Extraordinarily polymorphic microsatellite DNA in barley - Europe PMC

2 downloads 0 Views 2MB Size Report
Contributed by R. W. Allard, February 4, 1994. ABSTRACT. This study was undertaken to assess the ex- tent of genetic variation in barley simple sequence ...
Proc. Nat!. Acad. Sci. USA Vol. 91, pp. 5466-5470, June 1994 Population Biology

Extraordinarily polymorphic microsatellite DNA in barley: Species diversity, chromosomal locations, and population dynamics M. A. SAGHAI MAROOF*t, R. M. BIYASHEV*, G. P. YANG*, Q.

ZHANGf, AND R. W. ALLARD§

*Department of Crop and Soil Environmental Sciences, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061; tBiotechnology Center, Huazhong Agricultural University, Wuhan, China; and IDepartment of Agronomy and Range Science, University of California, Davis, CA 95616

Contributed by R. W. Allard, February 4, 1994

variants may be useful for genome mapping in plants and also in marker-directed plant breeding. SSR markers have been developed in several plant species including soybeans (10, 11) and rice (9, 12). Close linkage between an SSR marker and a gene conferring resistance to soybean mosaic virus was recently reported (13), and SSR markers in rice have identified several chromosome segments that may have significant effects on yield and its component traits (14). In this paper we report the extent of SSR polymorphism at four loci in cultivated barley (Hordeum vulgare ssp. vulgare; hereafter Hv) and in one of its wild relatives (H. vulgare ssp. spontaneum; hereafter Hs). We also report the Mendelian inheritance and chromosomal locations of these four SSR loci, and we discuss the evolutionary dynamics of allelic frequency changes that occurred over 52 generations in an experimental barley population (Composite Cross II, CCII).

This study was undertaken to assess the exABSTRACT tent of genetic variation in barley simple sequence repeats (SSRs) and to study the evolutionary dynamics of SSR alleles. SSR polymorphisms were resolved by the polymerase chain reaction with four pairs of pimers. In total, 71 variants were observed in a sample of 207 aIons of wild and cultivated barley. Analyses of wheat-barley addition lines and barley doubled hplolds identified these variants (alleles) with four lad, each located on a different chromosome. The numbers of alleles detected at a locus corresponded to the number of nudeotide repeats in the mictelite sequences. The numbers of alleles at two loci were 28 and 37; to our knowledge these are the largest numbers of alleles for single Mendelian loci reported in plants. Three allels were resolved by each of the other two loci. Allelic diversity was greater in wild than in cultivated barley and surveys of two generations (Fs and Fs3) of Composite Cross U, an exerimental population of cultivated barley, showed that few of the alleles present in the 28 parents survived into generation Fg3, whereas some infrequent alleles reached high frequencies. Such canges in frequency indiate that the chromosomal segments marked by the SSR alleles are under the influence of natural selection. The SSR variants allow spedflc DNA sequences to be followed through generations. Thus, the great resolving power ofSSR assays may provide clues regarding the predse targets of natural and man-directed selection.

MATERIALS AND METHODS Genetic Materials. Four sets of experimental materials were examined. The first set included 104 accessions of cultivated barley (Hv) from all of the major barley-growing areas of the world and 103 accessions of two-rowed wild barley (Hs) from many ecologically diverse locations in Israel. Both the Hv and the Hs accessions were obtained from the world barley collection of the United States Department of Agriculture. The second set included a barley variety ('Betzes'), a hexaploid wheat variety ('Chinese Spring'), and six wheat-barley chromosome addition lines. Each addition line contains one pair of barley chromosomes added to the 21 pairs of Chinese Spring. The third set included 79 doubled haploid (DH) lines derived from a cross between two barley cultivars, 'Steptoe' and 'Morex' (15). The fourth set of materials consisted of 168 families derived from random samples of stored kernels of generations F8 and F53 of barley CCII. CCII was synthesized in 1928 by pooling equal numbers of F2 seeds obtained by selfing F1 hybrid plants derived from the 378 possible pairwise intercrosses among 28 barley cultivars representing the major barley growing areas of the world. CCII has subsequently been grown annually under standard agricultural conditions in isolated plots containing >15,000 reproducing adults per generation and harvested in bulk at maturity without conscious selection; the next generation was sown from a random sample of seeds from the previous harvest. Microsatellite Sequences and Primers. All barley sequences in the GenBank and EMBL data bases were searched for tandem di- and tri-nucleotide repeats by using the Genetics Computer Group (GCG) Sequence Analysis Software Package (16). Four SSR-containing genes were selected for de-

