A2 was not included because it did not possess an intact ORF (Table 1). From Figure 4, it is clear that the expression of TaPMM-A1, B1, D1 or D2 restored the growth of sec53-6 cells at 30 or 37°C, and to a similar degree. By contrast, the expression of TaPMM-B2 complemented PMM deficiency of sec53-6 only weakly at 30°C, with no complementation observed at 37°C (Figure 4).

Figure 4
figure 4

Yeast complementation analysis of TaPMM genes. Five yeast constructs, harboring the cDNA of TaPMM-A1 (p181-A1), B1 (p181-B1), D1 (p181-D1), B2 (p181-B2) or D2 (p181-D2), were prepared using the expression vector p181AINE (containing a Leu2 selection marker). These constructs, together with the empty vector control (p181-VC), were transformed into the sec53-6 strain of S. cerevisiae, which does not grow at elevated temperatures (30°C or above) because of temperature sensitive mutation in the native PMM. The six recombinant strains, sec53-6, and RSY249 (the wild type progenitor of sec53-6) were serially diluted and spotted onto the medium lacking leucine, followed by incubation at three test temperatures (24, 30 or 37°C), The data shown are reproducible in three independent experiments.

Full size image

Subsequently, TaPMM-A1, B1, B2, D1 and D2 were individually expressed in the bacterial cells, and the resultant recombinant proteins were purified using nickel affinity chromatography (Additional file 11). The recombinant proteins of A1, B2, D1 and D2 were examined for catalytic activity using mannose-1-phosphate as the substrate. The B1 recombinant protein was not included in this analysis because its deduced amino acid sequence was identical to that of A1 (Additional file 2). In three independent experiments, B2 did not show detectable catalytic activity. By contrast, the recombinant A1, D1 and D2 proteins were all catalytically active. The average Km (mM, mannose-1-phosphate) values calculated for recombinant A1, D1 and D2 proteins were 0.46 ± 0.02, 0.40 ± 0.02 and 0.48 ± 0.02, respectively.

Temperature-activity profiles of the recombinant PMMs of rice, B. distachyon, barley and bread wheat

OsPMM, BdPMM and HvPMM-1 were expressed in the bacterial cells, and the resultant recombinant proteins were purified as described above. The main temperature-activity characteristics of the three PMMs and TaPMM-A1, D1 and D2 were compared using 1 mM mannose-1-phosphate as the substrate and at three test temperatures (24, 30 and 37°C). From Figure 5, it is evident that the activity level of OsPMM was significantly higher at 30 and 37°C than at 24°C. Moreover, increasing the test temperature from 30 and 37°C did not decrease OsPMM activity significantly (Figure 5). BdPMM was most active at 30°C, but its activity level was reduced by nearly 95% as the temperature increased from 30 to 37°C (Figure 5). HvPMM-1 was most active at 24°C, and its activity level also decreased sharply (by more than 90%) as the temperature shifted from 30 to 37°C (Figure 5). The three bread wheat PMMs exhibited different temperature-activity characteristics. TaPMM-A1 was less active than BdPMM and HvPMM-1 at 24 and 30°C, but its activity level was substantially higher than those of BdPMM and HvPMM-1 at 37°C, owing to a 51% increase of its activity as the temperature changed from 30 to 37°C (Figure 5). TaPMM-D2, although less active than BdPMM and HvPMM-1 at 24 and 30°C, displayed substantially higher activity levels than BdPMM and HvPMM-1 at 37°C, because the change of test temperature from 30 to 37°C did not reduce TaPMM-D2 activity as strongly as that occurred to BdPMM and HvPMM-1 (Figure 5). The activity level of TaPMM-D1 was lower than those of BdPMM and HvPMM-1 but higher than those of TaPMM-A1 and D2 at 24°C (Figure 5). However, TaPMM-D1 was highly active at 30°C, with an activity level statistically comparable to that of OsPMM (Figure 5). The shift of assay temperature from 30 to 37°C reduced the activity of TaPMM-D1 by about 40%, but the scale of this reduction was much lower compared to that observed for BdPMM, HvPMM-1 or TaPMM-D2.

