Our goal is to create the resource, expertise and infrastructure that will be required by internationally competitive BBSRC sponsored research in a market worth £75B. The availability of UNENCUMBERED RESOURCES will encourage UK researchers to transfer from model to agronomically important species. This will promote the exploration of key aspects of cereal biology. In recognition of the importance of cereals to Scottish agriculture, the Scottish Office has agreed to contribute to the cost of the barley research.

1. A hexaploid wheat BAC library of more than 600,000 clones with an average insert size of 125Kb. This would provide a 99% probability of having any target DNA sequence represented in the library. The library would be arrayed on nylon-filter membranes for screening.

2. A first class wheat transcriptome resource. This resource will consist of; A. A collection of single variety, high quality cDNA libraries prepared from various stages in wheat development (but focussing on grain development), B. EST sequences from approximately 40,000 clones, C. A high-density array of approximately 10,000 unique, selected ESTs and D. The continuing resource to both screen the arrays and analyse the data produced.

3. A first class barley transcriptome resource (mirroring that produced for wheat at IACR-LARS). It will consist of A. A collection of high quality cDNA libraries prepared from various stages of barley development from germination through to grain development and desiccation. Stress will be a secondary focus. B. EST sequences from ca. 40,000 clones. C. A high-density array of ca. 10,000 unique, selected EST clones.

4. A high density Gene Knock out array representing 10,000 Mu-active plants. This facility will allow IACR-LARS to make its existing maize Mu-tagged fragment resource more widely available to the UK research community.

5. A high-density deletion mutation grid. To our knowledge this facility will be unique to the UK barley research community and will provide the necessary tools to correlate gene function with phenotype.

The arrays described in 4. and 5. will be extremely valuable to researchers who wish to identify specific plants containing gene 'knock out' or 'change of function' mutations in those genes identified via both the transcriptome resource and other funding opportunities.

6. A password protected WWW accessible resource database.

In addition, a number of existing resources will also be incorporated. These will include; a barley BAC library from SCRI, two fast neutron wheat deletion mutant populations and a rice YAC library from the JIC and a maize YAC and Maize BAC library from IACR-LARS. Wheat, barley and rice transformation will be made available through collaboration with the contributing centres. Well equipped, state of the art laboratories capable of servicing the UK cereal science base will complete the cereal community resources.

Wheat and barley are members of the Triticeae and can be considered genetically similar. The worldwide production of wheat is put at 579 million tons. Despite its relatively low protein content, wheat is the single most important source of plant protein in the human diet. It is the most important agricultural crop in the UK. In 1998 more than 2 million ha were planted to wheat, producing more than 14 million tons with a farmgate value of £1.56B (26% for export). Around 5 million tonnes of wheat was milled in the UK of which 85% was home grown. About 4 million tons of flour is produced each year with 60% being used for bread making. Around 1.2 million ha are planted to barley producing more than 7 million tons with a value of £890 million. While the majority of the UK crop is used for animal feed, 50-70% of the Scottish crop, which is mainly spring barley, is malted for the production of beer and whisky. Whisky is in the top five export earners for the UK, earning over £2.4B in 1996 and is a mainstay of the Scottish economy employing over 13,000 people. Wheat research has traditionally fallen mainly within the remit of BBSRC, while most barley research is within the remit of the Scottish Office. The present proposal brings the research on both crops together.

Wheat and barley breeding programmes aim to improve productivity, grain quality and disease resistance. The advent of new molecular genetic technologies provides the opportunity to underpin plant-breeding programmes. The UK has been responsible for many of the advances in these technologies. Their application however requires significant resources in the development stages. Thus agri-biotech companies have and are continuing to invest significant sums of money in cereal genomics. Wheat in particular has emerged as a key target for the application of the new biotechnology. A number of key agri-biotech companies now have major centres for wheat research in the UK. These include Monsanto, Advanta, Dupont and Astra-Zeneca.

The observation that cereals share extensive genetic similarity means that scientific advances in one cereal (for example in rice or maize) can benefit research on another cereal. The recent expansion in interest by the agri-biotech companies in wheat has resulted in the loss of key researchers from the UK academic community to industry. This loss will be coupled with the retirement of nearly half of the major wheat researchers worldwide in the next five years (as reflected in the loss of around half of the speakers at the last International Wheat Genetics Symposium). It was the recognition of a similar situation developing for maize research, coupled with the observation that less than half the genes in Arabidopsis were showing any homology to cereal ESTs (reported at Comparative Plant Genomics 1999, Cold Spring Harbor), which provoked the US National Academy to lobby for more investment in crop plants, and in particular for maize. The objective was to provide a framework for training the next generation of maize researchers. In response the US government via NSF provided approximately 75% of its $40 million budget for 1998 to cereal programmes (mostly maize). In addition it has been suggested that a significant amount of the 1999 budget ($50 million) will be directed to the small grain cereals such as wheat, barley and rice. However the view of the National Academy was that Europe also needed to respond. Very recently the French government in collaboration with industry has launched Geneoplante, a 1,400 million-Franc plant genome initiative. The German government has similarly announced a large plant functional genomics initiative worth some Dm150 million, which will focus on Arabidopsis and small grain cereals. A Finnish government/industrial collaboration has also invested in genomics research in small grain cereals. However, these programmes do not guarantee that the information will enter the public domain (in fact Geneoplante specifically excludes this possibility). Therefore while this funding will provide a boost to national plant genomics it may not help UK academics. These developments, in addition to the loss of small grain researchers to industry or retirement poses a serious threat to the future UK competitiveness in a cereal agri-biotech field which has an estimated future value of £75B worldwide.

Despite their importance, to date UK small grain cereal genomics has been carried out piecemeal and without overall direction. The result has been some excellent research into individual systems, but no overall community based resource development. This failure to create community resources is a disincentive for young researchers to pursue a career in this field. The main thrust of this proposal is to create resources for existing small grain researchers in the UK. It is hoped that this will also make it attractive for other groups currently not working on cereals to do so.

Contributors

IACR-LARS: Keith Edwards, Donal O’Sullivan

SCRI: Robbie Waugh

Personnel and Equipment requested:

JIC: One band 6 postdoctoral researcher, Three band 7 technicians (3 years). BioRobotics Colony Picker and Arrayer, PhosphorImager, 5 -70°C Freezers, pulse-field gel apparatus.

In coming years, genes controlling a number of complex and simple traits in hexaploid wheat will require identification through positional cloning. These will include genes for disease resistance, yield, flowering time, stress (including a vernalisation requirement and frost tolerance) and metal tolerance, preharvest sprouting, quality (including protein content), grain hardness, non-specific bread making factors and processes which are fundamental to wheat breeding such as chromosome pairing and fertility. Although the genes controlling these traits are likely to be found in the genomes of other cereals, many phenotypes are likely to be specific to wheat.

The genome of hexaploid wheat is five times larger than the maize genome and 25 fold larger than that of rice. Its sheer size makes it less tractable to map-based cloning strategies. In contrast to rice and maize, it seems a remote possibility that the wheat genome will be sequenced either in the public or private sectors in the near future. The cloning of a gene or genes controlling complex characters requires a specific strategy. Such a strategy for wheat has been formulated (presented at AFRC Scientific Opportunities Meeting, Bath 1993; Royal Society Soiree, 1994; Centre for Exploitation of Science and Technology, Plant Biotechnology, 1994). It has been adopted within the USDA Plant Genome Programme (see USDA background document to its PGP). The order of genes in the rice genome is similar to that in the genomes of the Triticeae (wheat and barley), maize and sorghum. The cereals can be described as a single genetic system. Rice genes can be grouped into sets (conserved segments) which describe the gene content of chromosome regions in other cereals. Through an EU programme, extensive cereal comparative mapping is being undertaken in Europe and the study is being extended to include the grasses. The main implication of these studies is that the gene content of eg. a rice YAC can provide an indication of the gene content for the corresponding region in the wheat genome. Thus YAC contigs of the rice genome provide a framework for the structure of the wheat genome on which wheat/barley genomic clones can be placed. Such frameworks exist for the regions corresponding to Ph1 and vernalisation loci in wheat and Rpg1, Rpg4, and Ppd1 in barley. Combining the analysis of rice YACs with that of a wheat BAC library provides an efficient map-based cloning strategy for wheat. However it is important to note that there are some regions for which conservation of gene order is lost between the cereals. The strategy would not be appropriate for cloning genes in such regions.

There are of course other examples of rice being inappropriate as a model for wheat. In particular, data is emerging from phenotypic studies in hexaploid wheat and its diploid progenitors indicating that expression patterns in diploids may be altered when the genomes are combined giving rise to new phenotypes in the polyploid. It will therefore be important to be able to analyse the controlling elements of genes in hexaploid wheat. The analysis of wheat BACs will provide: an indication of whether the gene or genes have become duplicated or deleted with respect to other cereals; the flanking genomic elements controlling gene expression; candidate sequences in the three wheat genomes for generating PCR based assays for deletion analysis; the ability to analyse complex multigene families (such as disease resistance loci) and, of course, the actual wheat genes themselves.

