Nucleotide sequence, structural organization and length heterogeneity of ribosomal DNA intergenic spacer in Quercus petraea (Matt.) Liebl. and Q. robur L.
18S-5.8S-26S rDNA family comprises tandemly arranged, repeating units separated by an intergenic spacer (IGS) that contains transcription initiation/termination signals and usually repeating elements. In this study, we performed for the first time thorough sequence analysis of rDNA IGS region in two dominant European oaks, Quercus petraea and Q. robur, in order to investigate (1) if IGS sequence composition allows discrimination between these two species, and (2) if there is an rDNA length heterogeneity arising from IGS sequence. Two spacer length variants (slvs), 2 and 4 kb in length, were found in the genomes of both species. Inter-comparison of both slvs revealed no species-specificity in sequence or structural organization. Both slvs could be divided into four subregions; (1) the subrepeat region containing three repeated elements, (2) the AT-rich region containing matrix attachment sites and putative origin of replication, (3) the promoter region containing putative transcription initiation site and (4) the 5′ETS region. In the 4-kb slvs all four subregions are extended, and the subrepeat, AT-rich and promoter regions are duplicated. This is unique compared to other known IGS sequences where the variation in number of subrepeats is responsible for slvs creation. We also propose a possible evolutionary scenario to explain the formation of the subrepeat region in oak IGS. Results obtained in this work add to the previous picture of low-genetic differentiation of the two oaks and provide important data for further analyses of the function of IGS in control of rRNA gene expression.
KeywordsIGS Spacer length variants rRNA genes Repetitive elements Quercus petraea Q. robur
Ribosomal DNA (rDNA) is organized as tandem repeating units at one or more nucleolar organizing region(s) (NOR). Each repeating unit consists of coding (18S, 5.8S and 26S rRNA genes) and non-coding (internal transcribed spacers, ITS, and intergenic non-transcribed spacer, IGS) regions. Spacers are often different in closely related species. The greatest sequence divergence was reported in the large IGS that separates 18S and 26S rRNA genes. The IGS is also a region of particular interest, given the presence of transcription initiation sites, terminators of transcription and repetitive elements (subrepeats) shown to enhance transcription of rRNA genes from the adjacent promoter (Grimaldi et al. 1990; Kuhn et al. 1990; Pikaard 1994). Differences in the number and sequence of subrepeats in the IGS account for most of the length variation between rDNA repeat units in closely related species, among populations and even within an individual (Saghai-Maroof et al. 1984; Borisjuk et al. 1997; Reed et al. 2000; Reed and Phillips 2000). Mechanisms such as unequal crossing-over of iterated subrepeats, gene conversion and replication slippage are mostly responsible for creation of intra-specific spacer length variants (slvs) (Suzuki et al. 1986). Here we aimed at (1) determining the sequence and structural organization of rDNA IGS in two oak species, Q. petraea and Q. robur; (2) revealing if there is one or more types of rDNA repeat units due to different slvs in their genomes; (3) finding out if the IGS in Quercus shows typical repetitive structural features described in IGS of other eukaryotes; and (4) finding out whether the IGS region can discriminate these two closely related oaks.
Complete nucleotide sequence and internal structural organisation of IGS have been reported for many angiosperm dicotyledonous species. However, to our knowledge, no data of IGS sequence and structure for angiosperm and gymnosperm tree species has been published, except for olive tree (Maggini et al. 2008). The 28S-18S (18S) rDNA studies in trees have been generally limited to restriction mapping and Southern blot hybridisation, e.g. in Pinus (Cullis et al. 1988), Picea (Bobola et al. 1992) and in angiosperm tree genus Quercus (Bellarosa et al. 1990). The restriction maps for six Mediterranean oaks show several rRNA gene types for each species, resulting from the differences in IGS length. Nevertheless, no data are available for closely related sessile oak, Q. petraea (Matt.) Liebl., and pedunculate oak, Q. robur L., the two dominant oaks in European deciduous forests.
