Introduction

Satellite DNAs are tandemly repeated sequences that are present as long uninterrupted arrays in heterochromatic regions of most eukaryotes (Charlesworth et al. 1994; Ugarković and Plohl 2002). In the insect genus Tribolium (Tenebrionidae, Coleoptera) cytogenetic studies of eight species revealed a conserved karyotype of 20 or 18 chromosomes characterized by the presence of large blocks of pericentromeric heterochromatin (Juan and Petitpierre 1991). Molecular analyses suggested that these blocks are composed almost exclusively of satellite DNAs that comprise up to 40% of the whole genome and encompass the regions of functional centromeres (Ugarković et al. 1996; Durajlija Žinić et al. 2000). Tribolium species have sequence specific satellite DNA profiles: in each species a single highly represented satellite is detected and the satellites exhibit no significant sequence similarity except common structural features in the form of stable dyad structures and A + T rich blocks (Plohl et al. 1993; Ugarković et al. 1996; Mravinac et al. 2004). Due to the high turnover of Tribolium satellite DNAs it may be expected that the divergence of their profiles precedes the speciation process and could be detected even at the population level.

In the red flour beetle T. castaneum, TCAST satellite DNA has been previously characterized as the major satellite that comprises 17% of the whole genome (Ugarković et al. 1996). Since the haploid genome size of T. castaneum is 0.2 pg or 204 Mb (Alvarez-Fuster et al. 1991), this corresponds to around 80,000 copies (Ugarković et al. 1996; Wang et al. 2008). The satellite has a monomer size of 360 bp as revealed by sequencing of five randomly picked monomers, and encompasses centromeric as well as pericentromeric regions of all 18 chromosomes (Ugarković et al. 1996). Although the genome sequencing project of T. castaneum has been recently completed (Richards et al. 2008), most of the heterochromatic regions that are prevalently composed of TCAST satellite were excluded from the genome sequence due to technical difficulties associated with sequencing and assembling highly repetitive regions. Therefore, the number of clones in the GenBank sequence database that contain tandemly arranged TCAST repeats is very limited. Satellite DNA sequences characterized in different Tribolium species seem to be highly homogenous with the average variability of approximately 2% and absence of subfamilies, as revealed by comparison of randomly cloned monomers (Mravinac et al. 2004).

Tribolium castaneum is an important pest closely associated with stored cereal products or living in flour or feed mills. Although T. castaneum originated from India (Hinton 1948), it has been widely scattered by man, so that now it is considered cosmopolitan. For the study of T. castaneum satellite DNA sequence evolution at the population level, we used set of wild-type strains originating from diverse geographic locations in North- and South America, Europe and Japan (see “Materials and methods”). These locations are characterized by distinct environments ranging from continental and humid climate (Indiana, Missouri, USA), Mediterranean climate (Jerez, Spain) to subtropical and tropical climate (Kyushu, Japan, Peru, Ecuador). Samples of beetles were collected in a period from 1982 to 2010, and maintained as laboratory cultures.

In order to get more comprehensive insight into the structure, organization and variability of T. castaneum abundant satellite DNA we decided to sequence satellite DNA directly from genomic PCR products. Using this approach the average genomic sequence that represents the overall variability of the repeats is defined and the most variable positions in the satellite DNA are detected and characterized. The same method was previously used for the study of variability of satellite DNAs in frogs (Picariello et al. 2002; Feliciello et al. 2005) and rodents from the genus Ctenomys (Ellingsen et al. 2007; Caraballo et al. 2010). This method has some possible limitations such as primer specificity that can induce bias towards some variants while some others can remain hidden and underrepresented. This limitation can be partially avoided by the use of relaxed annealing temperatures during PCR and multiple primers, some of them with degenerated positions (see “Materials and methods”). The pattern of overall variability of particular satellite DNAs or satellite subfamilies could reveal evolutionary processes affecting these sequences while comparison of variability among different strains and populations of the same species could be useful to follow satellite DNA dynamics from the population to species level. In addition to sequence variability, changes in the organization and abundance of particular satellites can be examined. By combining these approaches, it could be possible to reconstruct the scenario leading to the formation of species-specific heterochromatin, characteristic for Tribolium species.