The following resources have already been created for this generic strategy: RFLP and SSR based genetic maps of hexaploid wheat - funded by industry and BBSRC; a Rice IR-20 YAC library with average inserts of approx. 400Kb (Rice BAC Library is being constructed) -funded through BBSRC PAGA; a Rice IR-20 x 63-83 mapping population; an RFLP and AFLP rice map - funded through BBSRC PAGA and now through an EU programme; Two fast-neutron irradiated hexaploid wheat populations - funded through JIC CSG; The Watkins collection of wheat germplasm held at the JIC - maintained through JIC CSG.

An example of the above strategy in practice is the analysis of the Ph1 locus on wheat chromosome 5B controlling homologous/homoeologous pairing. A deletion of the Ph1 locus (ph1b) identified 25 years ago has been used in wheat breeding over the last 20 years. The deletion is approximately 75Mb in size. A rice YAC contig constructed for this region covers more than 3Mb in size. For the first time in 25 years, new deletions of the Ph1 locus have been identified from the fast-neutron irradiated wheat populations using a series of PCR plus/minus screens. Wheat mutants exhibiting higher levels of homoeologous pairing than the original mutant have also been identified. The region containing the gene(s) controlling chromosome pairing has been defined by overlapping deletions to within a 100Kb section. The corresponding region in rice is currently being sequenced. The availability of wheat BACs would enable precise mapping of the breakpoint junctions in this region in the new mutants, in particular in those exhibiting a high level of homoeologous pairing. The BACs would also enable the regions controlling expression of the homoeologous genes in three genomes to be analysed to assess why combining the diploid genomes results in a different phenotype / expression.

Several systems have been developed which allow the cloning and maintenance of large insert fragments in either yeast or bacteria. These systems include YACs (yeast artificial chromosome vectors), PI and BACs (bacterial artificial chromosomes). YACs can maintain fragments greater than 500Kb. The availability of YACs provides a framework for assembling contigs of chromosome regions. However large fragments in YACs are subject to a higher chimerism and other rearrangements than those propagated in BACs. BAC vectors are derived from the E coli fertility plasmid and can reliably maintain up to 350Kb inserts. In species such as rice and Arabidopsis with fewer repetitive sequences, YACs can be used as probes to directly screen BAC libraries to detect the corresponding BACs to the fragment within the YAC. Thus the YACs can provide a framework for assembling BAC contigs. BACs are much easier to use being easier to prepare DNA from, subclone from and hence use in a sequencing programme.

The procedure for making BAC libraries is already available and published. The genome of hexaploid wheat is 15,900Mb in size. The number of clones required to have a 99% probability of containing a particular DNA sequence in a genomic library with an average insert size of 100Kb would be 740,000 or 490,000 if the average were 150kb. Our aim is to construct at least a 600,000 clone library with an average insert size of greater than 125Kb. The library will be constructed of the variety Mercia. If however there is not sufficient seed from this years harvest the variety Chinese Spring will be used.

Previously we have published a protocol for preparing high molecular weight DNA via protoplasting from young wheat leaves. The leaves are manually cut with a razor blade before incubating with cell-wall degrading enzymes. The released protoplasts can be respended in low melting point agarose. Such agarose blocks were used as a source of high molecular weight DNA for YAC cloning. A group making human YAC libraries at ICRF, were able to obtain 600 clones/ug with 500Kb inserts indicating that the wheat DNA was high quality. For the BAC procedure, the protoplasts are encapsulated agarose beads. Megabase DNA has also been isolated from plants by preparing isolated nuclei. The procedure involves grinding fresh young leaves into powder in liquid nitrogen with a mortar and pestle in the presence of Triton X-100. After removing the debris, the nuclei are encapsulated in agarose microbeads. Protoplasting is a much more time consuming procedure. However, both will be tried to assess which provides the better quality DNA for BAC cloning. Isolated nuclei or protoplasts are resuspended in low melting agarose with hot oil and shaken vigorously for a few seconds before pouring onto ice cold buffer to encapsulate in microbeads. The DNA in the microbeads, is released by incubation in a solution containing detergent and proteinase K. To determine the conditions for partial digestion, aliquots of microbeads are incubated with a range of HindIII concentrations. The partial digests are run on pulse field gels to identify the concentration giving the majority of DNA fragments in the 100 – 50Kkb range.

Once optimal conditions are established, the digestion reaction is performed on a large scale. This partially digested DNA is run on pulse field gel and a number of gel slices containing DNA in the 100-500Kb range of fragments are excised. The gel pieces are dialyzed, melted, incubated in the presence of agarase and the DNA quantified. The released DNA is then cloned directly into cut, dephosphorylated BAC vector, transformed into E. coli DH10B by electroporation and selected on LB chloramphenicol plates containing Xgal and IPTG. The recombinant frequency, total number of clones per ligation, average insert size and range are then calculated. If a high recombinant frequency is observed with a reasonable size average but broad range it may make sense to perform a second size selection. This will generally narrow the range and decrease the total number of recombinants in a ligation. However there is little point in performing a second size selection if the clone frequency or average size are too low from the first separation.

For all manipulations of the BAC library we have requested a BioRobotics colony picker (see Equipment section later). Initially it will be used to pick the clones into 384 well microtiter plates containing LB freezing medium. The plates are incubated overnight, two copies made and the copies stored separately at -80 degrees C. A racked -80 freezer will hold 1728, 384 well microtiter plates or 663,552 clones. Three freezers will be required to house copies of the library (one to house the master copy, one to house the working copy and the last one to house a copy elsewhere on the site.

There are several variations of the above protocol, which may be incorporated if necessary (eg electroelution of size selected DNA, precipitating ligations before transformation and other BAC vectors such as pBAC/SACB1 developed at the JIC). If they have clear advantages over the above basic protocol they will be incorporated.

Once the library is constructed, multiple sets of high-density filters will be produced. Using the requested colony picker / arrayer it is possible to grid over 36,000 colonies (double spotted) onto a single 22 x 22cm filter. The entire library will therefore be represented on just over 30 filters. The filters will be hybridised in batches. The possibility of generating microarrays of wheat BACs on glass slides for ease of screening has been explored. However the cost of purifying BAC DNA from minipreps for arraying using Qiagen based technology is prohibitive (£480K in Qiagen columns for the wheat library alone). For positional cloning strategies in the future it is clear that shotgun sequencing of the BAC is going to be the most effective and quickest route to analyse clones. This will require access to an ultra-HTP sequencer. This will be provided through this proposal to the 'Reduction to Practice Laboratory' at the JIC (see ESTs section later).

The Cereal Research Department has recently constructed a millet library in the pBAC/SACB1 vector. Researchers at the Centre have also constructed lotus and potato BAC libraries. An Arabidopsis BAC library is to be constructed by Ian Bancroft at the Centre in pBiBACII. Staff within Keith Edwards’s group have recently constructed a maize BAC library and Robbie Waugh was involved in constructing a barley BAC library and has other libraries under construction at the moment. Staff from Graham Moore’s group have been involved with Sean Mayes and Don MacDonald at the Genetics Dept, Cambridge in the construction of a rice YAC library and are now constructing a rice BAC library.

Triticum tauschii is proposed to be the D genome diploid progenitor and Triticum monococcum the A genome diploid progenitor of hexaploid wheat. A D genome BAC library has been built in Australia and an A genome BAC library is being built on a project funded by the USDA.

While no other large insert genomic libraries are to be built within the proposed project, the participants have already constructed, or are in the process of constructing, cereal libraries, which will be incorporated into the overall resource. Hence, a barley BAC library held at SCRI, a rice YAC library built by Genetics Cambridge and housed at the JIC, a maize YAC library built by Zeneca (KJE) and maize BAC library will be made available for screening to complete the cereal genomic library resource.

Contributors:

Personnel and Equipment:

IACR-LARS: One band 5 postdoctoral fellow, two band 7 technicians; a BioRobotics EST Arrayer, Fluorescent detector for high density arrays, MegaBACE sequencer, ABI9700 dual 384 well block, two upright -80 Freezers, one Biomek 2000HDR 384 pin system, microtitre plate vacuum centrifuge, microcomputer (Gateway G7-450).

SCRI: One band 5 postdoctoral fellow, one band 7 technician; BioRobotics Colony picker and Arrayer, Megabase sequencer, Eppendorf refrigerated centrifuge, ABI9700 dual 384 well PCR machine, two -80 freezers, micro computer, microtitre plate vacuum centrifuge, (Gateway G7-450).

