A Mitochondrial Phylogeographic Scenario

bs_bs_banner
Biological Journal of the Linnean Society, 2013, 108, 901–916. With 6 figures
A mitochondrial phylogeographic scenario for the most

widespread African rodent, Mastomys natalensis
PAOLO COLANGELO1*, ERIK VERHEYEN2,3, HERWIG LEIRS3,4, CAROLINE TATARD5,
CHRISTIANE DENYS6, GAUTHIER DOBIGNY7,8, JEAN-MARC DUPLANTIER7,
CARINE BROUAT7, LAURENT GRANJON9 and EMILIE LECOMPTE10,11
1
Department of Biology and Biotechnologies, University of Rome ‘La Sapienza’, Via Borelli 50, 00161
Rome, Italy
2
Department Vertebrates, Royal Belgian Institute of Natural Sciences, Vautierstraat 29, 1000
Brussels, Belgium
3
Evolutionary Ecology Group, University of Antwerp, Groenenborgerlaan 171, B-2020 Antwerpen,
Belgium
4
Danish Pest Infestation Laboratory, University of Aarhus, Faculty of Agricultural Sciences,
Department of Integrated Pest Management, Forsøgsvej 1, DK-4200 Slagelse, Denmark
5
INRA, UMR1062 CBGP, Campus International de Baillarguet, CS 30016, 34988 Montferrier-sur-Lez
cedex, France
6
UMR-CNRS 7205, Laboratoire Origine Structure Evolution de la Biodiversité (OSEB), Department
of Systematics and Evolution, CP51, 55 rue Buffon, 75005 Paris, France
7
IRD, CBGP (UMR IRD/INRA/CIRAD/MontpellierSupAgro), Campus International de Baillarguet,
CS 30016, 34988 Montferrier-sur-Lez cedex, France
8
Centre Régional Agrhymet, BP11011, Niamey, Niger
9
IRD, CBGP (UMR IRD/INRA/CIRAD/MontpellierSupAgro), Campus de Bel-Air, BP 1386, Dakar,
CP 18524, Sénégal
10
Université Toulouse III, Paul Sabatier, CNRS, ENFA; UMR5174 EDB (Laboratoire Evolution et
Diversité Biologique), 118 route de Narbonne, F-31062 Toulouse, France
11
CNRS, Université Paul Sabatier; UMR5174 EDB, F-31062 Toulouse, France
Received 8 July 2012; revised 30 October 2012; accepted for publication 30 October 2012
In order to evaluate the contribution of geological, environmental, and climatic changes to the spatial distribution
of genetic variation of Mastomys natalensis, we analysed cytochrome b sequences from the whole distribution area
of the species to infer its phylogeographic structure and historical demography. Six well-supported phylogroups,
differentiated during the Pleistocene, were evidenced. No significant correlation between genetic and geographic
distances was found at the continental scale, and the geographic distributions of the observed phylogroups have
resulted from extensive periods of isolation caused by the presence of putative geographic and ecological barriers.
The diversification events were probably influenced by habitat contraction/expansion cycles that may have
complemented topographic barriers to induce genetic drift and lineage sorting. According to our results, we propose
a scenario where climate-driven processes may have played a primary role in the differentiation among phylo-
groups. © 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 901–916.
ADDITIONAL KEYWORDS: African savannah – climate changes – geographic barrier – isolation by distance
– vicariance.
*Corresponding author. E-mail: paolo.colangelo@uniroma1.it
© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 901–916 901
902 P. COLANGELO ET AL.
INTRODUCTION sification of the Albertine Rift endemic rodents

