Lack of gene flow between Phytophthora infestans populations of two neighboring countries with the largest potato production

Abstract Gene flow is an important evolutionary force that enables adaptive responses of plant pathogens in response to changes in the environment and plant disease management strategies. In this study, we made a direct inference concerning gene flow in the Irish famine pathogen Phytophthora infestans between two of its hosts (potato and tomato) as well as between China and India. This was done by comparing sequence characteristics of the eukaryotic translation elongation factor 1 alpha (eEF‐1α) gene, generated from 245 P. infestans isolates sampled from two countries and hosts. Consistent with previous results, we found that eEF‐1α gene was highly conserved and point mutation was the only mechanism generating any sequence variation. Higher genetic variation was found in the eEF‐1α sequences in the P. infestans populations sampled from tomato compared to those sampled from potato. We also found the P. infestans population from India displayed a higher genetic variation in the eEF‐1α sequences compared to China. No gene flow was detected between the pathogen populations from the two countries, which is possibly attributed to the geographic barrier caused by Himalaya Plateau and the minimum cross‐border trade of potato and tomato products. The implications of these results for a sustainable management of late blight diseases are discussed.


| INTRODUC TI ON
Gene flow, referred to the movement of gametes, genotypes, or extranuclear segments of DNA such as mitochondria from one population to another through migration and hybridization event (Toews, Mandic, Richards, & Irwin, 2013), plays a dual role in the evolution of organisms (Slatkin, 1987). Evolutionary theory considers isolation as one of the essential steps leading to speciation. Regular gene flow acts as a constraining force on evolution by homogenizing genetic and phenotypic variation among populations (Booth Jones et al., 2017). On the other hand, occasional gene flow accelerates evolutionary processes by spreading successful genes or genotypes to neighboring populations (Paun, Schönswetter, Winkler, Tribsch, & Intrabiodiv, 2008). Under the shifting balance theory (Wright, 1990), gene flow is an essential factor in enforcing population replacements. In nature, biology (e.g., dispersal mechanism and reproductive mode), community composition (e.g., kinship and biotypes), and landscape structure (e.g., patch sizes and shapes) are among the most important elements influencing gene flow of organisms (Ropars et al., 2018).
Knowledge of gene flow is important to understand the evolutionary history of species, as well as to predict their adaptive responses to future ecological and environmental fluctuations such as climate change (Edelaar & Bolnick, 2012;Garant, Forde, & Hendry, 2007). In the field of plant pathology, knowledge of gene flow of pathogens is critical for developing preventive and eradicative strategies to mitigate the epidemiological and evolutionary risks of infectious diseases in major crops (McDonald & Linde, 2002). The extent of gene flow is usually estimated indirectly based on the variance of allele frequencies among populations using F-statistics (O'Donald, 1972). The indirect methods infer the average number of individuals that are successfully incorporated into the breeding system of resident populations (Korman et al., 1993;Panyamang, Duangphakdee, & Rattanawannee, 2018) and emphasize the long-term effects of gene flow on semi-isolated populations (Singh & Singh, 2008). However, these estimates are constrained by a large number of assumptions that are unlikely to be met in practice (Meirmans & Hedrick, 2011).
Alternatively, gene flow can be inferred directly by detecting identical genotypes using DNA-based molecular marker technologies.
In practice, many molecular marker technologies can rapidly differentiate almost all genetically distinct genotypes in a population, even if a small number of marker loci is used to assay a large number of individuals (Cornuet, Piry, Luikart, Estoup, & Solignac, 1999).
Furthermore, direct measurement of gene flow by genotype identification can provide important insight into the role of gene flow on the evolution of organisms over contemporary time scales (Rannala & Mountain, 1997).
However, the accuracy of direct estimations of gene flow by the detection of identical genotypes with molecular technologies is affected by marker resolution and mutation rate. Low-resolution markers based on fragment sizes tend to overestimate the extent of gene flow because size homoplasy can misassign nonhomologous individuals in different populations to a same genotype (Caballero, Quesada, & Rolan-Alvarez, 2008). High mutation rate in molecular markers can, on the one hand, facilitate the generation of analogous sequence structures from different ancestry genotypes, leading to an overestimate of gene flow among populations. This sequence convergence has been widely documented in all species kingdoms (Balloux, Lugon-Moulin, & Hausser, 2000). On the other hand, it can also lead to an underestimation of the gene flow, caused by enhanced divergent evolution of identical genotypes in different populations. DNA sequencing technology provides the highest resolution of genotyping species (Lexer et al., 2016). As a consequence, detecting identical genotypes among populations by sequencing genes with a low evolutionary rate should provide more accurate direct estimates of gene flow.
