Mitogenomic Insights into Adaptive Evolution of African Ground Squirrels in Arid Environments

Abstract

African ground squirrels (Xerus spp.), the inhabitants of African arid zones, face extreme heat and water scarcity driving selection for metabolic optimization. We assembled and annotated the first mitogenomes of Xerus inauris and Xerus rutilus (16,525–16,517 bp), revealing conserved vertebrate architecture with genus-specific traits. Key features include Xerus rutilus’s elongated ATP6 (680 vs. 605 bp), truncated ATP8–ATP6 spacers (4 vs. 43 bp), and tRNA-Pro control regions with 78.1–78.3% AT content. Their nucleotide composition diverged from that of related sciurids, marked by reduced T (25.78–26.9%) and extreme GC skew (−0.361 to −0.376). Codon usage showed strong Arg-CGA bias (RSCU = 3.78–3.88) and species-specific elevations in Xerus rutilus’s UGC-Cys (RSCU = 1.83 vs. 1.17). Phylogenetics positioned Xerus as sister to Ratufa bicolor (Bayesian PP = 0.928; ML = 1.0), aligning with African biogeographic isolation. Critically, we identified significant signatures of positive selection in key mitochondrial genes linked to arid adaptation. Positive selection signals in ND4 (ω = 1.8 × background), ND1, and ATP6 (p < 0.0033) correspond to enhanced proton gradient efficiency and ATP synthesis–molecular adaptations likely crucial for optimizing energy metabolism under chronic water scarcity and thermoregulatory stress in desert environments. Distinct evolutionary rates were observed across mitochondrial genes and complexes: Genes encoding Complex I subunits (ND2, ND6) and Complex III (Cytb) exhibited accelerated evolution in arid-adapted lineages, while genes encoding Complex IV subunits (COXI) and Complex V (ATP8) remained highly conserved. These findings resolve the Xerus mitogenomic diversity, demonstrating adaptive plasticity balancing arid-energy optimization and historical diversification while filling critical genomic gaps for this xeric-adapted lineage.

1. Introduction

The family Sciuridae, recognized as one of the most morphologically and ecologically diverse groups within Rodentia, has long been a critical focus in mammalian systematics due to its taxonomic classification and evolutionary dynamics. Comprising approximately 300 extant species distributed globally across terrestrial ecosystems (excluding Antarctica and certain oceanic islands), this family is categorized into the major subfamilies—Callosciurinae (tree squirrels), Xerinae (ground squirrels), and Pteromyinae (flying squirrels)—based on cranial morphology, dentition characteristics, and ecological adaptations [1,2]. Squirrels play indispensable roles in ecological processes: Sciurus carolinensis (North American gray squirrel) significantly influences deciduous forest succession through acorn caching [3], while the burrow systems of African Xerus spp., evolved as critical behavioral adaptations to extreme aridity, provide essential microclimatic refuges that reduce evaporative water loss and thermoregulatory costs during diurnal heat extremes, thereby conserving metabolic energy resources under chronic water scarcity [4,5]. Notably, certain species exhibit extraordinary adaptations to extreme aridity; for instance, Xerus inauris (South African striped ground squirrel) completes its life cycle in the Namib Desert with annual precipitation below 200 mm, establishing it as an ideal model for studying mammalian arid adaptation mechanisms [6,7].

While traditional molecular studies have advanced our understanding of sciurid phylogeny, critical gaps persist in resolving evolutionary relationships among arid-adapted lineages such as Xerus [8]. This is exemplified by mitochondrial cytochrome b-based phylogenies revealing complex parallel evolution within the arid-adapted subfamily Xerinae [9]. Despite insights from diversification models [10,11], a key methodological limitation remains, i.e., the absence of family-wide phylogenomic reconstructions using complete mitogenomes—particularly for xeric-adapted clades—impedes precise inference of macroevolutionary patterns and the origin of arid-adaptive traits [12,13].

