Syst. Biol. 63(1):96–110, 2014
© The Author(s) 2013. Published by Oxford University Press, on behalf of the Society of Systematic Biologists. All rights reserved.
For Permissions, please email: [email protected]
DOI:10.1093/sysbio/syt062
Advance Access publication October 22, 2013
Timeframes of Speciation, Reticulation, and Hybridization in the Bulldog Bat Explained
Through Phylogenetic Analyses of All Genetic Transmission Elements
FAISAL ALI ANWARALI KHAN1,2,∗ , CALEB D. PHILLIPS1 , AND R OBERT J. BAKER1
1 Department
of Biological Sciences and the Museum, Texas Tech University, Lubbock, TX 79409, USA and 2 Department of Zoology, Faculty of Resource
Science and Technology, Universiti Malaysia Sarawak, Kota Samarahan, Sarawak 94300, Malaysia
∗ Correspondence to be sent to: Department of Zoology, Faculty of Resource Science and Technology, Universiti Malaysia Sarawak,
94300 Kota Samarahan, Sarawak, Malaysia; E-mail: [email protected].
Received 22 February 2013; reviews returned 19 April 2013; accepted 16 September 2013
Associate Editor: Thomas Near
Abstract.—Phylogenetic comparisons of the different mammalian genetic transmission elements (mtDNA, X-, Y-, and
autosomal DNA) is a powerful approach for understanding the process of speciation in nature. Through such comparisons
the unique inheritance pathways of each genetic element and gender-biased processes can link genomic structure to
the evolutionary process, especially among lineages which have recently diversified, in which genetic isolation may be
incomplete. Bulldog bats of the genus Noctilio are an exemplar lineage, being a young clade, widely distributed, and exhibiting
unique feeding ecologies. In addition, currently recognized species are paraphyletic with respect to the mtDNA gene tree
and contain morphologically identifiable clades that exhibit mtDNA divergences as great as among many species. To test
taxonomic hypotheses and understand the contribution of hybridization to the extant distribution of genetic diversity in
Noctilio, we used phylogenetic, coalescent stochastic modeling, and divergence time estimates using sequence data from
cytochrome-b, cytochrome c oxidase-I, zinc finger Y, and zinc finger X, as well as evolutionary reconstructions based on
amplified fragment length polymorphisms (AFLPs) data. No evidence of ongoing hybridization between the two currently
recognized species was identified. However, signatures of an ancient mtDNA capture were recovered in which an mtDNA
lineage of one species was captured early in the noctilionid radiation. Among subspecific mtDNA clades, which were
generally coincident with morphology and statistically definable as species, signatures of ongoing hybridization were
observed in sex chromosome sequences and AFLP. Divergence dating of genetic elements corroborates the diversification
of extant Noctilio beginning about 3 Ma, with ongoing hybridization between mitochondrial lineages separated by 2.5
myr. The timeframe of species’ divergence within Noctilio supports the hypothesis that shifts in the dietary strategies of
gleaning insects (N. albiventris) or fish (N. leporinus) are among the most rapid instances of dietary evolution observed in
mammals. This study illustrates the complex evolutionary dynamics shaping gene pools in nature, how comparisons of
genetic elements can serve for understanding species boundaries, and the complex considerations for accurate taxonomic
assignment. [AFLP; genetic transmission elements; hybridization; Noctilio; X-chromosome; Y-chromosome.]
The process of speciation is frequently studied
through identifying and comparing patterns of
variability in morphology, ecology, and genetic
elements. A large portion of species-level taxonomy
is based on a narrow array of data types, such as
morphology and mitochondrial phylogenies, leaving
much evolutionary diversity and complexity yet to
be discovered. It is hypothesized that the number
of valid mammalian species yet to be described
remains in the thousands (Baker and Bradley 2006;
Clare et al. 2007; Francis et al. 2010; Clare 2011). This
estimate is based partially on the observation that
detailed morphological and mitochondrial studies often
identify cryptic species. However, complex evolutionary
histories involving introgression and hybridization
are also described, and the incorporation of genetic
data from differentially inherited genetic elements is
valuable for characterizing such phenomena (e.g., Good
et al. 2003, 2008; Buckley et al. 2006; Linnen and Farrell
2008; Maureira-Butler et al. 2008; Bossu and Near 2009;
Larsen et al. 2010; Reid et al. 2012). Results such as these
are refining and redefining our understanding of the
process of speciation and the complexity of systematic
biodiversity.
Inferring species trees is often complicated by
incongruency among gene trees (Knowles 2009). In
order to understand the process of speciation it is
important to identify the sources of phylogenetic
incongruency (Reid et al. 2012). The lineage sorting
process and the stochastic nature of genetic mutation are
well-known phenomena that can result in phylogenetic
incongruency (Pamilo and Nei 1988; Takahata 1989;
Degnan and Rosenberg 2006). Incongruency based
on these phenomena is not independent from effects
of divergence time, drift, selection, and demography.
In contrast, phylogenetic incongruency as a result of
hybridization identifies on-going or past gene flow
(Rieseberg et al. 2000). It is expected that incongruency
arising via hybridization events can be detected in the
study of recent adaptive radiations, as these complexes
are more likely to exhibit incomplete reproductive
isolation given recency of divergence. It is also true
that lineage sorting and stochastic mutational processes
can be influential in phylogenetic reconstructions of
recent divergences. Owing to the different inheritance
pathways of the various autosomal, sex, and organeller
chromosomes, phylogenetic comparisons among
these genetic elements can be used to identify and
characterize hybridization. Specifically, comparative
phylogenetic analysis of genetic variation distributed
across genetic elements yields perspectives about
the presence and nature of hybridization that is
distinguishable from sorting or mutational effects on
individual loci.
96
[14:29 4/12/2013 Sysbio-syt062.tex]
Page: 96
96–110
2014
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
Studies based on ecology, morphology, and genetics
indicate the bulldog bat genus Noctilio, of the monotypic
family Noctilionidae, is a young clade, vagile, and
widely distributed. Comparative phylogenetic analysis
of the mammalian genetic transmission elements could
document details of this history and timeframes of
divergence and hybridization. Noctilio consists of two
species, N. albiventris and N. leporinus. The two species
are broadly sympatric throughout their ranges in Central
and South America, and N. leporinus is also distributed
across the Caribbean Islands. Although these species
are sympatric and display minimal morphological shape
differences, they can be differentiated based on size, with
N. leporinus being considerably larger than N. albiventris
(Hood and Pitocchelli 1983; Hood and Jones 1984).
Although N. leporinus feeds broadly on fish, flying
insects, and aquatic invertebrates, N. albiventris primarily
is insectivorous. However, there is evidence from
stomach content analyses showing that N. albiventris
is able to consume fish and fruit (Howell and Burch
1974; Hood and Pitocchelli 1983; Hood and Jones 1984;
Bordignon 2006; Gonçalves et al. 2007). As no other
member of the superfamily Noctilionoidea (consisting of
7 families and an estimated 180 species) has developed
a piscivorous life history strategy, and N. leporinus still
retains a partially insectivorous dietary strategy, it is
likely that piscivory has evolved recently in this genus
(Lewis-Oritt et al. 2001; Pavan et al. 2013).
Systematic studies have indicated a complex yet
poorly defined evolutionary history for Noctilio. This
complexity has been partially characterized by multiple
subspecific phylogroups within N. albiventris, some of
which may be valid species and retain a primarily
insectivorous dietary strategy (Lewis-Oritt et al. 2001).
Noctilio leporinus consists of two moderately diverged
(≤ 3.2% cytochrome-b genetic distance) clades (LewisOritt et al. 2001; Clare et al. 2007). Taxonomic revisions
of both species have been based on bivariate analyses
of morphological characteristics (Davis 1973, 1976; Timm
and Genoways 2003). These studies divided N. albiventris
into the four subspecies N. a. affinis, N. a. albiventris,
N. a. cabrerai, and N. a. major. However, N. a. affinis was
later subsumed and considered as the junior synonym
of N. a. albiventris based on the finding of specimens
with intermediate morphology from French Guiana
(Simmons and Voss 1998; Simmons 2005; Gardner 2008).
Noctilio leporinus has been partitioned by morphology
into the three distinct subspecies N. l. leporinus,
N. l. mastivus, and N. l. refuscens (Davis 1973, 1976). Results
from these studies of subspecific variation document
the procession of a dynamic and ongoing evolutionary
radiation in bulldog bats, and the extent of ongoing
gene flow among these evolutionary clades remains
unknown.
Detailed phylogenetic characterization of this
complex would not only yield an improved estimate
of biodiversity for critical ecoregions, but would also
improve our understanding of the process of speciation
in recently radiated and highly vagile mammalian
lineages. The study of all genetic transmission
[14:29 4/12/2013 Sysbio-syt062.tex]
97
elements in this evolutionary background would
inform timeframes of divergence across the different
genetic transmission elements and would characterize
hybridization among diverging lineages. To this end,
patterns of genetic variation at mitochondrial, X-, Yand autosomal chromosome loci were compared using
several phylogenetic measures. These comparisons
provide a method for identifying and characterizing
the evolutionary legacies of stochastic sorting,
introgression, and hybridization in recent radiations
that avoids the pitfalls of less comprehensive genomic
sampling.
