- Email: [email protected]
Chromosome landing: a paradigm for map-based gene cloning in plants with large genomes
REVIEWS
M a n y , if not most, traits of interest to plant biologists and agriculturalists are controlled by genes whose products are unknown. Because of this, researchers have been anxious to devise and perfect gene isolation methods that do not require prior knowledge of the gene product(s). One such method is map-based or positional gene cloning1, 2. Map-based cloning has only two requirements: (1) that individuals within the population have genetically based differences in the trait of interest; and (2) that the gene(s) responsible for this difference can be mapped to a chromosomal position(s) adjacent to segments of DNA that have already been cloned (e.g. RFLP or microsatellite markers). This approach to gene cloning is very attractive, as the basic requirements can, in theory, be met for most traits, including those conditioned by multiple genes. Map-based cloning has been used successfully in a number of instances to isolate genes from plants3--9. Despite these successes, it is far from a routine technique in plant science, owing to the presence of several impediments, which can sometimes be overwhelming. One of these is the sheer amount of DNA in the genomes of most plants: plant species typically have more than 1 billion (109) base pairs of nuclear DNA, and some have genomes that are much larger; for example, the wheat genome has more than 16 billion base pairs 1°. Searching for a specific gene among these billions of base pairs is a daunting task. In the past, the paradigm used for map-based cloning has been one in which 'chromosome walking' played a central, and very time consuming, role (Fig. 1). First, either random molecular markers or established molecular linkage maps were used to conduct a search to localize the gene adjacent to one or more markers. A chromosome walk was then initiated from the closest linked marker and a series of overlapping clones (e.g. cosmids or YACs) isolated. This continued either until one arrived at another molecular marker known to be situated on the opposite side of the target gene, or until there was another indication that the walk had actually passed over the target gene. With this approach, most time and resources are devoted to isolating overlapping clones and preparing the resulting contig, and to searching for the target gene within the contig. Chromosome walking in large genomes is hampered not only by the large amounts of DNA being traversed, but also by the high frequency of repetitive DNA. Moreover, libraries of large inserts (e.g. YAC libraries) used for chromosome walking in large genomes have often contained chimeric clones, or clones that have become rearranged after cloning 11. Recent developments in both technology and analytical procedures are likely to shih the map-based cloning strategy in many species away from one dominated by chromosome walking. Instead, initial efforts will be directed towards selectively identifying molecular markers that are so close physically to the target gene that little or no chromosome walking is required. "l~nese tightly linked markers will be used to screen a genomic library to isolate directly a single clone containing the gene of interest. We refer to this approach as 'chromosome landing' and herein discuss its requirements, applications and limitations. In addition, we
Chromosome landing: a paradigm for map-based gene cloning in plants with large genomes STEVEND. TANKSLEY, MARTINW. GANALAND GREGORYB. MARTIN The original concept behind mop-based or positional cloning was to find a DNA marker linked to a gene of interest, and then to "walk'to the gene via overlapping clones (e.g. cosmids or YACs). Whilechromosome walking is straightforward in organisms with small genomes, it is did~cull to apply in most plant species, which typically have large, complex genomes. The strategy of chromosome walking is based on the assumption that it is d i g i t and time consuming to find DNA markers that are physically close to a gene of interest Recent technological developments invalidate this assumption for maW species. As a resug the mopping paradigm has now changed such that onefirst isolates one or more DNA marker(s) at a physical distance from the targeted gene that is less than the average insert size of the genomic library being used for clone isolatiog The DNA marker is then used to screen the library and isolate (or 7and'on) the clone containing the gene, without any needfor chromosome walking and its associated problems. Chromosome landin~ together with the technology that has made it possible, is likely to become the moin strategy by which mop-based cloning is applied to isolate both major genes and genes underlying quantitative traits in plant species. provide an example of how the method was recently used to isolate a gene involved in disease resistance in tomato, a plant with a relatively large genome.
Search for marker(s) tightly linked to a target gene When map-based cloning was fast proposed for use in species with large genomes, molecular linkage mapping was in its infancy and molecular maps were not available for most plant species. It could take months or even years to find a single molecular marker linked to a target gene. In the past ten years, the situation has changed dramatically. Molecular linkage maps have now been constructed for more than 30 plant species and work is under way to construct maps for at least as many again. Maps for some species are highly refined, containing hundreds of markers 12a3. The availability of high-density molecular linkage maps provides a ready starting point to search for molecular markers tightly linked to a target gene. For example, the genetic map of tomato contains more than 1000 DNA markers. Assuming a genome size of 1000 megabases, any gene is expected to lie, on average, within 1000 kb of a molecular marker 14.
Additional helpfrom comparativegenetic linkage maps Genomes of related species have evolved from a common ancestral genome and thus share portions of
TIG FEBRUARY1995 VOL. 11 No. 2 © 1995 ElsevierScienceLtd 0168 - 9525/95/$09.50
6 3
REVIEWS
(a)
O~ ¢.. ¢9
m
(b)
m
t-
I II
III I t
I I
I
I
III l
II
I 11
potato and pepper) and the mustard family22,23 (Arabidopsis, cabbage, cauliflower and rape). As regards map-based gene cloning, comparative maps allow researchers working on one species to identify and access probes from orthologous regions of related genomes. For some species (e.g. the grasses), comparative maps have expanded the repertoire of available mapped probes from a few hundred to several thousand. While high-density maps and comparative maps can provide some of the linked markers needed for chromosome landing in plants, it is clear that additional methods are needed by which new markers can be isolated selectively from specific regions of a genome.
Regional targeting witb bigb-vohtme marker tecbnologv
The probability of identifying one or more markers within a specified physical distance of a target gene decreases the closer one gets to the target. Ideally, one would like to idents,,. SS ~ ify a marker at a physical distance from a tar~ ~ | get that is less than the average insert size of A B C Candidate genes the genomic library from which one expects to isolate the gene. For example, working Fmu~ 1. Comparison of strategies for gene mapping using (a) chromosome with a YAC library with an average insert size walking and (b) chromosome landing. Blue boxes represent chromosomal of 200 kb, it would be desirable to identify at regions that contain a gene targeted for map-based cloning; yellow boxes least one marker within a 400 kb window represent overlapping genomic clones corresponding to the chromosomal containing the target (200 kb on each side). If region containing the target gene. Arrows indicate positions of linked DNA such a marker were to be identified, it wofild markers. Red boxes represent candidate genes within a genomic be possible to screen the YAC library with the clone that is known to contain a target gene. marker and to directly isolate a single YAC that contains the targeted gene withom any their linkage maps that are conserved in both gene conneed for chromosome walking. However, for species tent and order. The advent of molecular genome mapwith large genomes, the probability of identifying a r2nping has made it possible to determine the relative dom marker within 200 kb of a target is very low. For order of genes among related species that are no longer example, in a genome containing 1 billion base pairs, sexually compatible. Extensive comparative maps have the probability would be 4 x 102/106 = 4 × 104. To been constructed for a number of plant taxa, including have a 90% chance of identifying at least one marker species in the grass family15-19 (rice, maize, sorghum, within a 400 kb interval, ~ 6000 markers would need to barley and wheat), the nightshade family14,20,21 (tomato, be sampled (Fig. 2). It is not likely that many plant species will have molecular linkage maps of this density in the foreseeable future. Recently, several strategies have been developed 1.0J , _...... .................................. that allow one to screen a large number of random, ~¢.~ t ..-'"" .,~ -'" " " unmapped molecular markers in a relatively short time E ~ 0.8 ....." ~.~" and to select just those few markers that reside in the vicinity of the target gene. These methods rely on two principles: (1) the development of high-volume marker technology, which allows hundreds or even thousands o o~ 0.4. / / ,. of potentially polymorphic DNA segments to be gener"6 ,_= 100 Mb (Arabidopsis) ated and visualized rapidly from single preparations of ~'i / 400 Mb (Rice) DNA; and (2) use of genetic stocks to identify, among =~ ~ 0.2 -"/'/ ~/ 1000 Mb (Tomato) these thousands of DNA fragments, those few that are .o "_~ : 3000 Mb (Maize) ¢L derived from a region adjacent to the targeted gene. 0,0 High-volume marker technologies that have demon0 50'00 10()00 15000 20000 strable efficacy are random amplified polymorphic No. of markers DNAs24 (RAPDs), amplified fragment length polymorphisms 25 (AFLPs), RFLP subtraction x6 and repreFmuaE 2. Graph showing the number of random markers that sentational difference analysis27 (RDA). RDA and RFLP must be sampled to obtain at least one marker less than 200 kb subtraction were described recently by researchers from a given target gene. Calculations are for four genomic sizes. working in mammalian genetics and use DNA-DNA
li/
• . . . . . . . . . . . . . .
