¹Horticulture Research International, Wellesbourne, Warwick, CV35 9EF, UK;

ABSTRACT

Brassica genomes possess a more complex organisation than that of the related crucifer Arabidopsis. We have started to carry out systematic and detailed analysis of MADS-box, AUX-1, HSP-17 and other developmental gene families in Brassica oleracea. This is providing information about locus-specific variation in coding and non-coding sequences. We have observed contrasting patterns of sequence variation in 5' regulatory regions of MADS-box and HSP-17 loci. The ramifications of locus replication in crop brassicas are discussed in the context of developmental and environmental adaptations of regulatory genes. In addition, we address the need for development of appropriate methods for analysis and hypothesis testing when assigning gene function, in the context of increased genome complexity.

INTRODUCTION

Analysis of gene families in the crop species Brassica is complicated by the typical triplication, in diploid species, of large segments of the genome when compared with the closely related crucifer Arabidopsis. In an attempt to understand the extent of functional specificity or redundancy at different loci we are isolating sequences from B.oleracea which represent members of several gene families. In particular we are interested in identifying locus-specific differences in coding and promoter sequences, and associating these with variation in gene expression.

GENETIC AND PHYSICAL MAPS

Brassica oleracea includes the common vegetable crops cabbage, kales, cauliflower, Brussels sprout, broccoli, kohlrabi. The C (n=9) genome represents one half of the AC genome of B.napus which includes oil seed rape (canola) and Swedish turnip. The consolidated linkage map of B.oleracea currently consists of over 580 loci, including RFLP, AFLP and SSRs (R.L. Sebastian, personal communication). This is based on segregation data from two populations (Table 1) with good integration and agreement of marker order between populations.

Female parent	Male parent	Pop. Size	Code	Map ref.
var. alboglabra DH	var. italica DH	170	AG	Bohuon et al.
var. botrytis DH	var. gemmifera DH	110	NxG	Sebastian et al., personal communication
var. botrytis DH	var. italica inbred	70	NxB
var. capitata DH	var. italica DH	150	MHB
var. capitata	var. capitata	120	HxB

Table 1. Reference mapping populations of doubled haploid lines for B.oleracea

To facilitate physical mapping within Brassica oleracea, we constructed a genomic library from the doubled haploid line A12DH. The library consists of 26,000 clones in the bacterial artificial chromosome (BAC) vector pBeloBAC11, with an average insert size of 120kbp. This represents an estimated five-fold coverage of the genome.

Genetic mapping studies have indicated that each segment of the Arabidopsis genome is essentially represented in three copies in the diploid Brassica genomes, with extensive rearrangements in collinearity (Lagercrantz, 1998). Based on information from Arabidopsis, we have been using conserved sequences from coding regions to develop PCR-based assays in Brassica. The PCR products have been used in several ways. Polymorphism detected in PCR products from genomic DNA of doubled haploids has been used as the basis of mapping assays. In addition we have used the PCR products, both directly and as cloned sequences, as probes to screen the BAC-library and identify specific clones. Following probing of the B.oleracea BAC library, locus specific contigs have then been characterised. We are then able to match alleles, at the sequence level, from the BAC library (derived from A12DH alboglabra) with those segregating in the reference mapping population (AG).

CHARACTERISING GENE FAMILIES AT THE MOLECULAR LEVEL

Within the context of the consolidated map, we are currently locating members of several gene families important in regulation of plant development.

MADS-box gene families

The MADS-box domain gene family of AGAMOUS-like putative transcription factors (DNA binding proteins) has been well characterised in Arabidopsis, with at least 36 members described to date (Liljegren et al, 1998). We have carried out detailed analyses of the B. oleracea homologues of CAL, AP1 and AGL8. Several copies of each gene sequence have been identified. In two cases, the functional locus has been determined in agreement with segregation of crop morphotype (cauliflower vs Calabrese-type broccoli).

At the nucleotide level there is very high conservation of coding sequence between different copies of each of the BoCAL, BoAP1 and BoAGL8 genes in the B.oleracea genome. However, we have demonstrated that there is considerable locus-specific variation in the non-coding regions. As an example, for BoAGL8 we have isolated three genomic copies on distinct BAC contigs, thus accounting for all RFLP bands observed on Southern filters of genomic DNA. The coding sequence is highly conserved amongst these copies, although there is very low homology between intron and 5' flanking sequences. The promoter regions over 500bp appear extremely diverged, suggesting selection for locus-specific regulation.

