Phylogeny, species limits, and biogeography of the Brazilian lizards of the genus Eurolophosaurus (Squamata: Tropiduridae) as inferred from mitochondrial DNA sequences.

Abstract

Phylogenetic relationships and divergence times for 10 populations of the three recognized "species" of Brazilian lizards of genus Eurolophosaurus were estimated from 1229bp of cyt b, COI, 12S, and 16S rRNA mitochondrial gene segments. Eurolophosaurus is monophyletic and the basal split within the genus separates E. divaricatus from a clade comprising E. amathites and E. nanuzae. Three populations of E. divaricatus, which occurs along the western bank of Rio São Francisco, were consistently grouped together. On the east bank of the river, E. amathites and E. nanuzae from state of Bahia were recovered as the sister group of E. nanuzae populations from state of Minas Gerais. The paraphyly of E. nanuzae and the high divergence levels among populations of E. divaricatus strongly suggest that species limits in Eurolophosaurus should be revised. Even considering an extreme evolutionary rate of 2.8% sequence divergence per million years for the four gene segments analyzed together, E. divaricatus would have separated from the two other species by at least 5.5my ago, and E. amathites from E. nanuzae populations from Bahia and Minas Gerais, respectively, by 1.5 and 3.5my. The paleolacustrine hypothesis and changes in the course of the river potentially explain faunal divergence in the area, but divergences are much older than previously admitted.

Full text

Page 1

Phylogeny, species limits, and biogeography of the Brazilian lizards
of the genus Eurolophosaurus (Squamata: Tropiduridae)
as inferred from mitochondrial DNA sequences
Jose ´ Carlos Passonia, Maria Lu ´cia Benozzatia,*, Miguel Trefaut Rodriguesb
aDepartamento de Gene ´tica e Biologia Evolutiva, Instituto de Biocie ˆncias, Universidade de Sa ˜o Paulo, Rua do Mata ˜o,
277, Cidade Universita ´ria, 05508-900, Sa ˜o Paulo, SP, Brazil
bDepartamento de Zoologia, Instituto de Biocie ˆncias, Universidade de Sa ˜o Paulo, Brazil
Received 21 July 2006; revised 14 October 2007; accepted 23 October 2007
Available online 12 November 2007
Abstract
Phylogenetic relationships and divergence times for 10 populations of the three recognized ‘‘species” of Brazilian lizards of genus
Eurolophosaurus were estimated from 1229 bp of cyt b, COI, 12S, and 16S rRNA mitochondrial gene segments. Eurolophosaurus is
monophyletic and the basal split within the genus separates E. divaricatus from a clade comprising E. amathites and E. nanuzae. Three
populations of E. divaricatus, which occurs along the western bank of Rio Sa ˜o Francisco, were consistently grouped together. On the east
bank of the river, E. amathites and E. nanuzae from state of Bahia were recovered as the sister group of E. nanuzae populations from state
of Minas Gerais. The paraphyly of E. nanuzae and the high divergence levels among populations of E. divaricatus strongly suggest that
species limits in Eurolophosaurus should be revised. Even considering an extreme evolutionary rate of 2.8% sequence divergence per mil-
lion years for the four gene segments analyzed together, E. divaricatus would have separated from the two other species by at least 5.5 my
ago, and E. amathites from E. nanuzae populations from Bahia and Minas Gerais, respectively, by 1.5 and 3.5 my. The paleolacustrine
hypothesis and changes in the course of the river potentially explain faunal divergence in the area, but divergences are much older than
previously admitted.
? 2007 Elsevier Inc. All rights reserved.
Keywords: Cyt b; COI; 12S; 16S; mtDNA phylogeny; Species limits; Biogeography; Divergence times; Eurolophosaurus; Tropiduridae; Rio Sa ˜o Francisco
sand dunes; Brazil
1. Introduction
The tropidurid lizard genus Eurolophosaurus (Frost
et al., 2001) is a monophyletic group of three species
formerly called the Tropidurus nanuzae group (Rodrigues,
1986). It occurs along the Serra do Espinhac ?o, a large
mountain ridge extending along eastern Brazil in the states
of Minas Gerais and Bahia, and in sandy areas close to the
banks of the middle Rio Sa ˜o Francisco in Bahia (Fig. 1).
E. nanuzae is predominantly saxicolous and occurs in
several localities throughout the Serra do Espinhac ?o, from
Serra do Cipo ´ in the southern part of Minas Gerais to Rio
de Contas and Caetite ´, in Bahia, always at altitudes near or
above 900 m. The other two species are psammophilous
and occupy opposite banks of Rio Sa ˜o Francisco below
600 m, close to the northern portion of Serra do Espinhac ?o
in the state of Bahia. E. amathites is found at Santo Ina ´cio,
Gameleira do Assurua ´, and Lagoa de Itaparica, on the
sands of the right bank of the river, and E. divaricatus in
Queimadas, Ibiraba, Mocambo do Vento, Manga, and
Alagoado, on the opposite margin, in two dune fields sep-
arated by an area of rocky soils approximately 150 km
long. Additional data on distribution, physiography, and
ecology of the group have been reported by Rodrigues
(1986, 1996).
1055-7903/$ - see front matter ? 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.ympev.2007.10.022
*Corresponding author. Fax: +55 11 3091 7553.
E-mail address: benozzat@usp.br (M.L. Benozzati).
www.elsevier.com/locate/ympev
Available online at www.sciencedirect.com
Molecular Phylogenetics and Evolution 46 (2008) 403–414