TIG FEBRUARY1995 VOL. 11 NO. 2
64
REVIEWS
Box 1. Nearly isogenic ILnes and bulked segregant analysis (A) Nearly isogenic lines Nearly isogenic lines (NILs) are created when breeders cross a donor line P1 to a recipient line P2. The resulting F1 hybrid is then crossed back to the P2 recipient to produce the backcross 1 generation (BC1). From BC1, a single individual that contains the dominant allele of the target gene from P1 is selected. Selection for the target gene is normally made on the basis of phenotype. This BC1 individual is again backcrossed to P2, and the q'cle of backcrossselection is repeated for a number of generations. In the BC7 generation, most, if not all, of the genome will be derived from P2, except for a small chromosomal segment containing the selected dominant allele, which is derived from Pi. Lines homozygous for the target gene can be selected from the BC7 F2 generation. The homozygous BC7F2 line is said to be nearly isogenic with the recipient parent, P2. Such pairs of nearly isogenic lines (NILs) are frequently generated by plant breeders as fi~ey transfer major genes between varieties by backcross breeding.
(a)
P1
P2
F1
BC1
BC7F2 (NIL)
Tar e,_ilXII:IB. . . . . .
II
gene
(b)
P1
P2
li ' 11 ,arne, it.t1-1i--;t-'-2--iI--I1-II-ll-BI' Bulk I (c)
P1
Bulk 2 P2
NIL
Bulk 1
Bulk 2
(B) Bulked segregants analysis Bulked segregant analysis (BSA) requires the generation of populations of bulked segregants (bulks). When Pl and Marker P2 are hybridized, the F2 generation analysis derived from the cross will segregate for alleles from both parents at all loci throughout the genome. If the F2 population is divided into two pools of contrasting individuals on the basis of screening at a single target locus, these two pools (Bulk 1 and Bulk 2) will differ in their allelic content only at loci conrained in the chromosomal region close to the target gene. Bulk 1 individuals, selected for the recessive phenotype, will contain only P2 alleles near the target, while Bulk 2 plants (selected L, ,.he dominant phenotype) will contain alleles from both P1 and P2. Bulk 1 and Bulk 2 will both contain alleles from both P1 and P2 at loci unlJnked to the target.
(C) High-volume marker techniques High-volume marker techniques, such as those based on RAPDs and AFLPs, can be used to compare the genetic profiles of pairs of NILs or pairs of bulks to search for DNA markers near the target gene. DNA fragment(s) (indicated by arrows) that are observed in the NIL and P1 populations, but not in P2, or that are unique to either Bulk 1 or Bulk 2, should be derived from a locus tightly linked to the target gene. hybridization in solution to selectively enrich for portions of two genomic samples that differ in sequence 26.27. Currently, application of the use of RDA or RFLP subtraction in plants has not been reported, but it seems likely that these techniques could be used in combination with the strategies outlined below (see also Box 1). RAPDs, and more recently AFLPs, have been used with considerable success in plants species. Like RDA and RFLP subtraction, these methods allow a global comparison of genomes for genetic differences. Both RAPDs and AFLPs rely on the polymerase chain reaction (PCR) used in conjunction with primers that amplify
random sites throughout the genome. Amplification products are analysed on agarose or acrylamide gels to allow visualization of variation among individuals (see Box 1). By using one or mo~ of these high-volume marker technologies, thousands of loci scattered throughout the genome can be assayed in a matter of weeks or months. The next problem is to determine which of the amplified loci lie near the targeted gene. In the past few years, two strategies have proven effective for identifying from a large number of markers the few that reside near a targeted locus. Both involve the use of genetic stocks that are (almost) genetically identical, except in the regions flanking the targeted
TIG FEBRUARY1995 VOL. 11 NO. 2
65
REVIEWS
gene (Box 1). The in'st strategy, which uses nearly isogenic lines and is referred to hereafter as the NIL strategy, can be applied only where breeders or geneticists have generated inbred lines that differ at the targeted locus zS,zg. Generation of such stocks can take years but, fortunately for plant geneticists, many such stocks have been routinely generated as a byproduct of breeding. The second, and more generally applicable, strategy has been referred to as the bulked segregant analysis (BSA) method and relies on the use of segregating populations30,31. Both these methods, used in conjunction with high-volume DNA marker technology, permit one to screen thousands of loci and to selectively identify those adjacent to the gene of interest. Moreover, this can be accomplished without having a genetic map for the species. The efficacies of the NIL and BSA approaches have been demonstrated in plants 28-'34, and more recently these approaches have also proven effective in gene targeting studies in mammals35.
High resolution mapping High density molecular maps, comparative maps and regional targeting strategies permit relatively rapid identification of many markers that flank a targeted gene, with markers usually lying within a few cM of the gene. This creates the dilemma of determining which marker(s) lies closest to the target and should thus be used to screen a genomic library to isolate it. For example, if 6000 markers (see above) were screened using the NIL or BSA strategy, it is likely that 50 or more markers would be selected as being within 10 cM of the target. To determine which of these is actually closest to the gene, high-resolution genetic mapping is required.
Genetic mapping Genetic mapping depends on meiotic crossovers to orient loci along a chromosome. In a genome of 1 billion base pairs and 1000 cM, 1 cM is, on average, 1 million base pairs. To localize markers within distances of 100 kb or less means mapping to a resolution of 0,1 cM resolution or less. This would require analysis of thousands of segregating progeny in the hope of obtaining those few individuals in which crossovers had occurred very close to the target gene. For example, to have a 95% chance of recovering a single crossover less than 0.1 cM from a target gene, analysis of more than 3000 meiotic products would be required. While PCR-based marker technology and automation is making analysis of larger populations more practical, this is still a laborious task and one of the rate-limiting steps in map-based gene cloning. Below, we describe two strategies for improving the efficiency of highresolution mapping; further developments in this area are expected in the future.