In arid adaptation research, physiological ecologists have documented phenotypic evidence: sciurids achieve water efficiency via urine concentration and reduced basal metabolic rates, coupled with seasonal torpor strategies, to survive arid phases in grasslands [14]. Research across diverse arid-adapted mammals underscores the multifaceted nature of this challenge. Studies on desert rodents have identified adaptations ranging from behavioral strategies to molecular modifications in both nuclear and mitochondrial genes [15,16,17]. For example, molecular adaptations in mitochondrial Complex I genes enhancing oxidative phosphorylation efficiency [6,18,19] reflect conserved mechanisms for metabolic optimization under water scarcity, though their functional integration requires nuclear–mitochondrial coevolution [20,21]. Similarly, investigations in other taxa, such as camelids (Camelus dromedarius), have revealed coordinated changes in nuclear-encoded enzymes involved in metabolic pathways crucial for water and energy conservation [22]. Our approach leverages the mitogenome’s properties to detect lineage-specific adaptive evolution, particularly targeting genes under positive selection that optimize mitochondrial function—such as proton translocation efficiency and ATP synthesis—under the extreme energetic constraints imposed by aridity and heat stress. These findings highlight the potential advantages of mitogenomes for studying aspects of arid adaptation—mitochondrial genes directly encode core components of the electron transport chain, making them potentially sensitive to selective pressures related to energy metabolism under water scarcity. However, it is crucial to emphasize that mitochondrial function itself is profoundly shaped by the nuclear genome, which encodes most mitochondrial proteins and regulates their biogenesis, dynamics, and metabolic integration [20]. Complex physiological adaptations, such as those enabling survival in arid environments, inherently involve co-adapted interactions between nuclear and mitochondrial genomes over evolutionary time [23]. Therefore, while acknowledging the essential role of nuclear–mitochondrial coevolution, our analysis specifically targets the mitochondrial genome as a sensitive indicator of selective pressures acting on the core machinery of cellular energy production—a system under intense optimization pressure in arid-adapted mammals like Xerus. Additionally, the maternal inheritance and lack of recombination of mitogenomes simplify the detection of lineage-specific adaptive signatures [24], while their high mutation rate allows for the identification of recent selective sweeps associated with rapid environmental changes [25]. Therefore, while mitochondrial genome evolution, as a central regulator of cellular energetics, may reflect important facets of adaptation to energetic constraints in arid habitats, it represents only one component within a broader co-evolutionary framework governed primarily by the nuclear genome. Although mitochondrial markers have been used to resolve partial sciurid phylogenies [12], no study has systematically explored the nexus between mitogenomic adaptive evolution and arid ecological specialization within this integrated context.

Mitogenomes, characterized by maternal inheritance, conserved gene content, and elevated evolutionary rates, serve as powerful tools for deciphering species’ evolutionary histories and adaptive mechanisms [26]. Nevertheless, current research predominantly targets single species or specific genes, lacking systematic comparisons of selection pressures between arid-adapted and non-adapted clades within robust phylogenetic contexts. Few studies integrate multidimensional evidence—such as structural variations and codon usage bias—to dissect the molecular basis of adaptive evolution.

This study establishes the first integrative framework to resolve mitogenomic adaptations underlying extreme arid specialization in sciurids. By contrasting arid-adapted Xerus with non-adapted clades, we aim to decode evolutionary innovations for heat/water scarcity challenges through three primary objectives: (1) reconstructing a robust mitogenome-based phylogeny of Sciuridae, (2) identifying molecular signatures of positive selection associated with arid adaptation, (3) characterizing structural and functional implications of these adaptations.

2. Materials and Methods

2.1. Mitogenome Assembly and Structural Characterization

Based on the raw data of Xerus genome sequencing (X. inauris, SRX4562114 and X. rutilus, SRX12373276) that we obtained from the NCBI database, we assembled the two new complete mitochondrial genomes using NovoPlasty v4.3.1 [27,28] with a k-mer length of 39 and three iterative rounds to ensure circular genome integrity. Annotation was conducted via the MITOS2 web server [29], employing the Vertebrate Mitochondrial Code for automated identification of open reading frames and tRNAs, followed by manual calibration of gene boundaries through alignment with annotations from closely related species. Genome circular maps were generated using OGDRAW v1.3.1 [30], with visualization optimized by adjusting gene color codes, GC content gradients, and noncoding region highlights. Structural features—including gene order, overlapping regions, and intergenic spacers—were statistically analyzed using custom R v4.2.1 scripts incorporating the ape, ggplot2, and tidyverse packages to quantify gene length, GC skew, and noncoding region variation across all 28 mitogenomes. The two mitogenomes (X. inauris and X. rutilus) were deposited in Supplementary Materials File S1.