MATERIALS AND METHODS
Taxonomic Sampling
Tissue samples were obtained from the following
collections: Museum of Texas Tech University (TK,
Lubbock), Royal Ontario Museum (ROM, Toronto),
Louisiana Museum of Natural History (LSUMZ, Baton
Rouge), American Museum of Natural History (AMNH,
New York), National Museum of Natural History
(USNM, Washington, D.C.), Museum of Vertebrate
Zoology (MVZ, Berkeley), Sam Noble Oklahoma
Museum of Natural History (SNOMNH, Norman),
Museum of Southwestern Biology (MSB, Albuquerque)
and Carnegie Museum of Natural History (CMNH,
Pittsburgh). Specimens were selected based on each
museum’s recorded locality information in order to
sample the geographic range of the four subspecies
of N. albiventris and three subspecies of N. leporinus
(Davis 1973, 1976; Fig. 1). Online Appendix 1 (doi:
10.5061/dryad.77c1h) contains information about
museum voucher specimen/tissue number and
associated GenBank accession numbers for cytochromeb (Cyt b), cytochrome c oxidase-I (COI), zinc finger X
(ZFX), zinc finger Y (ZFY), and Amplified Fragment
Length Polymorphisms (AFLPs). Carollia perspicillata,
Mormoops megalophylla, Pteronotus parnelli, P. personatus,
P. quadridens were also sequenced and genotyped, and
used as outgroups (Online Appendix 1).
Mitochondrial DNA Sequencing
Total genomic DNA was extracted from muscle or
liver tissues by standard phenol/chloroform methods
(Sambrook et al. 1989). The entire Cyt b gene (1140 bp;
n = 83) and a portion of the COI gene (657 bp; n = 82) were
PCR amplified and sequenced to explore mitochondrial
(mtDNA) variation within the sampling of N. albiventris
and N. leporinus. Amplification for Cyt b was performed
using primers LGL765/LGL766 (Bickham et al. 2004).
Amplifications of COI included multiple primer pairs
(VF1/VR1, VF1i/VR1i, and VF1d/VR1d; Ivanova et al.
2006). Twenty-five L reaction volumes included ∼200
ng DNA, 0.12 M of each primer, 1.5 mM MgCl2,
0.012 mM deoxynucleoside triphosphates, 1X reaction
buffer, 0.32 mg/ml Bovine Serum Albumin, and 0.625U
Page: 97
96–110
98
SYSTEMATIC BIOLOGY
VOL. 63
30 L of buffer EB. Portions of ZFX (819–847 bp)
and ZFY (733–754 bp) were sequenced using internal
sequencing primers developed for Noctilio: ZFX-NocF
5 TGGAAATGAAAAATCCCTCAA3 and ZFX-NocR
5 TTTCTGAGATGCAAATGGGTA3 for ZFX, and
ZFY-NocF: 5 GGCATAATGCAGGGATCATT3 and
ZFY-NocR:
5 CTGAAACATTACTGGTTCCAAGA3
for ZFY. Allele-specific priming was not required for
ZFX sequences, as only one individual was observed to
be heterozygous.
FIGURE 1. Sampling localities for N. albiventris (left) and N. leporinus
(right). Sampling localities are indicated with symbols, and unique
symbols indicate the mitochondrial clade (Fig. 2) recovered at each
locality ( = Albiventris 1,
= Albiventris 2,
= Albiventris 3,
= Albiventris 4,
= Leporinus 1, and
= Leporinus 2).
Numbers on the symbols refer to the total number of samples from that
particular locality. Localities with one sample are shown by symbols
without numbers. Shaded regions on each map depict the geographic
range for subspecies following (Davis 1973; 1976) and are identified
with capital letters (N. albiventris: A = N. a. minor, B = N. a. affinis,
C = N. a. albiventris, D = N. a. cabrerai; N. leporinus: A = N. l. mastivus,
B = N. l. leporinus, C = N. l. rufescens)
Taq DNA polymerase (Promega Corporation, Madison,
Wisconsin). Thermal cycling conditions were 94◦ C for
2 min, followed by 34 cycles of 94◦ C for 45 s, 50◦ C
(45◦ C for COI amplification) for 1 min, and 72◦ C for 1
min 15 s, followed by a final extension of 72◦ C for 10
min. PCR amplicons were purified using QIAquick PCR
Purification Kits (Qiagen Inc., Chatsworth, California)
and sequenced using ABI Big Dye chemistry version 3.1,
and an ABI 3100-Avant Genetic Analyzer (PE Applied
Biosystems, Foster City, California). The entire Cyt b gene
of 1140 bp was sequenced using amplification primers
in addition to five internal primers: N360F and N956R,
(Lewis-Oritt et al., 2001); ART16, (Larsen et al., 2007);
G5L, (Hoffmann and Baker, 2001); MVZ04, (Smith and
Patton, 1993). COI amplicons were sequenced using
amplification primers.
Y- and X-chromosome Gene Sequencing
A total of 28 male Noctilio were sequenced for a
portion of the paternally inherited ZFY gene, and 33 (8
female and 25 male) were sequenced for a portion of the
bi-parentally inherited ZFX. Sampling included at least
one representative from each phylogroup identified in
mtDNA analysis. Introns from ZFX and ZFY were coamplified using primers LGL331 and LGL335 (Cathey
et al. 1998). The amplified fragments corresponded
to intron 10 of ZFX and intron 7 of ZFY compared
with human genome. Amplicons were separated on
1.5 % agarose gels. Males yielded two amplification
size products (ZFX ∼900 bp and ZFY ∼850 bp), and
females yielded a single product (ZFX). ZFX and
ZFY amplicons were gel excised and purified using
QiaQuick Gel Extraction Kits (Qiagen Inc., Chatsworth,
California) and final products were reconstituted in
[14:29 4/12/2013 Sysbio-syt062.tex]
AFLP Genotyping
Our AFLP protocol followed that of (Vos et al., 1995)
and (Phillips et al., 2007) with modifications on primer
sets, following (Khan et al., 2010). Fluorescently labeled
selective EcoRI primers in combination with 7 other
selective primers were used to generate unambiguous
and distinct AFLP for 56 individuals representing at
least four or more members of recovered mtDNA clades.
AFLP profiles were anonymously double read, and
only fragments (50–400 bp in length) with a signal
larger than 50 relative fluorescent units were retained
as confident genotype calls. GeneMapper version 4.0
software (Applied Biosystems) was used to manually
edit and verify all data. From the resulting binary
data matrix, descriptive statistics were calculated in
GenAlEx version 6 software (Peakall and Smouse 2006).
Error rates were calculated following (Bonin et al.,
2004) using 50 replicated samples (∼13% of the overall
sample size).
Sequence Variation and Study Comparisons
Initial analysis of the data addressed various genetic
summary statistics, substitution trends, and overall
comparisons of the geographic distribution of variation.
To this end, DnaSPv5 (Librado and Rozas 2009) was
used to resolve the number of unique haplotypes and
polymorphic sites for each locus. Substitution saturation
was evaluated for each locus using the Xia index,
calculated in DAMBE version 5.2.69 (Xia and Xie 2001;
Xia et al. 2003). For the mtDNA portion of the data set
(Cyt b and COI), Kimura 2-parameter (K2P) distance
matrices were generated in MEGA version 5 (Tamura
et al. 2011). K2P distance values for these genes were
calculated to describe genetic distances separating major
clades using markers and a metric commonly used
for species identification (Baker and Bradley 2006). For
this analysis, mtDNA clades compared were those that
through phylogenetic analysis retained 80% or higher
ML bootstrap values or 0.95 or higher Bayesian posterior
probabilities (see below). Sequences from Cyt b and
COI previously deposited on GenBank (Lewis-Oritt
et al. 2001; Clare et al. 2007; Cyt b: AF330794-806; COI:
EF080519–522, 524, 525, 527, 528, 530–533, JF448375)
were incorporated to increase sampling and improve
assessment of the geographic range for each phylogroup
Page: 98
96–110
2014
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
identified and relation to proposed distributions of
subspecies (Davis 1973, 1976).
Phylogenetic Analyses
Phylogenetic reconstructions were initially performed
for each mtDNA gene separately to assess phylogenetic
congruence among gene tree topologies. Mitochondrial
genes were subsequently concatenated into unique
composite haplotypes. Maximum-Likelihood and
Bayesian phylogenetic reconstructions were generated
for each gene, the concatenated mtDNA data set, and sex
chromosome markers. Appropriate models of molecular
evolution for each gene, concatenated mtDNA data set,
and sex chromosome markers were determined using
the Bayesian Information Criterion in MEGA version 5
(Tamura et al. 2011), and the best-fit model implemented
in MrBayes version 3.1.2 (Huelsenbeck and Ronquist
2001) was used. Sequences for Carollia, Mormoops,
Pteronotus, and GenBank sequences of Furipterus (Cyt
b: AY621004; COI: JF448846) were used as outgroups
in the mtDNA phylogenetic reconstructions; however,
Furipterus sequences were not available to include
in X- and Y-chromosome gene sequence analyses.
Phylogenetic analyses were conducted in PhyML
version 3.0 (Guindon and Gascuel 2003) and MrBayes.