Using direct genomic sequencing of TCAST satellite we showed that two related subfamilies Tcast1a and Tcast1b, together make up a significant portion of the genome and represent the major DNA component of heterochromatin of T. castaneum. The pattern of variability of the two subfamilies was characterized, as well as the organization and compartmentalization into more complex repeat units. Comparison among ten laboratory strains from different geographical locations revealed differences in mutational profiles, while no significant difference in the organization and copy number of the satellite subfamilies was detected.

Materials and methods

Beetle strains

The following strains of T. castaneum were used: GA-2 strain, originally used in the genome sequencing project and deriving from North American wild-type strain collected in Georgia in 1982, obtained from Dr. Dick Beeman (Manhattan, KS, USA); laboratory strain purchased from the Central Science Laboratory (CSL, Sand Hutton, York, UK) of European origin; seven wild-type strains obtained from Dr. Joachim Kurtz (Germany): 43—collected at Kyushu Island, Japan in 1988; 50—collected in Schegel Farm, Indiana, USA in 2005; 51—collected in Adrian, Missouri, USA in 2006; 52—collected in Bloomington, Indiana, USA in 2006; 55—collected in Jerez, Spain in 1991; 57—collected in Perù, in 2002; 61—collected in Banos, Ecuador in 1996; and a wild population collected near Zagreb, Croatia in 2010.

Each laboratory stock was established and maintained as a separate culture at a population size of >200 individuals on standard medium (20:1, flour:brewer’s yeast, by weight) in a dark incubator at 24°C and approximately 70% relative humidity.

PCR amplification

Because of the tandem organization of satellite DNA, whole Tcast1a and Tcast1b repetitive units could be amplified by PCR using primers with the same origin in the repeat sequence, but with opposite orientation. Different sets of primers were used (see Fig. 1 for schematic positions):

Fig. 1
figure 1

Schematic representation of monomers of satellite DNA subfamilies Tcast1a (specific region is marked in red) and Tcast1b (specific region is marked in yellow) and positions of PCR primers used for determination of genomic sequence as well as for study of organization of satellite repeats

Full size image

  • pr1 (GGACCAACCATAAGCGAG),

  • pr2 (GGTCYTACAARAAAAATGC),

  • pr3(CGTCTCAAACATTGAAAACTG),

  • pr4 (GACGAAGACGAAAATGCAGC),

  • a1(GCAARCGCTTWATTAAGGTAACAG),

  • a2(GCGYTTGCATATAAGCARATTTAG),

  • b1(CATTTTTGTTCWGTAACWGTCACAG),

  • b2 (GTAATTCAGTTTTCAATGTATGAG),

  • a3(CAGAATCGTTTTTAGTCGTC),

  • a4(ACTGTTACCTTAATAAAGCG),

  • b3(TAAATGCTGCAGTTTCGTC),

  • b4(GTAAAAACTGTTACCGTTTC),

  • a5(CGAGATATAAGCAARAAACC)

  • b5(TATAAGCAAAAAACCAAAGTG).

The primer pairs indicated with “a” or “b” are specific for Tcast1a and Tcast1b respectively. The PCR reaction consisted of a classical 40 cycles, starting from 30 to 50 ng of genomic DNA with annealing temperature of 55 or 52°C, depending on primer pairs.

Genomic sequencing

The average genomic sequence is obtained by direct amplification of genomic DNA using primer pairs corresponding to different regions of satellite monomer. The direct and inverse sequence electropherograms of each amplified repeat were compared and only bases present in both orientations were considered. The relative amount of each base in the positions containing two or more bases was considered to be proportional to the signal height. The average genomic sequences were converted into a genomic consensus sequence using IUBMB single letter code for multiple bases in the same position (e.g. Y = C + T, K = T + G, W = A + T, S = G + C, R = A + G, M = A + C, D = A + T + G, H = A + T + C). A minor base was included in the genomic consensus sequence only if it gave a signal at least one-tenth that of the main base. The most common sequence (MCS) was obtained by indicating the predominant base found at each variable position. If there was no predominant base at a particular position, and if a minor base gave at least 40% of the principal signal height, IUBMB single letter code for multiple bases was used. The sequencing was performed using the automatic sequencer ABI Prism 310 (Applied Biosystems).

