Anastrepha obliqua as a study organism

The definition of model organism has grown along with the increase in genome sequencing research. At first, model organisms were limited to species with characteristics that facilitated laboratory experiments, such as size and short generational cycle. However, model organisms nowadays include other species based on criteria such as agricultural importance, economic impact, or effects on human health (Hedges, 2002). Matz (2018) recently published an article on the relevance of genomic approaches with obscure model organisms (OMOs). These are defined as organisms for which no previous genome resource exists, but that constitute potentially good models to study genomic evolution, acclimatization, and adaptation in real ecological contexts. Although A. obliqua is not yet considered a model organism, it could be included as an OMO to study the adaptation mechanisms of phytophagous insects to host plants in the Neotropics, due to its economic impact on fruit crops and its status as an important pest.

Anastrepha obliqua is a species with a wide distribution, extending from Mexico to Brazil, including the Antilles. It is also occasionally reported in the USA (Fu et al., 2014). It is the most important mango (Mangifera indica L.) fruit fly pest in Latin America (Aluja et al., 2014; Fu et al., 2014; Rull et al., 2018), but has also been catalogued as a highly polyphagous species, attacking 24 fruit families (Carrejo & González, 1994). The species has an important economic impact on fruit crops (CONPES, 2008; Norrbom, 2008). No less than 30–40% of losses in fruit production in Colombia is related to fruit flies; 70% if adequate pest management is not in place (CONPES, 2008). Mexico uses sterile insect technique (SIT) programs to control populations of A. obliqua in some areas of its distribution (Rull et al., 2018), and Brazil has regulatory laws to quarantine this species (Passos et al., 2018). Being a highly mobile, opportunistic, multivoltine species, A. obliqua is thought to have great dispersive capacity and invasive potential for maintaining populations in new hosts and/or distribution ranges (Tejeda et al., 2016) that can influence diversification patterns.

As an important fruit pest, A. obliqua has been studied using various approaches to understand its biology. Anastrepha obliqua is in the fraterculus species group, and it shares morphological and behavioral characteristics with closely related species, resulting in the formation of natural hybrids (Scally et al., 2016). This fact is relevant for the diversification process, management strategies, and the interspecific relationships of A. obliqua. Research includes studies of the genetic population structure in some areas of its distribution (Ruiz-Arce et al., 2012, 2019; Aguirre-Ramirez et al., 2017; Passos et al., 2018); molecular mechanisms and phenotypic plasticity (i.e., the ability of one genotype to produce different phenotypes under different environmental conditions) (Velasco-Cuervo et al., 2021; Lemos-Lucumi et al., 2022); morphometric characteristics (Castañeda et al., 2015a); reproductive behavior (Aluja et al., 2001; Leal & Zucoloto, 2008); chemical ecology (Cruz-López et al., 2006; Aluja et al., 2020; De Aquino et al., 2021); relationships with microbiota (Gallo-Franco & Toro-Perea, 2020; Amores et al., 2021; Cárdenas-Hernández et al., 2022); and phylogenetic relationships (Smith-Caldas et al., 2001; Ruiz-Arce et al., 2012). These studies have revealed a marked genetic structure (Ruiz-Arce et al., 2012; Aguirre-Ramirez et al., 2017) and morphological differentiation (Castañeda et al., 2015) between populations of A. obliqua throughout its distribution range, and the importance of host plants in its genetic makeup and intra- and interspecific relationships.

The importance of A. obliqua interactions with various host plants has been documented, demonstrating host effects on various characteristics of the insect's life history (Table 1). For example, the composition of the aroma secreted by sexually mature adult males is influenced by larval development in specific fruit species (Aluja et al., 2020). These changes affect the interaction between males and females and could lead to mating preferences and host fidelity, bearing consequences on speciation. Aguirre-Ramirez et al. (2017) investigated the population structure and genetic diversity of A. obliqua in the southwest of Colombia and found different haplotype frequencies among flies belonging to different host plants. Two mitochondrial DNA markers, the subunit I of the cytochrome c oxidase (COI) gene, and the NADH dehydrogenase subunit 6 (ND6) gene, revealed a total of seven haplotypes, some of which exclusive to particular localities or host plants. Ruiz-Arce et al. (2019) found a lack of genetic structure associated with the host plant in populations of A. obliqua in Veracruz (Mexico); however, some populations did show genetic differentiation, which could be attributed to some degree to host fidelity. More recently, studies carried out on A. obliqua flies infesting red mombin (Spondias purpurea L.), mango, and carambola (Averrhoa carambola L.) have shown differences in the expression of genes associated with digestion and detoxification (Velasco-Cuervo et al., 2021) and in microRNAs (miRNAs) that have as target mRNA genes in these same categories (Lemos-Lucumi et al., 2022). In both studies, the main differences were found in flies infesting carambola, which could be related to the chemical composition of these fruits and their recent introduction to the American continent (Velasco-Cuervo et al., 2021). Aguirre-Ramirez et al. (2021) found similar results at the genomic level. Populations of A. obliqua infesting carambola have a greater genomic difference [based on restriction-site associated DNA sequencing (RAD-seq; Box 2] compared to populations infesting red mombin and mango. Additionally, 54 single nucleotide polymorphisms (SNPs) were found to be associated with host plant use, some of which in protein families that could be annotated in pathways related to nutrition (Aguirre-Ramirez et al., 2021).