Figure 5
figure 5

Temperature-activity characteristics of the recombinant PMMs of bread wheat (TaPMM-A1, D1, and D2), barley (HvPMM-1), B. distachyon (BdPMM), and rice (OsPMM). The purified recombinant PMMs were each tested under three temperature conditions (24, 30 and 37°C), with mannose-1-phosphate as the substrate. The OD340 readings, recorded at 10 min after the start of the reaction, were used to compare the relative activity levels of the six PMMs under different temperatures. Means ± s.d. were each calculated using the results of triplicate measurements. Different letters above the bars suggest that the means differ significantly (P £ 0.05). The data shown are representative of three independent experiments.

Full size image

Discussion

PMM genes and chromosomal loci in Triticeaespecies

Through the molecular cloning experiments in this work, the following suggestions may be made regarding PMM genes and chromosome loci in Triticeae species. First, there exist two distinct PMM gene members (PMM-1 and 2) in the progenitor diploid genomes (i.e., A, B and D) of bread wheat. There are four and six PMM genes (including both active and inactive members) in tetraploid and hexaploid wheats, respectively. Second, in bread wheat and its diploid and tetraploid progenitors, the loci containing PMM-1 or 2 genes are located on homoeologous groups 2 and 4 chromosomes, respectively. Third, two PMM genes (HvPMM-1 and 2) exist in diploid barley. The loci harboring HvPMM-1 or 2 are located on the 2H and 4H chromosomes, respectively. Finally, judging from the fragment analysis data of PMM genes described in Additional file 6, it is probable that the basic organization of PMM gene family, as defined for the A, B, D and H genomes (e.g., two distinct members contained in two separate chromosomal loci), is conserved in the diploid Triticeae species possessing AmAm, CC, MM, NN, RR, SS, SbSb, SlSl, SsSs, SshSsh, StSt or UU genomes. However, further work is needed to verify if in the 12 Triticeae genomes the two PMM genes may also be located on homoeologous group 2 and 4 chromosomes.

The existence of two PMM genes in diploid Triticeae species contrasts with the finding of a single PMM gene for Arabidopsis, rice and B. distachyon [[6, 12], this work]. A search of recently accumulated genomic data deposited in the phytozome website (http://www.phytozome.net) reveals that Medicago truncatula, Vitis vinifera, Sorghum bicolor and Zea mays each possess a single PMM gene. However, both Glycine max and Populus trichocarpa harbor two PMM copies in their genomes (http://www.phytozome.net). Two functional PMM copies have also been found present in several mammals (e.g., human, mouse and rat) [30–32].

Biochemically active PMM members in bread wheat and its progenitors

From the presence of intact ORF, the results of yeast complementation experiments, and the data of biochemical assays using recombinant proteins, we deduce that bread wheat expresses four biochemically active PMMs, TaPMM-A1, B1, D1 and D2. The primary structure and basic biochemical function of the four PMMs are similar to their orthologs in human, S. cerevisiae and Arabidopsis. In contrast to TaPMM-A1, B1, D1 and D2, TaPMM-B2 is probably not active in vivo because it did not complement PMM deficiency of yeast sec53-6 cells efficiently, and was not catalytically competent in the in vitro biochemical assay. The lack of detectable biochemical activity by TaPMM-B2 correlates with the presence of unusual amino acid substitutions in its deduced protein (Additional file 3). Further work is needed to determine the specific amino acid change(s) responsible for decreasing the biochemical activity of TaPMM-B2.

Owing to the possession of identical amino acid sequences (Additional file 2), it is likely that TuPMM-A1 (in T. urartu) behaves like TaPMM-A1, and AetPMM-D1 and D2 of Ae. tauschii may act like TaPMM-D1 and D2, respectively. Contrary to the above scenario, it is not possible to deduce the biochemical activity of TtPMM-B1 and B2 based on the data gathered for TaPMM-B1 and B2, because complete amino acid sequence identity did not exist between the two B1 or B2 members (Additional file 2).