All 28 mitogenomes analyzed in this study (

Table 1. Mitogenomes analyzed in this study, including genomic features and habitat groups.

Species	Accession Number	AT%	AT_Skew	GC%	GC_Skew	Group
Callospermophilus lateralis	NC_031210.1	63.19	0.022	36.81	−0.317	Arid
Cynomys leucurus	NC_026705.1	62.97	0.019	37.03	−0.308
Cynomys ludovicianus	NC_026706.1	62.63	0.020	37.37	−0.302
Lctidomys tridecemlineatus	NC_027278.1	63.68	0.008	36.32	−0.298
Spermophilus alashanicus	NC_071768.1	64.6	0.008	35.4	−0.300
Spermophilus citellus	NC_059784.1	64.59	0.015	35.41	−0.302
Spermophilus dauricus	NC_027283.1	63.12	0.017	36.88	−0.321
Spermophilus taurensis	NC_079844.1	64.63	0.011	35.37	−0.299
Urocitellus parryii	NC_059785.1	63.31	0.014	36.69	−0.302
Urocitellus richardsonii	NC_031209.1	63.21	0.013	36.79	−0.300
X. inauris	Supplementary Materials File S1	58.18	0.075	41.82	−0.361
X. rutilus	Supplementary Materials File S1	57.58	0.105	42.42	−0.376
Lariscus insignis	NC_030070.1	61.08	0.031	38.92	−0.327	Non-Arid
Hadrosciurus igniventris	NC_050027.1	61.94	0.011	38.06	−0.312
Hylopetes phayrei	NC_026443.1	62.63	0.032	37.37	−0.322
Glaucomys volans	NC_050026.1	62.39	0.028	37.61	−0.332
Guerlinguetus brasiliensis	NC_050010.1	62.58	0.004	37.42	−0.307
Exilisciurus exilis	NC_030072.1	59.69	0.061	40.31	−0.325
Callosciurus finlaysonii	NC_035817.1	60.6	0.036	39.4	−0.338
Dremomys rufigenis	NC_026442.1	60.89	0.065	39.11	−0.358
Marmota himalayana	NC_018367.1	63.48	0.013	36.52	−0.301
Microsciurus mimulus	NC_050020.1	62.2	0.006	37.8	−0.303
Petaurista hainana	NC_023089.1	60.38	0.052	39.62	−0.328
Pteromys volans	NC_019612.1	62.57	0.029	37.43	−0.328
Ratufa bicolor	NC_023780.1	60.32	0.051	39.68	−0.336
Sciurus vulgaris	NC_002369.1	62.97	0.020	37.03	−0.322
Sundasciurus brookei	NC_035812.1	61.98	0.045	38.02	−0.336
Tamiops swinhoei	NC_026875.1	61.3	0.065	38.7	−0.361

Total: 28 mitogenomes (28 species: 12 arid-adapted, 16 non-arid). Newly X. inauris and X. rutilus genomes marked as ‘Supplementary Materials File S1’.

) represent 28 sciurid species, including 12 arid-adapted and 16 non-arid species based on habitat classification. This classification into Arid and Non-Arid groups was explicitly designed to enable comparative analyses of mitogenomic evolution associated with adaptation to water scarcity. The Arid group (annual precipitation < 500 mm) primarily inhabits deserts and grasslands, while the Non-Arid group (>1000 mm) occupies forests and wetlands, based on WorldClim [31], IUCN Red List [32], and GBIF distribution data [33].

The mitochondrial genome accession numbers of Xerus are presented in Table 1. The composition skew values were calculated according to the following formulas: AT skew [(A − T)/(A + T)] and GC skew [(G − C)/(G + C)].

2.2. Relative Synonymous Codon Usage (RSCU)

For the codon usage bias analysis, all protein-coding genes (PCGs) from the 28 mitogenomes were extracted. The frequency of synonymous codons for each species’ 13 PCGs was calculated using the seqinr package in R v4.2.1 [34]. RSCU values were computed using the standard formula:

$$RSC{U}_{ij}={\displaystyle \frac{X_{ij}}{\frac{1}{n_j}\mathsf{\Sigma }{X}_{\ddot{y}}}}$$

where X_ij represents the occurrence count of the j-th codon for the i-th amino acid, and n_j denotes the total number of synonymous codons for that amino acid [35]. The codon usage patterns were visualized via the ggplot2 to highlight group-specific features. Patterns potentially linked to functional adaptations, such as those relevant to energy metabolism or oxidative stress response in arid environments, were specifically examined.