Microsatellites or simple sequence repeats (SSRs) are tandem repetitive DNA sequences with a repeat length of a few base pairs (1). Dinucleotide repeats are the most abundant microsatellites in mammals-e.g., (CA). repeats are reported to occur with a frequency of about 5-10 x 104 per genome in mammals (2), and such microsatellites appear to be randomly dispersed in the genome (3). SSRs of tri- and tetranucleotide repeats are also abundant in the human genome (4). Variation in the number of nucleotide repeats (SSR polymorphism) can be detected with the polymerase chain reaction (PCR) by selecting as primers the conserved DNA sequences flanking the SSRs. SSR polymorphisms have recently become genetic markers of choice (1, 5, 6) in mammalian genetics for several reasons, including their abundance and the simple experimental procedures by which they can be detected. Genetic linkage maps based solely on SSR variants have been constructed for several mammalian species (e.g., ref. 7). SSRs have also been reported in several species of higher plants. Screening of DNA libraries for the presence of dinucleotide repeats showed that there are about 5 x 103 to 3 x 10W (AC)n and (AG)n sites per genome in maize and in five tropical tree species (8). It is estimated that (GA)n repeats occur on average about once every 225 kb and that (GT)n repeats occur about once every 480 kb in the rice genome (9). These high estimates have led to the suggestion that SSR

Abbreviations: SSR, simple sequence repeat; RFLP, restriction

fragment length polymorphism; DH, doubled haploid; PCR, polymerase chain reaction; Hv, Hordeum vulgare ssp. vulgare, cultivated barley; Hs, Hordeum vulgare ssp. spontaneum, wild barley; CCII,

The publication costs of this article were defrayed in part by page charge

payment. This article must therefore be hereby marked "advertisement"

Composite Cross II. tTo whom reprint requests should be addressed.

in accordance with 18 U.S.C. §1734 solely to indicate this fact.

5466

Population Biology: Saghai Maroof et al. tailed study: (i) the ribulose-1,5-bisphosphate carboxylase (rubisco) activase (RcaA) gene (17) with (AT)., (ii) the starch synthase (Waxy) gene (18) with (AT)., (iii) an acyl carrier protein I (Acd)) gene (19) with (AT),, and (iv) a seed imbibition protein (Sip)) gene with (TCT)M. Four primer pairs (Table 1), one per locus, were then designed by using the flanking single-copy sequences and were synthesized by Operon Technologies, Alameda, CA. Henceforth, we designate these four primer pairs HVM3, HVM4, HVM7, and HVM9, respectively. SSR Assay. DNA samples were prepared (20) from greenhouse-grown seedlings of the 472 entries (Hv, Hs, DHLs, CCII, and wheat-barley addition lines). Procedures for SSR assays were as described (13) with minor modifications. Briefly, the PCR reaction was conducted in a volume of 20 A4 containing 50 ng of template DNA; 0.1 pM each primer; 200 ,uM each dATP, dGTP, and dTTP; 5 pM dCTP; 1 pCi (37 kBq) of [a-32P]dCTP; 50 mM KCI; 10 mM Tris (pH 8.3); 0.01% gelatin; 1.5 mM MgCl2, and 1.0 unit of Taq polymerase. Samples were covered with mineral oil, and the reaction was processed at 940C for 3 min, at 550C for 2 min, and at 7rC for 1.5 min for 1 cycle, followed by 29 cycles at 940C for 1 min, 550C for 2 min, and 720C for 1.5 min, with a final extension step of 720C for 5 min. After PCR the product was denatured at 940C for 8 min, and 10 A1 of stop solution containing 95% formamide and 20 mM EDTA was added. Samples (3-5 pA) were loaded on a 6% polyacrylamide denaturing gel containing 8 M urea and separated at 1500-V constant power by using a DNA sequencing unit (model STS-45, IBI). Gels were immediately covered with plastic wrap and exposed to x-ray film for 30-120 min. Multiplex PCR Assay. Large-sample surveys of HVM7 and HVM9, whose ranges of PCR-amplified fragment sizes do not overlap (Table 1), were carried out with multiplex PCR assays by including both pairs of primers in the same reaction. We also conducted multiplex PCR assays with the wheat-barley addition lines by including three primer pairs, HVM4, HVM7, and HVM9, in the same reaction to confirm the chromosomal locations of these loci.