As a housekeeping gene, elongation factor-1α (eEF-1α) is one of the most abundant and conserved sequences in eukaryotes (Hovemann, Richter, Walldorf, & Cziepluch, 1988). It encodes an isoform of the alpha subunit of the elongation factor-1 complex, an essential component of the protein synthesis process. During protein synthesis, the factor-1 complex forms a ternary structure with GTP and aminoacyl-tRNA and delivers appropriate amino acids to the ribosome (Moldave, 1985). In addition to protein synthesis, the eEF-1α subtype may be involved in functions such as organization of the mitotic apparatus, developmental regulation, signal transduction, aging, transformation, and immunoreactivity (Piedra-Quintero, Apodaca-Medina, Beltran-Lopez, Leon-Sicairos, & Lopez-Moreno, 2015;Riis, Rattan, Clark, & Merrick, 1990). In the life cycle of species, eEF-1α can be found in all developmental stages both in the cytoplasm and nucleus of cells (van't Klooster, 2000). Due to its universal occurrence, sufficient information, and slow rate of sequence evolution (Baldauf & Doolittle, 1997), the eEF-1α gene and its translated product are well suited for determining phylogenetic relationships among species and quantifying host-pathogen interaction (Chen & Halterman, 2011). It can also be used to directly estimate gene flow among populations by detecting identical genotypes of a species.
Phytophthora infestans (Mont.) de Bary, the cause of late blight in potato and tomato, can affect all parts of crops in the field and storage (Haas et al., 2009). Fast epidemics and rapid evolution are among the main challenges to effectively and sustainably control this disease. If uncontrolled, late blight can destroy entire crops within just a few days under favorable climatic conditions (Fry et al., 2015). Although many management strategies have been developed and deployed to control it over the last decades, P. infestans is still among the most destructive plant pathogens, causing approximately 8 billion US dollars annually of economic losses worldwide in potato production alone (Runno-Paurson et al., 2013). The pathogen is spread by rain-splash, infected plant materials, and wind-born sporangia (Fernández-Pavía, Grünwald, Díaz-Valasis, Cadena-Hinojosa, & Fry, 2004;Judelson et al., 2008), and increasing global trade in potato products facilitates long-distance spread and gene flow of the pathogen. This provides recurrent opportunities for new invasions of the pathogen (Montarry et al., 2010), enhancing its capacity of adaptation to changing environments and aggravating the difficulty to control it (Zhan, Thrall, & Burdon, 2014;Zhan, Thrall, Papaix, Xie, Burdon, 2015). Gene flow in a continental scale has been documented many times in P. infestans (Goodwin, Sujkowski, Dyer, Fry, & Fry, 1995). For example, the Blue-13 lineage found first time in the Netherlands in 2004 (Cooke et al., 2012) is believed to have been rapidly spread to many others countries including China and India (Chowdappa et al., 2014). However, the knowledge of gene flow in P. infestans is primarily derived indirectly from population analysis of the pathogen using fragment technologies such as isozyme and RFLP. In order to verify these indirect inferences derived, direct detection of gene flow by identifying genotypes shared among populations by DNA sequencing of conserved genes is important.
In this study, we compared sequence characteristics of eEF-1α gene generated from 245 P. infestans isolates originating from potato and tomato across wide geographic regions in China and India, the two largest potato production countries in the world. In 2017, the two countries produced 148 million tons of potato on around F I G U R E 1 Map showing the geographic locations (blue) of the Phytophthora infestans populations included in the current study. ArcGIS 10.0 software was used to create the map. Phytophthora infestans isolates from China and India are indicated by blue and pink 8,000,000 hectares (http://www.fao.org/), accounting for 38% of total global potato production and 41% of the global potato acreage.
Potato production in the two countries is still expanding.
The objectives of this study were to: (a) investigate population genetic structure of eEL-1α gene in the late blight pathogen P. infestans; (b) determine the types of sequence variation in eEL-1α gene; (c) infer the effect of host on the population genetic structure of P. infestans;and (d) infer gene flow of P. infestans between two countries with the largest potato production in the world and its implications for the sustainable management of the late blight disease.

