Phylogeny and geographic distribution of rock lizards (Lacertidae, Reptilia) in Alborz mountain range

. Rock lizards of the genus have long been an important model object for study of reptile evolution.To understand the overall picture, it is important to know how bisexual andparthenogeneticspeciesaredistributedwithinthisgenus.TherangesofCaucasianspeciesof rock lizards have been studied for a long time and in detail. owever, recent attention to the species inhabiting the territory of Iran has been attracted after the description of several new species in 2013. As part of the continuation of these studies, we studied the distribution and ge-netic diversity of five species of lizards of the genus along theAlborz mountain range in Iran: , , , and . In the course of this,wediscoverednewlocalitiesof thatareoutsidetheirknownrange.Weassessed the phylogenetic relationships between the studied species according to the ND4 – Leu tRNA region of mitochondrial DNA and found a relatively high level of genetic variability in

, while , despite its wide distribution, has low variabi lity.In general, the phylogenetic position of the studied species is somewhat different from that described byAhmadzadeh et al. (2013).
-Alborz mountain range is one of the main centres of origin of rock lizards ( Due to the diversity of the landscape and, consequently, the variety of biotopes, geographic and ecological isolation is formed between species and populations within them.In this regard, the study of the genetic diversity and distribution of rock lizards on Alborz will expand the knowledge of the evolution of the genus.Darevskia Ahmadzadeh et al Freitas et al ., 2013;., 2016).
Species ranges can change quite strongly and quickly within a short time for many reasons.Ranges of reptiles and amphibians may shrink or expand due to the direct anthropogenic influence, such as habitat modification (Doherty et al.,20 ), noise (Simmons Narins, 2018) and probably light pollution (Perry et al., 2008), and indirect influence, for example climate change (Erasmus et al., 2002;Peck et al., 2009).In this case, there are difficulties associated with preventing the extinction of species and climate change observation.

MATERIAL AND METHODS
Sampling and Mapping.We collected the material along the western and central parts of Alborz mountain range along the southern coast of the Caspian Sea in Iran from the spring-summer period of 2022.During the expedition we carefully searched for lizards along the roads and four times crossed the mountain ridge from North to South and vice versa in different parts; we were also guided by the distribution map in the work of Ahmadzadeh et al. (2013).Visually detected lizards were captured by noose or by hand; dorsal and ventral pictures of each lizard were taken by telephone photo camera (IPhone 12 mini).Iryshkov, Arakelyan, Moaddab, Milto and Galoyan participated in the capture of lizards.
Tail tips of the lizards were fixed in 96% ethanol.Following species were included: 3 sp.When determining the distance between the capture points and the range boundary of , we calculated the distance between our point and the north westernmost point presented in Ahmadzadeh et al. (2013) using Google maps.Then we rounded the distance up to integer values of kilometres.Using Google maps, we also obtained the approximate height of each GPS point presented in the supplementary material from Ahmadzadeh et al., 2013 and added altitudes measured by us (presented

D. defilippii -
We took GPS coordinates of each finding using GPS tracker (Garmin 64S, WGS 84) and put the point on the map.Also we added terrestrial ecoregions according to bioregions word map from oneearth.org(2001).The basis of species distribution map was taken from Wikimedia Commons author NordNordWest, licensed under CC BY-SA 3.0; colours were changed, labels, points, ecoregions, scale bar and north arrow were added for the article.
Phylogenetic trees were reconstructed under Bayesian criteria (BI) and using the maximum likelihood (ML) method.The optimal partitioning schemes and mo dels for Bayesian Inference analysis were identified with PartitionFinder software ing sche mes were selected: F81+F for 1 position of ND4, TN+F for 2d, TN+F+G4 for 3d position and separately for tRNAs partition (for full dataset); F81+F for 1 position of ND4, TN+F+G4 for 2d, HKY+F+G4 for 3d position and separa tely for tRNAs partition (for full dataset).
st st - (Lanfear et al., 2012) using greedy search algorithm under AIC criterion.Following schemes were identified: F81 for 1 position of ND4, GTR+G for 2 and 3 positions separately, and HKY+G for tRNAs (for reduced dataset); for the full dataset HKY+G for 1 position of ND4, GTR+G for 2 , HKY+I+G for 3 position and for tRNAs partition separately (for full dataset).BI was perfor med using MrBayes v3.2.6 (Ronquist, Huelsenbeck, 2003) with two simultaneous runs, each with four chains, for 5 million generations.We checked the convergence of the runs and that the effective sample sizes (ESS) were all above 200 by exploring the likelihood plots using TRACER v1.7.1 (Rambaut et al., 2018).The initial 10% of trees were discarded as burnin.Con dence in tree topology was assessed by posterior probability (PP) (Huelsenbeck Ronquist, 2001).The ML trees were generated using IQtree software ( 2015) with ultrafast boot strap = 10,000 (UFBoot) (Minh et al., 2013), partitioning schemes and models were selected using ModelFinder software (Kalyaanamoorthy et al., 2017).Follow For polymerase chain reaction (PCR) we used ND4 (forward) and Leu (reverse) primers and stuck to the amplification protocol described by Dutton et al. (1996).PCR products were sequenced by the Microsynth DNA company using the forward primer and the manufacture protocols.
We used two types of datasets for phylogenetic analysis.First was a reduced dataset, including only our sequences and outgroups from GenBank, and second was a full dataset, including our sequences and sequences from the work of Ahmadzadeh et al. (2013) with outgroups from GenBank.