Single loci of BoAP1 and BoCAL are implicated in the development of the arrested stage of floral meristem development, which is expressed as the common cauliflower curd. We have developed a genetic model to account for the variation observed in segregating populations from crosses between calabrese and cauliflower. The model also accounts for the intermediate forms seen in specific crop types important in the recent domestication of the curd form. It appears that the interaction of these genes is orthologous to that observed in double Ap1/Cal mutants of Arabidopsis (Kempin et al., 1995). However, in B. oleracea there is no corresponding mutant phenotype expressed at the later floral stage, and wild-type flowers are produced. This would suggest that replicated loci, other than those accounting for the arrest of the floral meristems at the curd stage, are fully functional at the later stage of development.

The arrest stage observed in cauliflower curds is responsive to environmental cues. We are currently investigating the role of the different loci in a range of phenotypes observed as 'physiological defects' in response to changes in temperature. For example, "riciness" of cauliflower curds, which is defined as growth of flower buds out of the curd surface, is induced when low temperatures follow curd formation. "Bracting" occurs under warm growing conditions, and results from growth of cauline leaves through the curd. These develop from axillary bracts of primary peduncles. "Hairyness", the result of bracteoles developing around each flower bud primordium, is also induced at elevated temperature.

The small-HSP gene family

Accumulation of products belonging to the small heat-shock protein gene family (e.g. Hsp17.4, Hsp 17.6) has been associated with variation in seedling development of rapid cycling brassicas (Bettey et al., 1998). RFLP probing suggested the presence of at least three conserved copies in the B. oleracea genome. To date we have isolated three of these as contigs from the BAC library. We have carried out detailed sequence comparisons of the genomic sequences, using both inverse PCR and sub-cloning of BAC clones to isolate flanking sequences. We have found that the coding sequences are highly conserved, not only between Arabidopsis, B. oleracea and B. rapa, but also between distinct loci within the B. oleracea genome. In addition, for two distinct copies within the B. oleracea genome lying on different BAC contigs, the promoter region over 800bp is also highly conserved, with 95% nucleotide homology. Conserved heat-shock element motifs have been detected upstream of the TATA box. This suggests that the promoter regions may be under strong selective pressure due to their specific regulatory response

The LAX gene family

The AUX1 gene encodes a permease-like protein, which regulates lateral root formation in Arabidopsis by acting as an auxin influx carrier (Bennett et al., 1996; Marchant et al., 1999). AUX1 is a member of a small conserved family termed LAX genes. The LAX gene family has been mapped, using a recombinant inbred population, to four independent loci on the Arabidopsis genome. Recent screening of the Arabidopsis IGF-BAC library, at low stringency, not only confirmed the map positions, but indicated that these four members are probably the extent of the LAX gene family.

We are currently using a range of approaches, similar to those described above, to identify and isolate functional homologues from B. oleracea. Clones have been isolated from the B.oleracea BAC library following screening with AtLAX gene probes. The characterisation of distinct loci of these genes will place particular emphasis on the regulatory regions. Functional differences may be investigated by complementation of mutant lines in Arabidopsis, although this will depend on the extent of sequence conservation amongst the regulatory regions.

ASSESSING ALLELIC VARIATION AT FUNCTIONAL LOCI

For the BoCAL-a locus we have been able to exploit a simple sequence repeat (SSR) polymorphism in an intron to carry out an allelic survey. This was carried out on over 200 accessions from the genetic resource collection at Horticulture Research International. The collection of Brassica accessions sampled represented land-races and cultivars from Italy, the centre of diversity for cauliflower and broccoli. It was apparent from previous collecting information and comparative studies (Massie et al., 1996) that different genotypes of these crops are well adapted to local growing conditions. The survey demonstrated that mutant alleles affecting curd development had been recruited from a range of crop types into the cauliflower gene pool.

THE RAMIFICATIONS OF LOCUS REPLICATION

With a recent crop such as cauliflower there appears to have been sufficient allelic variation in the domesticated gene-pool to enable selection of crop adaptations (landraces) within each region of Italy. Each landrace has been selected either to be morphologically distinct, or to reach harvestable maturity in specific months of the year. The genetic and transcriptional regulatory networks thus appear to be well adapted to local conditions. Specific genotypes are able to integrate local differences in temperature during each season to meet vernalisation requirements (Wurr & Fellows, 1998). In addition they appear to be adapted to day-length and levels of irradiance resulting from variation in latitude and season. The relative ease with which farmers and breeders have been able to select optimal combinations of alleles suggests that there may be considerable advantages conferred by replicated loci in crops such as Brassica. For key regulatory genes such as MADS-box genes, which are well conserved in eukaryote evolution, environmental adaptations are unlikely to arise from variation in the coding sequence.

