Size-selective harvesting impacts diversity at single-copy regions of the genome such as microsatellite loci (van Wijk et al., 2013) or SNPs within and among protein-coding regions (Therkildsen et al., 2019; Sadler et al., 2023a). However, whether size selective harvesting elicits other types of genomic change, such as variation in copy number, and whether any genomic changes affect the response to temperature is not known. Using an experimental zebrafish model, we found that directional selection (for small and large body size) associates with a reduction in telomere length and rDNA copy number, but has no significant effect on mtDNA content, compared to random-selection (i.e., no directional selection). Hence, relative telomere length and rDNA copy number exhibited a correlated response to directional selection, rather than the direction of selection, per se. While mtDNA content was not impacted by directional selection, fish reared at an elevated temperature exhibited an increase in mtDNA content. Counter to our hypothesis we found no evidence of an interaction between directional selection and thermal stress at any of the genomic regions, suggesting an independent action of these processes on copy number variation and telomere length.
Short relative telomere length associating with directional selection on body size is intriguing as, after ten generations of recovery, both directionally selected lines had lower growth rate and reached smaller adult body size than fish which had not experienced directional selection (Sadler et al., 2024). As telomeres shorten with cell division, unless repaired by telomerase (Chan and Blackburn, 2004), fast growing, larger individuals (random-selected fish) are expected to have shorter telomeres. Fast growing transgenic coho salmon (Oncorhynchus kisutch) are unable to maintain telomere length (Pauliny et al., 2015), and in brown trout (Salmo trutta), body size (but not compensatory growth) was negatively associated with telomere length (Debes et al., 2016) and a greater change in telomere length (Näslund et al., 2015). However, telomere length and expression of telomerase increase with development in zebrafish muscle such that telomeres do not shorten with growth in healthy zebrafish until old age (about 30 months) when telomerase expression declines (Lau et al., 2008; Anchelin et al., 2011). It could be speculated that the fish under directional selection were less capable of telomere maintenance than the line that experienced random selection, however the mechanisms for this would require further studies.
Loss of genetic diversity, (Poulsen et al., 2006; Therkildsen et al., 2010; Pinsky and Palumbi 2014; Sadler et al., 2023b) and potentially inbreeding (Hoarau et al., 2005; O’Leary et al., 2013), is a feature of many overharvested fish stocks. Directional selection may elicit a faster loss of genetic diversity than expected under a population reduction alone (Frankham, 2012). Although we do not measure inbreeding or genetic diversity, slower growth associated with directional selection may indicate that these lines experience some inbreeding depression whose effects extend to telomere maintenance (Sadler et al., 2024). However, the relationship between telomere length and inbreeding is controversial. Studies on wild vertebrate populations have shown that inbreeding/elevated levels of homozygosity is associated with short (Bebbington et al., 2016; Pepke & Eisenberg 2022) and long telomeres (Hemann and Greider, 2000), or have failed to uncover any significant effect of inbreeding on telomere length (Olsson et al., 2022). Nonetheless, the comparably short telomeres in both size-selected lines indicates that directional selection, at least for body size, can have an unintended, but important outcome on telomere length - an effect that has not been reported in teleost fish.
As it is not possible to quantify changes in telomere length using non-destructive sampling (e.g. from blood; Reichert et al., 2017) on such young zebrafish, we do not know whether the outcome of our experiment reflects an inherent difference in telomere length among the lines or whether all fish had similar length telomeres at hatching and the shorter telomeres are a consequence of poor telomere maintenance in the size-selected lines. Nonetheless, that size-selective harvesting can cause short telomeres and/or poor telomere maintenance is a potential cause for concern for the health of overharvested fish stocks given the widespread reports that short telomeres are a biomarker for stress exposure or reduced health in many animals (Monaghan and Haussmann, 2006; Horn et al., 2010; Bojesen 2013; Näslund et al., 2015; Bateson 2016; Wilbourn et al., 2018), including telomerase deficient zebrafish (Lex et al., 2020).
Variation in temperature did not impact telomere length in zebrafish in contrast with previous studies demonstrating a negative association between water temperature and telomere length in brown trout (Salmo trutto) (Näslund et al., 2015; Debes et al., 2016). In sticklebacks (Gasterosteus aculeatus), variation in temperature did not directly affect telomere length at individual level but had sex specific effects on telomere length in mature fish (Noreikiene et al., 2017). This is interesting as old (> 18 months) zebrafish have short telomeres (Evans et al., 2021), implying that a longer experimental duration might have revealed additional changes in telomere length in our lines. The lack of interaction between selection line and temperature stress on telomere length is surprising following previous interactions on phenotypic traits including growth (Morrongiello et al., 2021). But the lack of interaction is relevant if the size-selected lines experience some inbreeding depression, because a greater impact of inhabiting a poor environment on telomere length was uncovered in less genetically diverse juvenile birds (Bebbington et al., 2016).
