Virus Evolution, 2021, 7(1): veab017 doi: 10.1093/ve/veab017 Research Article Increasing temperature alters the within-host competition of viral strains and influences virus genetic variability Cristina Alcaide,1,† Josep Sardanyés,2,3 Santiago F. Elena,4,5,‡ and Pedro Gómez1,*,§ 1 Departamento de Biologı́a del Estrés y Patologı́a Vegetal, Centro de Edafologı́a y Biologı́a Aplicada del Segura (CEBAS), CSIC, PO Box 164, 30100 Murcia, Spain, 2Centre de Recerca Matemàtica (CRM), Edifici C, Campus de Bellaterra, Cerdanyola del Vallès, Barcelona 08193, Spain, 3Dynamical Systems and Computational Virology Associated Unit Instituto de Biologı́a Integrativa de Sistemas (I2SysBio) - CRM, Edifici C, Campus de Bellaterra, Cerdanyola del Vallès, Barcelona 08193, Spain, 4I2SysBio, CSIC-Universitat de València, Paterna, 46980 València, Spain and 5The Santa Fe Institute, Santa Fe, NM 87501, USA *Corresponding author: E-mail: pglopez@cebas.csic.es † https://orcid.org/0000-0002-7482-9810 ‡ https://orcid.org/0000-0001-8249-5593 § https://orcid.org/0000-0003-2830-4105 Abstract Environmental conditions can affect viral accumulation, virulence and adaptation, which have implications in the disease outcomes and efficiency of control measures. Concurrently, mixed viral infections are relevant in plants, being their epide- miology shaped by within-host virus–virus interactions. However, the extent in which the combined effect of variations in abiotic components of the plant ecological niche and the prevalence of mixed infections affect the evolutionary dynamics of viral populations is not well understood. Here, we explore the interplay between ecological and evolutionary factors during viral infections and show that isolates of two strains of Pepino mosaic potexvirus coexisted in tomato plants in a temperature-dependent continuum between neutral and antagonistic interactions. After a long-term infection, the muta- tional analysis of the evolved viral genomes revealed strain-specific single-nucleotide polymorphisms that were modulated by the interaction between the type of infection and temperature. These results suggest that the temperature is an ecologi- cal driver of virus-virus interactions, with an effect on the genetic diversity of individual viruses that are co-infecting an in- dividual host. This research provides insights into the effect that changes in host growth temperatures might have on the evolutionary dynamics of viral populations in mixed infections. Key words: plant virus, evolutionary ecology, genetic variability, mixed infections; environmental factors. 1. Introduction and political consequences (Vurro, Bonciani, and Vannacci 2010; Plant viral diseases cause widespread epidemics in crops and hamper the Lefeuvre et al. 2019). Among many others, notable examples of viral dis- sustainability of food production, with severe ecological, socio-economic, eases affecting crops are Tomato yellow leaf curl begomovirus, which is C The Author(s) 2021. Published by Oxford University Press. V This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unre- stricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 1 2 | Virus Evolution, 2021, Vol. 7, No. 1 devastating tomato production worldwide (Moriones and Navas-Castillo balance (Novella et al. 1995; Salda~ na et al. 2003; Froissart et al. 2004; 2000), Tomato spotted wilt orthotospovirus which re-emerged affecting Andino and Domingo 2015). It is therefore that multiple infections may more than 800 plant species, Yellow mottle sobemovirus disease affecting lead to an additional evolutionary advantage of rare variants, increasing irrigated rice (Fargette et al. 2008), and recently, the cassava mosaic gem- genetic diversity, as selection may favor those genotypes that require iniviruses that are threatening cassava production (Legg et al. 2015). The fewer stabilizing effects to allow coexistence (Lankau 2009). emergence and spread of infectious viral diseases is associated to intrinsic Together, all these aspects emphasize that the impact of mixed infec- viral and host components, as well as to environmental, agronomical and tions on the virus population genetic variability could be largely contin- socioeconomical factors (Anderson et al. 2004; Rojas and Gilbertson gent upon host ecology. This is particularly important since mixed 2008; Elena, Fraile, and Garcı́a-Arenal 2014; González, Butkovic, and infections may allow diverse and complex interactions within-host, which Elena 2019; González, Butkovic, and Elena 2020; González et al. 2021). in turn, could be affected by abiotic factors. Then, it is possible that tem- While environmental heterogeneity can shape plant–virus infection net- perature affect the strength and specificity of virus-virus interactions, works at spatio-temporal scales, which seems to be a key driver of the dis- influencing differentially the viral fitness, and hence, the evolutionary dy- ease emergence and dynamics (Hily et al. 2016; Rodrı́guez-Nevado et al. namics of those viruses that are simultaneously infecting the same host. 2018; Valverde et al. 2020), the impact of a fundamental plant niche di- Despite this is an epidemiologically relevant situation, there is currently mension, the environmental temperature, on virus evolution is unclear. It no information on the potential effect of within-host interactions between has been shown experimentally that symptom expression and virus accu- strains/genotypes of the same virus in the evolutionary dynamics of viral mulation can be influenced by temperature, with temporal fluctuations populations under different environmental conditions. during the plant infection process (Obre R palska-SteR plowska et al. 2015; In this study, we examined the interplay between temperature and Chung et al. 2016). In fact, changes in temperature affect the host plant– mixed infections using two isolates of pepino mosaic virus (PepMV; spe- virus interaction (Harrison 1956; Kassanis 1957) and have been corre- cies Pepino mosaic potexvirus, genus Potexvirus, family lated with modifications in host gene expression and metabolism changes Alphaflexiviridae). Both isolates, belonging to the European (EU) and that lead to variations on plant growth (Went 1953), subsequent symp- Chilean (CH2) strains (Gómez et al. 2009; Alcaide et al. 2020a), were tom severity and within-host virus accumulation (Király et al. 2008; used as a model of close-related viruses that naturally infect tomato Aguilar et al. 2015; Obre R palska-Ste R plowska et al. 2015; Chung et al. 2016; plants. PepMV genome consists of a positive-sense, single-stranded RNA Honjo et al. 2020). In this sense, it is thought that virus accumulation is molecule of approximately 6.4 kb in length, containing five open reading temperature-dependent, which in fact has been recently shown in natural frames (ORFs) flanked by two