RESULTS Number of Alleles. Assays ofthe 207 samples of Hv and Hs, carried out with the four primer pairs, identified 71 SSR variants. Henceforth, we refer to each primer pair as a locus and each variant as an allele. The numbers of alleles at the four loci varied widely: large numbers of alleles were detected at loci HVM3 and HVM4 and small numbers at loci HVM7 and HVM9. Twenty-eight different alleles were detected at locus HVM3 as specified in the top row of Fig. 1 (see also Table 2). Except for six missing variants (alleles) corresponding to PCR-amplified DNA fragment numbers 3, 4, 5, 26, 30, and 33 (Fig. 1, numbers in the diagonal), the 28 alleles (Fig. 1, top row) form a continuous ladder with adjacent steps differing by two base pairs (Fig. 1). Nine of the 28 alleles were Table 1. Oligonucleotide primer sequences designed for four microsatellite loci: Repeat patterns and number of repeat units in GenBank sequence Repeat Length, pattern Primers (5' to 3') Locus bp ACACCTTCCCAGGACAATCCATTG HVM3 AGCACGCAGAGCACCGAAAAAGTC 188 (AT), HVM4 AGAGCAACTACCAGTCCAATGGCA A 198 (ATh GTCGAAGGAGAAGCGGCCCTGGTA HVM7 ATGTAGCGGAAAAAATACCATCAT 174 CCTAGCTAGTTCGTGAGCTACCTC (A)7 CTTCGACACCATCACCCAG 221 (TCT)5 HVM9 ACCAAAATCGCATCGAACAT

Proc. Natl. Acad. Sci. USA 91 (1994)