(Huhndorf et al., 2007). Similarly, Faulkes et al.
The currently observed spatial organization of animal (2004) suggested that physical barriers, following the
communities in Africa results from long-term proc- formation of the Kenyan and Western rifts, together
esses, where geological, environmental, and climatic with palaeoclimatic cycles and associated vegetation
changes have played major roles. Since the Miocene modifications, have driven the diversification of some
the whole of the African continent has been undergo- lineages of African mole-rats of the family Bathy-
ing deep tectonic and climatic changes (deMenocal, ergidae during the Plio-Pleistocene. Finally, in
1995; Cerling et al., 1997), spanning the geological Southern Africa, both Rambau et al. (2003) and
period of the evolution and expansion of many mam- Russo et al. (2010) invoked climatic and/or associated
malian species, when the African climate was gov- vegetation changes of the last few millions years as
erned by changes in monsoon dynamics (Trauth, an explanation for the currently observed patterns of
Larrasoan & Mudelsee, 2009), and when periods of genetic structure in Rhabdomys pumilio and Rhab-
aridity coincided with the onset and intensification domys dilectus on the one hand, and Micaelamys
of northern latitude glaciations (deMenocal, 1995, namaquensis on the other.
2004). In general, during the coldest periods sub- All these pieces of evidence strongly support the
Saharan Africa was more arid and desert areas occurrence of geographical and ecological barriers
expanded, whereas forest habitats were severely across various regions of Africa as pivotal drivers of
constricted. During wetter periods, forest habitats rodent diversification; however, the investigated taxa,
replaced many savannahs (deMenocal, 1995, 2004), although sometimes quite common, show specific dis-
resulting in gene flow restriction among isolated tribution ranges that cover only a portion of the
populations of savannah species. continent and/or a specific habitat type. As such, no
A number of recent studies based on the analysis data allow one to draw a large-scale picture of the
of mitochondrial (mt) DNA sequences of mammals forces that shaped the phylogeographic structure of
have provided evidence for putative links between African rodents.
past eco-climatic oscillations and evolutionary events The multimammate rat Mastomys natalensis occurs
that resulted in the currently observed patterns of throughout sub-Saharan Africa, making it the African
genetic structure of large (Arctander, Johansen & rodent species with the largest distribution range
Coutellec-Vreto, 1999; Flagstad et al., 2001; Girman (Musser & Carleton, 2005), commonly found across a
et al., 2001) as well as small (Rambau, Stanyon & wide ecological range, with the notable exception of
Robinson, 2003; Huhndorf, Kerbispeterhans & Loew, deserts and very high mountainous areas. Previous
2007; Mouline et al., 2008; Nicolas et al., 2008a, b, phylogenetic analyses suggested that M. natalensis
2009; Brouat et al., 2009; Bryja et al., 2010; Russo, has recently differentiated from other congeneric lin-
Chimimba & Bloomer, 2010) species. In rodents, this eages following the expansion of sub-Saharan savan-
genetic structuring has been explained by various nah habitats during the Plio-Pleistocene (Lecompte
non-exclusive causal factors, such as physical barri- et al., 2002; Lecompte, Denys & Granjon, 2005). Con-
ers to dispersal, ecological constraints, and/or past siderable work has been conducted on various aspects
isolation in refugia (Rambau et al., 2003; Faulkes of the biology of M. natalensis (Leirs et al., 1997;
et al., 2004; Huhndorf et al., 2007; Mouline et al., Julliard et al., 1999; Vibe-Petersen, Leirs & De
2008; Nicolas et al., 2008a, b, 2009; Brouat et al., Bruyn, 2006; Sluydts et al., 2007); however, no com-
2009; Bryja et al., 2010; Russo et al., 2010; Castiglia prehensive study of the genetic structure of this
et al., 2012). Climatic and hydrogeological changes species at the scale of its entire range has been ever
are often invoked as the driving processes of these conducted. Here, we gathered an important and well-
phylogeographical patterns. In West Africa, for distributed number of individual samples over most of
instance, periods of intense aridity were associated the species distribution. This allowed us to draw a
with the differentiation of major genetic clades in picture of its phylogeography over almost all of its
the rocky habitat specialist Acomys chudeaui distribution range, namely most of sub-Saharan
(Nicolas et al., 2009). The western branch of the East Africa.
African Rift, the Volta, and the Niger River may We investigate the contribution of different diversi-
have contributed to the emergence of major clades in fication processes, i.e. dispersal and vicariance, to the
several species (e.g. Mouline et al., 2008; Nicolas shaping of the current phylogeographic pattern
et al., 2008b; Brouat et al., 2009). In Eastern Africa, observed in M. natalensis. We expect to find a cline of
the climatic fluctuations and volcanic activity of the continuous genetic variation, as predicted by an isola-
Virunga Mountains, a chain of volcanoes that caused tion by distance model, if dispersal had a primary role
the fragmentation of rainforest habitat in the Pleis- in the genetic diversification of this species. By con-
tocene, were supposed to have promoted the diver- trast, the presence of genetically divergent clades will
© 2013 The Linnean Society of London, Biological Journal of the Linnean Society, 2013, 108, 901–916
PHYLOGEOGRAPHY OF MASTOMYS NATALENSIS 903
be congruent with a scenario invoking allopatric frag- The cytochrome b sequences of 24 specimens were
mentation and vicariance. In particular we focused on downloaded from GenBank (Appendix S1).
the role, already suggested for other African rodents, of
the ecological shift in triggering the isolation of M. na-
talensis in patches of suitable habitat during different LABORATORY PROCEDURES
phases of Pleistocene climatic oscillation. DNA was isolated from ethanol-preserved tissues
using the CTAB protocol (Winnepenninckx, Backeljau
& De Wachter, 1993) or a QIAamp extraction kit
MATERIAL AND METHODS
(Qiagen). The mitochondrial cytochrome b gene (cyt b)
SAMPLING was amplified as described in Lecompte et al. (2002).
Mastomys natalensis individuals were sampled across Double-stranded PCR products were purified directly
a very large part of the distribution range of the from the PCR product or from agarose gel using the
species, i.e. from West Africa to East and Southern MinElute purification kit (Qiagen), and were
Africa. In total, 209 specimens from 121 different sequenced directly along both strands using the same
localities were used (Appendix S1; Fig. 1). Animals primers used for amplification.
were caught using live traps and subsequently tissue The obtained sequences (1115 base pairs long) were
samples were collected and preserved in 95% ethyl aligned using BioEdit 7.0.9 (Hall, 1999). Sequence
alcohol. Tissue samples are stored at the Museum of alignment was inspected by eye and no indel or stop
Comparative Anatomy of the University of Rome ‘La codon was detected. The sequences were deposited in
Sapienza’, at the Royal Belgian Museum of Natural GenBank under the accession numbers HE863974–
Science, and at the Museum of Natural History, Paris. HE864158.
Figure 1. Map of Africa showing Mastomys natalensis sampling localities. The major phylogroups identified by
maximum-likelihood, Bayesian, and Neighbour-Net analyses were reported.
PHYLOGENETIC ANALYSES AND DIVERGENCE was dated to 3.7 Mya (Denys, 1987). The same species
TIME ESTIMATIONS was found in an upper level of the same site dated to
2.5 Mya (Denys, 1987). Thus, we used the interval
Unique haplotypes were determined using DNAsp 5.0 3.7–2.5 Mya to calibrate the split between all the
(Librado & Rozas, 2009). The most suitable model of Mastomys species, except for the most basal species of
nucleotidic substitution for our data set was deter- the genus, M. kollmanspergeri. To take into account
mined using the Akaike information criterion (AIC) the uncertainty of the fossil record, we used a lognor-
implemented in ModelGenerator 0.85 (Keane et al., mal distribution with an offset at 2.5 Mya, a median
2006). All the phylogenetic reconstructions were per- at 3.5 Mya, and a 95% confidence interval ranging
formed assuming a Tamura–Nei model (Tamura & from 2.6 to 9.6 Mya.
Nei, 1993) with gamma-distributed rate variation Subsequently, we calibrated the split between the
across sites (G) and a proportion of invariable sites (I) genera Praomys and Mastomys on the basis of the
according to the AIC. earliest Praomys fossil found in Lissasfa, Morocco
First, phylogenetic relationships were inferred by (Geraads, 2002), dated back to 5.5–6.0 Mya. We used
maximum likelihood (ML) using PhyML 3.0 (Guindon an exponential prior for the split between the two
et al., 2010). Nodal support was assessed by bootstrap genera, considering that it occurred at least 5.5 Mya
resampling (BP; 1000 pseudoreplicates). The other (used as the offset value), had a median date of
species of the genus, Mastomys coucha, Mastomys 6.2 Mya, and a 95% confidence interval ranging
erythroleucus, Mastomys huberti, Mastomys koll- from 5.5 to 9.2 Mya. A uniform prior, based on litera-
manspergeri, Mastomys awashensis, as well as four ture (e.g. Faulkes et al., 2011; Horn et al., 2011), for
Praomys species, Praomys jacksoni, Praomys tull- the ucld.mean with minimum = 0.001 and maxi-
