Could bacterial associations determine the success of weevil species?

The weevil superfamily Curculionoidea is the largest insect group and so the largest animal group on earth. This taxon includes species which represent an important threat to many economically important crops and, therefore, pose a risk to agriculture and food security. Insect–bacteria associations have been recognised to provide the insect host with many benefits, such as ensuring the acquisition of essential nutrients or protecting the host from natural enemies. The role of bacteria associations within the weevil superfamily remains nonetheless understudied in comparison with other insect taxa. This review draws together existing knowledge on the influence of bacteria associated with weevils known to be agricultural pest species. The implications of these weevil–bacterial associations in determining pest status and their relevance to targeted pest management interventions are discussed. Specific consideration is given to the role of bacteria in cuticle formation, flight activity, reproduction manipulation and adaptation to different environments and food sources.


| WEEVILS, AN UNDERSTUDIED BUT EXTENSIVE INSECT TAXON
The nature of the association between insects and bacteria has been a controversial point for decades. In the early 1900s, a few scientists started to hypothesise that the presence of bacteria in insects was not a random event as bacteria seemed to be involved in important insect processes (reviewed by Steinhaus, 1940). However, these studies were limited by the available techniques at that time, mainly microscopy and culturing, to study microorganism morphology and physiology (reviewed by Handelsman, 2004). In the late 1990s, bacterial studies experienced a paradigm shift when Carl Woese determined that the 16S rRNA gene sequence from the small prokaryotic ribosome subunit could be considered a molecular chronometer which could be used as a taxonomy identifier (Woese, 1987). This new approach enabled the identification of prokaryotes taxonomically from complex samples in a culture-independent manner. At the same time, the techniques employed to sequence nucleotide molecules advanced rapidly. In 2005, technological advances allowed the automation of multiple sequencing reactions in parallel. This improvement created a platform for mass sequencing at an unprecedented time/ cost efficiency, which represented the beginning of high-throughput sequencing (reviewed by Heather & Chain, 2016). Technological innovation, together with the use of the 16S rRNA gene to identify prokaryotes, have enabled remarkable progress in studies focused on insect-bacteria associations, although less effort has been invested in the weevil taxon. For instance Web of Science contains 1,512 results for the search term "aphid*" and "bacteria*" despite there being only 5,000 aphid species described (Blackman & Eastop, 2000). By contrast, Web of Science returns 335 results for the search term "weevil*" and "bacteria*" even though there are more than 60,000 described weevil species (Alonso-Zarazaga & Lyal, 2002; Web of Science database searched on April 6, 2020, UK).
To date, many studies have demonstrated the potential importance of the insect-bacteria partnerships for host fitness. Some insects with restricted diets rely on bacteria to compensate nutritional deficiencies. For instance, the pea aphid Acyrthosiphon pisum (Harris) is provided with essential amino acids and the vitamin riboflavin by its obligate endosymbiotic bacterium Buchnera aphidicola (Nakabachi & Ishikawa, 1999; reviewed by Douglas, 2016) and the tsetse fly Glossina morsitans (Westwood) is provided with essential vitamins by the obligate endosymbiotic bacterium Wigglesworthia glossinidia (Akman et al., 2002;Nogge, 1981). Importantly, certain bacteria have been shown to render their insect hosts less susceptible to predators and pathogens. This has been illustrated for the pea aphid, which is protected from parasitism by the parasitoid wasp Aphidius ervi (Haliday) when aphids are infected with the facultative bacterium Hamiltonella defensa (Oliver, Moran, & Hunter, 2005;Oliver, Russell, Moran, & Hunter, 2003) and from infection by the entomopathogenic fungus Pandora neoaphidis (Remaud & Hennebert) when aphids harbour the facultative bacterium Regiella insecticola (Scarborough, Ferrari, & Godfray, 2005). Bacteria can also influence host reproduction as in the case of members from the genus Wolbachia, broadly recognised as reproductive parasites, which for instance increase fecundity of the fruit fly Drosophila simulans (Sturtevant; Weeks, Turelli, Harcombe, Reynolds, & Hoffmann, 2007; reviewed by Werren, Baldo, & Clark, 2008).