5467

observed in frequencies of 0.05-0.10, whereas the other 19 alleles were much less frequent. Thirty-seven alleles were detected at locus HVM4 (Table 2); these alleles formed a continuous ladder except for a few missing steps (the two alleles corresponding to the shortest and longest PCRamplified DNA fragments were not included in the ladder). Three of these 37 alleles were observed in frequencies of 0.10-0.14 and two in frequencies between 0.05 and 0.10; the remaining 32 alleles were much less frequent. Only three alleles were detected at each of the other two loci, HVM7 and HVM9; the frequencies of these alleles varied from 0.07 to 0.67. The sequences retrieved from the GenBank data bases indicated that there are 29, 9, and 7 AT dinucleotide repeats in the sequences targeted by HVM3, HVM4, and HVM7, respectively, and 5 TCT repeats in that of HVM9 (Table 1). Two of the lines used for GenBank sequence analysis were included in our survey; the data from these lines, in conjunction with the length of PCR products of various alleles (e.g., Fig. 1), allow the deduction that numbers of tandem repeats are about 9-43 among alleles of HVM3, 4-45 among alleles of HVM4, but ohly 6-8 among alleles of HVM7 and 4-6 among alleles of HVM9. Thus, for the SSR loci of the present study, the numbers of alleles detected by the primers are clearly correlated with the numbers of tandem repeats within the target sequences; the larger the repeat numbers, the larger the number of alleles detected. Comparison of Polymorphism in Hv and Hs. The large and approximately equal numbers of accessions sampled from Hv and Hs permit a direct comparison of the level of SSR variability in the two subspecies. Among the 71 alleles detected, 37 occurred in both subspecies, 10 were observed only in Hv, and 24 were observed only in Hs (Table 2); thus, 14 more alleles were detected in Hs than in Rv. The level of polymorphism of each locus was calculated by using the genetic diversity index, 1 - I pi, in which pi is the frequency of the ith SSR allele. Diversity values calculated for HVM7 and HVM9 (Table 3) fell within the range of previous estimates for isozymes and restriction fragment length polymorphisms (RFLPs) (e.g., ref. 21)-i.e., between 0:047 and 0.740. However, the diversity values for HVM3 and HVM4 were much higher than those of isozymes and RFLPs. A statistical test (22) indicated that the diversity valueof Hs was significantly higher than that of Hv for three of the four loci. At the fourth locus (HVM9), the diversity value was about equal in wild and cultivated barley (Table 2). The diversity of the total sample was partitioned into a within-group diversity component and a between-group differentiation component (Table 2). The component reflecting differentiation between Hv and Hs groups accounted for little (1.5-2.3%) of the total variation at loci HVM3, HVM4, and HVM9; however, for HVM7, 30.6%6 of the variation was due to between-group differentiation. Chronosomal Locations of the SSR Loci. Both uniplex and multiplex PCR analyses of the wheat-barley addition lines (Fig. 2) showed that sequences detected by HVM3 (RcaA), HVM4 (Waxy), HVM7 (Ac)), and HVM9 (Sip)) mapped to chromosomes 4, 1, 7, and 3, respectively. The exact map positions were assessed further by using a set of DH lines (Fig. 3). The parents (Steptoe and Morex) of the DH population differed in SSR alleles at only two (HVM3 and HVM4) of the four loci. Assays of the 79 lines of this DH population for these two SSR markers indicated good fits to 1:1 segregation ratios (X2l] values were 1.38 and 0.62 for HVM3 and HVM4, respectively); thus, the variants detected by each primer pair are allelic at a single Mendelian locus. The analyses located HVM3 in the region between ABG484 and Pgk2A on chromosome 4 in the RFLP linkage map (23), whereas the location of HVM4 coincided with that of the Waxy locus on chromosome 1. Among the two primer pairs

5468

~ ~.

Proc. Natl. Acad Sci. USA 91

Population Biology: Saghai Maroof et al.

(1994)

1 2 3 4 5 6 7 8 9, 10 11 12 13 14 i .16-17 1819 2021 2223.24.2-. 2

.

,~ ?~.,

,

"5

.'

B

aj

,

.' '. l?

l

i

i.

' :. ..

!e plaining.,,,'.

*

7

...

....'.. . .',: .

FIG. 1. Microsatellite DNA polymorphism in barley as detected by the primer pair HVM3. Genomic DNA from each plant was amplified by PCR using HVM3 in the presence of [32P]deoxynucleotides and detected by autoradiography. A total of 28 (top row) SSR variants (alleles) were observed. These 28 alleles, along with 6 that were expected on the basis of dinucleotide repeats but were not observed in our samples, compose a ladder of 34 steps (diagonal row), with each step differing from its immediately adjacent neighbors by a single dinucleotide repeat.

