Introduction

The genus Beta is divided into four sections: Beta, Corollinae, Nanae and Procumbentes. Section Beta includes the crop species B. vulgaris which contains sugar beet, fodder beet and chards. The systematics within this economically important genus have been subject to disagreement (Kishima et al., 1987; Santoni & Berville, 1992), but molecular data may provide a solution to the taxonomic problems, as well as information about the possible evolutionary relationships of sections and species within the genus. DNA sequence data are increasingly providing valuable information for evolutionary studies (Olmstead & Palmer, 1994) and, in plants, chloroplast genes and the 18S–5.8S–25S ribosomal DNA have been the main sequences used for such studies. The 18S–5.8S–25S rDNA is attractive for phylogeny reconstruction because of high copy number, universality and diverse rate of evolution within and among component subunits and spacers (Baldwin, 1994). Although the regions of rDNA that encode the mature rRNAs are useful for deep phylogeny inference within angiosperms as a whole, the two internal transcribed spacers (ITS1 and ITS2; Fig. 1) of nuclear rDNA have evolved more rapidly than the coding regions that flank them and are suitable for comparison of closely related taxa. For example, phylogenetic analysis of ITS sequences from several genera in the subtribe Masinae of the Compositae (Baldwin, 1992) has yielded results highly concordant with the phylogeny of species based on chloroplast DNA restriction site mutations. The ITS sequences have also revealed phylogeny in the genus Calycadenia which has close agreement with that based on the interpretation of cytological and morphological data (Baldwin, 1992). In sugar beet, the 18S–5.8S–25S rDNA repeats have been mapped onto chromosomes using fluorescent in situ hybridization (Schmidt et al., 1994). Restriction fragment length polymorphism (RFLP) analysis of Beta species has been carried out using rDNA probes, and variation in restriction sites was detected within the nontranscribed intergenic spacer (IGS) region but not in the transcribed ITS region (Santoni & Berville, 1992).

As a contrasting technique, RAPD (Williams et al., 1990) has been widely used to reveal genetic variation in crops. The technique has been used successfully for revealing polymorphism within species (Demeke & Adams, 1994). For the study of genetic relationships above the species level, the use of RAPD has been criticized for revealing unreliable phylogenies because of possible lack of homology of co-migrating bands (Brummer et al., 1995). However, several studies have used RAPD successfully to reveal relationships at the section level or above in several genera, including Oryza (Martin et al., 1997), Medicago (Brummer et al., 1995) and Rosa (Millan et al., 1996); in each case, the RAPD-derived phylogenies have been found to be in good agreement with those produced using other methods. In beets, RAPD analysis has so far been used in the taxonomic characterization of species/subspecies of wild annual beets within the section Beta (Shen et al., 1996) and five species within section Corollinae (Reamon-Buttner et al., 1996).

Here, we present results using ITS1 sequence data to illustrate the phylogenetic relationships of species in the genus Beta. We also show that similar genetic relationships are revealed using RAPD when species from each of the sections of the genus are analysed, confirming the reliability of RAPD for studying variation between species and sections of a genus.

Materials and methods

ITS (ITS1–5.8S–ITS2) amplification and sequencing

Thirteen accessions from the University of Birmingham Beet Germplasm Collection, representing members of all four sections, were used for ITS1 sequencing (Table 1). In addition, the ITS1 region of Chenopodium album was sequenced to provide an outgroup for the phylogenetic analysis. Leaf tissues from greenhouse-grown beet plants and a C. album plant growing naturally were used for DNA extraction according to the method of Sabir et al. (1992). Between two and 10 plants were sampled for each accession used in DNA sequencing. The yield of DNA was estimated by electrophoresis on an agarose gel (0.7%) along with λ phage DNA standards. PCRs were performed in 25 μL volumes containing 1.0 ng of genomic DNA, 200 μM of each dATP, dCTP, dGTP and dTTP, 2 μM of each primer, 1.0 U of Taq polymerase, 1×ammonium incubation buffer and 2.5 mM magnesium chloride. The primers were ITS5 (5§-GGAAGTAAAAGTCGTAACAAGG-3§) and ITS4 (5§-TCCTCCGCTATATGATATGC-3§) (White et al., 1990; Fig. 1). Amplifications were performed in a thermocycler (Hybaid-Omnigene) programmed as follows: one cycle at 95°C for 2 min; two cycles of 30 s at 95°C, 1 min at 57°C and 2 min at 72°C; two cycles of 30 s at 95°C, 1 min at 55°C and 2 min at 72°C; 31 cycles of 30 s at 94°C, 1 min at 55°C and 2 min at 72°C; and finally 72°C for 5 min. In order to separate the amplified product from the residual primers, 10–12 reaction mixtures from one accession were pooled and subjected to electrophoresis in a 1.2% (w/v) low-melting-point agarose gel using TAE buffer (Sambrook et al., 1989). The desired DNA fragment was cut out of the gel and recovered using a Geneclean II Kit (Bio 101, Vista, CA, USA) according to the manufacturer's instructions. The DNA was used for sequencing using an Applied Biosystem 373A Automatic DNA Sequencer. Only the ITS1 region (White et al., 1990; Fig. 1) was sequenced, using ITS5 as a forward primer and ITS2 (ITS2: 5§-GCTGCGTTCTTCATCGATGC-3§; White et al., 1990; Fig. 1) as reverse.

