GENETIC MAPS OF Saccharum officinarum L. AND Saccharum robustum

Genetic analysis was performed in a population composed 01 100 F, individuais derived lrom a cross between a cultivated sugarcane (S. officinarum 'LA Purple') and its proposed progenitor species (S. robustum 'Moi 5829'). Various types (arbitrarily primed-PCR, RFLPs, and AFLPs) 01 single-dose DNA markers (SDMs) were used to construct genetic linkage maps lor both species. The LA Purple map was composed 01 341 SDMs, spanning 74 linkage groups and 1,881 eM, while the Moi 5829 map contained 301 SDMs, spanning 65 linkage groups and 1,189 eM. Transmission genetics in these two species showed incomplete polysomy based on the detection 01 15% 01 SDMs linked in repulsion in LA Purple and 13% 01 these in Moi 5829. Because 01 this incomplete polysomy, multiple-dose markers could not be mapped for lack 01 a genetic model lor their segregation. Due to inclusion 01 RFLP ancho r probes, conserved in related species, the resulting maps will serve as uselul tools lor breeding, ecology, evolution, and molecular biology studies within the Andropogoneae.


INTRODUCTION
Saccharum L. is part of a polyploid complex within the Andropogoneae tribe. Cultivated forms of Saccharum (sugarcane) are most notably used for sugar and alcohol production worldwide, especially in the tropics. Sugarcane is the most genetically complex crop for which genome mapping has been achieved (AI-Janabi et al., 1993;daSilva et al., 1995). Polyploidy in Saccharum is widespread and is largely responsible for its genetic and taxonornic complexity. Studies using DNA markers and molecular cytogenetics revealed polysornic inheritance and octoploidy (x = 8) within S. spontaneum (2n = 64, from India) (AI-Janabi et al., 1993;daSilva et al., 1995;D'Hont et al., 1996). The basic chromosome number and levei of ploidy have not been conclusively determined for other Saccharum species.
Due to its genetic peculiarities, molecular genetic markers cannot be applied to sugarcane as they are to most plants. Use of DNA markers has recently allowed genetic mapping in polyploids (daSilva and Sobral, 1996). A novel genetic approach to direct mapping of polyploid plants was proposed by Wu et al. (1992). This approach is based on single-dose markers (SDMs). SDMs are present in one parent, absent in the other parent, and segregatel: 1 in the progeny. More recently, daSilva (1993) and Ripol (1994) presented a methodology for mapping multiple dose markers in polysornic polyploids, which greatly improved the accuracy of identification of homology groups (daSilva et al., 1993(daSilva et al., , 1995. Restriction fragment length polymorphisms (RFLPs) were the first DNA markers used to construct genetic maps of higher organisms (Botstein et al., 1980). DNA fingerprinting methods, based on amplification of random genornic DNA fragments by arbitrarily selected primers (Welsh and McClelland, 1990;Williams et al., 1990), have also been used for genetic mapping (Al-Janabi et al., 1993) among other applications (Welsh et al., 1991). More recently, amplified fragment length polymorphisms (AFLPs), a technique based on selective PCR amplification of genomic restriction fragments, have provided another very powerful tool for genornic research (Vos et al., 1995). When mapping with single-dose polymorphisms, all bands are scored as dominant markers, therefore the typical advantage of RFLPs, namely codominance of markers, is lacking. Thus, PCR-generated markers with an inherently higher data output per unit labor are good choices for generating and saturating linkage maps Vos et al., 1995). However, RFLPs remain the most informative marker to determine homologous relationships among chromosomes within Saccharum and among grasses, including maize and sorghum.
S. officinarum is a domesticated species, which is thought to have been derived primarily from S. robustum, a wild species in Papua New Guinea (Brandes, 1929). We herein report the development of SDM linkage maps for each of these species using RFLP-and PCR-based markers for progeny of an interspecific cross. These maps have also been used in comparative studies among sugarcane, sorghum and maize, and in the analysis of quantitative traits in these two species (Guimarães et al., 1997).

Plant materiais
Plant materials were kindly provided by the Hawaiian Sugar Planters' Association (Aiea, HI). The population consisted of 100 FI individuals produced by crossing S. offieinarum 'LA Purple' as female with 5. robustum 'MoI 5829'. Cytological evaluation of the population showed that parents and progeny displayed strict bivalent pairing at meiosis and had 2n = 80 chromosomes, as described previously by Al-Janabi et al. (1994a).

RFLPs
Genomic DNA was extracted according to the method of Honeycutt et ai. (1992). Fifteen ug of genomic DNA from parents and 100 progeny was restricted individually with DraI, EeoRI, HindID, and XbaI, and resolved in agarose gels. The gels were blotted and Southem hybridization was performed according to daSilva (1993). After hybridization, blots were exposed to BioMax film (Kodak) at -80°C for 3 to 7 days depending on signal intensity. One hundred and ninety probes were surveyed against parental DNA blots digested individually with the four enzymes to identify scorable polymorphisms. Subsequently, probes were hybridized to genomic DNA blots of the FI population that had been digested with the appropriate enzyme.