that did not detect polymorphism between Steptoe and Morex, HVM7 mapped to the same chromosome (chromosome 7) as the Acli locus (19), whereas the other, HVM9, mapped to the same chromosome (chromosome 3) as the Sip] locus (GenBank data). Barley-Specific Primers. The wheat-barley addition-line analyses also showed that for primer pair HVM4, the amplification product of addition line 1 (containing all 21 pairs of wheat chromosomes and one pair of barley chromosome 1) had four bands: one barley band and three wheat bands (Fig. 2). In contrast, there was only one barley band in the PCR product amplified by each ofthe other three primer pairs with their respective addition lines. Thus, the sequences of HVM4 are homologous in barley and all three wheat genomes, whereas at least one primer sequence of each of the other three SSR primer pairs is barley specific. Population Dynamics in CCII. In total, 33 alleles were detected at the four SSR loci in the 28 parents and two generations ofCCII. However, among these 33 alleles only 10 (5, 2, 1, and 2 of HVM3, HVM4, HVM7, and HVM9, respectively) survived into generation F53. Thus, more than two-thirds of the alleles originally present in the parents of CCII were lost during the span of 53 generations from parents to generation F53. The pattern of allelic frequency change differed for each locus. The pattern for locus HVM7 was the least complex: 27 Table 2. Number of alleles, level of diversity, and differentiation in Hs and Hv detected at four SSR loci No. of alleles Diversity index* Locus Hv Hs Common Total Hv Hs Total GsTt HVM3 23 22 17 28 0.94 0.93 0.95 1.6 HVM4 19 33 15 37 0.88 0.95 0.93 1.5 HVM7 3 3 3 3 0.11 0.46 0.59 30.6 HVM9 2 3 2 3 0.32 0.58 0.47 2.3 *Hs was significantly more diverse than Hv for loci HVM4, HVM7, and HVM9. tPercent of diversity due to differentiation between Hv and Hs.

of the 28 parents (%%) carried allele 2, whereas the 28th parent carried allele 3 of this locus. Note that allele 2 had become fixed but that allele 3 had been lost by generation F53 (Table 3). Also note that allele 3 was rare not only in the Table 3. Frequencies of the ultimately surviving alleles detected in the parents and two generations of CCII Locus HVM3

Allele designation* 19 20 22 25 29 Others

Hs 22t 0.02 0.08

Hv 23 0.05 0.10 0.03 0.08 0.02 0.72 19 0.09 0.15 0.18 0.13 0.16 0.29 3 0.03 0.95 0.02 2 0.81 0.19

Parents 16 0.04 0.14 0.07 0.04 0.04 0.67 12 0.18 0.07 0.21 0.18 0.04 0.22 2

CCII F8 12 0.05 0.23 0.14 0.01 0.15 0.42 9 0.22 0.09 0.28 0.03 0.03 0.35 1

F53 5 0.77 0.01 0.01 0.19 0.01

0.02 0.04 0.84 HVM4 33 2 3 0.03 5 0.02 0.75 6 0.09 0.25 7 0.09 8 0.08 Others 0.69 HVM7 3 1 1 0.23 2 0.08 0.% 1.00 1.00 3 0.69 0.04 HVM9 3 2 2 2 1 0.55 0.80 0.73 0.34 2 0.30 0.20 0.27 0.66 3 0.15 Alleles observed in relatively high frequencies in Hs and Hv are listed for comparison. Numbers of families for generations F8 and F53 were 87 and 81, respectively. *Aleles are identified by locus and numbers in columns 1 and 2. tThe top row gives total numbers of alleles for each locus for the five sets of genetic materials (Hs, Hv, parents, Fg, and F53). Other rows give the frequencies of designated alleles (or groups of alleles) in each of the five sets of genetic materials.

Population Biology: Saghai Maroof et aL E5

6

7

9

10 11

12 13

14

40

Proc. Natl. Acad. Sci. USA 91 (1994)

5469

study showed that allele 35 was also present in low frequency in generation F13. There are several possible explanations for the presence of allele 35 in the CCII population, including: (i) sampling accidents as a result of within accession diversity (24), (ii) migration of pollen and/or seeds from unrelated populations into the plots in which early generations of CCII were grown, and (iii) molecular events such as unequal crossing-over and transposition that occurred in early generations of CCII.