Grain development and germination are the major interests of the UK cereal biology community with areas such as root development and abiotic stress of secondary but significant interest. Our wheat and barley transcriptome programme takes account of these interests. The majority of our efforts (75%) will therefore focus on dissecting the important events of grain development and germination with the remainder split between variously stressed (cold, salt, drought, water logging) and normal root development. The work of UK researchers will greatly benefit from access to an unencumbered catalogue of genes expressed in different tissues during these important developmental processes. Currently a number of agri-biotech companies have invested significantly in obtaining EST collections from a wide range of tissues. However these cDNA libraries have been generated from gross tissue samples. There are only a few groups wordwide with the expertise to accurately identify and stage the wide range of tissues in grain development. SCRI has complementary expertise based upon starch composition and structure (Dr R. Ellis). The scientist employed at SCRI will take advantage of the skills by working at IACR-LARS for a period of six weeks within year one. While increasing the travel costs this will make best use of resources and expertise within the UK system. Using this experience, a detailed catalogue of temporal and spatial patterns of gene expression during grain development will be obtained.

It is important to note that by generating the resources described below, the UK will be able to participate fully in the current ITEC initiative. This initiative is designed to generate and make freely available a significant number of cereal ESTs derived from a range of whole plant, developmental stages. The organising committee of ITEC (Peter Langridge, Univ. of Adelaide, AUSTRALIA, Olin Anderson, USDA-ARS, USA, Mike Gale, John Innes Centre, UNITED KINGDOM, Perry Gustafson, USDA-ARS, USA, Pat McGuire, ITMI, UC Davis, USA and Cal Qualset, ITMI, UC Davis, USA) have suggested that each participating laboratory generates approximately 1,000 ESTs. If this current proposal is funded we would anticipate submitting ~4000 wheat ESTs and ~4000 barley ESTs (~10% of our total) to the ITEC database over the next two years. Such a contribution would enable the UK to have both a significant role within ITEC and enable us to interact with all the worlds key cereal laboratories on equal terms. Being seen as both the source of 8000 ESTs and non-redundant high density arrays (see below) will enable the UK to be at the forefront of future developments in cereal biology.

The initial part of this programme will focus on producing high quality cDNA libraries from (very) specific and agronomically important stages in grain development. Libraries to be made:

There is a close association between final weight of individual grains and the number of endosperm cells produced in both wheat and barley. In wheat, this relationship applies to ancient versus modern varieties grown in controlled conditions, positional differences in grain size within an ear of a given cultivar, effects of shading (decreased assimilate supply) on intact plants and variation in nutrient supply in detached, cultured ears. After fertilisation, endosperm nuclei divide synchronously, without wall formation, in the peripheral cytoplasm of a rapidly expanding central cell of the embryo sac. One important target of cereal biologists is to delay the onset of endosperm cellularisation with a viewpoint of increasing final cell number. To understand the processes occurring at this stage we will sample, by micro dissection, the embryo sac contents at 3-4 days post anthesis to look for messages involved in the early stages of cell wall formation. At the same time we will sample the maternal tissue involved in nuclear degradation and cell elongation.

The formation of cell walls commences ~4 days post anthesis when the endosperm contains ca. 5000 cells (compared with 100 in the embryo). Cell division continues at a slower rate until 16-20 days post anthesis giving a final population of 120-150,000 cells. A group of cells, called the antipodals, about 20-30 in number, are formed in the embryo sac 3 days before anthesis and become highly endopolyploid before degenerating 3-4 days post anthesis. These extremely large cells are thought to provide the protein synthesising machinery to support the free nuclear division stage and to provide the enzymes responsible for the digestion of the surrounding maternal nuclear tissue. Rapid longitudinal growth of the carpel also occurs during the first 6 days after anthesis accompanying the degeneration of the nuclear tissue, except in the region of the ventral groove. The remaining nuclear 'pillar' grows by cell elongation in the region of the degenerating antipodals and contains an activated intercalary meristem associated with the tip of the advancing vascular bundle in the nuclear pillar. This outlines some of the developmental complexities occurring during the very early stages of grain growth with degenerating tissues apparently supplying rapidly dividing and expanding tissues.

During the period 4 to 10 days post anthesis there is extensive synthesis of the cell wall components, for instance, the economically important arabinoxylans in wheat and $ -glucans in barley. These compounds are extremely important in terms of both the nutritional value of the specific crops but also in the industrial processing of the final grain. Knowledge of the genes expressed during this period would be extremely valuable.

1. Whole grain at 6, 8 and 10 days post anthesis

Around 10-14 days post anthesis, it will be important to sample the thick-walled cells of the ventral groove that become the 'modified groove aleurone' and carry out a comparison with those cells from the meristematic 'dorsal' aleurone. The groove aleurone cells act mainly as transfer cells that are responsible for uptake of assimilates into the endosperm during the main period of grain filling. This is the rate-limiting process for grain dry matter accumulation, rather than the movement of assimilate to the grain or its unloading into the apoplastic space of the endosperm cavity. Such tissue should be a rich source of several types of nutrient transporters. Development of the dorsal and lateral aleurone proper of the grain would be of great interest as this tissue stores a different range of reserves from the cells of the so-called starchy endosperm and remains viable to produce hydrolytic enzymes on germination. Messages associated with A-type starch granule formation and storage protein deposition should be abundant in 'starchy endosperm' at this time. The white/mint green maternal pericarp could also be sampled and the embryo would be differentiating tissues of the embryonic axis and the scutellum would be about one half its final size. The embryo is also digesting endosperm cells adjacent to the scutellum, these cells becoming the so-called crushed cell layer which, in barley, are important in restricting uniform hydration during the earliest stages of steeping and subsequent malting. Another sampling of these tissues around 21-28 days post anthesis, at maximum grain fresh weight, would find the maternal pericarp tissues beginning to senesce ie grain changing green to yellow and starch and protein accumulating at maximum rate. B-type starch granules would be present in the starchy endosperm. The embryo would be fully developed and accumulating storage reserves, mainly in the form of triacylglycerols, and producing 7S globulins within the aleurone cells associated with osmoprotection and desiccation tolerance. The same would also probably be true for the aleurone layer. Very little is known about when cell death occurs in the starchy endosperm but in plants grown in CE conditions at 18^oC day/14^oC night (16 hour day length) dry matter stops accumulating at day 40 and grains are dehydrating rapidly.

Libraries to be made: F. Thick-walled cells of the ventral groove, 10-14 days post anthesis.

I. Pericarp at 10, 14, 21 and 28 days

J. Embryo at 10, 14, 21 and 28 days

The initiation of seed desiccation provides a signal to arrest germinative development and to induce embryo (and aleurone ?) dormancy. Without desiccation, a pre-germination and germination program develops in the presence of sufficient grain moisture that may occur in damp seasons. Pre-harvest sprouting is a major agronomic problem. The so-called late embryogenesis-abundant (LEA) proteins are expressed during embryo maturation and are thought to be involved in binding water, maintenance of protein and membrane structure and ion sequestration. Genes encoding proteins involved in desiccation tolerance should be expressed in both embryo and aleurone at this stage of grain development. The plant hormone, abscisic acid (ABA) is abundant prior to seed desiccation and is thought to be a possible triggering molecule of the desiccation process. Transcripts relating to its production and action should also be abundant at this stage of development, as well as those involved in dormancy and quiescence. Samples of embryo and aleurone taken at the start of desiccation (30-35 days post anthesis, 45-50% moisture content) and again at , say, 40-45 dpa (30-40% moisture content) should yield transcripts involved in these processes. Under certain circumstances, before embryo dormancy has been induced or where dormancy induction is 'weak' and the period of after-ripening short, a pre-maturity germination program with all the detrimental consequences for wheat flour or barley malting quality may be induced.

Libraries to be made: L. Embryo tissue at 30 and 40 days post anthesis

M. Aleurone tissue at 30 and 40 days post anthesis

Following a short period of after ripening, or 'in the ear' in cool, damp seasons, a battery of hydrolytic enzymes are induced in the aleurone following imbibition of the grain and the transport of a gibberellin (GA) hormonal signal from the embryo. Hydrolytic enzymes are also produced initially, and to a much lower level, in the scutellar epithelium of the embryo, although in this case they may not be dependent on a hormonal signal. As well as the degradation of cell walls, starch and storage proteins in the endosperm, the triacylglycerols of the embryo and aleurone are also degraded following germination. Samples of embryo and aleurone of grains germinated for 2-3 days would be an abundant source of transcripts associated with hydrolytic enzyme production in addition to those involved in GA production and action. The uniformity of grain hydration and the extent of cell wall degradation are important aspects of barley malting quality.

Libraries to be made: N. 1 and 2 days germinated embryo tissue

O. 1 and 2 days germinated aleurone tissue

In addition to the various libraries described above the consortium will generate a further 6 libraries as follows:

5. Physiologically normal roots: As a direct comparison to 4. A single cDNA library will be prepared from root material obtained from whole seedlings subjected to normal physiological conditions.
6. Normal whole 21-day-old seedling: A single cDNA library will be prepared from whole seedling material grown under normal physiological conditions.

