Development of a diagnostic DNA marker for the geographic origin of Shorea leprosula

Study sites and plant material:

Shorea leprosula samples were collected from 14 natural forests in Indonesia: five populations in Sumatra and nine populations in Borneo. As a reference, S. parvifolia populations were sampled from 10 natural forests in Indonesia: four populations in Sumatra and six populations in Borneo. Shorea parvifolia and S. leprosula have very similar wood anatomies (Bosman and Baas 1996), but can be distinguished at locus-specific DNA markers (Nuroniah et al. 2010).

Information on the geographic location is shown in Table 1. Plants were collected from natural stands in an area of 100–300 ha keeping a minimum distance of 30 m from each other. Most samples were collected as leaves of adult trees, saplings or seedlings. Species identification in the field was done on the basis of leaf morphological characters. The leaf samples were dried and preserved in silica gel for DNA isolation. Leaf samples from three sites in Borneo (KB, WkB, MtB) were collected from seeds and identified according to the mother trees.

DNA extraction:

DNA was extracted from dried leaves using the DNeasy^® 96 Plant Kit protocol of Qiagen (Hilden, North-Rhine Westphalia, Germany).

Re-amplification and isolation of AFLP fragments:

AFLP analyses followed the protocol of Vos et al. (1995) with minor modifications (Cao et al. 2006). The original AFLP fingerprint including the AFLP fragment that highly differentiated between Sumatra and Borneo had been generated with the selective primer pair E35/M63 with the three selective nucleotides ACA and GAA, respectively (Cao et al. 2006). To increase the probability of successful excision of the desired fragments from the gel, an additional nucleotide (A, T, C, G) was added to the selective M primer (Brugmans et al. 2003; Gailing and Bachmann 2003). Two different selective amplifications were prepared: (1) using the MseI+4 primer and EcoRI primer labeled with the fluorescent dye Cy-5 for gel-extraction from a 7% polyacrylamide (PAA) gel and (2) using the MseI+4 primer and EcoRI primer labeled with the fluorescent dye 6-FAM for comparison with the original AFLP fingerprint after capillary electrophoresis. AFLP fragments were excised from 7% PAA gels following the protocol described in Nuroniah et al. (2010). Two microliters of the purified AFLP fragments were used for PCR.

A touch-down PCR with the selective primers was carried out (Gailing and Bachmann 2003), and the length of amplified products was analyzed by capillary electrophoresis on an ABI 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) and compared with the original AFLP fingerprint. Fragments having the same length as the original AFLP fingerprint were considered as homologous fragments.

Cloning and sequencing of the AFLP fragments:

The gel-excised fragments were re-amplified and either directly sequenced or sequenced after cloning of PCR products (Nuroniah et al. 2010). Sequences were analyzed with the Staden Package version 16.0 (Staden 1999; http://staden.sourceforge.net/), and sequence alignments were performed in BioEdit version 7.0.9 (Hall 1999) using ClustalW multiple alignment (Thompson et al. 1994). FASTA searches of sequences were done against the EMBL (Cochrane et al. 2008) and NCBI databases (Baker et al. 2000).

Inverse PCR:

