Comparative analysis of bat genomes provides insight into the evolution of flight and immunity
Associated Data
Abstract
Bats are the only mammals capable of sustained flight and are notorious reservoir hosts for some of the world’s most highly pathogenic viruses, including Nipah, Hendra, Ebola and SARS. To identify genetic changes associated with the development of bat-specific traits, we performed whole genome sequencing and comparative analyses of two distantly related bat species, fruit bat Pteropus alecto and insectivorous Myotis davidii. We discovered an unexpected concentration of positively selected genes in the DNA damage checkpoint and NF-κB pathways that may be related to the origin of flight, as well as expansion and contraction of important gene families. Comparison of bat genomes with other mammalian species has provided new insights into bat biology and evolution.
Bats belong to the order Chiroptera within the mammalian clade Laurasiatheria (1). While consensus has not been reached on the exact arrangement of groups within Laurasiatheria, a recent study placed Chiroptera as a sister taxon to Cetartiodactyla (whales + even-toed ungulates such as cattle, sheep and pigs) (2). The Black flying fox (Pteropus alecto), and David’s Myotis (Myotis davidii), represent the Yinpterochiroptera and Yangochiroptera suborders, respectively, and display a diverse range of phenotypes (Fig. 1). Captive colonies, immortalized cell lines and bat-specific reagents have been developed for these two species, however genomic data is currently unavailable.
P. alecto and M. davidii represent two distinct Chiropteran suborders and demonstrate diverse evolutionary adaptations. PNG: Papua New Guinea.
The most conspicuous feature of bats, distinguishing them from all other mammalian species, is the capacity for sustained flight. Positive selection in the oxidative phosphorylation (OXPHOS) pathway suggests increased metabolic capacity played a key role in its evolution (3), yet the byproducts of oxidative metabolism (such as reactive oxygen species, ROS) can produce harmful side effects including DNA damage (4). We hypothesize that genetic changes during the evolution of flight in bats likely included adaptations to limit collateral damage caused by byproducts of elevated metabolic rate. Another phenomenon that has sparked intense interest in recent years is the discovery that bats maintain and disseminate numerous deadly viruses (5). In this context, we further hypothesize that the long-term coexistence of bats and viruses must have imposed strong selective pressures on the bat genome, and the genes most likely to reflect this are those directly related to the first line of antiviral defense - the innate immune system.
We performed high-throughput whole-genome sequencing of individual wild-caught specimens of P. alecto and M. davidii using the Illumina HiSeq platform (6). More than 100x coverage high quality reads were obtained for P. alecto and M. davidii resulting in high quality assemblies (Tables S1–3, Fig. S1). The two bat genomes, at approximately 2 Gb, were smaller in size than other mammals (7) (Fig. S2), while the number of genes we identified was similar to other mammals (21,392 and 21,705 in P. alecto and M. davidii, respectively) (Fig. S3). Both species displayed a high degree of heterozygosity at the whole genome level (0.45% and 0.28% in P. alecto and M. davidii, respectively) (Tables S4–5), while repetitive content accounted for slightly less than a third of each genome (Tables S6–7). We identified a novel endogenous viral element derived from Saimiriine herpesvirus 2 that has expanded to 126 copies in P. alecto (Table S8, Fig. S4). Gene family expansion and contraction analysis (Tables S9–12) revealed significant expansion (p<0.05) of 71 gene families in M. davidii compared to only 13 in P. alecto, which may be related to a recent wave of DNA transposon activity (8).