Analysis offlanking markers In segregating populations of a modest size (e.g. 100 meiotic events), a target gene can usually be placed within an interval of 5-10 cM with a high degree of certainty. The markers defining this interval can then be used to screen a larger segregating population and identify individuals derived from one or more gamete(s) containing a crossover in the given interval. Only such individuals are useful in orienting other
markers closer to the target gene. Once identified, these individuals can be analysed in relation to all molecular markers within the region to identify those closest to the target. The flanking marker approach was recently used to refine the map position of the Cf-2gene, which encodes resistance to Cladosporiumfidvum in tomato, to a 40 kb region between two DNA markers 36. Flanking marker analysis is especially powerful ~, ".en the defined interval is small, for example, <10 cM, and the flanking markers are easy to score. Ideally, flanking markers could be screened without the need for DNA isolation or other time consuming activities. Morphological markers, which are common in most plants .and have in several instances been placed on existing molecular maps, are ideal. Unfortunately, the number of morphological markers available is limited and the probability that a specific target gene will lie within a small interval defined by two useful morphological markers is quite low. If the target gene is flanked by molecular markers, it then becomes necessary to extract DNA from all segregating individuals, unless PCR can be perforated directly on intact plant tissue57,38.
Pooled-sample mapping Pooled-sample mapping is another approach for selectively identifying those rare individuals in a large population that have a crossover near a target gene. This method does not rely on localization of the target gene to an interval between two flanking markers; instead, it requires either that the genotype of the target gene can be identified accurately on a large scale, or that there is a tightly linked marker that can be screened for on a large scale. Individuals from the large population are 'pooled' for DNA extraction and marker analysis and the pooled DNA subjected to analysis with all molecular markers in the vicinity of the target gene. On the basis of the proportion of the pools that contain a crossover with respect to the markers analysed, the marker(s) closest to the target can be identified and the order of all markers in the vicinity of the target can be determined. In theory, this method allows the screening of thousands of gametes for informative crossovers with a minimum of DNA extraction and molecular analysis. The pooled-sample strategy has been demonstrated in the fine mapping of a gene that regulates fruit ripening in tomato39.
Physical mapping High-resolution mapping can be used to identify the marker(s) closest (in terms of genetic recombination) to a gene targeted for cloning. Unfortunately, units of genetic recombination (cM) cannot be directly translated into distances in base pairs. For example, in tomato, 1 cM is expected to correspond to approximately 700 kb (Ref. 14); however, physical mapping of selected regions of the tomato genome has shown that 1 cM can correspond to physical distances that vary by up to two orders of magnitude40, 41. In the regions around centromeres and certain telomeres, in particular, meiotic recombination is suppressed, resulting in much higher kb/cM values TM. For chromosome landing to be successful, the marker used for library screening must lie within a physical distance of the target gene that is less than the average insert size of
TIG FEBRUARY1995 VOL. 11 NO. 2
66
I~.EV1[ E W S
the genomic library being screened. Unfortunately, the great variation in meiotic recombination along a chromosome makes it necessary to determine empirically the approximate kb/cM value within the targeted region. In plants, this has been accomplished most effectively through the use of longrange restriction mapping, using rare-cutting restriction enzymes and pulsed-field gel electrophoresis in conjunction with DNA markers that are closely linked to the target gene 4°--42. Long-range physical mapping can also help confirm the order of DNA markers in the vicinity of the target.
Identifying the target gene in a large genomic clone
t , .
CD~106 Markers
.CD127
TGS0;
I
I Centimorsans
10.2
I 1.4
I~
F
0.4
OA
<0.2
Identification of the target gene from a number of candidate genes encompassed by a large-insert FIGURE3. Chromosome landing in the Pto region of tomato. A 12 cM region of clone is the final daunting step in chromosome 5 containing the Progene and flanked by markers TG475 and TG504 is shown (bottom). A single YACclone was isolated using the marker TG538, which had map-based cloning. Although the been determined by genetic studies to be closest to Pto. The ends of the YACclone were density of genes will vary through- isolated and geneticallymapped (PTY538-1Land PTY638-1R)and shown to flank the out the genome, a 300--400 kb YAC Pto gene. Two candidate cDNAclones from the region (CD106and CD127)were isolated clone can easily encompass more using the YACas a probe. Autoradiographsof survey filtersof genomic DNA from than 30 genes. Identification of can- resistant (left lane of each pair) and susceptible(right lane) tomato plants digested with didate genes can be accomplished six restrictionenzymes and probed with CD106and CD127are shown (top). Note that by probing cDNA libraries with CD127 detects high levels of polymorphismbetween resistant and susceptible plants; either DNA from the whole YAC transformation-complementationexperimentslater showed that it was a member of the or with DNA from phage h or Pto resistance gene family. cosmid clones from a contig spanning the YAC. Recently, methods for isolating cDNAs Chromosome landing in plants: isolation of the Pro encoded by large genomic regions have been improved gene The tomato gene Pto confers resistance to the in several ways. The new methods, referred to as 'direct cDNA selection' or 'capture systems' obviate the net d bacterial pathogen Pseudomonas syringae pv. tomato. Recently, this gene was isolated using the chromosome for plaque or colony hybridizations and are likely to significantly ease isolation of candidate genes from landing approach 4. Intensive efforts resulted in the identification of 18 informative markers that were closely large genomic clones known to contain a target gene 43. linked to the Pto gene 45. High-resolution linkage analysis In most cases, numerous cDNA clones will be in more than 1200 F2 plants identified one marker that identified by the above methods. Evidence that a particular cDNA corresponds to the target gene is obtained cosegregated with the gene, and this marker was used by a variety of methods including: (1) high-resolution to isolate YAC clones derived from the target region mapping to demonstrate cosegregation of the candidate (Fig. 3). Genetic mapping of isolated YAC ends showed that one clone spanned the Pro locus. Gel-purified DNA clone with the phenotype; (2) demonstration that the expression pattern of the gene is consistent with the from a YAC clone was radiolabelled and used to probe a cDNA library. Two classes of cDNA clones were identphenotype, for