Avoiding Project Bankruptcy While Effectively Employing Markers

Tom Blake¹, Steve Larson², Joy Eckhoff³ and Vladimir Kanazin¹

¹ Department of Plant, Soil and Environmental Sciences

Montana State University, Bozeman, MT 59717

² USDA-ARS, Small Grains Germplasm Center

Box 307, Aberdeen, Idaho 83210

³ Eastern Agricultural Research Center

Montana Agricultural Experiment Station

P.O. Box 1350 Sidney, MT 59270

Abstract

Most university plant breeding projects share similarities with small businesses. We have customers, our producers and funding agencies, and we have spheres of interest which are largely defined by the needs of our customers. Unlike our more basic research colleagues, "it's interesting" is not sufficient justification for a project. Our projects must also provide practical benefits to crop producers, while remaining sufficiently topical to attract grant funds and contribute to our field.

Like a small business, we have enormous freedom to explore our working environment, and to find niches which are too small or specialized to be noticed by larger organizations. Many of these niches can be both lucrative to producers and fundamentally interesting. This paper will discuss and describe how my project is addressing the needs of my primary customers, the small grains producers of Montana, while helping to develop the tools of barley genetics for my secondary customers, research funding agencies.

General Characteristics of the Small Breeding Project

The economy of Montana is built upon agriculture- primarily livestock, wheat (Triticum aestivum) and barley (Hordeum vulgare) production. Our production system is an old one, built upon a crop/fallow rotation across much of the state. The intermountain valleys are enormously productive, although late season rains often damage grain quality. The eastern slope of the Rockies generally produces higher quality barley and wheat, with drought our greatest yield-limiting factor. Dark Northern Spring Wheat and 2-rowed malting barley are generally the most profitable crops in the state, while the high volume commodities produced in the state are hard red winter wheat and feed barley.

The job of the state barley breeder is to identify the most lucrative barley varieties which farmers can produce, and to develop varieties which further improve production profitability. Our research resources include a seven-farm research farm system staffed by competent agronomists, a molecular genetics laboratory, a state-of-the-art greenhouse facility, and a conservative but competent farming community consisting of a few thousand large land holders. Annually about 1.4 million acres are seeded to barley in Montana, with about 60% of the acreage seeded to varieties recommended as malting barley varieties by the American Malting Barley Association. Approximately 7 million acres are annually seeded to either winter or spring wheat.

Problems and Opportunities in Montana

Although barley has a large and well-explored primary germplasm pool each of the germplasm groups currently in production represents a narrow and specialized subgroup (Martin et al., 1991; Hayes et al., 1997). The upper Midwest has historically produced 6-rowed malting barley used predominantly for domestic malt. Montana and Idaho have been the primary sources of 2-rowed malting barley in the U.S. The reasons underlying the germplasm group differences are largely historical, although breeders have taken advantage of available genetic resources to create modified germplasm groups with excellent regional adaptation. The 2-rowed germplasm group descended largely from 'Betzes' and Betzes relatives shows excellent malting quality and good drought tolerance in the Pacific Northwest, while the Manchuria-derived 6-rowed germplasm group developed by Don Rasmussen produces excellent yields and superb quality grain in the upper midwest. Unfortunately for Montana producers, the 'Manchuria' -derived varieties which perform so admirably in the Red River Valley suffer from the short, dry growing conditions which predominate throughout Montana. The challenge of adjusting the adaptational characteristics of a germplasm group to fit a substantially different production environment, while maintaining the quality characteristics that define the value of the group provides impetus to an applied geneticist.

The recent increase in scab infestation in the upper Midwest has proven to be a significant production problem for U.S and Canadian wheat and barley producers, and the introduction of a notably better-yielding feed barley variety from Germany (Baronesse) has placed new emphasis on yield potential in the Pacific Northwest. These sorts of germplasm limitations generate the conditions needed to make genome research a productive addition to a small breeding program. When these events are combined with the current global scarcity of food- and feed-grains, the opportunities for the small breeding organization to make a significant contribution appears enormous.

