Chuan CHEN ,Long JIN ,Ying JIANG and Wenbo LIAO＊

1Key Laboratory of Southwest China Wildlife Resources Conservation (Ministry of Education),China West Normal University,Nanchong 637009,Sichuan,China

2Key Laboratory of Artificial Propagation and Utilization in Anurans of Nanchong City,China West Normal University,Nanchong 637009,Sichuan,China

3Institute of Eco-adaptation in Amphibians and Reptiles,China West Normal University,Nanchong 637009,Sichuan,China

Abstract Genome size changes significantly among taxonomic levels,and this variation is often related to the patterns shaped by the phylogeny,life histories and ecological factors.However,there are mixed evidences on the main factors affecting molecular evolution in animals.In this study,we used phylogenetic comparative analysis to investigate the evolutionary rate of genome size and the relationships between genome size and life histories (i.e.,hatchling mass,clutch size,clutches per year,age at sexual maturity,lifespan and body mass) among 199 squamata species.Our results showed that the evolutionary rate of genome size in Lacertilia was significantly faster than Serpentes.Moreover,we also found that larger species showed larger hatchling mass,more clutches per year and clutch size and longer lifespan.However,genome size was negatively associated with clutch size and clutches per year,but not associated with body mass we looked at.The findings suggest that larger species do not possess the evolution of large genomes in squamata.

Keywords genome size,body mass,evolutionary rate,life histories

1.Introduction

Genome sizes vary considerably across taxa in organisms(Cavalier-Smith,1978;Bennett and Leitch,2005;Lynch and Walsh,2007).This can be driven by the stochastic genetic and/or genomic processes associated with spontaneous deletions and/or insertions,polyploidization,prolonged tandem repeats length,transposable elements number and genetic drift,but can be also shaped by natural selection (Ogata

et al

.,1996;Petrov,2001;Sun

et al

.,2012;Lynch,2011;Whitney and Garland,2010).In particular,genome size variations are mainly explained by two important mechanisms including the duplication events and the proliferations of noncoding elements (Neiman

et al

.,2015).Establishing the association between genome size variation and organismal complexity has puzzled many evolutionary biologists and as such remains a classic problem in biology (Gregory,2005a).Previous studies across taxa have revealed positive associations between genome size and cell size,nucleus size,developmental time,nutrient requirements,tissue differentiation,life cycle complexity and body size (Vinogradov,1997;Olmo and Morescalchi,1978;Gregory,2005a;Gregory,2001;Gregory and Johnston,2008;Guignard

et al

.,2016).These positive associations have been suggested to be consequences of both the cytoplasm from more efficient mRNA transport and larger cells necessitating larger genomes based on structural causes (Cavalier-Smith,1985).Smaller cells for instance,usually divide faster and have a higher metabolic rate,evidenced by a negative correlation between metabolic rates and DNA amounts in mammals and birds (Hughes and Hughes,1995;Vinogradov,1995;Gregory,2002a;Hughes and Piontkivska,2005).However,a potential correlation which still needs to be explored between cell volume and genome size is body size(Gregory

et al

.,2000).Body size variation is often determined either by cell size and cell number,or both combination in organisms (Hessen

et al

.,2013;Koz?owski

et al

.,2003).For plants and animals,genome size displays a positive association with cell size (Gregory

et al

.,2000;Bennett,1987;Gregory,2005b).In addition,this correlation can be often linked to ecological factors.For example,genome size exhibits a positive association with body size in some invertebrates (e.g.,amphipods,copepods,crustaceans) due to low metabolic rate and temperatures in cold waters (Rees

et al

.,2008;Angilletta

et al

.,2004;Timofeev,2001;Jeffery

et al

.,2016;Leinaas

et al

.,2016).Genome size variation in frogs is indirectly affected by temperature and humidity as a result of its influence on the time of premetamorphic development (Liedtke

et al

.,2018).In birds,mitochondrial and nuclear of substitution rate in coding sequences reveal weak negative associations between the ratio of nonsynonymous and synonymous substitution rate and age at sexual maturity,lifespan and body mass associated with environmental factors (Weber

et al

.,2014;Lanfear

et al

.,2010;Nabholz

et al

.,2013),but it is not always the case (Figuet

et al

.,2017).Squamata constitutes the class of vertebrates with a small genome size due to a lower fraction of transposable elements and shorter introns (Organ