example, that it is transcribed in the ified. Placing the cDNAs on a high-resolution linkage appropriate tissue(s) and/or induced by the appropriate map eliminated one class, as recombination events stimulus; (3) determination of the DNA sequence of the gene and its comparison with those in sequence data- were identified that separated these clones from the disease resistance locus (Fig. 3). However, high-resolution bases to identify any homology with genes of known function; (4) probing mutant lines with the cDNAs to mapping indicated that the second cDNA class cosegreidentify alterations of the DNA or mRNA from the wild gated with the resistant phenotype. Agrobacteriummediated transformation of a susceptible tomato plant type; (5) complementation of the mutant phenotype by transformation with the gene. The last approach, comp- with this candidate cDNA demonstrated that it conlementation, provides the most direct and definitive ferred disease resistance. This work exemplifies the advantages of chromosome landing, in that initial evidence. Genetic complementation by Agrobacteriummediated transformation is routine in many plant emphasis on the isolation of many closely linked DNA markers eliminated the need for chromosome walking species and has been used in several instances to confirm that a candidate gene is responsible for a particular and the development of a high-resolution linkage map expedited the identification of candidate cDNAs. phenotypic trait3--9A4. TIG FEBRUARY1995 VOL. 11 NO. 2
67
REVIEWS
12 O'Brien, S.l., ed. (1993) Genetic Maps, Cold Spring Harbor
Conclusions and future prospects
Laboratory Press
Chromosome landing (and the associated techniques that have made it possible) can be applied in most sexually reproducing species, including most plants and several animals (including many domesticated animals). It minimizes or completely eliminates the need for chromosome walking, and thus greatly reduces the problems engendered by both the volume and repetitive nature of DNA analysis in such genomes. Currently, high-resolution genetic mapping is the most significant hurdle to the widespread application of chromosome landing. As discussed above, several strategic advances have been made in this area and we can reasonably expect that future developments will continue to improve the situation. It is important to point out that there are some species in which mapping by chromosome landing is not likely to play a major role, most notably in humans, where controlled matings are impossible and population sizes are inadequate for high-resolution mapping studies. Nonetheless, certain aspects of chromosome landing, in particular, high-volume marker technology and regional targeting, are beginning to play a role in map-based cloning in mammals 46. For the more intensively studied animals, such as humans and mice, and for plant species such as Arabidopsis, complete genomic contigs made up of YACs or bacterial artificial chromosomes will probably be constructed within the decade. When complete contig maps are available, the need for high-volume marker technology will be reduced. However, it is unlikely that the genomes of most plants, especially those of agronomic importance, will be organized into contigs in the near future. For these species, chromosome landing seems likely to become the main strategy by which map-based or positional cloning is applied to isolate agronomically important and botanically interesting genes whose products are not known. Finally, the past few years have seen numerous publications describing the use of molecular linkage maps to localize genes underlying quantitative traits ( Q ~ ) in plants and animals 47A8. The chromosome landing paradigm outlined here can readily be applied to cloning QTLs that can be mapped with a high degree of resolution. Progress has already been made in highresolution mapping of QTLs in plants, and it seems likely that in the foreseeable future, chromosome landing will begin to yield genes for characters that are quantitatively inherited 49.50.
13 Phillips, R.L. and Vasil, I.K. (1994) DNA-basedMarkers in Plants, IGuwer 14 Tanksley, S.D. etal. (1992) Genetics132, 1141-1160 15 Ahn, S.N. and Tanksley, S.D. (1993) Prec. Natl Acad. Sci. USA 90, 7980-7984 16 Ahn, S.N. et al. (1993) Mol. Gen. Genet. 241,483-490 17 Kurata, N. et al. (1994) Bio/Technology 12, 276-278 18 Whitkus, R. etal. (1992) Genetics 132, 1119-1130 19 Van Deynze, A.E. et al. Genome (in press) 20 Bonierbale, M. et al. (1988) Genetics 120, 1095-1103 21 Prince,J.P. etal. (1993) Genome36, 404--417 22 Kowalski, S.P. et al. (1994) Genetics 138, 499-510 23 Slocum, M.K. (1989)in Development a n d Application of Molecular Markers to Problems in Plant Genetics
24 25 26
27 28 29 30 31 32 33 34 35 36 37 38 39 40
41 42 43 44
45 46
47
References
48 49 50
1 Wicking, E. and WiUiamson, B. (1991) Trends Genet. 7, 288-293 2 Collins, F.S. (1992) Nature Genet. 1, 3--6 3 Arondel, V. et al. (1992) Science 258, 1353-1355 4 Martin, G.B. et al. (1993) Science 262, 1432-1436 5 Mindrinos, M. etaL (1994) Cell78, 1089-1099 6 Giraudat,J.B. etal. (1992)Plant Cell4, 1251-1261 7 Chang, C. et al. (1993) Science 262, 539-570 8 Leung,J. et al. (1994) Science264, 1448--1452 9 Meyer, K., Leube, M.P. and Grill, E. (1994) Science 264, 1452-1455 10 Arumuganathan, K. and Earle, E.D. (1991) PlantMol. Biol. Rep. 9, 208-218 11 Green, E.D. etal. (1991) Cell70, 1059--1068
(Helentjaris, T. and Burr, B., eds), pp. 73-80, Cold Spring Harbor Laboratory Press Williams,J.G.K. et al. (1990) Nucleic Acids Res. 18, 6531-6535 Zabeau, M. (1993) Eur. Pat. Appl. No. 92402629.7 Rosenberg, M., Przybylska, M. and Straus, D. (1994) Proc. Natl Acad. Sci. USA 91, 6113-6117 Lisitsyn,N., Lisitsyn, N. and Wigler, M. (1993) Science 259, 946-951 Young, N.D. et al. (1988) Genetics 120, 579-585 Martin, G.B. et al. (1991) Proc. Natl Acad. Sci. USA 88, 2336-2340 Michelmore, R.W. et al. (1991) P~c. Natl Acad. Sci. USA 88, 9828-9832 Giovanonni, J.J. et al. (1991) Nucleic Acids Res. 19, 6553--6558 Mackill, D.J. et al. (1993) Theor. Appl. Genet. 85, 536-540 Schiiller, C.G. et al. (1992) Theor. Appl. Genet. 84, 330-338 Pineda, O. et al. (1993) Genome36, 152-156 Lisitsyn,N.A. et al. (1994) Nature Genet. 6, 57-63 Dixon, M.S. et al. Mol. Gen. Genet. (in press) Klimyuk,V.I. et al. (1993) PlantJ. 3, 493-494 Chunwongse. j. et al. (1993) Theor. Appl. Genet. 86, 694-698 Churchill, G.A. et al. (1993) Proc. Natl Acad. Sci. USA 90, 16-20 Ganal, M.W. et al. (1989) Mo!. Gen. Genet. 215, 395--400 Segal, G. et al. (1992) Mol. Gen. Genet. 231, 173-185 Ganal, M.W. et al. (1990) AgBiotech News Inf. 2, 835--840 Lovett, M. (1994) Trends Genet. 10, 352-357 Whitham, S. etal. (1994)Cell78, 1101-1115 Martin, G.B. et al. (1993) Mol. Plant Micr. Interact. 6, 26-34 Aldhous, P. (1994) Science265, 2008--2010 Crabbe, J.C. et al. (1994) Science 264, 1715-1723 Tanksley, S.D. (1993) Annu. Rev. Genet. 27, 205-233 Paterson, A.H. et al. (1990) Genetics 124, 735-742 Ritsert,J.C. and Stam, P. (1994) Genetics 136, 1447-1455
S.D. TANgSLEFXS UN THE DEPARTMENTOF PLANT BREEDING AND BIOMETRY, CORN£LL UNIVERSt~, ITHACA, N Y 14853. USA, M.W. GANALIS IN THE INSTITOTEFOR PLANT GENETICS AND CROP PLANT RESEARCH, 0-06466 GATERSLEBEN, GERMANY,, AND G.R MARTIN IS IN THE DEPARTMENT OF AGRONOMY,PuP~U£ UNIVERSITy, WESTLAgFEZT~, I N 47907, USA.