Southern and dot blot hybridization

DNA was extracted from 50 to 100 mg of adult insects, which corresponds to 10–20 individuals, following the instruction of DNeasy Blood & Tissue Kit (QUIAGEN) for high molecular weight DNA purification. DNA was digested with different restriction enzymes (Fermentas-Life Science and SibEnzyme) under standard conditions, transferred to membrane and hybridized under high stringency conditions at 68°C.

For quantitative dot-blot analysis DNA samples at 1 μg/ml in denaturing buffer (0.4 M NaOH, 1 M NaCl) were serially diluted with an equal volume of herring sperm DNA (Promega) at the same concentration of 1 μg/ml in denaturing buffer. As a standard, a purified sample of a specific PCR-amplified Tcast1a or Tcast1b from genomic DNA was used at an initial concentration of 0.4 μg/ml + 0.6 μg/ml of herring sperm DNA in denaturing buffer. A 100 μl aliquot of each dilution was added per slot in a Dot-Blot apparatus (Bio-Rad), filtered on a nylon membrane Nytran N (Amersham Hybond—GE Healthcare) and fixed on a membrane by baking at 80°C. Hybridization and staining procedures were the same as those used for Southern blot analysis.

Biotinylated probes of the Tcast1a or Tcast1b satellite DNAs were obtained by PCR amplification of genomic DNA using primer pairs specific of each satellite DNA subfamily, respectively a3 + a4 and b3 + b4. The PCR conditions were: 95°C for 3 min; 30 cycles at 94°C for 30 s; 52°C for 10 s and 70°C for 10 s; 2 min at 72°C. The Tcast1a and Tcast1b repeat specific regions amplified from genomic DNA were isolated by gel electrophoresis and used as a template to prepare the subunit specific biotinylated probes by PCR, using bio-16-dUTP (Roche). The PCR conditions were: 95°C for 3 min; 30 cycles at 94°C for 30 s; 52°C for 30 s and 70°C for 1 min; 2 min at 72°C.

Results

Genomic consensus sequences and the most common sequences of Tcast1a and Tcast1b subfamilies

The average genomic sequence of TCAST satellite DNA was determined using the primer pairs shown in Fig. 1. Amplification was performed on T. castaneum DNA isolates from the GA-2 strain used for the complete genome sequence. The initial amplification was performed using primer pairs pr1 + pr2 and pr3 + pr4 that were constructed according to the previously cloned TCAST satellite monomers (Ugarković et al. 1996) as well as clones extracted from the whole genome shotgun (WGS) assembly of T. castaneum, using TCAST consensus as a query sequence. Sequencing of PCR product obtained by primer pairs pr1 + pr2 and pr3 + pr4 revealed a mixture of two TCAST satellite subfamilies named Tcast1a and Tcast1b. The subfamilies significantly differ in sequence within the approximately 100 bp located in the central region of the monomer, as indicated in Fig. 1. In order to better characterize the genomic sequence of each subfamily primer pairs a1 + a2 and a3 + a4 specific for the Tcast1a, as well as b1 + b2 and b3 + b4 specific for Tcast1b were constructed and used for the amplification. Amplification using primers b1 + b2 specific for Tcast1b gave a band of dimeric size more prominent than the monomer, and direct sequencing of the dimer revealed that it is composed of Tcast1b and Tcast1a subunits. This indicated that the TCAST satellite is prevalently organised into composite Tcast1a-Tcast1b repetitive units (Fig. 1). Primer pairs a3 + a5 as well as b3 + b5 were used to confirm the sequence in and around the specific Tcast1a and Tcast1b regions respectively. Primer pairs a3 + a4 and b3 + b4 were used also for the preparation of Tcast1a and Tcast1b specific probes.

The average genomic sequences of Tcast1a and Tcast1b were converted into genomic consensus sequences in which only the major variable positions were indicated (not shown). IUBMB single letter code for multiple bases in the same position was used and a minor base was included in the genomic consensus sequence only if it gave a signal at least one-tenth that of the main base. Genomic consensus sequences contained usually only two bases at each variable position and the most common sequences (MCS) were obtained by indicating the predominant base at each variable position (Fig. 2). Only at one position (237) in Tcast1a sequence bases A and G were present in almost same amount, and this position was indicated by IUBMB single letter code as “R” in the MCS. The results show that TCAST satellite DNA exists in the form of two subfamilies, Tcast1a and Tcast1b that have consensus sequences of 377 and 362 bp, respectively. Tcast1b corresponds to the previously characterized TCAST consensus sequence (Ugarković et al. 1996) while Tcast1a represents a new subfamily. Tcast1a and Tcast1b have mutual identity of 79%, but each has a sequence-specific region between positions 170 and 260 nt (Fig. 2).