Arbitrarily primed PCR
Genomic DNAs from parents and progeny (25 ng) served as templates for thermal cycling in a System Cycler 9600 (Perkin Elmer), using the protocol described by Sobral and Honeycutt (1993). Arbitrarily primed PCR products were resolved on either agarose or polyacrylamide gels. Agarose gel electrophoresis and recording and

Selective restrietion fragment amplifieation
AFLPs (Voset al., 1995) were generated using AFLP Analysis System I (Gibco-BRL). Two hundred and fifty ng of genomic DNA from parents and progeny was simultaneously digested to comp\etion with EeoRI and MseI. Restricted genomic DNA fragments were ligated to EeoRI and MseI adapters, diluted 1:10, and pre-amplified using AFLP core primers, each having one selective nucleotide. Preamplification products were then diluted to 1:50 and used as a template for selective amplification using the combinations of MseI-and EeoRI-specific primers, each containing three selective nucleotides. EcoRI-se\ective primers were labeled with 'f2P-ATP before amplification. The thermal profile for both steps of amplification, primer labeling, and selective primer combinations were performed as recommended by the manufacturer. The selective amplified products were resolved by electrophoresis in denaturing polyacrylamide gel, as described for arbitrarily primed PCR.

Marker identification
RFLPs were named by using the original probes' identification (ume, bnl, isu, esb, esc, esr or sg), followed by the first letter of the restriction enzyme used (d, e, h, or x for DraI, EeoRI, HindID, or XbaI, respectively), followed by a period and the molecular size (in base pairs). Size was a single-gel estimate calculated by linear regression and standardization against a l-kb ladder (Gibco, BRL) for each blot. Arbitrarily primed-PCR polymorphisms were named using the Operon denomination (from A to Z and from 1 to 20), or the RY-repeat primer designation (CG6-CG9), followed by a period and the molecular size (in base pairs). The arbitrarily primed PCR polymorphisms that are followed by the letter p were resolved in denaturing polyacrylamide gels, while the others were resolved in agarose gels. AFLPs were coded by the EeoRI (E) and MseI (M) selective primer combination and the respective molecular size (in base pairs).

Linkage analysis
Polymorphisms were scored for presence (1) and absence (O), and analyzed for dosage among FI progeny using chi-square tests (P < 0.05), as described by Wu et ai. (1992) and daSilva et ai. (1995). Because of the doublepseudo-testcross mating strategy used (reviewed in daSilva and Sobral, 1996), SDMs are identified in each of the parents, resulting in two maps: one for the male parent and one for the female parent. Linkage relationships among SDMs were deterrnined using MapMaker v 2.0 for the Macintosh (Lander et ai., 1987) by coding the data as haploid (as the population is resultant from a double pseudotestcross mating strategy). SDMs were grouped using a minimum LOD of7.0 and a maximum recombination fraction (r) of r < 0.25 (Wu et ai., 1992). Linkage groups were then ordered using multi-point analyses. Markers at r < 3 cM could not be ordered accurately because of the rel atively small sample size; however, the best possible order was always accepted, even if the LOD score supporting the order was not large. Map distance in centimorgans was calculated using the Kosambi mapping function. Linkages in repulsion phase were determined as described by AI-Janabi et ai. (1993).

Marker distribution and mapping output
The total number of polymorphisms between the two genomes analyzed was significantly different with X2  (Table I). S. officinarum showed a higher level of polymorphism for all marker types. Sixty-eight percent of the polymorphisms generated were single-dose. This percentage is similar tõ the results of daSilva (1993), that found 73% ofpolymorphisms in S. spontaneum to be single dose. Markers generated by different methods were not uniforrn1y distributed across the linkage groups. In LA Purple, 26% of the linkage groups had all three marker types, and in MoI 5829, just 9% of the linkage groups were covered by all types of markers. Lack of uniforrn distribution may be accounted for simply by the different numbers of each type of marker mapped on each genome (Table  I) and the incomplete saturation of both genomes with markers. daSilva et aI. (1995) mapped 208 AP-PCR markers and 234 RFLPs in S. spontaneum 'SES 208', and they did not find significant deviation from a random distribution of the markers among linkage groups.
Of the 190 maize probes surveyed against the parental sugarcane DNA, 131 probes produced a good hybridization pattem. The signal produced with maize genomic and cDNA probes suggests a high degree ofDNA sequence similarity among these species, despite at least 25 million years of evolution since they shared a common ancestor (AI-Janabi et aI., 1994b;Sobral et aI., 1994). A similar result was reported by daSilva et aI. (1993), in which 78% of maize probes surveyed produced a strong signal in S. spontaneum.
Chromosome assortment in S. officinarum and S. robustum Both repulsion and coupling phase linkages were observed in S. officinarum and S. robustum genomes. Fifteen percent of LA Purple markers were detected in repulsion phase and were assigned to 17 linkage groups having at least one repulsion phase SDMs. Similarly, 13% of MoI 5829 were in repulsion phase and were assigned to l l linkage groups.