Page 2

According to morphological and distributional data,
Rodrigues (1986) suggested that E. nanuzae (at the time
in genus Tropidurus) was the sister taxon to a clade com-
prising E. amathites and E. divaricatus. Based on morphol-
ogy, Frost (1992) placed E. nanuzae outside a clade
comprising the rest of the Tropidurus radiation (E. amath-
ites and E. divaricatus were not sampled in this study). This
was the reason why, in a later paper on tropidurid phylog-
eny using both morphological and molecular characters,
the genus Eurolophosaurus was created (Frost et al., 2001).
The chromosomal studies of Kasahara et al. (1987)
detected a nucleolar organizer region (NOR) on the long
arm of macrochromosome pair 6 in E. amathites and
E. nanuzae, but on the short arm of the same chromosome
pair in E. divaricatus. As NORs are absent from this chro-
mosome pair in phylogenetically related tropidurids, this
character was used to support the monophyly of the group
and a pericentric inversion was proposed to explain the
condition observed in E. divaricatus (Kasahara et al.,
1983, 1985, 1986a,b). Frost et al. (2001) considered loca-
Fig. 1. Above, South America inter-tropical region. Below, sand dune region of the middle Rio Sa ˜o Francisco, Brazil. Localities: (1) Caetite ´; (2) Rio de
Contas; (3) Pedra Menina; (4) Extrac ?a ˜o; (5) Serra do Cipo ´; (6) Alagoado; (7) Queimadas; (8) Vacaria; (9) Ibiraba; (10) Xique-Xique; (11) Ilha do Gado
Bravo; (12) Mocambo do Vento; (13) Lagoa de Itaparica; (14) Barra; (15) Santo Ina ´cio; (16) Gameleira do Assurua ´.
404
J.C. Passoni et al./Molecular Phylogenetics and Evolution 46 (2008) 403–414

