Genome size increases with the amount of TE and repetitive DNA contained in the genome. Transposons interact positively and negatively with DNA repair systems, and can influence the type of DSB repair system employed when damage occurs (Izsvak et al. 2009). The DNA repair hypothesis explains different rates of molecular evolution in terms of varying efficiencies of DNA repair (Britten 1986; Baer et al. 2007). Accordingly, genome size can affect mutation rates and speciation rates at two levels: (1) negatively, by enhancing DNA repair efficiencies due to genome size-dependent checkpoint-mediated extension of S-phase (GT), as proposed here; and (2) positively, by increasing the probability of either adaptive or deleterious transposition events. The balance between these two forces might adjust the molecular clock in a species-specific manner. The first proposal can be easily tested in yeast, for example; while the second is garnering a growing consensus as the hypothesis of transposon-driven evolution (Kazazian 2004; Feschotte and Pritham 2007).
In yeast, Ty transposition, although independent of the checkpoint, is stimulated by checkpoint function (Curcio et al. 2007). Mutants defective in the S-phase regulatory factor Rtt101 recover poorly from checkpoint function (Zaidi et al. 2008), and experience bursts of transposition. Moreover, transposition events depend on RNR and increase when RNR is up-regulated during DNA damage and repair (O’Donnell et al. 2010). Indeed, Ty transposons can participate in NHEJ-directed DNA repair (Teng et al. 1996); and can mediate gene amplification when unscheduled DNA replication occurs (Green et al. 2010), indicating that a disrupted S-phase can result in elevated adaptive, or maladaptive, mutation rates. A corollary to the transposon hypothesis of evolution then is that many, if not all, adaptive radiations in eukaryotes reflect DNA damage events, and are ultimately consequences of adaptations in checkpoint-mediated DNA repair. The evolution of DNA repair systems underlies, in that sense, the evolution of species (Britten 1986).
The increase in mutation rate with replication timing in eukaryotes correlates with two well-established phenomena: (1) reduced efficiency in DNA repair systems in heterochromatin; and (2) transition to a distinct late-S replication program after euchromatin replication is completed. The shift to a late replicating regime in both higher and lower eukaryotes is likely due to the change in chromatin status between early and late replicating DNA. Cdk1/cyclinA2, for example, might be required in addition to Cdk2/cyclinE to mediate histone modifications such as acetylation and demethylation that result in HP1 redistribution and relaxation of late replicating heterochromatin. Relaxation of heterochromatin is expected to result in an increase in the sites available for replication initiation, because later replicating heterochromatin contains more DNA (3X), and replicons are more clustered at mid S-phase (3X; Lima-de-Faria 1959; Frum et al. 2009). Consequently, higher levels of mutation-prone ssDNA at mid-to-late S-phase can contribute to an increase in mutations that will be repaired less efficiently (Prendergast et al. 2007; Stamatoyannopoulos et al. 2009; Herrick 2010).
In mice, the transition to a late S-phase replication program is associated with the intra S-phase checkpoint (Katsuno et al. 2009), suggesting that Chk1 activation at mid S-phase is coupled to its inactivation in late S/G2 phases (Herrick 2010). Evidence in support of that proposal comes from a recent study showing that Chk1 phosphorylation simultaneously activates Chk1 and promotes its degradation in a DDB1/Cul4A ubiquitin ligase and proteasome-dependent manner (Zhang et al. 2009a; Guervilly et al. 2011). Downregulation of Chk1 at mid-to-late-S/G2 provides an additional potential explanation for why DNA repair systems are less efficient in repairing DNA in heterochromatin. Downregulation of Chk1 during late S-phase in a normal cell cycle remains, however, to be fully demonstrated.