untranslated regions, with a poly(A) tail conditions, revealing that seasonality can affect virus–plant interaction at the 30 end of the genomic RNA (gRNA). This virus naturally spreads and virus dynamics during persistent infection (Honjo et al. 2020). Thus, through mechanical transmission on tomato plants and causes severe it is likely that temperature can impact on viral adaptive evolution by economic losses worldwide (Hanssen and Thomma 2010; Gómez-Aix both influencing the genetic variability and the effective population size. et al. 2019a; Alcaide et al. 2020a). Our previous studies show that the co- However, little is known about how and to what extent temperature can occurrence of both EU and CH2 types has been associated for long with affect viral genetic diversity within-host, and how it could affect the eco- recurrent epidemics in Spain (Gómez et al. 2009; Gómez-Aix et al. 2019a; evolutionary dynamics of viral populations. Alcaide et al. 2020a). In addition, despite viral accumulation may depend In parallel, multiple viral infections (i.e. mixed infections) have con- on the tested virus isolate as well as the analyzed plant host, it has been siderable significance on plant crops (Syller 2012; Syller and Grupa 2016; described that geographically different CH2 isolates had similar decrease Tollenaere, Susi and Laine 2016; Alcaide et al. 2020b; Moreno and in their accumulation levels in the presence of an EU type isolate in two López-Moya 2020). Viruses can interact when they are infecting the same different tomato cultivars, showing that the CH2 type is readily antago- host. In general, we can divide these interactions in synergistic, neutral nized by the EU type (Gómez et al. 2009 ; Alcaide et al. 2020a). We hy- and antagonistic, although they can be much more complex (Martı́n and pothesized that temperature could affect this virus–virus interaction, Elena 2009; Syller 2012; Syller and Grupa 2016; Alcaide et al. 2020b). which in turn could shape the within-host genetic diversity of viral popu- Thus, viruses can respond to the presence of other viruses co-infecting lations. To test this hypothesis, we experimentally inoculated tomato the same host, where selection processes through new environmental and plants at 20 C and 30 C with two isolates of PepMV belonging to the biotic conditions might shape viral population dynamics compared to EU and CH2 strains, either in single and mixed infections to assess the single infections (Barraclough 2015; Susi et al. 2015). Hence, within-plant extent in which temperature and mixed infections affect the within-host virus–virus interactions could have causal effects on the population struc- accumulation and genetic-variability of PepMV. ture and epidemiology during an infection (Alcaide et al. 2020b; Sallinen et al. 2020), leading to major consequences on virulence and fitness of the involved viruses (Read and Taylor 2001; Hodgson et al. 2004; de Roode 2. Materials and methods et al. 2005; Tollenaere, Susi and Laine 2016). Within this context, viral 2.1 Plant growth conditions, virus inoculation and co-infections could also affect the mutational load, which for example in sampling the case of the segmented bacteriophage u6 appears to be purged slower in co-infections than in single infections due to complementation of less- Two infectious PepMV clones, PepMV-Sp13 isolate (EU type) (Aguilar fit genotypes and defective mutants (Froissart et al. 2004). However, sim- et al. 2002) and PepMV-PS5 isolate (CH2 type) (Gómez et al. 2009; ilar genetic diversity in protein-coding regions was observed in popula- Sempere et al. 2011) were used to agroinfiltrate Nicotiana benthamiana tions of u6 at high and low levels of co-infection (Dennehy et al. 2013). Domin plants separately, in order to obtain purified virions of both iso- This difference can be attributed to the type of mutations assessed. Only lates (Gómez et al. 2009). After 14 days post-inoculation (dpi), viral par- deleterious mutations were considered by Froissart et al. (2004), while ticles were purified from the homogenized plant tissue, following a series Dennehy et al. (2013) included beneficial and neutral mutations, with of centrifuges and a PEG precipitation (AbouHaidar, Xu and Hefferon beneficial mutations maintained longer in asexual populations. 1998; Gómez et al. 2009). Seeds of tomato plants (Solanum lycopersicum Moreover, once new variants emerge in a population, viral interference L. cv. Money Maker, as a popular cultivated tomato variety) were pre- along with the segregation of the variants into different cell tissues could germinated in a greenhouse with a photoperiod of 16 h light:8 h dark and allow coexistence within-host either by a genetic trade-off between com- a temperature between 22  C and 26  C until seedling emergence. Then, petition and colonization (Ojosnegros et al. 2010), or minimizing the two sets of 33 seedlings were placed into a greenhouse with controlled strength of selection until reaching a within-host drift-mutation-selection temperature conditions at either 20  C or 30  C during 2 weeks. Both C. Alcaide et al. | 3 temperatures were selected based on the knowledge that, in general, the cent agarose gel. Two PCRs were made per sample (duplicates), mixed optimum temperature for tomato cultivation is 25  C, and plant growth R and purified using GENECLEANV Turbo Kit (MP Biomedicals). rate seems to be affected above and below of this temperature (Shamshiri et al. 2018). After 30 days post germination and for each temperature 2.4 Next-generation sequencing of full-length viral condition, six tomato plants were mock-inoculated and nine plants per genomes treatment were mechanically inoculated either with EU type, CH2 type or both isolates simultaneously in a proportion 1:1. Virus inoculations After full-length amplification, a total of 26 cDNA fragments were se- were carried out on the third and fourth true leaves by rubbing quenced: six cDNA fragments from EU type single infections (three from Carborundum-dusted and a suspension of virions particles in sodium plants grown at 20 C and other three from plants grown at 30 C), six phosphate buffer (30 mM), as previously described (Gómez et al. 2009). from CH2 type single infection (three for each temperature), twelve from During the experiment, PepMV infection was checked by the detection of mixed infected plants (three for each isolate and temperature), and also both type strains from a leaf-sample of each plant collected at 12, 30 and both genomes from the disassembled virions used as starting inoculum 58 dpi. Total RNA was extracted using Tri-reagent, and 1 ml of RNA was were included. Library preparation was performed using the Illumina placed on a nylon membrane and fixed with ultraviolet light (CL-1000 Nextera XT library preparation kit. Then, libraries were sequenced on Ultraviolet