Maximum-Likelihood analyses were initiated using the
BioNJ (Gascuel 1997) starting trees. Bootstrap support
values were obtained through 1000 subsampling
replicates. Bayesian phylogenetic reconstructions were
performed using four independent MCMC chains
consisting of two independent runs at 10 million
generations each, and trees were logged every 100th
iteration. As recommended by the program authors,
iterations were performed until the average standard
deviation of split frequencies was less than 0.01
(Ronquist et al. 2005). The stability of log likelihoods
were examined in TRACER version 1.5 (Rambaut
and Drummond 2007) and the first 10,000 trees were
discarded as burn-in.
Given the strong morphological signal separating
N. albiventris from N. leporinus (Davis 1973, 1976; Timm
and Genoways 2003), it was possible that previous
phylogenetic analyses on this group reporting polyphyly
(Lewis-Oritt et al. 2001; Pavan et al. 2013) were the result
of outgroup confounding effect, given the evolutionary
distance of available outgroups for Noctilio (Holland
et al. 2003). To examine this condition, additional
Bayesian phylogenetic analyses were conducted in
MrBayes without outgroups (Ronquist et al. 2005), and
implementing a strict molecular clock. Huelsenbeck et al.
2002 have previously demonstrated the ability of this
method to determine proper rooting.
The AFLP binary matrix was transformed into Nei–Li
genetic distances and used to generate a NeighborJoining tree (NJ) with 1000 bootstrap replicates in PAUP,
version 4.0b10 (Swofford 2002). Bayesian analysis was
performed for the AFLP data set using the restriction
site model in MrBayes that allowed correction for
[14:29 4/12/2013 Sysbio-syt062.tex]
99
coding bias (Ronquist et al. 2005). We used the option
“noabsencesites”, which most closely fits the bias in
AFLP data sets (Ronquist et al. 2005). The Dirichlet
prior for the state frequencies was set to (0.38, 0.62),
matching the actual 0/1 frequencies in the data set.
Iterations and final phylogeny estimation followed the
same steps described for the DNA sequence data.
Principal coordinate analysis (PCoA) was performed
on the AFLP data matrix using GenAlEx to describe
multidimensional relationships and major patterns of
genetic variance within and among mtDNA lineages.
The first three components from PCoA were used
in generating a 3D scatterplot with the minimum
spanning network option in MYSTAT version 12 (Systat
Software, Inc., USA). Ellipses were drawn around major
geographic clusters.
Because phylogenetic comparisons across loci
disclosed patterns of hybridization among subspecies
within each species, a formal statistical test for the
presence of hybrid individuals was conducted using
the AFLP data matrix. For this analysis, NewHybrids
version 1.1 (Anderson and Thompson 2002) was used
to determine posterior probabilities that individuals
belonged to a hybrid category (F1, F2, backcrosses).
Hybrid testing was conducted for N. leporinus and
N. albiventris separately. For N. leporinus this analysis
conformed to a two-parental-type system as only
two major genetic clades were disclosed through all
phylogenetic analyses. Within N. albiventris four mtDNA
clades were resolved and phylogenetic reconstructions
based on AFLP indicated mixed ancestry. Therefore
hybrid analysis for this species was also conducted as
a two-parental-type system. Although this design was
potentially too simplistic for N. albiventris, a history
which is most robustly diagnosed through phylogenetic
and variable reduction methods (described above),
formal hybrid classification is limited to a two-parentaltype system. The analysis consisted of a total of 5500
sweeps, in which the first 500 sweeps were removed
as burn-in. The default recommendation of a uniform
prior was imposed for priors on mixing proportions
and allele frequencies, and no a priori assumptions were
made for class membership of samples.
Divergence Time Estimates
Divergence times were estimated separately for
unique haplotypes from each locus including 67
concatenated mtDNA haplotypes, 8 ZFY haplotypes,
and 13 ZFX haplotypes. Molecular clock testing was
performed using Likelihood Ratio Tests (Felsenstein
1988) implemented in MEGA. Following the results
of these tests a relaxed lognormal molecular clock
was implemented for the mtDNA data set and strict
molecular clocks for ZFX and ZFY. BEAST version
1.7.1 (Drummond and Rambaut 2007) was used to
estimate timescales of diversification for Noctilio using
both paleontological and molecular constraints. The
oldest fossil indistinguishable from modern N. albiventris
Page: 99
96–110
100
is known from the middle Miocene La Venta fauna,
∼12.5 Ma in Colombia (Czaplewski 1996; Czaplewski
et al. 2003). However, this fossil was not used to define
a node prior, as it would have constrained the time of
diversification of Noctilio, which is uncertain due to gaps
in N. leporinus fossil record (Meredith et al. 2011). Fossil
priors demarcating the Mormoopidae/Phyllostomidae
split recovered from the Whitneyan land deposits in
Florida were used as an outgroup node prior (dated
late Oligocene; Morgan 2002). This node was constrained
with an exponentially distributed prior having a
minimum age of 28.3 Ma with an exponential mean of
3.2, allowing the maximum age of this divergence to
occur ∼40.6 Ma following (Meredith et al., 2011). The
molecular date (= 43.01; SD = 1.5) for the node shared by
the families Furipteridae, Mormoopidae, Noctilionidae,
and Phyllostomidae in (Meredith et al., 2011) was also
used in our analyses and was described by a normally
distributed prior.
Phylogeny and time to most recent common ancestor
(tmrca ) for Noctilio nodes were estimated for the
concatenated mtDNA data matrix partitioned by gene.
Previously estimated models of evolution (Cyt b:
HKY+I+G, COI: HKY+G) were applied with substitution
rate parameters, rate heterogeneity model, and base
frequencies unlinked across codon positions. Analyses
for ZFX and ZFY intronic sequences implemented the
HKY+I and HKY models of evolution, respectively.
Carollia, Mormoops, and Pteronotus were used as
outgroups in phylogenetic reconstructions. Node dates
were estimated using a birth–death process prior as
proposed for reconstructing phylogenies not including
fossil lineages (Gernhard 2008). Analyses consisted
of two independent runs of 50×106 iterations for
the mtDNA data set, and two independent runs of
10×106 iterations for the ZFX and ZFY data sets,
with every 1000th iteration logged for all analyses.
Independent runs were combined for each data set using
LogCombiner version 1.7.1 (Drummond and Rambaut
2007). TRACER version 1.5 (Rambaut and Drummond
2007) was used to determine appropriate burn-in
(∼10%), examine convergence, effective sample sizes
(ESSs), and 95% highest probability density intervals
(HPD) of constrained priors. To compare effects of
implementing different speciation models, additional
analyses were conducted using a Yule species prior (Yule
1925) for the mtDNA (including only two random unique
haplotypes per clade) and ZFY data sets (including all
unique haplotypes) with 50×106 iterations performed
for each data set, with every 1000th iteration logged.
Results from both speciation models were compared for
molecular dating incongruences.
Species Delimitations Using Coalescent Model
In order to identify independently coalescing lineages
that may correspond to species without a priori
assumptions of population membership, we used the
General Mixed Yule-Coalescent Model (Pons et al. 2006).
[14:29 4/12/2013 Sysbio-syt062.tex]
VOL. 63
SYSTEMATIC BIOLOGY
We used bGMYC, a Bayesian implementation of the
model that accounts for uncertainty in the phylogeny
and model parameters by sampling trees and conducting
MCMC (Reid and Carstens 2012). For this analysis
the ultrametric concatenated mtDNA trees for unique
haplotypes from the BEAST analyses were analysed. The
two original tree files with 50 million generations were
combined after discarding the first 5 million as burnin. All parameters for all replicates retained ESSs greater
than 200. The combined tree file was resampled at a lower
frequency resulting in 90 trees (one tree every 1,000,000
generations). Initial bGMYC analyses using only a single
tree from the population of trees suggested that the
MCMC reached stationarity after 10,000 steps. All 90
trees were analysed for 50,000 generations, discarding
the first 40,000 as burn-in and sampling every 100th
iteration. MCMC samples from each tree were pooled
together to calculate the probabilities that any two
leaves in the phylogeny were conspecific. We used a
probability of being conspecific of 0.5 or higher to cluster
individuals into putative species. The 50% threshold
value was selected because the number of clades
recognized as conspecifics using this threshold can be
readily supported by existing morphological data, and
corresponds to well-supported genetic clades. This was
considered a conservative measure as compared with
implementing larger threshold values which resulted in
identifying considerably shallower coalescing clades as
conspecific.
RESULTS
Mitochondrial Variation and Phylogenies
Cyt b and COI displayed similar levels of genetic
variation, with between 28 and 37 haplotypes recovered
for each gene within each species (Table 1). Tests
of substitution saturation yielded non-significant
results for both genes (Supplementary Table S1,
doi: 10.5061/dryad.77c1h). Average K2P sequence
divergence (Supplementary Table S2) within
N. albiventris was 0.05–0.07. Four mtDNA clades
(subsequently referred to as Albiventris 1–4; see
below for phylogenetic results) within currently
recognized N. albiventris were recovered with a pairwise
divergence value of 0.05 substitutions/site or greater.
The geographic range for these lineages generally
corresponds to previously recognized subspecies
(Figs. 1 and 2). Clades Albiventris 1, 2, and 4 were
restricted to South America, whereas Clade 3 was
restricted mainly to Central America with two samples
recorded from eastern Venezuela (El Manteco, Bolivar).