in
).We rounded the resulting height values up to tens.
Table 1 DNA Extraction, Amplification and Sequencing.We isolated genomic DNA from the 96% ethanol-fixed tissue samples using the SDS (Sodium Dodecyl Sulfate) DNA extraction method (Kabir et al., 2006), and the quality of extracted DNA was measured using NanoDrop 1000.We chose the mitochondrial DNA fragment ND4 -Leu tRNA (907 bp) which includes fragment of the NADH dehydrogenase 4 and His, Ser and Leu tRNAs (Arevalo et al., 1994) due to its sufficient variability and the presence of a large number of reptile sequences in the GenBank.
For visualisation we used dendrograms according to the Bayesian method (BI) and also added the values obtained E. S. Iryshkov E. N. Solovyeva, M. S. Arakelyan et al. ,

Phylogeny and geographic distribution of rock lizards
Table 1.Specimens used for phylogenetic analysis with localities and GenBank numbers of ND4 -Leu sequences obtained by us by maximum likelihood (ML) analysis.The dendrograms wa processed using FigTree v1.4.4 (tree.bio.ed.ac.uk/ software/figtree/).We calculated the uncorrected inter-and intragroup pairwise distances ( -distances) in the MEGA11 program using 1000 bootstrap replicas.ere p

RESULTS AND DISCUSSION
The topologies of the phylogenetic tree of our sequences, obtained by the BI and ML methods, are consistent in all nodes, but somewhat different in the levels of support.During visual sequence check we noticed a possible nuclear mitochondrial pseudogene (Bensasson et al., 2001) of ND4 -Leu tRNAfragment -the external parts of sequences were not completely read and a large number of double peaks on the electropherogram were observed (Chow et al., 2021).This was not noticed or reported in the closely related species and other distant species.

D. caspica
As shown in the dendrogram reconstructed from our sequences ( ig. 1), indeed has a fairly high genetic variability.This might be explained by the diverse habitats inhabited by (Ahmadzadeh et al., 2013).Phylogenetic relationships within reflect its geographical structure.Lineage consisting of specimens from Halu Dasht and Nilu villages from Gilan province diverged with the high level of support by BI, but with low support by ML -1/88.Specimen from Vadarbon village, Mazandaran province (OR340751) separates furthest with high node support -1/99.There is no connec tion between phylogenetic structure of and its geographic dispersal as in Unresolved nodes (Fig. 1) in and D. caspica D. schaekeli p together with their low within-group -distances indicate low genetic variability within these species (Table 2).2013) the topology of the tree changed (Fig. 2).Relationships between main clades and between D. defi lippii D. schaekeli and are resolved with a high support unlike the first tree, that contains only our sequences ( ig. 1).
is not monophyletic, it splits into three separate clusters.One cluster has high support, while two others form a trichotomy with within a lowsupported group (0.63/70).
specimens with high levels of support are divided according to localities from different provinces.We have also discovered two new locality records of D. defilippii beyond the distribution range (Ahmadzadeh et al., 2013).Records were taken 23 (OR340778 and OR340780 sequences) and 30 km (OR340724 and OR340725 sequences) northwest of previously documented locations.On the first point we caught individuals on a dry sandstone slope along an unpaved road surrounded by steppe and on the second point on a rocky slope with cracks along an asphalt road surrounded by mixed forest.In this regard, it can be concluded that either the range is initially wider, or it could expand towards the northwest during the last nine years.from Gilan province.Phylogenetic distribution of D. defilippii by localities has not been confirmed: nodes do not form distinct clusters and nodes supports are low.Phylogenetic relationships within still remain unclear: individual clusters by locality have not been formed, almost all nodes have low support, and many of them are unresolved.However, the most distant specimen within clade is from Dohezar, Mazandaran province (KF717320), and it is significantly divided from others.