The weevil superfamily Curculionoidea is the largest insect group (Alonso-Zarazaga & Lyal, 2002) and harbours agricultural pest species that are distributed around the world. For instance, the red palm weevil Rhynchophorus ferrugineus (Olivier) is an important pest of palm trees that causes considerable economic losses to crops grown in countries in the Gulf, Middle East and Europe (European Commission, 2011). The sibling weevil species Sitophilus oryzae (Linnaeus), the rice weevil, and Sitophilus zeamais (Motschulsky), the maize weevil, are important pests of stored grain, rice, maize, barley and wheat globally (Grahame, 2017), while the vine weevil Otiorhynchus sulcatus (Fabricius) damages a wide range of horticultural crops around the world (Moorhouse, Charnley, & Gillespie, 1992). Weevil species are, therefore, numerous and problematic for farmers globally. Studies focused on understanding weevil-bacteria associations are still scarce when compared with other insect families. Additionally, these are biased towards those weevil species that are relevant from an agricultural perspective. The present review has intended to conduct a comprehensive search of the literature but it is inevitably dominated by research focused on weevil agricultural pest species. In the following sections, research on weevil-bacteria associations is discussed and the implications of these associations for the biology of the weevil species are considered (Table 1). The intention here is also to assess the importance of these associations for the pest status of weevils, underlining the existing knowledge gaps and identifying priorities for future investigations in this field. Advancements in this research area will ultimately contribute to the development of improved weevil pest control strategies. (Linnaeus), although it remained undetermined as to whether the observed bacteria constituted a "symbiotic organ" or were simply "accessory cells" (Mansour, 1927(Mansour, , 1930Pierantoni, 1927). It was not until the beginning of the 21st century that Lefèvre et al. (2004), using a phylogenetic analysis of the 16S rRNA gene, identified this microorganism as a γ-proteobacterium and designated the new lineage Candidatus Nardonella. This lineage was subsequently found to be widespread throughout the weevil superfamily and it was estimated to have become associated with weevils 125 million years ago (Conord et al., 2008;Lefèvre et al., 2004). Nevertheless, some studies have found that Ca. Nardonella has been replaced in species of the genus Curculio and Sitophilus, highlighting the dynamic nature of insect-bacteria associations (Lefèvre et al., 2004;Toju et al., 2010;Toju, Tanabe, Notsu, Sota, & Fukatsu, 2013). Symbiont displacement may have occurred following genome degradation of the original symbiont and loss of function(s) essential for host survival, creating an opportunity for a facultative bacterium capable of substituting the function(s) to form a new obligate association (reviewed by Moya, Peretó, Gil, & Latorre, 2008;Moya, Gil, & Latorre, 2009;Sudakaran, Kost, & Kaltenpoth, 2017). There are well documented cases of this phenomenon for instance in aphids such as the cedar aphid Cinara cedris (Mimeur). The primary symbiont of this aphid, B. aphidicola, has undergone a process of genome degradation leading to the loss of functions essential for the aphid host. The abundant facultative symbiont Serratia symbiotica appears to fulfil the absent functions and is a putative candidate for substituting the primary symbiont (Pérez-Brocal et al., 2006). Subsequent studies focused on identifying Ca. Nardonella in other weevil species and on studying other features of its biology, such as population dynamics during different insect life stages or the location of the Ca. Nardonella bacteriocytes in insect tissues (Hosokawa et al., 2015;Huang et al., 2016;Mansour, 1930;Nardon, Lefevre, Delobel, Charles, & Heddi, 2002;Toju & Fukatsu, 2011). Importantly, Anbutsu et al. (2017) working on the black hard weevil Pachyrhynchus infernalis (Fairmaire) showed that Ca. Nardonella is involved in insect cuticle formation by contributing to tyrosine synthesis. Suppressing Ca. Nardonella, by

Egg development
Egg hatching rate after antibiotic treatment Chen, Lu, Cheng, Jiang, and Way (2012) and Son, Luckhart, Zhang, Lieber, and Lewis (2008) Gut and whole insect microbiota Nonetheless, in a more natural environment the developmental defects derived from the absence of this symbiont likely impair the fitness of the insect. The importance of the weevil cuticle is also revealed in Sitophilus weevils. In this weevil genus, Ca. Nardonella was substituted by another bacterial symbiont named Sodalis pierantonius or SPE. This symbiont, amongst other functions, provides the weevil host with tyrosine and phenylalanine, similarly to Ca. Nardonella.
These amino acids are then used in the formation of the adult cuticle.