The polyphagous insect status of A. obliqua, its importance as a species of economic impact in the Neotropics, and the evidence of the effect that host plants have on different aspects of its life history, make this species a good case study to understand the molecular mechanisms involved in the survival and/or diversification of Neotropical phytophagous insects associated with their host in sympatric conditions (i.e., when two populations or species have overlapping geographic distribution), and on a small geographic scale.

Adaptive molecular mechanisms

Host change in phytophagous insects can result in two scenarios, in which the change promotes (1) generalism by increasing its distribution range, or (2) differentiation in each or some of the hosts, and helps in specialization. This second scenario can even result in the generation of host races (i.e., populations that are partially reproductively isolated as a consequence of local adaptation to their hosts) or new species (Heard, 2012). Two non-mutually exclusive processes form the base for these scenarios: phenotypic plasticity and local adaptation through ecological specialization – the latter occurs when a population adapts to the specific conditions of its native environment and demonstrates higher fitness than individuals of the same species coming from other populations (Orsucci et al., 2018). A study of the molecular mechanisms of these two processes by which these insects adapt to different host plants should be undertaken.

Host plant-driven selection is generally associated with a plastic response. Some studies have associated differential gene expression (DGE) to herbivory in insects; DGE includes host-specific detoxification and digestion genes (De La Paz Celorio-Mancera et al., 2013; Ragland et al., 2015; Christodoulides, 2017; Etges, 2019; Pym et al., 2019; Velasco-Cuervo et al., 2021; Lemos-Lucumi et al., 2022). Comparative genomics has also helped in identifying differences in genomic architecture between poly- and monophagous insects. For example, some authors have found differences in the number of gene duplications for genes associated to chemical perception and detoxification (involved in resource use) (Christodoulides, 2017; Gouin et al., 2017). Other genomic patterns such as differences in allelic frequencies are usually found between species (e.g., poly- vs. monophagous) and also among populations of a single species that infests different hosts (Oroño et al., 2013). The differences that we can trace at the genome or transcriptome level will depend on factors such as gene flow (i.e., allelic movement) among populations, and the time at which the population continuum is sampled.

In a model such as A. obliqua, where populations infest different host plants in sympatry, it is assumed that gene flow will continue to occur among populations (Figure 1). Initially, differentiation is expected to occur on one locus or a few loci because the populations could have strong genetic similarity at this stage (Campbell et al., 2018). However, differentiation in these few loci could cause some degree of reproductive isolation. It is still strongly debated whether differentiation in this ecological context occurs due to many loci with a small effect or a few loci with a large effect (Feder et al., 2012). Even so, it has been predicted that differentiation could begin in loci directly associated with ecological differentiation, having an effect on particular traits of the phenotype. These characteristics can be subject to divergent selection – i.e., selection between two populations acts in opposite directions – and are generally related to a better use of the ecological resource or to non-random mating (Servedio et al., 2011).

Although selection can occur on loci undergoing fixation in the population, the process can also alter allelic frequencies of loci in neighboring or connected regions by what is known as hitchhiking (Smith & Haigh, 2008). These differentiated loci become a barrier to gene flow (barrier loci), causing local adaptation and ecological specialization in the presence of gene flow.

In a species such as A. obliqua, we would expect differentiation among populations that infest different host plants in sympatry. This would be reflected in F_ST (fixation index, a measure of population differentiation developed from Wright's F-statistics) peaks through the genome in loci that are related to characteristics under divergent selection associated with host use [which could be identifiable using methodologies such as quantitative trait locus (QTL) analysis and genome-wide association studies (GWAS); Box 2] (Caillaud & Via, 2012; Ragland et al., 2017; Doellman et al., 2018). In a more advanced phase, we would expect more homogeneous differentiation throughout the genome, and that not only the loci linked to regions under divergent selection increase their allelic frequency, but that poorly adapted genes are also being eliminated (Feder et al., 2012).