bergi, Praomys misonnei, and Praomys rostratus mum = 0.1 was set. The tree prior choice can signifi-
(accession numbers AF518335, AF518338, AF141220, cantly affect the estimation of divergence time along
FN393051, AF518344, AF518361, AF518362, the whole tree. Because our alignment contains
JQ735886, and JQ735866) were used as out-groups. a combination of interspecific and intraspecific
Second, phylogenetic relationships among all sequences, we implemented a Yule speciation prior
haplotypes and divergence time among phylogroups across the tree but with a separate coalescent expo-
were estimated using BEAST 1.7.4 (Drummond & nential prior for M. natalensis. Subsequently, two
Rambaut, 2007). This software relies on a Bayesian independent searches of 30 million generations were
Markov chain Monte Carlo (MCMC) method to infer run using the same model as for phylogenetic analy-
a relaxed phylogenetic topology that allow us to ses. Topologies were sampled every 1000 generations,
co-estimate phylogeny and divergence times (Drum- and the convergence between runs and the effective
mond et al., 2006). sample size (ESS) for each parameter was checked
The validity of a strict molecular clock against an using the program TRACER. Of the final 60 000 trees
uncorrelated lognormal relaxed-clock model (Drum- retained, the first 3000 of each independent run were
mond et al., 2006) was tested using Bayes factors discarded as burn-in, and the remaining trees were
(BFs). Bayes factors calculate the ratio of marginal used to obtain a maximum credibility tree. For each of
likelihoods between two given models. Considering the major splits recorded within M. natalensis the
twice the natural logarithm of the BF, a result > 10 mean value and the 95% highest posterior density
indicates that one model is more strongly supported interval (HPD) were reported.
by the data under consideration than another (Kass &
Raftery, 1995). We performed two preliminary runs of
10 million generations each for the two alternative NETWORK ANALYSIS AND GENETIC DIVERGENCE
molecular clock hypotheses, and calculated the ratio To further establish the relationships among haplo-
of marginal likelihoods between two given models types we used the program SPLITSTREE 4.10
using TRACER (http://tree.bio.ed.ac.uk/software/ (Huson & Bryant, 2006) to construct a network with
tracer/). The results rejected the existence of a strict Tamura–Nei distances, previously calculated using
molecular clock in favour of the uncorrelated lognor- PAUP* 4.0b10 (Swofford, 2003) and the Neighbor-Net
mal relaxed-clock model (logBF = 12.908). Conse- (NN) algorithm (Bryant & Moulton, 2004). The NN
quently, we estimated divergence time assuming a computes splits graphs based on the distance matrix,
relaxed molecular clock. producing a phylogenetic network that allows the
The fossil record of the genus Mastomys is rather visualization of conflict and ambiguous signals in the
scarce: the earliest record of Mastomys cinereus, a data set. The genetic divergences between and within
species that can be considered close to the ancestor of each clade were evaluated under the Tamura–Nei
M. huberti, M. coucha, and M. natalensis (Denys, model of nucleotide substitution using MEGA 4
pers. comm.), was found at Laetoli (Tanzania) and (Tamura et al., 2007).
DIVERSITY INDEXES, NEUTRALITY TEST, AND using a permutation procedure with 10 000 replica-
HISTORICAL DEMOGRAPHY tions. In the stratified Mantel test, the result of
Haplotype (Hd) and nucleotide (p) diversities were phylogenetic analyses was used to define the strata
estimated for each phylogroup using DNASP 5.0 (corresponding to different clades in our case).
(Librado & Rozas, 2009).
Demographic histories of the different M. natalen- RESULTS
sis phylogroups identified by previous analyses were
first inferred using a pairwise mismatch distribution SEQUENCE VARIATION
analysis (Rogers & Harpending, 1992), using the In total, 158 haplotypes were identified from
function MMD of the R package PEGAS (Paradis, 209 M. natalensis sequences. The whole data set
2010; R Core Team, 2012). Demographic stability displayed 303 variable (27.2%) and 182 (16.3%)
is expected to result in multimodal distribution, parsimony-informative sites, respectively. The base
whereas past sudden expansion would generate a composition (T 28.2%, C 26.9%, A 31.8%, and G
unimodal pattern (Slatkin & Hudson, 1991). Bayesian 12.5%) is in very good agreement with other data on
skyline plots (BSPs; Drummond et al., 2005) were also the rodent cyt b gene, including those obtained in the
drawn for the whole data set as well as for each Mastomys genus (e.g. Ducroz, Volobouev & Granjon,
phylogroup using BEAST 1.7.4. This coalescence- 1998; Mouline et al., 2008; Brouat et al., 2009).
based method aims to infer past population dynamics
without an ad hoc model of demographic history
PHYLOGENETIC ANALYSES AND DIVERGENCE TIMES
(Drummond et al., 2005). Here, it was performed
using an MCMC of 30 million generations, with a The M. natalensis haplotypes are distributed in six
sampling frequency of 1000, under the assumption of major lineages (Figs 2–4), the monophyly of which is
a strict clock as well as the same model of nucleotide well supported by bootstrap (BS) values (Fig. 2) and
evolution (see above). We checked for the effective posterior probabilities (PPs; Fig. 3), and each of these
sample size (ESS) for each parameter estimated, as phylogroups corresponds to an exclusive geographical
well as for the number of initial trees to discard using area (Fig. 1). The NN phylogenetic network (Fig. 4)
TRACER 1.5. Of the 30 000 recorded topologies, the shows the very marked phylogeographic structure of
first 10% were discarded and the remaining topolo- M. natalensis, reflecting, and hence confirming, the
gies were used to obtain the skyline plot. presence of six well-differentiated phylogroups.
Finally, two neutrality tests, Fu’s Fs (Fu, 1997) and The mean sequence divergence between the six
Ramos-Onsins and Rozas’ R2 (Ramos-Onsins & Rozas, phylogroups (hereafter labelled A-I, A-II, A-III, B-IV,
2002), were performed using DNASP 5.0. These tests B-V, and B-VI) ranges from 2.1 to 3.8%, whereas the
have been shown to be good indicators for detecting within-clade genetic divergence (within phylogroups)
population growth. The behaviour of R2 is better for ranges between 0.42 and 0.91%.
small sample sizes, whereas Fu’s Fs is better for Two main clades (hereafter designated A and B) can
bigger sample sizes (Ramos-Onsins & Rozas, 2002). be distinguished by phylogenetic analyses, each of
which includes three of the six phylogroups (Figs 2,
3). From a geographical point of view, clade A roughly
ISOLATION BY DISTANCE corresponds to West and Central Africa, with the
Under a model of isolation by distance (IBD), the inclusion of additional specimens from the Victoria
genetic distance between populations is expected to Lake region on the west side of the Rift Valley
increase with geographical distance. Conventionally, (Fig. 1). Clade B corresponds to the Eastern and
the most commonly used test for IBD is the Mantel Southern African regions (Fig. 1). However, the mono-
test. However, Meirmans (2012) suggested that phylies of clades A and B are not strongly supported
Mantel tests could be biased by hierarchical structure by phylogenetic trees (Figs 2, 3), and the presence of
in the populations under study, as it is often the case a basal reticulation is suggested by the NN network
for populations having undergone isolation in differ- (Fig. 4).
ent refugia. A stratified Mantel test is less sensitive to Within clade A, the phylogroup A-I includes speci-
this kind of bias (Meirmans, 2012). Another possibil- mens from Senegal, Guinea, Ivory Coast, Burkina
ity is to use a simple Mantel test in each cluster Faso, Mali, Benin, and Western Niger (Fig. 1),
separately. We used both approaches, performing a whereas the phylogroup A-II includes animals from
stratified Mantel test for the whole data set and a the easternmost localities of Niger, Cameroon, Congo,
simple Mantel test for each cluster identified by phy- Gabon, Chad, Central African Republic (CAR), and
logeographic analyses. Both tests were performed Democratic Republic of Congo (DRC). Here, phylo-
using the R package VEGAN (Oksanen et al., 2012). group A-III is represented by animals from the Lake
The significance of the Mantel test was assessed Victoria vicinity of Kenya only. This latter phylogroup
Figure 2. Maximum-likelihood tree based on Mastomys natalensis unique cytochrome b haplotypes. The tree was
obtained using a Tamura–Nei model with gamma-distributed rate variation across sites (G = 1.309) and a proportion of
invariable sites (I = 0.487). Major phylogroups are identified by arrows. Only bootstrap values > 50 were reported.
Figure 3. Phylogenetic Bayesian inference and molecular clock estimations based on Mastomys natalensis cytochrome b
sequences. Numbers at nodes represent posterior probabilities. The scale is in millions of years. Bars represent the 95%
highest posterior densities (HPDs). Divergence time was estimated for the nodes highlighted by letters. The means and
the 95% HPD intervals for major splits were reported. Black circles indicate fossil calibration points.
Figure 4. Mastomys natalensis Neighbor-Net phylogenetic network based on Tamura–Nei distances. Multiple connec-
tions represent conflict and ambiguous signal in the data set. Phylogroups are labelled according to the results of the
phylogenetic analyses.
forms a monophyletic assemblage with phylogroup HISTORICAL DEMOGRAPHY