2.3. Phylogenetic Reconstruction

The phylogenetic reconstruction was based on concatenated sequences of 13 mitochondrial PCGs from all 28 mitogenomes. Each gene was aligned using MAFFT v7.526 [36], followed by quality filtration with Gblocks v0.91b [37] under default parameters to remove poorly aligned regions; gaps were treated as missing data. Maximum likelihood (ML) analysis was performed in IQ-TREE v2.2.0 [38] under the optimal nucleotide substitution model (GTR+F+I+G4) selected by ModelFinder. Node support was assessed using 1000 ultrafast bootstrap replicates. Bayesian inference (BI) was implemented in MrBayes v3.2 with two independent Markov chain Monte Carlo (MCMC) chains, each run for 1 × 10⁷ generations, sampled every 1000 generations [39]. This robust phylogeny served as the essential framework for subsequent analyses of lineage-specific adaptive evolution, particularly positive selection associated with arid adaptation.

2.4. Positive Selection Analysis

The positive selection analysis was conducted using the codeml module in PAML v4.9 [40]. Branch models were employed to detect evolutionary patterns specifically associated with arid adaptation, with the phylogenetic tree as the framework. Species inhabiting arid environments were defined as foreground branches (Table 1, Arid group), while others served as background branches. A null model (Model 0, global single ω ratio) and an alternative model (allowing independent ω ratios for foreground branches) were compared via likelihood ratio tests (LRTs) to identify positive selection signals (significance threshold: p < 0.05). This approach specifically tests for accelerated evolution (dN/dS > 1) indicative of positive selection on the arid-adapted branches. To further resolve branch-specific selection pressures, a free-ratio model was implemented, permitting independent dN/dS (ω) values for each branch. Significantly positively selected sites were identified using Bayesian empirical Bayes (BEB) methods (posterior probability > 95%). All analyses employed the F3×4 codon frequency model, with initial parameters guided by neighbor-joining trees. MCMC chains were run three times to ensure stability, and computational reliability was verified using CodeML’s default convergence criteria (log-likelihood difference < 0.01). The primary goal of this analysis was to identify mitochondrial genes under positive selection that contribute to the adaptation of Xerus and other arid-adapted sciurids to their challenging environments.

3. Results

3.1. Mitochondrial Genome Annotation

The mitochondrial genomes of two ground squirrels (X. inauris and X. rutilus) exhibited typical vertebrate structures, with lengths of 16,525 bp and 16,517 bp, respectively (Figure 1). Both genomes contained 37 genes, including 13 protein-coding genes (PCGs), 22 tRNA genes, and 2 rRNA genes. The mitochondrial genomes of two ground squirrels exhibited typical vertebrate structures (Figure 1), with conserved gene content but some length variations in protein-coding genes. ATP6 showed a 75 bp extension in X. rutilus compared to X. inauris (680 vs. 605 bp), potentially influencing proton transport efficiency under arid stress. Start and stop codon usage showed conserved patterns with some exceptions (

Table 2. Characteristics of the mitochondrial genome (X. inauris: Xi, X. rutilus: Xr).