These two oaks are particularly interesting because they frequently hybridize in nature, but still preserve clear morphological features and ecological preferences even in overlapping habitats. Their genomes, on the other side, exhibit a low genetic differentiation and a high degree of allele sharing has been reported (Zanetto et al. 1994; Samuel et al. 1995; Bodenes et al. 1997; Muir et al. 2000; Coart et al. 2002) with no species-specific sequences found so far (Zanetto et al. 1994; Barreneche et al. 1996). Only by using microsatellites and AFLP was a clear differentiation of the gene pools in Q. robur and Q. petraea achieved (Muir et al. 2000; Coart et al. 2002). In addition, the genomes of the two Quercus species are of the same size, and the karyotype features are identical, including the number and position of 18S and 5S rDNA loci (Zoldos et al. 1998, 1999).
In order to better understand the differences between these two species and to set the stage for further analyses of the function of IGS in control of rRNA gene expression, we determined its complete nucleotide sequence and structural organization. We found no species-specificity in IGS of Q. petraea and Q. robur. Both species possessed two main rDNA gene types due to IGS length difference. Based on the presence of special sequence features two spacer length variants could be divided in four distinctive regions, and the length differences of all four regions were responsible for creation of the two slvs.
Materials and methods
Plant material and DNA isolation
Plant material was from a forest breeding station in Našice (north-east Croatia). Acorns of Q. petraea and Q. robur were collected in the location Krndija Našička by a professional forester and seedlings were grown under laboratory conditions. The species identity of individual trees was confirmed using the morphological features of leaves and fruit. Total genomic DNA was isolated from fresh leaves using Qiagen DNeasy Plant Mini Kit (Qiagen Co.) following the manufacturer’s instructions.
PCR amplification, cloning and sequencing
In order to amplify complete IGS region, two universal primers (PIGS1-26S, 5′-GATCCACTGAGATTCAGCCC-3′ and PIGS2-18S, 5′-TGGCAGGATCAACCAGGTAG-3′) were designed using the conserved regions of the 26S rDNA and 18S rDNA sequences, obtained from various plant species; 100 ng of genomic DNA was used as a template for Platinum PCR SuperMix High Fidelity Polymerase (Invitrogen). Cycling conditions were as follows: an initial denaturation of 3 min at 95°C followed by 35 cycles of 30 s at 95°C, 30 s at 58°C and 3 min at 72°C, with a final extension of 5 min at 72°C. The amplified PCR products were gel-purified with Perfect Gel Cleanup kit (Eppendorf). IGS variant of 2 kb was cloned using TA Cloning Kit with OneShot TOP10 Chemically Competent Cells (Invitrogen) following manufacturer’s instructions. Sequencing was performed by service of VBC-Genomics (Bioscience Research GmbH, Wien, Austria) and Macrogen LTD (Seoul, Korea) using vector-specific primers (M13R and T7) and IGS internal primers.
Restriction maps and Southern blot analysis
Restriction enzyme digestion of the 2-kb slvs was performed with several two-enzyme combinations: EcoRI + BamHI, EcoRI + HindIII, EcoRI + XbaI, EcoRI + SalI and EcoRI + SmaI in order to analyse intra-specific and inter-specific IGS variations. IGS length variants between different Quercus individuals were analysed using Southern blotting and hybridisation. Genomic DNA (10 μg) was digested with EcoRI. We used two different probes for Southern hybridization: heterologous 18S rDNA probe and an rDNA probe specific to the oak IGS. The 18S was SalI/SmaI fragment from the Cucurbita pepo 18S rDNA (Torres-Ruiz and Hemleben 1994). The oak-specific probe was a part of a 2-kb IGS sequence from Q. petraea, containing repetitive elements (a 545 bp BstXI fragment of pCR QP-IGS3cl 3). Both probes were labelled with fluorescein-11-dUTP by random priming method. Both probes detected the same bands after Southern hybridisation. The blot was hybridised at 65°C overnight, followed by two stringent washes in 1× SSC, 0.1% SDS at room temperature, and two washes in 0.1× SSC and 0.1% SDS at 65°C. Hybridisation signals were detected with Gene images CDP-Star detection module (Amersham Biosciences).