Main features of PMM evolution in diploid and polyploid Triticeaespecies

Despite its vital importance in eukaryotes, the molecular evolution of PMM has not been investigated in higher plants previously, and has been studied only recently in animals. Several investigations suggest that the presence of two functional PMM copies in human and murine cells may be caused by a gene duplication event that occurred 75 to 110 MYA [30, 32–34]. From the data presented in this work, the main features in the evolution of PMM genes in diploid and polyploid Triticeae species may be summarized below. First, a major event in PMM evolution in the diploid Triticeae species carrying AA, AmAm, BB, CC, DD, HH, MM, NN, RR, SS, SbSb, SlSl, SsSs, SshSsh, StSt or UU genomes is the duplication of PMM. This duplication event gave rise to two different PMM genes, PMM-1 and 2, located on groups 2 and 4 chromosomes, respectively. Out of the two PMM genes in diploid Triticeae species, PMM-1 is likely more ancestral. This possibility is mainly supported by 1) the phylogenetic tree shown in Figure 2, 2) the existence of a single PMM gene in rice, and 3) the established syntenic relationship between Triticeae group 2 chromosomes (on which PMM-1 homoeologs are located) and rice chromosome 4 (on which OsPMM resides) [27–29]. Second, the PMM duplication, as discussed above, occurred after the divergence between B. distachyon and wheat, but before the radiation of diploid Triticeae genomes. This is supported by not only the presence of two different PMM gene members in many diploid Triticeae genomes examined in this work, but also the estimated divergence time between TaPMM-1 and TaPMM-2 genes (Table 3). The two estimates, whether 16.6 ± 3.6 or 18.9 ± 4.0 MY, are both much larger than the divergence time between wheat and barley (about 11 MY) reported in past studies [19–22]. But they are substantially smaller than the divergence time between B. distachyon and wheat (35 to 37.8 MY) calculated by previous investigations [35, 36]. Third, substantial differences exist in the evolutionary patterns of PMM-1 and 2 genes. PMM-1 genes generally had intact ORF, their transcript levels in the vegetative and reproductive organs were comparatively higher (Figure 3), and their protein products were all biochemically active (Figure 5). By contrast, multiple types of defects occurred to PMM-2 members. The PMM-A2 copies in bread wheat and its diploid and tetraploid progenitors and HvPMM-2 in barley possessed mutated ORF (Table 1). TaPMM-B2 encoded a protein with undetectable biochemical activity (Figure 4). The recombinant protein of TaPMM-D2 was significantly less active than that of the product of TaPMM-D1 (Figure 5). Moreover, the three PMM-2 members in bread wheat all exhibited lower transcript levels relative to their PMM-1 paralogs (Figure 3). Fourth, the differences of PMM-1 and 2 genes established at the diploid level (i.e., in the Au or D genomes) are basically retained in polyploid wheats. Thus, PMM-A1 is active in diploid, tetraploid and hexaploid levels whereas the ORF of PMM-A2 is disrupted at all three ploidy levels. The two PMM genes encoded in D genome are all active in both Ae. tauschii and bread wheat. In contrast to the PMM genes in A and D genomes, the origin of the contrasting biochemical features of the two B genome PMM members in bread wheat is unclear at present, because the absence of a diploid species containing the B genome made it difficult to study the evolution of PMM-B1 and B2 members at the diploid level. Finally, from the data described in Additional file 6, the duplicated PMM genes in the G, C, N, M and U genomes are also likely retained in the tetraploid species possessing AAGG, DDCC, DDMM, DDNN or UUMM genomes, although further work is needed to study the PMM genes in these species in more detail.

Because PMM-1 and 2 genes were located on groups 2 and 4 chromosomes, respectively, in wheat and progenitors and barley, the duplication event that led to the formation of two PMM copies in the ancestral Triticeae species must have occurred interchromosomally. Strong evidence for the occurrence of interchromosomal duplications in wheat genome evolution has been obtained in recent years [37–39]. Importantly, it has been suggested that these duplications are largely conserved across different Triticeae species [39, 40]. However, it is still not well understood how these events might be accomplished in Triticeae genomes. Thus, the PMM duplication reported here may serve as a useful model for studying the molecular mechanisms underlying the interchromosomal duplications in Triticeae in the future.

The average nucleotide substitution rates of grass PMM-1 or 2 genes estimated based on the 60 MY divergence time between rice and wheat by this work are both lower than the commonly adopted mutation rate (6.5 × 10-9 per site per year) for grass genes [41–44]. This may not be unusual because evolutionary rates frequently differ among different grass genes [45]. For example, the mutation rates of ten randomly selected grass genes have been found to vary from 4.1 to 7.1 × 10-9 per site per year [45]. Furthermore, we estimated that the divergence time between BdPMM and wheat PMM-1 genes was about 40.2 ± 6.6 MY (Table 3) using the average nucleotide substitution rate calculated for grass PMM-1 genes. This estimate is comparable to the previously reported timeline (35 to 37.8 MY) for the divergence of the two species [35, 36].