This differentiation continuum corresponds to a four-phase speciation model with gene flow. In phases 1 and 2, strong selection over the selected loci is expected (a major contribution of divergent hitchhiking, which occurs when allelic frequencies surrounding barrier loci change as result of physical linkage and a selective sweep), whereas in phases 3 and 4, a barrier to gene flow is expected throughout the entire genome (a major contribution of genomic hitchhiking, which occurs when allelic frequencies change at the genome level as a result of a reduction in migration rates, including regions not linked physically) (Feder et al., 2012). Although the identification of these phases in natural populations can be challenging, genome patterns from which these phases can be inferred have been proposed. For example, Turner et al. (2005) had previously suggested genome differentiation patterns using a geographical metaphor corresponding to divergent genomic islands, and Feder et al. (2012) argued divergent genomic continents referring to the degree of differentiation that we can find through the genome. Nevertheless, these patterns in the genomic landscape can also originate from other evolutionary forces in a variety of geographic contexts (Wolf & Ellegren, 2017), so other methodologies and experiments must be integrated in order to correlate genetic variation with a phenotypic trait.

To verify molecular fingerprints of generalism or specialism, various tools, allowing exploration of differences at genome, transcriptome, proteome, and metabolome levels, can be used. In this way, we can observe changes reflected in allelic frequencies, F_ST peaks, mRNA levels, or miRNA gene expression, or differences in methylation patterns.

Tools for an evolutionary approach in the omics era

Although the contribution of classical genetics to the comprehension of basic cell processes and development goes without saying, next generation sequencing (NGS) changed the approach to understanding molecular mechanisms on large scale. NGS comprises a set of rapid and cost-effective platforms or technologies to generate thousands to many millions of sequencing reactions per run using DNA or RNA, to produce high-throughput sequencing data (van Dijk et al., 2018; Tiwary et al., 2021). NGS technologies have played an important role in many fields of research, including evolutionary biology, and enabled the development of new research questions by changing the way we think and carry out ecological and evolutionary studies, such as incorporating non-model organisms based on natural systems (Tautz et al., 2010).

Does genomic differentiation exist among insect populations that feed on various host plants?

When we approach phytophagous insect–host plant interactions for an organism such as A. obliqua, for which genomic information is scarce but there is some knowledge of population structure, it is necessary to study the biology of the species. Information on life history, behavior, and reproduction will help understand the possible influence of genomic differentiation in the insects at a given time during their interaction with host plants.

To track genomic differentiation and divergence patterns, techniques such as RAD-seq, WGS, pool-seq, and RNA-seq (Figure 2, Box 2) can be used. For example, DNA markers associated with restriction sites (RAD-seq) have been one of the most widely used methodologies to evaluate genomic differentiation, ecological specialization, and local adaptation (Catchen et al., 2013; Keller et al., 2013; Egan et al., 2015; Guo et al., 2016a; Marques et al., 2016; Picq et al., 2016; Koubínová et al., 2017). This methodology produces thousands of SNPs throughout the entire genome. RAD-seq has been used to characterize genomic variation in populations with gene flow and to identify genomic markers of differentiation or selection associated with different environmental parameters in non-model organisms within Lepidoptera (Nadeau et al., 2014) and fruit flies (Egan et al., 2015; Aguirre-Ramirez et al., 2021) (for more information on the family of RAD-seq methodologies, see Andrews et al., 2016). On the other hand, WGS has two widely used approaches: de novo sequencing and re-sequencing (WGR). A great advantage of WGS over other methodologies is that it will provide a complete genome assembly for the study subject, thereby providing a genomic resource that can be used for comparison with other individuals. WGS allows comparison of both genome content and architecture. Gouin et al. (2017) used WGS to obtain the complete genome of a highly polyphagous Lepidoptera species, Spodoptera frugiperda J.E. Smith (fall armyworm), and found an expansion (through gene duplication) of genes related to chemoreception, detoxification, and digestion, when compared to specialist Lepidoptera species. Findings suggested gene duplication is a possible adaptive mechanism for polyphagy (Gouin et al., 2017).

