A genome-wide investigation of microsatellite mismatches and the association with body mass among bird species

Introduction

It has long been recognized that the molecular evolutionary rates always covary with many life history traits of animals. Numerous studies have documented a negative relationship between the rate of molecular evolution and body mass (Nabholz, Glémin & Galtier, 2008; Bromham, 2011; Amos & Filipe, 2014), where genes in small-bodied species are likely to evolve faster than those in large-bodied species. This has been readily explained by higher number of mutation opportunities per unit time (generation length hypothesis, Li et al., 1996) or higher mutation probability in a round of DNA replication due to higher metabolic rate (metabolic rate hypothesis, Mindell et al., 1996) in small-bodied species. Although the precise reason for the pattern is not clear at present, to determine the universality of this pattern, we need to study additional form of genetic mutation besides mitochondrial DNA or nuclear ‘genes’ which are most frequently used. The first to consider is the fastest evolving components of the genome such as microsatellites.

Microsatellites, also known as simple sequence repeats (SSRs), are tandem repeats of simple nucleotide motifs, which have wide coverage in eukaryotic and prokaryotic genomes (Tóth, Gáspári & Jurka, 2000; Ellegren, 2004; Adams et al., 2016). One feature of microsatellites is that they have a high mutation rate (10⁻⁷ to 10⁻³ mutations per locus per generation), leading to high heterozygosity and extensive length polymorphisms (Kruglyak et al., 2000). It has long been assumed that the major cause of variation of microsatellite repeats is replication slippage (Kornberg et al., 1964; Bhargava & Fuentes, 2010), which will increase or decrease repeat copy numbers in microsatellites. Specifically, when it creates a loop in one of strands, a slippage error occurs. If the loop is formed in the replicating strand, it will introduce an insertion. If the loop is in the template strand, a deletion will emerge. Several mathematical models of microsatellite evolution have been proposed to represent the mutation processes of microsatellites, such as stepwise mutation model (SMM) of Ohta & Kimura (1973), which suggests that mutation in microsatellite loci occurs by one repeat unit at a time.

Many studies on microsatellites have explored the frequencies, abundance and polymorphism of microsatellites in the genomes (Wang et al., 2014; Qi et al., 2015; Adams et al., 2016). Few, if any, have correlated these microsatellite characters to the life history traits of a species. Specifically, microsatellites are hypothesized to experience a life cycle: start short (birth) and expand predictably due to mutation bias (expansion) until they become unstable and either collapse or degrade through internal point mutations (contraction and death) (Chambers & MacAvoy, 2000; Buschiazzo & Gemmell, 2006). Life history traits of species are expected to have an influence on the life cycle —‘birth’, expansion, and ‘death’—of microsatellites in the genome (Amos & Filipe, 2014). For example, in smaller species, higher mutation rate allows the ‘birth’ and expansion of microsatellites faster, due to higher mutation rate and slippage rate. Since the death rate is lower than the birth rate, microsatellites tend to accumulate in the genome (Buschiazzo & Gemmell, 2006). In that, the smaller species harbour a higher frequency of microsatellites across the genome, which has been proved in mammals (Amos & Filipe, 2014).

It is well known that except for repeat copy number variation, a microsatellite (e.g., ATATATATAT) also suffers from nucleotide substitutions and insertion/deletion mutations, hence becoming imperfect (e.g., ATATATCATAT: AT repeat with an insertion of C). Perfect and imperfect microsatellites are thus defined. It has been found that genomes possess a relatively small but significant number of imperfect microsatellites (Brinkmann et al., 1998). Mismatch variation of imperfect microsatellites is critical for their maintenance in the genome and imperfect microsatellites are more stable compared to perfect microsatellites since the former is less prone to slippage mutation (Sturzeneker et al., 1998). Several previous studies have already revealed the genome-wide motif imperfection pattern among species (e.g., Behura & Severson, 2015). Nevertheless, our understanding of motif mismatches in imperfect microsatellites is still very limited and their correlation with life history traits remains to be revealed and explained.

In this study, we used 65 avian genome sequences, employing SciRoKo (Kofler, Schlötterer & Lelley, 2007) to search SSRs in the whole genome. We chose avian genomes for this study because microsatellites have been widely used in population genetics of bird species, yet the pattern of microsatellites mismatches in birds is still not well understood, mostly owing to the lack of avian genomic information. With the advance of whole genome sequencing, evolution of microsatellites is attracting attention from researchers. With the genome-wide microsatellites data in hand, we presented the first detailed comparative study of microsatellites, aiming to reveal the patterns of motif mismatches across different bird species and to help understand the relationship with life history traits.

