A Review of Bacterial Interactions With Blow Flies (Diptera: Calliphoridae) of Medical, Veterinary, and Forensic Importance

Blow Fly Biology

• Development (Byrd and Butler 1996, Byrd and Allen 2001, Boatright and Tomberlin 2010, Owings et al. 2014)

• Genetics (McKenzie and Whitten 1982, Warman et al. 2000, Concha et al. 2010, Concha et al. 2011, Sze et al. 2012, Anstead et al. 2015, Kotze et al. 2015)

• Ecology (Stensmyr et al. 2002, van der Niet et al. 2011, Benbow et al. 2015)

• Role as vectors (Greenberg 1973)

• Control (Klassen and Curtis 2005)

• Behavior (Spivak et al. 1991, Boulay et al. 2015, Boulay et al. 2016, Brodie et al. 2016, Liu et al. 2016)

These data have been applied in various ways including, but not limited to, forensics (Smith 1986, Haskell and Williams 2008, Byrd and Castner 2010, Tomberlin and Benbow 2015), and maggot therapy (Sherman et al. 2000).

Blow Fly Attraction to Bacteria

Research during the early 20th century that explored microbial interactions with blow flies was conducted from a veterinary perspective due to key species causing myiasis of livestock (Tillyard and Seddon 1933, Mackerras 1936). Lucilia sp. (Cragg 1956), such as, but not limited to, L. sericata (Meigen) (Diptera: Calliphoridae) (Hobson 1935) and L. cuprina (Wiedemann) (Diptera: Calliphoridae) (Barton Browne 1958), were notorious for causing sheep strike in the United Kingdom and Australasian regions, respectively, while Cochliomyia hominivorax (Coquerel) (Diptera: Calliphoridae) was decimating cattle in the United States and southward (Bishopp et al. 1917, Hall et al. 2016). Gravid flies of C. hominivorax are attracted to larval-infested wounds and the navels of newborn livestock, where the flies oviposit (Bushland 1960). Infested wounds appear to release odors resulting in attraction and subsequent oviposition. Volatile organic compounds (VOCs) from wounds appear to play a role in attracting gravid flies and stimulating oviposition (Bromel et al. 1983, Hammack and Holt 1983, Hammack et al. 1987, Hammack 1991). During the early 1920s, researchers began speculating that bacteria were regulating blow fly attraction and colonization of sheep (Hobson 1935), but until recently, technological barriers limited the ability of researchers to explore these interactions and their relevance to the ecology and biology of this diverse group of flies.

DeVaney et al (1973) showed that bacteria-produced factors in bovine blood were highly attractive to primary screwworm (C. hominivorax). They determined cultures of several Gram-negative bacteria isolated from screwworm larval rearing media, homogenized third instars of C. hominivorax, and infested wounds on sheep produced attractive odors. These results suggested bacterial contamination of oviposition substrates was a prerequisite for attraction and oviposition by gravid flies (DeVaney et al. 1973, Eddy et al. 1975). These workers showed that the source of odors in incubated bovine blood that was attractive to gravid flies was bacteria (and compounds produced by them). Olfactometer tests showed that the blood containing Providencia (previously Proteus) rettgeri was the most attractive. Comparatively, Morganella (previously Proteus) morganii was somewhat more attractive than Proteus vulgaris and significantly more so than Proteus mirabilis. They also compared the attraction of these bacteria species alone with the four species in combination, and found that the combination was considerably more attractive than P. mirabilis, M. morganii, or P. vulgaris alone, but there was no significant difference between this combination and P. rettgeri (Eddy et al 1975). They also tested a combination of 12 Bacillus species against the combination of four species mentioned above. Results from these studies showed that the Bacillus combination trapped only 3% while the Morganella–Proteus–Providencia combination trapped 34% of the flies. Eventually, the cultures of P. rettgeri were found to be the most attractive (Eddy et al. 1975).

Screwworm flies normally deposit more egg masses when they have a suitable substrate such as horse meat, liver, bovine blood, etc. Eddy et al. (1975) compared oviposition on substrates incubated with each of the Morganella–Proteus–Providencia species and the combination of these four species. Significantly more oviposition occurred when a combination of Bacillus species was used than Proteus or a combination of Bacillus sp. and Proteus sp. (Eddy et al. 1975). According to these authors, many of the results were not consistent in their studies. For instance, peak attraction of incubated blood was sometimes after 7 d of incubation and sometimes after 14 or 21 d. Similar inconsistent results were obtained with homogenized screwworm larvae (Eddy et al. 1975). Most likely, these variations were caused by un-even presence of bacteria in the test materials used.

The above work was followed by Hammack et al. (1987) using a steam distillate of culture medium inoculated with P. rettgeri to assess the attraction of C. hominivorax adults. Results showed that females with previtellogenic ovaries and males were not attracted. Gravid mated females of 10–12 d old were most strongly attracted. These results confirmed that P. rettgeri produces an attractant for screwworm flies, and that this attractant lures females rather than males, and older females than younger ones, and more mated than virgin females (Hammack et al. 1987). Later, Hammack (1991) examined factors affecting oviposition by C. hominivorax using host-originated fluids in laboratory bioassays. She reported that fresh blood (with no attractive odor) was as attractive for oviposition as the other attractive fluids tested including fluids from screwworm-infested wounds and cultures of P. rettgeri. However, she noted that oviposition varied depending on the substrates to which the blood was applied, suggesting that an interaction exists between olfactory cues and tactile stimuli to bring about oviposition (Hammack 1991). More recent work indicates color could also play an important role in blow fly foraging behavior (Brodie et al. 2014, Brodie et al. 2015).