In summary, up to 35 different cDNA libraries will be constructed for both wheat and barley for the grain development and germination part of the programme. Another 6 will be constructed for the ‘other stages’ indicated. Initially, the libraries will be made in the reverse order, ie. grain germination then late grain development libraries first. From each of these libraries we will single pass sequence ca. 1,000 clones (41,000 in total). Both the original libraries (as colony filters, phage or plasmid stocks) and the information generated from this stage of the programme will be made available to UK researchers through password protected WEB access. Ultimately these will also be incorporated into both UK CropNet and ITEC databases. Insert sequences from the unique clones within this collection will be used to prepare high density arrays (see B. HIGH DENSITY ARRAYS below).

1. Generation and referencing of cDNA libraries.

Although the development of commercial kits has made the generation of cDNA libraries relatively easy, the generation of high quality full-length libraries is still relatively difficult, especially when the mRNA substrate is in short supply. Several groups at IACR-LARS and SCRI have experience in generating cDNA libraries from the developing grain of both wheat, maize and barley. For instance, the Crop Genetics Group at IACR-LARS has recently generated several libraries from the various stages in maize kernel development (7, 10 and 17 day post anthesis), and as part of an Industrial CASE studentship with Wolfgang Schuch at Zeneca Plant Sciences, the group is also developing similar libraries for developing grain in wheat (variety Savannah). The SCRI group has similarly made libraries from 28-day post anthesis and 2-day malted barley. However, a major problem with the generation of the high quality libraries required for this work will be the production of sufficient mRNA for library construction. During the later stages of grain development it will not be a problem either generating sufficient material or in separating the various parts of the grain required. In the initial stages, for instance, immediately after fertilisation, it will be extremely difficult to dissect the various components of the grain and therefore whole grain libraries will be produced.

For this public resource the cDNA libraries will be generated in the HybriZAP 2.1 two-hybrid system from Stratagene. Various groups at IACR-LARS have experience of this system. The HybriZAP2.1 vector is both suitable for directional cloning, mass excision, single pass sequencing and for inclusion in the yeast two hybrid system. This later fact is important as in the future (via further initiatives) we intend develop our array technology to include a two-hybrid array resource.

The most difficult part in the construction of the grain cDNA libraries will be the availability (and identification) of sufficient tissue at exactly the right stage of development. The use of PCR based or fully normalised libraries has been discounted as both ours and others experience of these technologies is mixed and suggest that the resulting material is biased and often not full length. Therefore we will need to isolate a significant amount of material for subsequent mRNA extraction. For the earliest stages of grain development this work will require a considerable amount of micro dissection work. This will be labour intensive and we have therefore requested sufficient technical support at the two active centres to complete this work. The construction of the 41-cDNA libraries will require a considerable amount of care and experience. The level of skill required will be beyond a post-doc with less than 10 years research experience. To successfully complete this stage of the work, at IACR-LARS and SCRI we have requested the resources of band 5 researchers. Although more expensive than a band 6 position the presence of experienced band 5 researchers should ensure that rapid progress is made at this most important stage of the work. Initially IACR-LARS and SCRI appointments and the IACR-LARS technical support staff will work under the close supervision of Professor Shewry and Dr Lenton, so that they can be fully trained in the art of staging the various tissues.

As little as 50 ng of mRNA (50 ug total RNA) for each tissue source will be required. However, the requirement to stage the various tissues to a very precise level will mean that considerably more material will need to be grown than utilised. In the case of the early stages in grain development we estimate that this will require up to 5,000 developing grains from which we will micro dissect approximately 500 to 1,000. Glasshouse resources and CE rooms at SCRI and IACR-LARS have therefore been requested to grow a considerable number of plants. Once produced the mRNA will be converted to double strand cDNA using the commercial kit produced by Stratagene. The cDNA will be produced so as to contain both an EcoRI linker and an XhoI linker such that the material can be directionally cloned into the HybriZAP 2.1 vector.

To reduce the level of redundancy within each library, before all of the material is cloned a small aliquot will be cloned and used to produce a small number (~100) of clones. Sequence analysis of these test clones will provide information on the redundancy of the RNA population. For instance, at certain stages of development (ie the later stages of endosperm development) we expect that a small number of sequences will make up the vast majority of clones. In these cases we will utilise those redundant clones which represent over 5% of the sequences to deplete the remaining uncloned cDNA sample. This will be achieved by hybridising the denatured uncloned cDNA to an excess amount of the redundant cDNA inserts, which have been bound to a solid support. Non hybridising cDNA will be converted back to double strand and cloned as before in the HybriZAP 2.1 vector. This two step process will reduce the speed that the libraries come on stream, but it will increase the efficiency of the sequencing step by partially eliminating redundant sequences. It will also provide preliminary information on the variation in ESTs during the different stages of grain development. Sequences generated during this step would be placed into the public domain under the ITEC initiative (see below). The cDNA libraries produced will constitute the first set of resources generated by the programme.

Once generated the lambda based clones will be converted into plasmids via mass excision. ca. 6,000 clones (16 x 384 well plates) from each stage specific library (ca. 250,000 in total) will be picked. Colonies will be picked by the colony picking robots requested in this application by the JIC and SCRI. It will therefore be necessary for staff to commute between IACR-LARS and the JIC to achieve this objective. Whilst this will increase the travel costs of the programme it will make full and best use of the requested resources and will therefore reduce the overall funds requested.

Two -80^oC freezers at each center (IACR-LARS and SCRI) have been requested to store the number of plates required. Each plate will be replicated before storage at -80^oC. Once produced the working stock plates will be used as follows:

i. The plates will be used to prepare high density COLONY filters using IACR-LARS's existing Biomek 2000 workstation and the colony picking / arraying robotic system requested by SCRI. At IACR-LARS we routinely prepare 12 x 8cm high-density filters for YAC and plasmids containing 6000 clones (16 x 384 well plates, 4 x 4 array) using the Biomek 2000. These filters will constitute the second set of resources generated by the programme. They will be used as representatives of the original libraries and will be provided to any UK researcher who wishes to use them. A single 384 Biomek replicator has been requested by IACR-LARS to augment our existing Biomek 2000 system.

ii. The plates will be used to prepare plasmid DNA for single pass sequencing. Plasmid DNA will be prepared using SCRI and IACR-LARS ‘in house’ protocols. Both utilise a Biomek 2000 robot in combination with filter-bottom microtitre plates to prepare sufficiently pure plasmid DNA for up to 50 single pass sequences (using BIG Dye chemistry). Based upon our existing procedures we will be able to prepare approximately 1,000 plasmids per working week (41 weeks in total). One fifth of the plasmid template will be used for the sequence analysis with the remaining material being stored for use in insert preparation for the generating of the insert arrays. We estimate that the cost of preparing each plasmid will be approximately £1.20 - £1.50 (£55,350 in total at each center).

Plasmid prepared from the previous step will be compressed back into 384 well format. Material will be kept in this format throughout the amplification stage and the preparation for gel loading stage. Based upon our existing protocol, a Biomek 2000 will be used to set up 10ul BIG DYE cycle sequencing reactions in the 384 well format. Following amplification the samples will be precipitated and prepared for gel loading in the micro plates. We estimate that the cost of sequencing each plasmid will be approximately £1.20 - £1.50 (£55,350 in total at each center).

At IACR-LARS, based upon existing ABI 377 resources we will be able to sequence no more than 250 clones per week. At SCRI, all available DNA sequencers are currently running at full capacity. Clearly for this programme this is inadequate.

There are three options:

1. Carry out our sequence analysis over a longer period of time.
2. Contract out the sequencing
3. Request as part of this proposal equipment to increase our sequencing capacity.

It is our belief that option 1 for IACR-LARS is unacceptable as it would both delay the construction of the high-density arrays by up to 12 months and increase the labour costs of the programme. For SCRI the program objectives would be unachievable. Option 2 is impractical for both institutes as we know of no contract sequencing company that can either deliver within the timescale requested or at a cost which is reasonable. On this matter the panel should be aware that KJE currently ‘contracts out’ as part of his microsatellite sequencing requirements and so he is aware of the current situation. Both IACR-LARS and SCRI have therefore chosen option 3.

Following discussions with the Stanford University Genome centre, IACR-LARS and SCRI has requested the resources to purchase a MegaBACE 96 well capillary sequencer (Amersham) at a cost of £137,000 (+VAT).

The handling and assessment of the information created by this project will require an integrated sample tracking, databasing and analysis pipeline. Details of the system we propose to adopt are given in section 4. BIOINFORMATICS.

Despite the relatively large number of reviews devoted to the generation and utilisation of high density arrays, it is still relatively difficult to make, use and analyse the results from high density ESTs array experiments. We do not wish to develop any array-based technology, however, we do wish to utilise commercial packages to generate the arrays as efficiently as possible. We have therefore based our proposal on both the available hardware and on a site visit to both the Stanford Genome centre including discussions with Shauna Sommerville and the Hinxton genome centre.

Initial development of array technology and procedures will be performed at IACR-LARS only. Arrays will be made using the requested BioRobotics array device. Using this tool, arrays of 10,000 unique ESTs will be generated on activated microscope slides. The BioRobotics instruments have been chosen because they have a proven track record. The system includes a purpose built fluorescent scanner and sample tracking and analysis software.