Page 3

tion of NOR on chromosome 6 and the XXAA:XAY/A
sexual determination type (Kasahara et al., 1987) as syna-
pomorphies of the genus Eurolophosaurus. In allozymic
analyses (Martins, 1995), however, monophyly of the nan-
uzae group could not be demonstrated with confidence.
Finally, in the study of Frost et al. (2001) monophyly of
Eurolophosaurus was corroborated, and E. divaricatus was
shown as the sister taxon to a clade comprising E. amath-
ites and E. nanuzae.
Based on geomorphological and paleoclimatic data on
the Sa ˜o Francisco sand dunes (Ab’Saber, 1969; Tricart,
1974), and assuming the close relationship of E. amathites
and E. divaricatus, Rodrigues (1986) proposed a vicariant
model of speciation to explain their origin. He hypothe-
sized an ancestral population bisected by the Rio Sa ˜o Fran-
cisco after the last glacial period, about 12,000 years ago.
According to the model, in the Pleistocene the river flowed
into a paleolake and, at the end of the Wu ¨rm–Wisconsin
glaciation it would have flowed to the Atlantic Ocean.
Other psammophilous endemic pairs of closely related spe-
cies of gymnophthalmid lizards, snakes, and amphisbae-
nians, which occur in the area and show the same
vicariant pattern of geographic distribution, lend support
to the model (Rodrigues, 1991a–d, 1996, 2003). More
recently it was also argued that, although the paleolacus-
trine hypothesis could account for the speciation in the
area, forsaken meanders of the river may have acted as bar-
riers (Rodrigues, 1996). Palaeoclimatic and geomorpholog-
ical data on this region support the model (Souza-Lima
et al., 2005). Nevertheless, allozymic data (Martins, 1995)
suggested that the E. amathites/E. divaricatus divergence
should be much older, of the order of millions of years,
and not of 12,000 years as originally proposed.
To obtain phylogenetic relationships and divergence
times between and within species of Eurolophosaurus, based
on DNA sequencing data, we analyzed mtDNA sequences
for populations of these three species.
2. Materials and methods
2.1. Taxonomic sampling and DNA extraction
Ten populations of the three species of Eurolophosaurus
and three outgroup species were sampled and assayed for
regions from mitochondrial genes encoding 12S and 16S
ribosomal RNAs (rRNA), cytochrome b (cyt b), and sub-
unit I of cytochrome c oxidase (COI) proteins. A total of
25 specimens were used. Technical problems in DNA
amplification and sequencing precluded the inclusion of a
larger number of individuals (see Section 2.2). Table 1
shows the sample sizes that we had as a result. The geo-
graphic distribution of populations studied is shown in
Fig. 1. Samples of E. divaricatus from Queimadas and Ibir-
aba were considered the same population (Queimadas/Ibir-
aba) according to Passoni et al. (2000). One specimen of
Uranoscodon superciliosus, one of Microlophus quadrivitta-
tus and one of Tropidurus hispidus were used as outgroups,
according to Frost et al. (2001). Heart, liver, and portions
of skeletal muscle were stored at ?120 or ?196 ?C. DNA
enriched in the mitochondrial fraction was obtained by
the methods of Hillis and Davis (1986) and Dowling
et al. (1990), as modified by Passoni et al. (2000). Approx-
imately 2–4 mm3of these tissues from each specimen were
homogenized in 800 ll STES (0.01 M NaCl, 0.01 M Tris,
0.1 M EDTA, 0.25 M sucrose, pH 7.5) and then centri-
fuged for 5 min at 1200g at 4 ?C to pellet the nuclei. The
supernatant was centrifuged at 23,000g for 20 min at 4 ?C
for mitochondrial precipitation. The pellet was suspended
in 250 ll STE (0.1 M NaCl, 0.01 M Tris, 0.1 M EDTA,
pH 7.5), with 1% sodium dodecyl sulfate (SDS) and
10 U/ml proteinase K. After incubation for 2 h at 55 ?C,
the preparation was extracted with one volume of phenol,
and then with one volume of 24:1 chloroform:isoamyl alco-
hol solution. DNA was precipitated from the supernatant
by the addition of 1/10 volume buffer (3 M NaCl, 0.25 M
Tris, 0.1 M EDTA) and 2.5 volumes 100% ethanol at
?20 ?C overnight. The DNA pelleted after 15 min centrifu-
gation at 23,000g at 4 ?C was washed in 70% ethanol, air
dried, and dissolved in 20–80 ll TE (0.01 M Tris,
0.001 M EDTA, pH 8.0), and stored at ?20 ?C.
2.2. Amplification and sequencing
DNA amplification and sequencing trials presented
many technical problems hindering the analysis of the
desirable number of individuals for each of Eurolophosau-
rus populations. Therefore, we decided to limit the present
analysis to phylogenetic issues. Segments of 12S and 16S
rRNA genes, and segments of cyt b and COI protein-cod-
ing genes of the mitochondrial genome were amplified via
Table 1
Number of Eurolophosaurus populations and outgroup samples success-
fully sequenced for different gene regions
Species Population mtDNA
Cyt b
COI 12S16S
Eurolophosaurus
divaricatus
Alagoado
Queimadas/Ibiraba
Mocambo do
Vento
Santo Ina ´cio
Gameleira do
Assurua ´
Caetite ´
Rio de Contas
Pedra Menina
Serra do Cipo ´
Manaus
4
3
3
3
2
1
3
2
2
1
2
1
Eurolophosaurus
amathites
2
1
2
—
1
—
1
1
Eurolophosaurus
nanuzae
2
2
2
2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Uranoscodon
superciliosus
Microlophus
quadrivittatus
Tropidurus hispidus
Punta Blanca1111
Santo Ina ´cio1111
All localities are in the Brazilian state of Bahia, except for Pedra Menina
and Serra do Cipo ´ (Minas Gerais), and Manaus (Amazonas). Punta
Blanca is located in Chile.
J.C. Passoni et al./Molecular Phylogenetics and Evolution 46 (2008) 403–414
405