TIG FEBRUARY1995 VOL. 11 NO. 2
68
M a n y , if not most, traits of interest to plant biologists and agriculturalists are controlled by genes whose products are unknown. Because of this, researchers have been anxious to devise and perfect gene isolation methods that do not require prior knowledge of the gene product(s). One such method is map-based or positional gene cloning1, 2. Map-based cloning has only two requirements: (1) that individuals within the population have genetically based differences in the trait of interest; and (2) that the gene(s) responsible for this difference can be mapped to a chromosomal position(s) adjacent to segments of DNA that have already been cloned (e.g. RFLP or microsatellite markers). This approach to gene cloning is very attractive, as the basic requirements can, in theory, be met for most traits, including those conditioned by multiple genes. Map-based cloning has been used successfully in a number of instances to isolate genes from plants3--9. Despite these successes, it is far from a routine technique in plant science, owing to the presence of several impediments, which can sometimes be overwhelming. One of these is the sheer amount of DNA in the genomes of most plants: plant species typically have more than 1 billion (109) base pairs of nuclear DNA, and some have genomes that are much larger; for example, the wheat genome has more than 16 billion base pairs 1°. Searching for a specific gene among these billions of base pairs is a daunting task. In the past, the paradigm used for map-based cloning has been one in which 'chromosome walking' played a central, and very time consuming, role (Fig. 1). First, either random molecular markers or established molecular linkage maps were used to conduct a search to localize the gene adjacent to one or more markers. A chromosome walk was then initiated from the closest linked marker and a series of overlapping clones (e.g. cosmids or YACs) isolated. This continued either until one arrived at another molecular marker known to be situated on the opposite side of the target gene, or until there was another indication that the walk had actually passed over the target gene. With this approach, most time and resources are devoted to isolating overlapping clones and preparing the resulting contig, and to searching for the target gene within the contig. Chromosome walking in large genomes is hampered not only by the large amounts of DNA being traversed, but also by the high frequency of repetitive DNA. Moreover, libraries of large inserts (e.g. YAC libraries) used for chromosome walking in large genomes have often contained chimeric clones, or clones that have become rearranged after cloning 11. Recent developments in both technology and analytical procedures are likely to shih the map-based cloning strategy in many species away from one dominated by chromosome walking. Instead, initial efforts will be directed towards selectively identifying molecular markers that are so close physically to the target gene that little or no chromosome walking is required. "l~nese tightly linked markers will be used to screen a genomic library to isolate directly a single clone containing the gene of interest. We refer to this approach as 'chromosome landing' and herein discuss its requirements, applications and limitations. In addition, we
Chromosome landing: a paradigm for map-based gene cloning in plants with large genomes STEVEND. TANKSLEY, MARTINW. GANALAND GREGORYB. MARTIN The original concept behind mop-based or positional cloning was to find a DNA marker linked to a gene of interest, and then to "walk'to the gene via overlapping clones (e.g. cosmids or YACs). Whilechromosome walking is straightforward in organisms with small genomes, it is did~cull to apply in most plant species, which typically have large, complex genomes. The strategy of chromosome walking is based on the assumption that it is d i g i t and time consuming to find DNA markers that are physically close to a gene of interest Recent technological developments invalidate this assumption for maW species. As a resug the mopping paradigm has now changed such that onefirst isolates one or more DNA marker(s) at a physical distance from the targeted gene that is less than the average insert size of the genomic library being used for clone isolatiog The DNA marker is then used to screen the library and isolate (or 7and'on) the clone containing the gene, without any needfor chromosome walking and its associated problems. Chromosome landin~ together with the technology that has made it possible, is likely to become the moin strategy by which mop-based cloning is applied to isolate both major genes and genes underlying quantitative traits in plant species. provide an example of how the method was recently used to isolate a gene involved in disease resistance in tomato, a plant with a relatively large genome.
Search for marker(s) tightly linked to a target gene When map-based cloning was fast proposed for use in species with large genomes, molecular linkage mapping was in its infancy and molecular maps were not available for most plant species. It could take months or even years to find a single molecular marker linked to a target gene. In the past ten years, the situation has changed dramatically. Molecular linkage maps have now been constructed for more than 30 plant species and work is under way to construct maps for at least as many again. Maps for some species are highly refined, containing hundreds of markers 12a3. The availability of high-density molecular linkage maps provides a ready starting point to search for molecular markers tightly linked to a target gene. For example, the genetic map of tomato contains more than 1000 DNA markers. Assuming a genome size of 1000 megabases, any gene is expected to lie, on average, within 1000 kb of a molecular marker 14.
Additional helpfrom comparativegenetic linkage maps Genomes of related species have evolved from a common ancestral genome and thus share portions of
TIG FEBRUARY1995 VOL. 11 No. 2 © 1995 ElsevierScienceLtd 0168 - 9525/95/$09.50
6 3
REVIEWS
(a)
O~ ¢.. ¢9
m
(b)
m
t-
I II
III I t
I I
I
I
III l
II
I 11
potato and pepper) and the mustard family22,23 (Arabidopsis, cabbage, cauliflower and rape). As regards map-based gene cloning, comparative maps allow researchers working on one species to identify and access probes from orthologous regions of related genomes. For some species (e.g. the grasses), comparative maps have expanded the repertoire of available mapped probes from a few hundred to several thousand. While high-density maps and comparative maps can provide some of the linked markers needed for chromosome landing in plants, it is clear that additional methods are needed by which new markers can be isolated selectively from specific regions of a genome.
Regional targeting witb bigb-vohtme marker tecbnologv
The probability of identifying one or more markers within a specified physical distance of a target gene decreases the closer one gets to the target. Ideally, one would like to idents,,. SS ~ ify a marker at a physical distance from a tar~ ~ | get that is less than the average insert size of A B C Candidate genes the genomic library from which one expects to isolate the gene. For example, working Fmu~ 1. Comparison of strategies for gene mapping using (a) chromosome with a YAC library with an average insert size walking and (b) chromosome landing. Blue boxes represent chromosomal of 200 kb, it would be desirable to identify at regions that contain a gene targeted for map-based cloning; yellow boxes least one marker within a 400 kb window represent overlapping genomic clones corresponding to the chromosomal containing the target (200 kb on each side). If region containing the target gene. Arrows indicate positions of linked DNA such a marker were to be identified, it wofild markers. Red boxes represent candidate genes within a genomic be possible to screen the YAC library with the clone that is known to contain a target gene. marker and to directly isolate a single YAC that contains the targeted gene withom any their linkage maps that are conserved in both gene conneed for chromosome walking. However, for species tent and order. The advent of molecular genome mapwith large genomes, the probability of identifying a r2nping has made it possible to determine the relative dom marker within 200 kb of a target is very low. For order of genes among related species that are no longer example, in a genome containing 1 billion base pairs, sexually compatible. Extensive comparative maps have the probability would be 4 x 102/106 = 4 × 104. To been constructed for a number of plant taxa, including have a 90% chance of identifying at least one marker species in the grass family15-19 (rice, maize, sorghum, within a 400 kb interval, ~ 6000 markers would need to barley and wheat), the nightshade family14,20,21 (tomato, be sampled (Fig. 2). It is not likely that many plant species will have molecular linkage maps of this density in the foreseeable future. Recently, several strategies have been developed 1.0J , _...... .................................. that allow one to screen a large number of random, ~¢.~ t ..-'"" .,~ -'" " " unmapped molecular markers in a relatively short time E ~ 0.8 ....." ~.~" and to select just those few markers that reside in the vicinity of the target gene. These methods rely on two principles: (1) the development of high-volume marker technology, which allows hundreds or even thousands o o~ 0.4. / / ,. of potentially polymorphic DNA segments to be gener"6 ,_= 100 Mb (Arabidopsis) ated and visualized rapidly from single preparations of ~'i / 400 Mb (Rice) DNA; and (2) use of genetic stocks to identify, among =~ ~ 0.2 -"/'/ ~/ 1000 Mb (Tomato) these thousands of DNA fragments, those few that are .o "_~ : 3000 Mb (Maize) ¢L derived from a region adjacent to the targeted gene. 0,0 High-volume marker technologies that have demon0 50'00 10()00 15000 20000 strable efficacy are random amplified polymorphic No. of markers DNAs24 (RAPDs), amplified fragment length polymorphisms 25 (AFLPs), RFLP subtraction x6 and repreFmuaE 2. Graph showing the number of random markers that sentational difference analysis27 (RDA). RDA and RFLP must be sampled to obtain at least one marker less than 200 kb subtraction were described recently by researchers from a given target gene. Calculations are for four genomic sizes. working in mammalian genetics and use DNA-DNA
li/
• . . . . . . . . . . . . . .