Fig. 2
figure 2

Alignment of the most common sequences (MCS) of Tcast1a and Tcast1b satellite subfamilies. The recognition sites for the restriction enzymes that are either present in the sequences or can be created by a single point mutation (indicated in red) are underlined. Part of the sequence specific for the Tcast1b subfamily is shown in red

Full size image

Variability and organization of Tcast1a and Tcast1b subfamilies

The organization and sequence variability of Tcast1a and Tcast1b satellite subfamilies were studied by Southern hybridization of DNA isolated from the GA-2 strain. Southern analysis represents a powerful method to validate the “genomic sequence” obtained by direct sequencing of a mixture of PCR-amplified repeats. Genomic DNA was digested with six restriction enzymes that recognize sequences of six base pairs: AatII, Eco47III, NsiI, PaeI, PstI and ScaI. Based on genomic consensus sequences and most common sequences of satellite subfamilies, Eco47III and ScaI have a single recognition site in Tcast1a, a single recognition site for PstI is present in a portion of Tcast1b repeats, while the recognition site for AatII does not occur within any of subfamilies (Fig. 2). Recognition sites for NsiI and PaeI can be created in Tcast1a by single point mutation. Blot was hybridized with mixed biotin labelled probes of Tcast1a and Ttcast1b monomers under high stringency conditions (Fig. 3a). A hybridization ladder-like profile characteristic for tandemly repeated satellite DNA is clearly visible on T. castaneum DNA digested with enzymes predicted to cut within Tcast1a genomic consensus sequence, while in the PstI digest, a faint ladder starting from the satellite dimer is visible (Fig. 3a). The restriction enzymes AatII, NsiI and PaeI do not produce a detectable ladder profile characteristic for satellite DNA. Digestion with restriction enzymes ScaI and Eco47III that cut once in Tcast1a gave a band of monomer size as well as a more prominent band of dimer size (Fig. 3a, lines 2 and 6). Presence of a band of monomer size indicates that a fraction of Tcast1a is organized as a homogenous tandem array. A prominent band of dimer size in ScaI and Eco47III digestions suggests however that Tcast1a is interspersed with Tcast1b, and that such interspersed pattern represents the prevalent type of organization of TCAST satellite in T. castaneum genome.

Fig. 3
figure 3

Organization and variability of TCAST satellite within GA-2 strain of T. castaneum. Southern hybridization analysis was performed a using a mixture of biotin labelled Tcast1a and Tcast1b monomers as a hybridization probe. DNA was digested with: AatII (line 1), Eco47III (line 2), NsiI (line 3), PaeI (line 4), PstI (line 5) and ScaI (line 6). b Southern hybridization analysis using probe specific for Tcast1b subfamily. DNA was digested with SatI (line 1), Eco47III (line 2) and Eco47III + RsaI (line 3)

Full size image

In order to analyse specific organization of Tcast1b we performed Southern hybridization using a probe for the Tcast1b subfamily (Fig. 3b). Digestions with restriction enzyme SatI that cuts once within Tcast1b revealed presence of monomer band (Fig. 3b, line 1), which suggests existence of a fraction of Tcast1b in the form of homogenous arrays. However, prominent multimers relative to the monomer band in SatI digestion indicate the interspersed organization of Tcast1b as a prevalent type of organization. Such an interspersed pattern of organization of Tcast1b and Tcast1a is further confirmed by an almost identical pattern of Eco47III digestion obtained by mixed probes (Fig. 3a, line 2) and a Tcast1b specific probe (Fig. 3b, line 2). Double digestion with restriction enzymes Eco47III + RsaI that both have single recognition site in Tcast1a revealed a strong band of dimer size, pointing also to the interspersed pattern of Tcast1a and Tcast1b.