Materials & Methods

Results

Discussion

In the present study, we did a genome-wide search of microsatellites using SciRoKo with the same parameters to ensure that the program can search all possible microsatellites with the same probability for every genome. Microsatellites search results showed that the frequency of microsatellites varies extensively among species. We have also found a positive relationship between microsatellites abundance and genome size among 65 bird species, which is consistent with earlier studies (e.g., Hancock, 1996). After providing a general description of the basic characteristics of microsatellites, we particularly focused on comparing the motif mismatches of imperfect microsatellites to body mass across bird species in a phylogenetic context.

We found a negative relationship between body mass and the ratio of frequency of imperfect repeats (mismatch = 1) to perfect (mismatch = 0) ones with the same length 20bp among the species. Moreover, it is known that mutations in microsatellites shorter than a critical length are generally gain or loss of single repeat units which cannot disturb the repeat tract (Buschiazzo & Gemmell, 2006). Whereas when it reached a critical length, mismatch was introduced, a perfect microsatellite became imperfect. Here, our result implied that the introduction of motif mismatches in imperfect microsatellites is significantly associated with the nature of point mutation in microsatellites. In small-bodied species, since more perfect microsatellites suffer from the introduction of mismatches due to the higher mutation rate, a larger number of imperfect microsatellites relative perfect ones can be observed.

We observed that the microsatellites harbouring mismatches higher than expected have lower abundance than that carrying mismatches lower than expected. Consistent with this result, we also found that the microsatellites ≥30 bp and <3 mismatches have lower abundance than that ≥30 bp and >3 mismatches, indicating that mismatches of motifs is a key determinant leading to a paucity of long imperfect microsatellites in the genome. That is to say, mismatches would stabilize the repeat array and impede the further expansion. When the extent of mismatches reached saturation point, the repetition pattern is interrupted, leading the microsatellites to degeneration and death. (Taylor, Durkin & Breden, 1999; Harr & Schlotterer, 2000; Yamada et al., 2002; Vowles & Amos, 2006). Although the exact details of death is still poorly understood, the relative number of older mismatches in an ‘average’ microsatellite is likely to reflect the mutability during its lifetime. It can be further confirmed by the finding that the average mismatches of imperfect SSRs decreases with increasing body mass.

Our results that higher average mismatches of imperfect SSRs in small-bodied species support a correlation between mutation rate and life history traits. The pattern is usually explained by a generation length model, where smaller species evolve faster due to higher number of mutation opportunities per unit time (Li et al., 1996). In addition, body mass might affect the mutation rate through a link with metabolic rate and/or body temperature, which can directly change the mutation probability in a round of DNA replication (Mindell et al., 1996). Apart from these two key hypotheses, a rising hypothesis which proposes mutation rates are influenced by heterozygosity (Amos, 2010) can better explain the intrinsic correlation of motif mismatches with body mass. Smaller species have larger number of imperfect microsatellites which has been demonstrated by our data (β ± SE =  − 0.008 ± 0.002, t = 4.008, P < 0.001; R² = 0.203). Meanwhile, more heterozygous sites at these imperfect microsatellites can be expected. Recognition and ‘repair’ of heterozygous sites during synapsis will cause additional rounds of DNA replication which in turn provide more opportunities for mutations (Amos, 2011) and introduce more motif mismatches at imperfect microsatellite sites. Therefore, a negative relationship between body mass and motif mismatches can be observed. We suggest that heterozygote instability hypothesis, which is supported by increasing evidence (Drake, 2007; Masters et al., 2011; Amos, 2013; Amos, 2016), could provide a potential link between body mass and motif mismatches. However, further studies are needed in order to examine carefully whether homologous imperfect microsatellites are generally more prone to introduce mismatches in smaller species with a detail comparison between sister species.

Conclusions

In conclusion, the present study is the first effort to explore the motif mismatch patterns of microsatellites in birds on a genomic level. The results we obtained from this study provide support for the long-standing correlation between mutation rate and life history traits and suggest that a negative body mass trend in mutation rate may be a general pattern of avian molecular evolution.

Supplemental Information