TIG FEBRUARY1995 VOL. 11 NO. 2
64
REVIEWS
Box 1. Nearly isogenic ILnes and bulked segregant analysis (A) Nearly isogenic lines Nearly isogenic lines (NILs) are created when breeders cross a donor line P1 to a recipient line P2. The resulting F1 hybrid is then crossed back to the P2 recipient to produce the backcross 1 generation (BC1). From BC1, a single individual that contains the dominant allele of the target gene from P1 is selected. Selection for the target gene is normally made on the basis of phenotype. This BC1 individual is again backcrossed to P2, and the q'cle of backcrossselection is repeated for a number of generations. In the BC7 generation, most, if not all, of the genome will be derived from P2, except for a small chromosomal segment containing the selected dominant allele, which is derived from Pi. Lines homozygous for the target gene can be selected from the BC7 F2 generation. The homozygous BC7F2 line is said to be nearly isogenic with the recipient parent, P2. Such pairs of nearly isogenic lines (NILs) are frequently generated by plant breeders as fi~ey transfer major genes between varieties by backcross breeding.
(a)
P1
P2
F1
BC1
BC7F2 (NIL)
Tar e,_ilXII:IB. . . . . .
II
gene
(b)
P1
P2
li ' 11 ,arne, it.t1-1i--;t-'-2--iI--I1-II-ll-BI' Bulk I (c)
P1
Bulk 2 P2
NIL
Bulk 1
Bulk 2
(B) Bulked segregants analysis Bulked segregant analysis (BSA) requires the generation of populations of bulked segregants (bulks). When Pl and Marker P2 are hybridized, the F2 generation analysis derived from the cross will segregate for alleles from both parents at all loci throughout the genome. If the F2 population is divided into two pools of contrasting individuals on the basis of screening at a single target locus, these two pools (Bulk 1 and Bulk 2) will differ in their allelic content only at loci conrained in the chromosomal region close to the target gene. Bulk 1 individuals, selected for the recessive phenotype, will contain only P2 alleles near the target, while Bulk 2 plants (selected L, ,.he dominant phenotype) will contain alleles from both P1 and P2. Bulk 1 and Bulk 2 will both contain alleles from both P1 and P2 at loci unlJnked to the target.
(C) High-volume marker techniques High-volume marker techniques, such as those based on RAPDs and AFLPs, can be used to compare the genetic profiles of pairs of NILs or pairs of bulks to search for DNA markers near the target gene. DNA fragment(s) (indicated by arrows) that are observed in the NIL and P1 populations, but not in P2, or that are unique to either Bulk 1 or Bulk 2, should be derived from a locus tightly linked to the target gene. hybridization in solution to selectively enrich for portions of two genomic samples that differ in sequence 26.27. Currently, application of the use of RDA or RFLP subtraction in plants has not been reported, but it seems likely that these techniques could be used in combination with the strategies outlined below (see also Box 1). RAPDs, and more recently AFLPs, have been used with considerable success in plants species. Like RDA and RFLP subtraction, these methods allow a global comparison of genomes for genetic differences. Both RAPDs and AFLPs rely on the polymerase chain reaction (PCR) used in conjunction with primers that amplify
random sites throughout the genome. Amplification products are analysed on agarose or acrylamide gels to allow visualization of variation among individuals (see Box 1). By using one or mo~ of these high-volume marker technologies, thousands of loci scattered throughout the genome can be assayed in a matter of weeks or months. The next problem is to determine which of the amplified loci lie near the targeted gene. In the past few years, two strategies have proven effective for identifying from a large number of markers the few that reside near a targeted locus. Both involve the use of genetic stocks that are (almost) genetically identical, except in the regions flanking the targeted
TIG FEBRUARY1995 VOL. 11 NO. 2
65
REVIEWS
gene (Box 1). The in'st strategy, which uses nearly isogenic lines and is referred to hereafter as the NIL strategy, can be applied only where breeders or geneticists have generated inbred lines that differ at the targeted locus zS,zg. Generation of such stocks can take years but, fortunately for plant geneticists, many such stocks have been routinely generated as a byproduct of breeding. The second, and more generally applicable, strategy has been referred to as the bulked segregant analysis (BSA) method and relies on the use of segregating populations30,31. Both these methods, used in conjunction with high-volume DNA marker technology, permit one to screen thousands of loci and to selectively identify those adjacent to the gene of interest. Moreover, this can be accomplished without having a genetic map for the species. The efficacies of the NIL and BSA approaches have been demonstrated in plants 28-'34, and more recently these approaches have also proven effective in gene targeting studies in mammals35.
High resolution mapping High density molecular maps, comparative maps and regional targeting strategies permit relatively rapid identification of many markers that flank a targeted gene, with markers usually lying within a few cM of the gene. This creates the dilemma of determining which marker(s) lies closest to the target and should thus be used to screen a genomic library to isolate it. For example, if 6000 markers (see above) were screened using the NIL or BSA strategy, it is likely that 50 or more markers would be selected as being within 10 cM of the target. To determine which of these is actually closest to the gene, high-resolution genetic mapping is required.
Genetic mapping Genetic mapping depends on meiotic crossovers to orient loci along a chromosome. In a genome of 1 billion base pairs and 1000 cM, 1 cM is, on average, 1 million base pairs. To localize markers within distances of 100 kb or less means mapping to a resolution of 0,1 cM resolution or less. This would require analysis of thousands of segregating progeny in the hope of obtaining those few individuals in which crossovers had occurred very close to the target gene. For example, to have a 95% chance of recovering a single crossover less than 0.1 cM from a target gene, analysis of more than 3000 meiotic products would be required. While PCR-based marker technology and automation is making analysis of larger populations more practical, this is still a laborious task and one of the rate-limiting steps in map-based gene cloning. Below, we describe two strategies for improving the efficiency of highresolution mapping; further developments in this area are expected in the future.