Gene	Nucleotide Positions				Size (bp)		Strand	Codon
	Xi		Xr					Xi		Xr
	Start	End	Start	End	Xi	Xr	Xi/Xr	Initiation	Amber	Initiation	Amber
tRNA-Phe	1	68	1	68	68	68	+
12SRNA	69	1041	69	1042	973	974	+
tRNA-Val	1042	1109	1043	1110	68	68	+
16SRNA	1110	2684	1111	2688	1575	1578	+
tRNA-Leu	2685	2759	2689	2763	75	75	+
ND1	2764	3719	2768	3723	956	956	+	ATG	TA	ATG	TA
tRNA-Ile	3720	3787	3724	3792	68	69	+
tRNA-Gln	3785	3856	3790	3861	72	72	-
tRNA-Met	3862	3930	3866	3934	69	69	+
ND2	3931	4972	3935	4976	1042	1042	+	ATC	T	ATC	T
tRNA-Trp	4973	5042	4977	5045	70	69	+
tRNA-Ala	5046	5114	5049	5117	69	69	-
tRNA-Asn	5132	5204	5136	5207	73	72	-
tRNA-Cys	5236	5302	5238	5304	67	67	-
tRNA-Tyr	5303	5368	5305	5370	66	66	-
COXI	5377	6918	5379	6920	1542	1542	+	ATG	TAG	ATG	TAA
tRNA-Ser	6921	6989	6923	6991	69	69	-
tRNA-Asp	6993	7061	6995	7063	69	69	+
COXII	7062	7745	7064	7747	684	684	+	ATG	TAA	ATG	TAA
tRNA-Lys	7748	7814	7750	7817	67	68	+
ATP8	7816	8022	7819	8025	207	207	+	ATG	TAG	ATG	TAG
ATP6	8052	8656	7980	8659	605	680	+	ATG	TA	ATG	TA
COXIII	8657	9440	8660	9443	784	784	+	ATG	T	ATG	T
tRNA-Gly	9441	9510	9444	9514	70	71	+
ND3	9511	9857	9515	9861	347	347	+	ATT	TA	ATT	TA
tRNA-Arg	9858	9926	9862	9930	69	69	+
ND4L	9928	10224	9932	10,228	297	297	+	ATG	TAA	ATG	TAA
ND4	10,218	11,595	10,222	11,599	1378	1378	+	ATG	T	ATG	T
tRNA-His	11,596	11664	11,,600	11,668	69	69	+
tRNA-Ser	11,665	11,724	11,669	11,728	60	60	+
tRNA-Leu	11,725	11,794	11,729	11,798	70	70	+
ND5	11,795	13,612	11,799	13,616	1818	1818	+	ATC	TAA	ATC	TAA
ND6	13,596	14,126	13,600	14,130	531	531	-	AGA	T	AGA	T
tRNA-Glu	14,121	14,189	14,125	14,193	69	69	-
Cytb	14,193	15,332	14,197	15,336	1140	1140	+	ATG	TAA	ATG	TAG
tRNA-Thr	15,333	15,400	15,337	15,405	68	69	+
tRNA-Pro	15,405	15,472	15,410	15,478	68	69	-

). While most genes used ATG as the start codon, ND3 and ND6 employed alternative initiation codons (ATT and AGA, respectively). Stop codon usage was more variable, with COXI exhibiting species-specific termination signals and ND6 consistently using incomplete stop codons. Intergenic regions also displayed notable length variations between species, particularly in the ATP8–ATP6 spacer and ND4L–ND4 overlap regions. Nucleotide composition exhibited distinct AT-rich patterns, with the tRNA-Pro control region showing exceptionally high AT content (78.1% in X. inauris; 78.3% in X. rutilus), significantly exceeding the genomic average (~65%) (Table 2), potentially stabilizing replication under the high-temperature regimes of their arid habitats.

Both species exhibited distinct nucleotide composition patterns relative to other sciurids. While their adenine content was consistent with related species, thymine content was notably reduced, resulting in elevated AT skew values that surpassed most relatives. The guanine–cytosine composition showed particularly pronounced guanine depletion, with cytosine content exceeding typical sciurid ranges and producing strongly negative GC skew. These patterns contrasted with the more moderate skew values observed in Callosciurus and Cynomys, underscoring the unique nucleotide bias characteristic of the Xerus species (Table 1). This distinct genomic signature may reflect selective pressures related to DNA stability or replication efficiency in their extreme environment.