Inverse PCR was performed according to Ochman et al. (1988) and Triglia et al. (1988) with slight modifications. Specifically, DNA genomic libraries were prepared by individually digesting 4 μl undiluted DNA with 0.4 μl [2–4 U (unit)] restriction enzymes DraI (5′-TTT↓AAA-3′), HindII (5′-GTY↓RAC-3′), MspI (5′-G↓CGG-3′), RsaI (5′-GT↓AC-3′), AvaII (5′-G↓GWCC-3′), EcoRV (5′-GAT↓ATC-3′), HinfI (5′-G↓ANTC), MluI (5′-A↓CGCGT-3′) and TaqI (5′-T↓CGA). All restriction reactions were then incubated at 37°C overnight, except for TaqI, where restriction was done at 65°C for 2 h. The self-ligations (ligation of the ends of the restricted DNA fragments into circular DNA fragments) were conducted in 50 μl reaction volumes containing 5 μl restricted DNA and 0.1 μl T4-ligase (4U μl−¹). The solution was incubated at room temperature overnight. The ligated products were precipitated with addition of 2.5 μl NaAc (3 M) and 70 μl ethanol (absolute) followed by 25-min centrifugation at 14.000 rpm. The resulting pellet was washed with 80% ethanol, air dried at 37°C for 1 h and re-suspended in 10 μl H₂O. The second restrictions were conducted with 1 U PstI (5′-CTGCA↓G-3′) added to 10 μl of each ligated DNA. The solution was incubated at 37°C overnight followed by incubation at 65°C for 20 min to inactivate the enzyme. Three microliters of this DNA were used as a template for inverse PCR.

Inverse PCR was carried out in 10 μl total volume containing 3 μl DNA, 5 μl HotStar Taq^® master mix Kit [10 mM Tris-HCl (pH 9.0), 1.5 mM MgCl₂, 50 mM KCl, 0.2 mM of each dNTPs, 0.5 U Taq DNA polymerase] (Qiagen, Hilden, North-Rhine Westphalia, Germany) and 5 pmol of each inverse PCR primer. Inverse PCR primers were designed from the resulting AFLP sequence with the Primer 3 software (Rozen and Skaletsky 2000; available at http://frodo.wi.mit.edu/). The following cycling parameters were applied for the touch-down PCR procedure: an initial denaturation of 15 min at 95°C, 1 min at 94°C, 1 min at 58°C and 1 min at 72°C. The annealing temperature was then lowered by 1°C until 54°C during the first five cycles, and then 30 cycles were performed for 1 min at 94°C, 1 min at 58°C, 1 min at 72°C and finished by an extension step of 10 min at 72°C. Inverse PCR products were separated electrophoretically, and the strongest bands were excised and purified using Geneclean^®kit (Q·BIOgene, Carlsbad, CA, USA). The fragments were cloned, and the recombinant plasmids were sequenced after colony PCR.

Primer design and primer optimization:

Primer design for SCAR markers with Primer 3 (Rozen and Skaletsky 2000) and marker screening on an ABI 3100 Genetic Analyzer was done as described in Nuroniah et al. (2010). In addition, restriction of the same amplified fragments (PCR-RFLP) was performed in a final volume of 7 μl with 2.5 μl of PCR reaction and 1 U of restriction enzyme MseI. Digestion products were resolved by electrophoresis in a 2.5% agarose gel.

Amplification of wood DNA using SCAR-172 primers:

Two samples of wood DNA of S. leprosula from Borneo were tested to illustrate the potential of the SCAR-172 marker for identifying the origin of timber. DNA isolated from leaves served as a PCR positive control. The wood DNA was isolated from three different areas based on transverse sections: outer sapwood (outer), the transition zone between sapwood and heartwood (middle) and heartwood (inner) (Rachmayanti et al. 2009).

Isolation and sequencing of the AFLP-172 fragment

The selective amplification using the MseI primers elongated by one base reduced the number of fragments and prevented co-isolation of contaminating DNA fragments (see also Brugmans et al. 2003; Gailing and Bachmann 2003). Thus, this protocol facilitates the excision of the target band from the gel. A new AFLP fingerprinting pattern of much lower complexity with the highly differentiating fragment AFLP-172 present was generated based on the primer combination E-ACA/M-GAAA (Figure 1b). The combinations of E-ACA with M-GAAC, M-GAAG or M-GAAT did not generate the target AFLP fragment (Figure 1c–e). Target bands in the gel were identified by comparing the patterns of AFLP fingerprints on PAA gels and after capillary electrophoresis.