Two monophyletic mtDNA lineages were identified
within N. leporinus (Leporinus 1–2) with an estimated
sequence divergence value of 0.024 substitutions/site.
Noctilio leporinus genetic lineages correspond to two of
the three proposed subspecies ranges but with extensive
geographic overlap (Figs. 1 and 2). Leporinus 1 was
restricted to Central America, western Andes, and
Jamaica (Greater Antilles). Leporinus 2 was restricted
Page: 100
96–110
2014
101
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
TABLE 1. Number of analysed specimens (N) and number of haplotypes for N. albiventris followed by N. leporinus, polymorphic sites, model
of evolution implemented in Maximum-Likelihood (ML) and Bayesian (BA) analyses, and the results of molecular clock testing for four different
genes and the concatenated mtDNA data set
Polymorphic sites (%)
N
Haplotypes
Indels
Conserved
Singleton
Parsimony
informative
Variables
ML
BA
44, 52
50, 43
42, 39
21, 12
16, 12
37, 34
28, 27
30, 37
7, 6
6, 2
—
—
—
6
2
913 (80.1)
536 (81.6)
1458 (81.1)
816 (98.1)
723 (96.7)
34 (3)
14 (2.1)
55 (3.1)
4 (0.5)
2 (0.2)
193 (16.9)
107 (16.3)
284 (15.8)
12 (1.4)
24 (2.9)
227 (19.9)
121 (18.4)
339 (18.9)
16 (1.9)
26 (3.2)
TN93+G
K2+G
HKY+G
HKY+G
HKY+G
HKY+I+G
HK+G
HKY+G
HKY+I
HKY
Gene
Cyt b
COI
Cyt b & COI
ZFX
ZFY
Model of evolution
Test of molecular
clocka
P < 1.0E-11*
P < 1.9E-126*
P < 0.0014*
P < 0.7067
P < 0.9877
a P-values marked with (*) indicate the null hypothesis of equal evolutionary rates throughout the tree was rejected at an = 0.05 significance
level.
to South America, Lesser Antilles (Dominica, Grenada,
Grenadines, and Montserrat), and Panama.
Independent phylogenetic analyses for mtDNA genes
recovered minor differences in tree topology, especially
for the position of Albiventris 1 clade, and relationships
among Albiventris 2, 3, and 4. Through Bayesian and
Maximum-Likelihood analyses of Cyt b, Albiventris 1
was placed sister to the rest of Noctilio lineages with low
statistical support, whereas Albiventris 2 is sister to a
clade containing Albiventris 3 and 4, but only supported
in Maximum-Likelihood analyses. Through phylogeny
reconstruction using COI, Albiventris 1 also also
positioned sister to the rest of Noctilio with low statistical
support, whereas Albiventris 4 was sister to a clade
containing Albiventris 2 and 3, but only supported in
Maximum-Likelihood analyses (Supplementary Fig. S1).
Bayesian strict clock analyses of Cyt b and COI phylogeny
positioned Albiventris 1 as sister to Leporinus 1 and
2 with high support values. Maximum-Likelihood and
Bayesian analyses of the concatenated data set provided
very similar results (Fig. 2), retaining relatively higher
support values at terminal nodes as compared with
deeper nodes. These phylogenies had <50% support for
reciprocal monophyly of N. albiventris and N. leporinus
as currently recognized (Fig. 2). In contrast, Bayesian
strict clock analyses of the concatenated data set yielded
strong support for almost all terminal and deeper
nodes and provided strong statistical support for a
paraphyletic relationship of currently recognized species
(Fig. 3).
X- and Y-variation and Phylogenies
For 33 individuals sequenced for ZFX, 16 variable
positions resulted in seven and six haplotypes for
N. albiventris and N. leporinus, respectively (Table 1). For
28 individuals sequenced for ZFY, 26 variable positions
resulted in six and two haplotypes for N. albiventris
and N. leporinus, respectively. The total number of
haplotypes recovered for ZFX was greater than that
observed for ZFY. ZFY retained higher variability (3.2%
vs. 1.9%) and number of polymorphic sites (2.9% vs.
1.4%) as compared with ZFX. Statistical testing indicated
[14:29 4/12/2013 Sysbio-syt062.tex]
neither of these markers were significantly saturated
(Supplementary Table S1).
Differences in branching order between ZFX and ZFY
were recovered, and both Maximum-Likelihood and
Bayesian reconstructions produced the same branching
order for each locus. Phylogenetic reconstructions
for neither marker fully recovered mtDNA defined
lineages into exclusive sex chromosome clades.
Although sister relationships between species were
not recovered, N. leporinus was supported through
both sex-linked phylogenies. Relationships among
N. albiventris clades differed between phylogenies and
were poorly supported. For both species, a pattern in
both phylogenies was sharing of sex loci alleles within
species. Figure 2 contains information on geographic
localities for samples found to contain sex loci sequences
indicating a history of hybridization or incomplete
lineage sorting. Similar to the rooted phylogenies,
Bayesian strict clock analysis for both sex chromosome
loci recovered monophyly of N. leporinus and low
statistical support for monophyly of N. albiventris
(Fig. 3).
AFLP-Nuclear Marker Phylogeny and Analysis
Three hundred thirty-nine AFLP loci were scored for
56 samples. An error rate of 4.9% was estimated based
on replicate amplification and genotyping, with most
discrepancies originating from poor amplification of
low quality DNA. Highest variability was observed for
Leporinus 2 (31.3%), and the lowest was observed for
Albiventris 1 (12.7%). The average number of private
bands per clade across the entire data set was found to
be 10 (ranging from 1 within Leporinus 1, to 20 within
Albiventris 4). AFLP unbiased Nei genetic distance
values ranged from 3.3% (Albiventris 2 vs. Albiventris
4) to 20.1% (Albiventris 4 vs. Leporinus 1).
The AFLP unrooted Bayesian phylogeny supported
reciprocal monophyly of species, provided overall
similar patterns as those recovered through the
phylogenetic analyses of other loci, and were most
similar at the mtDNA clade level to the sex chromosome
phylogenies (Fig. 2). For example, members from
Page: 101
96–110
102
SYSTEMATIC BIOLOGY
VOL. 63
FIGURE 2. Phylogenetic relationships among Noctilionidae hypothesized from Maximum-Likelihood consensus trees for three different data
sets: Concatenated Cyt b and COI (top left), ZFY (top right), ZFX (bottom right), and Bayesian topology for unrooted AFLP nuclear loci (bottom
left). The distributions of mitochondrial phylogroups across all phylogenies are identified with dotted vertical lines and corresponding symbols.
Node values for clades are Maximum-Likelihood bootstrap support followed by Bayesian posterior probabilities. AFLP node support values
refer to Neighbor-Joining bootstrap values followed by Bayesian posterior probabilities. Dashes (-) indicate nodes with low statistical support
for particular analyses. n = number of individuals. * = individuals identified as hybrids through membership analysis of AFLP data. Outgroups
are not shown to conserve space.
Leporinus 2 (CMNH 78256, 78259: Bolivar, Venezuela;
ROM 104207: Canal Zone, Panama) were found inside
Leporinus 1. Admixture among N. albiventris clades was
common and only Albiventris 3 (Central America and
western Venezuela) was recovered as monophyletic.
Analysis of multidimensional relationships within
and among lineages using PCoA yielded similar
results and elucidated a geographic component
of potential admixture indicated through various
phylogenetic analyses (Supplementary Fig. S2). Two
major N. albiventris clusters were resolved corresponding
to Central and South American localities. Two
N. leporinus clusters were resolved; one containing
South American and Lesser Antillean localities, and the
other containing Central American, western Andes, and
Venezuelan localities.
[14:29 4/12/2013 Sysbio-syt062.tex]
Separate hybrid classifications for N. albiventris and
N. leporinus subspecies identified individuals belonging
to parental forms or the F2 hybrid category. In
all instances, individuals were assigned to these
classifications with at least a 0.91 posterior probability,
and in 95% of classifications assignments were made
with >0.99 posterior probability. For N. albiventris,
individuals classified as F2 hybrids were among those
positioned on intermediate branches in the AFLP
phylogeny among different mtDNA clade members,
and were samples collected within areas of sympatry
following genetic and morphological data (this study;
Davis 1973, 1976). Similarly, for N. leporinus the F2
hybrid classification included samples indicated in
phylogenetic analyses as being of hybrid ancestry, were
collected within the Guiana Shield, or were collected
Page: 102
96–110
2014
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
103
FIGURE 3. Phylogenetic relationships among Noctilionidae hypothesized from Bayesian strict clock analyses for three different data sets:
Concatenated Cyt b and COI (top left), ZFY (bottom left), and ZFX (bottom right). The distributions of mitochondrial phylogroups across all
phylogenies are identified with corresponding symbols. Bayesian posterior probabilities for node support are provided. Clades are collapsed
and represented as triangles. Clades with high statistical support (≥ 0.95) are shaded in black and clades with low support are shaded with open
triangles. The depth/width of the triangles is proportional to genetic variation within each node.
within the proposed geographic distribution of the other
clade/subspecies. Two distantly collected Paraguayan
samples were also identified as F2 hybrids. Figure 2
indicates which samples were identified as F2 hybrids
for both species.