The EMBOSS suite of bioinformatics tools (available at http://emboss.bioinformatics.nl/) was used for sequence analysis and the BioEdit Sequence Alignment Editor for dot matrix analysis. Secondary structure was predicted using programmes for secondary structure prediction based on the Zuker method from the MFOLD programme. We also used the MAR-Finder (http://www.futuresoft.org/MAR-Wiz), which identifies DNA motifs (ORI motifs, TG-rich motifs, curved DNA motifs, kinked DNA motifs, DNA topoisomerase II recognition elements and AT-rich sequences) representing potential matrix attachment sites, and calculates a probability based on the number and distribution of these motifs. DNA unwinding elements (DUEs) were located using the programme Web-Thermodyn (accessed at http://www.gsa.buffalo.edu/dna/dk/WEBTHERMODYN/), which calculates the energy required to unwind a specific base pair in the context of a sliding window (Huang and Kowalski 2003).
Eleven 2 kb IGS sequences from Q. petraea and one sequence from Q. robur are deposited in EMBL/GenBank Data Libraries under accession nos.: EU555521, EU555522, EU555523, EU555524, EU555525, EU555526, EU555527, EU555528, EU555529, EU555530, EU555531, EU555532. In order to explore the degree of conservation between different repeat units, we sequenced six clones of 2 kb slvs from Q. petraea: QP-IGS3cl 3, QP-IGS3cl 4, QP-IGS3cl 6, QP-IGS3cl 7, QP-IGS3cl 8 and QP-IGS3cl 9 (accession nos. EU555524, EU555528, EU555529, EU555530, EU555531, EU555532). We also used the available 2 kb and 4 kb IGS of Q. robur from GenBank Database (EF208969 and EF208967, respectively) for alignments with the sequences obtained in this work.
Intra- and inter-specific variations in rDNA of Q. petraea and Q. robur
Molecular structures of the 2- and 4-kb IGS from Q. petraea and Q. robur
Motif TGCCC was repeated in both slvs, with the first repetition occurring around 55 bp from the beginning of the IGS. It was repeated 15 times in the subrepeat region, 3 times in the AT-rich and once in the 5′ETS region of the 2-kb slvs. Interestingly, this motif was not clustered, but was repeated irregularly 22 times throughout the entire 4-kb slvs (Figs. 4a, b).
Functional elements and domains
A pyrimidine-rich sequence CCCTCCCCCCTCTCCTCTCCC(C)T, found at the 5′ end of the oakf 2 and 4 kb slvs, was highly similar to the sequences found at the beginning of the IGS of some other plants (Kelly and Siegel 1989; Perry and Palukaitis 1990; Gruendler et al. 1991; Borisjuk and Hemleben 1992; Borisjuk et al. 1997), suggesting a function of transcription termination site (TTS). We designated the first cytosine of this sequence as position +1 of the oak IGS (Figs. 3, 4a, b). Downstream from this sequence, long stretches of C and A bases occurred.
Just downstream from the subrepeat regions 2-A and 4-A1 we identified several copies of imperfect complementary (i.e. antisense) TIS motif. A 9-bp core sequence identity to the 16-bp TTS sequence was found positioned 48 bp from TIS sites in both, the 2- and 4-kb, slvs (Figs. 4a, b). This motif might function as a proximal terminator responsible for readthrough enhancement. Proximal terminators are found upstream of TIS in S. cerevisiae, Xenopus, mouse and other vertebrates, and some plants such as maize and cucumber (review by Moss and Stefanovsky 1995).
Subrepeat and promoter regions were separated by an AT-rich region (Regions 2-B, 4-B1 and 4-B2) in both slvs (Figs. 3, 4a, b). Comparison of the region 2-B (position 466–861, 59.15% A + T), 4-B1 (position 1,519–1,918, 55.06% A + T) and 4-B2 (position 2,463–2,856, 55.61% A + T) revealed sequence homology at levels ranging from 80 to 87%. The AT-rich region consisted of three distinctive domains: AT-short and AT-long domains separated by a 104-bp long GC-block. The MAR-Wiz tool identified the highest probability for scaffold/matrix (SAR/MARs) attachment sites within AT-short and AT-long domains; here we found ORI elements (ATTA, ATTTA, ATTTTA and AAAAn7AAAn7AAAA), curved/bent DNA elements, TG di-nucleotides, DNA topoisomerase II recognition sites as well as intermingled runs of A and T tracts (of 3–6 As, and 3–8 Ts), interrupted by sequences never longer than 10 bp (Figs. 4a, b).
