Higher-order organization and compartmentalization of satellite DNA PIM357 in species of the coleopteran genus Pimelia.

Abstract

The PIM357 satellite DNA family is present in 26 Pimelia taxa (Tenebrionidae, Coleoptera) with endemic congeneric species from the Canary Islands showing higher interrepeat variability than continental ones. In this paper, we compare the repetitive DNA sequences of a Canarian species that has distinct subfamilies of repeat units, P. radula ascendens, with another without such subfamilies, P. sparsa sparsa. The chromosomal localization of the repeat units and the comparison of the variability of randomly cloned monomers to the one estimated by comparing repeat units from dimers and trimers suggest the absence of satellite subfamilies in P. sparsa sparsa. Hence, the repeat units of this species seem to be uniformly and randomly distributed throughout all chromosomes out of one chromosomal pair. On the contrary, P. radula ascendens shows four divergent subfamilies of repeat units supported by several diagnostic nucleotide substitutions. These subfamilies seem to form four distinct repeat units: monomer subfamily 1, monomer subfamily 4 and two higher-order units (dimer linking subfamily 1 and 4, and dimer linking subfamily 2 and 3). Moreover, monomers of subfamily 1 are present in three chromosomal pairs only. We discuss the effect of different potential factors acting in the concerted evolution and the genomic organization of stDNA sequences in these taxa.

Full text

Page 1

Complex structural features of satellite DNA
sequences in the genus Pimelia (Coleoptera:
Tenebrionidae): random differential amplification
from a common ‘satellite DNA library’
J Pons1,2, B Bruvo3, E Petitpierre1,4, M Plohl3, D Ugarkovic3and C Juan1
1Laboratori de Gene `tica, Departament de Biologia, Universitat de les Illes Balears, 07122 Palma de Mallorca, Balearic Islands, Spain;
2Molecular Systematics Laboratory, Department of Entomology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK;
3Department of Molecular Genetics, Ruder Boskovic Institute, Bijenicka 54, PO Box 1016, HR-Zagreb, Croatia;4Departament de
Recursos Naturals, Institut Mediterrani d’Estudis Avanc ¸ats (CSIC-UIB), C/. Miquel Marque ´s, 21, 07190 Esporles, Balearic Islands,
Spain
The major satellites of the nine species of the subgenera
Pimelia s. str. and Amblyptera characterised in this paper are
composed of longer monomers (500 and 700bp) than those
described previously in 26 Pimelia s. str. taxa (357bp, a
sequence called PIM357). Sequence analysis reveals partial
similarity among these satellites and with the PIM357
monomers. The discrepancy between the phylogeny ob-
tained based on three mitochondrial and two nuclear markers
and that deduced from satellite DNA (stDNA) sequences
suggests that the different Pimelia satellites were already
present in a common ancestor forming what has been called
a ‘satellite DNA library’. Thus, the satellite profiles in the
livingspecies resultfrom
sequences from that ‘library’ during diversification of the
a randomamplificationof
species. However, species-specific turnover in the se-
quences has occurred at different rates. They have included
abrupt replacements, a gradual divergence and, in other
cases, no apparent change in sequence composition over a
considerable evolutionary time. The results also suggest a
common evolutionary origin of all these Pimelia satellite
sequences, involving several rearrangements. We propose
that the repeat unit of about 500bp has originated from the
insertion of a DNA fragment of 141bp into the PIM357 unit.
The 705-bp repeats have originated from a 32-bp direct
duplication and the insertion of a 141-bp fragment in inverted
orientation relative to a basic structure of 533bp.
Heredity (2004) 92, 418–427, advance online publication,
3 March 2004; doi:10.1038/sj.hdy.6800436
Keywords: satellite DNA; Pimelia; Coleoptera; repeat unit rearrangements
Introduction
The species of the beetle genus Pimelia (Tenebrionidae,
Coleoptera) are flightless, saprophagous insects occur-
ring in xerophilic habitats in the southern regions of the
Northern hemisphere. Pimelia s. str., one of the five
recognised subgenera of the genus, is a polyphyletic
lineage originating in central Asia and colonising North-
east Africa (Kwieton, 1977). From there, at about 6–
12Mya, a lineage radiated in the North Western region of
Africa, colonising Morocco, the Iberian Peninsula, and
the Canary Islands. The subgenus Amblyptera is thought
to be derived from a lineage of the Pimelia s. str.
Moroccan radiation (Kwieton, 1977).
From previous work, we know that a large number of
congeneric Pimelia s. str. taxa possess a common major
satellite DNA (stDNA) family, which constitutes a
remarkably high percentage of their genomes (30–40%)
and it is localised in the heterochromatic regions of the
chromosomes (Pons et al, 1997, 2000a, b). The tandem
repeats of this family (PIM357) are similar in sequence
showing few insertions and deletions (consensus unit
357-bp long) and very similar nucleotide composition
(6971.4% AþT). PIM357 monomers show an intrinsic
bending related to a fairly conserved periodical distribu-
tion of runs composed of three or more adenines (Barcelo ´
et al, 1997; Pons et al, 2002a). These repeats have evolved
gradually as predicted by the molecular drive model
(Dover, 2002): that is, sequences are accumulating single
mutations and some of the new mutated sequences are
homogenised and fixed replacing completely the old
ones, in a continuous turnover, leading to species-specific
or group-specific stDNAs. Such a model of evolution
could explain as to why DNA divergence studies clearly
cluster PIM357 satellites into three sequence groups,
Iberian-Balearic, Moroccan, and Canarian, which are
mostly in accordance with the species biogeography and
mitochondrial phylogeny (Pons et al, 2002a).
The data presented in this paper concern nine species
of the Moroccan lineage of subgenera Pimelia s. str. and
Amblyptera having different major stDNA families with
longer repeats than the canonical PIM357 unit (500–
700bp), but which still represent a large fraction of
the genome (27–43%, Pons, 1999; this paper). The
monomers deduced for these longer families reveal
marked divergences from each other and also in
Published online 3 March 2004
Correspondence: J Pons, Department of Entomology, The Natural History
Museum, Cromwell Road, London SW7 5BD, UK.
E-mail: joap@nhm.ac.uk
Heredity (2004) 92, 418–427
& 2004 Nature Publishing Group All rights reserved 0018-067X/04 $25.00
www.nature.com/hdy

