The impact of tandem duplication on gene evolution in Solanaceae species

HUANG Yi-le,ZHANG Ling-kui,ZHANG Kang,CHEN Shu-min,HU Jian-bin,CHENG Feng

1 College of Horticulture,Henan Agricultural University,Zhengzhou 450002,P.R.China

2 Institute of Vegetables and Flowers,Chinese Academy of Agricultural Sciences,Beijing 100097,P.R.China

Abstract Whole genome duplication (WGD) and tandem duplication (TD) are important modes of gene amplification and functional innovation,and they are common in plant genome evolution.We analyzed the genomes of three Solanaceae species(Solanum lycopersicum,Capsicum annuum,and Petunia inflata),which share a common distant ancestor with Vitis vinifera,Theobroma cacao,and Coffea canephora but have undergone an extra whole genome triplication (WGT) event.The analysis was used to investigate the phenomenon of tandem gene evolution with (S.lycopersicum) or without WGT(V.vinifera).Among the tandem gene arrays in these genomes,we found that V.vinifera,which has not experienced the WGT event,retained relatively more and larger tandem duplicated gene (TDG) clusters than the Solanaceae species that experienced the WGT event.Larger TDG clusters tend to be derived from older TD events,so this indicates that continuous TDGs (absolute dosage) accumulated during long-term evolution.In addition,WGD and TD show a significant bias in the functional categories of the genes retained.WGD tends to retain dose-sensitive genes related to biological processes,including DNA-binding and transcription factor activity,while TD tends to retain genes involved in stress resistance.WGD and TD also provide more possibilities for gene functional innovation through gene fusion and fission.The TDG cluster containing the tomato fusarium wilt resistance gene I3 contains 15 genes,and one of these genes,Solyc07g055560,has undergone a fusion event after the duplication events.These data provide evidence that helps explain the new functionalization of TDGs in adapting to environmental changes.

Keywords:tandem duplication,whole genome duplication,Solanaceae species,gene retention,gene fusion

1.Introduction

Gene duplication plays an important role in plant evolution and provides material bases useful for the evolution of new functions.Duplicated genes provide greater and less-constrained chances for natural selection to shape novel functions (Longet al.2003).Gene duplication occurs in several different modes,such as whole genome duplication (WGD),single-gene duplications,and segmental duplications (Maereet al.2005;Patersonet al.2010).WGD is a major force in plants that drives biological complexity,evolutionary novelty,and adaptation to specific conditions (Peeret al.2009).For example,evolutionary patterns suggest thatArabidopsisand all Brassicaceae taxa have undergone three rounds (i.e.,γ,β,andα) of whole genome duplication or triplication (WGD or WGT) (Bowerset al.2003).The first WGT event (γevent) is shared by almost all dicotyledonous plants and has been linked to the diversification of eudicots and angiosperms (Bodtet al.2005).The existing Brassicaceae clade diverged after the subsequent two duplication events (αandβ)(Bowerset al.2003).

Single-gene duplications were also involved in the origin of many plant genes (Magadumet al.2013).Tandem gene duplication is an important source of single-gene duplication in plant genomes and refers to two or more homologous genes adjacent to each other in the genome (Jander and Barth 2007).Tandem duplicated genes (TDGs) are common in plant genomes and account for a significant proportion of all genes:17%inArabidopsis(Kaulet al.2000),14% in rice (Matsumotoet al.2005),16% in poplar (Tuskanet al.2006),and 35% in maize (Messinget al.2004).TDGs are important in plant evolution and adaptation to environmental changes.InArabidopsis,TDGs comprise almost as many genes as those duplicated by paleopolyploid events (～25%),and they represent a wide range of functional components of the genome (Kliebensteinet al.2001).

The retention of duplicated genes does not occur randomly.Dose-sensitive genes associated with biological processes,such as transcription factors and protein binding,are preferentially retained following WGDs (Blanc and Wolfe 2004a),while those genes related to abiotic and biotic stress are more likely to be over-retained following tandem duplications (Rizzonet al.2006;Hanadaet al.2008).Some functions related to TDGs are in the form of gene cluster arrays that lead to the expansion of gene families and an increase in gene dosage (Genget al.2019;Wanget al.2020).In some cases,sudden changes in the concentration of gene products have deleterious effects (dosage imbalance)on the physiology of the organism and will be selected against (Innan and Kondrashov 2010).

