Debugging: Strategies and Considerations for Efficient RNAi-Mediated Control of the Whitefly Bemisia tabaci

Simple Summary The whitefly Bemisia tabaci is a crop pest insect that is difficult to control through commercially available methods. Technology that inhibits gene expression is a promising avenue for controlling whiteflies and other pests. While there are resources available to make this method insect specific, and therefore more effective, it is currently being used in a way that targets insects broadly. This broad approach can cause potential harm to the surrounding environment. Here, we discuss considerations for using gene-silencing technology as a pest management strategy for whiteflies in a way that is specific to this pest, which will address short- and long-term issues of sustainability. We also provide a way of selecting target genes based on their roles in the life history of the insect, which will reduce the potential for unintended negative consequences. Abstract The whitefly Bemisia tabaci is a globally important pest that is difficult to control through insecticides, transgenic crops, and natural enemies. Post-transcriptional gene silencing through RNA interference (RNAi) has shown potential as a pest management strategy against B. tabaci. While genomic data and other resources are available to create highly effective customizable pest management strategies with RNAi, current applications do not capitalize on species-specific biology. This lack of specificity has the potential to have substantial ecological impacts. Here, we discuss both short- and long-term considerations for sustainable RNAi pest management strategies for B. tabaci, focusing on the need for species specificity incorporating both life history and population genetic considerations. We provide a conceptual framework for selecting sublethal target genes based on their involvement in physiological pathways, which has the greatest potential to ameliorate unintended negative consequences. We suggest that these considerations allow an integrated pest management approach, with fewer negative ecological impacts and reduced likelihood of the evolution of resistant populations.


Introduction
The whitefly Bemisia tabaci (Gennadius) (Hemiptera: Aleyrodidae) is a globally important cryptic species complex that causes billions of U.S. dollars in damages to many crops including vegetable and row crops and therefore presents a major threat to food security [1][2][3]. Over 1000 plant species serve as hosts to B. tabaci [4], and crops that are damaged range from common fruits and vegetables such as cassava (Manihot esculenta), squash (Cucurbita spp.), and tomato (Solanum lycopersicum) to with specific RNAi delivery methods applied to B. tabaci. However, conceptual frameworks that explore the scope of RNAi's functionality in pure biological terms, considering the appropriate target pathways and desired ecological outcomes, are lacking. Advancing these discussions demands closer attention to knowledge of species specificity, off-target effects, and the possibility that the target species will evolve resistance is lacking (discussed in [42]). Moreover, the central issue in the use of RNAi as a control strategy, the ideal target, is rarely discussed in terms of lethality without further evaluation of long-term impact, ecological interactions, or sustainability of this approach. This is especially important for complex recalcitrant pests like B. tabaci. Here, using B. tabaci as a model, we specifically highlight these considerations for the use of RNAi as a part of an integrated pest management strategy. Our goal is to consider sustainability of long-term use based on the unique features of B. tabaci and present a conceptual framework for selecting target genes based on their involvement in specific functional pathways.

RNAi as a Pest Management Strategy: Need for Greater Specificity
RNAi has been used for over 20 years to evaluate gene function in insects [43][44][45][46], but only recently has its utility in more commercial applications, including pest management, been considered [47][48][49][50][51]. Early studies on coleopteran and lepidopteran pests first demonstrated the potential of RNAi to induce lethality through the knockdown of critical functional genes [36,37]. Since this time, advancements in RNAi technology coupled with the increasing availability of genomic data for non-model insects have expanded the capabilities of researchers to develop more species-specific control mechanisms. Despite this, RNAi for pest management continues to rely on broad stroke approaches and molecular considerations with little attention given to species life history and physiology. Moreover, successful implementation of this new technology remains hindered by familiar pest management issues, including cost efficiency and ecological impacts, that ultimately limit the success of RNAi as a long-term solution. Whiteflies are remarkably well studied among insect pests, and many aspects of successful RNAi implementation in B. tabaci have been worked out [29][30][31][52][53][54][55][56][57][58]. According to overarching consensus, the formula for effective RNAi pest management of B. tabaci is to (1) select a target gene that is crucial for survival, (2) use a plant-mediated delivery method, and (3) prevent dsRNA degradation. While this workflow is useful for current implementations of RNAi, it only partially addresses the question, "What is successful application of RNAi in pest control?" We contend that successful RNAi-mediated control strategies should offer an effective and sustainable alternative or supplement to conventional methods. If true, the present considerations must be expanded to include strategies that minimize ecological impacts and maximize long-term efficacy. This sentiment echoes increasing calls for more holistic approaches to RNAi-mediated pest management [58]. For the remainder of this review, we focus on three key considerations that have been underrepresented in studies of RNAi in pest management: ecological impact, evolution of resistance, and choice of functional pathway. Each of these considerations relies fundamentally on an understanding of gene function, which we contend must be prioritized for the development of future, species-specific RNAi control strategies.