For both wheat and barley we will array EST inserts and not EST clones. Using inserts instead of clones will allow us to increase hybridisation sensitivity and will therefore allow us to detect the expression of clones, which would otherwise not be possible with clones. However, this approach does require that we amplify the inserts from the selected 10,000 clones. Clones for arraying will be chosen by the following criteria:

a. Generation of the substrate for the high density arrays

Using the information from single pass sequencing and the results of the database searches, plasmids prepared from 10,000 unique clones (the Unigene set) will be re-arrayed into new 384 well plates and diluted 1 in 100 with 1x TE buffer. Based upon existing work we estimate that the plasmid prepared by this procedure will be sufficient for over 1,000 PCR amplification or ~50,000 arrays.

Approximately 1 ng of plasmid DNA will be used to amplify the EST insert using amino modified universal forward and reverse primers and standard PCR conditions. PCR products will be purified using a proprietary PCR purification kit (eg from Qiagen) and quantified by gel electrophoresis. Products from one reaction will be re-suspended in 10 ul of array buffer and used in the arraying procedure. Arrays on activated/coated microscope slides will then be prepared as per the manufacture instructions. The PCR’d EST insert sequences from barley will be ready for depositing on glass slides approximately one year after the IACR-LARS. It is anticipated that by this time all of the technical difficulties in performing these types of assay will be ironed out and expression analyses can proceed rapidly.

We estimate that with the requested resources we will be able to make several thousand wheat and barley EST arrays. These arrays will be brought to the attention of the research community through the existing IACR-LARS/SCRI web pages and by 'advertising' through the UK CropNet and through the BBSRC research network.

b. Hybridisation, data capture and analysis

The procedures for hybridising labeled cDNA to EST arrays is still in its infancy, however, several protocols have been published which describe in detail the probe preparation and hybridisation procedure. For instance the group of Pat Brown has published a detailed protocol on: http://cmgm.standford.edu/pbrown/protocols/index.html

Initially we will follow the available protocols. Our first experiments will consist of both a single and a limited multiplex of known EST probes at varying concentrations. Once we have confirmed that we can both detect true positives and estimate the sensitivity of our system, we will progress to using complex RNA populations derived from specific stages in grain development. Although most current methods rely on cy3 and cy5 dye technology it our belief that by the time that this program would be operational, further dye technology will be available. This being the case we will also examine the issue of multiplexing the various probes. To create the initial expression database, we will undertake to screen the arrays, with probe prepared from each of the 41 stages used to make the original cDNA libraries. In itself this information will therefore represent a considerable resource to the cereal community.

Once hybridisation has taken place the slides will be analysed using the requested fluorescent scanner. This scanner has the ability to analysis multiple slides over a 30-minute period. Results from the scanner will be in a TIFF file format. Such a format can easily be transferred between either the consortia sites and any END USER site. However, under normal circumstances we expect to carry out at least a preliminary analysis of the data before submission to either the END USER or the WEB page. To date, we have not reached a conclusion on which software package will be used for the array analysis, several packages ranging from the free (for instance the Brown/Botstein software) to the very expensive (for instance GeneSpring by Silicon Genetics) are now available. However, again it is our belief that at the time of carrying out this step, the situation with regards to the analytical software will be clarified. It should be noted that the quotation from BioRobotics does include analytical software. Minimal requirements for the software eventually chosen will include both quantification of the TIFF images and a database function which will enable the user to perform complex data analysis such as highlighting all the genes that have a specific pattern of expression over a number of experiments. All the data generated via the chosen software will be made available on the IACR-LARS program web page.

With regards to the hardware requirements for the array analysis, it will be essential that each of the persons employed have full time access to a PC with web access. A Gateway G7-450 has been requested at each centre. These will be used throughout the program for processing the EST information and interpreting the results from the highly parallel expression studies.

Evers (1970) Ann. Bot. 34, 547-555; Briarty et al., (1979) Ann. Bot. 44, 641-658.;

Bosnes et al., (1992) Plant J. 2, 661-674.;Olsen et al., (1992) Seed Science Research 2, 117-131.; Smart and O'Brien (1983) Protoplasma 114, 1-13. Reviewed in Bennet et al (1975) Phil. Trans. R.Soc. B. 272, 199-227.

Contributors: Keith Edwards, David Edwards and Michael Holdsworth

Personnel and Equipment: One band 6 researcher and One band 7 technician at IACR-LARS

Background:

It is vital that UK researchers are able to determine the function of any cereal gene at will, as is becoming the case for Arabidopsis. Ideally as part of this programme, we would like include a wheat and barley based "gene machine". Unfortunately, with current technology, only maize has a well-developed gene machine capacity. By the time that such a system is developed for wheat (if it can be developed), it is likely that the function of the majority of cereal genes will have been determined via gene machines in other species. It is therefore important that UK researchers have access to the "alternative technology" right now.

Through BBSRC and EU funding, a working maize gene machine and a working low-density maize Mutator array is available at IACR-LARS (BBSRC Business Newsletter April edition). The IACR-LARS Mutator grid has already produced more than 150 knock out plants for various members of the UK and EU research community. This existing facility therefore provides the best opportunity to identify gene knock outs for wheat or barley ESTs identified as having an interesting expression profile. For example, IACR-LARS, in collaboration with Wolfgang Schuch at Zeneca, is investigating the development of wheat and maize endosperm using maize gene knock outs for wheat derived sequences.

The programme will utilise existing IACR-LARS Mutator plant collection which consists of ~7,500 individual plants. Additionally, during 1999 another 2,500 plants will be added, making a total of 10,000 plants. In the case of the existing material, heterozygote seed (via a B73 cross) and leaf tissue for each plant exists. The seed for this material is stored in the IACR-LARS seed store under cool and dry conditions. The seed should remain viable for approximately 10 years. In the case of the further 2,500 plants, these will be grown at IACR-LARS during the summer of 1999, using the facilities and protocols previously used for the 7,500 plants. Each Mutator active plant contains at least 30 randomised Mu insertions, then 10,000 plants should contain 300,000 insertion events or 6 fold coverage, based upon an estimated gene number of 50,000. A random, sequencing programme has confirmed that the Mu elements are randomly inserted within gene rich regions of the genome.

It has been shown that arrays of PCR generated Mu-tagged fragments when probed with cereal ESTs (both maize and oat ESTs have been employed so far) identify unique insertion events within the maize or oat homologs. The BioRobotics array device requested above, will be used to extend the scope of the existing work and generate multiple copies of the Mu-arrays for screening by any researcher who so wishes to use the facility (as described in the community access section).

Due to limited existing resources the IACR-LARS existing Mu-arrays consist of Mu-tagged fragments derived from pools of plants. Whilst this is convenient for in house purposes, to make the procedure amenable for all users and to allow a single screen to identify specific gene knock outs we intend generate 10,000 spot arrays of single Mu-tagged plants.

1. Preparation of genomic DNA from all of IACR-LARS Mu-plant collection

The first task in preparing the high density arrays will be the preparation of genomic DNA from each one of the 10,000 plants in our collection. Based upon our existing protocol genomic DNA (1-2 ug per sample) will be extracted using a mini-preparation procedure developed in our lab. This procedure will yield sufficiently pure genomic DNA for the next step. Based upon our experience we would expect to be able to generate DNA from at least 50 samples per day. Clearly, this part of the programme is labour intensive and would be best undertaken by both the researcher and a technician. Once produced the genomic DNA at a concentration of around 50 ng per ul will be stored in micro titre plates at -70^oC.

2. Preparation of Mu-tagged fragments

Mu-tagged fragments for each of the DNA samples will be prepared as follows, using our labs Biomek 2000 robot: Each genomic DNA (50 ng) sample will be digested separately with either HpaII or MseI (neither of these enzymes have restriction sites within the Mu inverted repeat). Following digestion, the appropriate vectorette linker is ligated to the sample. Following ligation the sample is amplified with both the appropriate vectorette primer (5'GAATCGTAACCGTTCGTACGAGAA3') and a biotinylated Mutator specific primer (5'BIOTINGCCTCCATTTCGTCGAATCC3') for 20 cycles. Following amplification the amplified products are captured using the Streptaviden magnetic beads. Following capture, a small amount of the beads will be used as a template for a further 35 cycles of amplification using both the original vectorette primer and an amino modified nested Mu primer (5'CAGAATTCCATAATGGCAATTATCTC3'). Following this round of amplification, the 2 PCR products derived from the same DNA sample will be combined and purified via ethanol precipitation. Finally the samples will be resuspended in 10ul ready for production of high density array. It is important to note that we have already shown that this procedure is capable of producing the desired fragments in a controlled and rapid manner. To accomplish this task, we will need to carry out 20,000 digestions, ligations and PCR amplifications. Such a task will require a considerable amount of both restriction enzyme, vectorette adapter, ligase and Taq DNA polymerase. This is reflected in the higher that normal consumable costs for this work in the first year. Approximately 1 ug of PCR product can be produced for each amplification. Because each sample will be derived from two amplifications ~2ug of PCR product will be produced for each of the original 10,000 plants. This amount of material should be sufficient for several hundred high density arrays. Using the two enzymes as described, Mu-tagged fragments will be generated for the majority of insertions present within the sample.