We screened all nuclear-encoded bat genes to identify those for which a single orthologous copy was unambiguously present in both bat species as well as human, rhesus macaque, mouse, rat, dog, cat, cattle and horse. From this, 2,492 genes were used to perform maximum likelihood and Bayesian phylogenomic analysis (Figs. 2, S5–7). All phylogenetically informative signals including concatenated nucleotides and amino acids vigorously supported bats as a member of Pegasoferae (Chiroptera + Perissodactyla + Carnivora) (9), with the bat lineage diverging from the Equus (horse) lineage approximately 88 million years ago (MYA), supported by findings at the transcript level (10). Interestingly, phylogenetic reconstruction with mitochondrial DNA sequences resulted in bats occupying an outlying position in Laurasiatheria (Fig. S8). The incongruence between nuclear and mitochondrial trees likely reflects rapid evolution of the mitochondrial genome of the bat ancestor during the evolution of flight (3).
Maximum likelihood phylogenomic analysis of 2,492 genes from M. davidii, P. alecto and eight mammalian species. Divergence time estimates in blue, gene family expansion events in green, and gene family contraction events in red. MRCA: most recent common ancestor.
To identify mechanisms that facilitated the origin of flight in bats, we surveyed genes involved in detection and repair of genetic damage. A high proportion of genes in the DNA damage checkpoint/DNA repair pathway were found to be under positive selection in the bat ancestor, including ATM, DNA-PKc, RAD50, KU80, and MDM2 (Fig. 3A, Table 1). We propose that these changes may be directly related to minimizing/repairing the negative effects of ROS generated as a consequence of flight. Additionally in this pathway, TP53 (p53) and BRCA2 were shown to be under positive selection in M. davidii, while LIG4 was under positive selection in P. alecto (Table 1). Bat-specific mutations in a nuclear localization signal in p53 and a nuclear export signal in MDM2 (Figs. 3B, S9) may affect subcellular localization and function in both species (11, 12). Other candidate flight-related genes under positive selection in the bat ancestor included COL3A1, involved in skin elasticity, and CACNA2D1, which has a role in muscle contraction (Table S13).
(A) Positive selection in the DNA damage checkpoint/DNA repair pathway. Genes under positive selection in the bat ancestor are highlighted in orange. Genes under positive selection in M. davidii only (p53, BRCA2) or P. alecto only (LIG4) are highlighted in blue. (B) Mutations unique to bats were detected in the functionally relevant regions of the p53 nuclear localization signal (NLS) and MDM2 nuclear export signal (NES) (black highlight).
Table 1.
DNA damage checkpoint and innate immune genes under positive selection in the bat lineages
| Lineage | Symbol | Gene | ω0 (average) | ω1 (other) | ω2 (target) | p-value |
|---|---|---|---|---|---|---|
| Ancestor | TLR7 | toll-like receptor 7 | 0.2821 | 0.2670 | 2.7778 | 3.54E-07 |
| ATM | ataxia telangiectasia mutated | 0.20096 | 0.19595 | 0.7163 | 1.34E-05 | |
| MDM2 | Mdm2 p53 binding protein homolog (mouse) | 0.13358 | 0.12615 | 0.81085 | 4.05E-04 | |
| NLRP3 | NLR family, pyrin domain containing 3 | 0.1788 | 0.1714 | 1.1884 | 1.93E-04 | |
| MAP3K7 | mitogen-activated protein kinase kinase kinase 7 | 0.0216 | 0.0194 | 0.4786 | 8.93E-03 | |
| RAD50 | RAD50 homolog | 0.09657 | 0.09343 | 0.28882 | 7.95E-03 | |
| PRKDC | protein kinase, DNA-activated, catalytic polypeptide | 0.23036 | 0.22768 | 0.45155 | 6.80E-03 | |
| KU80 | X-ray repair complementing defective repair in Chinese hamster cells 5 | 0.31145 | 0.30436 | 0.91747 | 3.75E-02 | |
| c-REL | v-rel reticuloendotheliosis viral oncogene homolog (avian) | 0.2495 | 0.2403 | 1.5717 | 1.11E-02 | |
|
| ||||||
| P. alecto | TBK1 | TANK-binding kinase 1 | 0.0643 | 0.0522 | 0.2930 | 1.29E-09 |