Botstein et al.(1980) proposed the development of restriction fragment length polymorphism-based linkage maps in H. sapiens. Although his data were limited, sufficient preliminary information was derived by 1987 (Blake and Kleinhofs, 1988) to rationally propose the same venture in barley. The North American Barley Genome Mapping Project (NABGMP) grew out of a broadly-based collaborative effort to create a practically worthwhile linkage map in barley, and to identify the location of agronomically significant genes which exhibit allelic variation. The first population studied involved a cross between the varieties 'Morex' (a malting barley line developed by Don Rasmussen) and 'Steptoe', a variety developed by Bob Nilan with excellent Western adaptation but notoriously poor grain quality. Pat Hayes (Oregon State University) derived the 150 doubled haploid lines utilized for map construction and QTL analysis, and now over 40 locations have provided data on the performance of these lines (in Graingenes: http://wheat.usda.pw.gov under several headings). The Montana barley genetics program remains a small, but significant provider of information to the NABGMP. In exchange for the relatively small service of replicated field trials and a small portion of the markers on the map, we receive access to the information generated by the NABGMP.

Blake's First Rule

• If somebody else wants to do your work for you, let them and thank them.

I gratefully acknowledge the contributions of my colleagues who participate in the NABGMP. From the Montana perspective, the most crucial information produced by the NABGMP was the elucidation of the genetics underlying the poor adaptation of Morex to Montana environments. (Kleinhofs et al., 1993; Hayes et al., 1994, Larson et al., 1996). Hopeful crosses between Steptoe (and other Coast-class barleys) and Morex (and more recent products of the Minnesota barley improvement program) have been made by myself and others since the early 1970's to little effect. While recovering yield potential is not terribly difficult, recovering yield potential and malting quality in Western environments has not yet been achieved.

Fig. 1. Barley Chromosome 3 from the Steptoe/Morex population. The upper QTL tracing shows the mean response over seven location years in Montana. The lower trace shows the response from one Oregon location. Genotype x Environment interaction is key to the expression of this yield gene.

The initial results of the Steptoe/Morex QTL analysis demonstrated that one or two genes reside on chromosome 3 which have a profound impact on local adaptation. (fig. 1). As these QTL scans demonstrate, one major (and perhaps one lesser) gene on chromosome 3 have an impact on yield in arid Western environments, while having no discernable impact on productivity in Minnesota, North Dakota, or Corvallis, Oregon. Although many characters were evaluated at many locations, the chromosome 3 yield effect showed neither a negative pleiotropic interaction nor a negative linkage effect with genes modifying any other measured character. This appears to be a relatively 'safe' gene to substitute. Molecular marker analysis, when coupled with relatively low cost yield trial evaluation, can help identify genes associated with location-specific (GxE) interactions. Larson et al. (1996, 1997) demonstrated the value of this single gene substitution

It had become apparent early in development of the NABGMP maps that RFLP analysis was too slow and skill-demanding to be a useful breeding tool in my small plant improvement program. A graduate student accepted the task of developing PCR derivatives of appropriate RFLP markers which might prove informative and straightforward to use in a backcrossing context. Steve
Larson developed the markers, and began the process of moving the Ras (rachis stability) gene from the Steptoe background into the Morex background. The next two tables show the result of this activity. Over five location*years replicated trials, backcross lines carrying the Steptoe allele outperformed those carrying the Morex allele at Ras by over 12% (Larson et al., 1996). The Great Western Malting Company performed malt analyses on these lines and found no significant interactions between the RasA gene and any of the characteristics significant to malting or brewing (Larson et al., 1997).

Our conclusion from these experiments is that capturing a large portion of the agronomic value of Steptoe in a Morex background may be accomplished relatively simply. Whether Montana can become a reasonable alternative to Minnesota, North Dakota and Saskatchewan in the 6-rowed malting barley market remains to be seen. However, from the breeder/geneticist's perspective, the potential is enticing.