Table 1 Beta material used
Full size table
Fig. 1
figure 1

Diagram of the organization of the ITS region of the 18S–5.8S–25S nuclear rDNA repeat. Arrows indicate approximate positions of primers for sequencing. Primer names follow White et al. (1990). The nontranscribed intergenic spacer (IGS) between 25S and 18S is not shown.

Full size image

Alignment of ITS1 sequences and phylogenetic reconstruction

The sequences of ITS1 obtained using the two different primers (ITS5 and ITS2) for each of the 13 samples of beet plus C. album were compared and checked. The sequences were aligned by eye using the LINE-UP program of the GCG sequence analysis software package (GCG, 1995). Divergence between ITS1 sequences in pairwise comparisons was calculated using the Kimura two-parameter method (Swofford & Olsen, 1990). In this method, different rates of transversions and transitions are taken into account and the gaps are not scored. The calculation was carried out using the DISTANCES program in the GCG software package (GCG, 1995). Phylogenetic trees were generated using the distances with the GROWTREE program and using both UPGMA and neighbour-joining options.

RAPD analysis

DNA was extracted from 25 accessions representing all the taxa in the Birmingham Beta collection (Table 1) as described above. PCR amplification and electrophoresis were carried out in duplicate following the method used for rice (Virk et al., 1995). The primers used were OPG-12, OPH-16, OPK-10, OPM-17 and OPM-18 (Operon Technology). Except for accessions from section Beta, in which DNA samples used for amplification were pooled, all the other DNA samples were from single plants. The DNA amplification reactions were performed in a volume of 25 μL containing approximately 1 ng of genomic DNA, 200 μM of each dATP, dCPT, dGTP and dTTP, 0.4 μM primer, 1.0 U of Taq polymerase, 1×ammonium incubation buffer and 2.5 mM magnesium chloride. The amplification was performed in a thermocycler (Hybaid-Omnigene) programmed as follows: one cycle of 95°C for 2 min; two cycles of 30 s at 95°C, 1 min at 37°C and 2 min at 72°C; two cycles of 30 s at 95°C, 1 min at 36°C and 2 min at 72°C; 41 cycles of 30 s at 94°C, 1 min at 36°C and 2 min at 72°C; and finally 72°C for 5 min. Aliquots of 16 μL of amplified products were loaded onto 1.4% (w/v) agarose gels for electrophoresis in 0.5×TBE buffer and run at 200 V for about 2 h. Gels were stained with ethidium bromide and photographed under UV light using the IS-500 Digital Imaging System (Alpha Innotech Corporation). Only strong bands that were observed in both duplicate amplifications were scored. RAPD bands showing variation across the 48 samples were used as polymorphic markers and were scored as present (1) or absent (0) for each accession. The data were then subjected to analysis using Jaccard's similarity coefficient, and a dendrogram was generated using UPGMA cluster analysis (NTSYS-pc: Rohlf, 1992).

Results

ITS fragments were efficiently amplified for all the 13 beet accessions and C. album. A band of the expected size (740 bp) was obtained, although some accessions also yielded a faint band of about 480 bp, probably because of weak, non-specific primer binding within the ITS region during PCR. The sequences of ITS1 fragments were aligned and, by comparison of the sequence data with published sequences in other crops (Yokota et al., 1989), the boundaries of ITS1 were determined. Because of insertions or deletions, four gaps of between one and four bases were included in the alignment (Fig. 2). Species in section Procumbentes have 238–239 bp sequences, in section Beta 241 bp and in sections Nanae and Corollinae 243–244 base sequences, whereas ITS1 in C. album is 221 bp long. Omitting base insertions and deletions leaves 211 bases, of which 86 (41%) are variable across the 14 sequences.

Fig. 2
figure 2

Aligned DNA sequences of the ITS1 region of the 18S–5.8S–26S nuclear ribosomal DNA from 13 beet accessions in the genus Beta and an outgroup species (Chenopodium album). Accessions abbreviations are indicated on the left and the gap positions are indicated as dots.