In this way, symbionts ensure that adult insects form a thick protective cuticle in shorter time (Vigneron et al., 2014;Wicker & Nardon, 1982). This similar functionality in two separate symbiont species illustrates the importance of weevil cuticle.
Cuticle colouration was shown to be important in deterring predatory lizards in the weevils Pachyrhynchus tobafolius (Kano) and Kashotonus multipunctatus (Kôno) (Tseng et al., 2014). The cuticle was also shown to reduce water loss in the weevil species Rhynchophorus cruentatus (Fabricius) (Weissling & Giblin-Davis, 1993). Although not a weevil, the saw-toothed grain beetle Oryzaephilus surinamensis (Linnaeus) associated symbionts were also shown to be involved in cuticle melanisation and resistance to desiccation (Engl et al., 2018).

The subterranean termites Reticulitermes flavipes (Kollar) and
Reticulitermes virginicus (Banks) secrete β-1,3-glucanase onto the cuticle that prevents infection by the pathogenic fungus Metarhizium brunneum (Petch) (Hamilton et al., 2011 Antimicrobial peptides from the coleoptericin family are important in the weevil immune system. These peptides are also known to be involved in regulating symbiosis in some weevil species. In the maize weevil, a member of the coleoptericins family, named Coleoptericin-A, controls and restricts the population of its primary symbiont SPE inside the bacteriocytes (Login et al., 2011). This coleoptericin seems to be conserved as it interacts not only with SPE but also with Ca. Nardonella. Weevils treated with dsRNA targeting the insects with insects that were bacteria-free as a result of a heat treatment. In this way, Wicker and Nardon (1982) and Wicker (1983) showed that bacteria in these weevils are involved in the metabolic route that provides the insect host with phenylalanine and tyrosine from stored forms, and in the synthesis of the vitamins pantothenic acid, biotin and riboflavin. A subsequent study by Gasnier-Fauchet et al. (1986) also suggested that bacteria were involved in the metabolism of the amino acid methionine and its derivatives sarcosine and methionine sulfoxide.
Several studies have shown that the influence of bacteria on host fitness goes beyond compensating for poor diets. Heddi et al. (1993) observed in rice weevils that mitochondrial activity was higher in weevils with bacteria compared with weevils without bacteria, although mitochondrial enzymatic activities were absent in bacteria isolated from these weevils. Hence, based on earlier discoveries that the bacteria were involved in amino acid and vitamin metabolism, it was suggested that these bacteria could be providing mitochondria with intermediary metabolites to maintain normal activity. Later studies confirmed that the intracellular symbiont of Sitophilus weevils SPE interacts with mitochondria by providing this organelle with pantothenic acid and riboflavin (Heddi et al., 1999;Heddi, Charles, Khatchadourian, Bonnot, & Nardon, 1998). Grenier et al. (1994) reported reduced or null flight activity in both the rice weevil and the maize weevil when individuals were deprived of bacteria by heat treatment, and that this effect could in some cases be partially Bacteria of the genus Wolbachia are intracellular α-proteobacteria initially observed by Hertig and Wolbach (1924) in the ovaries of the mosquito Culex pipiens (Linnaeus). Currently, Wolbachia is considered to be widely spread amongst arthropods and it has been estimated to infect 66% of species within this phylum (Hilgenboecker, Hammerstein, Schlattmann, Telschow, & Werren, 2008). This bacterium may inhabit host ovaries and testes, but it can also inhabit somatic tissues such as the brain, muscles, the midgut or the salivary glands (Dobson et al., 1999 (Stouthamer, Breeuwer, & Hurst, 1999).
The weevil superfamily is known for harbouring many polyploid species that reproduce by thelytokous apomictic parthenogenesis in which egg cells are generated by mitosis and develop into female offspring (Saura, Lokki, & Suomalainen, 1993;Suomalainen, 1962;Suomalainen, Saura, & Lokki, 1987). It was initially proposed that parthenogenesis in weevils emerged as a result of a two-step stochastic process. At the first step, insects hybridise with a closely related species generating a new lineage with higher genetic diversity. This step is then followed by the fusion of unreduced gametes generated by meiotic errors, consequently increasing the chromosomic load of the lineage. This is thought to promote the appearance of apomictic parthenogenesis as the most effective reproduction strategy because of meiotic problems caused by the increased chromosomic load (Saura et al., 1993). Polyploid parthenogenetic forms have, for example, been seen in Otiorhynchus scaber (Linnaeus) following hybridisation with O. nodosus (Robinson) and subsequent fusion of gametes that had not undergone meiotic chromosome reduction (Stenberg & Lundmark, 2004;Stenberg, Terhivuo, Lokki, & Saura, 2000). Nonetheless, hybridisation is not the only mechanism by which asexual reproduction can arise in species of weevil, as for instance the parthenogenetic triploid vine weevil is of non-hybrid origin (Lundmark, 2010). The polyploid lineage of this species of weevil originated from fusion of unreduced gametes within the same species or autopolyploidy. Wolbachia induces asexual reproduction in a variety of arthropod species, such as the thelytokous parthenogenetic spider mite species from the genus Bryobia (Weeks & Breeuwer, 2001).