Crosslinker). PepMV infections were detected by dot-blot the Illumina MiSeq platform, using 250 bp paired-end sequencing reads. molecular hybridization using specific RNA probes to specifically detect EU and CH2 types (Gómez-Aix et al. 2019b), following the Roche proto- 2.5 Next-generation sequencing data analysis col (RNA labelling and detection kit), and revealing the membrane in a The quality of the raw data was determined using FastQC (Andrews chemiluminescent detector Amersham Imager 600. Additionally, another 2014), and then trimmed by Trimmomatic (Bolger, Lohse, and Usadel set of tomato plants was mechanically inoculated with EU or CH2 types 2014), which allowed us to remove adapters and trimming low quality in single and mixed infections (five plants per treatment) under similar bases (Phred score > 30). The quality reads were mapped against the ref- conditions in order to estimate viral accumulation at early stages of the erence genome using bowtie2 (Langmead and Salzberg 2012) and align- infection (7 dpi). ments were processed by SAMtools (Li et al. 2009). The SNP calling was performed with different variant callers: LoFreq (Wilm et al. 2012), 2.2 Quantitative in Planta viral infections FreeBayes (Garrison and Marth 2012) and SNPGenie (Nelson, Moncla For two different groups of plants, all apical leaves, except the basal inoc- and Hughes 2015). Typically, we considered as true variants all those ulated leaves, of each plant within a group were collected to estimate the with a frequency >1 per cent, with a minimum raw depth of 3,000 and viral RNA accumulation either at 7 or 60 dpi. Reverse-transcriptase real- neglected those variants that were already present in the samples used as time quantitative polymerase chain reaction (RT-qPCR) was performed starting inocula. Finally, SNPGenie software was also used for estima- by using the Power SYBR Green RNA-to-CT 1-Step Kit (Applied tion of genetic variability indexes such as nucleotide diversity (p; mean Biosystems), including three technical replicates per sample and following number of pairwise nucleotide differences per site across the whole ge- the manufacturer’s recommendations. Ten-fold serial dilutions of viral nome) and observed heterozygosity (H; mean gene diversity at all poly- RNA from the disassembled EU and CH2 virions were stocked at morphic nucleotide sites in the genome). 80  C and used to generate standard curves during the absolute quanti- fication by RT-qPCR. The copy number of PepMV genomes was calcu- 2.6 Statistical analysis lated from the Ct values and using the molecular weight of PepMV Exploratory analyses were done with SPSS Statistics version 26 (IBM genome per ng of total RNA. Note that in mixed infection treatment two Corp., Armonk, USA) using generalized linear models (GLM). Viral RT-qPCR had to be done per sample, one with EU specific primers and load (VL) data were fitted to a model that incorporates three orthogonal another with those of the CH2 (Gómez et al. 2009). factors, one random nested factor, and one covariate. The orthogonal factors were the type of isolate (A 2 fEU, CH2g), the type of infection (I 2.3 Full-length viral amplification 2 fsingle, mixedg) and the experimental temperature (T 2 f20, 30g  C). Plant material was collected at 60 dpi. The top eight leaves of plants were The five replicate plants (R 2 f1, . . ., 5g) was nested within the interac- harvested and grinded in a mortar using liquid N2. Although, analysis of tion of the three orthogonal factors. Time was used as covariate (t 2 f7, individual samples may be more reliable, in this case, three pools were 60g dpi). The full model equation reads: made per treatment, with three plants per pool. A sample from each pool was taken and RNA extraction was performed using RNeasy Plant Mini VLijklm ðtÞ  k þ t þ Ai þ t  Ai þ Ij þ t  Ij þ Tk þ t  Tk þ ðA  I Þij Kit (QIAGEN) and stored at 20  C until use. One microgram of total þ t  ðA  I Þij þ ðA  T Þik þ t  ðA  T Þik þ ðI  T Þjk þ t RNA was used for retro-transcription (RT) (Expand Reverse  ðI  T Þjk þ ðA  I  T Þijk þ t  ðA  I  T Þijk Transcriptase, Roche). In the case of mixed infections, RT reactions were þ RðA  I  T Þijkl þ t  RðA  I  T Þijkl þ eijklm ðtÞ; performed in duplicate, one of them using a specific primer for the EU (1) strain (EU_rv: 50 -TTTTTTTTTTTTTTTTTCAAAGAAATAATTA-30 ) and another using a specific primer for the CH2 strain (CH2_rv: 50 - where k corresponds to the grand mean value of VL and e represents the TTTTTTTTTTTTTAGTAGATTTAGATACTA AG-3’). Then, cDNA sampling error. A Gamma distribution with a log-link function was as- was amplified using Phusion High-Fidelity DNA Polymerase (Thermo sumed based on its lowest BIC ¼ 5347.799 (Akaike’s weight W ¼ 1) com- Fisher Scientific) and specific primers for each isolate (EU_rv and pared to a Tweedie distribution with log-link (BIC ¼ 5415.938, EU_fw: 50 -CGCGGATCCGGAAAACAAAATAAATA W ¼ 1.599  1015) or identity-link functions (BIC ¼ 5584.627, AATAAATATAC-3 for EU and CH2_rv and CH2_fw: 50 - 0 W ¼ 3.745  1052) and with a Normal distribution with an identity-link GAAAACAAAACATAACACATAATATCAAAAGTGACC-30 for function (BIC ¼ 5979.395, W ¼ 7.090  10138). CH2). The initial denaturation was done at 98 C for 1 min, continued by The count of SNPs (SC) per sampled population 60 dpi were fitted to 20 cycles of denaturation at 98 C for 10 s, annealing at 48 C for 30 s and a model that incorporates A, I and T as orthogonal factors and uses three extension at 72 C for 3 min and 30 s, followed by a final extension at replicate plants to estimate the error. In this case, the model equation 72 C for 7 min. Amplification was checked by electrophoresis in 1 per reads 4 | Virus Evolution, 2021, Vol. 7, No. 1 SCijkl  r þ Ai þ Ij þ Tk þ ðA  I Þij þ ðA  T Þik þ ðI  TÞjk Third, the tree-ways interaction between the main factors was not sig- þ ðA  I  T Þijk þ nijkl ; (2) nificant per se, but only throughout a dependence with the duration of in- fection (Table 1). In the case of EU, VL was larger for mixed infections where r corresponds to the grand mean value of SC and n represents the at 7 dpi (the precise magnitude depending on T), while at 60 dpi it was sampling error. A Poisson distribution with log-link function was as- larger for single infections (again, the precise magnitude depending on sumed based on its lowest BIC ¼ 165.554 (W ¼ 0.793) compared to a T). The opposite pattern was observed for CH2: VL was larger in single Tweedie distribution with log-link (BIC ¼ 169.620, W ¼ 0.104) or infections 7 dpi and slightly larger in mixed infections after 60 dpi at identity-link functions (BIC ¼ 169.620, W ¼ 0.104) and a negative 20  C, yet the reverse was found at 30  C. Binomial with log-link function (BIC ¼ 218.456, W ¼ 2.579  1012). The significance of each term in the models was evaluated using a likelihood-ratio test (LRT) which asymptotically follows a v2 distribu- tion. In addition, the magnitude of the effect associated with each term 3.2 Mutation occurrence from full-length genome was evaluated using the g2P statistic which measures the proportion of to- sequencing tal variability in the trait attributable to each factor in the model. To examine the influence of differences in accumulation and competition Conventionally, values of g2P  0.15 are considered as large effects between CH2 and EU on the viral genetic variability within-populations, (Cohen 1988). Goodness-of-fit of the two fitted GLM models was evalu- we first identified a genomic region at the 30 end of the viral RNA that ated using McFadden’s pseudo R2 (McFadden 1974). allows to distinguish both PepMV types and to amplify full-length genomes from mixed infected samples. This allowed us to obtain whole- 2.7 Data availability genome sequences of each PepMV populations after 60 dpi, as well as in All sequencing information and data that support the findings of this the inocula. After cleaning and trimming the raw NGS data, we obtained study have been deposited in the NCBI GenBank with the BioProject between 809,756 and 1,777,280 reads per sample, resulting in a theoreti- code PRJNA639566, under accession numbers SRR12017758 - cal average coverage ranging from 31,750 to 69,687. These reads were SRR12017783. Data obtained from all the analyses performed in this further filtered out by quality, and using the initial inocula as reference, work are available from the corresponding author upon reasonable we identified SNPs (and evaluated their frequency in the samples) that request. appear during the infection experiments. After excluding SNPs present in the inocula and considering variants with a frequency >0.01, including 3. Results those variants found in the overlapping region between TGB2 and TGB3 genes as double variants because of the potential effect in both proteins, 3.1 Identifying significant drivers of PepMV we end up with a total of 519 SNPs (Supplementary Table S1, Figs. 2A accumulation and B). No indels or recombination events were observed. Attending to To study whether the PepMV isolates could respond differentially during the frequency of the new mutations, the vast majority were ranging from the infection process, we first investigate how temperature (T), type of in- 0.01 to 0.1, with only 3.28 per cent of the total SNPs displaying frequen- fection (I) and genetic differences among PepMV isolates (A) affected cies > 0.5 (Supplementary Table S1). Note that all these high-frequency gRNA accumulation in tomato plants at 7 and 60 dpi (Fig. 1A). The viral mutations were specific to the CH2 populations. load (VL) data were fitted to the Equation (1). Table 1 shows the statisti- SNP counts (SC) of PepMV within-host populations are summarized cal summary of the corresponding GLM analysis. Only the interaction in Fig. 3. These data were fitted to the model described by Equation (2) term A  T was not significant. However, a number of terms in Equation and the results of the analyses are shown in Table 2. Overall, and focus- (1) (Table 1, bold) had low statistical power (1  b < 0.800), thus ing first on the main effects, highly significant differences of large magni- wrongly failing to reject the null hypothesis (type II error), and the mag- tude existed between the average number of mutations fixed by the CH2 nitude of the effect classified as small according to the g2P < 0.15 crite- and EU within-host populations. CH2 populations accumulated 1.57- rion. Thus, hereafter we comment on highly significant and relevant fold more mutations than the EU populations (Fig. 3). No differences terms. First, regarding the main factors, only the A and T were relevant existed among single and mixed infections (I term in Table 2) in terms of by themselves to explain the observed variability in VL (Table 1, Fig. 1B the number of mutations accumulated. The average number of mutations and C). Overall, CH2 accumulates to higher concentrations than EU. per within-host population also significantly increased 1.57-fold with Remarkably, the interaction between the three main factors and t was temperature (T term in Table 2 and Fig. 3). The higher-order terms A  I also highly significant, indicating that the observed effects changed with and A  IT in Equation (2) were also significant (P  0.011) and their time. While VL was 3.1-fold larger for EU 7 dpi than for EU 60 dpi and effects could be considered of large magnitude (g2P  0.299), suggesting this pattern holds in mixed infections, it was 8.0-fold larger for CH2 that the observed effect associated with the viral genotype indeed would 60 dpi than for CH2 7 dpi. VL was, on average, larger in single than in depend on its interaction with the type of infection and temperature in a mixed infections, but the magnitude of this difference slightly decreased nonlinear way. However, given that the power of the two tests was 1  along the progress of infection (1.8-fold at 7 dpi, and 1.3-fold at 60 dpi). b < 0.8, this conclusion should be taken carefully (Table 2). At 7 dpi, infections at 20  C were 3.3 times more productive than at Furthermore, genetic diversity of CH2 and EU populations was evalu- 30  C, slightly less (2.7-fold) at 60 dpi. ated by using heterozygosity (H) and nucleotide diversity (p) estimators Second, the only relevant pairwise interaction was A  I (Table 1, (Table 3). We found that both, H and p, were higher in the CH2 than EU Fig. 1B and C): EU accumulated 1.6-fold more in mixed than in single populations, with averaged H values higher at 20 C in mixed infections. infections, whereas for CH2, the situation was the reverse: it accumulated However, averaged p values were higher at 30  C in mixed infections of 4.2-fold more in single than in mixed infections. Indeed, these differences the CH2 populations that, in turn, was consistent with significant strongly depend on t: while EU productivity in mixed infections was 2.6- increases of the numbers of synonynous than non-synonymous mutations fold higher 7 dpi, it was only 1.9-fold higher 60 dpi; in the case of CH2, it (Table 3). Hence, conservatively, we can conclude that populations of accumulated 8.9-fold more in single infections 7 dpi but 1.2-times more in both PepMV types differed in their mutational load in an infection- and mixed infections 60 dpi. temperature-dependent manner. C. Alcaide et al. | 5 Figure 1. (A) Predicted effect of mixed viral infections on the fitness and genetic diversity of viral populations within an individual host. (B and C) Barplots showing the viral load (viral genomes/ng RNA total) of each PepMV (CH2 and EU) strain in tomato plants grown at 20  C and 30  C under single (black) and mixed infection (grey) condition. Viral accumulation was inferred by absolute quantification using RT-qPCR after 7 dpi (B) and 60 dpi (C). 