3.2. Relative Synonymous Codon Usage (RSCU)

The RSCU analysis demonstrated both conserved and species-specific codon usage patterns. Both species exhibited strong preferences for certain codons (Arg-CGA, Leu-CUA) while completely avoiding others (Arg-AGA/AGG), with UGA exclusively serving as the stop codon (Figure 2). The comparative analysis of arid-adapted squirrel species and non-arid-adapted squirrel species revealed that the AGA/AGG avoidance phenomenon (p = 0.008) and the UGA termination advantage (p < 0.01) are common adaptive characteristics in species groups adapted to arid environments, while the unique UGC (cysteine) usage rate in X. rutilus (relative significance coefficient RSCU = 1.83) is significantly higher than that of other arid-adapted species (1.14–1.31; p = 0.003) (Figure 3). Interspecific variations were observed in cysteine and serine codon usage, while glutamine and lysine displayed bimodal distributions. Notably, proline and threonine codon preferences showed species-specific differences, which may represent fine-tuned adaptations in osmoregulation (proline) and protein-mediated energy homeostasis (threonine) under distinct ecological pressures associated with their arid niches (Figure 2).

3.3. Phylogenetic Reconstruction

The maximum likelihood (ML, Figure 4) and Bayesian inference (BI, Figure 5) analyses revealed highly concordant topologies with strong nodal support across most clades. Both methods robustly supported the monophylicity of Xerus (Bayesian posterior probability, PP = 1.0; maximum likelihood bootstrap support, BS = 1.0), with branch lengths (substitutions per site) to X. inauris and X. rutilus of 0.097 and 0.072 (BI) and 0.083 and 0.066 (ML), respectively. The analyses placed Xerus as sister to a clade containing Canis lupus familiaris and Ratufa bicolor with moderate support (Bayesian posterior probability, PP = 0.829; maximum likelihood bootstrap support, BS = 0.829). Within sciurids, North American ground squirrels formed a cohesive radiation: Cynomys species diverged at branch lengths of 0.020–0.026 (BI) and 0.021–0.026 (ML) substitutions per site, while Urocitellus species showed recent divergence (BI: 0.0117–0.0118 substitutions per site; ML: 0.0119–0.0120 substitutions per site). Notably, both methods consistently resolved Sciurus vulgaris and Spermophilus dauricus as sister taxa with maximum support (Bayesian posterior probability, PP = 1.0; maximum likelihood bootstrap support, BS = 1.0). Divergence events associated with Beringian migrations were evident in Callospermophilus lateralis (BI: 0.090 substitutions per site; ML: 0.072 substitutions per site) and Marmota himalayana (BI: 0.086 substitutions per site; ML: 0.080 substitutions per site), while Sunda Shelf dynamics influenced Southeast Asian radiations (Exilisciurus exilis: BI 0.202 substitutions per site; ML 0.157 substitutions per site). This congruent phylogenetic framework provides robust evolutionary context for identifying lineage-specific adaptations (Exilisciurus exilis: BI 0.202; ML 0.157).

3.4. Selection Pressure Analysis

Branch-model tests specifically contrasting arid-adapted lineages against background branches detected strong positive selection signals in key mitochondrial genes functionally linked to energy metabolism: ND4 (LR = 12.14, p < 0.001), ND1 (LR = 11.02, p < 0.001), ND2 (LR = 10.66, p < 0.001), and ATP6 (LR = 12.14, p < 0.001) (Figure 5, Table S1). Notably, ND4 in Xerus exhibited a 1.8-fold ω increase, which likely reflects adaptations in proton transport efficiency to meet heightened energy demands under the chronic water scarcity and thermoregulatory stress of arid conditions. ATP6 evolved rapidly in arid groups, potentially optimizing oxidative phosphorylation efficiency under water scarcity, while ATP8 remained conserved, indicating complementary functional constraints. Free-ratio models revealed modular rate shifts: Cytb and ND2/ND6 showed branch-specific heterogeneity (Figure 6), with elevated ω in cold-adapted (Urocitellus), gliding (Petaurista), and arid-adapted lineages (Xerus), likely due to divergent thermal or metabolic needs, whereas ATP8 remained strictly conserved (p = 0.39), underscoring its essential structural role (Figure 6, Table S1). These results provide strong evidence for positive selection acting on core mitochondrial energy production genes specifically in lineages adapted to arid environments.