A-II. Within clade B, phylogroup B-IV includes
sequences from central and northern Tanzania, All phylogroups show high haplotype diversity
southern Kenya, and Rwanda (Fig. 1), phylogroup (Table 1). However, the West and Central African
B-V clusters individuals from a restricted area lineages (i.e. A-I, A-II, and A-III) displayed a lower
of south-eastern Tanzania, and phylogroup B-VI nucleotide diversity (Table 1) than the Eastern and
includes animals from south-western Tanzania, Southern lineages (i.e. B-IV, B-V, and B-VI).
Zambia, Mozambique, Botswana, Zimbabwe, and The mismatch distributions for five of the six phy-
South Africa, thus covering a large part of Southern logroups (phylogroup A-III was excluded from the
Africa (Fig. 1). The latter two phylogroups (B-V and analysis because of a small sample size) present clear
B-VI) are associated in a monophyletic clade. bell-shaped curves, thus suggesting that the corre-
Our molecular clock estimates (Fig. 3) suggest sponding populations underwent past and sudden
that M. natalensis diverged from M. erythroleucus demographic expansions. Both Fu’s Fs and R2 statis-
c. 2.27 Mya (95% HPD 1.72–2.88). Within M. natalen- tics suggest a significant departure from neutrality
sis, the molecular clock suggests a rapid differentia- (Table 1), supporting an exponential growth model for
tion of different lines (points A, B, and C in Fig. 3) the phylogroups.
during the lower Calabrian age (0.781–1.806 Mya). The Bayesian skyline plots obtained for the
Two successive differentiation events within M. na- phylogroups A-I, A-II, B-IV, and B-VI (Fig. 5) suggest
talensis (points D and E in Fig. 3) then occurred episodes of demographic growth between 100 000 and
during the Ionian age (0.126–0.781 Mya). 150 000 years ago. Subsequently, phylogroups B-IV
Table 1. Diversity indexes: the number of specimens (N), the number of haplotypes (h), the haplotype (Hd), and
nucleotide (p) diversity were reported for each of the main phylogroups; Fu’s Fs and Rozas’ R2 test were calculated for
all the phylogroups, except for A-III, because of its low sample size
Phylogroup N h Hd p Fu’s Fs Rozas’ R2
A-I 85 60 0.991 0.00617 -60.0** 0.036*