DISCUSSION We have surveyed microsatellite DNA polymorphisms at four loci in barley materials representing a broad range of the

B W B WB WB 1 23 4 6 PT P9 P4 P:4.7. 9

W

7

FIG. 2. Mapping SSR markers to barley chromosomes using wheat-barley addition lines and multiplex PCR analysis. SSR assay was conducted by using DNA samples from wheat (W; lanes 1, 3, 5, and 7), barley (B; lanes 2, 4, 6, and 8), and wheat-barley addition lines 1, 2, 3, 4, 6, and 7 (lanes 8-14). Uniplex- (first six lanes) and multiplex- (last eight lanes) PCRs were carried out with primer pairs HVM7 (designated as P7), HVM9 (P9), and HVM4 (P4) as specified. Note that lanes 1 and 3 do not show any amplification product (band) for wheat. A separate PCR analysis was carried out for HVM3 (data not shown).

parents of CCII but also in the sample of 104 accessions of It is also interesting that allele 2 of this locus, although present at a high frequency in Hv, is the least frequent allele in the Hs group (Table 3). A second pattern was observed at the HVM9 locus: alleles 1 and 2 were frequent not only in Hs but also in Hv and in the parents of CCII. Both alleles 1 and 2 remained in high frequency in F53 but with frequencies reversed from those in the parents. Allelic frequency changes for HVM4 followed a third pattern. Initially, 12 alleles were present in CCII but only two alleles (alleles 5 and 6) survived into generation F53, in which generation these two alleles were present in f = 0.75 and 0.25, respectively. A fourth pattern, observed at locus HVM3, differed from the third pattern in that several rare alleles were present in F53 along with two frequent alleles, alleles 19 and 25, present inf = 0.77 and 0.19, respectively. It can also be seen from Table 3 that the allele ultimately favored at each locus was not the most frequent allele either among the 28 parents or in the total sample of Hs or Hv. As examples, three of the alleles that rose to high frequencies in F53 were present in only one or two of the 28 parents. One allele (allele 35) of the HVM4 locus was found in only one among the 103 Hs accessions, in none of the 104 Hv accessions, and in none of the 28 parents nor in generation F53 of CCII. Therefore, it was surprising to find that this allele was present in moderate frequency in F8 of CCII; a follow-up

Hv.

S Ma

,

i

;

I11

1'2 13 14 1i 14f 17 18

_ _ SW .

FIG. 3. Segregation patterns for the SSR marker HVM4 in parents Steptoe (S), Morex (M), and 18 DH lines.

habitats in which wild and cultivated barleys are found worldwide. The number of SSR alleles observed at loci HVM3 and HVM4, 28 and 37, respectively, are much larger than have been detected in previous studies of morphological variants (25), allozyme variants (24, 25), disease-resistance variants (26), chloroplast DNA variants (27), ribosomal DNA (DNA encoding rRNA) variants (28), and RFLP variants (unpublished data) in the same sets of barley materials. These are indeed startling numbers for such a predominately inbreeding species because selfling quickly exposes "deleterious" alleles to elimination due to selection. The large numbers of SSR alleles observed suggests that a high proportion of dinucleotide and trinucleotide repeats have no effect or relatively minor effect on survivability-i.e., that most of such repeats are adaptively "neutral." To our knowledge, the numbers of alleles observed at these SSR loci are the largest for single Mendelian loci that have been reported for any plant species. However, these numbers may not have reached the upper limits possible when one considers the observed fragment lengths of the PCR-amplified products. For example, for locus HVM4 the length difference between alleles corresponding to the shortest and longest PCRamplified products exceeds 100 bp. Hence, for a ladder with 2-bp steps, there may be >50 steps. The repetitive sequence nature of the microsatellite DNA probably allows for longterm generation of other SSR alleles through mismatch and unequal crossing-over in meiosis. Thus, 50 alleles or more are theoretically possible for the HVM4 locus. It is likely that SSRs will also provide a rich source of markers for species that are depauperate in variation detectable by other means. Furthermore, the ease of assaying large numbers of individuals for SSR polymorphism, especially through multiplex PCR reactions and/or multiple loadings of single gels, may reduce work loads to levels that make SSRs especially attractive for types of studies requiring very large samplese.g., studies of population and evolutionary dynamics, germ plasm assessment, or marker-aided selection in plant breeding. The close relationships between numbers of repeats in microsatellite sequences and the number of SSR alleles detected at a locus suggest that loci with large numbers of tandem repeats are likely to have larger numbers of alleles than loci with smaller numbers of tandem repeats. However, data bank sequences with smaller number of repeats should not necessarily be ignored; most data bank sequences have been obtained from only one among a large number of accessions that harbor highly variable numbers of repeats in the microsatellite DNA sequences. As an example, several accessions in the present study had only four or five AT repeats at HVM4, the locus with the largest number of alleles among the four loci. Consequently, it may be a good practice to carry out small-scale surveys for polymorphism before making judgments regarding the potential usefulness of particular repeat sequences. Our SSR survey showed that wild barley (Hs) is more variable than cultivated barley (Hv), both in numbers of SSR