Two AFLP-172 fragments from two samples originating from different geographic regions in Borneo were sequenced (sample 1408_SB and 212_TB). The size of AFLP fragments in this study was 172 bp, including the EcoRI and MseI adapters, but the length of genomic DNA was 146 bp. A high homology between the two samples indicates the absence of co-migrating DNA fragments. Some studies reported that conversion of AFLP to SCAR markers was complicated by contamination of the target fragment with non-target fragments, which co-migrated with the target fragments in the AFLP gel (Meksem et al. 2001).

Amplification and sequencing of the two AFLP fragments from the two populations in Borneo revealed no sequence variation (Suppl. Figure 1). Because of the lack of internal polymorphism, it was necessary to consider a different approach, beginning with the generation of larger DNA sequences including flanking genomic regions.

Inverse PCR

Earlier studies suggested that AFLP fragments with lengths between 150–300 bp needed information about the flanking regions for conversion to SCAR markers (De Jong et al. 1997; Negi et al. 2000). In the present study, the inverse PCR approach was carried out to obtain DNA sequences adjacent to the AFLP marker. Inverse PCR was chosen because of its simplicity and the simultaneous use of the two specific internal primers (Ochman et al. 1988; Brigneti et al. 1997; De Jong et al. 1997; Bradeen and Simon 1998; Qu et al. 1998) instead of genome walking (Negi et al. 2000; Brugmans et al. 2003).

The primers for inverse PCR were designed in such a way that they would amplify the unknown sequence upstream and downstream of the known fragment. The position of the primers was intentionally not placed directly at the beginning and the end of the known sequence (Suppl. Figure 1), so that a part of the known sequence could be used as a control in case of erroneous amplification.

Genomic DNA from two different samples from Borneo (sample 1408 and 212) were digested with nine restriction enzymes. After ligation, the circular DNA served as a template for a touch-down PCR with the primers for inverse PCR (see Suppl. Figure 1). Several strong bands were obtained, and the band obtained from restriction with TaqI was sequenced. From the 146-bp AFLP fragment corresponding to AFLP-172, a fragment of about 430 bp length was obtained and applied for the inverse PCR technique (Figure 2). The sequence was compared to the GenBank database but did not match with any known sequence after a homology search against the NCBI (http://ncbi.nlm.nih.gov/blast) or the EMBL (http://www.ebi.ac.uk/embl) databases.

Figure 2:

Sequences of the AFLP-172 marker with its flanking sites.

The original sequences of AFLP fragment are underlined. MseI restriction sites (TTAA) are in position 61–64 and 408–411, the EcoRI restriction site (GAATTC) is in position 261–266. Primers are shown by number 1–5. Primers 1 and 2 were used for inverse PCR. Primers 3–5 were tested for the development of SCAR markers.

Optimization and testing of SCAR-172 marker

Based on the flanking sequences, primers were designed for amplification of genomic DNA. The sequences of two samples from different geographic regions in Borneo obtained after inverse PCR had different lengths caused by an indel of three nucleotides. For primer development, it was desirable to identify sites that correlated with the presence and absence of the fragment in AFLP fingerprints to develop a SCAR marker that differentiates between Sumatra and Borneo.

Specifically designed primers that ended in diagnostic nucleotide positions have been used successfully to differentiate species of sturgeons (Birstein et al. 1998). The optimization of the SCAR-172 marker was performed based on two different approaches: (i) using a primer with the 3′ end identical to the selective nucleotide of the AFLP primers and optimizing the annealing temperature, (ii) using primers flanking the MseI restriction site and visualization of length differences of PCR products and by PCR-RFLP with the restriction enzyme MseI.

1. The forward primer (Figure 2, primer 3) was designed at the beginning of the sequence because of the presence of an indel near the EcoRI restriction site (position 256–258). The reverse primer was designed (1) with its 3′ end identical to the four selective nucleotides of MseI primer (Figure 2: primer 4) and (2) adjacent to the MseI restriction site so that the amplification product included the MseI restriction site (Figure 2: primer 5). In the present study, a gradient PCR with different annealing temperatures ranging from 60.8°C to 70°C with the primer pair 3 and 4 was tested in order to generate amplification conditions that would allow the amplification of products for samples from Borneo but not from Sumatra. Annealing temperature was raised in successive experiments, but it was not possible to establish a specific annealing temperature.