Page 4

polymerase chain reaction (PCR) in 25 or 50 ll reaction
volumes containing 20–30 ng DNA, 1? PCR buffer,
4 mM MgCl2, 0.2 mM each dNTP, 1 lM each primer,
and 2.5 UAmpliTaq DNA
Elmer, Roche). Amplification primers and thermocycler
conditions are specified in Table 2. Negative controls were
made to avoid contamination. Amplification products were
checked by electrophoresis on 0.8% agarose gel with
0.4 lg/ml ethidium bromide (the target fragment size was
estimated from molecular-weight markers). Products were
cleaned with Exonuclease and Shrimp Alkaline Phospha-
tase (Amersham Pharmacia Biotech.) per 1 h at 37 ?C
and, subsequently, per 10 min at 80 ?C. The Concert Rapid
PCR Purification System (Gibco BRL-Life Technologies)
was also used with some samples. In cases of non-specific
amplifications, the Concert Rapid Gel Extraction System
(Gibco BRL-Life Technologies) was used. Purified PCR
products containing about 100 ng DNA were utilized for
sequencing using the Big Dye Terminator Ready-Reaction
Kit (Perkin-Elmer Applied Biosystems). A Perkin-Elmer-
2400 thermocycler was used with the following parameters:
96 ?C (0:10), 50 ?C (0:05), 60 ?C (4:00) ? 25 (0:05 at 96 ?C
before first cycle). Both strands of each segment were
sequenced for each specimen. An ABI PRISM 310 Genetic
Analyzer (Perkin-Elmer Applied Biosystems) was used and
the sequences were edited using Sequence Navigator 1.0.1
(Perkin-Elmer Applied Biosystems).
Polymerase (Perkin-
2.3. Phylogenetic analyses
Multiple alignments were accomplished using Clustal X
1.64b (Thompson et al., 1997). For 12S and 16S segments,
gap-opening values of 6, 8, 10, and 12, and a gap-extension
valueof0.05wereusedtoassesstheeffectsonsequencealign-
ment according to Gatesy et al. (1993), Wiens and Reeder
(1997) and Reeder and Montanucci (2001). Regions of
ambiguous alignments were excluded. 12S rRNA secondary
structuremodelsforascincidlizard(Hicksonetal.,1996)for
the lizard family Opluridae (Titus and Frost, 1996) and for
hyperoliid treefrogs (Richards and Moore, 1996), and the
16S rRNA secondary structure models for humans (Gutell
and Fox, 1988) and for mammals (Burk et al., 2002) were
considered. Cyt b and COI protein-coding gene sequences
weretranslatedintoaminoacidsequencestocheckforunex-
pectedoccurrencesofstopcodons,whichmightindicatethat
pseudogenes had been amplified (Sorenson and Fleischer,
1996; Zhang and Hewitt, 1996). The degree of saturation
ineachgenesegmentwasinvestigatedbyplottingtransitions
and transversions against Lake 94 distances employing
DAMBEprogram(Xia,2000;XiaandXie,2001).Inthecase
of protein-coding segments, these analyses were accom-
plished considering all codon positions, the first and second