As rDNA copy number is sensitive to environment variation (Kobayashi, 2011; Paredes et al., 2011; Salim et al., 2017; Jernfors et al., 2021) it represents a hypothesised ‘environmental sensor’ that may regulate the molecular response to environmental cues (Kwan et al., 2013; Jack et al., 2015; Salim and Gerton, 2019; Symonová, 2019). It is therefore surprising that we found no significant impact of different temperature environments on zebrafish rDNA copy number. Future work could study different species and/or different environmental stressors, such as nutrient stress (Aldrich and Maggert, 2015) or exposure to pollutants (Jernfors et al. 2021), to uncover possible drivers of rDNA copy number variation in teleosts. Nonetheless, it is interesting that directional selection on body size affected rDNA copy number as significant differences in rDNA copy number among strains of inbred laboratory mice (Mus musculus; Veiko et al., 2007) raise the prospect that rDNA copy number might be impacted by reductions in population size/inbreeding (as discussed for telomere length). Indeed, the significant positive correlation between rDNA and relative telomere length supports the idea that these regions are sensitive to similar stressors (Valeeva et al., 2023). For example, rDNA and telomeres are both sensitive to changes in heterochromatin architecture (Mozgova et al., 2010) and oxidative stress (Valeeva et al., 2023). Our data highlight a need to quantify rDNA copy number and telomere length together to determine in what taxa, in what environments, and potentially why, copy number/length of these regions of the genome are co-associated. Moreover, it is important to understand these changes in the context of traits associated with growth as these regions are sensitive to cell division (Allsopp et al., 1995; Chan and Blackburn, 2004; Kobayashi, 2014). The negative association between mtDNA content and rDNA copy number in zebrafish is consistent with the negative association between these regions of the genome in humans (Gibbons et al., 2014). However, the weak relationships between mtDNA content and rDNA copy number likely reflects that mitochondria have a separate genome and mitochondrial density being dynamic and independent of cell division (Ding et al., 2021).
The significant effect of temperature on mtDNA content at 34°C adds to the diversity of biological impacts that occur in aquatic communities exposed to thermal stress. To our knowledge, only one previous study has examined the effect of temperature on mtDNA content in teleosts, in which there was an increase in mtDNA content in eggs at warmer winter temperatures (+ 5°C) in stickleback (Gasterosteus aculeatus) (Kim et al., 2022). Similarly, an increase in mtDNA content was reported in prawns (Palaemon carinicauda) raised in warm water (Li et al., 2018). An increase in mtDNA content corresponds with an increase in mitochondrial content (Lee and Wei, 2000) and is consistent with an expected rise in metabolic rate that accompanies an increase in temperature (Clarke and Fraser, 2004; Johansen and Jones, 2011). Studying mtDNA content in tandem with telomere length and rDNA copy number is relevant as mitochondrial density may positively associate with the production of reactive oxygen species (Abele et al., 2002; Olsson et al., 2018; Metcalfe & Olsson 2022) that can damage telomeres (von Ziglinki et al., 2002; Reichert & Stier 2017; Barnes et al., 2019) and rDNA (Kobayashi & Sasaki 2017). For example, an increase in mtDNA density was correlated with production of the free radical superoxide, which in turn influenced telomere length (Olsson et al., 2018). Indeed, we found negative correlations between mtDNA content and telomere length/rDNA copy number in the large-selected lines, which (1) reinforces the idea that telomere length (and rDNA copy number) should be quantified in tandem with mtDNA content, and ideally ROS production (Olsson et al., 2018), and (2) shows how any association between these regions of the genome can depend on genetic background.
Balanced harvesting is hypothesised to mitigate the effects of directional selection (Garcia et al., 2012), for example by retaining genetic diversity and lessening any effects of inbreeding (Zhou et al., 2019). Balanced harvesting can increase stock productivity (Zhou et al., 2015), aid the recovery of a stock’s natural size and age structure (Beamish et al., 2006), and improve the resilience of a stocks to natural disturbance (Hixon et al., 2014). Here, our random-selected lines experienced harvesting but no directional selection on body size and thus correspond with balanced harvesting. Our zebrafish model of overharvesting supports the idea that balanced harvesting can help maintain growth and body size. We also show how harvesting regime can impact regions of the genome that are associated with organismal health and fitness (Monaghan and Haussmann, 2006; Horn et al., 2010; Näslund et al., 2015; Wilbourn et al., 2018).
We show that directional selection (for body size) has a greater impact on these regions of the genome than an equivalent random reduction in population size. Identifying the mechanisms behind these results requires further work but may be related to a loss of genetic diversity that accompanies directional selection. Quantifying processes that drive variation in telomere length and rDNA copy number is important as maintenance of these loci is thought to be essential to genome integrity (O’Sullivan and Karlseder, 2010; Qiu, 2015; Symonová, 2019) and the rate of molecular aging (Kobayashi, 2011; Lu et al., 2018). Intriguingly, our data indicate relative telomere length and rDNA copy number are resilient to changes in temperature. In contrast, mtDNA content was not impacted by directional selection but was increased at elevated temperatures, presumably in response to a change in metabolic requirements. Our data open new avenues for future research of dynamics of telomere length, rDNA copy number, and mtDNA content in wild populations. For example, a next step would be to determine whether natural populations of exploited fish experienced similar genomic impacts and, if so, what are the mechanisms and do these genomic changes alter individual fitness. Crucially, direction of selection (either small- or large-selection on body size) appears less important that the act of directional selection itself, as directional selection reduced rDNA copy number and telomere length regardless of the direction compared to random selection. Our data suggest that selection regimes implicated by fisheries should be reconsidered, utilising alterative harvest strategies such as balanced harvesting to reduce any effects of directional selection on fitness.