This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Doebley, J. Search for Related Content PubMed PubMed Citation Articles by Doebley, J. Agricola Articles by Doebley, J.

An Evolutionist Visits the 43rd Annual Maize Genetics Conference

Laboratory of Genetics University of Wisconsin Madison, WI 53706

jdoebley{at}facstaff.wisc.edu

From modest roots of a few dozen maize geneticists sitting around a small room in the Allerton House in Illinois in the 1950s, the Maize Genetics Conference has grown this year to a less modest but still comfortable size of nearly 500 attendees, the largest maize meeting to date. It was held at the Grand Geneva Resort in southern Wisconsin, easily accessible but secluded enough to allow the participants to focus on maize. It featured four plenary talks and two workshops, with 21 additional invited talks, 25 contributed talks, and more than 170 posters. Packaged into just four days (March 15 to 18), the meeting had an intensity that kept the participants engaged, if not overwhelmed, leaving at least one of them ready to track back home to download and process the flood of new ideas, contacts, and information. In this report, I offer no pretense of providing a comprehensive abstract of all or even most of the talks and posters. Rather, this report provides the admittedly biased perspective of an evolutionary geneticist. Still, I hope to convey the excitement and energy of the meeting, and I offer my apology to the many fine speakers and poster presenters whose work I do not mention. Abstracts for all of the talks and posters can be found at http://www.agron.missouri.edu/Coop/Conf/2001.html.

PLENARY TALKS

Kelly Dawe (University of Georgia, Athens) kicked off the meeting with his talk on maize Abnormal chromosome 10 (Ab10) and its influence on chromosomal segregation. The Ab10 chromosome contains factor(s) that produces "meiotic drive" such that Ab10 is transmitted to progeny at frequencies in excess of Mendelian expectations. It was demonstrated by Marcus Rhoades in the 1940s that this system for meiotic drive involves heterochromatic regions (knobs) that can act as neocen-tromeres in the presence of Ab10. This allows chromosomes with knobs (Ab10 itself has a very large knob) to be pulled to the poles more rapidly than chromosomes lacking knobs and thus to be represented at greater than Mendelian proportions in the distal megaspore that forms the embryo sac. Dawe presented work from his laboratory demonstrating that the mode of neocen-tromere movement is distinct from that of the true centromeres, involving the action of neither MAD2 (a regulatory protein of the spindle) nor CENPC (a structural protein of the centromere). Dawe's laboratory also tested whether polar flux of the spindle microtubules might be involved by using taxol to stabilize the microtubules. Because taxol did not stop neocentromere movement, it does not appear that the neocentromeres are simply hitching a ride—via polar flux—on the microtubules. Dawe and colleagues are investigating the possibility that members of the kinesin superfamily of proteins, which are involved in chromosome and organelle movement in animals, might control the movement of the neocentromeres in maize.

June Nasrallah (Cornell University, Ithaca, NY) gave an overview of her exceptional work on the genetics of self-incompatibility (SI) in Brassica. SI is controlled by the S locus, a compound locus consisting of two separate genes: SRK, which encodes a receptor kinase, is the stigmatic element; SCR, which encodes a cysteine-rich peptide, is the pollen element and is predicted to be a ligand for SRK. A third gene, SLG, which encodes a glycoprotein localized to the cell wall of the stigma epidermis, is found in the majority of S haplotypes and appears to be required for the stabilization of SRK. Having worked out the control of SI in Brassica, Nasrallah and her group have been investigating the evolution of SI and the S locus in Brassica and other genera. In Brassica, they have shown that SRK (the female element) and SCR (the male element) have undergone a form of coevolution such that the allele phylogenies of these two genes are nearly identical. This is what one would expect if each change in the male component was matched by a change in the female component (or vice versa), as required for SI to be maintained over time. Allelic polymorphism at the S locus has been shown to date from 20 to 40 million years ago, a result expected under frequency-dependent selection. If one allele becomes too abundant in the population, plants carrying this allele have a disadvantage in finding a mate, whereas plants with rare alleles cross successfully with most others. Under frequency-dependent selection, genetic drift does not lead to the loss of alleles at the normal rate; therefore, ancient variants are maintained much longer. The Nasrallah group has found that in Arabidopsis the S locus is nonfunctional and contains pseudogenes for SRK and SCR, explaining the self-compatible mating of this colonizing weed.