Solving the Baronesse Problem

In 1991 a we recommended that Montana barley producers grow a new 2-rowed feed barley developed by 'Nordstadt', a German plant breeding company. A local plant breeding company, Western Plant Breeders, arranged to market this variety. It continues to be a remarkable success. This variety is every breeders' dream, superb yield potential and broad adaptation. Grain quality is a problem, especially with respect to malting, but its feedlot performance is adequate and testweight is generally high. As a consequence, this variety has spread across the Pacific Northwest, and it has the potential to become the predominant regional variety.

In addition to being essentially unmaltable, this variety has defects associated with drought tolerance. Having been introduced in the relatively wet years of 1992 to 1995, it remained largely unexposed to traditional dryland stress. The past cropping season of 1996, while an above average production year statewide, was erratic in terms of statewide moisture distribution. Several locations demonstrated Baronesse's limitations relative to yield stability.

Table 1.

1987-1996 SPRING BARLEY MULTI-YEAR OVERALL SUMMARY REPORT

Table 2. Yield results from 1996 Northern Ag. Research Center

Intrastate Barley Yield Trial.

The Technical Problem and its Solution

Baronesse derives from essentially the same germplasm pool (the Northern European Hannchen group) as do all of our Pacific Northwest 2-rowed varieties. While half of our STS markers (Blake et al., 1996) show polymorphism in the Steptoe/Morex population, only 10 produced reliable polymorphisms in a cross between Lewis, a 2-rowed well-adapted Montana product and Baronesse. We developed an RIL population in the hope that technical innovations would provide informative, useful markers.

Performance data were gathered from this RIL population in 1995 with and without supplemental irrigation at Bozeman, MT and in 1996 with supplemental irrigation at Bozeman and Sidney, MT and without supplemental irrigation at Huntley and Bozeman, MT. Data from these experiments are available at our website (http://hordeum.oscs.montana.edu).

AFLP

Amplified Fragment Length Polymorphism analysis was developed and patented by Keygene (Vos et al.l, 1995). This technique utilizes an ingenious approach to scan the entire genome for restriction fragment length polymorphisms. While only about 10% of our STSs showed polymorphisms between Lewis and Baronesse, we were able to read on average 9 AFLP polymorphisms per gel. In two weeks we built a 116 point linkage map which we anchored to six of the seven barley chromosomes using our STS markers. Although achieving our initial objective of identifying the location genes impacting yield, when examined closely the map appeared to miss significant regions of the genome. Although seven linkage groups were obtained, four were near 100cM, half the size of the linkage groups described by Kleinhofs et al.(1993). A colleague (Dr. Charles Stuber) suggested that this might be the consequence of the pervasive problem of clustering of AFLP markers.

Messeguer et al. (1991) found that in tomato unmethylated CG sites tended to be concentrated in high recombination frequency chromosomal regions, while their ^mCG counterparts tented to be more frequently found in recombinationally suppressed regions of the genome. Timmermans et al. (1996) further observed that recombination junctions cloned from maize were free of cytosine methylation. More recently, Powell et al. (1997) reported that methylation sensitive Pst1, when used with Mse1 in AFLP analysis provided markers in the recombinationally active, distal regions of barley chromosomes, while the methylation insensitive EcoR1/Mse1 restriction endonuclease combination provided a high frequency of markers clustered near the barley centromere.

The 'standard' AFLP kit (marketed by several providers, including Gibco and Perkin-Elmer) utilizes the restriction endonucleases EcoR1 and MseI to generate genomic restriction fragments which are then ligated to synthetic linkers. The linker sequences provide target sites for PCR amplification, and the complexity of the amplified products is limited to a useful level through the use of a number of 'selective' bases, in a fashion analogous to that used in differential display. Neither EcoRI nor MseI are particularly sensitive to target site methylation. MspI and HpaII provide an isoschizomer pair which differ in their ability to restrict methylated CCGG sequences. MspII restricts C^mCGG while HpaII does not. In a companion paper (Kanazin et al., submitted) we demonstrate that C^mCGG sites are most commonly found in the recombinationally suppressed regions surrounding the centromeres of barley chromosomes, while unmethylated CCGG sequences tend to be found in recombinationally active chromosomal regions. This can be easily observed in figure 2. Addition of HpaII restriction site polymorphisms to the Lewis/Baronesse map expanded our linkage groups and improved the fit of STS anchors to their previously described positions.