et al

.,2007).However,this may be a misconception caused by overlooking GC-rich regions,which are often hard to access (Botero-Castro,2017).Ploidy variations does not provide a major power of variation in genome size,and the whole-genome duplication events is not reported during the amniote evolutionary process in squamata (Van de Peer

et al

.,2009).Squamata mainly consists of two suborders(e.g.,Lacertilia and Serpentes) and displays complex life histories with prolonged developmental periods (hatching time),which likely constrains the variation of genome size because of a negative correlation between genome size and development time in invertebrates (Wyngaard

et al

.,2005).

To examine the selective mechanisms underlying genome size variation in squamata,we first estimated the evolutionary rates of genome size between Lacertilia and Serpentes in squamata.We also expanded our extent to which genome size can be considered as a determinant of life histories by investigating the relationships between variation in genome size and life histories among 199 squamata species.We tested whether larger bodies can promote evolution of larger genomes.

2.Materials and Methods

2.1

.

Data collection

The genome size of 199 squamata species was collected from genome size database (http://www.genomesize.com) (supplementary Information:Table S1).We extracted data on genome size for squamata species for which information on life histories can be found (see below),and obtain their average

C

-value.We used average values of genome size when more than one measurement per species was available.To avoid possible errors due to several methods being used to quantify genome size (Hardie

et al

.,2002),we used parallel analyses on a subset of genome size.We confirmed species names using the NCBI taxonomy database,and collapsed/pruned all synonyms from the phylogenetic tree.We rebuilt the phylogenetic tree using time-calibrated molecular phylogeny by Pyron

et al

.(2013) (Figure 1) and examined difference in the evolutionary rate of genome size between Lacertilia and Serpentes.Finally,we compiled information on hatching time,hatchling mass,clutch size per year,clutch size and body mass (see details in De Smet,1981;Feldman

et al

.,2016;Allen

et al

.,2017) and age at sexual maturity,lifespan from the AnAge databases (https://genomics.senescence.info/species/)(supplementary Information:Table S1).

2.2.Statistical analyses

The complementary approaches were used to evaluate the evolution rate of genome size for three suborders.For each suborder,we assessed phylogenetic signal using the

phylosig

function in the package of

phytools

in RStudio v.3.1.2 (Revell,2012).We then used the Blomberg’s

K

(Blomberg

et al

.,2003) in which genome size variation comparing on a null model is assumed genome size evolution under Brownian motion (BM) model.We also used the Pagel’s

λ

(Pagel,1999)in which phylogenetic signal is estimated on the basis of the phylogenetic dependence of genome size.

K

=1 indicated genome size evolved as expected under a BM model,while

K

＞ 1or

K

＜ 1 indicated less or more phylogenetic signal than expected under a BM model,respectively.We used Blomberg’s

K

and Pagel’s

λ

to estimate the phylogenetic signal and found qualitatively similar results (Table S2).We used the

fitContinuous

function in the R package-

Geiger

(Harmon

et al

.,2008) to compare genome size evolution on the basis of Brownian motion,Ornstein-Uhlenbeck and Early-burst models between the two suborders.Following the suggestions by Simmons and Fitzpatrick (2016),BM model of genome size evolution was regarded to be the best model due to small sample size.Moreover,to compare differences in evolutionary rate of genome size between the two suborders,we modified a likelihood method where a phylogeny can directly compare on the Brownian evolutionary rate (

σ

) of genome size (Adams,2013).To examine associations between genome size and life histories,we used the phylogenetic generalized least squares models where the phylogenetic structure of the model residuals was considered in the

caper

package (Orme

et al

.,2012;Huang

et al

.,2020).We used phylogenetic scaling parameter

λ

to estimate the phylogenetic influence on the associations between genome size and life histories based on a maximumlikelihood approach (Pagel,1999).The scale of

λ

-values ranges from zero (i.e.,phylogenetic independence) to one (i.e.,complete phylogenetic non-independence) (Freckleton

et al

.,2002;Mai

et al

.,2019)

.