Possible mechanisms involved in the evolution of PMM-1 and 2genes

Past studies have suggested several fates in the evolution of duplicated genes [46–48]. Neofunctionalization leads to the acquisition of new function by one copy, whereas subfunctionalization renders the progenitor and the duplicated copy acquiring non-overlapping but complementary expression patterns. Non-functionalization results in pseudogenization of one of the duplicates. In barley, the duplicate copy of HvHox2 (encoding a homeodomain-leucine zipper transcription factor) acquired a new function in controlling spikelet fertility [49]. In rice, evidence for subfunctionalization has been obtained for two C-class MADS box genes (OSMADS3 and OSMADS58) [50]. The evolutionary process of duplicated genes may be divided into two stages [51]. During the early stage, the duplicates generally evolve under purifying selection, with a phase of relaxed selective constraint immediately after the duplication [51, 52]. During the later stage, the great majority of the duplicates become non-functional pseudogenes that are no longer expressed, with only a few of them being involved in neofunctionalization or subfunctionalization [51–53].

From the data in Table 4 and the main evolutionary features of PMM-1 and 2 genes discussed above, we suggest that 1) purifying selection, but not positive selection, may have dominated the evolution of PMM gene family in wheat and relatives, and 2) despite the prevalence of purifying selection, the function of PMM-2 copies is not as strongly maintained as that of PMM-1 members. The first point is consistent with the findings that the vast majority of duplicated genes in cellular organisms experience purifying selection [51, 52], even though some of them are in the process of becoming pseudogenes [54]. Owing to purifying selection, the PMM-1 genes are generally and relatively stable. By contrast, TaPMM-A2 is close to becoming a non-functional pseudogene, because its ORF was mutated and its transcript level was decreased relative to PMM-1 genes (Figure 3). This scenario may also apply to PMM-A2 copies in the diploid and tetraploid progenitors of bread wheat and HvPMM-2 in barley although further work is needed to check if the transcript levels of these members may be lowered than those of PMM-1 genes. TaPMM-B2 and D2, although possessing intact ORF, also show signs of degenerative evolution, because their transcript levels were lower than corresponding PMM-1 members, and the PMM enzyme activity of their recombinant proteins was either undetectable or reduced (Figure 4 and 5). The degeneration of PMM-2 genes and the comparatively high mean evolutionary rate of these members indicate the occurrence of relaxed purifying selection in the evolution of PMM gene family in wheat and relatives, which may have affected PMM-2 copies more strongly than PMM-1 paralogs. However, further work is needed to verify this hypothesis.

In contrast to the degenerative evolution of the duplicated PMM-2 copies in diploid Triticeae species, the two mammalian PMM members, products of a duplication event occurred before the radiation of mammals (i.e., 75 to 110 MYA), all encode highly active enzymes [29–33]. However, unlike PMM2 whose mutation leads to hypoglycosylation of cellular proteins and severe disease [9, 12], the deletion of PMM1 in mouse does not cause obvious abnormalities in either growth or development under normal conditions [33]. Moreover, despite their overlapping expression profiles, PMM1 does not compensate for the functional loss of PMM2 [32, 33]. These findings have led to the suggestion that PMM1 may have diverged from PMM2 in physiological function [33]. This possibility has been confirmed with the recent discovery that PMM1 but not PMM2 acts as a glucose-1,6-bisphosphatase in the brain tissues [55]. Clearly, the unique PMM duplication event in Triticeae plants revealed in this work differs from the one in mammals in multiple aspects including the timing of the event, the evolutionary fates of PMM members after the duplication, and possibly the main mechanism involved in the evolution. The evolutionary knowledge generated in this work increases our understanding of PMM evolution in eukaryotes, and may be useful for future studies of PMM evolution in additional plant species such as G. max and P. trichocarpa, both of which have been found to contain two PMM genes (see above).