Southern hybridization analysis (Fig. 3a) revealed a difference in restriction profiles among enzymes ScaI and Eco47III that have a single recognition site in Tcast1a sequence (Fig. 2). In the ScaI hybridization profile, the dimer is the most prominent band as expected according to the highly conserved sequence at this restriction site and due to the interspersed organization of Tcast1a and Tcast1b. Differently, in Eco47III digestion, trimer and pentamer are the prevalent bands, indicating organization of the satellite into more complex higher order repeat units. Such a difference in profiles among the two enzymes could be explained by differences in sequence variability between the two restriction sites. Genomic sequence and most common sequence reveal the presence of a highly variable site (R) within the Eco47III recognition sequence (Fig. 2). According to the genomic consensus sequence, the variability occurs specifically within GCGC sequence present within the recognition site for Eco47III (AGCGCT). The predicted variability was confirmed by Southern hybridization after digestion of genomic DNA with restriction enzymes that recognize specifically this site: Eco47III, HaeII (recognition site RGCGCY) and Hin6I (recognition site GCGC) (Fig. 4a). The complete identity of digestion profiles obtained by the three enzymes demonstrates that the variability is restricted to GCGC sequence, as predicted by genomic sequencing.

Fig. 4
figure 4

Southern hybridization analysis of T. castaneum genomic DNA from strain GA-2, using as a hybridization probe mixture of biotin labelled Tcast1a and Tcast1b monomers. DNA was digested with: a Eco47III (line 2, recognition site AGCGCT), Hin6I (line 3, recognition site GCGC), HaeII (line 4, recognition site PuGCGCPy); b Hin6I (line 2) and GlaI (line 3, recognition site G5metCGC). Lines 1 represent undigested T. castaneum DNA

Full size image

An additional factor contributing to the specific variability in GCGC site could be methylation at a CpG site that blocks all three tested restriction enzymes. Digestion with GlaI, a restriction enzyme that cuts specifically methylated GCGC sequence (G5metCGC), demonstrated the absence of methylation at this site (Fig. 4b) which is in accordance with the previously reported absence of methylation in the T. castaneum genome (Zemach et al. 2010). Southern hybridization revealed that variable GCGC sites are not randomly distributed within the tandem satellite array, but regularly phased within a subset of the satellite. Prominent satellite trimers and pentamers observed in digestions with Eco47III, HaeII and Hin6I indicate presence of a more complex organization of TCAST satellite units in the form of higher order repeats.

TCAST satellite variability among T. castaneum strains

To characterize the overall sequence variability of TCAST satellite DNA in the species T. castaneum, the genomic consensus sequences for TCAST satellite were obtained from nine other laboratory strains/populations, using the same pairs of primers. Distribution of the major variable sites and types of variability at these sites was very similar among all strains, as demonstrated for Tcast1a subfamily (Fig. 5). Usually only two bases exchange at the same position and such a pattern was preserved in most of populations. Although the most common sequences (MCS) of both Tcast1a and Tcast1b subfamilies were identical in all populations and correspond to sequences presented at Fig. 2, the genomic consensus sequences differed in the percentage of the minor bases present at some variable positions (Fig. 5). For example, in the laboratory strain from the CSL, at position 236 of Tcast1a, in addition to a major peak corresponding to the nt A, a T nt peak appears which corresponds in intensity to approximately 20% of the main peak. Since this mutation creates two new restriction sites for NsiI (ATGCAT) and PaeI (GCATGC) (Fig. 2), its presence can be confirmed using Southern hybridization analysis (Fig. 6). In all other strains except strain 61 where T makes approximately 15% of the main peak, T is present below 10%, while in GA-2 strain it is almost absent at position 236, as is confirmed by Southern hybridization (Fig. 3a).

Fig. 5
figure 5

Alignment of the genomic consensus sequences (GCS) of Tcast1a from ten T. castaneum strains/populations: (1) GA-2, (2) 43, (3) 50, (4) 51, (5) 52, (6) 55, (7) 57, (8) 61, (9) ZAG-wild (10) CSL. Genomic consensus sequences are aligned relative to the most common sequence (MCS) of Tcast1a that is conserved in all populations. Changes in the genomic consensus sequences relative to the most common sequence are indicated, while dashed lines indicate identical nucleotides

Full size image
Fig. 6
figure 6

Organization and variability of TCAST satellite DNA within CSL strain of T. castaneum. Southern hybridization analysis was performed using a mixture of biotin labelled Tcast1a and Tcast1b monomers as a hybridization probe. DNA was digested with: AatII (line 1), Eco47III (line 2), NsiI (line 3), PaeI (line 4), PstI (line 5) and ScaI (line 6)