Laurie Smith (University of California, San Diego) presented an elegant talk about her work on the cell biology of leaf development. She briefly mentioned her published work on tangled (tan1), a mutant with abnormal cell divisions in all tissue layers of the leaf and elsewhere in the plant. Tan1 encodes a highly basic protein that can bind microtubules. She then discussed three classes of mutants (brick, discordea, and pangloss) that affect the proper orientation of asymmetric divisions in the leaf epidermis. Smith used most of her 40 minutes to discuss the brick mutants, in which the pavement cells of the leaf blade epidermis lack their characteristic lobed margins so that they superficially resemble the cells of the sheath epidermis (or bricks). The brick mutants enabled Smith and her student Mary Frank to test two models for the formation of the lobes during cell differentiation. One model is that the lobes are produced during cell expansion because of thick and thin bands of cellulose along the cell wall. When the cell expands, the regions of thinner cellulose deposition would bulge out to form the lobes. An alternate possibility is that the lobes are formed by tip growth, the same form of growth seen in pollen tubes or trichomes. Smith argued in favor of the tip growth model for lobe formation in part because the brick1 mutant also has short trichomes (reduced tip growth) and because actin is deposited in the lobes of epidermal cells, as expected for tip growth. Smith's group has cloned Brick1 and found that it encodes a small protein of only 86 amino acids that also is found in humans, Drosophila spp, and Caenorhabditis elegans.

In has long been known that nuclear genomes tightly regulate the expression of chloroplast-encoded genes, but how this is accomplished is not understood. In this context, Alice Barkan (University of Oregon, Eugene) presented work from her laboratory on the control of chloroplast gene expression. Barkan's group has used mutant screens to identify and clone nuclear genes that regulate the expression of genes encoded in the chloroplast genome. They initially screened Mutator populations for phenotypes suggestive of defects in photosynthesis (e.g., pale green seedlings) and performed secondary screens to identify mutants that failed to splice group II introns in the chloroplast. By this means, they identified and cloned the chloroplast RNA splicing (crs1 and crs2) genes. crs1 specifically functions in the splicing of a single group II intron, whereas crs2 assists in the splicing of multiple group II introns. crs2 is a homolog of peptidyl-tRNA hydrolase enzymes, but it has acquired several unique features that may account for its new function in group II intron splicing. crs1 is a member of a gene family of unknown function that is represented by 16 genes in Arabidopsis. Members of the crs1 family contain one or more copies of a 10-kD domain of ancient origin with features suggestive of a novel RNA binding domain. Evidence was presented that different members of the crs1 family in plants may function in the splicing of different group II introns, with one member of the family functioning in concert with crs2. The genetic approach being used by Barkan's group is revealing at a fine molecular level how eukaryotic genomes have taken over the control of gene expression in the chloroplast, which is a relic of a once free-living bacterium.

FUNCTIONAL GENOMICS MEETS POPULATION GENETICS

A decade ago, one of the leading lights of maize genetics declared at a conference that he did not want to hear any talk about "allele frequencies." Indeed, many academic departments were cleansing themselves of population genetics, so the discipline is now restricted largely to departments of evolutionary biology. The genomics era, with its mountains of sequence data and the puzzle of so many genes and so few known functions, has opened a crack through which population genetics is again slipping back into the mainstream of biology, although sometimes under the cloak of bioinformatics. The trend was visible at this year's maize meeting, with three separate groups presenting experimental results on the association of phenotypes with sequence polymorphisms in populations, so-called association analyses (Cardon and Bell, 2001). The logic behind association analyses is simple, even if the statistics are not. First, one measures a phenotype (or hundreds of phenotypes) on individuals in a population. Second, one determines the sequence of a gene (or hundreds of genes) on the same individuals. Finally, one looks for statistical associations between the phenotypes and allelic variants at the genes.

Ed Buckler and his group (United Agricultural Research Service/North Carolina State University, Raleigh) presented three posters and a talk at the conference related to association analysis. Artfully leaving out the details of his statistical analyses, Buckler presented compelling evidence that a tryptophan-to-arginine substitution in sugary (su1), which encodes an isoamylase, gave rise to sweet maize (a conclusion presented independently at the meeting by Jason Dinges from Martha James's group [Iowa State University, Ames] on the basis of biochemical evidence). Buckler went further and suggested that variation at su1 also is involved in making maize flour more palatable than was that of its ancestor, teosinte, by modifying the balance of different forms of starch in the endosperm. Jeffry Thornsberry of Buckler's group presented equally compelling results that Dwarf8 (a transcriptional regulator of the SCARECROW class) controls natural variation in silking date in maize. Finally, Bradley Rauh, also of Buckler's group, presented evidence that the prolamin binding box factor (Pbf) shows a strong signature of selection during maize domestication and is associated with natural variation in the abundance of protein in the endosperm.