4. Discussions

4.1. Mitochondrial Genome Annotation and Structural Specificity

The mitogenomes of X. inauris and X. rutilus exhibit the typical vertebrate structural framework but display genus-specific variations. Core gene lengths (COXI: 1542 bp; ND4: 1378 bp; Cytb: 1140 bp) remain conserved within Sciuridae. The 75 bp extension in the X. rutilus ATP6 gene elongates its C-terminal domain, representing a convergent trait with arid-adapted taxa like Spermophilus. This modification may optimize proton translocation during dehydration by expanding the channel geometry to reduce steric hindrance for hydrated proton flux, while repositioned residues could stabilize the proton-conducting water network. These adaptations appear to sustain ATP synthase activity under aridity, potentially maintaining energy production for elevated metabolic demands [41,42,43]. Intergenic regions show genus-specific dynamics: the ATP8–ATP6 spacer in X. rutilus is markedly shorter (4 bp vs. 43 bp in X. inauris), while the ND4L–ND4 overlap expands from 6 bp to 8 bp, suggesting selection-driven genome compaction [44]. The tRNA-Pro control region exhibits extreme AT enrichment (78.1–78.3%), significantly exceeding the genomic average (~65%), likely stabilizing replication origins under high-temperature stress [45]. Nucleotide composition further distinguishes Xerus: AT skew (0.075–0.105) exceeds that of closely related sciurids (<0.06), driven by reduced T content (25.78–26.9% vs. family average 28.6–31.8%); GC skew (−0.361 to −0.376) reflects extreme guanine depletion (G: 13.23–13.36%) and cytosine enrichment (C: 28.46–29.19%). This pattern is absent in Callosciurus (AT skew: 0.036) and Cynomys (GC skew: −0.30), highlighting Xerus’s adaptive fine-tuning within the conserved mitochondrial framework, potentially linked to environmental constraints of their arid habitats.

4.2. Codon Usage Bias and Metabolic Adaptation

The RSCU in Xerus reveals conserved preferences with species-level divergence. Both species exhibit extreme bias for Arg-CGA (RSCU = 3.78–3.88) and Leu-CUA (RSCU = 2.4–2.68) while completely avoiding AGA/AGG (Arg) codons, potentially optimizing translational efficiency under oxidative stress [30]. The exclusive use of UGA as the termination codon (RSCU = 3) further supports streamlined translation. Interspecific differences include elevated UGC (Cys) usage in X. rutilus (RSCU = 1.83 vs. 1.17 in X. inauris), which may enhance synthesis of cysteine-rich antioxidant proteins [31], and near elimination of UCG (Ser) in X. rutilus (RSCU = 0.06 vs. 0.2 in X. inauris), suggesting selective avoidance of rare tRNAs. Critically, comparative analysis with arid group (Spermophilus, Cynomys) reveals convergent adaptations: the elevated Cys-UGC bias (1.83 in X. rutilus vs. 1.14–1.31 in other desert taxa), systematic avoidance of AGA/AGG codons (RSCU = 0), and UGA termination codon dominance (RSCU ≈ 3) are significantly amplified in arid-adapted lineages compared to mesic species, providing direct evidence for molecular adaptation to oxidative stress. Bimodal distributions in Gln (CAA: 1.69–1.74; CAG: 0.26–0.31) and Lys (AAA: 1.82–1.9; AAG: 0.1–0.18) indicate environment-dependent codon optimization, while balanced variations in Pro (CCA: 0.92–1.09) and Thr (ACA: 1.57; ACC: 1.52) likely reflect species-specific metabolic demands (thermogenesis or lipid metabolism). Notably, overall codon preferences align with the extreme AT-biased nucleotide composition, facilitating efficient replication and transcription under arid conditions. In X. rutilus, the Cys-UGC characteristic of specific species has been enhanced, while it almost completely avoids the appearance of UCG. This indicates that this species has adapted to the intense ultraviolet radiation and water scarcity environment in arid habitats, possessing a sophisticated adaptation mechanism. In summary, these patterns represent a specific molecular adaptation to the energy and oxidative challenges in arid life, surpassing the potential mechanisms inferred from single-species studies.

4.3. Phylogenetic Concordance and Biogeographic Processes

The fully concordant BI and ML topologies confirm the monophy of Xerus. Xerus forms a moderately supported clade with Canis lupus familiaris and Ratufa bicolor. North American ground squirrels exhibit Pleistocene radiation patterns: Cynomys divergence aligns with habitat fragmentation [46], while recent Urocitellus divergence corresponds to glacial cycles [47]. The consistent sister relationship between Sciurus vulgaris and Spermophilus dauricus resolves previous topological uncertainties. Bering Land Bridge migrations [48] are reflected in Callospermophilus lateralis and Marmota himalayana divergences, while Sunda Shelf fluctuations shaped Exilisciurus exilis evolution [49]. These results underscore the significant role of biogeographic processes in sciurid diversification.