A-II 41 33 0.988 0.00623 -24.5** 0.040**
A-III 5 5 1 0.00416 NA NA
B-IV 24 22 0.993 0.00704 -13.53** 0.042**
B-V 12 11 0.985 0.00714 -3.64* 0.121
B-VI 42 37 0.993 0.00939 -26.32** 0.042**
*P < 0,05, **P < 0.01.
and B-VI reached stationary demographic phases separated by a pronounced level of genetic divergence
(Fig. 5), whereas phylogroup A-II displays a signature that contrasts with low within-group divergence. IBD
of slight population growth that seems to be continu- is not supported at the continental scale, and the
ing at present (Fig. 5). Phylogroup A-I was character- prominent genetic gaps observed among phylogroups
ized by a slight decline at c. 25 000–30 000 years ago, can be interpreted by the persistence of long-term
followed by a second episode of population expansion extrinsic barriers that triggered phylogroups isolation
(Fig. 5). and associated lineage sorting (Avise, 2000).
The BSP for phylogroup B-V shows initial slight This work is limited to a single mitochondrial
growth, which started c. 150 000 years ago, followed marker only, and we cannot completely exclude the
by an episode of demographic growth at c. 25 000– possibility that the exchange of nuclear DNA has
50 000 years ago (Fig. 5). This event almost corre- happened among M. natalensis phylogroups, and that
sponds to the second expansion episode recorded for the observed pattern is somehow related to phylopa-
phylogroups A-I, but this result should be interpreted try in females. However, it should be stressed that
with great caution as it may be greatly biased by the some of the observed phylogroups (particularly A-I
small sample size. and B-VI) correspond to very large geographic
regions, thus excluding strict female phylopatry on
the scale considered here. Thus, a break of the gene
ISOLATION BY DISTANCE
flow among phylogroups triggered by the occurrence
Figure 6 presents the scatter plot of uncorrected of long-term extrinsic barriers appears as the most
genetic distances versus geographic distance. At the plausible explanation for the extant structure
continental level, no IBD pattern is observed. By observed in M. natalensis.
contrast, the scatter plot shows the presence of dif- The six M. natalensis phylogroups apparently
ferent clusters, reflecting the previously detected phy- show a very small overlap of their respective distri-
logeographic structure. The stratified Mantel test bution ranges. Strikingly, only in two localities
confirms the absence of a significant relationship (at Muze and Masasi, situated in the south of the
between genetic and geographic distances for the Rift valley in southern Tanzania) did we detect hap-
whole data set (r = 0.48; P = 0.084). When the analy- lotypes from two different phylogroups. This very
sis was repeated for the clades separately, we found a strong spatial exclusion suggests an almost complete
significant positive correlation for clades A-I (r = 0.34, absence of mitogenetic exchanges between all of the
P < 0.001), B-IV (r = 0.46; P = 0.009), and B-VI regions concerned, although some of them are geo-
(r = 0.24, P < 0.001), but not for A-II (r = 0.13, graphically close to each other. Similar phylogeo-
P = 0.112) and B-V (r = 0.40, P = 0.101). The phylo- graphic patterns were reported for Lemniscomys
group B-III was excluded from the analysis because of striatus (Nicolas et al., 2008b), Acomys chudeaui
the limited sample size (only two localities). (Nicolas et al., 2009), and Mastomys erythroleucus
(Brouat et al., 2009), among others, and were
explained by the past occurrence of habitat frag-
DISCUSSION
mentation. As an IBD model is not supported for
The analysis of cyt b sequences reveals a strong and M. natalensis at the spatial scale under considera-
robust phylogeographic structure of M. natalensis tion, the explanation for the observed phylogeo-
throughout its range. Our phylogenetic analysis sup- graphic pattern must be sought in past occurrences
ports the presence of six distinct phylogroups, all of geographic and/or ecological barriers.
Figure 5. Mastomys natalensis Bayesian skyline plots (on the left) and mismatch distributions (on the right) based on
cytochrome b sequences. Analyses were reported for each of the major phylogroups: A-I, A-II, B-IV, B-V, and B-VI.
For the BSP the solid line is the median estimate, and the dotted lines show the 95% highest posterior density limits.
Time of expansion was calculated using the mean clock rate obtained for M. natalensis in molecular clock analysis
(mean.rate = 0.02994). For the mismatch distributions the observed distribution of pairwise distances are plotted
(histogram), as well as an empirical density estimate (solid line).
Figure 6. Scatter plot of genetic uncorrected distances versus geographic distances for the whole data set of Mastomys
natalensis. Different clusters, reflecting the phylogeographic structure of M. natalensis, can be observed.
The geological history of Africa has been character- et al., 2001; Nicolas et al., 2008b). As such, the respec-
ized by the appearance of different geographic barri- tive distribution of the two main clades resembles the
ers, mainly linked to past volcanic and tectonic pattern observed in other savannah mammal subspe-
activity, which may have had a role as a putative cies or species (e.g. Lorenzen et al., 2007; Lorenzen,
barrier to dispersion and gene flow among M. natalen- Arctander & Siegismund, 2008). It also mirrors the
sis populations. Moreover, the monsoon dynamics and genetic differentiation between populations of widely
glaciation cycles during the Pleistocene, which sig- distributed mammals, such as African bovids
nificantly affected the African biomes, may have (Arctander et al., 1999; Moodley & Bruford, 2007). The
contributed to the fragmentation of M. natalensis regional uplift of the western and eastern branches of
populations, triggering the expansion of unsuitable the rift dates back to the early–mid Miocene (Par-
habitats. tridge, Wood & deMenocal, 1995); however, the tec-
According to our phylogenetic analysis, the six phy- tonic activity that resulted in the major uplift
logroups are clustered into two separate assemblages, increased approximately 3 Mya, and reached a
with clear biogeographic significance, namely a maximum during what was formerly known as the
western (clade A) and an eastern (clade B) lineage. Plio-Pleistocene (2–1.6 Mya) interval (Denys, Chorow-
However, the NN phylogeographic network shows a icz & Tiercelin, 1986). According to our estimations,
star-like appearance of the phylogeographic structure the split between clades A and B occurred between
and an ambiguous phylogenetic signal in the initial 0.87 and 1.5 Mya, and is thus congruent with a vicari-
phases of the M. natalensis diversification. This is ance scenario resulting from this tectonic activity.
probably because of rapid diversification within During this very active period, both the Western and
M. natalensis, as suggested by the molecular clock the Eastern branches of the Rift Valley were frag-
analysis. mented into numerous different fault basins, and the
The split into two main clades may reflect the Rift Valley may then have greatly impaired migration/
separation of the African continent by the East African dispersal processes, resulting in high endemicity, par-
Rift Valley that runs southwards from the Afar ticularly in small mammals (Denys et al., 1986).
Depression in Ethiopia to Lake Malawi, as it has Today, M. natalensis is found in areas of the Rift
already been reported to be an efficient barrier to gene Valley characterized by high elevations. Thus, rather
flow in several other mammalian species (e.g. Girman than the presence of mountains and slopes, the
presence of widespread deep palaeolakes during Udzungwa Mountains and Mount Rungwe in the