We log-transformed life histories to linearize associations and used the phylogenetic tree of squamata species to correct for phylogenetic dependence (Mai

et al

.,2020).To test the associations between body mass and life histories,we treated body mass as response variable,hatchling mass,clutches per year,clutch size,age at sexual maturity and lifespan as predictor variables using the multivariate phylogenetic generalized least squares.To test whether genome size exhibited a association with body mass,we treated body mass as predictor variable,genome size as response variable,and hatchling mass,clutches per year,clutch size and lifespan as covariates using the multivariate phylogenetic generalized least squares.

3.Results

The average value of genome size was 2.11 pg,ranging from 1.19 to 3.93 pg among 199 species of squamata.Genome size in Lacertilia tended to be larger than that in Serpentes (Figure 2).The evolutionary rate of genome size in Lacertilia was faster than that in Serpentes (Table S3).

Figure 1 The phylogenetic tree of the 199 species of squamata used in the comparative analysis.

Figure 2 Genome size difference between Lacertilian and Serpentes for 199 species of squamata.

The multivariate phylogenetic generalized least squares model indicated that body mass was positively associated with hatchling mass,clutches per year,clutch size and lifespan among 199 species of squamata (Table 1) .The genome size was not associated with body mass when the effects of hatchling mass,clutches per year,clutch size and lifespan were removed(Table 2).We also found negative correlations between genome size and clutch size or clutches per year (Table 2).

For Serpentes in particular,body mass was positively and significantly associated with hatchling mass and clutch size,but not with clutches per year,age at sexual maturity and lifespan using the multivariate phylogenetic generalized least squares model (Table S4).However,there was no association between genome size and body mass when removing the hatchling mass and clutch size effects (Table S5).For Lacertilia,body mass was significantly associated with hatchling mass,clutches per year,clutch size and lifespan (Table S4).When the influences of hatchling mass,clutches per year,clutch size and lifespan were removed,we found no association between genome size variation and body mass (Table S5).

Table 1 The associations between body mass and life histories across 199 species of squamata.Phylogenetic scaling parameters (superscripts following λ denote P-values of likelihood ratio tests against models with λ=0 and λ=1,respectively).

Table 2 The associations between genome size and life histories across 199 species of squamata.Phylogenetic scaling parameters (superscripts following λ denote P-values of likelihood ratio tests against models with λ=0 and λ=1,respectively).

4.Discussion

Our results showed that genome size evolution in Lacertilia evolved significantly faster than that in Serpentes among 199 species of squamata.We found positive correlations between body mass and hatchling mass,clutches per year,clutch size,and lifespan.However,genome size was not associated with body mass when correcting for the effects of part life histories.For Lacertilia and Serpentes,genome size did not show a association with body mass.

Differences in transposable element accumulation rates in animals experienced may lead to substantial variation in genome size among species (Chalopin

et al

.,2015;Gibbs

et al

.,2004).For example,a number of DNA obtained by transposable element accumulation with strong changes among lineages,are counteracted by loss of DNA on the basis of large segmental deletion in birds (Kapusta

et al

.,2017).For 199 species of squamata,the rate of transposable element accumulation can also explain the marked variation in genome size,ranging from 1.19 to 3.93 pg.The evolutionary history of genome size in amphibians has been one of gradual,time-dependent variation (Brownian motion;Liedtke

et al

.,2018).In this study,evolutionary modelfitting showed that genomes in Lacertilia and Serpentes evolved under a shared processes of Brownian motion.The common ancestor of extant squamata was predicted to have similar size in genome in Lacertilia and Serpentes.We inferred that genome size in squamata evolved gradually as a function of time(Brownian motion).Herein we found that the evolutionary rate of genome size in Lacertilia evolved faster than Serpentes.Palaeontological data and genomic evidence display a similar pattern (Pyron

et al

.,2013).There are evidences that phylogeny is likely to promote the influences of genome duplications and transposons on genome size evolution in animals (mammals:Tang

et al

.,2019;insects:Alfsnes

et al

.,2017).For example,genome size is phylogenydependent when

λ

＞ 0.9 in all life-history traits is reported in mammals (Tang

et al

.,2019).However,phylogeny displays a weak correlation with genome size in crustaceans (Alfsnes

et al

.,2017).Likewise,there is a weak association between genome size and phylogeny among 240 species of birds when