The genomes of Solanaceae species contain many duplicated chromosomal fragments,which reflect WGT events.In addition to theγevent experienced by all dicotyledons,Solanaceae species underwent another WGT event (Tevent) about～65 million years ago (Satoet al.2012).These genes have subsequently diverged,but three sets of tomato sub-genomes can be recognized in comparison to other species,such asVitisvinifera,Theobromacacao,orCoffeacanephora,which only experienced theγevent.In this study,we identified the TDGs retained in the genomes ofSolanumlycopersicum,Capsicumannuum,andPetuniainflata,and compared them to the TDGs of three species that did not experience additional WGD events (V.vinifera,T.cacao,andC.canephora).We determined whether WGD had significant effects on TDG amplification and inferred the general patterns of TDG expansion in Solanaceae species.

2.Materials and methods

2.1.Data sources

We selected three sequenced genomes (S.lycopersicum,C.annuum,andP.inflata) representing the major lineages of the Solanaceae that have clear WGT records during their evolutionary history.Three non-Solanaceae genomes (T.cacao,C.canephora,andV.vinifera)that only experienced theγtriplication event were also selected.Genome data ofS.lycopersicum,C.annuum,andP.inflatawere downloaded from the Solanaceae Genomics Network (https://solgenomics.net) (Fernandez-Pozoet al.2015).TheT.cacaoandV.viniferagenomes were downloaded from Phytozome (version 11) (https://phytozome.jgi.doe.gov/pz/portal.html) (Goodsteinet al.2012).TheC.canephoragenome was downloaded from the Coffee Genome Hub (http://coffee-genome.org)(Denoeudet al.2014) (See Appendix A for the download information of these genomes).

2.2.Tandem duplicate identification

We used a refined method to identify tandem duplication arrays.Protein homology information for each gene pair was obtained by running self-to-self BLASTP (Altschulet al.1990).Each TDG cluster was composed of continuously distributed homologous genes(E-value＜1.0×10-5) and could not be interrupted by more than six non-homologous genes.Specifically,we sequentially detected the tandem genes of each gene according to its physical position in the genome.If the detected gene pair was homologous,it was classified into the gene cluster and continued to be detected.If the number of non-homologous genes exceeded six,the detection was stopped.When all the genes were detected,we were able to obtain the tandem duplication array of the species.This method can accurately identify the homology relationships of gene pairs in gene clusters and effectively reduce the false-positive genes introduced due to low gene pair homology.

2.3.ldentification of WGD events

It has been clearly documented that the coffee genome has not experienced any additional WGD events(Denoeudet al.2014).As coffee is closely related to the Solanaceae family,it was used as the reference genome in this study.We used SynOrths (Chenget al.2012) to conduct syntenic analysis of the tomato and coffee genomes,and 14 589 syntenic gene pairs were obtained.We found that one coffee gene corresponds to up to three copies in tomato,indicating that tomato underwent a WGT event after theγevent,and obtained 4 650 tomato paralogs.

2.4.Phylogenetic analysis

Single-copy gene families were used to reconstruct the phylogenetic tree forSolanumtuberosum,S.lycopersicum,Solanummelongena,C.annuum,Nicotianatabacum,P.inflata,C.canephora,T.cacao,V.vinifera,Betavulgaris,andArabidopsisthaliana.Proteins of single-copy gene families were aligned using MAFFT (Katohet al.2002).Based on the alignment results,each protein was tracked back to the coding sequence (CDS).CDS degenerate sites were extracted from each alignment and concatenated to one supergene forA.thalianaand the other species.Subsequently,maximum likelihood trees were estimated from the filtered alignments with RAxML (Stamatakis 2014),using the default rapid hill-climbing search algorithm and a GTR+Gamma nucleotide substitution model.

2.5.Calculation of KS values

ParaAT (Zhanget al.2012) was used to align the protein sequences of TDG pairs,and it translated the results into codons corresponding to accounting alignment results.Then the KaKs_Calculator (Wanget al.2010) was used to calculateKS-values.