humid periods, such as those reported in the Kenya south) are good candidates as past barriers to gene
Rift between 1.9 and 1.7 Mya, as well as between 1.0 flow among the three phylogroups of clade B (see
and 0.9 Mya (Trauth et al., 2005), may have contrib- Appendix S2).
uted to the fragmentation of an originally continuous The Malawi and Rukwa rifts started to develop
population distributed on either side of the Rift. 8 Mya, but both tectonic and volcanic activity con-
The first divergence within clades A (95% HPD tinued up to the mid-Pleistocene, accompanied by
0.74–1.33) and B (95% HPD 0.73–1.31) occurred significant water level fluctuations (Delvaux, 1995;
quickly after the first split. In western Africa, the Delvaux et al., 1998). In particular, Lake Rukwa
Niger river may have acted as a barrier between A-I flowed into Lake Tanganyika during the last humid
and A-II, as it has been suggested for other rodents period (Delvaux et al., 1998). Thus, during periods of
(Dobigny et al., 2005, 2010; Brouat et al., 2009; climate shift, with an increase in humidity, both
Nicolas et al., 2009, 2011; Hima et al., 2011), and Rukwa and Malawi rifts are likely to have repre-
earlier for mammal species distribution patterns sented efficient barriers to animal movements,
(Robbins, 1978). The Cameroon volcanic line, which which is compatible with the allopatric isolation of
reaches an elevation of more than 4000 m a.s.l. and phylogroups B-IV and B-VI during the Pleistocene.
extends along the border between Nigeria and Cam- The uplift of the Eastern Arc Mountains is ancient,
eroon, and Lake Chad, which was much larger during and the modern mountains are the result of a final
the Quaternary (Leblanc et al., 2006), have also been uplift that started c. 7 Mya (Griffith, 1993), largely
reported as putative barriers (Dobigny et al., 2005; pre-dating the estimated split of divergence among
Brouat et al., 2007, 2009; Hima et al., 2011; Nicolas the eastern African lineages. In addition to geomor-
et al., 2012), and may also have played a role in the phological barriers, climate-driven ecological vicari-
divergence of A-I and A-II (see Appendix S2). ance may also explain the phylogeographical
Unlike the geographically widespread phylogroups patterns found in M. natalensis, as suggested in
A-I and A-II, phylogroup A-III appears confined to a other mammals (Arctander et al., 1999; Nicolas
small region situated between the eastern and the et al., 2008b). For instance, the Tanzanian plateau,
western branches of the Rift Valley, where it may where phylogroup B-IV is found, is surrounded by a
have been isolated during the middle Pleistocene. ‘forest circle’ (see Appendix S2), covering the Eastern
However, our small sample does not allow us to define Arc Mountains as well as the southern and western
the real extent of phylogroup A-III and the possible Rift during wet Pleistocene phases. Forest expan-
contribution of geographic or climatic barriers to the sion may have generated a belt of unsuitable
isolation of this phylogroup. habitat for M. natalensis, as suggested by the
In clade B, geographic features may also have pro- perfect match observed between the locations of the
moted initial allopatric isolation between phylo- forest circle and locations of the contact zones
groups. The phylogroup B-IV is mostly confined to between the three eastern phylogroups. A similar
the East African Plateau, whereas B-V seems to be scenario may be invoked to explain the separation of
present only on the east side of the Eastern Arc western African phylogroups belonging to clade A
Mountains, at a lower altitude compared with phy- because of the expansion of Guineo-Congolian
logroup B-IV. The distribution pattern of these two forests and Cameroonian highland forests (see
phylogroups is remarkably similar to that found for Appendix S2). In M. natalensis, the hypothesis of
two new Acomys species (Verheyen et al., 2011). habitat expansion/contraction is reinforced by the
Phylogroup B-VI is mostly distributed in southern inferred historical demography. Indeed, the signa-
Africa, with its northernmost limits reaching South tures of population expansions found during the last
Tanzania along the southern part of the Rift Valley, Pleistocene suggest that the divergence among phy-
as observed in other African rodents such as Ger- logroups was followed by demographic bottlenecks,
billiscus leucogaster (Colangelo et al., 2007). The and by almost simultaneous demographic expan-
congruence between the phylogeographic patterns sions in phylogroups A-I, A-II, B-IV, and B-VI
found in M. natalensis and the geographic bounda- c. 100 000–150 000 years ago. Such timing roughly
ries of other rodent species suggests the occurrence coincides with the onset of the last glacial period
of common extrinsic factors that influenced the dis- when a drier climate favoured the expansion of
tribution of several African rodents. The western savannah (deMenocal, 1995). Phylogroups B-V and
branch of the Rift Valley (running along the border A-I underwent a second and more recent expansion
between Tanzania and DRC, southwards to Malawi) phase. In A-I, it was dated c. 25 000–35 000 year
and the Eastern Arc Mountains (which run from the ago, corresponding to the spread of ice sheets up to
Taïta Hills, close to the Kenyan border with Tanza- their maximum surface in the Northern hemisphere
nia in the North, through Eastern Tanzania to the (i.e. between 33 000 and 26 500 years ago; Clark
et al., 2009). In Africa, it is possible that M. natalen- (Tribe Alcelaphini). Molecular Biology and Evolution 16:
sis populations experienced bottlenecks because of 1724–1739.
the reduction of suitable habitat during Pleistocene Avise JC. 2000. Phylogeography: the history and formation of
interglacial phases, followed by demographic expan- species. Cambridge, MA: Harvard University Press.
sion and increased geographic range during Pleis- Brouat C, Loiseau A, Kane M, Bâ K, Duplantier JM.
tocene glacial phases. 2007. Population genetic structure of two ecologically
distinct multimammate rats: the commensal Mastomys
natalensis and the wild M. erythroleucus in south-eastern
CONCLUSIONS Senegal. Molecular Ecology 16: 2985–2997.
Brouat C, Tatard C, Bâ K, Cosson JF, Dobigny G, Fichet-
As a result of the high ecological tolerance and adapt- Calvet E, Granjon L, Lecompte E, Loiseau A, Mouline
ability of M. natalensis, we expected to observe weak K, Piry S, Duplantier JM. 2009. Phylogeography of the
phylogeographic structure for this species. However, Guinea multimammate mouse (Mastomys erythroleucus): a
we found significant genetic divergence among phy- case study for Sahelian species in West Africa. Journal of
logroups that was not explainable by just invoking Biogeography 36: 2237–2250.
IBD. The strict congruence between geography and Bryant D, Moulton V. 2004. Neighbor-net: an agglomerative
phylogeographic structure in M. natalensis suggest method for the construction of phylogenetic networks.
the occurrence of long-term barriers that isolated and Molecular Biology and Evolution 21: 255–265.
triggered the divergence among six main phylo- Bryja J, Granjon L, Dobigny G, Patzenhauerová H,
groups. Both geographic barriers and habitat vicari- Konečný A, Duplantier JM, Gauthier P, Colyn M,
ance may have played a role in the phylogeographical Durnez L, Lalis A, Nicolas V. 2010. Plio-Pleistocene
history of M. natalensis. Initial phases of divergence history of West African Sudanian savanna and the phylo-
might be related to rifting activity and the formation geography of the Praomys daltoni complex (Rodentia):
of palaeolakes. However, NN network, BSP, and the environment/geography/genetic interplay. Molecular
molecular clock estimations suggest that divergence Ecology 19: 4783–4799.