As Buckler's group has shown, part of the promise of association analyses is that one can study dozens (or hundreds) of phenotypes and genes simultaneously without the need to develop specialized stocks or populations for each phenotype and gene. In principle, any variable phenotype in maize can be tested for association with any gene for which there is allelic variation. Part of the power and appeal of association analyses is that they can provide far greater resolution than quantitative trait loci (QTL) mapping. In theory, if there has been sufficient recombination, then association analyses will have the power to identify specific nucleotide or amino acid substitutions that control phenotypic differences among individuals.

Nevertheless, there is a clear risk that association analyses could give false-positive results under some circumstances, such as when the controlling polymorphism is of recent origin or is located in a region of suppressed recombination. For this reason, it is essential to investigate some cases in which the link between gene and phenotype is well established so that any misleading associations with neighboring genes can be exposed. To explore the utility of association analysis for U.S. breeding lines, Kelly Palaisa of Antoni Rafalski's group (DuPont Agricultural Genomics, Newark, DE) is examining polymorphism for yellow (y1), a gene with a well-understood phenotypic effect. y1 encodes phytoene synthase and controls yellow versus white kernel color. The Rafalski group was able to show that an INS2 insertion in the promoter of y1 is associated with the yellow kernel phenotype, but because there was extensive linkage disequilibrium at y1, the phenotype is associated with other polymorphisms as well. The Rafalski group will be extending their study to surrounding genes to learn if the yellow kernel phenotype is associated with polymorphisms in neighboring genes. This type of study will make clear whether there has been sufficient historical recombination in U.S. maize to break up the linkages between the polymorphisms in genes that control phenotypes of interest and polymorphisms in neighboring genes.

THE RICE GENOME PROJECT: A WAKE-UP CALL FOR THE MAIZE COMMUNITY

The meeting included a workshop on the rice genome project that featured eight talks organized by Sue Wessler (University of Georgia, Athens). Takuji Sasaki (National Institute of Agrobiological Resources, Tsukuba, Japan) provided a progress report on the International Rice Genome Sequencing Project (IRGSP), which is proceeding ahead of schedule (see http://rgp.dna.affrc.go.jp). A draft version of the rice genome sequence is expected in mid 2002, with a full, high-quality version due in 2004. In the meantime, anxious users can BLAST or download the unfolding public rice sequence via GenBank. Rod Wing (Clemson University, Clemson, SC) and Robin Buell (Institute for Genome Research, Rockville, MD) reported IRGSP's progress on rice chromosomes 3 and 10, Gerard Barry provided an overview of Monsanto's draft sequence of the rice genome, and Steve Briggs discussed Syngenta's draft sequence of the rice genome. Other talks in the workshop centered on developing new tools for rice genetics or applying old ones more effectively in a world in which the full rice genome sequence is on line. David McElroy of Maxygen spoke about Maxygen's "Delete-a-Gene" technology, which combines fast neutron mutagenesis and polymerase chain reaction screening of DNA pools to identify deletions in genes of known sequence. Masahiro Yano (National Institute of Agrobiological Resources) beautifully outlined how QTL for phenotypes such as heading date can be readily cloned once the full genome sequence is on line (Yano et al., 2000). Importantly, his research revealed that to do this, one needs complete access to the full genomic sequence.

One announcement of interest in the rice workshop came from Steve Briggs, who estimated that the rice genome contains 50,000 genes, fully two-thirds more than are estimated to be nestled in the nucleotides of our own species. Given that maize, an ancient polyploid, could easily contain 50% more genes than rice, a "true" diploid, one might speculate that maize contains 75,000 genes. Is it conceivable that maize has twice the gene content of humans, with enough left over to encode the C. elegans genome? Venter et al. (2001), in their discussion of the surprisingly low gene content of humans, discuss an outdated population genetics theory that the maximum number of possible genes for humans is 30,000 (see Muller, 1967). This argument is based on a maximum amount of genetic load given a certain mutation rate and fitness effects of the mutations. This theory implies that the unexpectedly low gene number in humans may reflect a limit imposed by genetic load. The theory was incomplete, however, because it did not consider the possibility of epistasis between deleterious genes (see Eyre-Walker and Keightley, 1999). Therefore, there is no theoretical reason that maize should not have twofold or more the number of genes as humans or that the human number should not be greater than 30,000.

