Lophuromys, Peters, 1874

Kostin, Danila S., Kasso, Mohammed, Komarova, Valeria A., Martynov, Alexey A., Gromov, Anton R., Alexandrov, Dmitry Y., Bekele, Afework, Zewdie, Chemere, Bryja, Josef & Lavrenchenko, Leonid A., 2019, Taxonomic and genetic diversity of rodents from the Arsi Mountains (Ethiopia), Mammalia (Warsaw, Poland) 83 (3), pp. 237-247 : 239-242

Genus Lophuromys

Recently published African rodent distribution maps ( Lavrenchenko 2013 a, Monadjem et al. 2015) suggest that the AM should be inhabited by the short-tailed brush-furred rat L. brevicaudus Osgood, 1936 only, though there has also been a genetically unconfirmed record of

BM

%

genetic

AM-BM distance ±

% 0.003

±

0.003 %

±

0.008 %

±

% 0.003 the in Absent

BM

± 0.3

% haplotypes haplotypes haplotypes

±

% 0.41

±

% 0.005 from sequences ±

6.99

% 0.9

– 0.93 0.84 3.85 0.65 0.86 Mixed Mixed Mixed 0.71 2.39 No 4.44

sea above Ethiopia

) 2009 4500 – 3500 – 2760 – 4100 – – 3400 3200 – – 4100 4000 – 3800 – – 4100 – 3730 3500 – 3200 –

in

m

(

range

species Lavrenchenko 3100 2400 1200 2750 1900 1200 3000 2400 820 3170 3000 3100 1200

Altitudinal of)

the level after (.

haplotypes

elevation

3800

/ 10 /

47 21 22 59 24 18 6

/ 5 identical

nights 3400 1 shared

trapping 7

/ 1 / BM) (

number of

/ / 8 1

3331 1

Mountains locality

/

3294 / 2 2 10 9 10 1 Bale and of

7 /

1

AM)

Number

4

3200

/

/

9

9 1 2 15 4 1 2

(Mountains

3100 5 5 Arsi the

.

1

/

6

/

from surveys / 3100 2 20 2

field 3 /

9

populations during our

1

2 / /

2500 5 2

conspecific.

rodents trapped

List of

melanonyx breVicaUdUS chrySopUS

blicki imberbiS mahomet StenocephalemyS albocaUdata

griSeicaUda

StenocephalemyS albipeS

StenocephalemyS helleri nikolaUSi sp.

SplendenS ”

that haplotypes means are ordered elevation by

:

Table

2 Species LophUromyS LophUromyS LophUromyS ArVicanthiS MUS MUS OtomyS DendromUS DendromUS TachyorycteS “ Mixed Localities

Lophuromys melanonyx ( Kasso et al. 2010) . In our study, we recorded three brush-furred rat species in the AM. The first, L. melanonyx , an obligatory Afroalpine dweller ( Yalden and Largen 1992), was recorded at high-densities in the Badda region (site 5) at the highest elevation sampled in the AM. Almost all animals were trapped during the day within colonies co-inhabited by A. blicki . Simultaneous use of the same burrow by both species was repeatedly recorded by the camera-traps ( Figure 2 ). Similar external morphology, vocalisation and behaviour make this species pair an amazing example of convergent evolution in rodents adapted to extreme Afroalpine conditions. The second species, L. brevicaudus , is common in the ericaceous belt in the Ethiopian highlands, east of the ERV ( Lavrenchenko 2013a). The species was abundant in the Shirka region (site 1), the western slope of Mount Chilalo (site 4) and in the Badda region (site 5). The third species, the Ethiopian forest brush-furred rat L. chrysopus Osgood, 1936 , was trapped during a short survey of the montane forest near Assela (site 2). This species is typically a forest dweller of the southern Ethiopian highlands on both sides of the ERV ( Lavrenchenko 2013b) and our records confirmed previous observations.

Analysis of Lophuromys melanonyx CYTB sequences revealed two distinct haplogroups. Hereafter, the previously known haplogroup of L. melanonyx is termed “Melanonyx-II” (= L. melanonyx sensu Lavrenchenko et al. 2004 ), while the second haplogroup, reported here for the first time, is termed “Melanonyx-I” ( Figure 3 ). The mean p -distance between these two haplogroups was 6.54 ± 0.011%. We then analysed variability at four nuclear markers in order to test whether individuals bearing particular mitochondrial haplogroups represented two distinct gene pools (=species). The concatenated tree suggested that L. melanonyx (AM) represents a single genetic species as no monophyletic groups (defined by mitochondrial haplogroups) were evident on the nuclear markers ( Figure 4). The pattern observed is likely to be either a result of ancestral polymorphism or (more likely) mitochondrial introgression.