D. chlorogaster D. chlorogaster
The presented species also have different altitude distributions.
occupies Hyrcanian mixed forest (Fig. 3) at an altitude mainly between 130 and 580 metres a.s.l., but also there was a population from the altitude of about 1590 m in Savasraeh (KF717318 sequence).

D. caspica
Although it separated from the specimens (KF717316, KF717317 sequences) from Joybar (altitude is about 400 m a.s.l.), it forms with these specimens one group with 0.99/97 support (Fig. 2) which indicates the lack of strong isolation.D. chlorogaster has a range with a large elevation difference: from 10 to 1250 m a.s.l., and confined to Hyrcanian forest and forest-steppe ecozones.also occupies a wide elevation range of 840-2040 m a.s.l.within forest and forest-steppe zones.
lives in the high altitudes of 1290-2500 m a.s.l. in the forest-steppe.Moreover, an altitudinal distribution of this species is wider -some populations are known from the forested areas at 440 and 640 m a.s.l.
is also a high altitudinal species and, according to our and already published data, distributed only at 1560-2300 m a.s.l. in forest-steppe ecozones.

D. raddei D. defilippii D. schaekeli
The Hyrcanian forest consists of a large variety of woody plant species with high densities.The humidity level is high due to the large number of mountain streams flowing down the mountain slopes (Naqinezhad et al., 2008;Jafari et al., 2013), high precipitation (Heshmati, 2007) and mild climate due to low altitude (Beck et al., 2018).The foreststeppes are characterised by their mosaic vegetation: in some areas there are quite dense wet forest fragments and arid steppe spaces between them.Precipitation is generally low (Heshmati, 2007).Due to the high altitude, there is a high annual temperature range Beck et al., 2018).Desert basins are characterised by sparse vegetation and low precipitation.Daily and annual temperature differences there are quite high (Breckle, 2002).In this regard, it can be concluded that the existing distribution of species can be explained ( not only by altitude, but also by preferred humidity level and vegetation.Available data of distribution suggest that D. chlorogaster and D. caspica prefer wet habitats with dense vegetation, while D. defilippii and D. schaekeli prefer dry, sparsely vegetated but more rocky biotopes.D. raddei, apparently, can occupy an intermediate position between the presented species, but at the moment there is not enough data for accurate conclusions. It is important to note that the map presented in Safaei-Mahroo et al. (2015) was built using data of Olson et al. obtained in 2001.Therefore, the distribution of ecoregions at the moment of our data collection (2022) can be significantly different due to the desertification (Cherlet et al., 2018).In particular, the localities we found (Fig. 3) are located in the forest zone, but this zone, according to our observations, is a forest-steppe.For this reason, the range of D. defilippii could be shifted to the northwest.
Although studied species have some differences in their preferred biotopes, the borders between them are very fuzzy.Thus, further research should be aimed at clarifying the boundaries of the ranges of species and studying the reasons for such distribution.

CONCLUSIONS
1. We confirmed that the species D. defilippii has a high genetic variability mediated by geographic isolation.
2. D. chlorogaster is one of the most genetically variable among studied species.This may be explained by the possible presence of a nuclear mitochondrial ND4 -Leu tRNA fragment pseudogene.
3. Despite the wide distribution range and diverse habitat preferences D. raddei has a low variability of the ND4 -Leu tRNA fragment of mDNA.
4. The range of the species D. defilippii is broader than previously suggested.
rogaster is also represented in several localities in Gilan province.However, genetic variability (based on uncorrected -distances) within is two times stronger than in (2 and 1% respectively) ( able ).

Fig. 1 .Table 2 .Fig. 2 .
Fig. 1.Dendrogram showing phylogenetic relationships between rock lizards in different regions of Iran, based on the sequences of the ND4 -Leu tRNA fragment.Values over and under the nodes represent posterior probabilities by BI/bootstrap values by ML.The localities for each specimen are shown to the right of the vertical thin black line
D. schaekeli D. raddei and .belongs to the other clade Raddei, it has a large distances to other species and occupies basal position on trees topology (Fig .1, 2).Generally, is a monophyletic group (1/98 support s D. raddei Localities from Mazandaran province are divided separately -with individuals caught by us in Darreh-ye Malek Darreh and individuals from GenBank from Savasraeh (KF717317) and Kiasar (KF717316) which are the most distant from the others.