Could Wolbachia also be involved in weevil parthenogenesis?  (2004) recorded Wolbachia almost exclusively in sexual forms of O. scaber rather than in asexual forms meaning that this bacterium was not behind the origin of parthenogenesis, at least for this species of weevil. Similarly, Lachowska, Rożek, and Holecová (2008) suggested that weevil parthenogenesis originated from meiotic chromosome reduction failure in a sexual ancestor in weevils from the subfamily Entiminae. In this study, vestiges of meiosis were observed in developing eggs, indicating that apomictic parthenogenesis evolved from automictic parthenogenesis, as was previously proposed by Saura et al. (1993). Mazur et al. (2016) studied the genetic variability of the nuclear, mitochondrial and Wolbachia DNA in various populations of the parthenogenetic weevil Eusomus ovulum (Germar).
The same Wolbachia strain was shared by all weevil populations studied. However, this Wolbachia strain was present also in other parthenogenetic weevil species coinhabiting the same area. This suggests that Wolbachia has been acquired from the environment. This reproductive parasite then, seems to be benefiting from a higher transmission rate in asexual insects, rather than being the origin of parthenogenesis in this weevil species.

Research into the influence of Wolbachia on weevil reproduction
has not yet reached a final conclusion. Alternatively, other hypotheses to explain the prevalence of Wolbachia in parthenogenetic weevils have been proposed. Early studies of Wolbachia in weevils proposed that this proteobacterium causes cytoplasmic incompatibility in different populations of invasive alfalfa weevils Hypera postica (Gyllenhal) in the United States (Hsiao & Hsiao, 1985a, 1985bLeu et al., 1989).
The findings of Kotásková et al. (2018) also suggested that Wolbachia to adaptation and reproduction. In this study, Wolbachia was recorded almost twice as frequently in parthenogenetic weevils compared to sexual weevils. However, it was proposed that rather than inducing parthenogenesis, Wolbachia might have been again benefiting from infecting these weevils to increase its chances of transmission. Wolbachia could also increase its presence within a weevil population by being essential for the normal development of eggs. In the rice water weevil and the vine weevil for instance, reducing Wolbachia titre by antibiotic treatment in eggs decreased egg hatching rate (Chen et al., 2012;Son et al., 2008).
The function of Wolbachia for weevil biology and fitness to date harbours more questions than answers. Nonetheless, we expect that future research will bring a more comprehensive understanding of the influence of Wolbachia on development and fitness for this large insect group.
From an agricultural perspective, it would be interesting to test the possibility of exploiting Wolbachia induced cytoplasmic incompatibility for weevil control applying the incompatible insect technique (IIT). This technique uses males of the target pest which are artificially inoculated with a Wolbachia strain that creates a reproductive barrier with females of the target pest by cytoplasmic incompatibility. Mass release of infected (sterile) males that mate with wild females leads to decline in the pest population (Brelsfoard & Dobson, 2009). To date, the use of this technique to control populations of weevil pests has been limited because of poor prediction of the spread of Wolbachia but also because of lack of understanding of the role of this bacterium for the weevil biology. Wolbachia can be horizontally transferred within and between species as was seen between the rice weevil and the maize weevil, and between the maize weevil and its parasitoid wasp Theocolax elegans (Westwood) (Carvalho, Corrêa, de Oliveira, & Guedes, 2014). Hence, the newly introduced Wolbachia could ultimately spread to target as well as to non-target organisms within the same habitat. Although it remains unclear what role this bacterium plays in parthenogenesis in weevils, the application of IIT could inadvertently give rise to an asexual strain that may be better able to spread, as has been seen for a naturally occurring asexual strain of O. scaber (Stenberg & Lundmark, 2004). Further research is, therefore, needed to clarify if Wolbachia can be used as a safe IPM strategy, such as IIT. Diet has been shown to be a major factor in shaping the bacterial community for different weevil species in agreement with Colman et al. (2012). Merville et al. (2013) showed that four Curculio species coinhabiting oak trees had a very similar whole-body bacterial community composition despite being separate species, indicating that the food source may exert an important influence on the microbiota composition. Likewise, Berasategui et al. (2016) observed that the gut bacterial community of the pine weevil Hylobius abietis (Linnaeus) was closer in composition to bark beetles from different locations with a similar diet than to other weevil species feeding on non-conifer food sources. Changes in the diet in an experimental setup were also found to alter the bacterial community in the red palm weevil when considering the entire insect (Montagna et al., 2015). Similar results were found for the cotton boll weevil Anthonomus grandis (Boheman) when only considering the gut (Ben Guerrero et al., 2016). Cellulolytic activity has also been found in the bacterial microbiota of the Chinese white pine beetle larval gut Dendroctonus armandi (Tsai and Li) (Hu et al., 2014) and the gut of red palm weevil larvae (Muhammad et al., 2017), which is probably required to exploit their natural food sources.