Importantly, the error-prone DNA repair processes NHEJ and DNA translesion synthesis (TLS) can both occur independently of the ATR/Chk1 checkpoint response (Bi et al. 2006; Andersen et al. 2008), and TLS mediates checkpoint downregulation and resumption of DNA synthesis (Bi et al. 2006; Pillaire et al. 2007; Despras et al. 2010). TLS, which replicates through damaged DNA, takes place predominantly during and after S-phase, and is not dependent on genome duplication (Daigaku et al. 2010). In yeast, the Rev1 TLS factor increases slowly from early to mid S-phase, and then increases rapidly in late S-phase until it peaks in G2 (Waters and Walker 2006). Cell cycle regulation of Rev1 in higher eukaryotes, however, has not been observed, and the exact nature of the kinetics of TLS remains obscure (Shaheen et al. 2010). TLS has also recently been proposed as a possible explanation for the increase in non-CpG mutation rates associated with CpG mutations, the so-called “CpG effect” (Walser and Furano 2010). The observed changes in mutation rate within the genome are thus consistent with a growing reliance on Chk1-independent NHEJ (and possibly TLS (Lis et al. 2008)) in late S/G2 phases, whereas differences across species are consistent with greater reliance on NHEJ as genome size (TE-associated heterochromatin) increases (Table 1).
Together, the observations discussed above suggest that genome size has an impact on mutation rates that operates through checkpoint regulation of replication asynchrony and S-phase progression. Larger amounts of heterochromatin are associated with larger chromosomes and genomes, and larger chromosomes exhibit lower levels of recombination and higher rates of mutation in late replicating compared to early replicating DNA (3% increase in yeast versus 10% in rodents and 22% in primates). The unexpectedly low rates of mutation, for example, in the very late replicating Y chromosome can be attributed to lower recombination rates due to a higher relative fraction of heterochromatin (Pink and Hurst 2010). A negative correlation between chromosome size and sequence divergence has also been observed in birds (Nam et al. 2010). An unambiguous correlation between genome size and mutation rate, however, has not been observed to date (Lancaster 2010).
Nevertheless, it will be interesting to investigate if mutation rates are lower in genes embedded in larger genomes that are, paradoxically, more prone to mutation because of their sizes (Herrick 2011). The proopiomelanocortin gene in lungfish (C-value 74.8 pg), for example, is evolving at a much slower rate than in rodents or primates (C value 2.8 and 3.2 pg; Lee et al. 2006). Rates of diversity at other loci are also exceptionally low (Frentiu et al. 2001). Moreover, mutation rates in nuclear genes decrease up to threefold ranging from yeast, C elegans, D melanogaster to humans, whereas mutations in repetitive DNA increase several fold, respectively (Lynch et al. 2008). Interestingly, C. elegans has a relatively higher nuclear mutation rate compared to the other organisms; but a correspondingly lower repeat DNA mutation rate, suggesting an inverse relationship between the two rates. Together, these observations suggest a differential effect of chromatin status on gene location and mutation rate between early and late replicating DNA within the genomes of different species (Fig. 4B).
Here it has been argued that the late replicating status of heterochromatin and degree of replication asynchrony constitute important components—among many other factors—that affect DNA repair efficiencies, and can explain, in part, why eukaryotes have played host to so much junk in their genomes. Late-replicating heterochromatin acts as a substrate for checkpoint function that depends on Chk1/Rad53-mediated inhibition of late-firing replication origins during a normal S-phase and checkpoint-mediated upregulation of DNA repair systems during a perturbed cell cycle. This raises the intriguing possibility that checkpoint function and TE-associated heterochromatin coevolved symbiotically in a way that mutually benefited both host and parasite, a functional relationship made energetically feasible by the emergence of mitochondria (Lane and Martin 2010). Downregulation of DNA repair systems in late S/G2 phases, for example, might have been an adaptation that erodes TE genetic integrity, while at the same time a growing reliance on NHEJ repair incidentally retained TE sequences in the genome (Table 1). Far from being parasites, TEs and other forms of junk DNA, in competing with gene-rich euchromatin, had the fortuitous consequences of promoting the DNA damage response and repair systems that maintain genome integrity, while increasing species diversity through the adaptations that so-called “junk DNA” made possible.