λ

≤ 0.564 is recorded in all life-history traits (Yu

et al

.,2020).We found that genome size was not associated with phylogeny,suggesting that the phylogeny did not a strong power in driving transposons and duplications of genome in squamata.Genome size variation can be explained by the more mechanistic and/or short period effects which is regarded as the proximate causes.Moreover,the evolutionary powers (i.e.,selection),regarding as the ultimate causes,can also explain the genome size variation (Hessen

et al

.,2013;Alfsnes

et al

.,2017;Yu

et al

.,2020).For birds,variations in genome size are positively related to the length of developmental period (Kapusta

et al

.,2017;Yu

et al

.,2020),providing evidence for the associations between life histories and genome size evolution.Indeed,genome size displays markedly and directly effects on cell size and cell replication rate (Gregory,2002b),so larger genomes are expected to be positively correlated with larger egg size and smaller clutch size.However,large datasets have indicated that variations in genome size are not associated with offspring number and size in mammals (Tang

et al

.,2019) and life history complexity of amphibians (Liedtke

et al

.,2018).In this study,there were negative correlations between genome size variation and life histories such as clutch size and clutches per year in squamata,suggesting that less offspring number or larger offspring size can promote evolution of larger genomes.Body mass is positively associated with genome size in vertebrates (Liedtke

et al

.,2018;Tang

et al

.,2019;Yu

et al

.,2020)and invertebrates (Gregory

et al

.,2000;McLaren

et al

.,1989;Hessen and Persson,2009;Alfsnes

et al

.,2017).Such positive associations between cell size and genome size (Gregory,2005a;McLaren and Marcogliese,1983) have indicated that variations in body size among the related species can partly respond to variation in cell size (Hessen

et al

.,2013).Indeed,genome size exhibits positively correlations with body mass in birds and mammals (Tang

et al

.,2019;Yu

et al

.,2020).Across 199 species of squamata,there were not associations between genome size and body mass,suggesting that diversity in genome size was not response of variation in cell size.

In conclusion,we illustrated the relationships between genome size and life histories in squamata.The hatching time,hatchling mass,clutch size per year and clutch size cannot shaped the genome size variation,and species with larger bodies did not possess larger genomes in squamata.Our future research would need more species to reveal the relationships between genome size evolution and life histories.

Acknowledgements

We thank C.L.MAI and J.P.YU to help the data collected.Financial support was provided by the National Natural Sciences Foundation of China (31 772451;31970393) and the Science and Technology Youth Innovation Team of Sichuan Province (2019JDTD0012).

Appendix

Table S1 TSpecies,body mass (g),genome size (pg),hatchling mass (g),clutch size,clutches per year among 199 species of squamata from the references of De Smet (1981),Feldman et al.(2016),Allen et al.(2017),and age at sexual maturity (years) and lifespan (years) from AnAge (https://genomics.senescence.info/species/).

(Continued Table S1)

(Continued Table S1)

(Continued Table S1)

References for Table S1

Allen W.L.,Street S.E.,Capellini I.2017.Fast life-history traits promote invasion success in amphibians and reptiles.Ecol Lett,20(2):222-230

De Smet W.H.O.1981.The nuclear Feulgen-DNA content of the vertebrates (especially reptiles),as measured by fluorescence cytophotometry,with notes on the cell and chromosome size.Acta Zool Pathol Antverp,76(1):119-167

Feldman A.,Sabath N.,Pyron R.A.,Mayrose I.,Meiri S.2016.Body sizes and diversification rates of lizards,snakes,amphisbaenians and the tuatara.Glob Ecol Biogeogr,25(2):187-197

Table S2 Evaluation of phylogenetic signal in genome size examined.

Table S3 Comparison of model parameters and fit for each suborder examined under Brownian motion,Ornstein-Uhlenbeck and Early-burst evolutionary models.

Table S4 Associations between body mass and life histories for the two suborders in squamata using phylogenetic generalized least squares models.Phylogenetic scaling parameters (superscripts following λ denote P-values of likelihood ratio tests against models with λ=0 and λ=1,respectively).

Table S5 Associations between genome size and life histories in squamata using phylogenetic generalized least squares models.Phylogenetic scaling parameters (superscripts following λ denoteP-values of likelihood ratio tests against models with λ=0 and λ=1,respectively).