3. Preparation of the High Density Arrays

The high density arrays will be produced using the requested array device. The initial experiments will be designed to address how little material can be spotted onto the array and still obtain a significant signal. Individual arrays will be produced containing an amount of DNA equivalent to 1, 2, 5 and 10 ul of the original PCR product. Before spotted the DNA onto the arrays, the appropriate amount of PCR product will be arrayed onto the activated microscope slides.

The 10,000 DNA samples will be spotted onto single microscope slides. Once produced, the test arrays will be probed with three cDNA sequences which IACR-LARS other Mu based programmes have already shown are contained within the grid. These cDNAs are: 1. Glutamine Synthetases, 2. A mannosidase II related sequence and 3. AMP deaminase. In the case of 1 and 3 several plants in the grid are known to contain Mu inserted into the gene whereas in the case of 2, only a single plant is believed to contain Mu inserted into the gene. These sequences will therefore act as positive controls and will allow the determination both the optimum amount of DNA to spot onto the arrays, the reproducibility of the procedure and the optimum hybridisation and washing conditions to increase the longevity of the arrays. Once the optimum conditions are determined for both making and using the arrays, they will be further test as below.

4. Testing the arrays

The arrays will be tested in one of three ways:

Firstly, known and unknown maize cDNA clones will be used to screen the array. Here screening will both be carried out using individual cDNAs and mixtures of cDNA under a variety of hybridisation conditions.

Secondly, known and unknown wheat and barley sequences (derived from the EST programme) will be used to screen the array in a similar way as described above. Here the hybridisation conditions will have to be determined on a case by case bases.

Thirdly, arrays will be supplied to UK academics along with a full screening protocol (including suitable control probes). Once the hybridisation procedure has been completed the arrays will be returned to IACR-LARS and scanned using the fluorescent scanner as described for the EST arrays. The various collaborators will then be informed of the results.

In all cases, plants shown to contain specific Mu insertions will be processed via our standard procedure. This involves each plant being first crossed with a Mutator suppressor line. This line has been shown to switch off the Mutator transposon and so will eliminate further transposon events, which may confuse the further analysis of the material. The plants will then be backcrossed to F2 (a maize line which cycles in 9 weeks and which is suitable for growing outside in the UK) for four generations (to remove the unlinked Mu elements) before being selfed to uncover any phenotype associated with the insertion. It will be a condition of screening the grid that all interesting data is placed as soon as possible on the existing BBSRC sponsored Mutator Web page: (http://www.maize.bbsrc.ac.uk).

Once the quality of the arrays is confirmed, they will be advertised through the IACR-LARS web site and through the various UK and worldwide web pages (ITMI, Maizedb etc). If there is sufficient interest in the Mu array we will also seek the co-operation of a commercial organise to explore the possibility of marketing the arrays.

Mutator References:1. Database of Expressed Sequence Tags at NCBI and unpublished data. 2. Newman et al (1994). Plant Physiol. 106:1241-1255. 3. Meeley and Briggs (1995) Maize Genetics Newsletter. 69: 67-82. 4. Bensen et al (1995). Plant Cell. 7:75-84. 5. Mena et al (1996). Science. 274:1537-1540. 6. Edwards, K., Johnstone, C. and Thompson, C. (1991). Nucleic Acids Research. 19, no 6, 1349.7. Ogilvie, DJ., Butler, R, Riley, JH., and Markham, AF. (1991). Cytogenetics and Cell Genetics. 58, no 3-4, 2151-2159. 8. Macas, J., Nouzova, M. And Galbraith, DW. (1998). BioTechniques. 25, no 1, 106-110.

Contributors: Robbie Waugh, Gordon Machray, Bill Thomas, Roger Ellis

Rapid progress in genome sequencing projects has propelled a shift in genome analysis from structural genomics to functional genomics (ie. the genome-driven systematic study of gene function). Rapid methods for ascertaining gene function by targeted gene inactivation are therefore highly desirable. One large-scale approach to germline gene-inactivation is to induce random mutations in a population of plants. Historically this has been achieved by chemical or irradiation mutagenesis followed by screening for visible phenotypes in M2 populations. Barley has been particularly amenable to this approach and currently boasts one of the largest available phenotypically characterised mutant collections (SCRI has carried out extensive work on barley starch mutants). However, with these mutants, isolation of the gene responsible for a given phenotype relies solely on a 'forward genetics' approach (ie mapping followed by positional cloning). In contrast, biological transposon-based mutagenesis tools are particularly useful in some plant species (such as maize (see contribution 3A) and Arabidopsis) where they have facilitated the development of 'reverse genetics' approaches. Insertion causes knock-outs, which can be easily characterised using the known DNA sequence of the transposon to identify the mutated gene and study the phenotype of the null mutation. However, for wheat and barley, transposon-based mutagenesis has not yet been successfully established and therefore remains 'pre-technology'.

The potential of chemically or physically mutagenised populations to facilitate reverse genetics has increased recently with the development of sensitive methods for the detection of 'aberrant' DNA fragments, especially in complex mixtures. In particular, strategies deploying chemical mutagens, which induce random short deletions in DNA ("deletogens") have been identified and these allow a simple PCR-screen to detect deletions in a specified target region (Yandell et al 1994). Using this general approach over 100 deletions in different targets in C. elegans (Liu et al, High throughput isolation of deletion mutants of C. elegans, in preparation) have recently been identified. Chemical mutagenesis has a number of inherent attractions, such as the ability to manipulate the mutagen and its dose (influencing mutation type and frequency) and the ability to scale up or down easily. Importantly, chemical mutagens also generate an allelic series at any target locus, revealing both change of function and knock-out mutations which can be very valuable when attempting to assign function to a given gene. For example, an allelic series of point mutations was recently exploited in barley (Buschges et al 1997) to confirm that an open reading frame, with no clues as to function from database searches, was the Mlo powdery mildew disease resistance gene. Given these attractions (and the lack of a biological mutagenesis system), we propose to generate a chemically mutagenised barley population to facilitate reverse genetics in this crop.

Deletions rather than mutations per se are the goal of this approach (although point mutations may ultimately prove just as useful if 'SNP detection' in pooled samples can be established). Thus the first objective will be to identify an optimal 'delotogen' chemical for barley and its most effective dose. Deletogens will be selected initially on the basis of information arising from mutation spectrum analyses (MSA) on animals and to a lesser extent in plants. For example, in C. elegans, a collection of mutagens have been identified which generate a high frequency of deletions of 100 - 5000bp (Carl D. Johnson, Axys Pharmaceuticals, pers. comm.). These include ethyl methyl sulphonate (EMS), ethylnitrosourea (ENU), trimethylpsoralen followed by UV irradiation and diepoxyoctane. In mice and humans MSA has shown that epoxybutane and in particular diepoxybutane (DEB) induces point mutations and large scale deletions of hundreds up to thousands of base pairs in equal frequency (Cochrane and Skopek 1994, Steen et al., 1997a,b). DEB is also an effective "deletogen" in rice and has recently been used successfully to construct mutant rice populations in IRRI (Hei Leung, IRRI, pers. comm.)

1. Choice of "Deletogen"

a. Establishment of Optimal Mutagenesis Conditions

We will perform pilot studies on three of the above chemicals, EMS, ENU, and DEB, to determine the optimal "deletogen". For each chemical the approach will be similar. Therefore, the following description outlines only the details of choosing DEB. DEB is a bi-functional alkylating agent that induces DNA interstrand cross-links between adjacent Guanine bases. A modified but similar approach will be taken if either of the other chemicals proves more appropriate.

