Evolutionary Determinants of Host and Vector Manipulation by Plant Viruses

Abstract

Plant viruses possess adaptations for facilitating acquisition, retention, and inoculation by vectors. Until recently, it was hypothesized that these adaptations are limited to virus proteins that enable virions to bind to vector mouthparts or invade their internal tissues. However, increasing evidence suggests that viruses can also manipulate host plant phenotypes and vector behaviors in ways that enhance their own transmission. Manipulation of vector–host interactions occurs through virus effects on host cues that mediate vector orientation, feeding, and dispersal behaviors, and thereby, the probability of virus transmission. Effects on host phenotypes vary by pathosystem but show a remarkable degree of convergence among unrelated viruses whose transmission is favored by the same vector behaviors. Convergence based on transmission mechanism, rather than phylogeny, supports the hypothesis that virus effects are adaptive and not just by-products of infection. Based on this, it has been proposed that viruses manipulate hosts through multifunctional proteins that facilitate exploitation of host resources and elicitation of specific changes in host phenotypes. But this proposition is rarely discussed in the context of the numerous constraints on virus evolution imposed by molecular and environmental factors, which figure prominently in research on virus–host interactions not dealing with host manipulation. To explore the implications of this oversight, we synthesized available literature to identify patterns in virus effects among pathogens with shared transmission mechanisms and discussed the results of this synthesis in the context of molecular and environmental constraints on virus evolution, limitations of existing studies, and prospects for future research.

Introduction

Vector-borne plant viruses are obligate intracellular parasites that can fundamentally change the physiology of their host plants. The outcomes of these changes, such as reductions in crop yield or quality, have driven much of the research on virus–host interactions. But virus effects on host plants extend well beyond agronomically relevant metrics. There is now increasing evidence that viruses can alter aspects of the host plant phenotype (cues) that mediate interactions with other organisms, including the mobile insect vectors responsible for much of virus transmission (Casteel and Falk, 2016; Eigenbrode and Bosque-Perez, 2016; Mauck, 2016; Mauck et al., 2012, Mauck et al., 2016). These cues include visual and tactile characteristics, odors, induced defenses, secondary metabolites, sugars, free amino acids, and likely other undescribed factors (Bosque-Pérez and Eigenbrode, 2011; Casteel et al., 2014; Mauck et al., 2014a, Mauck et al., 2014b). Insect vectors make their initial foraging decisions by integrating visual and odor cues, which convey information about plant presence, identity, and quality. After contacting a host plant, vectors assess additional cues from leaf or stem surfaces, parenchyma, and vascular tissues through olfactory and gustatory sensory systems. The insects’ probing, feeding, and dispersal behaviors in response to plant cues directly determine the probability that virions will be acquired, retained, and transported (Fereres, 2016; Fereres and Collar, 2001; Hogenhout et al., 2008; Madden et al., 2000; Moreno et al., 2012; Ng and Falk, 2006). Thus, if a plant virus alters aspects of the host plant that provide cues for its herbivorous vectors, these changes have potential to influence rates of host–vector contact and vector-feeding behaviors that determine virus transmission. Given that transmission is critical to the fitness of vector-borne plant viruses, it has been proposed that viruses evolve traits that induce (or at least maintain) host phenotypes and effects on vectors that encourage virus spread.

Consistent with this hypothesis, there are now more than 100 published reports of plant viruses purportedly “manipulating” host plant phenotypes to increase vector attraction to infected plants, or elicit transmission-conducive feeding behaviors (reviewed in Casteel and Falk, 2016; Eigenbrode and Bosque-Perez, 2016; Fereres and Moreno, 2009; Heil, 2016; Mauck, 2016; Mauck et al., 2012, Mauck et al., 2016). More recently, there is evidence that some plant viruses can also manipulate vector behaviors to favor virus transmission to new hosts by interacting with the vector's tissues following acquisition from infected hosts or artificial substrates (Ingwell et al., 2012; Moreno-Delafuente et al., 2013; Rajabaskar et al., 2014; Stafford et al., 2011). The idea that plant viruses can manipulate hosts and vectors to enhance transmission is not unique. Manipulation of host phenotypes by parasites is well documented across a wide range of taxonomic groups (Heil, 2016). The fitness advantages of manipulation as a strategy are captured by the “adaptive host manipulation hypothesis,” which proposes that parasites can evolve to control elements of their host's phenotype that help maintain or enhance rates of transmission (Poulin, 2010). Thus, “manipulated” hosts exhibit additional hallmarks of infection beyond those associated with the basic need for a parasite to attenuate host immunity and use host resources for reproduction (Lefèvre et al., 2009). Although there are hundreds of reports of putative host manipulation by parasites (Lafferty and Shaw, 2013; Lefèvre et al., 2009; Mauck et al., 2012, Mauck et al., 2016; Poulin, 2010), only a handful of these studies have made progress in pinpointing the parasite as the “manipulator”—that is, the organism having genetic control over the altered host phenotype. An equally likely explanation is that the phenotype of the infected host is due to an immune response under genetic control of the host. Alternatively, observed phenotypes could represent by-products of pathology, or even the residual influence of inherited ancestral traits that were adaptive in one host–parasite context but have become maladaptive in another context (Heil, 2016). Parsing these explanations has proven difficult for eukaryotic parasites that provide the most charismatic examples of host manipulation, but prove to be intractable laboratory models. The growing evidence of host and vector manipulation by plant viruses provides new opportunities to explore the adaptive significance of parasite manipulation in the context of environmental variation using pathosystems that are more amenable to experimental methods involving functional genomics.