|
The bacterial microbiota in some weevil species is also involved in detoxifying plant secondary metabolites produced to deter herbivorous insects. Caffeine is a naturally occurring plant secondary metabolite toxic to insects and abundant in coffee beans (Nathanson, 1984 At a fitness level, fecundity was negatively affected when weevils were deprived of bacteria following antibiotic treatment. However, survival was not significantly affected. Based on these results, Berasategui et al. (2017) suggested that rather than detoxifying terpenes, the pine weevil gut microbiota seems to be degrading terpenoids to provide the weevil host with an additional source of nutrients. Given the role that plant toxins exert in deterring herbivorous pests, the role of bacteria in degrading these toxins merits further attention. Future research should, however, aim at considering not solely microbiota analysis but also include insect fitness measurements.
Various studies have revealed the presence of a core microbiota shared by populations of the same weevil species found at separate geographic locations. This is the case for the gut microbiota of the While still a developing field, studies of bacterial communities will advance through combining knowledge of ecology, physiology, genetics and evolution (Christian, Whitaker, & Clay, 2015;Douglas & Werren, 2016). In terms of pest control, there is scant research focused on applying knowledge gleaned through microbiota studies to design pest control strategies. For instance, the characterisation of the bacterial community of various native and invasive weevil species in New Zealand identified candidate bacteria involved in resistance to the parasitic wasp Microctonus aethiopoides (Loan), used as a biological control against these species of weevil (White et al., 2015). As this is a fast-moving area of research, it is likely that discoveries in the near future will begin to find their way into IPM programmes targeting weevils that are agricultural pests around the world.

| THE ROSETTA STONE: TRANSLATING MICROBIOTA ANALYSIS INTO ECOLOGICAL INSIGHTS
In this review we have shown how weevil-bacteria associations are relevant for weevil adaptation and evolution, but also for determining the detrimental effect as agricultural pest species. However, additional research is still needed to deepen our understanding of weevilbacteria associations. For instance, the symbionts Ca. Nardonella and SPE share a common function in providing the host with tyrosine, which is required in cuticle formation (Anbutsu et al., 2017;Kuriwada et al., 2010;Vigneron et al., 2014). It is possible that the extra tyrosine produced by weevils carrying symbionts provided an evolutionary advantage by ensuring more rapid polymerisation of the exoskeleton even on nutritiously poor diets. Did these associations enable weevils to broaden their range of host species? Or, what is the mechanism used by Wolbachia to manipulate egg development in the rice and vine weevils (Chen et al., 2012;Son et al., 2008)? The technological innovations developed through studies of other groups of organisms should be applied to study bacteria in weevils.
In the "era of omics," the application of metagenomics and metatranscriptomics could provide valuable information to understand the function of the associated bacteria for the weevil host. This could allow the identification of candidate bacteria influencing the development and/or adaptation of the weevil host, which could be confirmed with functional analyses. Weevil phenotypes derived from the manipulation of these candidate bacteria, for example by selective removal or introduction, could reveal meaningful associations. Ultimately, this could allow knowledge acquired from characterising weevil microbiota to be translated into an understanding of the role of bacteria in weevil ecology and provide valuable information to design more efficient and sustainable pest control strategies.