4.4. Positive Selection and Arid Adaptation

The central finding of this study is the detection of strong positive selection signals in key mitochondrial genes (ND4, ND1, ND2, ATP6) specifically within arid-adapted sciurid lineages, including Xerus. Branch model analyses detected strong positive selection signals in arid-adapted lineages for ND4 (LRT = 12.14), ND1 (LRT = 11.02), ND2 (LRT = 10.66), and ATP6 (LRT = 12.14). The ω of ND4 in Xerus exhibited a 1.8-fold increase, potentially enhancing proton gradient efficiency to meet heightened ATP synthesis demands under the combined stresses of water scarcity and high temperatures [41]. Cytb showed global heterogeneity (LRT = 128.29) but weak selection signals in arid lineages (LRT = 4.89), suggesting convergent selection across ecotypes [50]. Accelerated evolution of ATP6 in arid taxa—including the observed elongation in X. rutilus—contrasted with conserved ATP8, reflecting subunit-specific functional optimization of the ATP synthase complex for energy production under aridity [51]. Free-ratio models confirmed modular evolutionary patterns: Cytb and ND2/ND6 displayed branch-specific rate heterogeneity (LRT = 128.29–114.89), with elevated ω values in cold-adapted (Urocitellus), gliding (Petaurista), and arid-adapted lineages (Xerus), while ATP8 and COXI remained strictly conserved (p > 0.8), highlighting the stability of core respiratory complexes. These patterns align with mitochondrial adaptation strategies documented in other arid-adapted mammals like camels [52] and jerboas [53], supporting conserved mechanisms for optimizing mitochondrial energy metabolism under water scarcity across phylogenetically diverse taxa. While branch models identified strong positive selection signals in ATP6 and ND genes, two critical constraints merit emphasis: (1) The functional consequences of positively selected sites (e.g., ATP6 elongation) require experimental validation (e.g., CRISPR-based mutagenesis assays) to confirm their roles in thermotolerance. (2) Our exclusive reliance on mtDNA precludes analysis of mito-nuclear coadaptation—nuclear genomic data (e.g., OXPHOS genes) are essential to resolve compensatory mechanisms. Nevertheless, the identification of significantly elevated positive selection signals in core mitochondrial energy production genes specifically in arid-adapted lineages relative to background branches suggests that mitogenomic adaptation is a key component of the evolutionary response to extreme aridity in sciurids.

4.5. Limitations and Methodological Considerations

While this study provides compelling evidence for mitogenomic adaptation to aridity in sciurids, several methodological limitations warrant consideration. First, our definition of “arid-adapted” species relies on macroclimate proxies derived from distribution-wide climate data, which serve as standardized proxies for comparative analyses at a macroecological scale. However, these proxies do not fully capture species’ microhabitat preferences, behavioral adaptations (diel activity patters or active foraging strategies in riparian zones within arid landscapes), or physiological tolerance thresholds (e.g., species reliant on hibernation or estivation to bypass extreme dehydration periods). Species such as Cynomys spp. inhabiting dry grasslands but exploiting subterranean microclimates or water-rich foraging areas may thus exhibit adaptations not fully aligned with macro-scale aridity metrics. Second, our analysis focused exclusively on mitochondrial genomes. Mito-nuclear coadaptation, particularly in OXPHOS complexes involving nuclear-encoded subunits, is critical for functional optimization under environmental stress; future studies incorporating nuclear genomic data are required to resolve compensatory mechanisms across gene compartments. Third, the functional significance of candidate sites under selection (e.g., the extended ATP6 domain in X. rutilus) remains statistically inferred. Experimental validation through molecular assays (e. g., CRISPR-based mutagenesis testing proton translocation efficiency under dehydrating conditions) is necessary to confirm the phenotypic consequences of these molecular adaptations. Despite these constraints, the convergent patterns of selection in key mitochondrial genes across arid-adapted lineages remain robust indicators of evolutionary pressures associated with water scarcity and oxidative stress.