Full size image

Pairwise comparisons of transitions and transversions of the ITS1 regions and the Kimura two-parameter distances were calculated. The distances among pairwise comparisons ranged from 0 to 48.85. The distances between species within sections were usually less (0–2.53) than those between sections, except for those between species in sections Nanae and Corollinae.

The phylogenetic relationships among species were the same in the trees generated using the distance data and both UPGMA and neighbour-joining methods. To illustrate the distances between accessions, a phylogenetic tree was drawn using the data from the neighbour-joining analysis (Fig. 3). Three groups were formed: species from section Procumbentes formed a distinct group, species from section Beta formed another and species from sections Nanae and section Corollinae formed the third group (closer to Beta than to Procumbentes). In section Beta, two accessions from B. vulgaris and one accession from B. maritima did not show any variation in the ITS1 region and grouped as one. Beta macrocarpa is more distantly related to B. vulgaris than is B. adanensis. In section Procumbentes, two accessions from B. procumbens were separated; one of them had the same sequence as B. webbiana and the other differed by six bases. In section Corollinae, B. macrorhiza and B. trigyna did not show sequence variation in the ITS1 region and clustered together. Beta nana, the single species in section Nanae, was grouped with species from section Corollinae; B. nana was closer to B. trigyna and B. macrorhiza than to B. corolliflora or B. lomatogona.

Fig. 3
figure 3

Phylogenetic tree based upon data using the neighbour-joining method with the Kimura two-parameter distances displayed.

Full size image

Forty-eight DNA samples derived from 25 accessions were analysed using RAPD (see Table 1). The primers used were selections of those that had been used in a previous study (Shen et al., 1996), and all five primers produced reliable and reproducible banding patterns. Accessions from the four sections gave rise to characteristic RAPD profiles, which were so obviously different as to allow identification at the section level by eye. An example of the profiles is shown in Fig. 4. For numerical analysis, 31 polymorphic bands were scored, and the data were analysed using the simple matching coefficient and UPGMA clustering (NTSYS-pc: Rohlf, 1992) to produce a dendrogram (Fig. 5). All accessions from section Procumbentes formed one group (A) well separated from the others. Within this group, plants from each accession were generally clustered together. However, accessions putatively belonging to the same species were not always grouped together. The remainder of the accessions formed another main group consisting of three subgroups: accessions from section Beta formed one group (B), although the four accessions (B0317, B0424, B0588 and B0731) were well separated; accessions from section Nanae showed no polymorphism and formed a tight group (C), which is closer to section Corollinae than section Beta. Within section Corollinae (Fig. 5), accessions from B. lomatogona clustered together and formed a distinctive group (E). Accessions from the other three species grouped together (D) but, although plants from the same accessions did cluster together, the subgrouping did not correlate well with the putative species identification.

Fig. 4
figure 4

RAPD profiles showing variation in band patterns between samples. The amplification was carried out using primer OPM17. M indicates molecular size markers. Lanes 1–24 (from left to right): (a) B0205–1, B0205–2, B0213–1, B0213–2, B0221–3, B0224–2, B0234–1, B0234–2, B0234–3, B0317, B0349–1, B0367–1, B0367–3, B0368–1, B0368–2, B0397–1, B0397–2, B0403–1, B0403–2, B0424, B0534–1, B0535–1, B0535–2 and B0535–3; (b) B0536–1, B0536–2, B0536–3, B0537–1, B537–2, B0555–1, B0555–2, B0566–1, B0566–2, B0566–3, B0576–1, B0576–2, B0576–3, B0576–4, B0576–5, B0588, B0731, B1108–1, B1108–2, B1108–3, FD19–1, FD24–1, FD24–2, FD25–1 (numbers after the hyphens are plant numbers within accessions). Species are identified using the abbreviations included in Table 1.

Full size image
Fig. 5
figure 5

Dendrogram of genetic similarity (Jaccard's) and the UPGMA cluster method using the 48 wild beet samples and 31 RAPD markers. Numbers after the hyphens are individual plant numbers; species abbreviations are also included.

Full size image

Discussion

ITS1 base sequence and RAPD banding data have been obtained for 11 species or subspecies of Beta and for Chenopodium album. Both the phylogeny obtained using the ITS1 sequence data and the genetic relationships revealed using RAPD are in general agreement with the relationships defined using other methods; these include relationships revealed from taxonomically more restricted analyses of RFLP studies of chloroplast (Kishima et al., 1987, 1995), mitochondrial (Senda et al., 1995) and total genomic DNA (Jung et al., 1993) and the distribution of satellite DNA families (Schmidt et al., 1991; Schmidt & Heslop-Harrison, 1993). There is also general agreement with the classifications produced by Santoni & Berville (1992) using variation in four restriction enzyme sites in the IGS region of rDNA in nine Beta species; no differences in restriction sites for the enzymes EcoRI, BamHI, HindIII or SacI were found in the ITS1 region of these species, and this is in agreement with our data.