| LIG4 | ligase IV, DNA, ATP-dependent | 0.12033 | 0.11376 | 0.24797 | 8.91E-04 | |
| IL18 | interleukin 18 (interferon-gamma-inducing factor) | 0.5298 | 0.4532 | 1.7647 | 2.66E-04 | |
| IFNG | interferon, gamma | 0.5010 | 0.4527 | 1.3282 | 4.89E-03 | |
| ISG15 | ISG15 ubiquitin-like modifier | 0.2069 | 0.1909 | 0.4387 | 2.63E-02 | |
| DDX58 | DEAD (Asp-Glu-Ala-Asp) box polypeptide 58 | 0.3040 | 0.2923 | 0.4661 | 1.23E-02 | |
|
| ||||||
| M. davidii | IFNAR1 | interferon (alpha, beta and omega) receptor 1 | 0.4954 | 0.4723 | 1.0924 | 7.00E-03 |
| TP53 | tumor protein p53 | 0.25623 | 0.23933 | 0.48123 | 7.00E-03 | |
| BRCA2 | breast cancer 2, early onset | 0.49002 | 0.47732 | 0.64213 | 1.31E-03 | |
| IRAK4 | interleukin-1 receptor-associated kinase 4 | 0.1670 | 0.1583 | 0.3531 | 1.96E-02 | |
The rate ratio ω of non-synonymous to synonymous substitutions (dN/dS) was calculated using multi protein alignments of P. alecto and M. davidii sequences with orthologous sequences from human, rhesus macaque, mouse, rat, dog, cattle and horse. ω0 is the average ratio in all branches, ω1 is the average ratio in non-bat branches, and ω2 is the ratio in the bat branch. A low p-value indicates that the ω2 model fits the data better than the ω1 model.
We next examined genes of the innate immune system (Table 1). Positively selected genes in the bat ancestor included c-REL, a member of the NF-κB family of transcription factors, which also contained amino acid changes potentially affecting IκB binding (Fig. S10). In addition to diverse roles in innate and adaptive immunity (13), c-REL plays a role in the DNA damage response by activating ATM (14) and CLSPN (15), while ATM is also an upstream regulator of NF-κB (16). The DNA damage response plays an important role in host defense and is a known target for virus interaction (17), raising the possibility that changes in DNA damage response mechanisms during selection for flight could have influenced the bat immune system.
Intriguingly, both P. alecto and M. davidii have lost the entire locus containing the PYHIN gene family, including AIM2 and IFI16; both of which are involved in sensing microbial DNA and the formation of inflammasomes (Fig. S11). The association between PYHIN genes and cell cycle regulation in other species (18) hints that loss of the PYHIN family in bats may be connected to changes in the DNA damage pathway; since at least one PYHIN gene is present in all other major groups of eutherian mammals (19). NLRP3, triggered by both viral infection and ROS in other mammals (20), plays an analogous role to AIM2 in inflammasome assembly and was also under positive selection in the bat ancestor (Table 1).
Natural killer (NK) cells provide a first line of defense against viruses and tumors and include two families of NK cell receptors; killer-cell immunoglobulin-like receptors (KIRs), encoded by genes in the leukocyte receptor complex (LRC) and killer cell lectin-like receptors (KLRs, also known as Ly49 receptors), encoded within the natural killer gene complex (NKC). KLRs and KIRs were entirely absent in P. alecto and reduced to a single Ly49 pseudogene in M. davidii (Table S14). KIR-like receptors identified in other species (21) were also absent from both P. alecto and M. davidii genomes, supported by transcript analysis in P. alecto (10). This likely indicates that bat NK cells use a novel class of receptors to recognize classical MHC class I molecules. Furthermore, additional LRC members of the immunoglobulin superfamily (including SIGLECS, LILRs, CEACAMs and LAIRs) have undergone considerable gene duplication in M. davidii and other mammals; yet have almost completely failed to expand in P. alecto (Fig. S12). As the genes encoded within the LRC bind a variety of ligands and play multiple roles in immune regulation, these observations have diverse implications for differences in immune function between P. alecto and M. davidii and between bats and other mammals.