5470

Population Biology: Saghai Maroof et al.

alleles and in diversity levels of individual loci. These SSR results are consistent with the results of earlier studies with other types of markers; only about half as many alleles per locus were observed in Middle Eastern landraces (primitive cultivars) of cultivated barley (Hv) as in Middle Eastern wild barley (Hs) and still fewer alleles were found in modern elite cultivars than in landraces (for reviews, see refs. 29 and 30). Overall, only about one-third as many allozyme and ribosomal DNA alleles found in wild barley were observed in modern elite cultivars adapted in one or more of the major barley growing regions of the world. Interestingly, a majority of the alleles that were rare or infrequent in wild barley were not observed in cultivated barley. In contrast, a high proportion of the alleles that are present in moderate to high frequency in wild barley survived into landraces and ultimately into modern elite cultivars adapted in one or more barley-growing regions. We also surveyed allelic frequencies of the four SSR loci in the 28 parents and in two generations (F8 and F53) of CCII. Allelic frequency changes in CCII revealed two main features of the dynamics of evolutionary change in this population: (i) among 32 SSR alleles detected in the 28 parents, only 24 (75%) survived into generation F8 and only 10 (31%) into generation F53, although population size was sufficiently large (>15,000 reproducing adults per generation) to circumvent loss due to sampling accidents of even infrequent neutral alleles; and (ii) several alleles (e.g., alleles 19 and 25 of HVM3 and allele 5 of HVM4) that were present in only one or two of the 28 parents of CCII (f = 0.04 or 0.07) had increased dramatically in frequency by generation F53 (Table 3). Similar population behavior has been observed with other markers, and, as suggested previously (e.g., ref. 25), such changes can be explained on the basis of natural selection acting on the chromosomal segments in which the marker loci reside. The higher resolving power of microsatellites than previously available markers may make it possible to monitor the population behavior of an allele derived from a particular parent. As mentioned above, the origin of several of the apparently favored alleles can be traced to only one or two of the parental lines. In particular, allele 19 at the HVM3 locus, which was detected in only one parent ("Trebi") had become the predominant allele in generation F53; the cumulative frequency increase of this allele between generations F8 and F53 was thus 0.73; if one assumes no reversals in the direction of allele frequency changes in the intermediate generations, the most parsimonious estimate for the rate of directional frequency changes is 0.016 per generation. This represents strong selection, suggesting that the chromosome segment of Trebi marked by allele 19 at the HVM3 locus may possess a large advantage in the Davis environment over its homolog in the other 27 parents. The situation appears to be similar but less extreme for two other alleles (allele 25 of HVM3 and allele 5 of HVM4). Therefore, the SSR assays provide detailed information concerning the fate of specific alleles and specific chromosome segments over generations. Thus, the high resolving power of the SSR alleles appears to offer opportunities for identifying the precise targets of selection in barley.