codon positions together, and each position separately.
Maximum parsimony (MP) analyses based on all indi-
viduals of Table 1 along with molecular (Martins, 1995;
Passoni et al., 2000) and morphological data (Rodrigues,
1986, 1996; Skuk, 1994) have evidenced monophyletic pop-
ulation units in Eurolophosaurus (see Section 3.2). There-
fore, the four gene segments of one or two individuals of
each monophyletic population unit could be used for main
phylogenetic analyses, as follows.
The partition homogeneity test (Farris et al., 1994) using
PAUP*(1000replicates)indicatedthatthefourgeneregions
couldbeanalysedtogether.Phylogeneticsignalwasdetected
for the combined data set using the g1statistic (Sokal and
Rohlf, 1981; Hillis, 1991; Huelsenbeck, 1991; Hillis and
Huelsenbeck,1992),whichmeasuredtheskewnessofthedis-
tributionof10,000randomcladograms.Phylogeneticanaly-
ses were conducted on ribosomal sequences and protein-
coding sequences separately, and on combined mitochon-
drial data for all populations, using maximum parsimony
(MP), maximum likelihood (ML), and neighbor joining
(NJ) methods implemented in PAUP*4.0b10 (Swofford,
2002). Parsimony trees were constructed using a branch-
and-bound search, and the confidence tested by 1000 boot-
strap replicates (Felsenstein, 1985). Due to saturation on
third codon positions, MP analyses were done considering
two different weightings for those characters, 0 and 1. In
thecaseof theribosomalsegments, gap wastreatedas afifth
character state. For ML heuristic searches and NJ tree esti-
mates, the evolutionary model that best fits each data set
was selected employing PAUP*and MODELTEST 3.06
Table 2
List of PCR and sequencing primers, and a summary of the PCR conditions for all four gene segments
Primer labelSequence (50–30)PCR conditions: denaturation/annealing/extension ? 35*
95 ?C (1:00), 55 ?C (1:00), 72 ?C (0:30)B1a
B2a(Cyt b)
L7354b
CCATCCAACATCTCAGCATGATGAAA
GCCCCTCAGAATGATATTTGTCCTCA
TACCAACACCTATTCTGATT 92 ?C (0:45–1:00), 47 ?C–53 ?C (0:45) or 54 ?C (1:00),
68 ?C (1:00) or 72 ?C (0:45)
H7794b(COI)
L1091c
H1478c(12S)
Ar-50d
Br-30d(16S)
ATAATGGCAAATACTGCCCC
AAAAAGCTTCAAACTGGGATTAGATACCCCACTAT
TGACTGCAGAGGGTGACGGGCGGTGTGT
CAAACCCCGCCTGTTTACCAAAAACAT
CCGGTCTGAACTCAGATCACGT
95 ?C (1:00), 54 ?C (1:00), 72 ?C (0:30)
95 ?C (1:00), 54 ?C (1:00), 68 ?C (1:00)
Incubation times are given in parentheses in minutes:seconds.
*At 92 or 95 ?C for 5:00 before first cycle, and for 7:00 at 68 or 72 ?C after the last cycle. Reference for primers are:aFu et al. (1997);bFrost et al. (1998);
cKocher et al. (1989);dPalumbi (1996).
406
J.C. Passoni et al./Molecular Phylogenetics and Evolution 46 (2008) 403–414