If the rice workshop made anything clear to me, it was that the future of maize as a model organism is in jeopardy without the complete genome sequence. The elegant work of Yano et al. (2000) on the cloning of QTL and similar work in Arabidopsis (Alonso-Blanco and Koornneef, 2000) underscore the ease with which such work can be accomplished when the genome is fully sequenced; they also show the futility of working in systems that lack this advantage. Although the colinearity among the grass genomes will allow maize geneticists to play off the rice genome to some degree, it is no substitute for having the maize genome sequence itself.

A WORKSHOP ON MAIZE GENOMICS AND FUTURE DIRECTIONS

Torbert Rocheford (University of Illinois, Champaign-Urbana) moderated a workshop on maize genomics that featured a series of 12 five-minute updates on a selection of maize genome projects. Here are a few of the topics: Donal O'Sullivan (University of Bristol, UK) reported on his group's analysis of the rp1 disease resistance locus; Michele Morgante provided an overview of DuPont's bacterial artificial chromosome and expressed sequence tag libraries for maize; Pablo Rabinowicz from Rob Martienssen's group (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY) presented their strategy for sequencing hypomethylated portions of the maize genome; Tom Brutnell (Boyce Thompson Institute, Ithaca, NY) and Paul Chomet (Mystic Research, Mystic, CT) discussed improvements in technologies for forward and reverse genetics using Ac and Mu; Kan Wang (Iowa State University) discussed protocols for maize transformation using Agrobacterium tumefaciens; and Antoni Rafalski (Dupont Agricultural Genomics) discussed the effort to assemble a single nucleotide polymorphism database for maize.

After this whirlwind tour of maize genomics, Jeff Bennetzen (Purdue University, West Lafayette, IN) and the other members of the Maize Executive Committee stepped onto the stage to discuss future directions for maize. Bennetzen outlined the committee's priorities for maize, and topping the list was sequencing of the maize genome. The announcement was followed by remarkably little discussion, reflecting the consensus among maize geneticists that the genome must be sequenced. The plan would involve sequencing 350 Mb of the 2500-Mb maize genome, restricting the effort to hypomethylated or gene-rich regions. The combination of the sequence of the gene-rich regions for maize and the complete sequence for rice would keep maize genetics competitive with other model organisms. The greatest challenge may be to avoid the problems that have arisen with other genome projects when commercial and public efforts were either uncoordinated or antagonistic. With this in mind, the Executive Committee has proposed to organize a workshop on maize genome sequencing involving the key public laboratories and companies such as DuPont, Syngenta, and Monsanto.

POSTERS AND SHORT TALKS

It would be hopeless for me to try to summarize the 200 posters and short talks and worthless for me to present a single sentence on each of a dozen or so. So with my apology to those not mentioned, let me discuss three of my favorites at slightly greater length.

Quantitative genetics (the inheritance of quantitative traits) operates on the assumption that variation is under genetic control with some environmental noise. Epigenetics has not been factored into the linear models usually used in quantitative genetics, yet epigenetic mechanisms may contribute to quantitative variation. Ken McWhirter, Julie Zinnert, and Bill Eggleston (Virginia Commonwealth University, Richmond) presented a poster on the interface of quantitative genetics and epigenetics. Their work is on the R locus, which is a regulatory gene controlling anthocyanin deposition in the aleurone of developing maize kernels. R is one of several loci in maize that have served as models for the investigation of paramutation. It is known that the level of R gene expression for the paramutable alleles is inversely correlated with the level of methylation. McWhirter and colleagues isolated a series of R alleles that no longer participated in paramutation but that exhibited a range of phenotypes from pale to dark purple kernels. They further reported that this quantitative variation was correlated with the level of methylation at the 3' end of the gene. No structural differences were found among these derived alleles, suggesting that differences in their epigenetic state (i.e., methylation) underlie the observed quantitative variation. Because the variation was stable, heritable, and likely epigenetic, it raises the possibility that quantitative variation in populations may be controlled in part in an epigenetic manner. McWhirter and colleagues speculate that this phenomenon may occur at genes throughout the genome. This mechanism, if it is found to be general, will require the extension of quantitative genetic models to accurately describe the inheritance of quantitative variation.