among the major phylogroups occurred very rapidly, Castiglia R, Solano E, Makundi RH, Hulselmans J, Ver-
heyen E, Colangelo P. 2012. Rapid chromosomal evolu-
and almost at the same time in different parts of
tion in the mesic four-striped grass rat Rhabdomys dilectus
Africa. These results support a climate-driven diver-
(Rodentia, Muridae) revealed by mtDNA phylogeographic
sification scenario on a whole-continent scale, rather
analysis. Journal of Zoological Systematics and Evolution-
than local emergences of phylogroups as a result of
ary Research 50: 165–172.
geographic barriers or regional climatic shifts.
Cerling TE, Harris JM, MacFadden BJ, Leakey MG,
Quadek J, Eisenmann V, Ehleringer JR. 1997. Global
ACKNOWLEDGEMENTS vegetation change through the Miocene/Pliocene boundary.
Nature 389: 153–158.
We are grateful to numerous collectors as well as the Clark PU, Dyke AS, Shakun JD, Carlson AE, Clark J,
many colleagues that helped us in the field. This Wohlfarth B, Mitrovica JX, Hostetler SW, McCabe AM.
study was supported by the Institut de Recherche 2009. The last glacial maximum. Science 325: 710–714.
pour le Développement (IRD), the Muséum National Colangelo P, Granjon L, Taylor PJ, Corti M. 2007. Evo-
d’Histoire Naturelle (MNHN), the European Commu- lutionary systematics in African gerbilline rodents of the
nity (INCO-DEV grant ICA4-CT2002-10050, the genus Gerbilliscus: inference from mitochondrial genes.
Marie Curie Intra-European Fellowship (MEIF- Molecular Phylogenetics and Evolution 42: 797–806.
CT2003-506686), and the French ANR programme Delvaux D. 1995. Age of Lake Malawi (Nyasa) and water
(ANR Biodiversité – IFORA, ANR-JC-JC0548631). level fluctuations. Musée royal de l’Afrique centrale
E.V. was partially funded by FWO-V project (Belgique), Département de Géologie et de Minéralogic,
Rapp. Ann. 1995–1996: 99–108.
2.0004.91, a Belgian Science Policy Office (BELSPO)
Delvaux D, Kervyn F, Vittori E, Kajara RSA, Kilembe E.
project entitled ‘Evaluating the effect of Pleistocene
1998. Late Quaternary tectonic activity and lake level
climate changes on speciation patterns in selected
change in the Rukwa Rift Basin. Journal of African Earth
African vertebrates’. The molecular work was partly
Sciences 26: 397–421.
carried out in the Joint Experimental Molecular Unit
Denys C. 1987. Fossil rodents (other than Pedetidae) from
(JEMU) laboratory of the Royal Belgian Institute of Laetoli. In: Leakey MD, Harris JM, eds. Laetoli: a Pliocene
Natural Sciences. site in northern Tanzania. London: Oxford Science Publica-
tions, 118–170.
Denys C, Chorowicz J, Tiercelin JJ. 1986. Tectonic and
REFERENCES
environmental control on rodent diversity in the Plio-
Arctander P, Johansen C, Coutellec-Vreto MA. 1999. Pleistocene sediments of the african rift system. In: Frostick
Phylogeography of three closely related African bovids LE, Renaut RW, Reid J, Tiercelin JJ, eds. Sedimentation
in the African rifts. Geological Society Special Publication performance of PhyML 3.0. Systematic Biology 59: 307–
No. 25, Oxford: Alden Press, 363–372. 321.
Dobigny G, Aniskin V, Granjon L, Cornette R, Volo- Hall T. 1999. BioEdit v7.1.3: biological sequence alignment
bouev V. 2005. Recent radiation in West African Taterillus editor for Win95/98/NT/2K/7. Available at: http://www.
(Rodentia Gerbillinae): the concerted role of chromosome mbio.ncsu.edu/BioEdit/bioedit.html
and climatic changes. Heredity 95: 358–368. Hima K, Thiam M, Catalan J, Gauthier P, Duplantier
Dobigny G, Catalan J, Gauthier P, O’Brien PCM, JM, Piry S, Sembène M, Britton-Davidian J, Granjon
Brouat C, Bâ K, Tatard C, Ferguson-Smith M, Duplan- L, Dobigny G. 2011. Extensive Robertsonian polymor-
tier JM, Granjon L, Britton-Davidian J. 2010. Geo- phism in the African rodent Gerbillus nigeriae: geographic
graphic patterns of inversion polymorphisms in a wild aspects and meiotic data. Journal of Zoology 284: 276–
African rodent, Mastomys erythroleucus. Heredity 104: 378– 285.
386. Horn S, Durka W, Wolf R, Ermala A, Stubbe A, Stubbe
Drummond AJ, Ho SYW, Phillips MJ, Rambaut A. 2006. M, Hofreiter M. 2011. Mitochondrial genomes reveal slow
Relaxed phylogenetics and dating with confidence. PLoS rates of molecular evolution and the timing of speciation in
Biology 4: e88. beavers (Castor), one of the largest rodent species. PLoS
Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evo- ONE 6: e14622.
lutionary analysis by sampling trees. BMC Evolutionary Huhndorf M, Kerbispeterhans J, Loew S. 2007. Compara-
Biology 7: 214. tive phylogeography of three endemic rodents from the
Drummond AJ, Rambaut A, Shapiro B, Pybus OG. 2005. Albertine Rift east central Africa. Molecular Ecology 16:
Bayesian coalescent inference of past population dynamics 663–674.
from molecular sequences. Molecular Biology and Evolution Huson DH, Bryant D. 2006. Application of Phylogenetic
22: 1185–1192. Networks in Evolutionary Studies. Molecular Biology and
Ducroz JF, Volobouev V, Granjon L. 1998. A molecular Evolution 23: 254–267.
perspective on the systematics and evolution of the genus Julliard R, Leirs H, Stenseth NC, Yoccoz NG, Prevot-
Arvicanthis (Rodentia Muridae): inferences from complete Julliard AC, Verhagen R, Verheyen W. 1999. Survival-
cytochrome b gene sequences. Molecular Phylogenetics and variation within and between functional categories of the
Evolution 10: 104–117. African multimammate Rat. Journal of Animal Ecology 68:
Faulkes CG, Bennett NC, Cotterill FPD, Stanley W, 550–561.
Mgode GF, Verheyen E. 2011. Phylogeography and Kass RE, Raftery AE. 1995. Bayes factors. Journal of the
cryptic diversity of the solitary-dwelling silvery mole-rat, American Statistical Association 90: 773–795.
genus Heliophobius (family: Bathyergidae). Journal of Keane TM, Creevey CJ, Pentony MM, Naughton TJ,
Zoology 285: 1469–7998. McInerney JO. 2006. Assessment of methods for amino
Faulkes CG, Verheyen E, Verheyen W, Jarvis JUM, acid matrix selection and their use on empirical data
Bennett NC. 2004. Phylogeographic patterns of genetic shows that ad hoc assumptions for choice of matrix are not
divergence and speciation in African mole-rats (Family: justified. BMC Evolutionary Biology 6: 29.
Bathyergidae). Molecular Ecology 13: 613–629. Leblanc MJ, Leduc C, Stagnitti F, van Oevelen PJ,
Flagstad Ø, Syvertsen PO, Stenseth NC, Jakobsen KS. Jones C, Mofor LA, Razack M, Favreau G. 2006.
2001. Environmental change and rates of evolution: the Evidence for Megalake Chad, north-central Africa, during
phylogeographic pattern within the hartebeest complex as the late Quaternary from satellite data. Palaeogeography,
related to climatic variation. Proceedings of the Royal Palaeoclimatology, Palaeoecology 230: 230–242.
Society London 268: 667–677. Lecompte E, Denys C, Granjon L. 2005. Confrontation of
Fu Y-X. 1997. Statistical tests of neutrality of mutations morphological and molecular data: the Praomys group
against population growth, hitchhiking and background (Rodentia Murinae) as a case of adaptive convergences and
selection. Genetics 147: 915–925. morphological stasis. Molecular Phylogenetics and Evolution
Geraads D. 2002. Plio-pleistocene mammalian biostratigra- 37: 899–919.
phy of atlantic morocco. Quaternaire 13: 43–53. Lecompte E, Granjon L, Petherans JK, Denys C. 2002.
Girman DJ, Vilà C, Geffen E, Creel S, Mills MGL, Mcnutt Cytochrome-b based phylogeny of the Praomys group
JW, Ginsberg J, Kat PW, Mamiya KH, Wayne RK. 2001. (Rodentia Murinae): a new African radiation? CR Biologie
Patterns of population subdivision gene flow and genetic 325: 827–840.
variability in the African wild dog (Lycaon pictus). Molecu- Leirs H, Stenseth NC, Nichols JD, Hines JE, Verhagen
lar Ecology 10: 1703–1723. R, Verheyen W. 1997. Stochastic seasonality and nonlinear