The main active PMMisozymes in bread wheat tissues

Among the four biochemically active TaPMMs, functional similarity may be limited to only TaPMM-A1 and B1 because their deduced proteins were identical, and the transcriptional patterns of their coding genes were very similar (Figure 3). By contrast, functional differences are likely to exist among A1(B1), D1 and D2 members because of dissimilarities in their deduced amino acid sequences and transcriptional patterns. For further analysis of the physiological function of PMM genes in bread wheat, it is necessary to find out the main active PMM isozymes in vivo. Considering the fundamental importance of PMM in the temperature adaptability of eukaryotes [1, 12], the differences in the temperature-activity profiles of the four enzymatically active TaPMMs are of particular interest. From the data shown in Figure 5, it is suggested that TaPMM-D1 may be highly active in a wider range of temperatures (from 24 to 37°C). The activity of TaPMM-A1 is less than that of TaPMM-D1, but it is unique in exhibiting the highest activity at 37°C (although this activity level was still below that of TaPMM-D1 at 37°C). The activity of TaPMM-D2 is generally lower than that of TaPMM-D1 and A1, especially at 37°C (Figure 5). This finding and the observation that the transcript levels of TaPMM-D2 were considerably lower than those of TaPMM-A1, B1 and D1 in multiple organs (Figure 3) indicate that TaPMM-A1, B1 and D1 are likely to be the main active PMM isozymes in bread wheat tissues. Moreover, judging from Figure 3, at least two main PMM isozymes are expressed in each of the five organs (root, seedling leaf, culm, flag leaf, and developing spike) examined in this work.

As displayed in Figure 5, rice, a tropical cereal and adaptable to both tropical and temperate regions [56], possesses a PMM showing very high activity levels at elevated test temperatures (30 and 37°C). By contrast, B. distachyon, barley and bread wheat, originated from temperate regions [57–59], have the PMMs exhibiting relatively high activity levels at 24 or 30°C (Figure 5). A general decrease in the activities of the PMMs of B. distachyon, barley and bread wheat at 37°C seems to correlate with the temperate origin of these species. However, compared with BdPMM and HvPMM-1, at 37°C, the magnitude of activity decrease was much smaller for TaPMM-D1 and D2, and the activity of TaPMM-A1 even increased relative to its performance at 24 and 30°C. This indicates that bread wheat PMM isozymes may be generally more stable than their counterparts in B. distachyon and barley at elevated test temperatures (such as 37°C), which correlates well with the wide temperature adaptability of the bread wheat crop [13, 14].

Conclusion

This work has generated new information on PMM genes and their evolution in bread wheat and related Triticeae species. The main active PMM isozymes (e.g., TaPMM-A1, B1 and D1) in bread wheat tissues are revealed. The three isozymes may be more tolerant to elevated temperatures than their counterparts in barley and B. distachyon. The insights gained in this work have broadened the understanding of PMM evolution in eukaryotic organisms. The resources produced here may aid future investigations of the physiological function of this important gene in the temperature adaptability of bread wheat.

Methods

Plant materials, yeast strains and oligonucleotide primers

The wheat and barley varieties and the T. urartu, T. turgidum ssp. dicoccoides, Ae. tauschii and B. distachyon genotypes used for cloning PMM genes are listed in Additional file 1. The NT lines of CS, the D genome substitution lines in the tetraploid wheat variety LDN background, and the barley chromosome addition lines in CS background, employed for mapping the chromosomal locations of PMM genes, are also listed in Additional file 1. The growth of wheat and related Triticeae species was accomplished as described previously [6]. B. distachyon was cultured in the greenhouse under a constant temperature (23°C) and with a 16 h light/8 h dark photoperiod. Two S. cerevisiae strains, sec53-6 (MAT a, sec53-6, leu2-3, -112, ura3-52) and its wild type progenitor RSY249 used for the yeast complementation experiment, were described previously [6]. The oligonucleotide primers used in this work are all listed in Additional file 4. The general molecular methods for handling nucleic acid and protein samples, PCR amplification, DNA cloning, and protein expression in bacterial cells were adopted from Sambrook and Russell [60]. High fidelity Taq DNA polymerases, LA-Taq (TaKaRa, China) and Advantage 2 Polymerase Mix (Clontech, USA), were used in PCR amplifications. The accuracy of the yeast or bacterial expression constructs prepared in this work was all confirmed by DNA sequencing.