Full size image

Southern hybridization analysis of DNA from the CSL strain digested with the restriction enzymes AatII, Eco47III, NsiI, PaeI, PstI and ScaI and hybridized under the same conditions as GA-2 DNA (Fig. 3a), revealed very similar satellite profiles relative to the GA-2 strain. The only difference was in NsiI and PaeI digestions where a clear ladder pattern is visible in the CSL strain (Fig. 6, lines 3–4). This confirms the partial replacement of A at position 236 of Tcast1a in GA-2 strain with T in the CSL strain, as predicted by direct genomic sequencing.

We also checked by Southern hybridization if the pattern of variability observed at GCGC sequence that corresponds to the Hin6I restriction site is conserved among all tested strains (Fig. 7). Based on genomic consensus sequences (Fig. 5) a major variable site (R) at position 237 in Tcast1a is preserved in all strains and overlaps the Hin6I recognition site. Southern hybridization analysis using mixed probes of Tcast1a and Tcast1b reveals very similar variability at this site among ten strains. The hybridization pattern also indicates the same type of organization of TCAST satellite in all tested strains, as well as very similar amount of this satellite DNA in all strains (Fig. 7). Hybridization of the same blot using probes specific for either the Tcast1a or Tcast1b subfamily revealed almost identical restriction profiles, confirming an interspersed organization of Tcast1a and Tcast1b subunits as the prevalent type of satellite organization in all strains (not shown). In addition to main hybridization ladder composed of Tcast1a and Tcast1b subfamilies, existence of another ladder composed of weak, intermediate size bands is also visible (Fig. 7). Such a ladder could indicate presence of a third TCAST satellite subfamily within T. castaneum genome.

Fig. 7
figure 7

Organization of TCAST satellite in ten strains/populations of T. castaneum. Southern hybridization analysis was performed with a mixture of biotin labelled Tcast1a and Tcast1b monomers. The same amount (1 μg) of DNAs from the following strains/populations was digested with Hin6I restriction enzyme: 50 (line 1), 51 (2), 52 (3), 55 (4), 57 (5), 61 (6), ZAG-wild (7) CSL (8), 43 (9), GA-2 (10)

Full size image

We used dot blot hybridization with the probes specific for Tcast1a and Tcast1b to determine the relative amount of the two subfamilies in T. castaneum genome. The amount of Tcast1b in strain GA-2 is estimated to 15–17% of total DNA matching the previously determined amount of 17% for TCAST major satellite (not shown). Using the hybridization probe specific for Tcast1a, the amount of Tcast1a is estimated to be almost the same as those of Tcast1b, corresponding to approx. 20% of the genomic DNA in the all strains. This quantification shows that the newly described Tcast1a represents a highly abundant subfamily of TCAST satellite and that the total amount of TCAST satellite in the genome is much higher than previously determined, ranging approximately between 35 and 40% of the whole genome.

Discussion

Tandemly repeated satellite DNAs are subject to gene conversion, unequal crossing-over, transposition, slippage replication and other exchanges (Dover 2002). Because mutations that change the sequence of one repeat are less common than recombination-induced replacement of one repeat by another, the repeats resemble each other much more than they would if they had been evolving independently. This phenomenon is known as concerted evolution. Because concerted evolution is usually unbiased and random, every version of repeat has an equal probability of being the one that replaces the others (Malik 2009). This simple mutation-recombination balance model, where recombination is the underlying force that homogenizes satellite repeats, does not seem to be able to completely explain the observed features of satellite DNAs (Ugarković 2005, 2009). Several studies have also pointed out the inadequacy of mutation and recombination alone to explain variability of tandem repeats suggesting that selection must play a role in their evolution (Walsh 1987; Stephan 1989; Charlesworth et al. 1994; Stephan and Cho 1994; Mravinac et al. 2005; Feliciello et al. 2006).