2.6.Functional enrichment analysis

We obtained GO annotations for each of the TDGs from the tomato annotation file (ITAG4.0).The GO items and KEGG pathway of gene enrichment were identified using the clusterProfiler package (Yuet al.2012).We used Fisher’s exact test and adjusted theP-values according to the Benjamini and Hochberg (1995) (false discovery rate)procedure,in order to evaluate the statistical enrichment of these genes.An adjustedP-value ＜0.05 was set as the cut-off criterion for the significance of the gene enrichment.

2.7.Conserved domain analysis of the genes

We used the CD-search tool (Marchler-Bauer and Bryant 2004) on the NCBI website (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) to query the domain of genes.The search was based on the Conserved Domain Database (CDD) database with an expected value threshold of 0.01.Composition-based statistics is a new“composition-aware”measure implemented in RPS-blast,and it largely eliminates the necessity of using a lowcomplexity filter.

3.Results

3.1.T event is associated with relatively fewer TDGs in Solanaceae genomes

Significant quantitative differences were found in the TDGs of Solanaceae species and other species that did not experience additional WGD events.We selectedS.lycopersicum,C.annuum,P.inflata,C.canephora,T.cacao,andV.viniferato reveal the general patterns in their amplification of TDG.Their phylogenetic tree indicated thatP.inflatawas differentiated during the early stage of Solanaceae evolution,andC.canephorais closely related to the Solanaceae species (Fig.1-A).The number of clusters and the number of TDGs are shown in Table 1,and the proportion of the number of TDGs in each cluster to the total number of genes is shown in Fig.1-B.We identified the greatest number of TDGs in theV.viniferagenome,and they occupied～25% of the total number of genes.We found that species that did not experience theTevent have significantly higher proportions of TDGs (Fig.1-B;Appendix B),especially when the number of genes in TDG clusters was ＞6.In those cases,the difference was the most significant(independentt-test,P-value=2.80×10-3).These data suggest that WGD events provided adaptive resources for the evolution of Solanaceae species.Species that have not experienced additional WGD events tend to increase the number of genes through tandem duplication.

Table 1 Number of tandem duplicated genes (TDGs) identified in six plant species

Fig.1 Phylogenetic relationships of eudicots and the number of tandem duplicated genes (TDGs) in clusters.A,S.tuberosum,S.lycopersicum,S.melongena,C.annuum,and P.inflata were selected to represent the Solanaceae family in the construction of the phylogenetic tree.Arabidopsis was designated as an outgroup.The red triangles mark whole genome triplication (WGT)events.B,the proportion of TDGs to the total number of genes in each species.Different numbers of genes in clusters are represented by different shades of blue.

3.2.A TDG burst occurred after a long period of evolution since the T event in Solanaceae

Solanaceae species experienced one relatively recent large-scale TDG outbreak,while the TDGs of non-Solanaceae species were sustainably amplified.TheKSvalue distribution of duplicated genes has been used to infer the polyploid origin and pattern of plants (Blanc and Wolfe 2004b).We compared theKS-value distribution of TDGs from Solanaceae and non-Solanaceae species.Solanaceae species experienced a large-scale TDG burst atKS=0.3 (Fig.2-A),while theTevent occurred atKS=0.65 (Satoet al.2012).In contrast,non-Solanaceae species continuously accumulated more TDGs between the occurrences of theTevent and theγevent.This implies that after the occurrence of theγevent,non-Solanaceae species have increased genetic diversity and environmental adaptability by amplifying a large number of TDGs,thus eliminating the need for duplication at the whole genome level.In contrast,the Solanaceae species depended on theTevent for preservation.

Both Solanaceae and non-Solanaceae species showed that the larger tandem duplicated gene (TDG)clusters accumulated from older duplications.Tomato,the representative species of Solanaceae,and grape,with a significant quantity advantage,were selected to compare the TDG amplification patterns.We used the largestKS-value in a cluster of TDGs to represent the duplication time of that cluster (Fig.2-B;see Appendix C for other species).Some genes of Solanaceae and non-Solanaceae species have accumulated and formed larger TDG clusters through long-term evolution,while some smaller TDG clusters appeared to have been duplicated recently.In large clusters (gene number ＞6),theKSvalues of TDGs were uniformly distributed (Fig.2-C;Appendix D),reflecting the tendency of these genes to continuously undergo new amplification steps.