It can be postulated that there is likely to be a non-linear increase in degrees of freedom for gene expression in Brassica crop plants when compared with Arabidopsis. As a result of the overall triplication of gene loci, for any one class of gene there is increased scope for variation in timing and level of transcription in response to different developmental and environmental cues. Differential expression of alleles, loci and members of gene families are particularly relevant for putative transcriptional factors such as MADS-box genes, where specific interactions have been suggested between the K-domains of these proteins (Fan et al., 1997). Variation observed in regulatory regions of different loci of the conserved gene BoAGL8 may reflect the capacity for different developmental or environmental responses by this regulatory genes.

TRAIT DETECTION IN Brassica

Qualitative traits such as flower colour, anther-spot and self-incompatibility alleles map to single loci on the integrated linkage map of B. oleracea. Using the reference doubled haploid mapping populations as the starting material we are now carrying out a range of studies to characterise quantitative traits affecting development. Quantitative phenotypic data from replicated field and glasshouse trials have been scored for a wide range of traits. The traits include flowering and vernalisation requirement, water use efficiency, photosynthetic potential, seedling vigour, responses in tissue and microspore culture, and a range of morphological traits. The morphological traits include variation in leaf, internode, and flowering head characters. For many of these traits significant QTL have been detected which account for the genetic variation in the populations. In order to resolve such loci further it is important to increase the effective number of recombinants in a population, either through pedigree analysis from breeding populations and genetic resources, or through the use of substitution libraries of recombinant backcross lines (Ramsey et al., 1996). A further strategy being adopted is to exploit local collinearity and sequence synteny between the Brassica and Arabidopsis genomes. Based on the complete Arabidopsis genome sequence, we can carry out detailed analysis by using a large numbers of markers or candidate genes to saturate a locus in Brassica.

CONCLUSION

In a heterozygote plant of B.napus there may be up to 12 copies of a gene when compared with Arabidopsis. This increase in genomic complexity raises questions for development of efficient strategies for assigning and searching for gene function. It is important to assign complete gene sequences to specific loci, and understand the interaction between regulatory regions and phenotype at different developmental stages, and in different environments. Current methods for hypothesis testing, which in Arabidopsis are based on complementation via transformation, may need to be refined in the context of locus and allelic replication in the brassicas.

ACKNOWLEDGEMENTS

This work is funded primarily by the Biotechnology and Biological Sciences Research Council. Mapping of RFLP and AFLP markers in the NxG population was carried out by Rachel Sebastian at Birmingham University.

REFERENCES

1. Bennett MJ, Marchant A, Green HG, May ST, Ward SP, Millner PA, Walker AR, Schulz B, Feldmann KA (1996) A permease-like regulator of root gravitropism. Science 273:948

2. Bettey M., Sinniah UR, FinchSavage WE, Ellis RH (1998) Irrigation and seed quality development in rapid-cycling brassica: Accumulation of stress proteins. Annals Botany 82:657-663.

3. Bohuon et al. (1996) Alignment of the conserved C genomes of Brassica oleracea and Brassica napus.

4. Theoretical and Applied Genetics 93:833-9

5. Fan H-Y, Hu Y, Tudor M and Ma H (1997) Specific interactions between the K domains of AG and AGLs, members of the MADS box domain family of DNA binding proteins. The Plant Journal 12:999-1010

6. Kempin, SA, Savidge B, and Yanofsky MF (1995) Molecular basis of the cauliflower phenotype in Arabidopsis. Science 267: 522-525

7. Lagercrantz U (1998) Comparative mapping between Arabidopsis thaliana and Brassica nigra indicates that Brassica genomes have evolved through extensive genome replication accompanied by chromosome fusions and frequent rearrangements. Genetics 150:1217-1228

8. Liljegren SJ, Ferrandiz C, Alvarez-Buylla ER, Pelaz S, Yanofsky MF (1998) Arabidopsis MADS-box genes involved in fruit dehiscence. Flowering Newsletter 25.

9. Marchant A, Kargul J, May ST, Muller P, Delbarre A, Perrot-Rechenmann C, and Bennett MJ (1999) AUX1 regulates root gravitropism in Arabidopsis by facilitating auxin uptake within root apical tissues. The EMBO Journal 18:2066-2073.

10. Massie, I H, Astley D and King GJ (1996) Patterns of genetic diversity and relationships between regional groups and populations of Italian landrace cauliflower and broccoli (Brassica oleracea L. var. botrytis L. and var. italica Plenck). Acta Horticulturae 407: 45-53.

11. Ramsay, L.D., Jennings, D.E., Bohuon, E.J.R., Arthur, A.E., Lydiate, D.J., Kearsey, M.J. and Marshall, D.F. (1996) The construction of a substitution library of recombinant backcross lines in Brassica oleracea for the precision mapping of quantitative trait loci. Genome, 39, 558-567

12. Wurr, DCE and Fellows JR (1998) Leaf production and curd initiation of winter cauliflower in response to temperature. Journal of Horticultural Science & Biotechnology 73: 691-697.