Proc. Nad. Acad. Sci. USA 91

(1994)

We thank Yong G. Yu for technical assistance in designing the primers used in this study. This research was supported in part by the North American Barley Genome Mapping Project and the U.S. Department of Agriculture Grant 93-37300-8840. 1. Litt, M. & Lutty, J. A. (1989) Am. J. Hum. Genet. 44, 397-401. 2. Stalings, R. L., Ford, A. F., Nelson, D., Torney, D. C., Hilderrand, C. E. & Moyzis, R. K. (1991) Genomics 10, 807815. 3. Weber, J. L. (1990) Genomics 7, 524-530. 4. Edwards, A., Civitello, A., Hammond, H. A. & Caskey, T. C. (1991) Am. J. Hum. Genet. 49, 746-756. 5. Tautz, D. (1989) Nucleic Acids Res. 17, 6463-6471. 6. Weber, J. L. & May, P. E. (1989) Am. J. Hum. Genet. 44, 388-3%. 7. Weissenbach, J., Gyapay, G., Dip, C., Virgnal, A., Morissette, J., Millasseau, P., Vaysseix, G. & Lathrop, M. (1992) Nature (London) 359, 794-801. 8. Condit, R. & Hubbell, S. P. (1991) Genome 34, 66-71. 9. Wu, K. S. & Tanksley, S. D. (1993) Mol. Gen. Genet. 241, 225-235. 10. Akkaya, M. S., Bhagwat, A. A. & Cregan, P. B. (1992) Genetics 132, 1131-1139. 11. Morgante, M. & Olivieri, A. M. (1993) Plant J. 3, 175-182. 12. Zhao, X. & Kochert, G. (1993) Plant Mol. Biol. 21, 607-614. 13. Yu, Y. G., Saghai Maroof, M. A., Buss, G. R., Maughan, P. J. & Tolin, S. A. (1994) Phytopathology. 84, 60-64. 14. Zhang, Q., Gao, Y. J., Yang, S. H., Ragab, R. A., Saghai Maroof, M. A. & Li, Z. B. (1994) Theor. Appl. Genet., in press. 15. Chen, F. & Hayes, P. M. (1989) Theor. Appl. Genet. 77, 701-704. 16. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acid Res. 12, 387-395. 17. Rundle, S. J. & Zielinski, R. E. (1991) J. Biol. Chem. 266, 4677-4685. 18. Rhode, W., Becker, D. & Salamini, F. (1988) Nucleic Acids Res. 16, 7185-7186. 19. Hansen, L. & von Wettstein-Knowles, P. (1991) Mol. Gen. Genet. 229, 467-478. 20. Saghai Maroof, M. A., Soliman, K. M., Jorgensen, R. A. & Allard, R. W. (1984) Proc. Natl. Acad. Sci. USA 81, 80148018. 21. Zhang, Q., Saghai Maroof, M. A. & Kleinhofs, A. (1993) Genetics 134, 909-916. 22. Zhang, Q. & Allard, R. W. (1986) J. Hered. 77, 54-55. 23. Kleinhofs, A., Kilian, A., Saghai Maroof, M. A., Biyashev, R. M., Hayes, P., Chen, F. Q., Lapitan, N., Fenwick, A., Blake, T. K., Kanazin, V., Ananiev, E., Dahleen, L., Kudrna, D., Bollinger, J., Knapp, S. J., Liu, B., Sorrells, M., Heun, M., Franckowiak, J. D., Hoffman, D., Skadsen, R. & Steffenson, B. J. (1993) Theor. Appl. Genet. 86, 705-712. 24. Kahler, A. L. & Allard, R. W. (1970) Crop Sci. 10, 444-448. 25. Allard, R. W. (1988) J. Hered. 79, 225-238. 26. Allard, R. W. (1990) J. Hered. 81, 1-6. 27. Neale, D. B., Saghai Maroof, M. A., Allard, R. W., Zhang, Q. & Jorgensen, R. A. (1988) Genetics 120, 1105-1110. 28. Saghai Maroof, M. A., Allard, R. W. & Zhang, Q. (1990) Proc. Natl. Acad. Sci. USA 87, 8486-8490. 29. Allard, R. W. (1991) in Plant Breeding in the 1990s, eds. Stalker, H. T. & Murphy, J. P. (CAB, Wallingford, U.K.), pp. 119-146. 30. Allard, R. W. (1992) in Reproductive Biology and Plant Breeding, eds. Dattee, Y., Dumas, C. & Gallais, A. (Springer, Heidelberg), pp. 325-334.