4. Discussion hardly inform about the link between environmental factors and virus evolution during epidemic processes. In our experiments, where tempera- The interactions between viruses and their host plants are expected to ture was controlled and PepMV infections checked at different time- change by altering temperature (Kassanis 1957; Obre R palska-Ste R plowska points, we found that the magnitude and sign of the interactions between et al. 2015; Honjo et al. 2020). In parallel, accumulating evidence shows a both isolates appeared to vary according to time and temperature. In the high prevalence of mixed viral infections in plants (Syller 2012; Syller and early stage of the infection, 7 dpi, EU is facilitated by CH2 in mixed infec- Grupa 2016; Tollenaere, Susi and Laine 2016; Alcaide et al. 2020b; tions, whilst CH2 is strongly antagonized by EU at both temperatures. Moreno and López-Moya 2020). In this work, we have explored the com- However, at the end of the experiment, CH2 showed a higher viral load bined effect of a rise in temperature and mixed viral infections. We have than EU in single and mixed infections (Fig. 1), and either CH2 or EU connected both, abiotic and biotic factors, and more explicitly, we have are facilitated at 20  C. Considering these results, along with the efficient sought to explore the relationship between temperature and evolutionary PepMV mechanical transmission in the field, it is possible that the ob- dynamics of viral populations in single and mixed infections. In our stud- served maintenance of the EU type on tomato crops may be mainly due ies, different PepMV-CH2 isolates experienced a fitness cost in the pres- to its dispersion at early states of the plant growth. Whereas, it seems not ence of an EU-type isolate, suggesting that an asymmetrical antagonistic to be the case in the later stages. In this sense, a better understanding of interaction between these two strains provide a partial explanation for the population dynamics at different stages of the plant growth could the maintenance of the genetic structure of the PepMV populations in to- provide valuable information about the causes of establishment and dis- mato (Gómez et al. 2009; Alcaide et al. 2020a). Both strains have been placement of viral strains. co-occurring in mixed infections and their long-term co-circulation may Increasing temperature may influence the accumulation of individual have contributed to the evolutionary dynamics of PepMV populations viruses co-infecting the same host plant and affect their potential viral (Alcaide et al. 2020a). However, environmental factors may also affect interactions. This temperature effect on virus accumulation can be likely the virus genetic diversity, and sequencing data from several field isolates due to the potential alterations on the plant physiological and biological 6 | Virus Evolution, 2021, Vol. 7, No. 1 Table 1. Results of the GLM fitting of the viral load data to Equation resilience to temperature changes may be playing an important role, sug- (1). gesting that this framework could be extended to investigate how the tol- erance of plants to warmer temperatures affects the viral genetic Source of variation LRTa df P g2P 1 – bb variability and competitive interactions in mixed infections. Further re- Full model 698,703 61 <0.001 search can help in anticipating the impact of plant breeding programs for k 5399.943 1 <0.001 0.993 1.000 thermotolerance and how new plant cultivars will drive the eco- t 128.340 1 <0.001 0.483 1.000 evolutionary dynamics of viral populations in plant crops. A 298.137 1 <0.001 0.455 0.889 Most of the information on virus evolution relies on single infections. tA 417.126 1 <0.001 0.913 1.000 However, mixed infections are highly prevalent in nature, thus limiting I 129.466 1 <0.001 0.165 0.342 the value and generality of qualitative and, to a much lesser extent, quan- tI 67.675 1 <0.001 0.352 1.000 titative inferences about emerging viral diseases are gathered from single T 294.315 1 <0.001 0.410 0.826 infections. Viral occurrence and prevalence in plant crops seems to be tT 70.774 1 <0.001 0.352 1.000 largely explained by either neutral, synergistic or antagonistic interac- AI 280.915 1 <0.001 0.416 0.836 tions among co-infecting viral species (Martı́n and Elena 2009; Syller t  AI 260.109 1 <0.001 0.750 1.000 2012; Syller and Grupa 2016; Alcaide et al. 2020b; Moreno and López- AT 3.630 1 0.057 0.003 0.054 Moya 2020). However, whether these interactions show some degree of t  AT 12.626 1 <0.001 0.044 0.649 genetic plasticity in response to environmental stresses has not been IT 71.393 1 <0.001 0.071 0.161 tested so far. This lack of information results from the hardness to distin- t  IT 8.810 1 0.003 0.025 0.422 guish among genetically similar strains of the same virus. This particular A  IT 146.124 1 <0.001 0.176 0.363 bias happened with both CH2 and EU strains. Fortunately, we could t  AI  T 81.264 1 <0.001 0.391 1.000 R(A  IT) 511.786 14 <0.001 0.857 1.000 solve this problem by focusing into a genomic region at the 30 end of the t  R(A  IT) 257.093 14 <0.001 0.779 1.000 viral RNA genome that allowed to obtain the full-length genomes from mixed infections, capturing the nucleotide temporal variation of two re- MacFadden’s pseudo R2 ¼ 0.871. lated strains. Pending a more reliable approach that may explore genetic a Likelihood ratio test, asymptotically distributed as a v2. variability in mixed infections without ruling out variants with very low b Power of the test. Shadowed terms are not further considered because their low statistical frequency within populations, our results can be summarized as follow. power (1  b < 0.800) or small magnitude effect (g2P < 0.150). First, no recombinant variants were found in the mixed infections experi- ments, neither in the previous field tomato surveys since 2005 (Alcaide et al. 2020a). This suggests that inter-strain recombinants seem to be dele- traits (Went 1953). For instance, changes on viral titers of Potato potyvi- terious in the virus populations. Second, SNPs were the only mechanism rus YO and Potato potyvirus A were temperature-dependent, since at contributing to the genetic variability within PepMV populations after high temperatures (25  C–30  C), they were able to accumulate in the long-term infections. Third, the type of infection (single and mixed) did early stages of the infection but disappeared over time (Chung et al. not influence the average population mutational load at 20 C. The last 2016). In the case of Turnip mosaic potyvirus infecting Brassica campest- observation is in accordance with previous studies where the viral popu- ris, viral accumulation was lower at elevated temperatures (33  C) as well lation diversity did not differ between mixed and single infections as an inhibition