[14:29 4/12/2013 Sysbio-syt062.tex]
Divergence Dating
Phylogeny, divergence times, and 95% HPD for
the radiation of Noctilio estimated using BEAST are
reported in Figure 4. ESSs for all statistics in both
data sets were more than 1000. MtDNA, ZFX, and
Page: 103
96–110
104
VOL. 63
SYSTEMATIC BIOLOGY
ZFY analyses dated the basal radiation of Noctilio to
3.3 Ma (95% HPD: 2.3 and 4.3 Ma), 3.41 Ma (95%
HPD: 1.8 and 5.2 Ma), 3.42 Ma (95% HPD: 2.1 and 4.9
Ma), respectively. The mean rate of mtDNA sequence
evolution was estimated at 0.014 substitutions/site/my
(95% HPD: 0.011–0.018). The clock rate for ZFX and
ZFY was estimated at 0.0016 substitutions/site/my (95%
HPD: 0.0013–0.0019), and 0.0030 substitutions/site/my
(95% HPD: 0.0025–0.0034), respectively. Mitochondrial
diversification among subspecies began as early as 2.8
Ma (Albiventris 2, 3, and 4) and as recent as 1.0 Ma
(Leporinus 1 and 2). In comparison, ZFX diversification
among subspecies started around 2.6 Ma (Leporinus 1, 2,
and Albiventris 1) and as recent as 0.34 Ma (Albiventris 1,
2, and 4). However, diversification of paternal subspecific
lineages was estimated to have occurred as long ago as
2.7 Ma (Albiventris 1, 2, 3, and 4) and as recent as 0.56 Ma
(Albiventris 1 and 4). Independent analyses using either
birth–death or Yule species priors resulted in similar tree
topology and divergence times (results not shown).
Species Delimitations
The bGMYC analyses identified seven clades
containing species-level probabilities >0.95 (Fig. 5).
These clades corresponded to major mitochondrial
clades identified within the recognized species in
addition to a lineage represented by two samples (with
identical haplotypes) from Venezuela, which was also
identified as a presumed species. Dated to ∼1 Ma, the
most recent node representing a speciation event was
the node separating Albiventris 3 (Central America)
from the Venezuelan lineage.
DISCUSSION
FIGURE 4. Bayesian divergence time estimations of Noctilionidae
for mtDNA (Cyt b and COI; top), ZFX (middle), and ZFY (bottom).
MtDNA chronogram clades with high posterior probabilities are
collapsed and represented as solid triangles. Node values are mean
tmrca . Stars demark nodes with high posterior probabilities (> 0.95).
Gray bars across nodes represent 95% HPD intervals for divergence
estimates. Mio- = Miocene. Outgroups are not shown to conserve
space.
[14:29 4/12/2013 Sysbio-syt062.tex]
Recent evolutionary radiations are difficult to resolve
phylogenetically and frequently exhibit signatures
interpretable as effects of the lineage sorting process.
Implicating lineage sorting is often sufficient to explain
patterns derived from single gene investigations (e.g.,
mitochondrial), which have been the foundation for
much of the DNA sequence-based descriptions for parts
of the tree-of-life. However, species trees can contain a
history of hybridization among lineages. Characterizing
these events informs temporal components of the
process of speciation. Comparative phylogenetics of
the genetic transmission elements indicates timeframes
of divergence and genetic exchange via hybridization
and can implicate gender-biased processes underlying
observed patterns. By whichever combination of
mechanisms genetic divergences between populations
arise, phylogenetic methods can be used to identify
instances of genome capture, ongoing gene flow, and
gender-biased processes.
Phylogenetic reconstructions of the recent radiation
of Noctilio chronicle a dynamic evolutionary history.
Signatures of reticulation, mitochondrial genome
capture, ongoing hybridization, divergence, and genderbiased dispersal all contribute to the evolutionary
Page: 104
96–110
2014
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
105
FIGURE 5.
Heat map representation of potential species limits from the bGMYC analysis. The phylogenetic tree is the maximum clade
credibility tree from BEAST analysis and the lightly shaded clades represent the Maximum-Likelihood species limits (Fig. 1). The table is a
sequence-by-sequence matrix in which cells are shaded by the posterior probability that corresponding sequences are conspecific, allowing for
visualization of uncertainty in species limits. Outgroups are not shown to conserve space.
history of extant Noctilio lineages. Such occurrences were
documented across species and among morphotypes
that conform to monophyletic mtDNA lineages. Geneto-gene comparisons are important for understanding
the extent to which genomic components have sorted or
the extent to which genetic exchange still occurs across
genetic elements. These comparisons provide resolution
at the level of incipient speciation and describe ongoing
gene flow among lineages.
Species Divergence and Mitochondrial Capture
Recent systematic studies of the genus have been
unable to recover monophyletic relationships of
[14:29 4/12/2013 Sysbio-syt062.tex]
these ecologically and morphologically distinct species
(Lewis-Oritt et al. 2001; Pavan et al. 2013). The
results of the current study indicate that difficulty
in understanding basal Noctilio relationships has been
a consequence of lack of suitable outgroups taxa
and a likely instance of ancient mtDNA capture
between species. The genus Noctilio belongs to a
monotypic family separated from the most recent
common ancestor, Furipteridae, by ∼33 myr (Teeling
et al. 2005; Meredith et al. 2011). Increasing evolutionary
time separating ingroup from outgroup can lead
to the well-documented phenomenon of long-branch
attraction in which recurrent substitution drives clade
misplacement among ingroup taxa (Bergsten 2005
and references therein). Current data indicated this
Page: 105
96–110
106
SYSTEMATIC BIOLOGY
phenomenon through discrepancies in mtDNA clade
relationships inferred using outgroups versus Bayesian
strict clock analysis. The use of outgroups in our analysis
often resulted in an unresolved polytomy at the base
of Noctilio, through uncertainty in placement of the
Albiventris 1 clade. Bayesian strict clock analysis placed
this clade sister to N. leporinus and separate from the
remaining N. albiventris with high statistical support.
Outgroup rooted reconstructions of sex chromosome
data sets also yielded poor basal resolution. Bayesian
strict clock rooting analyses of sex chromosomes resulted
in low resolution for the monophyly of N. albiventris,
but the placement of Albiventris 1 within N. albiventris
was strongly supported. AFLP analysis recovered
monophyly of N. albiventris. Thus, comparison of
phylogenies based on all genetic transmission elements
provides strong evidence for mitochondrial genome
capture as opposed to incomplete lineage sorting.
Specifically, the genome capture event as depicted
in the current data involved the acquisition of an
early diverging N. leporinus mtDNA genome lineage
by Albiventris 1. Although evidence for ongoing
hybridization among subspecies was also documented
in this study, this instance of mitochondrial capture is
unique in that it represents relatively ancient reticulation
in which the mitochondrial genome lineage has been
retained in only the capturing species. Given the basal
relationship of the captured genome lineage relative to
N. leporinus, and the tmrca of ∼2 Ma between these clades,
we hypothesize that this mitochondrial capture is a likely
remnant signature of ancient hybridization during early
Noctilio species formation.
Subspecific Distributions, Hybridization,
and Geographic Zones
The geographic distributions of morphologically
recognized subspecies and those of mtDNA clades
were found to be generally geographically coincident.
Congruence of independent marker types, such as
genetics and morphology, has been proposed as support
for subspecies status (Avise and Ball 1990). Congruence
of marker types is also an indicator of species status,
and distinguishing between subspecific and specific
taxonomic rank often relies on prescribed thresholds. In
this study stochastic coalescent modeling of the mtDNA
data set with a 50% confidence threshold, defined
seven presumed conspecific lineages. Perspectives from
the other genetic transmission elements provided a
powerful assessment of the genetic complexity of these
lineages as incipient species (Figs. 2 and 3).
Within currently recognized N. albiventris, the
recovered phylogenetic patterns are generally congruent
with mtDNA divergences, yet also document putative
hybridization. Specifically, Albiventris 1 corresponds to
the geographic range of morphologically recognized
N. a. affinis (Amazonia, northern Venezuela, and the
coastal Guianas), Albiventris 2 to N. a. albiventris
(southeastern Venezuela and southern Guyana to
[14:29 4/12/2013 Sysbio-syt062.tex]
VOL. 63
eastern Brazil), Albiventris 3 to N. a. minor (southern
Mexico to northwestern Venezuela: west of the
Cordillera Oriental), and Albiventris 4 to N. a. cabrerai
(from southwestern Brazil, Paraguay, and northern
Argentina; Fig. 1; Davis 1976; Gardner 2008). Comparing
the geographic distributions indicated by morphology
and mtDNA clades identified putative areas of sympatry.
Specifically, Albiventris 4 samples were obtained within
the southern distribution (Bolivia) of Albiventris 1.