Briefly, weighed, sterilised seeds will be incubated in various concentrations of DEB with shaking for 12 hours, washed extensively and planted (0.004% - 0.006% DEB is effective for rice in IRRI). We will use the spring barley cultivar Optic that has excellent malting quality and is sensitive to mildew (Erisyphe graminis). % kill (survival) will be measured directly in the "soil" environment by sowing treated seeds plus controls and comparing the total number of viable plants at the end. Sowing directly into soil gives the natural survival rate without counting the very weak plants. The M0 seeds will give rise too much weaker M1 seedlings than the wild type because mutagens generally damage many things in addition to nucleic acids, such as cytoplasmic proteins. These are not heritable but show up in early stages of the M1. The appearance of chlorophyll deficient sectors at low frequency among the M1 plants (c. 1 in 100 - 1000 plants in Arabidopsis) will be used as initial evidence that the mutagenesis has been successful. A portion of the most promising test mutagenesis will be grown through to maturity to check for seed setting. In rice, with a 70% kill, ~60% of the M1 generation set sufficient seed to proceed with developing and screening the M2 population. Approximately 100g of seed (1000-2500 seed) will be used for each pilot experiment. These studies to optimise mutagenesis conditions will be carried out in the glasshouse. The optimal mutagen concentration and treatment will be estimated from a compromise between the % kill, the appearance of visible mutant sectors, and seed setting in the M1 (these pilot studies will use ~20,000 plants).

b. Effective Mutation Frequency (EMF) and Mutation Spectrum Analysis (MSA)

i). EMF: From the treatments deemed most effective by the above criteria, three new M1 populations (ca: 5000 M0 seed) will be developed and grown to maturity. The number of M0 seed will be estimated to yield ca. 1000 seed setting M1 plants. Seed from the different treatments will be bulk harvested separately. M2 populations (3 populations, c. 5000 - 10000 seed each) will be grown in the glasshouse and scored for a variety of naked eye polymorphisms (NEPs - chlorophyll deficient, dwarf, leafy, lesion mimic etc.). Both dominant (eg 'hooded') and recessive (eg 'dwarf') mutations should be visible in the M2. In Arabidopsis, a successful EMS mutagenesis will typically show 1-3% of M1 plants segregating chlorophyll deficient plants in their M2 progeny and a similar frequency in barley will be expected. To estimate the EMF we will first establish the frequency of obvious phenotypic mutants. The test M2 populations will also be examined closely to identify potential lesion mimic mutants. If apparent, these will be isolated and be screened for resistance to powdery mildew (E. graminis). If not the entire M2 may be screened after artificial inoculation. The race, non-specific mlo resistance gene is recessive and mutations in this gene will be expressed in the M2. Resistant genotypes will be isolated, fungicide treated and grown to maturity. a priori we expect a frequency of 1:1000 - 1:10000 mlo plants in the M2. At this stage of the work we will use 3 test M1 populations of ~5,000 MO seed and 3 test populations of ~5,000 M1 seed (30,000 plants).

ii). MSA As the region encompassing mlo has been entirely sequenced (Panstruga et al, 1998), mutation spectrum analysis at this locus will be determined by a combination of PCR from the resistant genotypes, followed by restriction digestion and DNA sequencing of the cloned PCR products. This will quantify the frequency of point mutations vs. deletions and the average size of the latter and allow identification of the optimal 'deletogen'.

If mlo mutations are not clearly recovered, the average frequency of detectable DNA mutations a small random selection of 10 M2 plants from the most promising treatment will be genotyped using a minimum of 40 AFLP combinations and compared to the wild type. An average single AFLP primer combination in barley using EcoRI or Pst1 and MseI will survey approximately 1200nt for point mutations (ave. = 80 amplified fragments, 6bp restriction site + 2 selective nucleotides one side, 4bp restriction site +3 selective nucleotides on the other) and ca. 16,000bp DNA sequence for small deletions (80 fragments x 200nt average length). 40 combinations on 10 genotypes will therefore survey c. 480,000 base pairs for point mutations and 6.4 x 10⁶ bp for deletions (ie >1% of the barley genome). Using this latter approach it will be impossible to estimate the relative frequency of deletions vs. point mutations and a judgment will have to be made on the most appropriate mutagen.

2. Large scale Mutagenesis

Having established optimal conditions, a full-scale mutagenesis will be performed. After mutagenesis and prior to sowing the large M2 population, the percentage survival of the M2 will be checked on a small scale in the glasshouse to confirm that the experimental conditions mimic the 'pilot' studies. Our objective is to obtain a final M2, which is approximately 20 times the size of the M1. n M0 seed (number determined empirically from the pilot studies) will be sown in Plantpak seed trays with 6x4 cells in a tray. Each cell is approximately 125 cm³ and under glasshouse conditions will produce a plant with 1-2 spikes, each spike setting 10-20 seed. We have a machine to mechanically sow the Plantpaks and the resources to grow the trays in a glasshouse out of season under optimal conditions so that M2 plants can be sown in the spring. We are aiming for approximately 10,000 surviving, seed setting M1 plants. A single spike will be harvested from each. Each spike will then be individually threshed and automatically loaded into a cell of a magazine for a seed drill using the specialised threshing equipment at SCRI. We have a Wintersteiger Precision Spaced Planter seed drill which can be programmed to automatically sow the seed from one cell of the magazine in one row. The length and spacing within the row is programmed into the drill, which is also programmed to exhaust seed from the sowing mechanism at the end of each row and the drill is then primed for the next row. We propose to use the capabilities of this drill to sow a 96 x 96 matrix, each cell of which will be a short row containing the progeny of a single M1 plant - ie about 20 M2 seed. The rows will be protected from foliar pathogens by fungicide applications. Once the majority of the plants are growing vigorously, leaf samples will be collected from all plants within a row, identified with their X/Y co-ordinates and freeze-dried for pooled DNA extraction. At maturity, each cell within the matrix will be harvested as a bundle of plants for later threshing in the laboratory, after which the seed will be identified by its X/Y co-ordinates and stored. A range of obvious NEPs will be scored within the matrix at various times over the growing period. Scoring the NEPs will rely heavily on the expertise of Drs. Bill Thomas and Roger Ellis (over 40 years experience of barley breeding and genetics).

While this is a large-scale operation the sowing and threshing is much less than that of a plant breeding programme and is well within SCRIs capabilities given sufficient casual labour for the routine operations such as harvesting and threshing. The 96 x 96 M2 matrix provides 9216 twenty plant cells or c. 400,000 mutagenised genomes. Assuming a conservative mutation frequency of c.30 - 50 short deletions (incorporating coding sequences) per M2, this represents c. 300,000 - 500,000 independent events. Assuming a gene number of 50,000 this is 6 - 10 fold genome coverage. Redundancy will be more accurately characterised after determination of the EMF and MSA.

3. Utilisation of the Deletion Grid

a. DNA Isolation

DNA will be isolated from the pooled leaf material from each of the cells in the array (9216 DNA preps) by standard procedures. After extraction, aliquots of the pooled DNA will be pooled again to yield superpools for analysis (size will have to be determined empirically). The minimum number of superpools would represent the rows and columns of the 96 x 96 format (ie 192 DNA super-pools forming a 2-D grid). At the other end of the spectrum, c. 1000 superpools containing 10 cells could also be handled. Superpools this size should be appropriate for all three mutation detection strategies outlined below (dHPLC has been used to detect SNPs in pools this size (see below)). The super-pools will be used for mutation detection PCR.

Isolation of 10,000 plant DNA samples at c. 48 samples per person per day it will take c. 1 year to isolate the DNA from the entire grid. The DNA concentration of all samples will be estimated by running on a gel next to Lambda concentration standards. Isolated DNA will be stored in 96 x 96 well plates at a standard concentration c. 200 ng/ul (in 2 copies) and stored at -80oC (in separate freezers). Pooled DNA samples will be constructed from approximately equal quantities (100-200ng) of DNA from each cell in the pool. This will allow c. 400 - 800 PCR screens using 25ng DNA template per reaction (see below) each time a new set of 192 super pools is constructed (40 - 80 for 10 cell pools).

The individual cell DNAs and super-pooled M2 DNA samples will thus be the major deliverable of this project and the primary access point to gene 'function search'. The second major deliverable will be the M2 seed stocks from the individual cells. A database will be constructed which will hold housekeeping and phenotypic / genotypic information from the mutation grid (see 4. BIOININFORMATICS).

b. Mutation detection.

Exploitation of the grid for reverse genetics will require that we are able to detect mutations using a sequence based approach. We plan to explore 3 different approaches for mutation detection. We will also keep abreast of new developments - particularly in the field of SNP detection - as other approaches will likely emerge as more appropriate / sensitive over the duration of the project. In each case we will use mutations / deletions in mlo as our test target sequence (either known mutations or new mutations detected in the grid).

Briefly, the approaches will be:

1. Preferential PCR amplification: A relatively simple nested-PCR approach will be tested for detecting deletions in a given gene (after Westlund et al, 1999) from superpooled DNA samples. PCR amplification conditions are designed to preferentially amplify shorter than wild type DNA fragments, which are the product of a deletion between the primer pairs. Mutant and wild type products can be identified by standard procedures.

2. Denaturing HPLC: Denaturing HPLC is emerging as a good candidate for SNP detection and has the advantage over 1. in that it will detect point mutations as well as short deletions. Both approaches could be regarded as complementary. Strategies for the discovery of mutation generated SNPs and short deletions in pooled samples will be explored through collaboration.

3. Matrix Assisted Laser Desorption Ionisation - Time of Flight spectrometry (MALDI-TOF): MALDI-TOF is mainly used for the separation of proteins or synthetic polymers. Analysis of DNA suffers from sample fragmentation during flight. Thus, at the moment, only short DNAs are suitable for analysis. However instrument manufacturers are currently focussing on developing the technology to extend the applications of their instruments for the analysis of DNA. We will work with Bruker Instruments and Genomics Solutions to test the efficacy of MALDI-TOF for mutation detection using target DNA's PCR'd from individual mutants and from super pools.