An alternative to maximizing resource use is to pool samples (pool-seq). Pool-seq provides a better cost–benefit strategy compared to approaches in which samples are analyzed or sequenced individually (Schlötterer et al., 2014), especially when the research question is population-based and aimed to determine genomic variation patterns. Pool-seq has been used in variant filtering (Anand et al., 2016) and for association mapping of non-model organisms (Micheletti & Narum, 2018). The use of this methodology implies the loss of individual genotypes, which must be taken into account before conducting the study (for more information on pool-seq see Schlötterer et al., 2014). When pooling samples, it is important to keep the original individual DNA samples so that variations can be demultiplexed in case interesting variation is detected. Gautier et al. (2013) directly estimated differences in allelic frequencies of genotyped individual vs. pooled samples and found that pool-seq is a precise and suitable strategy for comparing genomic patterns of genetic variation.

The need for multidisciplinary approaches

The answers to research questions regarding the study of phytophagous insect–host plant interactions ought to be found through an interdisciplinary approach. Omics technologies provide a large amount of information to produce robust conclusions on genomic differentiation, and the relationship between the phenotype and regions of the genome under selection. These technologies should be used together with other methodologies such as selection experiments, reciprocal transplants, fitness measurements, and experimental evolutionary studies that can help understand phytophagous insect–host plant interactions holistically. For example, selection experiments contribute to elucidating adaptation and the genetic basis of groups with adaptive traits (Fuller et al., 2005). When combined with other genomic tools, selection experiments have also helped determine which specific regions, genes, and/or SNPs seem to be affected by selection (Doellman et al., 2018).

Reciprocal transplant experiments and fitness measurements can also provide strong evidence for divergent selection. These experiments consist of the evaluation of organism fitness under various controlled environments. Together with omics approaches, they can provide evidence of regions under which phenotypic traits are associated with phytophagous insect adaptation to the host plant. We propose a methodology based on reciprocal transplants, fitness measures, and omics technologies to understand the molecular mechanisms in insects associated with host use, for example A. obliqua (Figure 3). In the first stage, a reciprocal transplant experiment is carried out with two A. obliqua populations colonizing mango and carambola. First instars of either population can be collected and transplanted to the opposite fruit host (Figure 3A). This type of experiment allows various analyses, such as fitness measurements in terms of number of larvae that survive in reciprocal transplants vs. controlled conditions (Figure 3A). The experiments also allow understanding direct local adaptation in order to compare native to non-native population fitness (Savolainen et al., 2013). The expected result, with local adaptation, is that the population performance of A. obliqua coming from mango would be greater in mango than in carambola. This could be tested using reaction norms (Figure 3B).

In a second stage, differential gene expression can be measured or selected SNPs can be detected (e.g., at population level) after the populations have been submitted to the contrasting environments. This step requires the implementation of omics tools and can provide information, for example, on gene expression for a population or a species in the native vs. the new environment. In this manner, (1) the extent of the plastic response, (2) the genes directly involved in survival within the host, and (3) whether selection has occurred, can be determined. Reduction of the original plastic response can be determined (e.g., in one of the populations), thereby contributing to knowledge of the genetic bases for ecological specialization. At a final stage, methodologies such as GWAS and/or CRISPR/Cas9 can be included. For example, once divergent gene candidates or divergent genomic regions are identified by host, CRISPR/Cas9 could be used to silence these genes and validate their function and effect on the phenotype. This workflow would answer holistic questions from an evolutionary approach, thus uncovering the mechanisms by which insects are able to use and adapt to their host plants.

Reciprocal transplant experiments are popular in phytophagous insect–host plant studies. For example, preference and performance for native vs. alternative hosts have been evaluated in economically impacting borer moths, Ostrinia nubilalis Hübner and Ostrinia scapulalis Walker (Orsucci et al., 2016). In these experiments, preferences for native hosts were found and the importance of these traits as barriers to reproductive isolation in divergence processes were discussed (Orsucci et al., 2016). Other studies have shown the effect of diet on the colonization of new hosts. For example, in cotton aphid, Aphis gossypii Glover, garden experiments and reciprocal transplants demonstrated that host range expansion can be measured by individual feeding experience and, in some cases, by the presence of specific genotypes (genetic background) (Ma et al., 2019). In the family Tephritidae (Ragland et al., 2015), larval performance and gene expression in Rhagoletis zephyria Snow and R. pomonella were evaluated by alternating host fruits. The results showed high plasticity differential gene expression and a greater performance in species survival in its native host.

Multidisciplinary approaches, together with omics approaches, provide an invaluable platform for corroborating phenotype association, local adaptation, or manipulation of model system evolution under controlled conditions (De Villemereuil et al., 2016). A multidisciplinary approach contributes to knowledge of the genetic basis and molecular mechanisms in phytophagous insect–host plant interactions.