Although lack of extensive genetic sampling in this
area precludes a detailed description of gene flow in
this contact zone, one of these samples was identified
through AFLP analysis as a hybrid individual. In
comparison, eastern Venezuela and Guyana were found
to be a contact zone for Albiventris 1, 2, and 3. The
occurrence of the single specimen of Albiventris 3
in eastern Venezuela indicated a range extension for
this subspecies (this individual was placed within
Albiventris 3 clades for all marker types indicating
lineage sorting or hybridization did not produce this
pattern). Comparing phylogenies for ZFX, ZFY, mtDNA,
and AFLP collectively support a complex evolutionary
history for N. albiventris. Of note is the observation
that supported clade relationships varied among sex
chromosome and mtDNA phylogenies. The AFLP
phylogeny revealed Albiventris 1, 2, and 4 distributed
and co-associated among various short branches of
the phylogeny. Individuals from these clades were also
frequently classified as hybrid individuals. All data
combined provided good support for ongoing gene
flow between N. albiventris subspecies. The pattern of
incongruency between the mtDNA and sex chromosome
phylogenies indicated a male-biased component to this
hybridization. Geographic areas of hybridization were
locations in which sympatry among hybridizing lineages
is supported (Bolivia and Guiana Shield). In addition,
the Guiana Shield is a known South American dispersal
corridor, represents a broad environmental ecotone from
neighboring regions (e.g., Fouquet et al. 2012), and
has been previously proposed as a contact zone for
N. albiventris subspecies (Davis 1976).
Within N. leporinus the clade designated Leporinus
1 corresponded to the proposed geographic range of
N. l. mastivus (Mexico and the southernmost Bahamas
south to western Ecuador and Venezuela). However,
none of our samples from eastern Venezuela and
Lesser Antilles were recovered within the Leporinus 1
clade as proposed for N. l. mastivus. Rather, 1 sample
from Central America (Canal Zone, Panama), and
12 samples from Lesser Antilles (Dominica, Grenada,
Grenadines, Montserrat) grouped inside Leporinus 2
mtDNA clade, conflicting with proposed subspecies
distributions based on morphology. In addition, the
geographic distribution of Leporinus 2 was inclusive of
the combined geographic ranges of N. l. leporinus and
N. l. rufescens, and no genetic data type presented in this
study was able to discriminate between these proposed
subspecies. The geographic area of hybridization occurs
within the Guiana Shield, again highlighting the
biological importance of this area of South America.
Page: 106
96–110
2014
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
Similar to N. albiventris, patterns of incongruency
between the mtDNA and sex chromosome phylogenies
indicated a male-biased component to hybridization. X-,
Y-, and AFLP data all place two Leporinus 2 individuals
collected in eastern Venezuela within the Leporinus 1
clade. These findings might indicate the presence of
Leporinus 1 in Guiana Shield, although not sampled
in this study. Evidence of hybridization in Guiana
Shield is more robust in the AFLP data, in which the
above mentioned individuals occupy an intermediate
position in the phylogeny. Another explanation, not
mutually exclusive with the first, is that the distribution
of Leporinus 2 is progressing northward, hybridizing
with and displacing the Leporinus 1 lineage. The
observed direction of hybridization and the occurrence
of Leporinus 2 in the Lesser Antilles and in Panama are
compatible with this hypothesis. Samples in addition
to these three individuals also collected in the Guiana
Shield were frequently identified as hybrid individuals
through AFLP analysis, further suggesting extensive
hybridization within this region. The observation that
a few samples from distant locations (i.e., Honduras,
Paraguay), or those with relatively divergent genotype
frequencies, were identified as potential hybrids does
not preclude the identification of hybridization in
Guiana Shield, but rather presents the possibility of
incorrect assignment of individuals with relatively
unique genotypes.
Diversification Timescale and Implications
Divergence estimates following the fossil and
molecular priors imposed in this study place the tmrca
of extant Noctilio to ∼3 Ma. This estimate sheds light
on several aspects of the noctilionid diversification.
The oldest reported fossil for Noctilio was recovered
from La Venta, Colombia, dated to 12.5 Ma, which
was morphologically similar to extant N. albiventris
(Czaplewski 1996; Czaplewski et al. 2003). Another fossil
of a hypothesized extinct species, N. lacrimaelunaris,
was recovered from the Villavieja Formation and was
dated to 5–9 Ma (Czaplewski 1997). The molecular
divergence time estimated in this study supports the
age of extant Noctilio is considerably younger than
lineages recovered in the fossil record. Furthermore, the
estimated divergence of extant Noctilio is placed during
and/or after the uplift of the Isthmus of Panama. As
the fossils were recovered in the northern Amazonian
Basin, these data support a South American origin
of Noctilio and a subsequent colonization of Central
America, if not through colonization of the Caribbean
Islands (see below).
(Iturralde-Vinent and MacPhee, 1999) hypothesized
that the Caribbean colonization occurred during the
existence of an ancient land bridge during the Oligocene.
Our data do not support this hypothesis, but rather
that of a relatively recent over-water colonization, which
has been hypothesized through multiple studies (Baker
and Genoways 1978; Koopman 1989; Morgan 2001).
[14:29 4/12/2013 Sysbio-syt062.tex]
107
Considering the fact that closely related Leporinus 1
haplotypes were observed in Jamaica and Central
America, invasion of the Greater Antilles likely occurred
through Central America. Similarly for Leporinus 2,
minimally diverged haplotypes were observed between
South America and the Lesser Antilles, suggesting
multiple and recent colonization routes to the Caribbean
Islands. Additionally, the observation that haplotype
diversity was lower in the Caribbean Islands as
compared with mainland populations suggests the
mainland as the source of the insular invasion (Larsen
et al. 2007), rather than Caribbean populations as the
source of Central and South American populations.
Interpretation from the current data, which includes
Central American sampling, is likely more robust
than that of (Pavan et al., 2013), who postulated the
Caribbean as the location of evolution of piscivory
(and N. leporinus). This interpretation was based on the
occurrence of haplotypes in the Caribbean which are also
found on the mainland, presence of N. leporinus fossils in
Cuba and Puerto Rico, and the absence of N. albiventris
across the Antilles.
The estimated divergence time of between 1 and 3
Ma for N. leporinus supports the findings of (LewisOritt et al., 2001) who postulated that piscivory in
N. leporinus represents a recently derived morphological,
ecological, and behavioral state for the genus. Based
on the size of Noctilio fossils, which corresponds
to that of N. albiventris, the ancestral condition was
likely insectivorous (Czaplewski 1996; Czaplewski 1997;
Czaplewski et al. 2003). The most parsimonious
explanation is that the origin of piscivory arose
subsequent to the radiation of extant Noctilio. Divergence
estimates for the different genetic elements place the
tmrca of N. leporinus to ∼1 Ma. Uncertainty about the
dietary condition existing on the branch leading to
extant N. leporinus does not permit an exact timepoint for the origin of piscivory. Regardless of the
exact timing of this dietary evolution, such a rapid
rate of change has not been documented in even
the most ecologically diversified family of bats, the
Phyllostomidae, which have rapidly radiated and exhibit
a variety of dietary strategies (e.g., Datzmann et al. 2010;
Rojas et al. 2011; Baker et al. 2012). This evolutionary
achievement may have been possible because the life
history characteristic of gleaning was already in place,
and extensive modifications to morphology (shape) or
physiological features (echolocation) were not required
to transition into piscivory (Liem 1973).
Taxonomy
Because species are interbreeding lineages that
should be evolutionarily isolated from all other such
lineages, the rank of species often is recognized
as the most biologically meaningful designation.
Species can be morphologically, ecologically, or
genetically unique, or any combination of these
biological characteristics. Phylogenetic patterns
from the different genetic elements are capable of
Page: 107
96–110
108
clearly identifying divergent lineages that continue to
exchange genetic material. Should clades for which
mtDNA and morphology are generally congruent
be recognized as species although there is clear
evidence for hybridization between their nuclear
genomes? The observation that well over 20 species
concepts exist (Mayden 1997), and that these concepts
are themselves classified into conceptual categories
(Wilkins 2011), illustrates the difficulty of defining
species. Supplementary Table S3 provides a summary
of the various characters (morphology, cytogenetics,
mitochondrial, Y-chromosome, X-chromosome, and
genotypes), in addition to ecological differences
that could be used to make decisions about species
identification following many of the proposed species
concepts. For currently recognized species of Noctilio, the
striking observation is the similarity among subspecies
for characteristics most focused on in many species
concepts. Subspecies are shown to be genetically and
reproductively compatible, exhibit little-to-no genetic
divergence at autosomal and sex chromosome loci and
no ecological differences have been identified. Table
3 also supports that the subsumed taxon N. a. affinis
(Simmons and Voss 1998; Simmons 2005; Gardner 2008)
is at least as distinct as other recognized subspecies.
The intermediate morphology observed for specimens
from the Guiana Shield (Simmons and Voss 1998), an
established geographic region of hybridization between
morphotypes, is in fact an expected characteristic of
subspecies. For these reasons using all available data,
although polyphyletic, the species-level taxonomy for
Noctilio as currently defined is supported. The study
of Noctilio clearly illustrates the complexity of defining
species in the tree-of-life with any simple criterion.
SUPPLEMENTARY MATERIAL
Supplementary material, including data files and/or
other supplementary information related to this paper
have been deposited at Dryad (http://datadryad.org/)
under doi:_10.5061/dryad.77c1h. Matrices and trees
generated/inferred during phylogenetic analyses can
be retrieved at TreeBASE: http://purl.org/phylo/
treebase/phylows/study/TB2:S13946.