Gene duplication might be followed by the introduction of other genes,which results in interspaced gene pairs.Therefore,we investigated theKS-values of TDG pairs with different numbers of interspaced genes (Fig.2-D;Appendix E).Gene distance (measured by the number of interspaced genes between two TDGs) was positively correlated withKS-values and conform to a linear regression (the linear regression coefficientR2values of tomato and grape were 0.892 and 0.955,and the correspondingP-values were 3.84×10-5and 1.13×10-6,respectively),indicating that the TDGs were preferentially duplicated in an adjacent location.

Fig.2 KS-value distributions of tandem duplicated genes (TDGs) among different species,different lengths and different distances.A,tandem duplication in six plant genomes revealed by KS analyses.B,the distribution of the largest KS-values in TDG clusters with different numbers of genes.The examples of Solanaceae and non-Solanaceae species shown here are tomato and grape,respectively (see Appendix C for other species).C,in large TDG clusters (gene number ＞6),the distribution of KS-values in TDG clusters with different numbers of genes (see Appendix D for other species).D,the KS-values of different numbers of interspaced genes between TDG pairs (see Appendix E for other species).

3.3.Adaptation-related genes are over-retained in Solanaceae TDGs

Over-retained genes derived from different duplication mechanisms,i.e.,WGD and TD,have distinct functional preferences,indicating a strong bias.Dose-sensitive genes are often lost through selection sweep after tandem duplication to ensure normal biological functions.Therefore,the retention of TDGs is often biased.Genes involved in biological processes of Gene Ontology (GO),such as response to endogenous stimulus,response to stress,and response to chemicals,were over-retained after tandem duplication in tomato (Fig.3-A).Besides,the genes related to peroxidase activity,oxidoreductase activity,and hydrolase activity participate in redox reactions in plants and respond to changes in the external environment by regulating the synthesis or content of hormones (Loughranet al.2014;Mindreboet al.2016;Sahaet al.2016).Large TDG clusters (gene number＞6) in tomato are biased to retain genes involved in the response to endogenous stimulus (Fig.3-A).The duplication of these genes contributes to the adaption of tomato to cope with environment changes or infection by pathogens,and is retained under natural selection.Similarly,genes involved in stress responses,material transportation,and enzyme activity were also enriched in the coffee genome that did not undergo theTevent(Fig.3-B),and this reflects the consistency of the TDG amplification rules.

We identified 5 651 paralogous genes that resulted from the WGD event.These paralogous genes were located in different sub-genomes of tomato and were retained after theTevent.WGD genes preferentially retain dose-sensitive genes such as those with GO terms DNA-binding transcription factor activity and regulation of metabolic processes (Fig.3-A).DNA binding transcription factors participate in many basic processes of higher plants and are expressed in most plant tissues.For example,the transcription factors RIN and DOF regulate tomato maturation and the formation of vascular tissue,respectively (Zhuet al.2008;R?thet al.2017;Rojas-Graciaet al.2019).According to the gene balance hypothesis,these dose-sensitive genes tend to be retained after WGD (Birchler and Veitia 2012),which is consistent with our enrichment results.

Similarly,KEGG pathway enrichment analysis also verified this point.TDGs in tomato are enriched in a large number of hormone pathways related to stress response(Fig.3-C).The MAPK signaling pathway cascades are highly conserved signaling modules downstream of receptors/sensors that transduce extracellular stimuli into intracellular responses in plants (Meng and Zhang 2013).TDGs in the coffee genome were similarly enriched in KEGG pathways such as the MATE family and chitinases(Fig.3-D).Transcription factors,GTP-binding proteins,and other dose-sensitive genes in tomato also showed excessive retention in WGD events (Fig.3-C).

