Ph1 Locus at a Molecular Level
Given Ph1 is a deletion phenotype effect, its mapping required the screening of mutagenised hexaploid and tetraploid wheat populations to identify a set of overlapping deletions covering chromosome 5B. Over a 15-year period ten mutagenised populations were developed and screened in hexaploid wheat (Roberts et al. 1999), and one mutagenised population in tetraploid wheat with the help of Shrahryar Kianian. The deletion breakpoints needed to be located and the gene content of the regions covered by these deletions revealed. The hexaploid wheat genome is 5x larger than the human genome and was unsequenced. To solve this problem, the use of the small rice genome, as a model for the larger wheat genome was postulated, should there be conservation of gene order (Moore et al. 1993). This needed to be confirmed before the Ph1 cloning strategy could be implemented (and funded). Conservation in gene order, was confirmed first at the genetic and then at the physical level (Moore et al. 1993; Kurata et al. 1994; Foote et al. 1997; Griffiths et al. 2006). Brachypodium with its small genome was also added to this concept (Moore et al. 1993; Foote et al. 2004; Griffiths et al. 2006). Taking this approach further, it should be possible to reconstruct the ancestral genome from which the genomes of present day cereals and grasses have evolved (Moore et al. 1993). To this end, mapping data for rice from Japan, for maize and sorghum from North America, for sugarcane from France and finally for rye, wheat and millets from the UK was used. From these datasets, there was indeed a pattern of genomic building blocks or groups of genes within the rice genome, which could be used to describe the structure of all the other cereal chromosomes (Moore et al. 1995). The comparison of the order of blocks within the different cereal chromosomes, revealed that they could all be derived from the cleavage of a single structure, a hypothetical 'ancestral' genome, formed from the blocks, and a diagrammatical framework for comparing the order of all the major cereal genomes unified cereal genetics (Moore et al. 1995). This concept was purposefully termed “Synteny”, which in classical genetics had been used in a different context, but which is today the widely used term for the concept, indicating its cereal origin (see Encyclopaedia Britannica). With the development of this concept, funding was made available from BBSRC for the Ph1 cloning strategy, in particular the development of genomic libraries for wheat and Brachypodium in order to generate a physical contig of the Ph1 locus. A 1.2 million clone BAC library was constructed with INRA (Allouis et al. 2003). The Ph1 deletion effect region was delineated by phenotyping the deletion lines and mapping the deletion breakpoints using Synteny. The breakpoints of deletions lacking Ph1 clustered nonrandomly either side of a 2.5 Mb region carrying a large segment of satellite DNA, located within an amplified Cdk locus (Griffiths et al. 2006). There is also an anther specific gene within the delimited Ph1 region, which is the homologue of RA8 in rice, now named Raftin1 protein. We initially named the annotated wheat gene, as RA8, then Raf1 in subsequent analyses (Griffiths et al. 2006; Al-Kaff et al. 2008). The Raftin genes are anther specific and have been extensively characterized in rice, and now in wheat (Jeon et al. 1999; Wang et al. 2003; Sheng et al. 2011). They are mainly expressed in tapetal cells and are responsible for transporting lipids and cell wall proteins to the developing meiocytes. Mutation or deletion of the genes produces male steriles, as a result of the microspores becoming stressed (dehydrated). Stressed meiocytes exhibit chromosome clumping or clustering at metaphase I. Thus the genes have been patented in rice, maize and wheat for making male steriles in hybrid production. We excluded this gene as being responsible for the Ph1 effect because: it is only expressed on the male side and not on the female side; it is not expressed during the stages when recombination occurred on the male side; and finally if the 5B copy is functional, its deletion would result in male steriles, which are not observed with deletion of the Ph1 region. Consistent with this observation, the 5B copy of RA8/Raf1 carries an early stop codon, and the transcripts derived from this copy are antisense and not sense. The transcripts run into the promoter regions and contain exonintron junctions in the incorrect orientation.
Subsequently, we identified two additional deletion mutants which possessed wild type pairing in wheat itself, and therefore both retained the Ph1 locus (Al-Kaff et al. 2008). One of these deletions encompassed the RA8/Raf1 gene. Thus by a process of exclusion, the analysis delineated the Ph1 locus to a region where nearly half the genes are a cluster of kinases, including Cdk2-like genes. Expression analysis revealed that many of these Cdk2-like genes are expressed during meiotic prophase I, where the processes of pairing and recombination occur. To take the molecular study further required a working hypothesis for Ph1's mode of action. Given nearly half the genes in the delimited region are kinases, our hypothesis is that Ph1 affected kinase activity and hence overall phosphorylation levels. Amongst these kinases is a cluster of defective kinase genes (Cdk-like), with similarity to Cdk2 (Griffiths et al. 2006; Al-Kaff et al. 2008; Yousafzai et al. 2010). Therefore the deletion of Ph1 region could result in either an increase or decrease Cdk activity and phosphorylation levels, and that this altered phosphorylation levels could induce pairing between related chromosomes. We were able to test whether increasing Cdk-type activity phenocopies the effect of deleting the Ph1 locus. Treatment with okadaic acid, a serine-theonine phosphatase inhibitor increases Cdk-type activity. Treatment of detached tillers from Ph1 wheat-rye hybrids with okadaic acid from the onset of meiosis, does indeed phenocopy the effect of Ph1 deletion by inducing metaphase I pairing between related chromosomes (Knight et al. 2010). Thus increased phosphorylation levels overcomes the stalling of MLH1 sites on paired homoeologues in the presence of Ph1, enabling some of the sites to progress to crossovers which are visualized as pairing between related chromosomes at metaphase I. However does deleting the Ph1 region actually increase phosphorylation levels during meiosis? Our mapping of Ph1 region reveals the presence of a defective Cdk2-like kinase complex, which therefore could suppress active Cdk2-like genes via a dominant negative effect. Consistent with this proposal, phosphoproteomics revealed that phosphorylation at Cdk2 consensus sites on Histone H1 is increased in the absence of Ph1 (Greer et al. 2012). As indicated previously, phosphatases directly dephosphorylate proteins including Cdks, and are inhibited by Okadaic acid, which therefore can increase Cdk2 type activity and hence Cdk2-type phosphorylation. Okadaic acid treatment during meiosis mimics the effect of deleting Ph1 by inducing pairing and recombination between homoeologues even in the presence of Ph1 (Knight et al. 2010). This treatment also increases phosphorylation of the same Cdk2 consensus sites on Histone H1 as deleting Ph1 (Greer et al. 2012). Thus the reduced phosphorylation levels at Cdk2 consensus sites (hence Cdk2-type activity) in the presence of Ph1 and the stalling of MLH1 complex (which in other species has been shown to contain CDK2) on Double Holliday Junctions between paired homoeologues, are all entirely consistent. The Ph1 data implies that the MLH1 complex needs to be more active to resolve junctions on paired homoeologues than it does for junctions between paired homologues. This is consistent with the observations of Dvorak and colleagues. They found that in the absence of Ph1, recombination occurred between a pair of wheat chromosomes composed of combinations of homoeologous and homologous segments, but in the presence of Ph1, recombination was restricted to homologous segments (Dubcovsky et al. 1995). Interestingly mutating the Ph1 Cdk homologue in Arabidopsis also affects meiotic chromosome pairing (see Wen 2011 for initial studies on models).