We also used programme Thermodyn to calculate the difference in free energy between the single- and double-stranded DNA within the entire IGS. Sequences with lower free energy requirements for unwinding (compared to adjacent sequences) correspond to DNA unwinding elements (DUEs), and they were identified at positions 701 (94.52 kcal/mole), 749 (94.13 kcal/mole) and 801 (96.35 kcal/mole) within AT-long domain of the 2-kb slvs. They occurred at positions 2,703 (113.64 kcal/mole), 2,744 (96.60 kcal/mole) and 1,801 (100.19 kcal/mole) within the AT-long domain of the 4-kb slvs. These positions were 140–240 and 190–230 bp distant from TIS and TIS1 in the 2- and 4-kb slvs, respectively. The DUE minimums in both slvs corresponded to two alternating purine/pyrimidine sequences longer than 20 bp containing four T-tracts (T7, T8, T5 and T6) and three A-tracts (A4, A6 and A3) (Figs. 4a, b). Such sequence composition greatly favours DNA bending by minimizing base stacking interactions. In addition, we found a cluster of eight near matches and one perfect match to the 11-bp (A/T)TTAT(A/G)TTT core consensus sequence (ACS) of the yeast autonomously replicating sequence (ARS) within the potential bent locus. The ACS has been found within the origins of replication in different eukaryotes, including plants (Hernandez et al. 1993).
The distance from the putative oak TIS to the first nucleotide of the 18S rDNA coding region was 1,052 bp in the 2-kb slvs (Region 2-D, positions 947–1999, 60.97% G + C), and this distance was 1,302 bp in the 4-kb slvs (Region 4-D, positions 2,942–4,244, 61.78% G + C). These comprised the oak 5′ external transcribed spacer (5′ETS) (Figs. 3, 4a, b). Regions 2-D and 4-D shared over 90% sequence identity. We identified a 909-bp CpG island (starting at position 1097) in the 2-kb slvs and a 1,062 bp CpG islands (starting at position 3,190) in the 4-kb slvs. The islands were rich in G + C bases and TG di-nucleotides. The sequences TTACCC in the 2-kb slvs and TTGCCC in the 4-kb slvs, located about 70 bp upstream from the first nucleotide of the 18S rRNA gene, represented potential splice sites.
Three possible CpG methylation sites were found within each of the promoters. CpG sites were found at positions −24, −8 and +17 within the 2-kb slvs, and at positions −34, −24 and +17 within the 4-kb slvs relative to the initiating A of the gene promoter (Figs. 4a, b). These were at the same relative positions within the duplicated spacer promoter. Up to 17 CpG sites were found at positions from −114 to −387 relative to the initiating A in both slvs, the location that corresponded to the AT-rich region.
MFOLD, which uses free-energy minimalization to predict secondary structures, revealed a potential of the entire 2- and 4-kb IGS regions to form strong and extensive secondary structures. We submitted the subrepeat region and 5′ETS region separately, and promoter and AT-rich region together to higher-order prediction. This grouping was according to presumed biological relevance: the subrepeat region most probably gives rise to non-coding RNA transcripts involved in transcriptional regulation, the promoter and the AT-rich region are involved in initiation of replication and transcription, while the 5′ETS region is a part of 45S pre-rRNA transcript.