Page 5

(Posada and Crandall, 1998) programs. The substitution
model for all four concatenated segments was found to be
the General-Time-Reversible model (GTR + C + I; Rodrı ´-
guez et al., 1990). In ML analyses, tree bisection-reconnec-
tion (TBR) branch swapping, MULTREES, and random
addition of sequences (100 replicates) were employed. Boot-
strap values (BVs) were estimated by 100 replicates with 10
replicates of random addition of sequences. In NJ analyses,
BVswerecalculatedby1000replicates.Theinterpretationof
BVs is still a controversial issue. We followed Shaffer et al.
(1997):BVs P 90%wereconsideredashighlysignificantval-
ues; 70 6 BVs < 90% as marginally significant proportions
and 50% 6 BVs < 70% as suggesting limited evidence of
monophyly. For discussion purposes, we consider only
highly significant BVs.
2.4. Divergence time analyses
Absence of fossils and geomorphological records pre-
cludes independent testing of divergence times for the
Eurolophosaurus radiation. Previous molecular divergence
time estimates were based only on a reduced sample of
allozymic loci (Martins, 1995). Therefore, in the present
Eurolophosaurus evolutionary history context, it is worth-
while determining time intervals in order to have at least
rough date estimates and also to test the vicariant hypoth-
esis of 12,000 years ago. With this purpose, the corrected
genetic distances based on the gene sequences of the main
phylogenetic analyses were used.
The hypothesis of rate constancy was tested by the like-
lihood-ratio test (Muse and Weir, 1992) for each gene
sequence separately, for ribosomal sequences, for protein-
coding sequences, and for all mitochondrial sequences.
This test compares the log likelihood (ln L) of a tree using
an assumption of clock-like evolution with the log likeli-
hood of a tree calculated without this assumption
(P = 0.05). PAUP*and MODELTEST were used to select
the most appropriate model of evolution for each data set
and PAUP*was used to calculate each log likelihood value.
Evolutionary rate constancy was detected only for cyt b,
16S and the four concatenated gene segments.
Since we obtained the first considerable amounts of
mtDNA sequence data for Eurolophosaurus populations,
we have an opportunity to obtain secure time limits for the
evolutionary history of this genus. The highest rates of
molecular evolution found in literature, along with lower
rates, were appliedfor thecombineddataset. Toallow com-
parison with published data, the Kimura Two-Parameter
(K2P) model (Kimura, 1980) was used to correct the cyt b
genetic distances.
3. Results
3.1. Molecular characterization
About 1360 mtDNA base pairs (bp) were sequenced for
22 specimens of Eurolophosaurus and the three outgroup
individuals. Appendix I shows the aligned 12S DNA
sequences used in the main phylogenetic analyses, with
the identified stem and loop regions, from which 33 ambig-
uously aligned characters were removed. Seven insertion/
deletion events—indels—(characters 123, 124, 128, 129,
189, 194, and 275) could be precisely aligned in 12S
sequences which, along with others, did not disrupt the
RNA secondary structure. For 16S ribosomal sequences
(Appendix II), a region of 98 positions was excluded due
to ambiguous alignment. Indels in 16S DNA sequences,
six of which could be unambiguously aligned (characters
16, 17, 317, 331, 355, and 365) were all located in unpaired
regions. Cyt b and COI protein-coding gene sequences con-
tained no indels. The resulting phylogenetic data matrix
was composed of 282 characters of 12S rRNA gene, 389
characters of 16S rRNA gene, 209 bp of cyt b and 349 bp
of COI genes. These sequences, used in the main phyloge-
netic analyses, have been deposited in GenBank under
Accession Nos.: cyt b (DQ848725–DQ848735); COI
(DQ848747–DQ848757);12S
16S (DQ848736–DQ848746). The remaining sequences,
used in other analyses, have been deposited in GenBank
under Accession Nos.: cyt b (DQ848713–DQ848724);
COI (DQ848769–DQ848772);
DQ848776); 16S (DQ848777 and DQ848778). Several
observations indicate that the sequences analyzed here cor-
respond to functional mitochondrial genes (see Zhang and
Hewitt, 1996). Protein-coding sequences do not have pre-
mature stop codons, and ribosomal sequences apparently
code for rRNAs with stable secondary structures. Nucleo-
tide frequencies shown in Table 3 were estimated from all
sequences referenced in Table 1. The bias against guanine
on light strands of the protein-coding sequences and the
bias against guanine and thymine on light strands of the
ribosomal sequences support the conclusion that the
sequences do not represent nuclear-integrated copies of
mitochondrial genes (e.g. Zhang and Hewitt, 1996; Macey
et al., 1997, 1999; Honda et al., 2000a, b; Tu et al., 2000;
Ast, 2001; Daniels et al., 2002; Schulte et al., 2002).
(DQ848758–DQ848768);
12S(DQ848773–
3.2. Phylogenetic relationships
Our data diagnose at least six monophyletic population
units in Eurolophosaurus. Although our cladograms based
Table 3
Mean base frequencies, X2and P values in homogeneity test for base
composition among taxa for each DNA segment
GeneCyt b
COI12S16S
Bpa
%A
%C
%G
%T
X2
P
209
27.1
29.1
15.7
28.1
6.0346
1.00
348–349
28.7
22.8
19.8
28.7
10.9451
0.99
307–313
36.2
26
19.2
18.6
3.18
1.00
456–468
32.8
27.2
20.1
19.9
6.0138
1.00
aNumber of bases.
J.C. Passoni et al./Molecular Phylogenetics and Evolution 46 (2008) 403–414
407