Cloning of PMMcDNA and genomic DNA sequences

Total RNA samples were prepared from the desired leaf materials using RNeasy plant mini kit (Qiagen, Germany), and were converted into cDNAs using Moloney Murine leukemia virus (M-MLV) reverse transcriptase (Promega, USA). Public nucleotide and protein data banks were searched using the cDNA and deduced amino acid sequence of rice PMM (OsPMM) [6], leading to the finding of the ESTs derived from wheat and barley PMM genes. By analyzing the obtained ESTs, the oligonucleotide primers permitting the amplifications of full length cDNA and genomic DNA sequences of the PMM genes from bread wheat and related Triticeae species (T. urartu, T. turgidum, Ae. tauschii, barley) were developed (Additional file 4). With these primers and the reverse transcription products described above, the PMM cDNAs from bread wheat and relatives were amplified by RT-PCR, cloned and sequenced. The PMM cDNA of B. distachyon was cloned using a similar strategy, except that the development of the oligonucleotide primers specific for BdPMM was facilitated by the genome sequence of Bd21 (http://www.phytozome.net/brachy.php). Genomic DNA samples were extracted from the relevant leaf materials as detailed previously [61], and used for isolating the genomic coding sequences of the PMM genes of bread wheat and related Triticeae species and B. distachyon with the same sets of primers employed for cDNA cloning.

During the above experiments, at least 70 positive PCR clones derived from three separate cloning experiments were sequenced for each of the plant species examined in this work. This ensured the identification of the cDNA and genomic DNA sequences corresponding to different PMM genes. The final nucleotide sequences for the cDNA and genomic DNA coding regions of a given PMM gene were each constructed from the sequencing information of at least three independent clones. The nucleotide sequences of the PMM genes cloned in this work have all been submitted to GenBank. The accession numbers are listed in Additional file 12.

Investigation of PMMcopy number and chromosomal location

The genomic PCR products amplified with the IF primers (Additional file 4, Figure 1) from Xiaoyan 54 were confirmed to be derived from PMM genes by cloning and sequencing. The amplified fragments differed in size, which was thus useful for distinguishing the different PMM genes and estimating the copy number of PMM in bread wheat, its Triticeae relatives, and B. distachyon. The forward IF primer was labeled with D4 fluorescence by a commercial company (TaKaRa, China), allowing the separation and identification of PCR products via capillary electrophoresis (CE) and fragment analysis in the CEQ/GeXP Genetics Analysis System (GAS) (Beckman, USA). The preparation of PCR products for CE and the subsequent fragment analysis in CEQ/GeXP GAS were conducted following the supplier's instructions. Using the same strategy, PMM copy numbers in the Triticeae species carrying the diploid AmAm, CC, MM, NN, RR, SS, SbSb, SlSl, SsSs, SshSsh, StSt or UU genomes, or the tetraploid AAGG, DDCC, DDMM, DDNN or UUMM genomes (Additional file 6), were also investigated.

For investigating PMM chromosomal location, the PMM gene specific fragments, amplified from CS, the different NT lines of CS, LDN, and the D genome substitution lines in LDN background, were separated using CE and differentiated by fragment analysis. The IF primers and fragment analysis were also used for mapping the chromosomal locations of two barley PMM genes (HvPMM-1 and 2). In this case, the DNA samples extracted from the barley variety Betzes and its derivative chromosome addition lines in CS background were utilized as templates for genomic PCR amplifications.

Bioinformatic, phylogenetic and evolutionary analysis of PMM

Bioinformatic investigations of cloned PMM genes were carried out using the softwares installed in public websites, including EBI (http://www.ebi.ac.uk/Tools/), ExPASy (http://www.expasy.ch/tools/) and NCBI (http://www.ncbi.nlm.nih.gov/). Full length genomic DNA sequences of PMM genes were aligned using the Clustalw2 program (EBI website). The resulted alignment was used for constructing phylogenetic trees by MEGA4 [62]. The number of substitutions per synonymous site (dS) was calculated as described previously [63]. The average nucleotide substitution rates for PMM-1 and 2 genes were estimated by the formula k = dS/2t [64], assuming a divergence time of 60 MY between rice and wheat [22]. The obtained rates were then used to estimate the timing of PMM duplication in wheat.