It is worth mentioning here that two distinct mitochondrial haplogroups were also found in another mountain species, L. simensis ( Lavrenchenko et al. 2004) . As one (termed “North II”) was very close to Lophuromys melanonyx (“Melanonyx-II”) and L. menageshae ( Figure 3 ), Lavrenchenko et al. (2004, 2007) proposed ancient reticulate speciation processes, involving past hybridisation between the species. Our discovery of “Melanonyx-I” may well shed light on the evolutionary history of this complex. Phylogenetic analysis revealed strong support (by 97% bootstrap value) for a monophyletic clade consisting of

0.02

Lophuromys menageshae 868 1 Lophuromys menageshae 873 1 Lophuromys simensis North I 1119 1 Lophuromys simensis North I 1120 0.55 Lophuromys simensis North II 1149

Lophuromys simensis North II 1147 0.008

Lophuromys brevicaudus Arsi 2146 1 Lophuromys brevicaudus Arsi 2152 Lophuromys melanonyx I Arsi 2672 1 1 Lophuromys melanonyx I Arsi 2289 1 Lophuromys melanonyx II Bale 2390 Lophuromys melanonyx II Arsi 2657 Lophuromys chrysopus 3079 Acomys wilsoni 3148

three mitochondrial lineages belonging to three different species ( L. melanonyx “Melanonyx-II”, L. menageshae and L. simensis “North II”). Mitochondrial sequences of L. melanonyx “Melanonyx-I” and L. simensis “North I” formed separate clades, differing from the composite clade by 6.74 ± 0.011% and 6.83 ± 0.012%, respectively. Based on these results, the following evolutionary scenario is proposed. “Melanonyx-I” and “North I” presumably represent the original species-specific mitochondrial haplotypes for L. melanonyx and L. simensis , respectively. The “Melanonyx-II” and “North II” mitochondrial haplogroups could have been introgressed into L. melanonyx and L. simensis through hybridisation with L. menageshae living at lower elevations in Central Ethiopia. While the distribution of the two high-elevation species (i.e. L. melanonyx , east of ERV, and L. simensis , west of ERV) is presently separated from that of L. menageshae (see distribution maps in Lavrenchenko et al. 2007), Pleistocene climatic fluctuations are likely to have repeatedly shifted the distribution of vegetation zones, along with their associated biota ( Osmaston et al. 2005, Umer et al. 2007, Bryja et al. 2018).

For both L. brevicaudus and Lophuromys melanonyx (in the case of L. melanonyx , we only took the more widespread “Melanonyx-II” haplogroup), the negligible genetic distance between the AM and BM populations (Table 2) suggests a relatively recent split (late Pleistocene). In contrast, specimens of L. chrysopus captured in the AM displayed a more pronounced genetic distance from their BM conspecifics (Table 2). Topology of the reconstructed phylogenetic tree (without bootstrap support) suggests a closer relationship between L. chrysopus from the AM and conspecific populations from the Beletta and Sheko forests in the northwestern plateau ( Figure 1 ) than with population from the geographically neighbouring BM ( Figure 3 ). Hence, the evolutionary history of L. chrysopus was probably more complex and most likely involved two independent crossings of the ERV, though the hypothesis needs further testing.

Figures

Figure 1: Map of the Arsi Mountains with localities where rodents were collected (1 = Galama Ridge, Shirka region; 2–4, 8 = Mount Chilalo, 5–7 = Galama Ridge). The map of Ethiopia shows the location of places mentioned in the text: AM = Arsi Mountains, BM = Bale Mountains, BF = Beletta Forest, SF = Sheko Forest. The dotted line denotes the border of the Ethiopian Rift Valley.

Figure 2: Black-clawed brush-furred rat LophUromyS melanonyx (left) and Blick’s grass rat ArVicanthiS blicki (right) photographed in the same burrow with a time difference of several minutes. Badda region of the Galama ridge.

Figure 3: Maximum likelihood phylogenetic tree of LophUromyS lineages based on alignment of the CYTB gene (1140 bp) with a reduced dataset. The remaining specimens were analysed by neighbour-joining analysis in Mega 7.0 (Kumar et al. 2016) and excluded from the final analysis as they were almost identical. The values above the branches indicate statistical support based on 1000 bootstrap replications. Bootstrap values below 75% are not shown.