2. The approach was changed, and the reverse primer was designed to include the MseI site (Figure 2, primers 3 and 5). Consequently, the SCAR primers amplified both samples from Sumatra and Borneo.

Preliminary results showed that samples of S. leprosula from Sumatra had only one band with a fragment length of 427 bp. Samples from Borneo had more than one allele with fragment lengths of 418 bp and 421 bp. Both homozygotes and heterozygotes were observed in samples from Borneo. Thus, a dominant AFLP marker was converted into a co-dominant marker (Figure 3); this SCAR marker was designated as SCAR-172 and was tested in all populations from both Borneo and Sumatra (Table 1). The probability to convert a dominant marker to a co-dominant marker is not high because the AFLP method primarily generates dominant rather than co-dominant markers. One study based on the AFLP method for linkage analysis found that readily identifiable co-dominant markers ranged from 6% to 12.6% among all polymorphic AFLP markers (Waugh et al. 1997). Santos and Simon (2002) reported that only one out of eight (12.5%) SCAR markers developed from AFLP markers can be converted into a co-dominant marker, whereas Paran and Michelmore (1993), who introduced SCAR markers, developed three co-dominant markers from nine RAPD markers.

Figure 3:

Capillary electrophoresis of the PCR products obtained after amplification using SCAR-172 markers.

Genotyping of the SCAR-172 marker

SCAR markers were used to screen 151 out of 163 samples (because of lack of amplification) from 14 populations. One out of 53 investigated trees from Sumatra was homozygous for a rare allele with a length of 426 bp, whereas the remaining 52 trees were homozygous for the allele with the fragment length of 427 bp. Alleles 418 bp and 421 bp were found as common alleles in samples from Borneo. Three out of 98 samples of Borneo had genotypes with the rare alleles, i.e fragments with a length of 409 bp, 426 bp and 427 bp, in combination with other common alleles as shown in Figure 4. An analysis of molecular variance (AMOVA) showed that a highly significant proportion (66.7%) of the total genetic diversity is distributed among islands. More importantly, no genotype was shared between populations of both the islands allowing for an unambiguous assignment of samples to either Borneo or Sumatra. Only one sample from Borneo (BkB) was heterozygous for the 427 bp allele (418 bp/427 bp). Additionally, one heterozygous individual for allele 426 bp was found on Borneo (421 bp/426 bp) and one individual from Sumatra was homozygous for the 426 bp allele. While the marker was variable in Borneo, diagnostic markers for populations or regions were not found (Figure 4).

Figure 4:

Genotype distribution of SCAR-172 in 14 populations from Sumatra and Borneo.

Fragments from 16 samples (six samples from Sumatra and 10 samples from Borneo) were sequenced in order to find sequence differences among alleles. The variation was characterized by size variation due to insertions/deletions (Suppl. Figure 2). The single nucleotide polymorphism (SNP), that originally caused the AFLP polymorphism, was found in the MseI restriction site where nucleotide A is replaced by nucleotide G (Suppl. Figure 2: position 412). All samples from Borneo had nucleotide A at position 412, including the rare allele 427 bp in Borneo. Another SNP was found at position 170 bp where samples from Borneo were characterized by nucleotide A (Suppl. Figure 2).

PCR-RFLP of SCAR-172 marker

Size differences after restriction of SCAR markers were visualized as a result of a polymorphism within the restriction site (De Jong et al. 1997; Brugmans et al. 2003). Differences in size of the restricted fragments were expected and observed after restriction because of the difference in the MseI restriction site (Suppl. Figure 2, position 412). The variation in size revealed after restriction completely differentiated samples from Sumatra from those of Borneo. Samples from Borneo (Suppl. Figure 3, pattern 1) have shorter fragments than samples from Sumatra (Suppl. Figure 3, pattern 0).