A number of key points emerge from our results with respect to Beta taxonomy and phylogeny. For example, it is clear from both ITS1 sequence data and RAPD banding profiles that there is a considerable divergence of section Procumbentes from the other three sections; this agrees with results from other molecular studies (Kishima et al., 1987; Mita et al., 1991 ; Santoni & Berville, 1992; Senda et al., 1995). With regard to section Nanae, represented by the single species B. nana, the RAPD banding patterns suggest a lack of genetic variation within this species and that it is closely related to species of section Corollinae. This relationship has also been demonstrated in studies using RFLPs of minisatellite DNA (Jung et al., 1993) and hybridization of satellite DNA family probes cloned from species B. corolliflora and B. trigyna (Schmidt & Heslop-Harrison, 1993). In the studies of Jung et al. (1993), Nanae was grouped within section Corollinae, which is also the case for our ITS1 results, whereas our RAPD data and the study by Schmidt & Heslop-Harrison (1993) showed Nanae to be separated from, but closely allied to, section Corollinae. In either case, the conclusion must be that sections Nanae and Corollinae are closely related phylogenetically.

Section Procumbentes includes three species, B. patellaris, B. procumbens and B. webbiana. Beta patellaris has been regarded as an allotetraploid (Abe et al., 1987;; Senda et al., 1995), with B. procumbens and B. webbiana as diploids. Molecular studies have recently suggested that the last two are extreme forms of the same species (Wagner et al., 1989; Mita et al., 1991; Jung et al., 1993; Kishima et al., 1995 ), a situation originally proposed by Curtis (1968) based on morphological characteristics. Our ITS1 sequence data suggest that one accession (B0535), designated by the original collector as B. procumbens, may actually be B. webbiana. The RAPD results show that considerable variation exists within this section. The subgroupings do not correlate with the original classification into species, which might result from natural gene exchange and could support the single-species hypothesis, and an autopolyploid origin for B. patellaris (Santoni & Berville, 1992). Section Procumbentes is, however, the most genetically distinct of all sections within the genus, and this is supported by both RAPD and sequence data. Phylogenetic analysis of the sequence data, with C. album included as an outgroup to root the tree, indicates that section Procumbentes diverged from the other forms of beet at a relatively early stage in the evolution of the genus, a conclusion that is also supported by Santoni & Berville (1992) .

As for section Corollinae, accessions of B. lomatogona formed a distinctive group using RAPD data, but much variation was observed in the group composed of the other three species. This may, in part, result from difficulty with taxonomic identification within this section. It seems just as likely, however, that it is because B. macrorhiza, B. corolliflora and B. trigyna are closely related phylogenetically. Reamon-Buttner et al. (1996) indicated that these three species are distinct from B. lomatogona and that B. macrorhiza is ancestral to both B. corolliflora and B. trigyna, which have evolved through various hybridizations. If this is the case, then one might expect a lack of discrimination between these species in our RAPD analysis. Santoni & Berville (1992) also separated B. lomatogona from B. macrorhiza, B. trigyna and B. corolliflora using variation in restriction sites within the IGS region of rDNA.

The sequence data provide direct information about mutations in DNA that accompany divergence of species. In further work, we have developed protocols that exploit these ITS1 sequence differences to define specific primers, which allow the identification of Beta taxa using single locus PCR (data not shown). However, in the context of genetic relationships, it is important to remember that the sequence data reflect variation at only a single locus (ITS1), whereas the RAPD analysis used data from 31 loci. A potential problem with the RAPD data is some co-migrating bands, which may not be allelic. The results obtained here indicate that the patterns of interspecific relationships revealed by the RAPD data are broadly similar to those revealed using sequence data, although there are some differences. For instance, the plants of B. macrorhiza and B. trigyna possessed identical ITS1 sequences and were closely linked to B. nana; however, all three were clearly separated using RAPD data. This suggests that the RAPD data may provide a more accurate picture of relationships at the species level. This is supported by the concordance of these results with those of Reamon-Buttner et al. (1996) even to the extent that two forms of B. macrorhiza may exist; the two accessions used here appear to be very distinct. Overall, the results demonstrate the value of using more than one molecular technique for the analysis of genetic relationships.