Fig.3 Functional enrichment of tandem duplicated genes (TDGs) and whole genome duplication (WGD) genes.A,the top section shows the significantly enriched GO terms for all TDGs and WGD genes in tomato,and the bottom section shows the significantly enriched GO terms for large TDG clusters (gene number ＞6) in tomato.The length of each bar indicates the ratio of enriched genes to background genes.The TDGs and WGD genes are represented by green and red,respectively.P-adjust corresponds to different shades of colors.B,significantly enriched GO terms for the TDGs in coffee.C,significantly enriched KEGG pathways for the TDGs and WGD genes in tomato (similar to the illustration in A).D,significantly enriched KEGG pathways for the TDGs in coffee.

3.4.A case of TDG functional innovation through domain rearrangement

The classical model of gene evolution in plants proposes that new genes arise,in large part,through the duplication of existing loci or genomic regions,which then diverge to acquire new functions (Panchyet al.2016).We identified a large number (154) of NBS-LRR class resistance genes among the tomato TDGs.It is well known that the NBSLRR domains are highly variable and tend to be under diversifying selection to adapt to continually changing pathogen proteins (Meyerset al.1998;Michelmore and Meyers 1998).However,we found a TDG cluster that was continuously duplicating and did not encode the NBSLRR domain in tomato,and this TDG cluster was newly functionalized by gene fusion.Fusarium wilt of tomato causes high yield losses,and studies have successfully identified and cloned the resistance genesI2andI3(Oriet al.1997;Ajilogba and Babalola 2013;Catanzaritiet al.2015).I3encodes the S-receptor-like kinase gene and is located in a TDG cluster comprised of 15 genes.To reveal the duplication pattern of this TDG cluster,we found that the large TDG cluster in tomato corresponded to two separate gene clusters inC.canephora.Distinct functional domains were contained in these two gene clusters (Fig.4-A).We found a putative gene fusion event in the tomato TDG cluster,which contributes to the identification of the complicated gene cluster derived from the two clusters.Detailed domain analyses show that this fused gene,Solyc07g055560,possessed both of the domains inherited from the two gene clusters.We divided theSolyc07g055560into two pseudogenes based on its domain position to illustrate the amplification patterns of the two TDG clusters.According to the phylogenetic tree of cluster 1,we found that all tomato genes clustered withpseudogene1as the root,and all coffee genes clustered withGSCOCT00016928001as the root of another branch (Fig.4-B).The phylogenetic tree of cluster 2 showed that tomato geneSolyc07g055690,coffee geneGSCOCT00016940001,andpseudogene2were distributed on the outermost part of the tree.Other coffee genes and tomato genes were divided into two clusters,withGSCOCT00016935001andSolyc07g055670as the roots,respectively (Fig.4-B).

Based on the phylogenetic tree (Fig.4-B),thepseudogene1in the tomato genome was the ancestor of the gene cluster carrying the P450 superfamily domain,whilepseudogene2,Solyc07g055670,andSolyc07g055690were the ancestors of another gene cluster.The two ancestor genes duplicated,and this resulted in two nested gene clusters.Then,the adjacentpseudogene1andpseudogene2fused to formSolyc07g055560(Fig.4-C).By querying the public RNAseq data (Wenet al.2019),we confirmed that the gene could be covered by sequencing reads (Appendix F),ruling out the possibility of incorrect annotation of the reference genome.These results demonstrate that tandem gene clusters can serve as generators of novel gene fusions that innovate the formation of complex loci and/or the inactivation of genes that can ultimately affect ecologically relevant traits.

Fig.4 A tandem duplicated gene (TDG) cluster in tomato that has undergone gene fusion.A,the conserved domains and physical position of the TDG cluster before and after gene fusion.The left section shows the TDG cluster of tomato after whole genome duplication (WGD),which contains the fused gene (indicated in red) and the fusarium wilt resistance gene I3 in tomato.The right section shows two clusters of genes detected from the Coffea canephora genome before the gene fusion event.B,the phylogenetic tree after Solyc07g055560 was divided into two pseudogenes.C,the patterns of tandem duplication of tomato and coffee clusters.The squares represent genes,and the different colors in the squares represent the domains carried by the genes,which correspond to the domains in A.