FUNDING
This work was funded by the Texas Biological
Database Program. FAAK was supported through the
length of this study by the Universiti Malaysia Sarawak,
Higher Education Ministry of Malaysia and Department
of Biology, Texas Tech University.
ACKNOWLEDGMENTS
We thank the Natural Science Research Laboratory
(NSRL, Texas Tech University), Carnegie Museum of
Natural History, Royal Ontario Museum, Museum of
Vertebrate Zoology (University of California, Berkeley),
[14:29 4/12/2013 Sysbio-syt062.tex]
VOL. 63
SYSTEMATIC BIOLOGY
National Museum of Natural History (Smithsonian),
Louisiana Museum of Natural History (Louisiana State
University), American Museum of Natural History,
Museum of Southwestern Biology (University of New
Mexico), Oklahoma Collection of Genomic Resources
(The University of Oklahoma), and Universidad
Nacional Mayor de San Marcos. We thank all personnel
that were involved in the collection of materials
deposited at these museums. We thank Miguel Pinto
for requesting tissues and laboratory assistance. Tissue
samples and voucher specimen loans were prepared
by H. J. Garner and K. MacDonald at the NSRL of
the Museum of Texas Tech University. We are grateful
to L. Bradley, R. D. Bradley, J. Sullivan, N. Reid, and
an anonymous reviewer for valuable comments which
greatly improved this manuscript.
REFERENCES
Anderson E.C., Thompson E.A. 2002. A model-based method for
identifying species hybrids using multilocus genetic data. Genetics
160:1217–1229.
Avise J.C., Ball R.M. Jr. 1990. Principles of genealogical concordance
in species concepts and biological taxonomy. Oxf. Surv. Evol. Biol.
7:45–67.
Baker R.J., Bininda-Emonds O.R.P., Mantilla-Meluk H., Porter C.A.,
Van Den Bussche R.A. 2012. Molecular timescale of diversification
of feeding strategy and morphology in New World leaf-nosed
bats (Phyllostomidae): A phylogenetic perspective. In: Gunnell
G.F., Simmons N.B., editors. Evolutionary history of bats: Fossils,
molecules and morphology. UK: Cambridge University Press, In the
CUP series Cambridge Studies in Morphology : New Paradigms in
Evolutionary Biology (No. 2). p. 385–409.
Baker R.J., Bradley R.D. 2006. Speciation in mammals and the genetic
species concept. J. Mammal. 87:643–662.
Baker R.J., Genoways H.H. 1978. Zoogeography of Antillean bats. Spec.
Publ. Acad. Nat. Sci. Philadelphia 13:53–97.
Bergsten J. 2005. A review of long-branch attraction. Cladistics
2:163–193.
Bickham J.W., Patton J.C., Schlitter D.A., Rautenbach I.L., Honeycutt
R.L. 2004. Molecular phylogenetics, karyotypic diversity, and
partition of the genus Myotis (Chiroptera: Vespertilionidae). Mol.
Phylogenet. Evol. 33:333–338.
Bonin A., Bellemain E., Eidesen P.B., Pompanon F., Brochmann C.,
Taberlet P. 2004. How to track and assess genotyping errors in
population genetics studies. Mol. Ecol. 13:3261–3273.
Bordignon M.O. 2006. Diet of the fishing bat Noctilio leporinus
(Linnaeus) (Mammalia, Chiroptera) in a mangrove area of southern
Brazil. Rev. Bras. Biol. Zool. 23:256–260.
Bossu C.M., Near T.J. 2009. Gene trees reveal repeated instances of
introgression in orange throat darters (Percidae: Etheostoma). Syst.
Biol. 58:114–129.
Buckley T.R., Cordeiro M., Marshall D.C., Simon C. 2006.
Differentiating between hypotheses of lineage sorting and
introgression in New Zealand alpine cicadas (Maoricicada Dugdale).
Syst. Biol. 55:411–425.
Cathey J.C, Bickham J.W., Patton J.C. 1998. Introgressive hybridization
and nonconcordant evolutionary history of maternal and paternal
lineages in North American deer. Evolution 52:1224–1229.
Clare E.L. 2011. Cryptic Species? Patterns of maternal and paternal gene
flow in eight Neotropical bats. PLoS ONE 6:e21460.
Clare E.L., Lim B.K., Engstrom M.D, Eger J.L., Hebert P.D.N. 2007. DNA
barcoding of Neotropical bats: species identification and discovery
within Guyana. Mol. Ecol. Notes 7:184–190.
Czaplewski N.J. 1996. Opossums (Didelphidae) and bats
(Noctilionidae and Molossidae) from the late Miocene of the
Amazon Basin. J. Mammal. 77:84–94.
Page: 108
96–110
2014
KHAN ET AL.—SPECIATION IN THE GENUS NOCTILIO
Czaplewski N.J. 1997. Chiroptera. In: Kay R.F., Madden R.H., Cifelli
R.L., Flynn J.J., editors. Vertebrate paleontology in the neotropics:
The Miocene Fauna of La Venta, Colombia. Washington (DC):
Smithsonian Institution Press. p. 410–431.
Czaplewski N.J., Masanaru T., Naeher T.M., Shigehara N., Setoguchi T.
2003. Additional bats from the middle Miocene La Venta fauna of
Colombia. Rev. Acad. Colomb. Cienc. Exactas Fís. Nat. 27:263–282.
Davis W.B. 1973. Geographic variation in the fishing bat, Noctilio
leporinus. J. Mammal. 54:862–874.
Davis W.B. 1976. Geographic variation in the lesser Noctilio, Noctilio
albiventris. J. Mammal. 57:687–707.
Datzmann T., von Helversen O., Mayer F. 2010. Evolution of nectarivory
in phyllostomid bats (Phyllostomidae Gray, 1825, Chiroptera:
Mammalia). BMC Evol. Biol. 10:165.
Degnan J.H., Rosenberg N.A. 2006. Discordance of species trees with
their most likely gene trees. PLoS Gen. 2:e65.
Drummond A.J., Rambaut A. 2007. BEAST: Bayesian evolutionary
analysis by sampling trees. BMC Evol. Biol. 7:214.
Felsenstein J. 1988. Phylogenies from molecular sequences: Inference
and reliability. Ann. Rev. Genet. 22:521–565.
Fouquet A., Noonan B.P., Rodrigues M.T., Pech N., Gilles A., Gemmell
N.J. 2012. Multiple quaternary refugia in the eastern Guiana shield
revealed by comparative phylogeography of 12 frog species. Syst.
Biol. 61:1–29.
Francis C.M., Borisenko A.V., Ivanova N.V., Eger J.L., Lim B.K. et al.
2010. The role of DNA barcodes in understanding and conservation
of mammal diversity in Southeast Asia. PLoS ONE 5:e12575.
Gardner A.L. 2008. Genus Noctilionidae Gray, 1821. In: Gardner A.L.,
editor. Mammals of South America, Vol. 1: Marsupials, Xenarthrans,
Shrews, and Bats. Chicago (IL): University of Chicago Press.
p. 384–389.
Gascuel O. 1997. BIONJ: an improved version of the NJ algorithm based
on a simple model of sequence data. Mol. Biol. Evol. 14:685–695.
Gernhard. T. 2008. The conditioned reconstructed process. J. Theor. Biol.
253:769–778.
Gonçalves F., Munin R., Costa P., Fischer E. 2007. Feeding habits of
Noctilio albiventris (Noctilionidae) bats in the Pantanal, Brazil. Acta
Chiropterol. 9:535–538.
Good J.M., Demboski J.R., Nagorsen D.W., Sullivan J. 2003.
Phylogeography and introgressive hybridization: Chipmunks
(genus: Tamias) in the northern Rocky Mountains. Evolution
57:1900–1916.
Good J.M., Hird S., Reid N., Demboski J.R., Steppan S.J.,
Martin-Nims T.R., Sullivan J. 2008. Ancient hybridization and
mitochondrial capture between two species of chipmunks. Mol. Ecol.
17:1313–1327.
Guindon S., Gascuel O. 2003. A simple, fast, and accurate algorithm
to estimate large phylogenies by maximum likelihood. Syst. Biol.
52:696–704.
Hoffmann F.G., Baker R.J. 2001. Systematics of bats of the genus
Glossophaga (Chiroptera: Phyllostomidae) and phylogeography
in G. soricina based on the cytochrome-b gene. J. Mammal.
82:1092–1101.
Holland B.R., Penny D., Hendy M.D. 2003. Outgroup misplacement
and phylogenetic inaccuracy under a molecular clock—a simulation
study. Syst. Biol. 52:229–238.
Hood C.S., Jones J.K. 1984. Noctilio leporinus. Mamm. Species 216:1–7.
Hood C.S., Pitocchelli J. 1983. Noctilio albiventris. Mamm. Species
197:1–5.
Howell D.J., Burch D. 1974. Food habits of some Costa Rican bats. Rev.
Biol. Trop. 21:281–294.
Huelsenbeck J.P., Bollback J.P., Levine A.M. 2002. Inferring the root of
a phylogenetic tree. Syst. Biol. 51:32–43.
Huelsenbeck J.P., Ronquist F. 2001. MrBayes: Bayesian inference of
phylogeny. Bioinformatics 17:754–755.