Although we do not yet have a thorough mechanistic understanding of plant virus genes that confer manipulative traits, there is still evidence to support the hypothesis that host and vector manipulation is adaptive for plant viruses. Theoretical studies demonstrate that “manipulative” viruses inducing transmission-enhancing effects in hosts (or vectors) will spread more rapidly, and from lower starting frequencies, relative to viruses that have neutral effects, or viruses that elicit changes that deter virus acquisition by vectors (Jeger et al., 2004; McElhany et al., 1995; Roosien et al., 2013; Shaw et al., 2017; Sisterson, 2008). Thus, under ideal conditions, manipulative virus genotypes are expected to enjoy higher fitness than nonmanipulative virus genotypes. Additionally, the nature of virus influence on vectors generally corresponds with the virus transmission mechanism regardless of virus phylogeny (Mauck et al., 2012). In other words, virus-induced changes in host cues and quality for vectors are not uniform, but differ depending on the requirements for virion uptake and transmission that are inherent to a given virus (Mauck et al., 2012, Mauck et al., 2016). Each insect-borne plant virus is classified into one of the four transmission mechanism groups depending on requirements for acquisition, retention, and inoculation (Table 1). These requirements are based on virus localization within hosts and the nature of associations with vectors, which range from transient binding to cuticular surfaces of vector mouthparts (noncirculative, nonpersistent viruses), to invasion of the hemocoel (circulation) and retention (persistence) of ingested virions in salivary glands (circulative-persistent viruses). Within the circulative-persistent category, some viruses undergo active replication in salivary glands and other internal vector tissues (propagative), while others localize to salivary glands but do not replicate (nonpropagative) (Table 1). These associations determine what vector behavioral sequences most favor efficient virus acquisition, retention, and eventual inoculation. Thus, transmission mechanism groups can serve as a basis for generating predictions about how any given virus might be expected to alter host phenotypes (or vector physiology) to influence probing, feeding, and dispersal behaviors in ways that are conducive to its own transmission. The adaptive significance of virus effects on hosts and vectors can then be explored by evaluating evidence for convergence in virus effects across phylogenetically diverse viruses that share a transmission mechanism group.

We took this approach for our earlier quantitative synthesis of 55 papers reporting putative instances of virus manipulation (Mauck et al., 2012), which was the first to show that phylogenetically divergent plant viruses transmitted via the same sequences of vector behavior induce similar phenotypes in their host plants. By demonstrating convergent effects based on transmission mechanism group, this synthesis provided support for the hypothesis that virus effects are the result of adaptations and not just by-products of infection. Since the publication of this synthesis, the number of empirical reports of putative plant virus manipulation has more than doubled. Despite the popularity of this topic and its clear relevance for understanding virus epidemiology, there has not been any subsequent attempt to comprehensively reevaluate virus effects with regard to transmission mechanism groups, or to place this body of work within the context of constraints on virus evolution imposed by molecular and environmental factors, which should not be lightly dismissed. Plant viruses have small genomes that often encode less than 10 functional proteins, sometimes through overlapping open reading frames. Virus proteins perform multiple functions in the host plant, interact with each other extensively, and may play a dual role in facilitating interactions with both plant and vector tissues. These features enable rapid replication and maintain vector transmissibility but impose major limitations on virus evolution because most mutations are likely to be deleterious and will be rapidly purged. Molecular constraints will further interact with environmental factors to shape virus evolution. In a field context, plant viruses are subject to heterogeneous host environments at intraspecific and interspecific levels, as well as variation in the frequency of transmission-conducive contacts with vectors (Elena et al., 2014; Gutiérrez et al., 2013; Pagán et al., 2012; Rodelo-Urrego et al., 2013; Roossinck and García-Arenal, 2015). These constraints will influence the evolution and maintenance of manipulative traits in plant viruses, and thus all reports of putative manipulation of hosts and vectors must be considered within the context of these constraints.

To explore the extent to which the existing literature considers molecular and environmental axes of virus evolution, and to revisit the question of whether viruses exhibit convergence in effects within each transmission mechanism group, we performed a comprehensive review and quantitative synthesis of all studies reporting putative instances of virus manipulation of hosts and vectors following the guidelines used in Mauck et al. (2012). Here, we discuss this synthesis in the context of the methodologies employed and molecular and environmental factors that may facilitate, or hinder, the evolution of manipulative functions. Our results provide evidence of convergence in virus effects within transmission mechanism groups while revealing a number of inadequacies in the current literature that provide a roadmap for future research directions.