35. Abundance of PCR-based RFLP for marker-aided selection in rice

Behzad GHAREYAZIE, Ning HUANG, Gerard SECOND, John BENNETT and Gurdev S. KHUSH

International Rice Research Institute, P. 0. Box 933, Manila, Philippines

Concept of marker-aided selection (Tanksley et al. 1989) and recent preliminary reports (Abenes et al. 1993) show that target genes can be identified in a segregating population at early plant growth stage based on linked DNA markers. The linked DNA markers can be identified by RFLP analysis using either Southern hybridization or polymerase chain reaction (PCR). Practical application of marker-aided selection requires that markers be identified with high level of accuracy and efficiency, be cost effective, and easy to use. Southern hybridization requires application of either radiochemicals or very expensive nonradioactive materials and is laborious, time-consuming and relatively large quantities of high quality DNA are needed. PCR overcomes these drawbacks and therefore is a method of choice.

To use PCR in marker-assisted selection, the generation of Specific Amplicon Polymorphism (SAP) such as PCR-based RFLP (PBR) (Williams et al. 1991; Ghareyazie et al. 1993) to follow a target gene is required. In the most desirable cases, Amplicon Length Polymorphism (ALP) between gene donor and recipient can be obtained, and represents the simplest and fastest way of detecting polymorphism. Digestion of PCR products with restriction endonucleases (4- base cutters) has been found to increase the level of polymorphism as well as the number of alleles in rice (Ghareyazie et al. 1993). However it is not clear how often ALP and PBR can be observed in rice for marker aided-selection.

To answer this question we analyzed the data obtained from a rice germplasm consisting of 35 Iranian, 3 indica and 2 japonica varieties. Total genomic DNA was extracted from the rice leaves. The DNA was then amplified with 14 sets of primers synthesized based on mapped RFLP markers. Amplicons were digested with 9 different restriction endonucleases (4-base cutters), regardless of the presence or absence of polymorphism (ALP). We then performed pairwise comparison among the 40 varieties for polymorphism. To calculate the abundance of PBR for marker-aided selection, we assumed the segregating populations for selection in a breeding program derived from pairwise crosses. Alleles in PBR are defined as polymorphisms detected by one enzyme or enzyme combination. For example the PCR products at the waxy locus before digestion are monomorphic (1 allele only), but after digestion with Taq1, three alleles were detected, and after digestion with HinFI, 2 alleles were obtained (data not shown). Combination of these alleles gives rise to 5 different alleles (polymorphism) at this locus. The total number of polymorphic pairs (NP) for each marker locus were then calculated as follows:

                  NP = n-1Summation`i=1`(Xi*nSummation`j=i+1 Xj)      
Where, n is the total number of alleles in a given locus, Xi and Xj are the number of varieties carrying the ith and jth alleles respectively.

The abundance of ALP and PBR is shown in Table 1. Out of the 9895 pairwise comparisons we detected ALP in 1277 pairs of varieties (13%). This percentage of ALP in rice is obviously not high enough for general use in marker-aided selection. Additional polymorphism was generated by restriction digestion of PCR products. The number of alleles increased for many marker

Table 1.  The abundance of ALP and PCR-based RFLP in 14 rice loci
_______________________________________________________________________________
                    Allele                               Allele
                 _________                          ___________
Marker  #Entrya   1  2  3  ALP Total ALP% #Entrya   1  2  3 4 5 6 PBR Totl PBR%
_______________________________________________________________________________
PTA248     39     24 15    360  741  49       39   21 15  3       360  741  49
RG13       37     22 10 5  380  666  57       26   14  4  4 2 1 1 221  325  68
RG64       40     32  7 1  263  741  34       40   30  7  2 1     263  741  34
RG100      40     40         0  780   0       40   40               0  780   0
RG118      39     39         0  741   0       38   20 11  4 2 1   451  703  64
RG120      39     39         0  741   0       39   39               0  741   0
RG173      39     39         0  741   0       40   36  3  1       147  780  19
RG235      40     35  5    175  780  22       40   35  5          175  780  22
RG241      39     39         0  741   0       34   31  2  1        95  561  17
RG257      40     40         0  780   0       40   38  2           76  780  10
RG329      40     40         0  780   0       22   21  1           21  231   9
RG365      36     33  3     99  630  16       36   33  3           99  630  16
RG386      23     23         0  253   0       26   17  9          153  325  47
Waxy       40     40         0  780   0       39   24  6  4 3 2   440  741  60
_______________________________________________________________________________
Total                     1277 9895  13                          2501 8859  28
_______________________________________________________________________________
a=Number of entries examined for ALP and PBR.
loci. For example, there are 6 alleles for RG13 in PBR while only 3 in ALP. For Waxy locus a total of 5 alleles were revealed in PBR. Increased number of alleles gives rise to more polymorphic pairs. Out of 8859 pairwise comparisons we found 2501 pairs being polymorphic at these loci (28%). Since the total number of enzymes available is much higher than used in this study, and considering the fact that studied varieties did not include all isozyme groups (Glaszmann 1987), we expect more than 28% PBR in rice. Since marker-aided selection is used to complement phenotype-based selection, we do not expect to perform marker-aided selection on all target genes. If we assume that marker-aided selection will be performed on one third or half of the target genes, we can basically rely on ALP or PBR for marker aided-selection. Accuracy, ease and speed of ALP and PBR therefore are attractive and should be included in the plant breeding programs when marker-aided selection is used.

References

Abenes, M. L. P., E. R. Angeles, G. S. Khush and N. Huang, 1993. Selection of bacterial blight resistant rice plants in the F`2` generation via their linkage to molecular markers. RGN 10: 120-123.

Ghareyazie B., N. Huang, G. Second, J. Bennett and G. S. Khush, 1993. Comparison between PCR-based RFLP and Southern-based RFLP as DNA markers for germplasm classification in rice. RGN 10: 129-132.

Glaszmann J. C., 1987. Isozymes and classification of Asian rice varieties. Theor. Appl. Genet. 74: 21-30.

Tanksley S. D., N. D. Young, A. H. Paterson and M. W. Bonierbale, 1989. RFLP mapping in plant breeding: new tools for old sciences. Biotechnology 7: 257-264.

Williams, M. N. V., N. Pande, S. Nair, M. Mohan and J. Bennett, 1991. Restriction fragment length polymorphism analysis of polymerase chain reaction products amplified from mapped loci of rice (Oryza sativa L.) genomic DNA. Theor. Appl. Genet. 82: 489-498.