Iturralde-Vinent M.A., MacPhee R.D.E. 1999. Paleogeography of the
Caribbean region: implications for Cenozoic biogeography. Bull.
Am. Mus. Nat. Hist. 238:1–95.
Ivanova N.V., deWaard J.R., Hebert P.D.N. 2006. An inexpensive,
automation-friendly protocol for recovering high-quality DNA. Mol.
Ecol. Notes 6:998–1002.
Khan F.A.A., Solari S., Swier V.J., Larsen P.A., Abdullah M.T., Baker
R.J. 2010. Systematics of Malaysian woolly bats (Vespertilionidae:
[14:29 4/12/2013 Sysbio-syt062.tex]
109
Kerivoula) inferred from mitochondrial, nuclear, karyotypic, and
morphological datasets. J. Mammal. 19:1058–1072.
Knowles L.L. 2009. Estimating species trees: methods of phylogenetic
analysis when there is incongruence across genes. Syst. Biol.
58:463–467.
Koopman K.F. 1989. A review and analysis of the bats of the West
Indies. In: Woods C.A., editor. Biogeography of the West Indies,
Past, Present, and Future. Gainesville (FL): Sandhill Crane Press.
p. 635–644.
Larsen P.A., Hoofer S.R., Bozeman M.C., Pedersen S.C., Genoways
H.H., Phillips C.J., Pumo D.E., Baker R.J. 2007. Phylogenetics
and phylogeography of the Artibeus jamaicensis complex based on
cytochrome b DNA sequences. J. Mammal. 88:712–727.
Larsen P.A., Marcha’n-Rivadeneira M.R., Baker R.J. 2010. Natural
hybridization generates mammalian lineage with species
characteristics. Proc. Natl Acad. Sci. USA 107:11447–11452.
Lewis-Oritt N., Van Den Bussche R.A., Baker R.J. 2001. Molecular
evidence for evolution of piscivory in Noctilio (Chiroptera:
Noctilionidae). J. Mammal. 82:748–759.
Librado P., Rozas J. 2009. DnaSP v5: A software for comprehensive
analysis
of
DNA
polymorphism
data.
Bioinformatics
25:1451–1452.
Liem, K.F. 1973. Evolutionary strategies and morphological
innovations: cichlid pharyngeal jaws. Syst. Biol. 22:425–441.
Linnen C.R., Farrell B.D. 2008. Phylogenetic analysis of nuclear
and mitochondrial genes reveals evolutionary relationships and
mitochondrial introgression in the sertifer species group of the
genus Neodiprion (Hymenoptera: Diprionidae). Mol. Phylogent. Evol.
48:240–257.
Maureira-Butler I.J., Pfeil B.E., Muangprom A., Osborn T.C., Doyle
J.J. 2008. The reticulate history of medicago (Fabaceae). Syst. Biol.
57:466–482.
Mayden, R.L. 1997. A hierarchy of species concepts: the denouement in
a saga of the species problem. In: Claridge M.F., Dawah H.A., Wilson
M.R., editors. Species: the units of biodiversity. London: Chapman
and Hall Ltd. p. 381–424.
Meredith R.W., Janecka J.E., Gatesy J., Ryder O.A., Fisher C.A.,
Teeling, E.C., Goodbla A., Eizirik E., Simão T.L.L., Stadler T.,
Rabosky D.L., Honeycutt R.L., Flynn J.J., Ingram C.M., Steiner
C., Williams T.L., Robinson T.J., Burk-Herrick A., Westerman M.,
Ayoub N.A., Springer M.S., Murphy W.J. 2011. Impacts of the
Cretaceous Terrestrial Revolution and KPg extinction on mammal
diversification. Science 334:521–524.
Morgan G.S. 2001. Patterns of extinction in West Indian bats. In: Woods
C.A., Sergile F.E., editors. Biogeography of the West Indies. Boca
Raton (FL): CRC Press. p. 369–407.
Morgan G.S. 2002. New bats in the Neotropical families
Emballonuridae and Mormoopidae from the Oligocene and
Miocene of Florida, and the biochronology of Florida Whitneyan,
Arikareean, and Hemingfordian faunas. J. Vert. Paleont. 22:90A.
Pamilo P., Nei M. 1988. Relationships between gene trees and species
trees. Mol. Biol. Evol. 5:568–583.
Pavan A.C., Martins F.M., Morgante J.S. 2013. Evolutionary history of
bulldog bats (genus: Noctilio): recent diversification and the role of
the Caribbean in Neotropical biogeography. Biol. Journ. Linn. Soc.
108:210–224
Peakall R., Smouse P.E. 2006. GENALEX 6: genetic analysis in Excel.
Population genetic software for teaching and research. Mol. Ecol.
Notes 6:288–295.
Phillips C.D., Henard C.A., Pfau R.S. 2007. Amplified fragment length
polymorphism and mitochondrial DNA analyses reveal patterns
of divergence and hybridization in the hispid cotton rat (Sigmodon
hispidus). J. Mammal. 88:351–359.
Pons J., Barraclough T.G., Gomez-Zurita J., Cardoso A., Duran D.P.,
Hazell S., Kamoun S., Sumlin W.D., Vogler A. 2006. Sequence-based
species delimitation for the DNA taxonomy of undescribed insects.
Syst. Biol. 55:595–609.
Rambaut A., Drummond A.J. 2007. Tracer v.1.4. Available from: URL
http://beast.bio.ed.ac.uk/Tracer.
Reid N., Carstens B. 2012. Phylogenetic estimation error can decrease
the accuracy of species delimitation: a Bayesian implementation
of the general mixed Yule-coalescent model. BMC Evol. Biol.
12:196.
Page: 109
96–110
110
SYSTEMATIC BIOLOGY
Reid N., Demboski J.R., Sullivan J. 2012. Phylogeny estimation of the
radiation of western North American chipmunks (Tamias) in the
face of introgression using reproductive protein genes. Syst. Biol.
61:44–62.
Rieseberg L.H., Baird S.J.E., Gardner K. 2000. Hybridization,
introgression, and linkage evolution. Plant Mol. Biol. 42:205–224.
Rojas D., Vale Á., Ferrero V., Navarro L. 2011. When did plants become
important to leaf-nosed bats? Diversification of feeding habits in the
family Phyllostomidae. Mol. Ecol. 20:2217–2228.
Ronquist F., Huelsenbeck J.P., van der Mark P. 2005. MrBayes v.3.2.
Available from: URL http://mrbayes.csit.fsu.edu/index.php.
Sambrook J., Fritsch E.F., Maniatis T. 1989. Molecular cloning: A
laboratory manual. 2nd ed. Cold Spring Harbor (NY): Cold Spring
Harbor Laboratory Press.
Simmons N.B. 2005. Order Chiroptera. In: Wilson D.E., Reeder D.M.,
editors. Mammal species of the world: A taxonomic and geographic
reference. 3rd ed. Baltimore (MD): Johns Hopkins University Press.
p. 312–529.
Simmons N.B., Voss R.S. 1998. The mammals of Paracou, French
Guiana: A Neotropical lowland rainforest fauna. Part 1. Bats. Bull.
Am. Mus. Nat. Hist. 237:1–219.
Smith M.F., Patton J.L. 1993. The diversification of South American
murid rodents: evidence from mitochondrial DNA sequence data
for the Akodontine tribe. Biol. J. Linn. Soc. 50:149–177.
Swofford D.L. 2002. PAUP*: phylogenetic analysis using parsimony
(*and other methods), version 4.0 b10. Sunderland (MA): Sinauer
Associates.
[14:29 4/12/2013 Sysbio-syt062.tex]
VOL. 63
Takahata N. 1989. Gene genealogy in three related populations:
consistency probability between gene and population trees. Genetics
122:957–966.
Tamura K.P.D., Peterson N., Stecher G., Nei M., Kumar S. 2011.
MEGA5: Molecular evolutionary genetics analysis using maximum
likelihood, evolutionary distance, and maximum parsimony
methods. Mol. Biol. Evol. 28:2731–2739.
Teeling E.C., Springer M.S., Madsen O., Bates P., O’Brien S.J., Murphy
W.J. 2005. A molecular phylogeny for bats illuminates biogeography
and the fossil record. Science 307:508–584.
Timm R.M., Genoways H.H. 2003. West Indian mammals from the
Albert Schwartz collection: Biological and historical information.
Scientific Papers, Natural History Museum, The University of
Kansas. 29:1–47.
Vos P., Hogers R., Bleeker M., Reijans M., Vandelee T., Hornes
M., Frijters A., Pot J., Peleman J., Kuiper M., Zabeau M. 1995.
AFLP—A new technique for DNA-fingerprinting. Nucl. Acids Res.
23:4407–4414.
Wilkins J.S. 2011. Philosophically speaking, how many species concepts
are there? Zootaxa 2765:58–60.
Xia X., Xie Z. 2001. DAMBE: software package for data analysis in
molecular biology and evolution. J. Hered. 92:371–373.
Xia X., Xie Z., Salemi M., Chen L., Wang Y. 2003. An index of
substitution saturation and its application. Mol. Phylogenet. Evol.
26:1–7.
Yule G.U. 1925. A mathematical theory of evolution based on the
conclusions of Dr. J.C.Willis. Phil. Trans. R. Soc. London B 213:21–87.
Page: 110
96–110