Desired Outcomes: Lethal versus Sublethal Effects
Of primary importance for designing and implementing species-specific RNAi control is the careful consideration of desired effects. When it comes to pest management, the desired effect is most often assumed to be lethality [29,[59][60][61]. However, in practice, such blunt-force approaches potentially create more problems than they solve. Below, we expand on these shortcomings and also present the case for substituting sublethal forms of RNAi control as practically effective, long-term sustainable alternatives to current approaches. In contrast to conventional approaches that manage lethality through dosage administration, we use the term "sublethal" to refer specifically to the magnitude of outcome predicted when knocking down particular gene pathways. In other words, gene targets are selected based on their additive contributions to total fitness as opposed to their necessity for survival.

Short-Term Considerations: Off-Target Effects and Ecological Impact
The first major issue with RNAi applications that result in lethal outcomes lies in the ecological impact of totally eradicating B. tabaci from agroecosystems. One possible outcome of this is that a secondary pest species will increase in population density and be more harmful to the cropping system than B. tabaci [62]. A second potentially harmful effect is the disruption of complex community interactions in which whiteflies are involved outside of agricultural ecosystems [63,64]. Finally, managing pests through the use of lethal RNAi strategies means that any off-target effects are likely to be especially detrimental [36,65,66]. One of the ways that RNAi can become a biological hazard is that the dsRNA designed to silence specific target genes aligns, imperfectly or perfectly, to gene regions in another species that share high sequence similarity [67,68]. Such effects become increasingly likely when the targets of RNAi are housekeeping genes with critical functions for survival, as these tend to be highly conserved across arthropods [69,70]. For example, one study observed that RNAi-induced silencing of core metabolic genes (v-ATPases) in a pest beetle species had significant, unintended effects on survival of non-target beetles [36]. Deployed in a mixed agroecosystem, such off-target effects could have devastating consequences.
An alternative to lethal RNAi strategies is to develop a framework that more closely aligns with the goals of integrated pest management. One approach currently being developed for RNAi is based on the use of sublethal methods to control pest populations [71][72][73][74]. Together with integrated approaches that target pests only within controlled, locally affected areas, sublethal RNAi approaches offer a promising solution to the issue of unintended ecological consequences arising from the total eradication of whiteflies. A second major advantage of this approach is that it reduces the severity of unintended effects on non-target species [71,75]. This is true because genes that are less crucial for basic cellular functions are also less likely to be strongly conserved. In addition, in the improbable case of perfect sequence alignment, off target effects are less likely to cause mortality. Further, by using sublethal strategies there is the potential to combine RNAi with more traditional and ecologically-based approaches, including the use of insecticides and natural enemies, which may enhance efficacy at reduced cost [71]. The use of sublethal strategies could make it possible to utilize parental RNAi for long-term control and could reduce the number of treatment applications. However, this strategy may not be feasible for all gene targets, specifically those that affect reproductive fitness.

Long-Term Considerations: Evolution of Resistance
Though the development of resistance is not considered an immediate issue for RNAi-mediated pest control, it remains an important consideration for long-term sustainability [42]. The main way that insects become resistant to RNAi appears to be through the acquisition of more efficient dsRNA degradation capabilities [76]. However, field insect strains that are resistant to RNAi due to an inability to uptake dsRNA in the gut cells have been reported [77]. Such concerns are pronounced when it comes to managing pests with fast generation times.
Whiteflies exhibit very rapid population growth cycles, which are augmented by their ability to flexibly switch between sexual and asexual modes of reproduction [5]. Under optimal conditions, B. tabaci females reproduce sexually, resulting in diploid daughters [78] (Figure 1). Because most genetic mutations, including those that confer insecticide resistance, are recessive [79], female offspring must inherit two copies of this new variant to become functionally resistant to RNAi. However, B. tabaci is haplodiploid, meaning that females also have the ability to reproduce asexually when suboptimal conditions make it difficult to locate mates [5] (Figure 1). This results in the production of haploid male offspring, which inherit only a single copy of a new variant and therefore always express this new resistance phenotype. Due to their hemizygous condition, males inheriting mutant (resistant) copies are exposed to selection regardless of dominance or recessiveness [79]. The result of this is that can rise to fixation rapidly in populations containing a large proportion of haploid individuals, whereas diploidy delays the evolution of resistance, particularly when beneficial mutations are recessive and their effects are "shielded" in heterozygous states [79,80] (Figure 1B). Hence, systems expressing flexible haplodiploidy are already more likely to evolve resistance to RNAi than systems that rely on strict diploidy for their reproduction [79,80]. Selecting RNAi targets with sublethal effects, as opposed to lethal effects, offers an intuitive path for slowing the evolution of resistance-by relaxing selection against susceptible individuals (Box 1). In the majority of cases where an insect's pestiferous status can be traced to its high densities, methods of biocontrol that dramatically suppress population growth, rather than inducing mass mortality, are likely to be as effective in mitigating economic damage.

Box 1. A population genetics perspective on the evolution of resistance.
Imagine a hypothetical scenario in which a random mutation creates a genetic variant for resistance (a "resistant allele"), which initially segregates at low frequencies in the population. Two factors-the magnitude of effect of the allele on an individual's phenotype and the strength of natural selection on resulting phenotypes-will determine the speed at which a resistance allele rises to fixation in a population. The efficiency of these dynamics is particularly pronounced in instances where control methods cause lethality, as this mimics a phenomenon referred to in nature as 'hard selection' [75,81] Hard selection occurs when alleles are inherited as lethal equivalents, causing immediate death of an organism independent of other selection pressures [82]. The result is strong and efficient purging of non-resistant alleles segregating in the population [79]. Consider alternatively a scenario whereby non-resistant alleles do not cause lethality, but instead reduce organismal fitness and gradually suppress population growth. This is referred to as 'soft selection' [75,81] and in our example will be applied to any continuous trait imparting sublethal, but severe, effects on fitness. By permitting variation in relative fitness among individuals that carry resistant alleles, soft selection increases the likelihood that some susceptible alleles will be passed on to subsequent generations, thereby relaxing 'hard selective' pressures that drive rapid evolution.

Selecting Suitable Targets Based on Pathway Involvement
To prioritize pathways containing genes that are less highly conserved across arthropods and whose knockdown is less likely to cause lethality, an understanding of species-specific gene function is required. Unfortunately, we do not always have gene function defined across species. Overwhelmingly, given the status of Drosophila melanogaster as a long-standing genetic model organism, dipteran model systems serve as the standard from which entomologists have come to understand gene function in insects [83]. Even though genomic and transcriptomic resources are increasingly available for common hemipteran pests, including whiteflies [22][23][24], the function of many insect genes are rarely assigned based on functional genomics or other functional studies. Rather, many of the assigned gene functions have been extrapolated from laboratory experiments on D. melanogaster and/or inferred from sequence similarity with Drosophila homologues [84,85]. These approaches are imperfect, as hemipteran pests, including whiteflies, exhibit distinct life history strategies and physiology. Thus, current and ongoing functional genomic studies will provide invaluable insight into selecting genes to silence. A further reason that selecting RNAi targets in pathways of known functions is a desirable long-term option over reliance on annotated functions extrapolated from Drosophila is that this would allow for more precise scientific monitoring and troubleshooting of RNAi behavior.
Categories of sublethal target pathways include those involved in gametogenesis, immunity, movement, mating, and appetite stimulation. Targeting genes from pathways involved in the detoxification of plant defenses has already shown to be successful at reducing, but not eradicating, B. tabaci populations [58]. Table 1 provides examples of potential targets. For B. tabaci, genes related to mating behavior may not be effective because of their ability to reproduce parthenogenically. More promising are genes targeting gametogenesis and immunity. Knocking out reproductive capacity is attractive because B. tabaci reproduction can be knocked out at several stages. Potential genes include those involved in gamete viability such as Boule, Vasa, or DNA methyltransferase 1 (Dnmt1) [86][87][88][89]. The latter suggestion is a good example of the importance of understanding specific gene function. Conventionally, Dmnt1 is expected to maintain methylation in the genome. However, methylation in insects does not appear to be universal, and some species such as the genetic model species D. melanogaster and Tribolium confusum lack Dmnt1 altogether [90]. Yet in Oncopeltus fasciatus, a hemipteran, co-authors have shown through RNAi knockdown that Dnmt1 has an essential function in gamete production in both males and females and that the silencing of Dnmt1 results in sterility without causing mortality or disrupting somatic tissue [86,87]. The Dnmt1 gene sequence has been identified in B. tabaci and successfully knocked down using RNAi [91].  [86,87] Another possible sublethal target is immunity. Targeting immunity genes has the two-fold benefit of reducing fitness and increasing efficiency of the dsRNA construct. The immune system barriers that evolved to protect B. tabaci from infection also have the capability to degrade the dsRNA before it can be processed by the endogenous RNAi machinery. In many cases, the dsRNA is ingested and subsequently degraded by nucleases in the saliva or the gut [59]. This phenomenon has been observed in many insects and contributes to variation in RNAi efficiency seen between and among species [100][101][102]. It has been demonstrated in B. tabaci, as well as other insect orders, that dsRNA-mediated silencing of digestive nucleases enhances the efficiency of silencing target genes [53,[103][104][105].

Conclusions
The remarkable diversity of pest management strategies is matched only by the ability of insects to overcome them. RNAi strategies that take into consideration life history and population genetics are likely to be valuable control strategies. Success in B. tabaci control will depend not only on building on our knowledge of the success and failures of current approaches, but also on how well we address long-i.eterm issues. Using RNAi to resolve these challenges offers a unique opportunity to explore these novel aspects of B. tabaci, including evolution of resistance and reproductive biology, while also managing their economic damage. It also makes it possible to address other issues such as application effects and cost issues, which are crucial factors for the development of practical biopesticides. For the best long-term outcomes, future RNAi applications using B. tabaci should consider selecting non-lethal target genes that confer continuous traits such as fecundity. Further developing RNAi for use as a sustainable pest management strategy will require extensive collaboration between IPM specialists, laboratory scientists, and public policy makers. Funding: This work was funded by the USDA Non-Assistance Cooperative Agreement #58-6080-9-006 "Managing Whiteflies and Whitefly-transmitted Viruses in Vegetable Crops in the Southeastern U.S."