To investigate the potential mechanisms behind the evolution of PMM genes in bread wheat and its progenitor Triticeae species, two complementary evolutionary tests (one for difference in the evolutionary rates of PMM-1 and 2 genes, and the other for the codon sites under positive selection) were carried out by the CODEML program implemented in the PAML package [65]. The NJ tree, constructed using the cDNA sequence of 11 active PMM-1 and 2 genes, was employed for these tests. The one-ratio and two-ratio models [66, 67] were used for investigating evolutionary rate, whereas the site specific models M1a, M2a, M3 (k = 2), M3 (k = 3), M7 and M8 [68, 69] were deployed for testing potential diversifying selection at individual codon sites. Statistical analysis of the evolutionary data was conducted as described previously, with the confidence level set at or above 0.95 [70, 71].

Semiquantitative RT-PCR

The seeds of Xiaoyan 54 were germinated for 48 h at 25°C, followed by 4 weeks vernalization treatment at 4°C. The vernalized seedlings were then grown in the greenhouse as described above. The root and seedling leaf samples were collected after two weeks growth in the greenhouse. The stem, flag leaf, and immature spike samples were collected at the bolting stage (after 5 weeks growth in the greenhouse). The isolation of total RNA samples from the collected materials and cDNA synthesis were conducted as described above. Semiquantitative RT-PCR analysis of the relative transcript levels of 6 TaPMM genes in different wheat organs was performed using gene specific primers (Additional file 4), and following the method detailed previously [72]. The amplification of wheat tubulin transcripts served as an internal control, and the reproducibility of the transcriptional patterns revealed was verified in three independent experiments.

Yeast complementation

The complementation assay was conducted as described previously [6]. Briefly, five complementation constructs (p181-A1, B1, D1, B2 and D2) were prepared by cloning the cDNAs of TaPMM-A1, B1, D1, B2 and D2 individually into the yeast expression vector p181AINE. The oligonucleotide primers used for amplifying TaPMM cDNAs during the cloning are listed in Additional file 4. The five constructs, together with the empty vector control (p181-VC), were individually introduced into the sec53-6 strain whose endogenous PMM harbors a temperature sensitive mutation [1]. The recombinant strains were serially diluted and plated onto agar medium, followed by incubation at three temperatures (24, 30 and 37°C). As controls, sec53-6 and its wild type progenitor RSY249 were also included in the growth assays. Successful complementation by TaPMM expression was indicated by restoring the growth of sec53-6 cells at elevated temperatures (30 and 37°C). The complementation assay was repeated three times to verify the reproducibility of the data obtained.

Biochemical analysis of recombinant PMMs

The PMM cDNAs from bread wheat (TaPMM-A1, B1, D1, B2 and D2), barley (HvPMM-1), B. distachyon (BdPMM) and rice (OsPMM) were individually cloned into the bacterial expression vector pET-30a (Novagen, USA), in order to express recombinant PMM proteins with a C-terminal histidine tag. The oligonucleotide primers used for amplifying PMM cDNAs during preparing the bacterial expression constructs are provided in Additional file 4. The induction of PMM cDNA expression in the bacterial cells and the purification of histidine tagged PMMs through metal chelate affinity chromatography were accomplished as described previously [6]. The purified PMMs were characterized using 10% SDS-PAGE, and were confirmed to be the desired histidine tagged proteins by protein blotting using an anti-histidine antibody (Roche Diagnostic GmbH, Germany) [6].

The kinetic properties of the purified recombinant TaPMM-A1, B2, D1 and D2 were investigated using mannose-1-phosphate as the substrate, and following the method detailed in our earlier work [6]. The assay was repeated three times for each of the four recombinant TaPMMs. The resultant data were used to calculate average Km values. The temperature-activity characteristics of OsPMM, BdPMM, HvPMM-1, and TaPMM-A1, D1 and D2 were examined using 1 mmol/L mannose-1-phosphate as the substrate, with the assays conducted at three different temperatures (24, 30 and 37°C). The OD340 readings were collected at 10 min after the start of the enzyme assays. At this time point, the reactions catalyzed by the six recombinant PMMs were in the linear phase. The assay was repeated three times per recombinant PMM per test temperature, with each assay consisting of three technical replicates. The OD340 readings obtained were statistically analyzed using the software SPSS 10 for comparing the relative activity levels of the six recombinant PMMs.