Page 2

comparison to PIM357 sequences, with long insertions
and deletions, but nevertheless show partial sequence
similarities. The homogeneous length and nucleotide
composition, and experimental evidence of gradual
evolution within PIM357 sequences seems to indicate
that they are a different stDNA family from the divergent
and longer ones. For this reason, those longer sequences
were not included in the analysis of PIM357 sequences
(Pons et al, 2002a).
Dot blot analysis reveals that the monomer of the
major stDNA family of P. monticola (PMON 533-bp long)
is also found at low copy numbers in the genome of other
Pimelia species having different major stDNAs (Bruvo
et al, 2003). Six Pimelia species, from both subgenera,
were screened using primers specifically designed to
amplify PMON sequences by PCR. The phylogenetic
analysis of the amplified sequences revealed the pre-
sence of two groups of sequences: PIM357 in Iberian-
Balearic species plus P. monticola, and the other PIM357,
in African species plus P. fornicata. These results from
Pimelia were additional evidence of the ‘library hypoth-
esis’ postulated by Fry and Salser (1977) and confirmed
experimentally by Mestrovic et al (1998). This hypothesis
suggests that major stDNAs found in a related group of
living species were already present in the common
ancestor of those species, forming a pool of stDNA
sequences at low copy number (‘a satellite DNA library’).
During the speciation process, an occasional amplifica-
tion of a stDNA repeat from that low copy number pool
would drive this repeat to form major stDNAs in one
species, while the other repeats from the pool would
remain in all the others species at low copy number.
In the present paper, we characterise the new major
stDNA families in nine Pimelia species, comparing them
to the PIM357 stDNA family previously described (Pons
et al, 2002a). The goal of this paper is to understand the
mechanisms involved in the formation and persistence of
these stDNA sequences. The phylogenetic analysis of the
stDNA sequences very likely does not reflect the
diversification of Pimelia species, due to their marked
sequence divergences, and the possibility of random
amplification from a pool of stDNA sequences. There-
fore, an independent phylogeny, based on mitochondrial
and nuclear markers, was obtained to give a broader
picture of the evolution of stDNA sequences and
distribution of particular stDNA families among spe-
cies-groups.
Material and methods
Sampling and genomic DNA extraction
The individuals were collected at localities in the Iberian
Peninsula and North Africa. Five species are classified
as part of the subgenus Pimelia s. str.: P. monticola (Pico
Veleta, Granada, Spain), P. cordata (Kraatg, Morocco),
P. echidna (Kelaa des Sraghna, Morocco), and two were
unclassified species from Nefta, Tunisia (Pimelia sp 1 and
Pimelia sp 2). Another four species are of the subgenus
Amblyptera: P. fornicata (Zahara de los Atunes, Ca ´diz,
Spain), P. scabrosa (Zahara de los Atunes, Ca ´diz, Spain),
P. rotundipennis (Kelaa des Sraghna, Morocco), and
P. rugosa (Tizi-N-Tichka, Morocco). DNA was isolated
from adults by standard phenol extraction and ethanol
precipitation procedures.
Isolation, cloning, and sequencing of satellite DNA
Digestions of genomic DNA with restriction enzymes
were performed according to the instructions of the
manufacturer (Roche), and the fragments separated
by electrophoresis on 1.5% agarose gels. The DNA
bands corresponding to putative monomeric sequences
were cut from the agarose gel, purified with the
Gene Clean Kit (Bio 101 Inc.), ligated into the Sma I site
of plasmid pUC18 vector (Amersham) and used to
transform Escherichia coli DH5a. Clones were screened
using the b-galactosidase blue-white colour system.
Positive clones were sequenced on both strands by
the dideoxy sequencing method using the Dig Taq
DNA Sequencing Kit for Cycle Sequencing (Boehringer
Mannheim) and the semiautomatic sequencing system
GATC1500-SystemDirect
(Boehringer Mannheim). The sequenced stDNA repeat
units have been deposited in EMBL databank under
Accession Numbers AJ247374-AJ247404, AJ307969, and
AJ307973.
BlottingElectrophoresis
Southern blot and estimation of stDNA percentage
For Southern analysis, 5mg of genomic DNA from each
species was digested with different restriction enzymes
and blotted on nylon membranes. P. monticola mono-
mers, from genomic DNA or clones, were purified from
the gel, labelled with digoxigenin, and used as probes in
a Southern hybridisation under high (85–90% similarity)
and low (60% similarity) stringency conditions. Digoxi-
genin labelling of the probe, filter hybridisations, and
detection of the hybridisation signals were performed as
described in the manual for the DIG High Prime DNA
Labelling and Detection Starter Kit I (Roche). The relative
amount of stDNA was determined from digestion of
genomic DNA (Hae III or Rsa I) electrophoresed on an
agarose gel. The digitisation and densitometric measure-
ments from the gel photographs were performed with
the help of the Sun View program (Pharmacia).
Sequence analysis of monomers
Multiple alignment was performed using Clustal W v. 1.7
(Higgins et al, 1996). Details of the nucleotide composi-
tion, sequence divergences (measured as the proportion
of nucleotide sites at which two sequences are different
or p distance), and parsimony tree searches (1000
random replicates) were performed using PAUP* 4.05
(Swofford, 2002). Bremer support values were deter-
mined performing 200 random replicate searches with
constraint files obtained with TreeRot.v2b (Sorenson,
1999). The nucleotide diversity was calculated using
DnaSP v. 3.51 (Rozas and Rozas, 1997). Curvature
analysis of monomers, searches for A or T X3 runs
and mobility on nondenatured polyacrylamide gel
electrophoresis, were performed as described in Barcelo ´
et al (1997).
PCR amplification and sequencing of mitochondrial and
nuclear fragments
Mitochondrial fragments were obtained using primers
and methods described elsewhere: a 376-bp fragment of
cytochrome oxidase subunit I (COI) (Juan et al, 1995), a
358-bp segment of cytochrome b (Vogler and Welsh,
1997), and a fragment of about 510bp from the large
ribosomal (16S) subunit gene (Funk, 1999). A 328-bp
Complex Pimelia satellite DNAs
J Pons et al
419
Heredity