Griffith CJ. 1993. The geological evolution of East Africa. density-dependent factors regulate population size in an
In: Lovett JC, Wasser SK, eds. Biogeography and ecology of African rodent. Nature 389: 176–180.
the rain forests of Eastern Africa. Cambridge: Cambridge Librado P, Rozas J. 2009. DnaSP v5: a software for
University Press, 9–21. comprehensive analysis of DNA polymorphism data.
Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk Bioinformatics 25: 1451–1452.
W, Gascuel O. 2010. New algorithms and methods to Lorenzen ED, Arctander P, Siegismund HR. 2008. Three
estimate maximum-likelihood phylogenies: assessing the reciprocally monophyletic mtDNA lineages elucidate the
taxonomic status of Grant’s gazelles. Conservation Genetics Partridge TC, Wood BA, deMenocal PB. 1995. The influ-
9: 593–601. ence of global climate changes and regional uplift on large-
Lorenzen ED, de Neergaard R, Arctander P, Siegis- mammalian evolution in East and Southern Africa. In: Vrba
mund HR. 2007. Phylogeography, hybridization and Pleis- E, Denton G, Partridge TC, Burckle L, eds. Paleoclimate
tocene refugia of kob antelope (Kobus kob). Molecular and evolution with emphasis on human evolution. New
Ecology 16: 3241–3252. Haven: Yale University Press, 331–356.
Meirmans PG. 2012. The trouble with isolation by distance. R Core Team. 2012. R: a language and environment for
Molecular Ecology 21: 2839–2846. statistical computing. Vienna, Austria: R Foundation for
deMenocal PB. 1995. Plio-Pleistocene African climate. Statistical Computing. ISBN 3-900051-07-0. Available at:
Science 270: 53–59. http://www.R-project.org/
deMenocal PB. 2004. African climate change and faunal Rambau RV, Stanyon R, Robinson TJ. 2003. Molecular
evolution during the Pliocene–Pleistocene. Earth and genetics of Rhabdomys pumilio subspecies boundaries:
Planetary Science Letters 220: 3–24. mtDNA phylogeography and karyotypic analysis by fluores-
Moodley Y, Bruford MW. 2007. Molecular biogeography: cence in situ hybridization. Molecular Phylogenetics and
towards an integrated framework for conserving pan- Evolution 28: 564–575.
African biodiversity. PLoS ONE 2: e454. Ramos-Onsins SE, Rozas J. 2002. Statistical properties of
Mouline K, Granjon L, Galan M, Abdoullaye D, new neutrality tests against population growth. Molecular
Ag Atteyine S, Duplantier JM, Cosson JF. 2008. Biology and Evolution 19: 2092–2100.
Phylogeography of a Sahelian rodent species Mastomys Robbins CB. 1978. The Dahomey Gap. A revaluation of its
huberti: a Plio-Pleistocene story of emergence and colonisa- significance as a faunal barrier to West African high forest
tion of humid habitats. Molecular Ecology 17: 1036– mammals. Bulletin of the Carnegie Museum of Natural
1053. History 6: 168–174.
Musser GC, Carleton MD. 2005. Superfamily Muroidea. In: Rogers AR, Harpending H. 1992. Population growth makes
Wilson DE, Reeder DM, eds. Mammal species of the world: waves in the distribution of pairwise genetic differences.
a taxonomic and geographic reference, Vol. 2. Baltimore: Molecular Biology and Evolution 9: 552–569.
Johns Hopkins University Press, 894–1531. Russo IRM, Chimimba CT, Bloomer P. 2010. Bioregion
Nicolas V, Bryja J, Akpatou B, Konecny A, Lecompte E, heterogeneity correlates with extensive mitochondrial
Colyn M, Lalis A, Couloux A, Denys C, Granjon L. DNA diversity in the Namaqua rock mouse, Micaelamys
2008a. Comparative phylogeography of two sibling species namaquensis (Rodentia: Muridae) from southern Africa –
of forest-dwelling rodent (Praomys rostratus and P. tull- evidence for a species complex. BMC Evolutionary Biology
bergi) in West Africa: different reactions to past forest 10: 307.
fragmentation. Molecular Ecology 17: 5118–5134. Slatkin M, Hudson RR. 1991. Pairwise comparisons of mito-
Nicolas V, Granjon L, Duplantier JM, Cruaud A, chondrial DNA sequences in stable and exponentially
Dobigny G. 2009. Phylogeography of spiny mice (genus growing populations. Genetics 129: 555–562.
Acomys Rodentia: Muridae) from the southwestern margin Sluydts V, Crespin L, Davis S, Lima M, Leirs H.
of the Sahara with taxonomical implications. Biological 2007. Survival and maturation rates of the African rodent
Journal of the Linnean Society 98: 29–46. Mastomys natalensis: density-dependence and rainfall.
Nicolas V, Mboumba JF, Verheyen E, Denys C, Integrative Zoology 2: 220–232.
Lecompte E, Olayemi A, Missoup AD, Katuala P, Swofford DL. 2003. PAUP*. Phylogenetic analysis using par-
Colyn M. 2008b. Phylogeographic structure and regional simony (*and other methods). Version 4. Sunderland, MA:
history of Lemniscomys striatus (Rodentia: Muridae) in Sinauer Associates.
tropical Africa. Journal of Biogeography 35: 2074–2089. Tamura K, Dudley J, Nei M, Kumar S. 2007. MEGA4:
Nicolas V, Missoup AD, Colyn M, Cruaud C, Denys C. molecular evolutionary genetics analysis (MEGA) software
2012. West-Central African Pleistocene lowland forest evo- version 4.0. Molecular Biology and Evolution 24: 1596–
lution revealed by the phylogeography of Misonne’s soft- 1599.
furred mouse. African Zoology 47: 100–112. Tamura K, Nei M. 1993. Estimation of the number of nucle-
Nicolas V, Missoup AD, Denys C, Kerbis Peterhans J, otide substitutions in the control region of mitochondrial
Katuala P, Couloux A, Colyn M. 2011. The roles of rivers DNA in humans and chimpanzees. Molecular Biology and
and Pleistocene refugia shaping genetic diversity in Evolution 10: 512–526.
Praomys misonnei in tropical Africa. Journal of Biogeogra- Trauth MH, Larrasoan JC, Mudelsee M. 2009. Trends,
phy 38: 191–207. rhythms and events in Plio-Pleistocene African climate.
Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin Quaternary Science Reviews 28: 399–411.
PR, O’Hara RB, Simpson GL, Solymos P, Stevens Trauth MH, Maslin MA, Deino A, Strecker MR. 2005.
MHH, Wagner H. 2012. vegan: community ecology Late Cenozoic moisture history of East Africa. Science 309:
package. R package version 2.0-4. 2051–2053.
Paradis E. 2010. pegas: an R package for population genetics Verheyen W, Hulselmans J, Wendelen W, Leirs H,
with an integrated-modular approach. Bioinformatics 26: Corti M, Verheyen E. 2011. Contribution to the
419–420. systematics and zoogeography of the East–African Acomys
spinosissimus Peters 1852 species complex and the descrip- dynamics in Tanzanian maize fields. Journal of Animal
tion of two new species (Rodentia: Muridae). Zootaxa 3059: Ecology 75: 213–220.
1–35. Winnepenninckx B, Backeljau T, De Wachter R. 1993.
Vibe-Petersen S, Leirs H, De Bruyn L. 2006. Effects of Extraction of high molecular weight DNA from molluscs.
predation and dispersal on Mastomys natalensis population Trends in Genetics 9: 407.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article:
Appendix S1. Mastomys natalensis specimens included in the analyses. The identification number, museum
code, country and locality of origin, haplotype number (used in the analysis), and haplogroup were reported for
each specimen. Specimens denoted with an asterisk are those for which cytochrome b was downloaded directly
from the GenBank.
Appendix S2. Putative geographic and ecological barriers in Western and Eastern Africa.

A Mitochondrial Phylogeographic Scenario

Uploaded by

Copyright:

Available Formats

A Mitochondrial Phylogeographic Scenario

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Mitochondrial Phylogeographic Scenario

Uploaded by

Copyright:

Available Formats

bs_bs_banner

A mitochondrial phylogeographic scenario for the most

*Corresponding author. E-mail: paolo.colangelo@uniroma1.it

INTRODUCTION sification of the Albertine Rift endemic rodents

forms a monophyletic assemblage with phylogroup HISTORICAL DEMOGRAPHY

Phylogroup N h Hd p Fu’s Fs Rozas’ R2

A-I 85 60 0.991 0.00617 -60.0** 0.036*

*P < 0,05, **P < 0.01.

You might also like