Results of the Lewis/Baronesse Experiment

This relatively simple field experiment, grown at 6 locations over two years led to an improved understanding of the genetic differences delimiting our current working germplasm group and the remarkable introduction, Baronesse. If accurate, these results suggest that approximately 80% of the heritable variance for yield can be attributed to the action of genes or gene clusters on barley chromosomes 2,3 and 6. Backcrossing is now underway to move these genes into genetic backgrounds with improved malt potential, and enhanced feedlot performance.

Fig. 2. Chromosome 2 yield QTL. The chromosome is anchored at ABG602 and TB33/34, and the presumed centromeric cluster of unmethylated MspI restriction site polymorphisms is indicated by 'C'. The converted sequence-tagged-site is labeled STSCAC, and the location of the AFLP polymorphism from which it derived is the adjacent ACCACB1. Markers whose names begin with 'H' are derived from methylation-sensitive HpaII restriction site polymorphisms.

AFLP analysis per se did not in our hands seem to be an ideal technology for single gene introgression. Sequencing gels and radioisotopic detection make this technique time-consuming and expensive. However, we found that cloning specific AFLP products was relatively simple (Kanazin et al., submitted), and conversion isolated, cloned AFLP bands to their single locus PCR derivatives was relatively straightforward. One of the AFLP markers in the chromosome 2 'yield' QTL region was cloned, sequenced, and converted to its single locus PCR counterpart. This primer pair was then used in segregation analysis, and given the name STSCACB1. One apparent recombinant was observed in the Lewis/Baronesse population between the AFLP marker from which STSCACB1 derived and the sequence-tagged-site derivative. I attribute that difference to error, although other explanations are also reasonable. To determine whether this marker would be productive in crosses between Pacific Northwest germplasm and Baronesse, we tested 60 lines derived from a cross between Baronesse and Clark (CI15657), which were also grown in a 2 replication randomized complete block design at the A.H. Post Research Farm, near Bozeman, in 1996. The polymorphism detected in the Lewis/Baronesse population was also observed to segregate in this population, and an analysis of yield*STSCACB1 phenotype was performed. Lines carrying the Baronesse allele showed on average a 350 Kg/ha yield advantage over those carrying the 'Clark' allele (table 3). This result was very similar to that observed in the more thoroughly mapped Lewis/Baronesse population. This suggests that AFLP analysis, when done well, can provide adequate maps even in relatively narrow germplasm pools, and that when a marker is found to be linked to an important QTL, marker conversion is a reasonable objective.

Table 3. The Effect of QTL linked to STSCACB1 on Yield and Plant Morphology in a population derived from a cross between 'Clark' and 'Baronesse'.

Marker Technology, Heterosis and Small Research Projects

Many of the fundamental concepts of genetics are meaningless outside of context. Heterosis is only a meaningful term when use to describe the genetics underlying trait expression within a particular germplasm. Our initial experiments have failed to show that there exists an optimum barley genotype, but rather that several very productive barley genotypes exist, each of which performs differently in different environments. Optimizing economic and agronomic performance within environments seems more practically useful than attempting to identify the ultimate genotype.

Heterosis, per se, is either the result of dominant genes linked in repulsion or overdominance. While we in the inbred small grains have little opportunity to observe true overdominance, repulsion linkages are commonly observed. In the case of the Lewis/Baronesse population, an obvious replusion linkage between a gene modifying head emergence and a gene affecting yield had an important impact on my breeding program. Early generation selection for head emergence is easy and reliable. In selecting for this highly heritable, observable character, we generally selected against a desirable yield factor. Now that markers are available which permit characterization of lines which carry genes for optimal yield, we can identify recombinants with nearly ideal plant height and peduncle extension.