Two fast neutron irradiated populations have been created at JIC and lines taken through to M3. Seed are available for lines through Professor John Snape and Steve Reader at Genetic Resources Unit JIC (refer to section 1 for the application of wheat deletion lines).

Buschges et al. (1997) Cell 88: 695-707, Cochrane and Skopek (1994a,b) Carcinogenesis 15: 713-717, 719-723, Panstruga et al. (1998) NAR 26:1056-1062, Steen et al. (1997) Mutagenesis 12:61-67, Westlund et al. (1999) PNAS USA 91:1381-1385, Yandell et al. (1994) PNAS USA 96: 2497-2502,

Contributors

JIC: Graham Moore

IACR-LARS: Keith Edwards

SCRI: David Marshall, Robbie Waugh

Personnel and Equipment Requested: Two Band 6 positions, one each at both IACR-LARS & SCRI. Each of the band 6 posts will undertake a shared development/management role in addition to responsibility for local data management. The studentship at JIC will focus on integration with the UKCropNet and USDA databases whereas the development roles of the IACR-LARS & SCRI posts will focus on database schema, interactions with the MENDEL database and procedures to handle the mutant and EST data respectively. Three Sun UNIX Ultra 60 Workstations and an ORACLE 8 license to cover use at all 3 sites for the duration of the project and Web-based Access to the information resource for the user community are also requested.

A key component of the Cereal Community Resource is the development and operation of a Bioinformatics Infrastructure to support not only the operation of the experimental components of the programme but also the long term maintenance and development of the associated data. Our aim is to develop an Integrated Bioinformatics infrastructure to provide a framework for the coordination of the project between the three sites. Such integration of the data from the different components of the project and with relevant external data sources (including the UK CROPNET and USDA Plant Genome databases and the EMBL and NCBI Sequence databases) will maximise the value of the biological and data resources that we will develop. The efficiency of data capture, analysis, storage and access i.e. the bioinformatics infrastructure, is crucial for the success of the project. If it is not right then all subsequent manipulations will be flawed. A key element will be the provision of access to this data resource to the UK academic community, which will be achieved through a password-protected web interface.

We propose to adopt a pragmatic approach to relevant bioinformatics by the use of a combination of in-house development and free & commercial third party software to support the different components of the project. The key data resource for the project will be a central database CCRDB (Cereal Community Resources DataBase) based on a series of ORACLE RDMS modules running on SUN Unix workstations at each site. In order to provide essential resilience, this database will be replicated at least daily at each of the partner sites as will user-access permissions. In addition, we will make use of a series of software tools and scripts for data analysis and information processing. We intend to make use wherever possible of robust third-party software in order to focus software development on specific gaps in existing software provision. For example PE Informatics are currently developing a range of ORACLE modules which would support a number of aspects of the system we would like to develop. The PE SQL*GT module could provide the LIMS support for sample tracking and the BioLIMS and BioMERGE modules could provide a platform to process and manage the sequence data allowing effective integration with the instruments and data analysis software modules. There is also a wide range of public domain Unix tools currently available for the processing of sequence data. However there is, as yet, no comprehensive set of tools available that will provide a complete cost-effective solution to our needs.

This project will enable us to share our resources and experience to develop a coordinated bioinformatics system based on existing tools and new software development that will maximise the value of the project to the user community.

The four major elements of the proposed RDMS system are:

1. A Laboratory Information Management System (LIMS). This will be a key element of the project and will be required for efficient control of work plans for the laboratory components of the project ensuring precise tacking of samples, plates, filters etc and the efficient control of laboratory work plans.
2. A Sequence Information Management System (SIMS). We need to establish an efficient system, which will enable us to automate the basic processing of sequence data (e.g. trimming vector etc) as well as homology searching both within the growing wheat and barley resource and against internal and external databases. A range of analysis tools are freely available e.g http://www.sanger.ac.uk/software. However, the appropriate components will require integration / adaptation for our specific needs. A robust system will be required for the long term maintenance of this data and updating by automatic searching at repeated intervals and the updating of relevant annotation. Dr. Dave Lonsdale will support the automatic interrogation of the MENDEL plant gene family database and reporting back to the central information database.
3. A Plant Information Management Systems (PIMS). This element will responsible for the tracking of plant lines from the mutagenised populations both in terms of the pedigree structure for their development and the maintenance and distribution of stocks.
4. An Hybridisation Array Management System (HAiMS). As this is a new and rapidly developing area it is still currently difficult to predict what is either the most cost effective or the most efficient method to analyse and/or store this class of data. For example, there is currently a range of 'free' software such as that from the Brown lab at Stanford as well as a developing range of alternative commercial (often very expensive) options. We will endeavor to ensure that our storage of array data conforms to appropriate standards as they develop.

A key element of the informatics infrastructure will be the annotation of both sequence and resources. With the volume of data proposed it is clear that as much automated annotation as possible is desired and will enhance the value of the information. However, it is important that annotation is not only automated, e.g. by the regular inclusion of updated information on the results of current BLAST searches against public databases, but is also fully integrated across the project e.g. ensuring that information on a particular knockout also can be linked to the relevant sequence and expression data. Therefore, provision will be made (under strictly defined data rules) to allow all users of CCRDB to provide both annotation of sequence and resources based on their experience and knowledge and, where relevant, new data of their own e.g. for the sequence or expression arrays classes. Where appropriate, links to external data sources (e.g. the CROPNET and EMBL databases) will be provided through the provision of a CORBA data interface to the CCRDB database. This will also facilitate the use of JAVA tools developed through UK CROPNET and other related projects with the CCRDB.

Though these resources have been requested for the duration of the development phase of the project, we will undertake to support the Bioinformatics infrastructure and maintain the data resources (i.e. CCRDB) and user community access well beyond this period.

Dr Graham Moore will be the overall co-ordinator. Dr Robbie Waugh and Dr Keith Edwards will be the site co-ordinators for SCRI and IACR-LARS. It was decided for ease of management of the programme that the cereal researchers at the three sites be represented by one researcher. It is intended that the co-ordinators / or personnel employed should meet once every 4 months. This has been costed into the programme.

We recognise that programmes generating resources in the public sector do not generally produce publications. However it is still necessary to provide the scope for pursuing small projects with the potential to generate publications, particularly for the postdoctoral researchers within the programme who require publications for their career. Failure to do so could lead to a high turnover of staff within the programme. To that end, we will make a commitment to provide where possible sufficient scope for this to be achieved. Within the limits of space we are not able to describe in detail these specific projects.

However at JIC, the postdoctoral researcher will have the opportunity to characterise the Ph1 (homoeologous) locus using the wheat BACs generated in conjunction with the new deletion lines available. At SCRI, the postdoctoral researchers will attempt different approaches for mutation detection (reverse genetics) and preliminary experiments with high density arrays. Postdoctoral researchers at IACR-LARS will carry out pilot transcriptome studies with the arrays. In particular, the profile of genes expressed at high levels in very early grain development (3-4 days post fertilisation) and in the very early stages of germination will be studied. In each case RNA derived from single seeds or seed will be used. The barley arrays will also be tested using RNA from a germinated seed.

A number of companies including Zeneca and Dupont have indicated their interest in using the resources generated. Researchers at SCRI, JIC and IACR-LARS will make immediate use of the resources. BBSRC, institutes and university researchers will make increasing use of the resources as research on cereal biology expands.

For each of the components incorporated into the resource, access will be provided through implementation of a 'Hotel' based system, with researchers coming from 'client' labs to the resource to perform their experiments and record their results. When required, distribution of filters or arrays or DNA pools will also be considered. This service will be provided free of charge during the lifetime of the programme to UK academics and BBSRC/SOAEFD employees. A small charge will be levied on industrial users. After this, a small charge will be made to cover the costs and contribute towards maintenance of the resources. This charge will be at three levels and will be set towards the end of the project:

i. UK academic researchers,

ii. Any academic researcher and

iii. Commercial organisations.

All services will be undertaken on the basis that ALL the information generated will be made public within six months of the data being generated.

If funded, this programme will be just the first step in updating the UK cereal research community with the resources that will be required for the 21^st century post-genomics era. To date, cereal community planning for the post-genomics era has been inhibited by the lack of publicly available sequence information and large insert libraries. Such resources have been available but at a price for future academic freedom and UK plc intellectual property. This programme will go some way to alleviating this situation and should therefore result in a significant expansion in the number of researchers actively involved in cereal based programmes. We hope that due to this programme two types of research will be promoted:

1. Further large scale, resource intensive work on aspects such as cereal proteomics, contig mapping and the development of technologies for the large scale characterisation of gene function.
2. Numerous individual research programmes, which directly or indirectly utilise the tools and information generated via this initiative. Such programmes will be initiated on an individual bases by researchers using both responsive mode funding and specific initiatives.