of symptoms was observed, which can be explained by (Dennehy et al. 2013; Ali and Roossinck 2017). However, the situation an increase of siRNAs involved in the plant silencing pathway (Szittya changed at the higher temperature regardless viral accumulation: both et al. 2003; Chung et al. 2015). However, temperature can also reduce strains displayed higher genetic variation and greater differences in accu- plant resistance against different pathogens; for example the hypersensi- mulation at higher temperatures (Fig. 2 and Fig. 3). In particular, the tive response induced by different R genes against Tobacco mosaic toba- CH2 populations displayed a larger number of SNPs in mixed than in movirus (TMV) and Potato potexvirus X was reduced by an increase of single infections at higher temperature, while the opposite seems to be the the temperature, although no determinants of this temperature sensitivity case for the EU population. Importantly, a rise in temperature in mixed in the plant defence response were identified (Wang et al. 2009). Another infections resulted in a higher occurrence of synonymous than non- study using TMV showed that the stimulated activity of antioxidant synonymous substitutions in the CH2 population. However, in the EU enzymes at elevated temperatures (above 28  C) was associated with the population and under the same conditions, it was observed a reduction suppression of TMV-induced HR-type necrosis and could have a role in of mutations, although the ratio between both type of substitutions the suppression of the resistance at high temperatures (Király et al. remained at similar levels to those in single infections. This suggests that 2008). most of mutations are deleterious, and there is limited understanding of We found that virus–virus interactions varied in a temporal scale in a the long-term impact of those mutations per se. For instance, analyses of temperature-dependent manner, which had an effect on the nucleotide RNA virus species from in vitro and natural populations have found that variation of both viral isolates, depending on the type of infection. genetic variation mostly comprises transient deleterious mutations that Compared to the single infection, the genetic variation of the CH2 popu- are purged by purifying selection (Elena and Moya 1999; Pybus et al. lation increased in the mixed infections at 30  C, while the opposite pat- 2007; Cuevas et al. 2015). However, these mutations could also interact tern was found for the EU population (Fig. 3). However, at 20  C, no through positive epistasis with a relative effect on fitness (Lalic and Elena differences were found in genetic variability between single and mixed 2012; Hillung, Cuevas and Elena 2015; Kutnjak, Elena and Ravnikar infections in both virus populations. Despite the observed differences be- 2017; Minicka et al. 2017) and even do so in a host-dependent manner tween 20  C and 30  C tell us nothing about the shape of the relationship (Lalic and Elena 2013). Thus, uncovering the relationship between those between evolution rate and temperature (i.e. whether it is monotonously mutations found in this study and fitness, and whether they can contrib- increasing or has an optimum or a minimum at some intermediate value), ute to adaptation under temperature and mixed infection conditions, the conclusion that increasing temperature could indeed modify the vi- would be the subject of future investigations. ruses’ evolution rates is still robust. As a summary, temperature likely Theoretical and computational studies have shown the importance of has an impact on the origin and maintenance of the PepMV genetic di- competitive ability and viral fitness trade-off in determining the evolution- versity throughout a relaxation of competitive interactions in mixed ary dynamics of the viral populations (Sardanyés, Elena, and Solé 2008; infections. In this sense, we speculate that plant phenotypic plasticity and Ojosnegros et al. 2010). We speculate that a trade-off between replication C. Alcaide et al. | 7 Figure 2. Occurrence and distribution of sites exhibiting mutations in the full-length genome of both PepMV isolate populations; (A) PepMV-PS5 (CH2 type) and (B) PepMV-Sp13 (EU type). The SNPs are marked as synonymous mutations in single (䊊) and mixed (h) infections, and non-synonymous mutations in single ( ) and mixed ( ) infections for each replicate (n ¼ 3) under 20  C and 30  C conditions. The size of the marker indicates the frequency of each SNP. rate and competition could have influenced this different mutational effect. replication and competition in mixed infections under high temperatures, Under the same framework, it has been theoretically demonstrated that the reducing its genetic variability compared to single infections. Different viral trade-off between replication and competition can affect diversity, with a combinations with different fitness responses (Alcaide et al. 2020b and moderate trade-off promoting diversity (Farahpour et al. 2018). For exam- references therein) may modify the outcomes of this eco-evolutionary dy- ple, it is likely that the EU population had an increased cost between namics, thus emphasizing the need of further research. 8 | Virus Evolution, 2021, Vol. 7, No. 1 Figure 3. Boxplot displaying the average number of SNPs for each PepMV-CH2 and -EU population in single (black) and mixed infections (grey) at 20  C and 30  C of plant growth temperature. The horizontal line shows the median value, and whiskers show the minimum and maximum values. Table 2. Results of the GLM fitting of the mutation count data to Acknowledgements Equation (2). We thank M.C. Montesinos for technical assistance and three anony- Source of variation LRTa Df P g2P 1 – bb mous reviewers for constructive comments. We acknowledge support of the publication fee by the CSIC Open Access Publication Support Full model 67.633 7 <0.001 Initiative through its Unit of Information Resources for Research r 1416.585 1 <0.001 0.955 1.000 A 23.662 1 <0.001 0.478 0.949 (URICI). I 0.403 1 0.526 0.000 0.050 T 23.155 1 <0.001 0.514 0.971 AI 6.993 1 0.008 0.299 0.690 Funding AT 0.991 1 0.320 0.000 0.050 CA was supported by the Ministry of Economy, Industry and IT 0.019 1 0.890 0.013 0.072 Competitiveness (MINECO, Spain) within a PhD programme grant A  IT 6.457 1 0.011 0.311 0.713 (FPU16/02569). This work was supported by the Agencia Estatal de Investigación-FEDER grants AGL2014-59556-R and AGL2017- MacFadden’s pseudo R2 ¼ 0.952. a Likelihood ratio test, asymptotically distributed as a v2. 