Page 3

fragment of the nuclear Histone 3 gene was obtained as
described elsewhere (Colgan et al, 1998). Finally, a partial
region of about 650bp of the rRNA 28S containing the
variable domains D3–D6 was amplified under standard
PCR conditions using the following primer combination:
forward 50-GGG ACC CGT CTT GAA ACA C-30and
reverse 50-TTA CAC ACT CCT TAG CGG AT-30.
Sequencing of both strands in each case was performed
with the dideoxy sequencing method using the Big Dyet
Terminator Cycle Sequencing Kit and an ABI PRISMt
3700 DNA Analyzer (Applied Biosystems). Sequences
have been deposited in EMBL databank under Accession
Numbers: COI (X97209–X97223, AJ248198–AJ248217,
and Z71727), 28S (AJ565939–AJ565974), cytochrome b
(AJ565975–AJ566009), Histone 3 (AJ566010–AJ566044),
and 16S (AJ566045–AJ566080). The corresponding se-
quences from Tenebrio molitor (X88966, U65184, and
AJ438153) and Tribolium castaneum (NC003081) were
obtained from Gene Bank.
Phylogenetic analysis of mitochondrial and nuclear
sequences
Our preferred approach for tree searches was the
combined analysis with all characters equally weighted
and gaps treated as fifth character under the parsimony
criterion. Parsimony tree searches (1000 random repli-
cates), and incongruence length difference test (ILD test,
Farris et al, 1994; with 200 random replicates) to estimate
the incongruence between the different markers were
carried out in PAUP*v. 4.05 (Swofford, 2002). Bremer
support and partitioned Bremer support (PBS) values
were determined performing 200 random replicate
searches with constraint files obtained with TreeRot.v2b
(Sorenson, 1999). ML branch lengths based on the MP
tree and model parameters calculated using MODELT-
EST (Posada and Crandall, 1998) were modified using
the nonparametric rate smoothing (NPRS) algorithm as
implemented in TreeEdit v. 1.0a9 (Rambaut and Charles-
ton, 2002).
Results
Cloning and sequence analysis of satellite DNAs
Genomic DNAs of P. fornicata, P. scabrosa, P. rotundipennis,
P. rugosa, P. monticola, P. cordata, P. echidna, and the two
unclassified species P. sp 1 and P. sp 2 were digested with
several restriction enzymes. Hae III and, in the case of
P. cordata, Rsa I were selected for further studies. Seven
out of the nine species checked showed conspicuous
bands of about 500bp. Moreover, P. scabrosa and
P. rotundipennis had longer prominent bands of about
700bp (Figure 1a). Some of the species, such as P. rugosa,
P. scabrosa, P. rotundipennis, and P. sp 2, also revealed
DNA bands twice as large as the monomers interpreted
as uncut contiguous units. Densitometric quantification
of the corresponding electrophoretic bands in the DNA
digestions revealed that these stDNAs represent a high
proportion of the genome in the species studied, ranging
from 26.1 to 43.6% of the total DNA (Table 1). Given the
DNA content of the species screened (Pons, 1999), the
stDNA sequences obtained correspond to from 2.1 to
7.5?105copies per haploid genome, depending on the
species (Table 1).
Southern blots of genomic DNA digested with restric-
tion enzymes were hybridised using monomers from the
genomic DNA of P. monticola as a probe. The repeat units
of P. rugosa, P. scabrosa, P. cordata, P. echidna, Pimelia sp. 1,
and Pimelia sp. 2, and several species whose major
stDNA is PIM357 displayed cross-hybridisation even at
high stringency conditions (85–90% similarity, Figure 1).
P. fornicata and P. rotundipennis also revealed cross-
hybridisation (not shown) indicating that all the Pimelia
stDNA sequences studied to date are related in sequence.
We cloned the monomeric bands of each taxon and
randomly sequenced between 3 and 8 units (see Table 1).
The Southern mentioned above shows identical results
when a cloned monomer of P. monticola was used as a
probe. The sequenced repeats showed striking intraspe-
cific similarities in length, nucleotide composition, and
nucleotide diversity (Table 1).
At the interspecific level, the repeat units displayed
differences in nucleotide sequences (Table 2 and Figure 2)
except P. rotundipennis and P. scabrosa, which share
almost identical repeat units (PROT and PSCA respec-
tively). An alignment of the consensus monomeric
sequences of the nine species described in this work
with the PIM357 consensus sequence characterised
previously, reveals the presence of four sequence
segments conserved in the monomers of all species
(segments I–IV, Figure 2). Interestingly, these conserved
regions comprise almost the complete sequence of
monomer PIM357. Moreover, all the repeat units, with
the exception of PIM357, share a common central region
of about 140bp (Figure 2). The unrooted tree based on
parsimony retrieved three different groups of sequences
Figure 1 (a) Agarose gel electrophoresis of genomic DNA restricted with Hae III, except P. lutaria, P. baetica, P. atlantis frigioides, P. sparsa
serrimargo, P. laevigata validipes, and Tenebrio molitor, which were cut with Eco RI: Pimelia lutaria (1), P. baetica (2), P. rugosa (3), P. scabrosa (4),
P. cordata (5), P. echidna (6), Pimelia sp. 2 (7) and Pimelia sp. 1 (8), P. atlantis frigioides (10), P. sparsa serrimargo (11), P. laevigata validipes (12),
Tenebrio molitor (13), and Timarcha balearica (14) as negative control. Lane 9 shows a DNA molecular standard (Marker VI Roche) with bands
ranging from 2100 to 150bp. (b) Southern blot of the gel after hybridisation using monomers from genomic DNA of P. monticola as a probe.
Complex Pimelia satellite DNAs
J Pons et al
420
Heredity

Page 4

(Figure 3). One group includes the sequences of 705bp of
P. scabrosa and P. rotundipennis, plus the sequence of
533bp of P. monticola. The second contains PIM357
monomers, with the sequences from Canary Islands
appearing as basal. The last group joins the repeats of
about 500bp where the sequences of P. echidna are basal
relative to two sister groups: (1) P. rugosa and P. fornicata,
and (2) P. cordata and the two unclassified Pimelia species
from Tunisia. A parsimony analysis based on the most
conserved segments (I and II), revealed identical se-
quence groupings to those obtained from the complete
sequences.
The consensus sequences of P. rotundipennis and
P. scabrosa (both 705-bp long) differ by two nucleotide
substitutions only (positions 124 and 467; Figure 2) but
neither of these is completely fixed within the species.
The shorter consensus sequence of P. monticola (PMON),
of 533bp, shows sequence identity with the 532-bp-long
segment of the two former species given that there is an
indel at position 441 (Figure 2). The related segments
shared by PMON and PSCA–PROT show about 80%
similarity (see Table 2) with point mutations distributed
throughout the monomers. The remaining 173bp of the
PSCA and PROT sequences are made of partial duplica-
tions of the common 533-bp segment: a 32-bp-long direct
duplication (90.6% similarity, Figure 2), and a 141-bp-
long duplicated segment in an inverted orientation
(78.7% similarity, Figure 2).
The species P. rugosa, P. fornicata and P. echidna show
related sequences of 502bp. The sequences of P. echidna
are more dissimilar to the remaining repeats (differing by
around 20%, see Table 2) due to many single diagnostic
(fixed) substitutions distributed throughout the mono-
meric sequence. However, the sequences of P. fornicata
and P. rugosa are very similar (the interspecific diver-
gence is very similar to the intraspecific divergence, see
Table 2). In fact, monomers from these species differ by
10 nucleotide substitutions, but only four of these are
fixed at the species level (positions 214, 227, 248, and 417;
Figure 2). Finally, the repeat units of the two unclassified
species from Tunisia (P1NF and P2NF) and those from
P. cordata, are different in sequence from the ones
described above and also from that of PIM357 family.
The monomers P1NF and P2NF differ by two gaps of 10
and 12bp, showing 75.4% sequence similarity (Figure 2).
P. cordata monomers are the most divergent relative to
the repeats of any other species (Figure 2 and Table 2).
Analysis of sequence-induced curvature
All the species have AþT-rich stDNA sequences (Table 1)
and high frequencies of periodically distributed runs,
composed of three or more adenines or thymines,
throughout the monomers. All the isolated repeat units
migrated slower than expected, based on their length, on
nondenaturing polyacrylamide gels. The repeat units
Table 1 List of specimens, species-clones code, restriction enzyme used to isolate the repeats, number of sequenced repeats, percentages of
satellite DNA, copy number of the repeat units (?103), intraspecific consensus repeat unit length (bp), percentages of A+T, and the nucleotide
diversity (p) of the repeat units
SpeciesCodeRestriction enzymeRepeats sequenced
% stDNA
Copy numberLength% A+T
p
P. echidna
P. rugosa
P. sp 2
P. sp 1
P. cordata
P. rotundipennis
P. scabrosa
P. monticola
P. fornicata
26 taxa
PECH
PRUG
P2NF
P1NF
PCOR
PROT
PSCA
PMON
PFOR
PIM357
Hae III
Hae III
Hae III
Hae III
Rsa I
Hae III
Hae III
Hae III
Hae III
Eco RI Hae III
3
3
3
3
2
3
8
5
3
33.372.9
32.873.1
35.970.6
31.272.1
34.473.7
26.172.5
30.773.0
28.173.1
43.772.9
27.1?43.6
380
360
460
250
—
—
210
170
750
450
502
502
509
512
515
705
705
533
502
357
69.32
64.44
57.96
60.16
57.09
68.75
68.86
71.18
64.28
69.00
0.016
0.027
0.013
0.025
0.017
0.021
0.023
0.039
0.023
0.02–0.2156
Average, or maximum and minimum values, for 26 previously described taxa which repeat units are about 357bp long (Pons et al, 2002a) is
also included.
Table 2 Matrix of pairwise divergence values of stDNA consensus sequences based on alignment of Figure 2 (above diagonal), and pairwise
divergences based on the two first more conserved segments only (I and II)
PSCAPROTPMON357AF357IB357CI PRUGPFORPECHPCOR6P1NFP2NF
PRUG
PSCA
PROT
PMON
357AF
357IB
357CI
PFOR
PECH
PCOR6
P1NF
P2NF
0.3180.318
0.003
0.270
0.221
0.219
0.327
0.353
0.350
0.300
0.285
0.320
0.317
0.283
0.154
0.295
0.290
0.287
0.238
0.209
0.202
0.020
0.382
0.382
0.371
0.382
0.367
0.348
0.173
0.362
0.361
0.349
0.344
0.333
0.320
0.181
0.361
0.483
0.485
0.483
0.514
0.510
0.502
0.363
0.363
0.354
0.479
0.481
0.474
0.473
0.460
0.481
0.360
0.376
0.440
0.308
0.467
0.468
0.477
0.472
0.460
0.448
0.323
0.329
0.415
0.246
0.382
0.382
0.366
0.378
0.366
0.348
0.016
0.190
0.376
0.328
0.324
0.003
0.179
0.278
0.236
0.218
0.312
0.286
0.421
0.400
0.396
0.173
0.273
0.231
0.213
0.312
0.279
0.426
0.406
0.401
0.220
0.210
0.149
0.275
0.254
0.416
0.400
0.390
0.151
0.208
0.327
0.284
0.461
0.420
0.429
0.188
0.279
0.274
0.428
0.378
0.382
0.289
0.271
0.421
0.421
0.403
0.201
0.366
0.328
0.335
0.344
0.339
0.351
0.410
0.4060.151
Species codes are as indicated in Table 1.
Complex Pimelia satellite DNAs
J Pons et al
421
Heredity

Page 5

Figure 2 Nucleotide sequence alignment of the consensus monomeric sequences from: Pimelia scabrosa (PSCA), P. rotundipennis (PROT),
P. monticola (PMON), stDNA family from 26 Pimelia taxa characterised elsewhere (Pons et al, 2002a, PIM357), P. rugosa (PRUG), P. fornicata
(PFOR), P. echidna (PECH), clone pCOR6 from P. cordata, and Pimelia sp. 1 (P1NF) and Pimelia sp. 2 (P2NF) from Tunisia. Dots denote
nucleotides that are the same as the first sequence; dashes denote gaps. Boxes denote the four fragments conserved in all nine consensus
monomeric sequences (I–IV). The central segment (shadowed) is not present in the consensus PIM357 unit. Repeat units of P. scabrosa and
P. rotundipennis exhibit a 32-bp-long direct repeat (indicated in bold and italics), and inverted one of 141bp in length (indicated in lower case).
Complex Pimelia satellite DNAs
J Pons et al
422
Heredity

Page 6

showed similar RLvalues ranging from 1.71–2.17 (RLis a
ratio of apparent to actual length in an acrylamide gel at
41C, and it is an estimation of the retarded mobility).
Pimelia phylogeny based on mitochondrial and nuclear
markers
Alignment of the five genes (COI, cytochrome b, 16S, 28S
and Histone 3) was straightforward with few gap sites
(seven in 28S and 16 in 16S), mostly nonambiguous. The
combined analysis of the five markers in tree searches
resulted in a single tree of 2633 steps (Figure 4). The two
subgenera, Pimelia s. str. and Amblyptera, appear as
monophyletic sister groups. Within Pimelia s. str., there is
one clade composed of African species whose major
stDNAs belong to the PIM357 family with P. monticola
basal to them. The other one is composed of two clades:
(1) Iberian-Balearic species whose major stDNAs belong
to the PIM357 family; and (2) endemic Canarian species
(with the major stDNAs being the PIM357 family) with
African species whose stDNAs are about 500-bp long
basal to them (P. echidna, P. cordata, and the two
unclassified species from Tunisia). Within Amblyptera,
two clades are resolved: one composed of species with
the major satellite of 705bp (P. rotundipennis and
P. scabrosa), and the other with species P. fornicata and
P. rugosa, sharing the highly similar major satellite of
502bp.
The five partitions were studied separately to test for
the phylogenetic signal and potential conflict. Character
incongruence between the different markers as deter-
mined by the ILD test was nonsignificant for all possible
combinations of the five genes (data not shown), except
when 16S gene was included (Po0.01). Nevertheless, the
incongruence between the three mitochondrial partitions
was reduced when they were analysed together with the
nuclear data (ILD/change¼0.0362 vs 0.0072), demon-
strating that combining all data in a simultaneous
analysis produced a more consistent character distribu-
tion than any of the partitions separately. Most of the
signal and support was provided by COI, 16S, and
Histone 3 (PBS 177, 87, and 61, respectively). Despite its
signal, cytochrome b supported only 40% of the nodes
with no global PBS. Finally, 28S showed no global
positive PBS probably due to its low number of
informative sites (14), supporting four nodes only (the
Amblyptera clade, the African PIM357 clade, the P. radula
radula and P. radula granulata clade, and the outgroups
T. molitor and T. castaneum).
ML branch lengths based on the parameters selected
by MODELTEST (GTRþIþG) and the MP tree were
used to estimate node ages. Likelihood ratio tests
between rate-constant and rate variable models, for the
combined data and for each one of the partitions,
revealed deviation from the molecular clock (data not
shown). Therefore, ML branch lengths of the combined
data were fitted using the NPRS algorithm. Absolute
node ages were estimated by constraining the split of
P. laevigata laevigata and P. laevigata costipennis to 1Mya
since the island El Hierro, from where the latter species is
endemic, is not older than 1Mya (Juan et al, 1995). This
age represents a maximum estimate since these beetles
not necessarily colonised El Hierro immediately after its
emergence. An older precolonisation age also seems to
be unlikely because there is no evidence of an earlier split
of the lineage in La Palma Island (with an age of about
2Mya; Juan et al, 1995) from where P. laevigata laevigata is
endemic. Using that split of 1Mya as a calibration point,
we can conclude that the maximum estimate for the split
between the two subgenera Pimelia s. str. and Amblyptera
was about 11Mya. On this basis, Pimelia s. str. lineages
originated about 9–6Mya: the African lineage harbour-
ing the major stDNA PIM357 about 9Mya, the Iberian
and Canarian lineages at about 7Mya, and the species
with major stDNA sequences of 500bp about 8–6Mya.
However, estimation based on mitochondrial ML branch
lengths, fitted on NPRS, and using the considered
standard insect mtDNA clock of around 2% per My
(Brower, 1994) nearly doubled node ages.
Discussion
The phylogeny based on mitochondrial (COI, cyto-
chrome b, and 16S) and nuclear markers (28S and
Figure 3 Unrooted phylogram representing the single shortest tree
obtained from the parsimony analysis of the Pimelia monomeric
sequences (1264 steps). Branch lengths represent the average
number of parsimony changes (gaps treated as 5th character).
Numbers at each node indicate Bremer support value out of 200
replicates (above node) and the percentage of trees representing the
particular node out of 1000 bootstrap replicates (below node).
Bootstrap values under 50 are not indicated.
Complex Pimelia satellite DNAs
J Pons et al
423
Heredity

Page 7

Histone 3) reveals a different scenario of diversification
of Pimelia species than that based on the major stDNA
sequences. The presence of units similar in length
(502bp) and sequence, in species from two different
subgenera P. rugosa and P. fornicata (Amblyptera), and
P. echidna (Pimelia s. str.) could be explained by an
independent amplification from the same pool of satellite
sequences, that is, the ‘library hypothesis’ (Fry and
Salser, 1977; Ugarkovic and Plohl, 2002). The same
applies for the case of P. monticola (Pimelia s. str.) and
the P. scabrosa–P. rotundipennis pair (Amblyptera) whose
major stDNA sequences, though different in length (533
and 705bp, respectively), are clearly related. The
existence of monophyletic groups having PIM357 as
the major stDNA in three independent nodes of the
phylogeny (Moroccan, Iberian-Balearic, and Canarian)
could also be explained by independent amplification.
Nevertheless, the phylogeny based on mitochondrial and
nuclear markers could, depending on character transfor-
mations, support any of three different hypotheses, one,
two or three independent amplifications from the
‘library’.
The existence of a ‘library of stDNA sequences’ at low
copy number has been previously demonstrated experi-
mentally in six Pimelia species (Bruvo et al, 2003), and
also in other tenebrionid species of the genus Palorus
(Mestrovic et al, 1998). The Pimelia phylogeny based on
mitochondrial and nuclear markers and the results of
previous work (Bruvo et al, 2003) suggest a common
ancestor whose genome harboured all or most of the
Figure 4 Parsimony tree representing the single shortest tree (2633 steps) obtained from simultaneous analysis of COI, cytochrome b, 16S,
28S, and Histone 3 (2236 characters). Numbers at each node indicate Bremer support values out of 200 replicates (above node) and the
percentage of trees representing the particular node out of 1000 bootstrap replicates (below node). Bootstrap values under 50 are not
indicated. Relative node ages were estimated calculating the ML branch lengths (GTRþGþI) on the MP tree and then made ultrametric by
NPRS. The scale bar shows the calibration of absolute ages (My). Absolute node ages were estimated by constraining the split of P. laevigata
laevigata and P. laevigata costipennis to 1Mya since the island El Hierro, from where is endemic the last subspecies, is not older than 1Mya
(Juan et al, 1995). The biogeographic regions and major satellites are indicated for each species.
Complex Pimelia satellite DNAs
J Pons et al
424
Heredity

Page 8

major stDNA families present in the living species at low
copy numbers. Speciation and concerted evolution of the
repeats may have changed dramatically their copy
number by differentially amplifying from the ‘library’
one low copy satellite variant to compose the major
satellite. However, in contrast to the data obtained in
Palorus species, whose major satellite sequences showed
no similarity, the monomers of the major stDNA families
of Pimelia still conserve some regions of sequence
similarity despite their marked differences, which
suggests a common evolutionary origin.
The persistence of intrinsic bending in Pimelia mono-
mers indicate its potential role in heterochromatin
condensation (Brutlag, 1980; Martı ´nez-Balba ´s et al, 1990;
Barcelo ´ et al, 1998). Nevertheless, it has been shown that
there is a tremendous variation in the magnitude of
intrinsic bending among satellite DNAs (Fitzgerald et al,
1994), and it is this bending, rather than DNA sequence
per se, which could be the motif being recognised by
specific binding proteins (Lobov et al, 2001). The
importance of bending could explain the partial con-
servation of the primary sequence in Pimelia monomers,
since curvature is linked to a periodical distribution of
runs composed of three or more adenines, but not
conservation of length (which is 357, 500, or 705bp). A
similar trend has been suggested for other tenebrionid
stDNAs (Ugarkovic et al, 1995; Barcelo ´ et al, 1998).
A schematic representation of a plausible evolutionary
scenario for the origin of satellite repeats of Pimelia in the
ancestor, that is, the acquisition of the ‘library’, is shown
in Figure 5. Since repetitive sequences from outgroups
have no similarity with Pimelia satellites, and the
independent phylogeny could support either scenario,
we cannot decide whether Pimelia repeats originated
from 357 or 500bp sequences. However, most of the cases
described to date demonstrate that long and complex
repeats have been created from shorter ones (Bigot et al,
1990; Rojas-Rousse et al, 1993; Ugarkovic et al, 1996).
Therefore, we suggest that the insertion of a 141-bp DNA
fragment into the unit of about 357bp shared for all
species created the ancestral units of 500bp. The
sequences of 500-bp split in two lineages: one has
preferentially been amplified in P. monticola (533bp)
and the other is present in several species. The repeats of
P. scabrosa and P. rotundipennis (705bp) have likely
originated by a 32-bp direct duplication and the insertion
of a 141-bp duplicated segment in inverted orientation
with respect to the basic structure of the 533-bp unit
currently present in P. monticola (Figure 5). Other
insertion/deletion events seem to have occurred in a
Figure 5 A schematic portrayal of a plausible evolutionary scenario for the satellite repeats of Pimelia.
Complex Pimelia satellite DNAs
J Pons et al
425
Heredity

Page 9

species-specific way, as they do not show similarity with
any other region of a repeat unit, although mutational
processes might have obscured the similarity between
the original and duplicated sequences. Rearrangement of
the PIM357 monomers to create 500-bp units has been
previously described in some Canarian Pimelia (Pons et al,
2002a), but these are different from the sequences
described in this paper (eg they do not include the 141-
bp motif shared by 500 and 700 repeats). Similar
rearrangements have been suggested as mechanisms
involved in the evolution of two related stDNA families
in the coleopteran Tribolium madens, which are composed
of different subunits of about 100bp with 60–80%
homology (Ugarkovic et al, 1996). Recombination is
thought to be one of the main forces driving concerted
evolution (Dover, 2002), but may be also a force
generating new monomers by rearrangement of the
repeats.
The two sister clades of the subgenus Amblyptera,
which split about 10–20Mya, amplified different repeats:
in P. rugosa and P. fornicata a sequence of 502bp, and in
P. scabrosa and P. rotundipennis, a 705-bp unit. Interest-
ingly, the sequences show very little variation within the
clades despite having diversified about 6–12Mya. On the
other hand, the three paraphyletic clades showing
PIM357 sequences that split between 10–20Mya, Mor-
occan, Iberian-Balearic, and Canarian, have evolved
gradually, homogenising and fixing mutations within
each clade during the last 6–12Mya. Our results show
that satellite sequences are evolving at different rates in
two manners: both gradually and saltatory. Moreover,
the partial conservation of the periodical distribution of
runs composed of three or more adenines in Pimelia
satellites could be related to the role of stDNA in tight
heterochromatin condensation.
In summary, the results presented in this paper
suggest that most of the major stDNAs of Pimelia present
in the extant species were already present in a common
ancestor composing a ‘library of satellite sequences’ of
common origin characterised by extensive sequence
rearrangement. During diversification, different species
randomly amplified different examples of these low copy
sequences to form the abundant major satellites. This
turnover occurred at different rates: in relatively short
periods of time (abrupt-saltatory replacement), in a
gradual manner (consistent with a molecular drive
model, Dover, 2002), or simply with no apparent change
for long evolutionary time.
Acknowledgements
We are very grateful to Drs P Oromı ´, J Gomez-Zurita,
M Palmer, and C Garin who provided many of the
Pimelia samples. Drs Oromı ´ and Go ´mez-Zurita also
helped in the taxonomic determination. This work was
supported by the Spanish Research Funds REN2000–
0282/GLO, REN2003–00024/GLO to CJ and BOS2000–
0822 to EP.
References
Barcelo ´ F, Pons J, Petitpierre E, Barjau I, Portugal J (1997).
Polymorphic curvature of satellite DNA in three subspecies
of the beetle Pimelia sparsa. Eur J Biochem 244: 318–324.
Barcelo ´ F, Gutie ´rrez F, Barjau I, Portugal J (1998). A theoretical
perusal of the satellite DNA curvature in tenebrionid beetles.
J Biomol Struct Dyn 16: 41–50.
Bigot Y, Hamelin MH, Periquet G (1990). Heterochromatin
condensationandevolution
families in two parasitic wasp species: Diadromus pulchellus
and Eupelmus vuilleti (Hymenoptera). Mol Biol Evol 7:
351–364.
Brower AVZ (1994). Rapid morphological radiation and
convergence among races of the butterfly Heliconius erato
inferred from patterns of mitochondrial-DNA evolution. Proc
Natl Acad Sci USA 91: 6491–6495.
Brutlag DL (1980). Molecular arrangement and evolution of
heterochromatin DNA. Ann Rev Genet 14: 121–144.
Bruvo B, Pons J, Ugarkovic D, Juan C, Petitpierre E, Plohl M
(2003). Evolution of low-copy number and major satellite
DNA sequences coexisting in two Pimelia species-groups
(Coleptera). Gene 312: 85–94.
Colgan DJ, Mclauchlan A, Wilson GDF, Livingston SP, Edge-