| Phytophthora infestans eEF-1α sequences
A total of 245 eEF-1α sequences were included in the current analysis of population genetic structure in P. infestans. Of these, 165 sequences were generated from 156 potato isolates and nine to-  (Table S1, Nirmal Kumar, Chowdappa, & Krishna, 2016). The isolates from China were pregenotyped with molecular amplification of eight SSR markers (Knapova & Gisi, 2002;Lees et al., 2006), restriction enzyme-PCR amplification of mitochondrial haplotypes (Flier et al., 2003), mating type (Zhu et al., 2015), and partial sequence analysis of three genes (b-tubulin, Cox1 and Avr3a) (Cardenas et al., 2011). Only isolates with a distinct genotype were selected for sequencing. In both hosts, leaves with a typical late blight symptom were collected from plants separated by at least one meter and transported to the laboratory within 24 hr for pathogen isolation. Detailed information on the pathogen isolation can be found in previous publications Zhu et al., , 2015. Briefly, infected leaves were first rinsed with running water for 60 s and then with sterilized distilled water for 30 s.
A piece of diseased tissues was cut from the margin of leaf lesions and placed abaxial side up on 2.0% water agar for 20-30 hr. A single piece of mycelium was removed aseptically from the sporulating lesions using an inoculating needle, transferred to a rye B agar plate supplemented with ampicillin (100 μg/ml) and rifampin (10 μg/ml), and maintained in the dark at 19°C for 7 days to allow a colony to develop. The isolates were purified by two sequential transfers of a single piece of mycelium hyphae tipped from the colony to a fresh rye B plate and maintained at 13°C until use.
To extract DNA, P. infestans isolates retrieved from a long-term storage were cultured on rye B agar supplemented with ampicillin (100 μg/ml) and rifampin (10 μg/ml) at 19°C in the dark for 15 days.
The lyophilized mycelia were ground to powder with a mixer mill

| Data analysis
Nucleotide sequences were visually assessed to remove potential mutations caused by PCR artifacts (Suzan et al., 2007). Amino acid haplotypes were deduced from nucleotide sequences. The multiple sequence alignment of eEF-1α gene was performed using the ClustalW algorithm embedded in MEGA 7.0.21 (Kumar, Stecher, & Tamura, 2016), and the mutation site map was generated by BioEdit Sequence Alignment Editor (Hall, 1999 Phylogenetic trees were reconstructed from unique eEF1-α nucleotide haplotypes as well as all eEF1-α sequences using the neighbor-joining (NJ) method (Saitou & Nei, 1987)

| Sequence variation in the eEF-1α gene of Phytophthora infestans
A total of 245 partial eEF-1α sequences from China and India were included in the analysis of population genetic structure in P. infestans.
Multiple sequence alignment indicates that all sequence variations were generated by point mutations (Figure 2 (Table 3). When the nucleotide sequences were considered according to individual country, higher genetic variation was found in the eEF-1α gene from India than China. A higher variation in the eEF-1α gene was also found in samples from tomato compared to samples from potato when the samples from the two countries were combined (   (Table 2).
Only three amino acid haplotypes (isoforms) were deduced from the 17 nucleotide haplotypes (Figure 2). The main amino acid haplotype (AAH1) deduced from nucleotide haplotypes H1-H9, H11, H13-15, and H17 was found in both P. infestans populations from China and India (Table 4). It was the only amino acid haplotype detected in China (100%) and also accounted for 90.12% of Indian P. infestans population. AAH2 deduced from H12 and H16 was generated by the substitution of isoleucine in the 78th amino acid of AAH1 with valine, while AAH 3, deduced from H10, was generated by the substitution of glycine in the 268th amino acid of AAH1 with cysteine. AAH2 and AAH3 were only found in the P. infestans population from India.