Analysis offlanking markers In segregating populations of a modest size (e.g. 100 meiotic events), a target gene can usually be placed within an interval of 5-10 cM with a high degree of certainty. The markers defining this interval can then be used to screen a larger segregating population and identify individuals derived from one or more gamete(s) containing a crossover in the given interval. Only such individuals are useful in orienting other
markers closer to the target gene. Once identified, these individuals can be analysed in relation to all molecular markers within the region to identify those closest to the target. The flanking marker approach was recently used to refine the map position of the Cf-2gene, which encodes resistance to Cladosporiumfidvum in tomato, to a 40 kb region between two DNA markers 36. Flanking marker analysis is especially powerful ~, ".en the defined interval is small, for example, <10 cM, and the flanking markers are easy to score. Ideally, flanking markers could be screened without the need for DNA isolation or other time consuming activities. Morphological markers, which are common in most plants .and have in several instances been placed on existing molecular maps, are ideal. Unfortunately, the number of morphological markers available is limited and the probability that a specific target gene will lie within a small interval defined by two useful morphological markers is quite low. If the target gene is flanked by molecular markers, it then becomes necessary to extract DNA from all segregating individuals, unless PCR can be perforated directly on intact plant tissue57,38.
Pooled-sample mapping Pooled-sample mapping is another approach for selectively identifying those rare individuals in a large population that have a crossover near a target gene. This method does not rely on localization of the target gene to an interval between two flanking markers; instead, it requires either that the genotype of the target gene can be identified accurately on a large scale, or that there is a tightly linked marker that can be screened for on a large scale. Individuals from the large population are 'pooled' for DNA extraction and marker analysis and the pooled DNA subjected to analysis with all molecular markers in the vicinity of the target gene. On the basis of the proportion of the pools that contain a crossover with respect to the markers analysed, the marker(s) closest to the target can be identified and the order of all markers in the vicinity of the target can be determined. In theory, this method allows the screening of thousands of gametes for informative crossovers with a minimum of DNA extraction and molecular analysis. The pooled-sample strategy has been demonstrated in the fine mapping of a gene that regulates fruit ripening in tomato39.
Physical mapping High-resolution mapping can be used to identify the marker(s) closest (in terms of genetic recombination) to a gene targeted for cloning. Unfortunately, units of genetic recombination (cM) cannot be directly translated into distances in base pairs. For example, in tomato, 1 cM is expected to correspond to approximately 700 kb (Ref. 14); however, physical mapping of selected regions of the tomato genome has shown that 1 cM can correspond to physical distances that vary by up to two orders of magnitude40, 41. In the regions around centromeres and certain telomeres, in particular, meiotic recombination is suppressed, resulting in much higher kb/cM values TM. For chromosome landing to be successful, the marker used for library screening must lie within a physical distance of the target gene that is less than the average insert size of
TIG FEBRUARY1995 VOL. 11 NO. 2
66
I~.EV1[ E W S
the genomic library being screened. Unfortunately, the great variation in meiotic recombination along a chromosome makes it necessary to determine empirically the approximate kb/cM value within the targeted region. In plants, this has been accomplished most effectively through the use of longrange restriction mapping, using rare-cutting restriction enzymes and pulsed-field gel electrophoresis in conjunction with DNA markers that are closely linked to the target gene 4°--42. Long-range physical mapping can also help confirm the order of DNA markers in the vicinity of the target.
Identifying the target gene in a large genomic clone
t , .
CD~106 Markers
.CD127
TGS0;
I
I Centimorsans
10.2
I 1.4
I~
F
0.4
OA
<0.2
Identification of the target gene from a number of candidate genes encompassed by a large-insert FIGURE3. Chromosome landing in the Pto region of tomato. A 12 cM region of clone is the final daunting step in chromosome 5 containing the Progene and flanked by markers TG475 and TG504 is shown (bottom). A single YACclone was isolated using the marker TG538, which had map-based cloning. Although the been determined by genetic studies to be closest to Pto. The ends of the YACclone were density of genes will vary through- isolated and geneticallymapped (PTY538-1Land PTY638-1R)and shown to flank the out the genome, a 300--400 kb YAC Pto gene. Two candidate cDNAclones from the region (CD106and CD127)were isolated clone can easily encompass more using the YACas a probe. Autoradiographsof survey filtersof genomic DNA from than 30 genes. Identification of can- resistant (left lane of each pair) and susceptible(right lane) tomato plants digested with didate genes can be accomplished six restrictionenzymes and probed with CD106and CD127are shown (top). Note that by probing cDNA libraries with CD127 detects high levels of polymorphismbetween resistant and susceptible plants; either DNA from the whole YAC transformation-complementationexperimentslater showed that it was a member of the or with DNA from phage h or Pto resistance gene family. cosmid clones from a contig spanning the YAC. Recently, methods for isolating cDNAs Chromosome landing in plants: isolation of the Pro encoded by large genomic regions have been improved gene The tomato gene Pto confers resistance to the in several ways. The new methods, referred to as 'direct cDNA selection' or 'capture systems' obviate the net d bacterial pathogen Pseudomonas syringae pv. tomato. Recently, this gene was isolated using the chromosome for plaque or colony hybridizations and are likely to significantly ease isolation of candidate genes from landing approach 4. Intensive efforts resulted in the identification of 18 informative markers that were closely large genomic clones known to contain a target gene 43. linked to the Pto gene 45. High-resolution linkage analysis In most cases, numerous cDNA clones will be in more than 1200 F2 plants identified one marker that identified by the above methods. Evidence that a particular cDNA corresponds to the target gene is obtained cosegregated with the gene, and this marker was used by a variety of methods including: (1) high-resolution to isolate YAC clones derived from the target region mapping to demonstrate cosegregation of the candidate (Fig. 3). Genetic mapping of isolated YAC ends showed that one clone spanned the Pro locus. Gel-purified DNA clone with the phenotype; (2) demonstration that the expression pattern of the gene is consistent with the from a YAC clone was radiolabelled and used to probe a cDNA library. Two classes of cDNA clones were identphenotype, for example, that it is