89550-R to PG and PID2019-103998GB-I00 to SFE. JS has been par- b Power of the test. Shadowed terms are not further considered because their low statistical tially funded by the CERCA Programme of the Generalitat de power (1  b < 0.800) or small magnitude effect (g2P <0.150). Catalunya, MINECO grant MTM2015-71509-C2-1-R, Agencia Estatal de Investigación grant RTI2018-098322-B-I00, and by a Ramón y Cajal contract (RYC-2017-22243). In conclusion, we have observed a tight connection between tempera- ture and viral interactions in mixed infections that can result in a driver Conflict of interest: None declared. of viral diversity in plants. Due to the changing PepMV virus–virus inter- actions over time, a rise in temperature may affect the within-host genetic diversity of viral populations in a strain-dependent manner regardless of References the viral accumulation in mixed infections. A noteworthy remark is that AbouHaidar, M. G., Xu, H., and Hefferon, K. L. (1998) ‘Potexvirus pathogens evolve not only within hosts but also between hosts, and thus, Isolation and RNA Extraction’, Plant Virology Protocols: From Virus the frequency of mixed infections and the success of re-infections in one Isolation to Transgenic Resistance, 81: 131–43. host must be also considered within the evolutionary context. Finally, Aguilar, E. et al. (2015) ‘Effects of Elevated CO2 and Temperature on assessing the mechanistic basis of selective differences will allow us to un- Pathogenicity Determinants and Virulence of Potato Virus derstand the importance of abiotic and biotic factors in wider X/Potyvirus-Associated Synergism’, Molecular Plant-Microbe pathosystems. Interactions, 28: 1364–73. Aguilar, J. M. et al. (2002) ‘Complete Sequence of the Pepino Mosaic Virus RNA Genome’, Archives of Virology, 147: 2009–15. Alcaide, C. et al. (2020a) ‘ ’Long-Term Cocirculation of Two Strains of Supplementary data Pepino Mosaic Virus in Tomato Crops and Its Effect on Population Genetic Variability’, Phytopathology, 110: 49–57. Supplementary data are available at Virus Evolution online et al. (2020b) ‘Implications of Mixed Viral Infections on Plant Disease Ecology and Evolution’, Advances in Virus Research, 106: 145–69. Ali, A., and Roossinck, M. J. (2017) ‘Analysis of Quasispecies Variation Author Contributions in Single and Mixed Viral Infection’, Virus Evolution, 3. CA, SFE and PG conceived and designed the study. CA and PG con- Anderson, P. K. et al. (2004) ‘Emerging Infectious Diseases of Plants: ducted the experiments. CA, JS, SFE and PG analysed data and wrote Pathogen Pollution, Climate Change and Agrotechnology Drivers’, the manuscript. Trends in Ecology & Evolution, 19: 535–44. C. Alcaide et al. | 9 Table 3. Average number of SNPs and estimation of genetic variability indexes (p, nucleotide diversity and H, observed heterozygosity) for each viral strain population grouped by the temperature and type of infection. PepMV strain Temperature ( C) Type of infection syn nsyn nc Total p (104) SD (p) (104) H SD (H) CH2 20 Single 21 44 0 65 4.584 0.240 0.090 0.008 Mixed 26 33 2 61 5.400 0.408 0.107 0.016 30 Single 37 33 2 72 6.591 2.399 0.079 0.043 Mixed 71 35 8 114 6.910 1.335 0.057 0.023 EU 20 Single 25 13 1 39 3.549 0.758 0.049 0.019 Mixed 20 15 1 36 4.264 0.555 0.071 0.018 30 Single 50 31 3 84 5.101 1.191 0.065 0.013 Mixed 29 17 2 48 3.345 0.562 0.043 0.005 Mutations are classified according to the type of change [synonymous (syn) and non-synonymous (nsyn), as well as non-coding variants (nc)]. Andino, R., and Domingo, E. (2015) ‘Viral Quasispecies’, Virology, González, R., Butkovic, A., and Elena, S. F. (2019) ‘Role of Host Genetic 479–480: 46–51. Diversity for Susceptibility-to-Infection in the Evolution of Virulence Andrews, S. (2014) ‘FastQC: A Quality Control tool for High of a Plant Virus’, Virus Evolution, 5: vez024. Throughput Sequence Data’. March 2018. Expand the Limits of Host Phenotypic Plasticity’, Advances in Virus Barraclough, T. G. (2015) ‘How Do Species Interactions Affect Research, 106: 85–121. Evolutionary Dynamics across Whole Communities?’, Annual Review et al. (2021) ‘Plant Virus Evolution under Strong Drought of Ecology, Evolution, and Systematics, 46: 25–48. Conditions Results in a Transition from Parasitism to Mutualism’, Bolger, A. M., Lohse, M., and Usadel, B. (2014) ‘Trimmomatic: A Proceedings of the National Academy of Sciences USA, 118: Flexible Trimmer for Illumina Sequence Data’, Bioinformatics, 30: e2020990118. 2114–20. Hanssen, I. M., and Thomma, B. P. (2010) ‘Pepino Mosaic Virus: A Cohen, J. (1988) Statistical Power Analysis for the Behavioral Sciences, Successful Pathogen That Rapidly Evolved from Emerging to Endemic 2nd edn. Hillsdale, NJ: Lawrence Earlbaum Associates. in Tomato Crops’, Molecular Plant Pathology, 11: 179–89. Chung, B. N. et al. (2016) ‘The Effects of High Temperature on Infection Harrison, B. D. (1956) ‘ ’Studies on the Effect of Temperature on Virus by Potato Virus Y, Potato Virus A, and Potato Leafroll Virus’, The Multiplication in Inoculated Leaves’, Annals of Applied Biology, 44: Plant Pathology Journal, 32: 321–8. 215–26. et al. (2015) ‘Effects of Temperature on Systemic Infection and Hillung, J., Cuevas, J. M., and Elena, S. F. (2015) ‘Evaluating the Symptom Expression of Turnip Mosaic Virus in Chinese Cabbage within-Host Fitness Effects of Mutations Fixed during Virus (Brassica campestris)’, The Plant Pathology Journal, 31: 363–70. Adaptation to Different Ecotypes of a New Host’, Philosophical Cuevas, J. M. et al. (2015) ‘ ’Temporal Dynamics of Intrahost Molecular Transactions of the Royal Society B: Biological Sciences, 370: Evolution for a Plant RNA Virus’, Molecular Biology and Evolution, 20140292. 32: 1132–47. Hily, J.-M. et al. (2016) ‘Environment and Host Genotype Determine the Dennehy, J. J. et al. (2013) ‘Frequent Coinfection Reduces RNA Virus Outcome of a Plant–Virus Interaction: From Antagonism to Population Genetic Diversity’, Journal of Heredity, 104: 704–12. Mutualism’, New Phytologist, 209: 812–22. Elena, S. F., Fraile, A., and Garcı́a-Arenal, F. (2014) ‘Evolution and Hodgson, D. J. et al. (2004) ‘ ’Host Ecology Determines the Relative Emergence of Plant Viruses’, Advances in Virus Research, 88: 161–91. Fitness of Virus Genotypes in Mixed-Genotype Nucleopolyhedrovirus , and Moya, A. (1999) ‘Rate of Deleterious Mutation and the Infections’, Journal of Evolutionary Biology, 17: 1018–25. Distribution of Its Effects on Fitness in Vesicular Stomatitis Virus’, Honjo, M. N. et al. (2020) ‘Seasonality of Interactions between a Plant Journal of Evolutionary Biology, 12: 1078–88. Virus and Its Host during Persistent Infection in a Natural Farahpour, F. et al. (2018) ‘Trade-off Shapes Diversity in Environment’, The Isme Journal, 14: 506–18. Eco-Evolutionary Dynamics’, eLife, 7: e36273. Kassanis, B. (1957) ‘Effects of Changing Temperature on Plant Virus Fargette, D. et al. (2008) ‘Diversification of Rice Yellow Mottle Virus and Diseases’, Advances in Virus Research, 4: 221–41. Related Viruses Spans the History of Agriculture from the Neolithic to Király, L. et al. (2008) ‘Suppression of Tobacco Mosaic Virus-Induced the Present’, PLoS