Heterogeneity in the IGS of Q. petraea and Q. robur
In order to explore degree of conservation between the IGS of different rDNA repeat units, the complete spacer sequences of six clones of Q. petraea individual QP-IGS3 were aligned (Supplements, Fig. S2). Five of them were of the similar length (1,999 bp–2,003 bp) and corresponded to the main 2-kb slvs, while the clone QP-IGS3cl 4 was 1,770 bp long and corresponded to the double band of 2 kb slvs on Southern blot (Fig. 1). Sequence of this clone was shorter for the entire eight subrepeats, normally located in the inner part of the subrepeat region of other clones. Sequence comparison among six repeating units counted 39 variable sites suggesting an overall high level of identity (Supplements, Table S1).
We aligned the complete 2 kb IGS sequences (Supplements, Fig. S2) and the sequences corresponding to the subrepeat region only (Supplements, Fig. S3) from six Q. petraea and two Q. robur individuals. A high level of identity was found in both cases (>90%). The length of eight complete IGS sequences varied from 1,945 to 2,034 bp, mainly due to variation in the number of repeated elements. Some sequences had small gaps of 2-3 or 6–9 nucleotides, or small insertions of 2-3 nucleotides within the AT-rich region and/or CpG island. The C-tract at the 5′end of the IGS showed length variability (supplements, Fig. S2). Among six different IGS sequences there were a total of 107 nucleotide variations (5.21% divergence, Supplements, Table S1).
This study of rDNA IGS showed that there were no species-specific differences in the sequence and structural organization of IGS in Q. petraea and Q. robur. Comparison of IGS from several different individuals of both species showed over 90% of sequence identity and identical molecular anatomy. The spacer regions represent faster evolving parts of an rDNA cluster compared to coding regions; however, the two oak species share the same ITS (Muir et al. 2000) and IGS (this work), unlike some other white oaks where spacer sequences differentiated between species despite ongoing hybridization (Bellarosa et al. 1990; Whittemore and Schaal 1991; Bellarosa et al. 2005). Thus, the results obtained in this work support the conclusion of Muir et al. (2000) that the split between the two species was too recent for rDNA to have diverged.
Southern hybridization, PCR and sequence analysis revealed two main rRNA gene types within the genomes of Q. robur and Q. petraea and these resulted from the difference in the IGS length. No species-specific IGS length variant was identified, i.e. the same gene types were present in the two genomes. We estimated the entire length of the rDNA repeat units to be 8 and 10 kb, using sequences available in the public database for the 26S rRNA gene (GenBank Acc. no. AY428812) and ITS1 + 5.8S + ITS2 (GenBank Acc. no. AY283026) in Quercus suber, combined with the 18S rRNA sequences from Fagus grandifolia (GenBank Acc. no. AF206910), and using the oak IGS sequences obtained in this work. Two additional rRNA gene types of 7.8 and 10.5 kb were identified by Southern hybridisation in most Q. petraea and some Q. robur individuals. These gene subtypes arose from deletion/addition of several repeats within the subrepeat region of the 2- and 4-kb slvs.
Insight into the rDNA IGS structure is important in understanding the importance of the spacer in achieving appropriate transcription level of rRNA genes. IGS structure has been studied in angiosperms; however, no IGS was characterised for tree species except for olive tree (Maggini et al. 2008). Here, we report for the first time a thorough analysis of IGS structural organization in two oak species. The petraea/robur IGS consists of several functional regions probably involved in initiation of transcription, transcriptional regulation and initiation of replication. The length difference between 2 and 4 kb slvs was due to (1) different number of the repetitive elements within the subrepeat region and these elements might act as rDNA transcription control elements, (2) a large duplication of the entire AT-rich region, which is probably implicated in the initiation of replication and transcription, and involved in rDNA architecture, (3) the almost perfect duplication of the promoter region followed by an insertion of a subrepeat block, and (4) a longer 5′ETS region. Most studies consider type and copy number variations of repetitive elements as the main reason for intra-specific IGS length heterogeneity (reviewed by Moss and Stefanovsky 1995). Here, we report that the length difference of all four distinct regions, and not only the subrepeat region, is responsible for creation of the two petraea/robur slvs analysed in this work.