transcribed in the ified. Placing the cDNAs on a high-resolution linkage appropriate tissue(s) and/or induced by the appropriate map eliminated one class, as recombination events stimulus; (3) determination of the DNA sequence of the gene and its comparison with those in sequence data- were identified that separated these clones from the disease resistance locus (Fig. 3). However, high-resolution bases to identify any homology with genes of known function; (4) probing mutant lines with the cDNAs to mapping indicated that the second cDNA class cosegreidentify alterations of the DNA or mRNA from the wild gated with the resistant phenotype. Agrobacteriummediated transformation of a susceptible tomato plant type; (5) complementation of the mutant phenotype by transformation with the gene. The last approach, comp- with this candidate cDNA demonstrated that it conlementation, provides the most direct and definitive ferred disease resistance. This work exemplifies the advantages of chromosome landing, in that initial evidence. Genetic complementation by Agrobacteriummediated transformation is routine in many plant emphasis on the isolation of many closely linked DNA markers eliminated the need for chromosome walking species and has been used in several instances to confirm that a candidate gene is responsible for a particular and the development of a high-resolution linkage map expedited the identification of candidate cDNAs. phenotypic trait3--9A4. TIG FEBRUARY1995 VOL. 11 NO. 2
67
REVIEWS
12 O'Brien, S.l., ed. (1993) Genetic Maps, Cold Spring Harbor
Conclusions and future prospects
Laboratory Press
Chromosome landing (and the associated techniques that have made it possible) can be applied in most sexually reproducing species, including most plants and several animals (including many domesticated animals). It minimizes or completely eliminates the need for chromosome walking, and thus greatly reduces the problems engendered by both the volume and repetitive nature of DNA analysis in such genomes. Currently, high-resolution genetic mapping is the most significant hurdle to the widespread application of chromosome landing. As discussed above, several strategic advances have been made in this area and we can reasonably expect that future developments will continue to improve the situation. It is important to point out that there are some species in which mapping by chromosome landing is not likely to play a major role, most notably in humans, where controlled matings are impossible and population sizes are inadequate for high-resolution mapping studies. Nonetheless, certain aspects of chromosome landing, in particular, high-volume marker technology and regional targeting, are beginning to play a role in map-based cloning in mammals 46. For the more intensively studied animals, such as humans and mice, and for plant species such as Arabidopsis, complete genomic contigs made up of YACs or bacterial artificial chromosomes will probably be constructed within the decade. When complete contig maps are available, the need for high-volume marker technology will be reduced. However, it is unlikely that the genomes of most plants, especially those of agronomic importance, will be organized into contigs in the near future. For these species, chromosome landing seems likely to become the main strategy by which map-based or positional cloning is applied to isolate agronomically important and botanically interesting genes whose products are not known. Finally, the past few years have seen numerous publications describing the use of molecular linkage maps to localize genes underlying quantitative traits ( Q ~ ) in plants and animals 47A8. The chromosome landing paradigm outlined here can readily be applied to cloning QTLs that can be mapped with a high degree of resolution. Progress has already been made in highresolution mapping of QTLs in plants, and it seems likely that in the foreseeable future, chromosome landing will begin to yield genes for characters that are quantitatively inherited 49.50.
13 Phillips, R.L. and Vasil, I.K. (1994) DNA-basedMarkers in Plants, IGuwer 14 Tanksley, S.D. etal. (1992) Genetics132, 1141-1160 15 Ahn, S.N. and Tanksley, S.D. (1993) Prec. Natl Acad. Sci. USA 90, 7980-7984 16 Ahn, S.N. et al. (1993) Mol. Gen. Genet. 241,483-490 17 Kurata, N. et al. (1994) Bio/Technology 12, 276-278 18 Whitkus, R. etal. (1992) Genetics 132, 1119-1130 19 Van Deynze, A.E. et al. Genome (in press) 20 Bonierbale, M. et al. (1988) Genetics 120, 1095-1103 21 Prince,J.P. etal. (1993) Genome36, 404--417 22 Kowalski, S.P. et al. (1994) Genetics 138, 499-510 23 Slocum, M.K. (1989)in Development a n d Application of Molecular Markers to Problems in Plant Genetics
24 25 26
27 28 29 30 31 32 33 34 35 36 37 38 39 40
41 42 43 44
45 46
47
References
48 49 50
1 Wicking, E. and WiUiamson, B. (1991) Trends Genet. 7, 288-293 2 Collins, F.S. (1992) Nature Genet. 1, 3--6 3 Arondel, V. et al. (1992) Science 258, 1353-1355 4 Martin, G.B. et al. (1993) Science 262, 1432-1436 5 Mindrinos, M. etaL (1994) Cell78, 1089-1099 6 Giraudat,J.B. etal. (1992)Plant Cell4, 1251-1261 7 Chang, C. et al. (1993) Science 262, 539-570 8 Leung,J. et al. (1994) Science264, 1448--1452 9 Meyer, K., Leube, M.P. and Grill, E. (1994) Science 264, 1452-1455 10 Arumuganathan, K. and Earle, E.D. (1991) PlantMol. Biol. Rep. 9, 208-218 11 Green, E.D. etal. (1991) Cell70, 1059--1068
(Helentjaris, T. and Burr, B., eds), pp. 73-80, Cold Spring Harbor Laboratory Press Williams,J.G.K. et al. (1990) Nucleic Acids Res. 18, 6531-6535 Zabeau, M. (1993) Eur. Pat. Appl. No. 92402629.7 Rosenberg, M., Przybylska, M. and Straus, D. (1994) Proc. Natl Acad. Sci. USA 91, 6113-6117 Lisitsyn,N., Lisitsyn, N. and Wigler, M. (1993) Science 259, 946-951 Young, N.D. et al. (1988) Genetics 120, 579-585 Martin, G.B. et al. (1991) Proc. Natl Acad. Sci. USA 88, 2336-2340 Michelmore, R.W. et al. (1991) P~c. Natl Acad. Sci. USA 88, 9828-9832 Giovanonni, J.J. et al. (1991) Nucleic Acids Res. 19, 6553--6558 Mackill, D.J. et al. (1993) Theor. Appl. Genet. 85, 536-540 Schiiller, C.G. et al. (1992) Theor. Appl. Genet. 84, 330-338 Pineda, O. et al. (1993) Genome36, 152-156 Lisitsyn,N.A. et al. (1994) Nature Genet. 6, 57-63 Dixon, M.S. et al. Mol. Gen. Genet. (in press) Klimyuk,V.I. et al. (1993) PlantJ. 3, 493-494 Chunwongse. j. et al. (1993) Theor. Appl. Genet. 86, 694-698 Churchill, G.A. et al. (1993) Proc. Natl Acad. Sci. USA 90, 16-20 Ganal, M.W. et al. (1989) Mo!. Gen. Genet. 215, 395--400 Segal, G. et al. (1992) Mol. Gen. Genet. 231, 173-185 Ganal, M.W. et al. (1990) AgBiotech News Inf. 2, 835--840 Lovett, M. (1994) Trends Genet. 10, 352-357 Whitham, S. etal. (1994)Cell78, 1101-1115 Martin, G.B. et al. (1993) Mol. Plant Micr. Interact. 6, 26-34 Aldhous, P. (1994) Science265, 2008--2010 Crabbe, J.C. et al. (1994) Science 264, 1715-1723 Tanksley, S.D. (1993) Annu. Rev. Genet. 27, 205-233 Paterson, A.H. et al. (1990) Genetics 124, 735-742 Ritsert,J.C. and Stam, P. (1994) Genetics 136, 1447-1455
S.D. TANgSLEFXS UN THE DEPARTMENTOF PLANT BREEDING AND BIOMETRY, CORN£LL UNIVERSt~, ITHACA, N Y 14853. USA, M.W. GANALIS IN THE INSTITOTEFOR PLANT GENETICS AND CROP PLANT RESEARCH, 0-06466 GATERSLEBEN, GERMANY,, AND G.R MARTIN IS IN THE DEPARTMENT OF AGRONOMY,PuP~U£ UNIVERSITy, WESTLAgFEZT~, I N 47907, USA.
TIG FEBRUARY1995 VOL. 11 NO. 2
68