When we began the quest for markers appropriate for marker-assisted-selection, our first need was informative markers. Several technologies proved useful, although RFLPs remained the dominant map-making tools throughout the early 1990s. Single locus PCR technologies have proven useful, although any technique which permits only one (or a handful) of loci to be scored per day is both tedious and expensive. AFLP analysis provides a multilocus approach to map construction while still utilizing restriction site polymorphisms as its source of variation. Through careful selection of the types of restriction sites surveyed, it is likely that recombinationally normalized maps may be developed in many species. It appears as though barley is among these fortunate organisms.

As relatively inexpensive technologies like AFLP analysis become popularized across institutions and crops, questions regarding genetic mechanisms within specific contexts will be more commonly addressed, and more frequently answered. The next decade should be a lot of fun.

Literature Cited

Blake TK,D. Kadyrzhanova, K.W. Shepherd , A.K.M.R. Islam, P.L.Langridge, C.L. McDonald , J. Erpelding, S. Larson , N.K. Blake , L.E. Talbert . (1996). STS-PCR markers appropriate for wheat-barley introgression. Theor. Appl. Genet. 93:826-832

Blake, T. K., A. Kleinhofs. 1988. Applications of molecular biology to crop improvement. p. 991-999. In R. J. Summerfield (ed.) World Crops: Cool season food legumes. Martinus Nijhoff, Dordrecht, Netherlands.

Blake TK, S. Larson, J. Eckhoff , J.Jeung , B. Jones ,V. Kanazin.. The positives and negatives of 'Baronesse' barley. Submitted to JQTL/Crop Science.

Botstein D, R.L. White, M. Skolnick, R.W. Davis. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32:314-331.

P.M. Hayes, B.H. Liu, S.J. Knapp, F. Chen, B. Jones, T. Blake, J. Frankowiak, D. Rasmussen, M. Sorrells, S.E. Ullrich, D. Wesenberg and A. Kleinhofs. 1994. Quantitative trait locus effects and environmental interaction in a sample of North American barley germplasm. Theor. Appl. Genet. 87:397-401

Kanazin V.,J. Jeung , T.K. Blake. (Submitted). Distribution of methylated and unmethylated CCGG sites across the barley genome. Genetics (Submitted).

Kleinhofs, et al. 1993. A saturated medium-density map of the barley genome. Theor. Appl. Genet. 86:705-712.

Kunzel G and L. Korzun. 1996. Physical mapping of cereal chromosomes with special emphasis on barley. Proc. Of the Seventh International Barley Genetics Symposium. 197-206. Eds. Scoles G. and Rossnagel B. Univ. of Saskatchewan, Saskatoon, Sask.

Larson S., C. McDonald ,T.K. Blake. (1996) Evaluation of barley chromosome 3 yield QTL in a backcross F₂ Population using PCR-STS markers. Theor. Appl. Genetics 93:618-625.

Larson S, D. Habernicht, T. Blake and M. Adamson. 1997. Backcross Gains for Six-Rowed Grani and Malt Qualities sith Introgression of a Feed Barley Yield QTL . J. Am. Soc. Brew. Chem. 55(2) 52-57.

Martin, J. M., T. K. Blake, and E. A. Hockett. 1991. Diversity among North American spring barley cultivars based upon coefficients of parentage. Crop Sci. 31(5):1131-1137.

Messeguer R, M.W. Ganal, J.C. Steffens and S.D. Tanksley. 1991. Characterization of the level, target sites and inheritance of cytosine methylation in tomato nuclear DNA. Plant Mol. Bio. 16:753-770

Powell L, W.T.B. Thomas, E. Baird , P. Lawrence , A. Booth , B. Harrower,J.W. McNicol and R. Waugh . 1997. Analysis of quantitative traits in barley by the use of Amplified Fragment Length Polymorphisms. Heredity 79:48-59.

Timmermans MCP, O.P. Das, J. Messing . 1996. Characterization of a Meiotic crossover in maize identified by restriction fragment length polymorphism-based method. Genetics 143:1771-1783.

Vos P,R. Hogers, M. Bleeker,M. Reijans,T. van de Lee, M. Hornes , A. Ftijters,J. Pot,J. Peleman ,M. Kuiper,M. Zabeau . 1995. AFLP: a new technique for DNA fingerprinting. Nucl. Acids Res. 23:4407-4414.