Repeated elements were found grouped in one or two blocks within the 2- and 4-kb slvs, respectively. In both slvs, the subrepeat region occurred at the 5′-end of IGS, just downstream from the 26S rRNA gene. Relative to promoter, the subrepeats were positioned at 472, 479 and 465 bp upstream from the TIS (2 kb slvs), TIS2 and TIS1 (4 kb slvs), respectively (Figs. 3, 4a, b). Most of the eukaryotic rDNA IGS analysed so far contain subrepeats at the same relative positions as the subrepeats in the oak IGS, i.e. upstream of the gene promoter or multiple promoters, and they are shown to enhance transcription from cis-located Pol I promoter (reviewed by Moss and Stefanovsky 1995). Indeed, transcription enhancement by promoter adjacent repetitive elements seems to be a common feature of rRNA transcription in plants, insects and vertebrates. The repetitive nature of elements A, B and C in the petraea/robur 2-and 4-kb slvs, their location upstream from the gene and spacer promoters and their low level of divergence among clones and/or individuals (Supplements, Fig. S3, Table S1) suggest that these elements might be important as Pol I enhancers. The antisense nature of several copies of the TIS motifs, found downstream from the subrepeat regions 2-A and 4-A1, suggests a potential for transcription of repetitive elements. Indeed, our ongoing experiments have shown that the subrepeat region within the oak IGS is transcribed. Transcripts originating from the spacer promoters are known to regulate the epigenetic state of rRNA genes (Mayer et al. 2006).
The petraea/robur 2 and 4 kb IGS length variants contain only one type of subrepeats and the first subrepeat starts at less than 100 bp apart from the 3′-end of the 26S rRNA gene, similar to simple organization of the maize IGS (McMullen et al. 1986). Closely related A, B and C elements of the subrepeat region in the oak IGS represent one of the shortest repeated elements found in plant rDNA IGS. A high similarity between their core sequences enabled us to propose an evolutionary scenario (Fig. 8a) for the formation of the subrepeat regions in the oak 2 and 4 kb slvs, even though the time points and correct sequences of events that might support the model are hard to predict. We propose two possible scenarios explaining how the promoter sequence (TCTTTAGGGGGGG), after being modified by T–C transition(s), multiple T deletions and T/C insertions, gave rise to an element CCCATGGGGG that might have been evolutionary exploited to establish patterns in the oak IGS subrepeat region, thus contributing to overall IGS variability and creating elements that might entail specific functions. We base our assumptions on evolutionary studies in Xenopus, Drosophila and Mus, which show that subrepeat region within their IGS were partly, if not entirely, created from partial or full promoter amplifications (review by Moss and Stefanovsky 1995). A part of the modified promoter sequence (CCATGG), regularly found as a part of the element C in oak subrepeats, could have undergone two substitution events (G → C and A → T), thus creating core sequences of the elements A, B and C. This sequence also contains the inner CATG motif, which gave rise not only to complete element A and C after a substitution event, but might also have been used as a starting point for duplication events in the course of evolution of the element B (Fig. 8b). Since this is the first report of the full IGS sequence in Quercus, the proposed evolutionary model should be reinforced with determination of more IGS sequences from both closely and remotely related oak species.
AT-long domain within the AT-rich region of both oak slvs contained only around 32 to 37% GC base pairs. Due to such extreme AT-richness, duplex stability is probably lower here than elsewhere in the oak IGS, which might affect the kinetics of DNA melting during the initiation of replication and/or transcription. Indeed, regions containing stretches of homopolymeric dAdT base pairs, about half a helical turn long and repeated at 10–11 bp intervals, such as found within AT-long domain, result in intrinsically bent or curved DNA molecule identified in various gene regulatory regions (Crothers et al. 1990 and references herein) and in replication (Coffman et al. 2006) and transcription (Miyano et al. 2001) initiation sites. Recently, Coffman et al. (2006) suggested that the highly preferred of the multiple replication initiation sites within the human 43-kb rRNA gene unit is the site characterized by AT-richness and juxtaposition of MARs and DUEs. Individual MAR motifs were found in the entire AT-rich region of the petraea/robur IGS. Complete MAR, DUE and ARS-like sequences were found in close proximity only in AT-long domain, 250 bp in length, preceding the promoter, suggesting that these might be cis-acting elements influencing the activity of origin of replication and transcription. Also, the AT-long domain contains TG di-nucleotides and DNA topoisomerase II recognition sites, which represent SAR/MAR attachment sites known to hold rDNA in appropriate position in interphase nucleus (Gonzalez and Sylvester 1995).