There remains much to learn about the fine structure of the maize genome and the extent to which the structure varies among regions. Pioneering work from Jeff Bennetzen's laboratory has indicated that the maize genome has its genes scattered between long runs of retrotransposons, the latter making up 60% or more of the genome (San Miguel et al., 1996). Bennetzen's work confirmed data from DNA reassociation kinetics that also indicated that in maize, the genes are like islands in a sea of repetitive DNA (Hake and Walbot, 1980). Although this model certainly still holds, Huihua Fu of Hugo Dooner's group (Rutgers University, Piscataway, NJ) found a striking exception to the rule. In a 32-kb stretch surrounding the bronze locus on chromosome 9, there are 10 genes without any intervening retrotransposons. The average distance between transcripts was a mere 1 kb. Dooner's group speculates that there may be a connection between the high gene content of this region and the fact that bronze has the highest known recombination rate for a maize gene. Their observation also raises a number of questions about the structure and history of the maize genome. How frequent are such regions of high gene density? During the expansion of the retrotransposon families, why did the bronze region escape being infiltrated by retrotransposons, or were they once there and subsequently lost?

An interesting question for a large genome plant such as maize is: how far upstream of a coding sequence can regulator elements be found? Generally, in plants, all elements necessary for proper tissue-specific expression are located within a few kilobases of the coding sequence, and because Arabidopsis has one gene approximately every 5 kb, it would seem that regulatory elements are usually nearby in plants. In this context, Maike Stam from Vicki Chandler's laboratory (University of Arizona, Tucson) reported the mapping of both an enhancer and sequences for paramutation of the b1 locus between 90 and 103 kb upstream of the coding sequence. This observation suggests that there is long-range communication between these distant regulatory elements and the proximal promoter. If such distant regulatory elements are a common feature in maize and other large genome plant species, they will pose a special challenge to molecular quantitative geneticists, because it is a clear possibility that such distant enhancers may figure prominently in the subtle quantitative effects that contribute to natural variation and evolutionary change.

SEE YOU NEXT YEAR

Next year, the Maize Genetics Conference moves south to Orlando, Florida, for a warmer venue. The Maize Steering Committee has reserved extra rooms, anticipating that many will bring their families along for a premeeting or postmeeting trip to Disney World. With the anticipated record-breaking attendance, it would be best to register early.

Acknowledgments

I thank Richard Clark for comments and many of the speakers for clarifying parts of their presentations for me on short notice.

References

Alonso-Blanco, C., and Koornneef, M. (2000). Naturally occurring variation in Arabidopsis: An under exploited resource for plant genetics. Trends Plant Sci. 5, 22–29.[CrossRef][ISI][Medline]

Cardon, L., and Bell, J. (2001). Association study designs for complex diseases. Natl. Rev. Genet. 2, 91–99.

Eyre-Walker, A., and Keightley, P. (1999). High genomic deleterious mutation rates in hominids. Nature 397, 344–347.[CrossRef][Medline]

Hake, S., and Walbot, V. (1980). The genome of Zea mays, its organization and homology to related grasses. Chromosoma 79, 251–270.[CrossRef][ISI]

Muller, H. (1967). The genetic material as the initiator and the organizing basis of life. In Heritage from Mendel, R.A. Brink, ed (Madison, WI: University of Wisconsin Press), pp. 419–447.

San Miguel, P., Tikhonov, A., Jin, Y.-K., Motchoulskaia, N., Zakharov, D., Melake-Berhan, A., Springer, P.S., Edwards, K.J., Lee, M., Avramova, Z., and Bennetzen, J.L. (1996). Nested retrotransposons in the intergenic regions of the maize genome. Science 274, 765–768.[Abstract/Free Full Text]

Venter, J.C., et al. (2001). The sequence of the human genome. Science 291, 1304–1351.[Abstract/Free Full Text]

Yano, M., Katayose, Y., Ashikari, M., Yamanouchi, U., Monna, L., Fuse, T., Baba, T., Yamamoto, K., Umehara, Y., Nagamura, Y., and Sasaki, T. (2000). Hd1, a major photoperiod sensitivity quantitative trait locus in rice, is closely related to the Arabidopsis flowering time gene Constans. Plant Cell 12, 2473–2483.[Abstract/Free Full Text]

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Doebley, J. Search for Related Content PubMed PubMed Citation Articles by Doebley, J. Agricola Articles by Doebley, J.