| Haplotype network of eEF-1α
The nucleotide haplotype network of eEF-1α gene formed two major groups and there was a clear geographic association among haplotypes ( Figure 3). The less diverse group comprised of all five nucleotide haplotypes from China diverged by a maximum of three mutation steps among haplotypes. The other group was more diverse. It included all 12 nucleotide haplotypes from India and diverged by a maximum of eight mutation steps among haplotypes.
The two groups were connected by five mutation steps. H1, the most abundant nucleotide haplotype in China, and H6, the most abundant nucleotide haplotype in India, were distanced by 10 mutation steps.
All three nucleotide haplotypes (H10, H12, and H16) coding the rare amino acid haplotypes (AAH2 and AAH3) were located in the tip of the network tree. The nucleotide haplotype network also contained two reticulating structures. One reticulation was formed by four haplotypes (H1, H2, H3, and H5) from China, and another one was formed by other four haplotypes (H6, H9, H11, and H13) from India.
Most nucleotide haplotypes were unevenly distributed between potato and tomato hosts.

| Phylogenetic analysis
Similar to network analysis, phylogenetic cluster analysis by a neighbor-joining approach also produced a dendrogram that divided the  Positions  and types of  substitution  H1  H2  H3  H4  H5  H6  H7  H8  H9  H10  H11  H12  H13  H14  H15  H16  H17 87s T  T  T  T  T  T  T  T  T   918v  T  T  T  T  T  T  G  T  T  T  T  T  T  T  T  T  T 948s Note: s = transition and v = transversion.