Comparison of IGS from diverse eukaryotes suggested conservation of higher-order structure potential for this rDNA region, which is probably related to evolutionary and functional constraints on chromatin organization, transcriptional regulation and processing of rRNA genes, as well as the stability of transcripts involved in epigenetic control of rDNA loci (Baldridge et al. 1992) The entire oak 2 and 4 kb IGS has the potential to form strong and extensive secondary structures. The most interesting was the higher-order structure able to put the conserved CCAAAAAAGA motif, which delimited the 5′-end of the petraea/robur promoter (and was also found at a conserved position at the border of promoters of different Brassicaceae (Delcasso-Tremousaygue et al. 1988; Rathgeber and Capesius 1990; Gruendler et al. 1991), in close proximity to petraea/robur TIS. The entire higher-order structure might represent a structural element in formation of functional initiating complex through binding of UBF, which recognizes specific DNA structures rather than a sequence (Kuhn et al. 1994). The most extensive secondary structures were found at oak 5′ETS. Indeed, helical elements, likely to have a role in regulation of rRNA transcription and processing, have been found within most eukaryotic rDNA ETS regions (Fernandez et al. 2000; Schnare et al. 2000).
The oak IGS overall GC content (53.83% G + C for the 2 kb slvs and 59.53% G + C for the 4 kb slvs) and the GC content for each of four distinct regions was higher than that of Q. petraea and Q. robur genome average (39.90% G + C, (Zoldos et al. 1998). The GC content of the oak 45S pre-rRNA coding region is not known; however, the GC-richness of the IGS corresponds to Chromomycin-positive NORs in karyotypes of the two species (Zoldos et al. 1999). The subrepeat and 5′ETS regions were the GC-richest regions in the petraea/robur IGS. Indeed, the whole region between TIS and 3′-end of the 18S rRNA gene represents a large CpG island (909 and 1,062 bp in length within the 2- and 4-kb slvs, respectively). CpG islands have GC content significantly higher than that of the genome average; they are nonmethylated and are associated with the genomic regions implicated in gene regulation. CpG islands have also been found within mouse and human rRNA genes (Grozdanov et al. 2003). The IGS base composition in plants is reported only for Arabidopsis. Compared to petraea/robur 5′-ETS, the same region in Arabidopsis IGS is moderately GC-rich, while CpG islands coincide with subrepeats (Gruendler et al. 1991).
Sequence comparison among different petraea/robur individuals as well as among different clones of the single Q. petraea individual showed that single base changes were not evenly distributed within the 2-kb IGS. Most substitutions were located in the AT-rich region; nevertheless, the elements characteristic of the SCAR/MAR sites stayed highly conserved. Also, functional elements such as TTS, promoter including TIS and the potential splice site within the 5′ETS showed high nucleotide conservation (Supplements, Fig. S2). Nucleotide conservation within the subrepeat region was striking (Supplements, Fig. S3, Table S1). It is remarkable that individuals QP-IGS6 and QR-IGS1 lack repetitive elements at the same positions and share the identical nucleotide changes compared to other individuals, even though these two individuals represent different species, suggesting a low differentiation of this genomic region in the two oaks.