Pathogens, 4: e1000125. Hypersensitive-Type Necrotization in Tobacco at High Temperature is Froissart, R. et al. (2004) ‘ ’Co-Infection Weakens Selection against Epistatic Mutations in RNA Viruses’, Genetics, 168: 9–19. Associated with Downregulation of NADPH Oxidase and Superoxide Garrison, E., and Marth, G. (2012) ‘Haplotype-Based Variant Detection and Stimulation of Dehydroascorbate Reductase’, Journal of General from Short-Read Sequencing’, arXiv:1207.3907. Virology, 89: 799–808. Gómez-Aix, C. et al. (2019a) ‘Genetic Diversity and Population Structure Kutnjak, D., Elena, S. F., and Ravnikar, M. (2017) ‘Time-Sampled of Pepino Mosaic Virus in Tomato Crops of Spain and Morocco’, Population Sequencing Reveals the Interplay of Selection and Genetic Annals of Applied Biology, 174: 284–92. Drift in Experimental Evolution of Potato Virus Y’, Journal of et al. (2019b) ‘In Situ Hybridization for the Localization of Two Virology, 91: e00690–17. Pepino Mosaic Virus Isolates in Mixed Infections’, Journal of Lalic, J., and (2012) ‘Magnitude and Sign Epistasis among Virological Methods, 267: 42–7. Deleterious Mutations in a Positive-Sense Plant RNA Virus’, Heredity, Gómez, P. et al. (2009) ‘Mixed Infections of Pepino Mosaic Virus Strains 109: 71–7. Modulate the Evolutionary Dynamics of This Emergent Virus’, Journal , and (2013) ‘Epistasis between Mutations is Host-Dependent of Virology, 83: 12378–87. for an RNA Virus’, Biology Letters, 9: 20120396. Langmead, B., and Salzberg, S. L. (2012) ‘Fast Gapped-Read Alignment with Bowtie 2’, Nature Methods, 9: 357–9. 10 | Virus Evolution, 2021, Vol. 7, No. 1 Lankau, R. A. (2009) ‘Genetic Variation Promotes Long-Term de Roode, J. C. et al. (2005) ‘Virulence and Competitive Ability in Coexistence of Brassica Nigra and Its Competitors’, The American Genetically Diverse Malaria Infections’, Proceedings of the National Naturalist, 174: E40–53. Academy of Sciences USA, 102: 7624–8. Lefeuvre, P. et al. (2019) ‘Evolution and Ecology of Plant Viruses’, Salda~na, J., Elena, S. F., and Solé, R. V. (2003) ‘Coinfection and Nature Reviews Microbiology, 17: 632–44. Superinfection in RNA Virus Populations: A Selection-Mutation Legg, J. P. et al. (2015) ‘Cassava Virus Diseases: Biology, Epidemiology, Model’, Mathematical Biosciences, 183: 135–60. and Management’, Advances in Virus Research, 91: 85–142. Sallinen, S. et al. (2020) ‘Interspecific Host Variation Plays a Key Role in Li, H., 1000 Genome Project Data Processing Subgroup. et al. (2009) Virus Community Assembly’, Nature Communications, 11: 5610. ‘The Sequence Alignment/Map Format and SAMtools’, Sardanyés, J., Elena, S. F., and Solé, R. V. (2008) ‘ ’Simple Quasispecies Bioinformatics, 25: 2078–9. Models for the Survival-of-the-Flattest Effect: The Role of Space’, Martı́n, S., and Elena, S. F. (2009) ‘Application of Game Theory to the Journal of Theoretical Biology, 250: 560–8. Interaction between Plant Viruses during Mixed Infections’, Journal of Sempere, R. N. et al. (2011) ‘Development of Expression Vectors Based General Virology, 90: 2815–20. on Pepino Mosaic Virus’, Plant Methods, 7: 6. McFadden, D. (1974) ‘Conditional Logit Analysis of Qualitative Choice Shamshiri, R. R. et al. (2018) ‘Review of Optimum Temperature, Behavior’, Frontiers in Econometrics, New York: Academic Press, Humidity, and Vapour Pressure Deficit for Microclimate Evaluation 105–142. and Control in Greenhouse Cultivation of Tomato: A Review’, Minicka, J. et al. (2017) ‘Strain-Dependent Mutational Effects for Pepino International Agrophysics, 32: 287–302. Mosaic Virus in a Natural Host’, BMC Evolutionary Biology, 17: 67. Susi, H. et al. (2015) ‘Co-Infection Alters Population Dynamics of Moreno, A. B., and López-Moya, J. J. (2020) ‘When Viruses Play Team Infectious Disease’, Nature Communications, 6: 5975. Sports: Mixed Infections in Plants’, Phytopathology, 110: 29–48. Syller, J. (2012) ‘Facilitative and Antagonistic Interactions between Plant Moriones, E., and Navas-Castillo, J. (2000) ‘Tomato Yellow Leaf Curl Viruses in Mixed Infections’, Molecular Plant Pathology, 13: 204–16. Virus, an Emerging Virus Complex Causing Epidemics Worldwide’, , and Grupa, A. (2016) ‘Antagonistic within-Host Interactions be- Virus Research, 71: 123–34. tween Plant Viruses: Molecular Basis and Impact on Viral and Host Nelson, C. W., Moncla, L. H., and Hughes, A. L. (2015) ‘ ’SNPGenie: Fitness’, Molecular Plant Pathology, 17: 769–82. Estimating Evolutionary Parameters to Detect Natural Selection Using Szittya, G. (2003) ‘Low Temperature Inhibits RNA Silencing-Mediated Pooled Next-Generation Sequencing Data’, Bioinformatics (Oxford, Defence by the Control of siRNA Generation’, The EMBO Journal, 22: England), 31: 3709–11. 633–40. Novella, I. S. et al. (1995) ‘Size of Genetic Bottlenecks Leading to Virus Tollenaere, C., Susi, H., and Laine, A.-L. (2016) ‘Evolutionary and Fitness Loss is Determined by Mean Initial Population Fitness’, Epidemiological Implications of Multiple Infection in Plants’, Trends Journal of Virology, 69: 2869–72. in Plant Science, 21: 80–90. ObreR palska-Ste R plowska, A. et al. (2015) ‘Effect of Temperature on the Valverde, S. et al. (2020) ‘Coexistence of Nestedness and Modularity in Pathogenesis, Accumulation of Viral and Satellite RNAs and on Plant Host–Pathogen Infection Networks’, Nature Ecology & Evolution, 4: Proteome in Peanut Stunt Virus and Satellite RNA-Infected Plants’, 568–77. Frontiers in Plant Science, 6: 903. Vurro, M., Bonciani, B., and Vannacci, G. (2010) ‘Emerging Infectious Ojosnegros, S. et al. (2010) ‘ ’Competition-Colonization Dynamics in an Diseases of Crop Plants in Developing Countries: Impact on RNA Virus’, Proceedings of the National Academy of Sciences USA, Agriculture and Socio-Economic Consequences’, Food Security, 2: 107: 2108–12. 113–32. Pybus, O. G. et al. (2007) ‘Phylogenetic Evidence for Deleterious Wang, Y. et al. (2009) ‘Analysis of Temperature Modulation of Plant Mutation Load in RNA Viruses and Its Contribution to Viral Defense against Biotrophic Microbes’, Molecular Plant-Microbe Evolution’, Molecular Biology and Evolution, 24: 845–52. Interactions, 22: 498–506. Read, A. F., and Taylor, L. H. (2001) ‘The Ecology of Genetically Went, F. W. (1953) ‘The Effect of Temperature on Plant Growth’, Diverse Infections’, Science, 292: 1099–102. Annual Review of Plant Physiology, 4: 347–62. Rodrı́guez-Nevado, C. et al. (2018) ‘The Impact of Host Genetic Wilm, A. et al. (2012) ‘LoFreq: A Sequence-Quality Aware, Diversity on Virus Evolution and Emergence’, Ecology Letters, 21: Ultra-Sensitive Variant Caller for Uncovering Cell-Population 253–63. Heterogeneity from High-Throughput Sequencing Datasets’, Nucleic Rojas, M. R., and Gilbertson, R. L. (2008) ‘Emerging Plant Viruses: A Acids Research, 40: 11189–201. Diversity of Mechanisms and Opportunities’, Plant Virus Evolution, 27–51.