4.Discussion

WGD amplifies all genes in the genome in a balanced manner.This may be beneficial for altering the entire biological pathway and processing (Birchler and Veitia 2007) and is associated with longer halflives of the resulting gene duplicates (Lynch and Conery 2000).However,it is not clear whether these advantages outweigh the greater availability of new tandem and proximal duplication that can result from drastic environmental changes.Different types of gene duplication lead to the retention or loss of new duplicates,so that some gene families expand while others do not.The consistency between our functional analysis (Fig.3)of different types of genes and the balanced gene drive hypothesis indicates the expansion and stability of gene families during evolution (Freeling 2009).

Classical theories of population genetics suggest that duplicated genes have identical sequences immediately following duplication and then gradually diverge over evolutionary time (Panchyet al.2016).Structural divergence between tandem duplicates increases with time,and this is consistent with classical theory (Fig.2-A).Since most TDGs are younger than the recently generated WGD gene pairs (as most TDGs have lowKS-values),it was difficult to determine whether a TDG has also experienced WGD in this study.We used species that had not experienced additional WGD events,and this might represent the genomic state of Solanaceae plants prior to theTevent.

Multigene families with tandemly located gene members are especially prone to rearrangements,including duplications,deletions,and more complex rearrangements.These can produce chimeric genes that allow for the origin (and loss) of new functionalities and novel genes (Caicedoet al.2009).Tandem duplication and segmental duplication are major factors in R gene amplification.Duplication,sequence variation due to mutation,disparity in the length of leucine-rich repeats,and rearrangement of different domains can lead to variations within and between R genes.These genes can exhibit variations in recognition specificity that are sufficient to resist new pathogens (McDowell and Simon 2006;Joshi and Nayak 2013;Liuet al.2020).We presented an example of a gene that does not encode an LRR domain but is derived from the fusion of adjacent tandem genes (Fig.4).The proximity of these different TDG distributions may give rise to a variety of gene functional innovations.The tomato genome underwent two WGT events,and its internal chromosome rearrangement and gene losses were more severe than in coffee,which did not undergo theTevent.Although coffee and tomato genes are differentiated,collinear segments within the genomes are likely to remain connected.Using theI3TDG cluster as an example,the P450 superfamily domain in the coffee and tomato clusters underwent considerable tandem duplication (Fig.4-A).Many P450s are involved in the syntheses of pigments,antioxidants,structural polymers,and defense-related compounds such as flavonoids,phenylpropanoids,and phenolic esters.Some P450s are also involved in the detoxification of pollutants,herbicides,and insecticides (Yuet al.2017).The PAN_APPLE domain,which is widely duplicated in the coffee genome,tends to be lost (or not copied) in the tomato genome (Fig.4-AandC).Two sets of the B_lectin superfamily and PKc_like superfamily domains were retained in theI3gene,which may indicate thatI3also experienced a gene fusion event and retained its important function of resisting fusarium wilt disease.The fused geneSolyc07g055560may also have similar functions,but this needs to be verified.

5.Conclusion

TD and WGD are important forces in plant evolution.In this study,we identified TDGs of Solanaceae core species and explored their relationships with theTevent.The TDG outbreak event of Solanaceae species occurred relatively recently,and the content of large TDG clusters was less than is found in non-Solanaceae species that did not undergo theTevent.Some genes have accumulated and retained more TDGs after long-term evolution,while some smaller TDG clusters appear to have been copied recently.The gene bias retention is consistent with the evidence that TDGs retained more hormone-related stress resistance genes,while the WGD genes retained more dose-sensitive genes.By revealing the occurrence of tandem repeats and the fusion of adjacent TDGs,we provide new insights into the development of gene functionalization.Our study provides data useful for further exploration of the genome evolution of Solanaceae.

Acknowledgements

The work was supported by the National Natural Science Foundation of China (NSFC;31972411 and 31722048),the Program for Scientific and Technological Innovative Talents in Universities of Henan Province,China(20HASTIT035),the Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences,and the Key Laboratory of Biology and Genetic Improvement of Horticultural Crops,Ministry of Agriculture and Rural Affairs,China.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm