Key message

  • Insect vector-borne plant diseases have a great importance worldwide in crops, particularly those transmitting viruses or phloem- and xylem-restricted bacteria.

  • Insect vectors harbor numerous facultative symbionts that could be used to control them.

  • During the last decade, researches on human and animal vector-borne epidemics have stimulated interest to experience vector control through symbionts manipulation.

  • We review potential candidates, advantages and limitations of the development of such techniques in vector-borne plant diseases management in agriculture.

Introduction

Since the beginning of agriculture, crops have suffered high losses due to pest injuries. Even now, the potential losses due to crop pests are estimated to range from 50 to 80 % depending on the crops and the planting latitude. Pest control only limits the outbreaks by 25–40 % (Oerke 2006). Pesticides, including insecticides, have been extensively used for less than a century but with important ecological side effects on targeted agrosystems and non-targeted ecosystems: the emergence of animal and plant parasite resistance (Hazarika et al. 2009; Powles 2008; van den Bosch and Gilligan 2008), the resurgence or emergence of new pests (Hardman et al. 1991; Hazarika et al. 2009), and the destruction of natural enemies (Croft and Brown 1975; Desneux et al. 2007). The impact of pesticides on non-target species, including humans, has been corroborated (Coats 1994; Elbaz et al. 2009) since Rachel Carson’s pioneering publication (Carson 1962). Additionally, the increased use of pesticides in agriculture over the last 40 years has not significantly decreased yield losses (Oerke 2006). Thus, alternative pest control approaches are urgently needed. The public’s perception of the environmental impact of pesticides in agriculture intensifies the need to develop more sustainable agriculture and practical integrated pest management solutions that reduce the dependence on insecticides. Crop protection has to be based on a more rigorous and better-targeted use of chemicals and/or the development of alternative techniques.

Insect vector-borne diseases, particularly those with viral, or phloem- and xylem-restricted bacterial causative agents, have a great, and yet underestimated, role in crop losses. The epidemiology of vector-borne diseases greatly depends on their vectors, which are the only route of dissemination for many pathogens. The effectiveness of a vector-borne pathogen depends on the vectorial capacity, which is related to the vector’s density, feeding activity on hosts, longevity before and after pathogen ingestion, duration of the incubation period, and vector competence (Cook et al. 2008). During the last decade, research on human vector-borne epidemics has stimulated interest in novel control strategies targeting different parts of the vector cycle. Recent research also highlighted the role of symbiotic microorganisms in the coevolution of hosts and their parasites (Dheilly et al. 2015), which stresses its potential importance in biological control programs. Our purpose is to draw a comprehensive parallel between this field of research and agronomy to increase our understanding of plant disease vectors and to present future control scenarios.

Many pest insects that feed on poor and unbalanced food sources, such as phloem and xylem, harbor symbiotic bacteria (endosymbionts) within specialized host cells, called bacteriocytes, which can group together to form the bacteriome (Buchner 1965). Symbionts supply the host with components lacking in the host’s diets, thereby, improving the host’s physiology and invasive power. Symbionts can be divided into two categories: (i) obligate symbionts, which contribute to the nutrition of the host and are essential for its survival, and (ii) facultative symbionts that are non-essential and have diverse effects on host fitness (Buchner 1965). Obligate and facultative symbionts differ in many respects, including their maintenance and transmission across host generations. Obligate symbionts are directly transmitted to a host from its parents (vertical transmission) and have essential functions in the host’s biology, including its metabolism and reproduction (Heddi et al. 1999). Facultative symbionts usually have less specialized interactions but also have intimate relationships with their host. They are transmitted both vertically and horizontally, and seem to infect host populations repetitively depending on environmental stress conditions, including parasitism and high temperature (Moran et al. 2008).

Two control strategies against plant vector-borne diseases based on insect–symbiont manipulation could be developed: (1) vector eradication targeting the essential host-physiological functions in which obligate symbionts are involved, or (2) facultative symbiont-mediated manipulation of insect traits with the aim of suppressing the vector’s transmission ability without the objective of vector eradication. Facultative symbionts appear to be the best candidates for manipulating symbiont control in insect vectors. In this review, after a brief overview of the insect vectors of pathogens causing crop diseases and the symbionts of these vectors, we attempt to identify how symbiont manipulation could contribute to an ecological engineering control approach and suggest research perspectives that should be developed.

Insect vectors of pathogens causing crop diseases

Most pathogenic viruses (>200) and tissue-associated bacteria are transmitted from plant to plant by insects and nematodes (Hogenhout et al. 2008). Vector-borne microorganisms, including viruses, bacteria, and phytoplasmas, are mostly associated with hemipteran insects that feed on plant vascular tissues. In the Sternorrhyncha sub-order, aphids (Aphididae) and whiteflies (Aleyrodidae) are the most important vectors of viruses (Hogenhout et al. 2008), while psyllids (Psyllidae) transmit bacteria and phytoplasmas (Haapalainen 2014). In the Auchenorrhyncha sub-order, leafhoppers (Cicadellidae), froghoppers (Cercopidae), and planthoppers (Delphacidae, Cixiidae, and Dictyopharidae) are able to transmit both viruses, bacteria, and phytoplasmas (Hogenhout et al. 2008; Weintraub and Beanland 2006). Thrips (Thysanoptera) transmit only a few viruses but these have significant economic impacts (Hogenhout et al. 2008; Whitfield et al. 2005). All of the plant pathogens transmitted by these insects cause important diseases in crop plants, including grapes, citrus, coffee, fruit trees, rice, vegetables, and ornamentals. The rice-infecting viruses that are transmitted by planthoppers and leafhoppers have resulted in significant disease outbreaks in Asia (Hibino 1996). Severe yield losses caused by begomoviruses transmitted by whiteflies have been reported for several crop plants, including tomato, cassava, and cotton (Briddon and Markham 2000; Costa 1976; Czosnek and Laterrot 1997; Legg et al. 2014). Barley yellow dwarf virus, transmitted by aphids, causes considerable losses of small grain cereals worldwide, including yield losses of up to 46 % in Triticum aestivum (wheat), 25 % in Hordeum vulgare (barley), and 15 % in Avena sativa (oat) (Larkin et al. 2002; Ordon et al. 2009). Phytoplasmas, wall-less bacteria, are associated with more than 700 diseases in hundreds of plant species (Weintraub and Beanland 2006). The bacterium Xylella fastidiosa that causes Pierce’s disease of grape and citrus variegated chlorosis, two economically important diseases, also affects many other plant species, such as alfalfa, peach, plum, almond, elm, coffee, sycamore, oak, maple, olive, and pear (Chatterjee et al. 2008).

Vector transmission is a complex process and, irrespective of the pathogen, different modes of transmission are encountered (Nault 1997 and Table 1): (1) non-persistent and semi-persistent, (2) persistent, circulative, and (3) persistent, circulative propagative. These systems of transmission imply several, narrow, and specific interactions between a plant pathogen and its insect vector. Indeed, some of these plant-pathogenic agents are considered symbionts of their insect vector, while also being plant pathogens (Caspi-Fluger and Zchori-Fein 2010). In some cases, the pathogen can invade the oocytes within the female vector insect and be transmitted to the offspring, which is the case for the rice stripe virus transmitted by the planthopper Laodelphax striatellus (Huo et al. 2014) and phytoplasmas (Hanboonsong et al. 2002; Kawakita et al. 2000; Tedeschi et al. 2006). The specificity of the transmission can be very strict or, alternatively, several viruses or bacteria can be transmitted from plant to plant by the same insect (Hogenhout et al. 2008; Lee et al. 2000). The latent period, the time between the vectors’ acquisition of the pathogen and the moment when the vector is able to inoculate plants, may be considerably different among the vectors and the pathogens, ranging from no latent period to a long period of days or weeks.

Table 1 Representative examples of pathogens transmitted by insect vectors to cultivated plants
Full size table

Non-persistent transmission

Non-persistent transmission occurs through insects that can transmit the pathogen only during a short time after acquisition, i.e., up to 12 h. In general, the non-persistent transmission of viruses occurs after a brief period of acquisition and no latent period is required. This is the case for the potato virus Y and other potyviruses (Quenouille et al. 2013). In such cases, viruses are restricted and linked to the aphid’s stylet tip and, more rarely, to the foregut. The interactions between viruses and insects is mediated by direct interactions between capsid and insect proteins, as with cucumber mosaic cucumovirus (Bricault and Perry 2013), or by “helper” proteins distinct from capsid proteins, which facilitate the transmission by aphids, as with tobacco vein mottling virus (Maia et al. 1996). Since aphids rapidly lose their infectivity, such viruses cannot be dispersed over a large geographical area. However, due to their feeding behavior and their capacity to probe different plants, including non-hosts, aphids can transmit viruses to many different plant species.

Bacteria of the genus Xylella are transmitted by xylem sap-feeding insects, such as sharpshooters, leafhoppers, and spittlebugs (Chatterjee et al. 2008). Xylella transmission does not require a latent period (Purcell 1979). The bacteria are localized in the vector foregut where they attach and multiply, and can be transmitted for several weeks or months by the insect depending on its lifespan, or on the occurrence of molting during which the insects become uninfected (Almeida and Purcell 2006; Fuente et al. 2007).

Persistent, circulative transmission

In persistent circulative transmission, vectors can transmit the pathogen for several days or weeks, depending on how long the insects feed on the infected plants, and thus, the amount of pathogen ingested. In this case, the plant-pathogenic agent is not able to multiply in its insect vector but has to circulate in it to be transferred to the plant. For example, the beet curly top virus can be transmitted by the leafhopper Circulifer tenellus after a latent period of few hours, but the transmission period can continue for several weeks (Carter 1973; Soto and Gilbertson 2003). Luteoviruses that cause barley yellow dwarf and potato leaf roll are other examples of viruses transmitted over a long period by aphids without multiplication in their vector (Harrison 1958; Miller and Rasochová 1997). The potato leaf roll virus specifically binds the symbionin, a symbiotic bacterial protein present in the hemolymph, allowing the virus to be carried to the accessory lobes of the aphid’s salivary glands (van den Heuvel et al. 1999). Maize streak virus, which causes a serious disease of maize in many regions of Africa, is transmitted in a persistent manner by leafhoppers in the genus Cicadulina (Bosque-Pérez 2000).

Persistent, propagative transmission

In this type of transmission, microorganisms have to multiply in their insect host to complete their persistent, propagative transmission to the plant. They often multiply in several organs that are not directly linked to the acquisition/inoculation (gut and salivary glands), thus, they can also affect the physiology or behavior of the insect vector. The length of the latent period depends on the multiplication and colonization of the insect by the plant pathogen. It can be sensitive to temperature, such as in the case of phytoplasma multiplication (Salar et al. 2013). Bacteria using this mode of transmission belong to the Mollicutes (phytoplasmas and spiroplasmas), Alphaproteobacteria (“Candidatus Liberibacter sp.”), and Gammaproteobacteria (“Candidatus Arsenophonus phytopathogenicus” and “Candidatus Phlomobacter fragariae”) classes (Bressan 2014; Grafton-Cardwell et al. 2013; Weintraub and Beanland 2006). They are restricted to the phloem of the plant. Phytoplasmas are responsible for a wide array of diseases on many crops of great economic interest, whereas spiroplasmas cause the citrus stubborn (Spiroplasma citri) and corn stunt (Spiroplasma kunkelii) diseases (Garnier et al. 2001). Both phytoplasmas and spiroplasmas are transmitted by species of insects among the leafhoppers, planthoppers, and psyllids. The psyllids Diaphorina citri and Trioza erytreae are the vectors of Liberibacter that causes huanglongbing in Citrus (Grafton-Cardwell et al. 2013; Van den Berg 1990). Viruses are also transmitted using persistent, propagative transmission. At least 14 rice viruses transmitted by planthoppers and leafhoppers cause rice diseases. Vector-borne reoviruses replicate in several organs of their insect vectors and some can spread to the ovary and be transmitted to offspring. Viruses are also able to multiply in whitefly, such as tomato yellow leaf curl virus in Bemisia tabaci (Ghanim 2014).

Symbionts

Eukaryotes, from plants to mammals, maintain complex interactions with microorganisms. In insects, these microorganisms are most frequently bacteria living intracellularly, but viruses and yeast-like microorganisms can also occur (Heddi and Gross 2011). When those microorganisms form intimate associations with host species and form chronic infections, they are commonly called “symbionts” (Moran 2006). The effects of symbionts on the host can vary from mutualism to pathogenicity, depending on the symbiont–host association and may have beneficial and detrimental effects at the same time (Moran 2006; Zug and Hammerstein 2015). Symbionts play an important role in the evolution of their hosts. Two main interactions are encountered: obligate and facultative relationships. In obligate symbiosis, symbiont acquisition is ancient and co-diversification between symbionts and their hosts is reported (Baumann 2005). While obligate symbionts are essential for the host’s survival, reproduction, and development, the facultative symbionts may instead play important roles in the ecology and adaptation of hosts in a given environment. The principal mode of symbiotic transmission is vertical, i.e., transmission through the maternal line, but horizontal transmission, i.e., intra or interspecific transmission from one insect to another, can occur and it permits colonization of insects by facultative symbionts. The lack of genetic divergence among some facultative symbionts within and between herbivorous insect species indicates that horizontal transmission is an important mode of dissemination (Russell et al. 2003; Sandström et al. 2001; Thao et al. 2000). Horizontal transmission has also been involved in the replacement of obligate symbionts (Conord et al. 2008; Lefèvre et al. 2004; Pérez-Brocal et al. 2006). This review focuses on facultative symbionts that are the easiest targets for vector control.

Facultative symbionts can be inherited vertically (Frydman et al. 2006; Sacchi et al. 2008) but are more often horizontally transmitted. For example, horizontal transmission can occur through food plants (Caspi-Fluger et al. 2012), parasitoid wasps (Ahmed et al. 2015; Gehrer and Vorburger 2012) or during mating (Moran and Dunbar 2006). This allows the symbionts to colonize diverse species as is the case for Hamiltonella and Arsenophonus, which are found in different species of aphids, psyllids, whiteflies, and planthoppers (Baumann 2005; Dale et al. 2006; Gherna et al. 1991; Russel et al. 2003; Sémétey et al. 2007). These symbionts are erratically distributed in the insect population as for example within the different putative species whitefly Bemisia tabaci, which harbor contrasting symbionts (Chiel et al. 2007; Gueguen et al. 2010; Gnankiné et al. 2012; Tajebe et al. 2015). They reside in the hemolymph and/or colonize a wide variety of organs, including reproductive organs (Moran et al. 2005). When localized inside the host cells, they were found either in a vacuole or free in the cytoplasm. Genomes of facultative symbionts often contain elements related to the type III secretion system (TTSS) that are implicated in the pathogenesis of many enteric bacteria. These genes may help the symbiont to invade insect cells by injecting effectors directly into the cytoplasm of host cells. The Hamiltonella genome contains homologs of Salmonella enterica Pathogenic Island 1 and 2 genes that code for TTSS proteins (Moran et al. 2005), and Arsenophonus nasoniae has two complete TTSSs that are very similar to that of Yersinia in terms of gene content and order (Wilkes et al. 2010). Facultative symbionts have a variety of biological characteristics, ranging from mutualism to parasitism. Facultative symbionts can be cultivated outside their natural hosts in insect cell lines, such as Arsenophonus and Rickettsia (Dale et al. 2006; Darby et al. 2005; Pontes and Dale 2006; Welburn et al. 1987).

Some symbionts can colonize a wide variety of arthropods. This includes Arsenophonus, which infects approximately ~5 % of arthropods, including arachnids, such as ticks, and insects, such as aphids, psyllids, whiteflies, planthoppers, cockroaches, true bugs, bees, wasps, lice, flies, and beetles (Duron et al. 2008). Acetic acid bacteria of the genus Asaia are symbionts of phylogenetically different insect species, including anopheline mosquitoes and the phytoplasma vector Scaphoideus titanus (Hemiptera: Cicadellidae) (Crotti et al. 2010). Cardinium has been found in leafhoppers, as well as in S. titanus (Marzorati et al. 2006) and whiteflies (Thierry et al. 2011; Weeks et al. 2003). Wolbachia is an important symbiont that infects most arthropod species (Hilgenboecker et al. 2008; Zug and Hammerstein 2012) and among them are insect vectors, including aphids (De Clerck et al. 2014; Wang et al. 2014). Additionally, one organism can host many different prokaryotes. Symbiont communities in aphids have been found to be diverse and dynamic, with up to eight facultative symbionts, Serratia, Hamiltonella, Rickettsia, Spiroplasma, Wolbachia, Arsenophonus, Regiella, and Rickettsiella (Oliver et al. 2010; Tsuchida et al. 2010). Seven facultative symbionts, Arsenophonus (Chiel et al. 2007; Thierry et al. 2011; Zchori-Fein and Brown 2002), Hamiltonella (Chiel et al. 2007; Thierry et al. 2011), Wolbachia (Chiel et al. 2007; Weeks et al. 2003; Zchori-Fein and Brown 2002), Cardinium (Thierry et al. 2011), Fritschea (Everett et al. 2005), Rickettsia (Chiel et al. 2007; Gottlieb et al. 2006; Thierry et al. 2011), and Hemipteriphilus (Bing et al. 2013), were found in the whitefly B. tabaci, depending on its genotype (Chiel et al. 2007; Thierry et al. 2011), and the symbiont community of Auchenorrhyncha includes Asaia, Cardinium, Rickettsia, and Wolbachia (Gonella et al. 2012, 2011).

Facultative symbionts and insect life history traits

Facultative symbionts can drive the development of their host by improving its survival, development rates, and/or influencing its sex ratio (Table 2). Symbiotic bacteria can also have huge effects on insect physiology, such as providing a better resistance to abiotic stresses, such as heat (Brumin et al. 2011; Montllor et al. 2002) or speeding up larval development (Chouaia et al. 2012). Interestingly, facultative symbionts also have roles in typical parameters, such as adaptation to the host plant, pathogen transmission, vector relations with natural enemies, and insecticide resistance that could be used in managing vectors of crop diseases (Table 2).

Table 2 Representative examples of facultative symbiont effects on insect vector life history traits
Full size table

Sex-ratio disturbance

Insect population dynamics can be hugely distorted by symbiont strains that skew host offspring sex ratios. The sex-ratio disturbance is mainly due to the killing of males by symbionts that are maternally inherited through the egg cytoplasm (Engelstadter and Hurst 2009; Ma et al. 2014). Among them, diverse clades of bacteria, such as Arsenophonus, Cardinium, Rickettsia, Spiroplasma, and Wolbachia, are able to manipulate their host’s reproductive system to increase the proportion of infected individuals in the female populations, and thus, their spread (Engelstadter and Hurst 2009; Himler et al. 2011, Ma et al. 2014). There are four main reproductive manipulations: (1) killing of male symbiont hosts (Hurst and Jiggins 2000), (2) cytoplasmic incompatibility, in which matings between infected and uninfected individuals, or between two individuals infected by different strains of the same bacterium, do not produce viable offspring (Perlman et al. 2006; Werren 1997), (3) feminization of males, in which infected males reproduce as females (Rousset et al. 1992), and (4) thelytokous parthenogenesis, in which non-fertilized eggs produce females (Groot and Breeuwer 2006; Pannebakker et al. 2004).

Adaptation to the host plant

The role of symbionts in the affiliation of aphids with different host plants has been investigated and provided interesting results. The study of the relationships between Acyrthosiphon pisum populations collected from different plant species and their symbiotic microbiota showed a strong link between symbiont occurrence and pea aphid host plant specialization (Ferrari et al. 2004; Leonardo and Muiru 2003; Simon et al. 2003; Tsuchida et al. 2004). Thus, correlations between certain facultative symbionts and the use of particular host plants by aphid and psyllid species have been demonstrated (Ferrari et al. 2012; Russell et al. 2013). For example, genotypes of A. pisum harbored Serratia symbiotica and Rickettsia when sampled from pea or bean and they harbored Hamiltonella defensa when collected on alfalfa or lotus (Ferrari et al. 2004; Frantz et al. 2009; Simon et al. 2003). Interestingly, aphid interspecific Regiella insecticola transfer from A. pisum to the vetch aphid Megoura crassicauda confers to M. crassicauda the ability to develop on a non-host plant, i.e., a plant on which a phytophagous insect species cannot grow or reproduce (Tsuchida et al. 2011). Additionally, the host plant-symbiont association can have an effect on insect fitness (Chen et al. 2000). The presence of R. insecticola in pea aphid lines specialized on white clover provides a greater fecundity compared to R. insecticola-free clones. The experimental manipulation of an R. insecticola infection in a single aphid genotype indicated that this symbiont enhanced female fecundity on clover compared with on vetch, but induces a higher mortality when fed alfalfa (Tsuchida et al. 2004). The roles and relative importance of symbionts in plant specialization greatly depend on the genetic backgrounds of the insects and bacteria (Ferrari et al. 2007; Leonardo 2004).

Symbionts can also help their insect host to exploit a resource by regulating the plant’s defense-related genes (Barr et al. 2010; Casteel et al. 2012), detoxifying plant allelochemicals (Dowd and Shen 1990; Karban and Agrawal 2002) and/or manipulating plant physiology to enhance the resource quality. For example, Wolbachia symbionts of the leaf-mining moth Phyllonorycter blancardella inhibit apple tree leaf senescence by increasing cytokinin production, which provides P. blancardella a longer feeding period (Kaiser et al. 2010). In another example, several resistance genes from some varieties and wild rice species have been identified and incorporated into high-yielding rice varieties to reduce planthopper and leafhopper damage to rice in Asia (Ferrater et al. 2013). However, insects quickly adapted to the resistant rice and virulent vector populations are now common. The implication of a symbiont is a probable explanation for such a quick adaptation (Ferrater et al. 2013).

Facultative and obligate symbionts can also improve host plant exploitation by providing nutrients that insect hosts alone cannot synthesize. Thus, yeast-like symbionts recycle uric acid synthesized by the planthopper Nilaparvata lugens (Hemiptera: Delphacidae) as an amino acid source (Sasaki et al. 1996). The same phenomenon was described with an Erwinia-like symbiont and its host Parastrachia japonensis (Hemiptera: Parastrachiidae; Kashima et al. 2006).

Transmission of pathogens

Facultative symbionts are not confined to bacteriocytes and can be found in various parts of their insect host. Thus, they can be co-localized with plant pathogens in some insect tissues, such as the midgut and salivary glands (Marzorati et al. 2006; Rana et al. 2012), and can affect the transmission of plant pathogens. For instance, in aphids and whiteflies, symbionts produce proteins that bind to virus particles and protect them in the insect hemolymph to allow a high transmission efficiency level (Gottlieb et al. 2010; Rana et al. 2012).

Breaking the life cycle of vector-borne human diseases by blocking the development of the pathogen in the mosquito is a promising research area (Riehle and Jacobs-Lorena 2005). A pathogen-blocking effect of symbionts present in the vector has been demonstrated for a wide variety of pathogenic organisms, such as nematodes, viruses and protozoans (Kambris et al. 2009; Moreira et al. 2009). The density and/or the distribution of the symbiont in the tissue of the vector can efficiently reduce the pathogen’s success (Lu et al. 2012; Moreira et al. 2009). Several hypotheses have been proposed for this phenomenon, (i) stimulation by the symbiont of the vector’s innate immunity (Kambris et al. 2009; Rancès et al. 2012), (ii) competition for key cellular locations or resources between the symbiont and the pathogen (McGraw and O’Neill 2013), and/or (iii) the production by the symbiont of toxins which inhibit the pathogen (Aksoy 2003; Carter et al. 2013). However, the successful use of mass release of engineered vectors depends upon how the toxin produced by symbiont affects the vector fitness and allow them to compete with wild populations of vectors (Carter et al. 2013).

Protection against natural enemies and pathogens

Facultative symbionts can also provide protection against insect pathogens and parasitoids. The efficiency of the resistance afforded by symbionts against natural enemies varies greatly depending on host and parasitoid genotypes (Ferrari et al. 2001; Vorburger et al. 2009). Symbionts belonging to the genera Regiella, Rickettsia, Rickettsiella, and Spiroplasma can directly reduce mortalities due to fungal infections, but also indirectly by decreasing sporulation on dead aphids (Łukasik et al. 2013). The presence of the symbiont R. insecticola allows a better resistance to the fungal pathogen Pandora neoaphidis in the aphid A. pisum (Scarborough et al. 2005). This symbiont is also associated with resistance in two other aphid species, Myzus persicae and Aphis fabae, to the parasitoid wasp Aphidius colemani (Vorburger et al. 2010). H. defensa is a symbiont that is well-known to offer protection against parasitoid wasps (Vorburger 2014). Although the mechanisms involved in the resistance to natural enemies are globally poorly known, the protection provided by H. defensa is due to a toxin-encoding bacteriophage carried by a symbiont (Oliver et al. 2009). The presence of this bacteriophage is essential in the protection provided by H. defensa, but it also provides a higher weight at adulthood and numerous offspring (Weldon et al. 2013). Protection by Wolbachia against viruses was also reported in dipterans (Chrostek et al. 2013; Hedges et al. 2008). This protection provided by symbionts, however, has a cost and the presence of such symbionts in insect populations decreases without parasitic pressure (Oliver et al. 2008).

Using symbionts to enhance predation and parasitism on insect vectors represents an interesting opportunity for crop protection. The susceptibility of aphids to parasitoids and predators is partly related to their color and some studies showed that green aphids suffer higher rates of parasitoid attacks while red ones were more susceptible to predation (Libbrecht et al. 2007; Losey et al. 1997). The body color of the pea aphid varies according to the amount of blue-green polycyclic quinones, which are modified by a Rickettsiella infection (Tsuchida et al. 2010). In addition to color, kairomonal prey odors are important for parasitoids and predators during their food search. Symbionts can make their host more or less attractive to their natural enemies. Staphylococcus sciuri, which inhabits some aphid guts, when present in honey dew, produces odors that attract and stimulate oviposition of females of the larval aphid predator Episyrphus balteatus (Diptera: Syrphidae) (Leroy et al. 2011).

Resistance to insecticides

The extensive use of chemical insecticides for decades has led to the development of insecticide resistance. The mechanisms involved in this resistance are diverse but are often linked to changes in pest genomes. They include the over-expression of detoxifying enzymes, alterations of pesticide target loci, reductions in the sensitivity of the sodium channels to insecticide binding, and/or the up-regulation of degrading esterases (Denholm and Rowland 1992; Hemingway and Ranson 2000). Insecticide resistance due to an insecticide-degrading bacterial symbiont was recently reported (Kikuchi et al. 2012). The pest and fungal vector Riptortus pedestris (Hemiptera: Alydidae) acquires gut symbiotic bacteria of the genus Burkholderia from soil during the nymphal instar stages. Some strains of this symbiont are able to degrade fenitrothion and confer resistance to the host insects that harbor it. Burkholderia symbionts can live in the absence of pest insects and their development is favored under fenitrothion applications. Thus, under insecticide pressure, the level of fenitrothion-degrading Burkholderia strains increases, which in return increases their probability of being acquired by the insect, increasing the proportion of insecticide resistance in the pest population. On the other hand, the insecticidal activity of Bacillus thuringiensis on Lymantria dispar (Lepidoptera: Erebidae) depends on the presence of midgut symbionts. In fact, individuals of L. dispar that harbor gut symbionts are not directly killed by B. thuringiensis, but the B. thuringiensis toxin enables the midgut bacteria to reach the hemocoel and cause septicemia by permeabilizing the gut epithelium (Broderick et al. 2006). In other systems, however, symbionts can have protective roles against B. thuringiensis and gut flora-cured mosquitoes have a high mortality rate when exposed to B. thuringiensis (Patil et al. 2013).

Symbiont manipulation in integrated vector management

In disease vectors, the vectorial capacity (V), i.e., the ability to spread a disease, has been related to five main variables (Chuche and Thiéry 2014; Cook et al. 2008):

$$V = \frac{{ma^{2} p^{n} b}}{ - \ln p},$$

where m is the vector density, a is the probability of the vector feeding on a host, which can be extended to all behaviors related to feeding, such as leaf choice, plant choice, and spatial dispersion, p is the probability of vector survival, n is the duration of the latent period, b is the vector competence, and 1/(–lnp) is the longevity of the vector’s life after the latent period.

A higher vectorial capacity is thus attained when the insect vector has a wide range of host plants, is mobile to locate new host plants and new habitats, and has sufficient longevity to optimally acquire and transmit the pathogen. Feeding behaviors, like those of sucking insects with specialized mouthparts, allow for the efficient acquisition and transmission of pathogens. In vector-borne pathogens, the rate of transmission varies depending upon the vector or pathogens’ life history traits and environmental ecosystem parameters, such as resource availability and distribution of main hosts, potential alternative hosts, natural enemies, and climatic conditions (Daugherty et al. 2010). Thus, the transmission rate of vector-borne pathogens is strongly correlated with the size and dynamics of the vector population (Jeger et al. 2004), as well as its biology and behavior.

By using symbionts to affect these parameters, it is possible to influence the vectorial capacity and propose alternatives to chemical controls (Fig. 1). Shortening vector lifespans would reduce the transmission of pathogens that require long latent periods. This period, between pathogen acquisition and the ability to inoculate plants, implies that older insects contribute disproportionately to pathogen transmission (Cook et al. 2008). Consequently, small changes in vector lifespan could have huge impacts on the transmission dynamics of a disease (Cook et al. 2008; Rasgon and Scott 2004). To succeed in insect vector population control using insect microflora, the choice of the symbiont species/strains is critical. The symbiont’s features will determine its ability to be modified, its manner of dissemination and the limits of its efficiency (Table 3).

Fig. 1
figure 1

Example of symbionts able to influence the insect vectorial capacity. The five variables, related to the vectorial capacity, are effective drivers of action in the reduction of the transmission of pathogens by insect vectors. Useful symbionts can act on a vector species by modifying its feeding behavior (feeding on plant), shortening its lifespan (vector survival), increasing the latent period (latent period), blocking the pathogen transmission (vector competence), affecting its reproduction, and increasing its susceptibility to natural enemies (vector density). The arrows show published effects on each variable (see Table 2 and text for details)

Full size image
Table 3 Most promising facultative symbionts for vector-borne disease control
Full size table

Paratransgenic control

Selecting or modifying a symbiont that is able to produce a molecule to inhibit or kill the vector-borne pathogen or the vector itself is probably the easiest part of the process using the paratransgenic approach. Paratransgenesis consists of genetically modifying a symbiotic microorganism of an arthropod to express an anti-plant pathogen effector and then re-introducing it into the vector to generate a pathogen-resistant phenotype (Caljon et al. 2013). According to Beard et al. (2002), a successful paratransgenic control strategy has six requirements that could be extended to other symbiont-associated control approaches: (1) a symbiotic relationship between bacteria and vector, (2) bacterial cultivation, (3) a feasible and stable genetic modification, (4) maintenance of modified bacterial fitness, (5) production of an anti-parasitic molecule by the bacteria, and (6) an efficient dispersion system for the bacteria. Thus, best candidate microorganisms are free-living and acquired by arthropods from their environment, such as from water or plant surfaces. Symbionts that are dependent on the arthropod’s environment need to use insect host cells as cultivating media (Welburn et al. 1987), which is more complicated. Because Wolbachia is present in a high number of insect species (Hilgenboecker et al. 2008; Zug and Hammerstein 2012) and has been well studied (reviewed in Stouthamer et al. 1999; Werren 1997), it is probably the best candidate for vector-borne disease control. Symbiont paratransgenesis is based on the use of episomal plasmids that can express the anti-pathogen effector in an autonomous way, or integrative plasmids that allow for the insertion of genes into the symbiont’s chromosomal DNA (Caljon et al. 2013). Another good control candidate would be Asaia sp. that were suggested as potential targets for the paratransgenic control of anopheline mosquitoes (Riehle and Jacobs-Lorena 2005) and that are found in plant disease vectors, which acquire them during feeding (Crotti et al. 2009). Paratransgenic control is now being considered for vector-borne plant diseases. Thus, genetically modified yeast may be used against the planthopper vector of Fiji disease virus, Perkinsiella saccharicida, in sugarcane in Australia (Hughes et al. 2011). A Candida yeast was isolated from P. saccharicida, cultured and transformed. The transformation was accomplished by electroporation of Candida albicans codon-optimized plasmids, designed to integrate into the genome via homologous recombination. Transgenic lines were observed but no stably transformed yeast lines could be isolated. Overcome the obstacles to the production of stable engineered yeast lines will be necessary before considering the transformation of yeast symbionts of vector to produce a molecule to inhibit or kill the vector-borne pathogen as a potential way to control vector-borne diseases.

Symbiont dissemination

The most challenging step in the control of plant vector-borne diseases based on host–symbiont manipulation is successfully distributing the selected or modified symbiont in the vector’s population. The selected symbiont must survive in the environment prior to vector acquisition, and then spread and survive in the vector’s body. To ensure a better spread, symbionts should be horizontally and vertically transmitted. But, even if horizontal transmission enhances the symbiont spread in the vector population, this way of transmission allows in some cases the transfer from one species to another. Thus, vertical transmission should be avoided in cases of risk towards non-targeted species. The ability of the symbiont to be transmitted, and the manner of transmission, vertical and/or horizontal, depend on the biology of the symbiont.

Much of the research into the use of symbiotic bacteria to control human vector-borne epidemics was performed on malaria and its mosquito vector. The non-biting aquatic larval stage in the life cycles of some insects, such as mosquitoes, raised the possibility of bacterial acquisition by spraying the selected bacteria in water, even though more than 99 % of mosquito bacteria are lost between the last larval and the adult instar (Moll et al. 2001). Similarly, some Enterobacter species are being used to block Plasmodium sp. in Anopheles sp. vectors (Riehle and Jacobs-Lorena 2005). The goal is to prevent the pathogen from passing through the intestinal barrier after acquisition during feeding. Thus, the symbiont chosen for pathogen blocking should be localized in the midgut. Even though it appears promising, this approach is limited because midgut bacteria are not transovarially transmitted to the offspring. This method may also increase target specificity since the bacteria’s adaptation to the mosquito gut could restrict the bacteria’s ability to colonize non-target organisms. Additionally, the symbiont acquisition by feeding during larval life is critical in most plant pathogen vectors that are hemimetabolous, with nymphs and adults having the same feeding habits or ecological niche. Thus, it is more difficult to transfer symbionts to hemipteran than to dipteran because all of the instars feed on plant sap. Choosing the best manner to disseminate a symbiont in a species requires a good knowledge of the vector’s behavior. Some hemipteran, vectors of human and plant pathogens, can acquire facultative symbionts during feeding. For example, Rhodnius prolixus (Hemiptera: Reduviidae), a vector of Trypanosoma cruzi, which causes Chagas disease, acquires Rhodococcus symbionts by ingesting fecal deposits (Durvasula et al. 2003). The newly emerging first-instars are aposymbiotic and acquire their symbionts through coprophagy on fecal droplets deposited by adults. Trials aimed to infect R. prolixus first instar nymphs using simulated fecal paste containing modified Rhodococcus were successful (Durvasula et al. 2003). Since aphids can acquire bacteria from the plant surfaces or from the honeydew of infected aphids (Stavrinides et al. 2009), infecting vectors by spraying cultures containing a symbiont able to survive until ingestion could be tested. Attempts to establish horizontal transmission between infected and uninfected aphid feeding on the same plant have been unsuccessful (Chen et al. 2000), while laboratory techniques, such as rearing vectors on artificial diets and microinjection, have been, to date, more successful (Pontes and Dale 2006).

Symbiont dissemination could occur by parasitoids transferring them horizontally when sequentially stabbing infected and uninfected insects. In this way, H. defensa and R. insecticola were experimentally transferred among aphids (Gehrer and Vorburger 2012), and Wolbachia among B. tabaci individuals, which resulted in stably infected B. tabaci lines (Ahmed et al. 2015). The use of parasitoids as vaccine agents to transfer selected symbionts to vector populations should be considered even if the rate of transmission and symbiont establishment in the prey population is affected by the mortality induced by the parasitoids.

The best source of symbionts for acquisition could be the crop targeted for protection from the vector-borne disease. Thus, endophytic bacteria could be used as insect symbionts. Some bacteria living in plants are highly genetically similar to hemipteran symbionts, and vector-borne plant pathogens are good examples of microorganisms that can live both in plants and insects (Caspi-Fluger and Zchori-Fein 2010). Thus, the production of plants harboring target symbionts that act as an inoculum source for insect vector infections is possible. This plant-mediated transmission was achieved in the lab with the acquisition of Rickettsia by the whitefly B. tabaci that had fed on Rickettsia-containing plants previously infected by Rickettsia infected B. tabaci (Caspi-Fluger et al. 2012). Exploiting the abilities of some bacteria to live in plants and animals for vector control was investigated. The cultivable bacterium Alcaligenes xylosoxidans denitrificans, which can be found in grapes, was isolated from the sharpshooter Homalodisca vitripennis, a vector of X. fastidiosa (Bextine et al. 2004). This bacterium seems to be a good candidate to control Pierce’s disease because it can easily be modified to produce anti-X. fastidiosa factors, and colonizes both host plants and insect vectors, allowing for its potential employment in driving anti-pathogenic factors. In theory, according to the microbial ecological data available, such factors could operate within the insect or directly inside the plant, to cure infected grapevines by neutralizing or eliminating existing X. fastidiosa colonies (Bextine et al. 2005, 2004). This could reduce economic costs due to the removal and replacement of diseased grapevines and reducing costs associated with lost yields in subsequent seasons.

Advantages and limitations

A great advantage of using engineered symbionts is the easiness of producing large amounts of bacteria or yeasts as compared with producing equivalent amounts of insect vectors that would be required to ensure population replacement. Furthermore, the release of a weak population of vectors carrying selected symbionts into the wild can be sufficient to settle an engineered symbiont in wild vector population. Rapid spread in natural populations was hence obtained with a limited number of vectors carrying engineered Wolbachia in mosquito vector population (McGraw and O’Neill 2013). Bacteria can also be rather easily genetically modified to express several different effector molecules against vectors and/or pathogens (Riehle and Jacobs-Lorena 2005).

Another great advantage of using symbionts is their sustainability, except when their use leads to local population eradication followed by re-colonization. Indeed, well selected symbionts should be able to spread and maintain themselves in the vector population. For example, the life-shortening symbiont Wolbachia wMel, which was discovered in Drosophila melanogaster, has been successfully introduced into Australian mosquito populations, 100 % of individuals being infected 1.5 years after their initial release (McGraw and O’Neill 2013). The sustainability of this method could be reinforced by using symbionts as gene vectors to modify the life history traits of insects. Indeed, lateral gene transfer from symbionts to host has been demonstrated in four insect species and suspected in three others (Dunning Hotopp et al. 2007). Wolbachia is an intracellular symbiont that can colonize germinal cells, transfer up to their entire genome (>1 megabase) into the insect genome and provide new heritable traits to the insect host. Because bacterial sequences are usually considered as contamination during eukaryote genome sequencing, the importance of gene transfers from symbionts to eukaryote hosts is probably underestimated and could be a promising way to modify insect vectors traits (Dunning Hotopp et al. 2007). Selected or engineered strains of Wolbachia could be used to disseminate interesting traits, such as pathogen transmission-blocking patterns, in insect vector populations. With this technique, even if the Wolbachia infection is not stable over time, selected traits may persist and spread. The complexity of the pathogen-blocking mechanism could also be an advantage in avoiding or delaying the emergence of insect resistance against this trait. If insect hosts take advantage of colonization by symbiont-mediated pathogen blocking, a co-evolutionary process maintaining the blocking trait could be expected (McGraw and O’Neill 2013).

Even though symbiont infection often has a fitness cost for the vector, it is not necessarily an obstacle to the success of control strategies against plant vector-borne diseases based on host–symbiont manipulation. Models showed that Wolbachia causing cytoplasmic incompatibility was able to spread in uninfected insect populations even with a 50 % reduction in host fitness, although the dissemination rate was affected (Turelli 2010). The success of such techniques depends on the effectiveness of the selected symbionts and their ability to be spread in the vector population. This success can also be under environmental control. The temperature hinders the efficacy of the wAlbB strain of Wolbachia which reduced Plasmodium falciparum prevalence and oocyst intensity at 28 °C, had no effect on prevalence and increased oocyst intensity at 24 °C, and had no effect on prevalence or intensity at 20 °C in Anopheles stephensi even though the vector was stably transinfected with the wAlbB strain (Murdock et al. 2014). Another factor that can limit the successful use of symbiont-modified vectors is the existence of reproductively isolated strains. Reproductive isolation prevents the members of two different populations that cross or mate from producing offspring, and thus, prevents the spread of the modified symbiont in vector population through vertical and horizontal transfer. In Anopheles gambiae sensu stricto, the most important vector of the malaria parasite in sub-Saharan Africa, there are two populations with different molecular forms of ribosomal DNA that coexist without cross-breeding (della Torre et al. 2002). Thus, the infection of individuals from one molecular form with a target symbiont will impact only this form and not lead to the infection of all A. gambiae s.s. The existence of such reproductively isolated strains is also known in plant vectors, such as the sympatric host-race evolution in Hyalesthes obsoletus (Hemiptera: Cixiidae) populations, a vector of the Stolbur phytoplasma in Germany (Imo 2013). The existence of H. obsoletus host plant races (bindweed and nettle) causes different disease epidemics (Imo 2013). The stinging nettle-specific phytoplasma strain leads to outbreaks of the grapevine disease Bois Noir and, in this case, it would be preferable to infect H. obsoletus–nettle populations with selected symbionts.

Potential risks associated to symbiont uses

Even though application perspectives could reduce the use of pesticides in crop protection, using symbionts could bear some ecological risks. While the potential of controlling plant diseases using vector or pest symbionts seems promising, the risk of create stronger vector insects may arise for example in case of fitness benefits conferred to them. Resistance to predators is one of the risks, as recently demonstrated in aphids (Polin et al. 2015). Another obvious risk is changing the a variable in the vectorial capacity model presented above (Cook et al. 2008). This probability of the vector feeding on a host, which can be extended to all behaviors related to feeding, such as leaf choice, plant choice, and spatial dispersion Insect harboring selected symbiont could be better adapted to their host plant crop or attack new crop species by becoming more specialized or reversely more generalist (Leonardo and Muiru 2003; Tsuchida et al. 2004). Vector feeding behavior, a key point in pathogen transmission, could also be affected and make the insect more efficient in disease spread. Such feeding behavior modification was observed for Wolbachia-infected mosquitoes A. aegypti that are less able to obtain blood meals in old age than uninfected ones due to a defect in the insect’s proboscis (Turley et al. 2009).

Recent findings (Bressan 2014; Salar et al. 2010) support the hypothesis that some plant-pathogenic bacteria transmitted by insect vectors originating from insect symbionts (Bové and Garnier 2002; Caspi-Fluger and Zchori-Fein 2010). Such a symbiotic origin of plant pathogens concerns mainly Mollicutes (Phytoplasma and Spiroplasma), γ-(Arsenophonus and Phlomobacter) and α-Proteobacteria (Liberibacter and Rickettsia). Hence a symbiont engineered decreasing the vectrorial capacity of the vector could potentially become a new plant pathogen and be responsible of new epidemics. Moreover, the existence of symbionts transferred horizontally through the plant from one vector species to another (Gonella et al. 2015) increases the risk of dissemination of a “plant-pathogenic insect-symbiont”. Indeed, other species than the targeted vector could spread the pathogen to different host plant species and create new uncontrolled epidemies in other crops. To minimize such risk, the engineered symbionts should not be chosen in clades where plant pathogens occur.

The main objective of control strategies against plant vector-borne diseases based on insect–symbiont manipulation is to have a sustainable alternative to the widely use of chemical spraying. But, in some cases, symbiont manipulation could have an antagonism effect on existing integrated pest management (IPM) approaches as it is observed between chemical pesticide use and IPM (Guesdes et al. 2016; Desneux et al. 2007). In this way, well established IPM, such as biological control, could be negatively impacted by the use of symbionts.

Hamiltonella sp. is a good symbiont candidate because it have genes coding for effectors helping it to invade insect cells (Moran et al. 2005), can be horizontally transferred, e.g., during feeding (Darby and Douglas 2003), mating (Moran and Dunbar 2006) and by parasitoid (Gehrer and Vorburger 2012). Thus, their spread in nature after the release of insect vectors harboring engineered Hamiltonella would be efficient and allow a stable establishment of this symbiont. This symbiont is also involved in plant adaptation (Frantz et al. 2009) and could be used in order to divert the insect vector from the crop to another plant. But Hamiltonella also offer protection to aphids against parasitoid wasps (Vorburger 2014) that are the main biological control agents against aphid (Boivin et al. 2012). This protection is not negligible and can reduce parasitism rate by 41.5 % in pea aphids exposed to Aphidius ervi (Oliver et al. 2003) and provide almost complete resistance against Lysiphlebus fabarum (Vorburger et al. 2009). Because, insect vector protection against parasitoids by symbionts occurs naturally in the field, the use of biological control agents would be act as a selection pressure and increase the proportion of insect vector harboring defensive symbionts. Biological control using parasitoid should be lead to the outbreak of resistance, as for insecticides (Vorburger et al., 2009). The mechanisms involved in the protection provided by H. defensa begin to be understood and are due to a toxin-encoding bacteriophage (Oliver et al. 2009). Because the presence of this bacteriophage is essential in the protection provided by H. defensa, inactivating it in engineered Hamiltonella could be a way to avoid providing parasitoid protection to insect vector. The most challenging issue will be to confer desired traits to insect vectors while avoiding traits allowing them to counter existing pest management techniques.

Conclusion

The use of symbionts to control vector-borne diseases appears to be within reach. The first success was not recent: the eradication of local populations of the filariasis vector by cytoplasmic incompatibility induced by Wolbachia in the 1960’s (Laven 1967). The further, stable introductions of such symbionts in vector populations against different species of mosquito (Hoffmann et al. 2011; O’Connor et al. 2012) confirmed the practicality of this strategy. Due to technical progress, understanding the basis of molecular and biochemical mechanisms that drive insect–symbiont interactions, and their consequences on host ecology, has improved. Thus, the use of symbionts in insect control is becoming realistic. Such promising techniques, which have sometimes demonstrated their success against human diseases, could be adapted to insect-borne diseases of plants. This, however, requires, as a first step, progress in our knowledge of insect vector microflora. Many plant pathogen vectors have similar habits, such as feeding on the same tissues. Therefore, using a single symbiont strain could affect many vector species. In addition to scientific and technical progress, such innovative techniques would need the approval of the public and the authorities. Because the use of emerging technologies can be rejected by the public, even when they aim at controlling human pandemics (McGraw and O’Neill 2013), it will be essential to anticipate possible public concerns.

Author contribution

All authors collected and analyzed the literature and wrote the manuscript. All authors reviewed and approved the final manuscript.