combe GD, Macaranas J et al (1998). Histone H3 and U2
snRNA DNA sequences and arthropod molecular evolution.
Aust J Zool 46: 419–437.
Dover G (2002). Molecular drive. Trends Genet 18: 587–589.
Farris JS, Ka ¨llersjo ¨ M, Kluge A G, Bult C (1994). Testing
significance of incongruence. Cladistics 10: 315–319.
Fitzgerald G, Dryden EC, Bronson EC, Williams JS, Anderson,
JN (1994). Conserved pattern of bending in satellite and
nucleosome positioning DNA. J Biol Chem 269: 2312–2314.
Fry K, Salser W (1977). Nucleotide sequence of HS-a satellite
DNA from kangaroo rat Dipodomys ordii and characterisation
of similar sequences in other rodents. Cell 12: 1069–1084.
Funk DJ (1999). Molecular systematics of cytochrome oxidase I
and 16S from Neochlamisus leaf beetles and the importance of
sampling. Mol Biol Evol 16: 67–82.
Higgins DG, Thompson JD, Gibson TJ (1996). Using clustal
for multiple sequence alignments. Methods Enzymol 266:
383–402.
Juan C, Oromı ´ P, Hewitt GM (1995). Mitochondrial DNA
phylogeny and sequential colonization of Canary Islands by
darkling beetles of the genus Pimelia (Tenebrionidae). Proc R
Soc B 261: 173–180.
Kwieton E (1977). Esquisse phylogenetique du genre Pimelia F.
Acta Entomol Mus Natl Pragae 39: 559–589.
Lobov IB, Tsutsui K, Mitchell AR, Podgornaya OI (2001).
Specificity of SAF-A and lamin B binding in vitro correlates
with the satellite DNA bending state. J Cell Biochem 83:
218–229.
Martı ´nez-Balba ´s A, Rodriguez-Campos A, Garcia-Ramirez M,
Sainz J, Carrera P, Aymani J et al (1990). Satellite DNAs
contain sequences that induce curvature. Biochemistry 29:
2342–2348.
Mestrovic N, Plohl M, Mravinac B, Ugarkovic D (1998).
Evolution of satellite DNAs from the genus Palorus –
experimental evidence for the ‘library’ hypothesis. Mol Biol
Evol 15: 1062–1068.
Pons J (1999). Evolucio ´n del DNA sate ´lite en el ge ´nero Pimelia. PhD
Thesis,. University of Balearic Islands.
Pons J, Bruvo B, Juan C, Petitpierre E, Plohl M, Ugarkovic D
(1997). Conservation of satellite DNA in species of the genus
Pimelia (Tenebrionidae, Coleoptera). Gene 255: 183–190.
Pons J, Petitpierre E, Juan C (2002a). Evolutionary dynamics of
satellite DNA family PIM357 in species of the genus Pimelia
(Tenebrionidae, Coleoptera). Mol Biol Evol 19: 1329–1340.
Pons J, Petitpierre E, Juan C (2002b). Higher-order organisation
and compartmentalisation of satellite DNA PIM357 in
species of the coleopteran genus Pimelia. Chromosome Res
10: 597–606.
Posada D, Crandall KA (1998). MODELTEST: testing the model
of DNA substitution. Bioinformatics 14: 817–818.
Rambaut A, Charleston M (2002). TreeEdit. Phylogenetic Tree
Editor v1.0 alpha 10. Oxford University: Oxford.
ofuniquesatellite-DNA
Complex Pimelia satellite DNAs
J Pons et al
426
Heredity

Page 10

Rojas-Rousse D, Bigot Y, Periquet G (1993). DNA insertions as a
component of the evolution of unique satellite DNA families
in two genera of parasitoid wasps: Diadromus and Eupelmus
(Hymenoptera). Mol Biol Evol 10: 383–396.
Rozas J, Rozas R (1997). DnaSP version 2.0 a novel software
package for extensive molecular population genetic analysis.
Comput Appl Biosci 13: 307–311.
Sorenson MD (1999). TreeRot, Version 2. Boston University:
Boston, MA.
Swofford DL (2002). PAUP*. Phylogenetic Analysis using Parsi-
mony, Version 4.0b. Sinauer Associates: Sunderland, MA.
Ugarkovic D, Plohl M (2002). Variation in satellite DNA profiles
– causes and effects. EMBO J 21: 5955–5959.
Ugarkovic D, Petitpierre E, Juan C, Plohl M (1995). Satellite
DNAs in tenebrionid species: structure, organization and
evolution. Croat Chem Acta 68: 627–638.
Ugarkovic D, Durajlija S, Plohl M (1996). Evolution of Tribolium
madens (Insecta, Coleoptera) satellite DNA through DNA
inversion and insertion. J Mol Evol 42: 350–358.
Vogler AP, Welsh A (1997). Phylogeny of North American
Cicindela tiger beetles inferred from multiple mitochondrial
DNA sequences. Mol Phylogenet Evol 8: 225–235.
Complex Pimelia satellite DNAs
J Pons et al
427
Heredity