Subfamilies of satellite DNAs or different satellites can coexist within a genome forming a so called satellite library (Meštrović et al. 1998; Ugarković 2009). In the case of T. castaneum, two major abundant satellite subfamilies Tcast1a and Tcast1b are discovered that together make up 35% of the genome. Southern hybridization analysis confirmed the structures of both subfamilies obtained by direct genomic sequencing and revealed their mutual organization. The approach of direct sequencing of genomic PCR products used in this study does not exclude existence of other, less abundant subfamilies of TCAST. Despite limited number of TCAST satellite clones present in GenBank, using Blast search with the most common sequences of Tcast1a and Tcast1b as query sequences, it is possible to retrieve numerous clones highly homologous to each subfamily. This suggests abundance of the two subfamilies within the T. castaneum GenBank sequence database and T. castaneum genome, respectively. The two families are mutually interspersed and a composite Tcast1a + Tcast1b dimer seems to be the prevalent repeat unit. The pattern of digestion with some restriction enzymes indicates that Tcast1a and Tcast1b can form even more complex repeat units such as trimers and pentamers. Satellite subfamilies and higher order repeats are not often found in insects, in particular in beetles, although there are a few examples revealing the presence of complex dimers and trimers in some beetle species (reviewed in Palomeque and Lorite 2008). The interspersed pattern of two related satellites is characteristic for heterochromatin in the related species T. madens (Durajlija Žinić et al. 2000). However, the two satellites of T. madens are not regularly arranged like TCAST subfamilies, but form homogenous arrays of approx. 15 kb in length, composed of a single satellite, and those arrays are mutually interspersed within heterochromatin.

The sequences of the two TCAST satellite subfamilies are relatively homogenous and are characterized by major variable sites that are restricted to some positions and are preserved among ten strains from different geographical locations. However, differences in the percentage of minor bases present at some of the positions among T. castaneum strains were detected. In the case of the Tcast1a subfamily it is possible to follow partial replacement of the bases at a particular position among T. castaneum strains by restriction enzyme digestion. Such replacement and spreading of a new mutation represents the first step in the homogenization process and indicates the effect of concerted evolution on TCAST satellite DNA at the population level (Strachan et al. 1985). However, the major, predominant bases at all positions of either Tcast1a or Tcast1b subfamily were still identical in all strains. Satellite DNA sequence evolution at the population level was studied in different insect species such as cave crickets Dolichopoda schiavazzi (Bachmann et al. 1994), leaf beetle Xanthogaleruca luteola (Lorite et al. 2002), and Drosophila buzzatti species cluster (Kuhn et al. 2003; Kuhn and Sene 2004; de Franco et al. 2006). In all of these cases no specific mutations or other sequence features of satellite DNAs that could discriminate a population or groups of populations were discerned. Most of the analysed satellites exhibited high sequence homogeneity on both intra- and interpopulation levels, as revealed from the comparison of randomly cloned satellite repeats. Since all of these studies were based on the analysis of a relatively low number of cloned repeats that did not represent the whole population of repeats present within a genome, it is possible that slight changes in the mutational profiles remained hidden. Direct sequencing of PCR amplicons of a satellite DNA seems to represent a more suitable approach to study overall satellite DNA variability, as demonstrated here for TCAST satellite. To avoid bias in the estimation of variability due to primer specificity, different pairs of primers were combined and relaxed annealing conditions in PCR used. Comparison of genomic consensus sequences discerned discrete changes in the mutational profiles between populations and demonstrated the effect of concerted evolution during the early phases of satellite sequence divergence. Complete homogenization and fixation of population-specific mutations in satellite DNA was only observed in fishes (Elder and Turner 1994).

Expansions and contractions of satellite arrays within a library could also occur due to unequal crossingover and might contribute to continuous change of heterochromatin content and amount. Addition and loss of heterochromatic satellite DNA repeats is characteristic for strains of Drosophila and correlates with variation in the genome size (Bosco et al. 2007). Significant intraspecific heterogeneity of genome size was also detected in Tribolium species (Alvarez-Fuster et al. 1991). However, our present study did not reveal change in the amount or organization of the two TCAST subfamilies at the population level. This indicates that the genome size heterogeneity is not related to this satellite DNA, although the change in the amount of some other repetitive DNAs could not be excluded. In conclusion, the analysis of major TCAST satellite DNA of beetle T. castaneum reveals that divergence of satellite DNA sequence characterized by a change in mutation profile occurs at the level of the population and precedes the evolution of species, representing the first step in the formation of a population-specific satellite DNA profile.