We identified seven complete and two partial copies of the digestive enzyme RNASE4 in M. davidii (Table S15), while P. alecto RNASE4 has acquired a frame-shift mutation resulting in loss of catalytic residues (Fig. S13). We also identified critical amino acid changes in M. davidii RNASE4 genes (relative to the mammalian consensus) that suggest diversification of substrate specificity (Fig. S13). With a proven role in host defense against RNA viruses (22), RNASE4 expansion in M. davidii may have implications for virus resistance, but may also reflect the insectivorous diet of M. davidii, which contrasts that of P. alecto which consumes predominantly fruit, flowers and nectar.
M. davidii also differs from P. alecto in aspects including hibernation and echolocation (Fig. 1). Bile salt-stimulated lipase (BSSL), capable of hydrolyzing triglycerides into monoglycerides and subsequently releasing digestible free fatty acids, has been specifically expanded in M. davidii compared to P. alecto and other mammals (Fig. S14). In addition, we observed six candidate genes related to hibernation showing positive selection in M. davidii and three other hibernating species, relative to non-hibernators (Table S16). Seven echolocation related genes, including new candidates WNT8A and FOS (a subunit of the AP-1 transcription factor) had significantly higher dN/dS in the echolocating M. davidii branch relative to non-echolocating branches (Table S17). Of note, the third exon in M. davidii FOXP2 had even greater variation from the mammalian consensus than two previously identified variable sites (Fig. S15) suggesting a specific transcript variant is involved in echolocation (23).
In summary, comparative analysis of P. alecto and M. davidii genomes has provided insight into the phylogenetic placement of bats, and has revealed evidence of genetic changes that may have contributed to their evolution. Gene duplication events played a particularly prominent role in the evolution of Myotis bats and may have helped contribute to their speciation. Concentration of positively selected genes in the DNA damage checkpoint pathway in bats may indicate an important step in the evolution of flight, while evidence of change in components shared by the DNA damage pathway and the innate immune system raises the interesting possibility that flight-induced adaptations have had inadvertent effects on bat immune function and possibly also life expectancy (24). The data generated by this study will help to address major gaps in our understanding of bat biology and provide new directions for future research.
Supplementary Material
Supplementary Materials
Supplementary Tables
Acknowledgments:
We thank H. Field, C. Smith and M. Yu for helping source genomic DNA, K. Itahana and J.J. Boomsma for constructive discussion and M. Cowled for graphics assistance. We acknowledge financial support from the China National Genebank at Shenzhen, CSIRO (OCE Science Leaders Award, Julius Award), The Australian Research Council (FT110100234), State Key Program for Basic Research (2011CB504701), National Natural Science Foundation of China (81290341), and the Defense Threat Reduction Agency of USA. The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the U.S. Government. KBL and KGF are contractors for the U.S. Government. This work was prepared as part of their official duties. Title 17 U.S.C. §105 provides that ‘Copyright protection under this title is not available for any work of the United States Government.’ Title 17 U.S.C. §101 defines a U.S. Government work as a work prepared by a military service member or employee of the U.S. government as part of that person’s official duties. P. alecto and M. davidii genomes have been deposited at DDBJ/EMBL/GenBank under the accession numbers ALWS01000000 and ALWT01000000. Short read data have been deposited into the Short Read Archive under accession numbers SRA056924 and SRA056925. Raw transcriptome data has been deposited in Gene Expression Omnibus as GSE39933. Tree files and alignments have been submitted to TreeBASE under Study Accession URL: http://purl.org/phylo/treebase/phylows/study/TB2:S13654. We also wish to thank the editor and two anonymous reviewers for their helpful comments and suggestions.
Footnotes
The authors declare no competing financial interests.