SCAR primers were used to screen 158 out of 163 samples (because of lack of amplification) from 14 populations. A complete differentiation between Sumatra and Borneo was revealed, including the sample of Borneo that was heterozygous for the 427 bp allele (426 bp/427 bp) which is predominant in Sumatra.

SCAR-172 as diagnostic marker for species S. leprosula

The AFLP fragment 172 bp is present in all S. leprosula samples from Borneo and showed a very strong but not complete differentiation between Borneo and Sumatra (Cao et al. 2006). It was also found in one sample of S. leprosula from Sumatra (586-AS, Asialog Sumatra) and in five samples of S. parvifolia as follows: 298-BB (Batu Ampar Borneo), 564-AS (Asialog Sumatra) 572-AS (Asialog Sumatra), 753-PS (Pasir Mayang Sumatra) and 1167-NS (Nanjak Makmur Sumatra) (Cao et al. 2006). Sample 586-AS has been tested with SCAR-172 (PCR-RFLP) and showed the characteristic pattern for S. leprosula from Sumatra. No amplification of SCAR-172 (primers 3 and 5) was found in the five samples of S. parvifolia and in the other 87 samples of S. parvifolia from ten populations (8–12 samples/population; four populations from Sumatra and six populations from Borneo). Thus, SCAR-172 is suitable as a specific marker to differentiate between the species S. leprosula and S. parvifolia.

The presence of fragment AFLP-172 in S. parvifolia was most likely caused by homoplasy, i.e. the occurrence of non-homologous fragments of the same size. Homoplasy in AFLP fingerprints has already been reported (Shan et al. 1999; Vekemans et al. 2002). A simple test to prove homoplasy was performed based on a set of primers with one extra nucleotide (O’Hanlon and Peakall 2000). Shorea leprosula from Borneo has nucleotide adenine (A) after the selective nucleotides (M+GAAA) (Suppl. Figure 4, A1). Four samples (S. leprosula 586-AS, S. parvifolia 298-BB, S. parvifolia 564-AS and S. parvifolia-1167-NS) also have nucleotide adenine (A) after the selective nucleotides (M+GAAA, Suppl. Figure 4, B1), but two samples (S. parvifolia-572-AS 572 and S. parvifolia-753-PS) have nucleotide cytocine (C) as the fourth nucleotide (M+GAAC, Suppl. Figure 4, C2). Different nucleotides following the three selective nucleotides prove that the fragments have different sequences. Generally, AFLP-172 fragments of S. leprosula from Borneo showed consistently high intensity (high peaks), whereas AFLP-172 of S. leprosula from Sumatra (S. leprosula 586-AS) and S. parvifolia samples showed weak intensities (Suppl. Figure 4, A and B). It might be assumed that different relative intensities (peak height relatively to other peaks) of AFLP fragments in the same size range are an indication of homoplasy.

Amplification of wood DNA using SCAR-172 primers

Only DNA from the outer area of wood discs (sapwood) could be amplified successfully (Suppl. Table 1), even though fragments of about 600 bp length can often be obtained using wood DNA of dipterocarps isolated from the outer and middle area (Rachmayanti et al. 2009). It has been reported that the best results were obtained from sapwood because quantity and quality of DNA extracted from sapwood are higher than from heartwood (Deguilloux et al. 2002, 2003). However, dipterocarp DNA can be contaminated by PCR-inhibitory substances in particular in the outer sapwood reducing the chance of successful amplification (Rachmayanti et al. 2009). The amplification of oak and dipterocarp wood DNA has been reported to be possible even for amplification products up to a size of 1500 bp, but the amplification success rate being negatively correlated to the size of products (Deguilloux et al. 2002; Rachmayanti et al. 2009).