| D ISCUSS I ON
Overall, a low genetic variation was found in the eEF-1α gene. Only 17 nucleotide and three amino acid haplotypes were detected in the 245 sequences. The genetic variation is substantially lower than in many functional genes in P. infestans (Yang et al., 2018) and other pathogens (Marisa et al., 2018). For example, from a subset of the same Chinese collection, 51 nucleotide haplotypes were identified in the 96 Avr3a sequences (Yang et al., 2018). The eEF-1α is a housekeeping gene playing multifaceted roles in the biochemical and physiological processes of life (Chang et al., 2002;Kato, Sato, Nagayoshi, & Ikawa, 1997). Low genetic variation in eEF-1α is consistent with evolutionary hypothesis that genes important to cell functions evolve at a reduced rate (Jordan, Rogozin, Wolf, & Koonin, 2002).
Housekeeping genes routinely experience purified selection through which genetic variation is largely reduced (Viscidi & Demma, 2003). In addition, a lack of intragenic recombination may also contribute to the low genetic variation in the eEF-1α gene. Intragenic recombination can generate new haplotype variation (Watt, 1972) and has been commonly found in genes responsible for the interaction of pathogens with hosts and other environments (Stergiopoulos et al.., 2013) including effector and fungicide resistance genes of P. infestans (Chen, Zhou, Qin, Li, & Zhan, 2018;Yang et al., 2018). Although some reticulation structures exist in haplotype network (Figure 3), no signals of intragenic recombination were identified in the eEF-1α gene by any of the seven algorithms implemented in the RDP4 suite (data not shown). Thus, it is reasonable to believe that the reticulation structures were generated by convergent evolution of nucleotide sequences (Ralph & Coop, 2015), suggesting that mutations occur frequently in the eEF-1α gene and the observed low genetic variation was likely caused by other mechanisms such as purifying selection rather than low mutation rate of the gene.
Higher genetic variation was found in the eEF-1α sequences of P. infestans populations originating from India compared to populations from China. This result is consistent with previous surveys using a similar set of neutral markers. With the SSR markers, 24 multi-locus genotypes were detected among 59 P. infestans isolates sampled from India (Dey et al., 2018), while only 26 multi-locus genotypes were identified among 279 isolates sampled from China (Tian, Yin, Sun, Ma, Ma, Quan, et al., 2015;Tian, Yin, Sun, Ma, Ma, Wang, et al., 2015). China produces potato and tomato on larger acreages than India, and therefore, it is expected to host a larger P. infestans population than India.
Pathogens with a larger population size tend to have a higher genetic variation due to more alleles being generated by mutations and fewer alleles being lost by genetic drift (Lázaro-Nogal, Matesanz, García-Fernández, Traveset, & Valladares, 2017 prescreened molecularly and phenotypically and only isolates with distinct genotypes were selected for sequencing in this study. The fact that a higher genetic variation was still found in the Indian P. infestans population is likely caused by other natural factors and agricultural practices promoting the accumulation of genetic variation in the pathogen populations such as conducive environmental conditions, diversifying selection for different ecosystems, increasing international trade of plant materials, and reduced field hygiene during production (Dey et al., 2018). Indeed, it is reported that farmers in many parts of India tend to use potato tubers saved from previous years as seeds (Chowdappa et al., 2014;Dey et al., 2018)  The P. infestans populations sampled from tomato displayed a higher genetic variation than those sampled from potato (Table 2), and this pattern of genetic difference among host origins was found in both China and India (data not shown). This is consistent with a previous result from France (Wangsomboondee, Groves, Shoemaker, Cubeta, & Ristaino, 2002) in which the higher genetic variation of P. infestans from tomato than potato was thought to be resulted from the different mating systems adopted by the pathogen on the two hosts, that is, sexual reproduction on tomato versus asexual reproduction on potato (Lebreton & Andrivon, 1998). However, this cannot explain the current finding because no evidence of sexual reproduction has occurred in P. infestans populations on tomato in either China or India (Chowdappa et al., 2014;Yang et al., 2009). The difference more likely reflects the difference in selection pressure imposed by the two crops. In potato, P. infestans recurrently moves from and forth between foliages and tubers, posing a strong selection pressure on the pathogen and reducing its genetic variation. On the other hand, no such selection occurs in the tomato production system.
Phytophthora infestans is a pathogen with a great potential for international migration (Fry et al., 2015). Successful clonal lineages originating from a regional population can quickly spread globally.
For example, Blue_13 first reported in Europe in 2004 has been detected in many parts of the world (Cooke et al., 2012) including China and India (Chowdappa et al., 2014;Li et al., 2013). Interestingly, no nucleotide haplotypes of the eEF-1α gene were shared between the P. infestans populations from China and India, suggesting cross-border movement of the pathogen in our study may not be as frequent as reported previously (Fry et al., 1993).
India is bordered by the Himalaya Plateau with Tibet of China, and the majority of lands in Tibet areas are not conducive for agricultural production. Furthermore, cross-border trade of potato and tomato products between the two countries is also very limited.
China exports potato and tomato mostly to Russia, Japan, and South Korea (Wang & Zhang, 2004) and potato imports mainly from the USA and Europe (Huang, 2004). On the other hand, the biggest potato export from India goes to Nepal, Sri Lanka, Pakistan, Mauritius, and Bangladesh (Kumarasamy & Sekar, 2014) and Indian potato imports mainly from Germany and France (https ://www.infod rivei ndia.com/). The unique landscape structure coupled with reduced anthropogenic activities related to potato and tomato materials largely disconnect the P. infestans populations between the two countries, contributing to the observed population genetic differentiation. Blue-13 lineage is consisted of many genotypes. Different Blue-13 genotypes from Europe by different migration events could also generate the observed spatial pattern of haplotype distribution (Chowdappa et al., 2014;Li et al., 2013).
Our results have implications to the management of late blight in potato and tomato. Even though no evidence of direct gene flow was shown to have occurred between the P. infestans populations in the two biggest potato production countries in the study reported here, necessary precautions should be taken to prevent the indirect movement of the pathogen through third-party countries by implementing strict quarantine procedures.

CO N FLI C T O F I NTE R E S T
None declared.

Associated data have been deposited in GenBank: Accession
Numbers MN422761-MN422925.