rDNA units are prone to homogenization through the process of concerted evolution, whereby one particular rRNA gene type overwrites pre-existing units. There are species showing only one rRNA gene type; however, many species reveal incomplete homogenization or rDNA repeats, so that length variants would be detected within an individual. A very interesting correlation is given recently in the study of Nicotiana allotetraploids. Decondensed and transcriptionally active, nucleolus-associated, rDNA units are vulnerable to recombination processes and thus homogenized, while inactive condensed rDNA loci remain unconverted perhaps because of reduced levels of somatic recombination (Dadejova et al. 2007). Q. petraea and Q. robur have two 18S rDNA loci (Zoldos et al. 1999) and since there are only two main IGS length variants, each of the two loci would contain its own slvs. There are approximately 2,200 copies of rRNA repeat units per diploid genome in both species (Zoldos et al. 1998). The intensity of hybridisation signals after fluorescence in situ hybridisation (FISH) suggests that the major 18S rDNA locus comprises at least double the number of rRNA genes than the minor locus (Zoldos et al. 1999). In Southern hybridisation, using 18S rDNA as a probe, bands corresponding to the 6-kb rRNA gene type (4 kb slvs) are twice as intense as bands corresponding to the 4-kb rRNA gene type (2 kb slvs); thus it is rather likely that the major 18S rDNA locus contains the 4-kb slvs. FISH has shown that the major locus is uniquely associated with nucleolus and with considerable level of decondensation (Zoldos et al. 1999), suggesting transcriptional activity. Indeed, Muir et al. (2000) have shown that only one rDNA family in genomes of Q. petraea and Q. robur is active. Inactivity of the minor locus, possibly containing the 4-kb gene type (2 kb slvs), would not therefore allow homogenization through inter-chromosomal recombination of these two IGS variants in the rDNA of Q. petraea and Q. robur. Indeed, our ongoing study is directed to unravel the molecular organization of rDNA-repeating units within the major 18S rDNA locus after microdissection. Determination of the nucleotide sequence and structural organization of IGS in Quercus, thus, provide a useful data in setting the stage for future analysis of the function of spacer in control of differential expression of rRNA genes within the genome and/or within the major 18S rDNA locus.
This work was funded by the Ministry of Science, Education and Sport of the Republic of Croatia, grants 119-1191196-1224 and 119-1191196-1225. We thank prof. Ž. Borzan for providing biological material.
- Barreneche T, Bahrman N, Kremer A (1996) Two dimensional gel electrophoresis confirms the low level of genetic differentiation between Quercus robur L. and Qurecus petraea (Matt.). Liebl For Genet 3:89–92Google Scholar
- Coart E, Lamote V, De Loose M, Van Bockstaele E, Lootens P, Roldan-Ruiz I (2002) AFLP markers demonstrate local genetic differentiation between two indigenous oak species (Quercus robur L. and Quercus petraea (Matt.) Liebl.) in Flemish populations. Theor Appl Genet 105:431–439PubMedCrossRefGoogle Scholar
- Cullis CA, Creissen GP, Gorman SW, Tiasdale RD (1988) The 25S, 18S and 5S ribosomal RNA genes from Pinus radiata D.Don. In: Cheliak WM, Yappa AA (eds) Molecular genetics of forest trees. Canadian Forest service, Petawawa National forest Institute, pp 34–40Google Scholar
- Samuel R, Pinsker W, Ehrendorfer F (1995) Electrophoretic analysis of genetic variation within and between populations of Quercus cerris, Q. pubescens, Q. petraea and Q. robur (Fagaceae) from eastern Austria. Bot Acta 108:290–299Google Scholar
- Schnare MN, Collings JC, Spencer DF, Gray MW (2000) The 28S-18S rDNA intergenic spacer from Crithidia fasciculata: repeated sequences. Length heterogeneity, putative processing sites and potential interactions between U3 and small nucleolar RNA and the ribosomal precursor. Nucleic Acids Res 28:3452–3461PubMedCrossRefGoogle Scholar
- Suzuki H, Miyashita N, Moriwaki K, Kominami R, Muramatsu M, Kanehisa T, Bonhomme F, Petras ML, Ze-Chang Y, De-Yuan L (1986) Evolutionary implication of heterogeneity of the nontranscribed spacer region of ribosomal DNA repeating units in various subspecies of Mus musculus. Mol Biol Evol 3:126–137PubMedGoogle Scholar
- Zanetto A, Roussel G, Kremer A (1994) Geographic variation of inter-specific differentiation between Quercus robur L. and Quercus petraea (Matt.) Liebl. For Genet 99:111–123Google Scholar