Chaudhury et al. (2002) continued the above research on screwworms using eight species of bacteria that were isolated from screwworm-infested animal wounds. The species of bacteria were: Enterobacter cloacae, E. sakazakii, Klebsiella oxytoca, P. mirabilis, P. vulgaris, P. rettgeri, P. stuartii, and Serratia liquefaciens. Both fertile and sterile (irradiated) male and female C. hominivorax were tested in a cage bioassay system for assessing attraction and oviposition using bovine blood inoculated with all eight species of bacteria and incubated for varying time periods. Substrates incubated for 48–72 h attracted more (>50% of flies were attracted to the resource) 7-d-old fertile females than did the substrates incubated for 24 and 96 h (<25% attraction). Significantly more fertile females were attracted to these substrates than sterile females (<20% attraction) of the same age group. Males of all tested age groups were unresponsive (<1% attraction). Oviposition tests lasting for 1 h resulted in significantly more oviposition in treated substrates compared to untreated control. Results indicate volatiles from five individual species of bacteria (K. oxytoca, P. mirabilis, P. vulgaris, P. rettgeri, and P. stuartii) were responsible for attracting more females resulting in more oviposition than volatiles from the remaining three species (E. cloacae, E. sakazakii, and S. liquefaciens) (Chaudhury et al. 2010). Volatiles from the same five species were also tested in a two-choice bioassay to study landing response and oviposition of the secondary screwworm, C. macellaria (Chaudhury et al. 2016). These tests showed that significantly more flies landed on substrates containing P. mirabilis than on substrates with other species of bacteria. Substrates treated with K. oxytoca attracted the least flies. Substrates containing bacteria incubated for 72 h attracted significantly more flies than those incubated for 24, 48, or 96h period. In 3-h duration oviposition tests, substrates with P. rettgeri attracted significantly more flies to oviposit than the other four species. The most oviposition events were recorded from substrates treated with all five species of bacteria. At least 72 h of incubation seems to be required to obtain the most active volatiles. These results suggest that C. macellaria uses similar chemical cues as C. hominivorax from bacteria volatiles as an oviposition attractant, supporting the generalizability of at least some blow fly interactions with microbes.

Interestingly, several of the key microbial players in the Cochliomyia studies also appear in the Lucilia system where myiasis manifests in association with animal feces in wool. The larvae of sheep blow flies L. cuprina and L. sericata are the primary cause of sheep myiasis in Australia, Europe, and New Zealand. The responses of these two species are similar and attraction to host and subsequent oviposition appear to involve volatile chemicals resulting from bacterial decomposition. Emmens and Murray (1982) selected four species of bacteria isolated from the fleece of Merino sheep, namely, Pseudomonas aeruginosa, Bacillus subtilis, E. cloacae, and P. mirabilis for their study. Extracts from cultures of these bacteria were incorporated into nutrient agar and exposed to females of L. cuprina in cages. Odors from the cultures of bacteria attracted L. cuprina females to oviposit but the attraction was not consistent for all the cultures. The order of the bacterial species with respect to decreasing overall response was P. mirabilis with highest number of eggs laid, followed by E. cloacae, B. subtilis and least being P. aeruginosa (Emmens and Murray 1982). In another experiment, extracts from unsterile sheep fleeces treated with these four species of bacteria singly stimulated oviposition of L. cuprina equally during 24-h period; however, with increasing length of incubation, the cultures of P. mirabilis, E. cloacae, and B. subtilis became contaminated with increasing numbers of P. aeruginosa resulting in greater responses of the flies (Emmens and Murray 1983). The maximum number of eggs was deposited over 4-d-old cultures. This response was significantly associated with the presence of P. mirabilis and E. cloacae but not B. subtilis. The relatively lower oviposition responses to P. aeruginosa in pure culture as seen in these experiments and in previous studies (Emmens and Murray 1982) was thought to be due to different culturing methods used; however, the collective results obtained from these studies indicate that the effects of attractants are enhanced when bacteria are mixed. This was also evident in studies with the screwworm flies described in the previous section (Chaudhury et al. 2010). Others have treated baits with P. mirabilis cultured in a commercial medium or a gut mucus mixture and examined their level of attraction to blow flies in New Zealand sheep pastures (Morris et al. 1998). Females of L. cuprina, Calliphora stygia (F.) (Diptera: Calliphoridae), and Chrysomya rufifacies (Macquart) (Diptera: Calliphoridae) were attracted to media culture. The gut mucus culture was significantly less effective. This is the first published record of sheep blow flies being attracted to bacterial odors in the field (Morris et al. 1998). In a study of interkingdom swarming signals, Ma et al. (2012) used P. mirabilis isolated from the salivary glands of L. sericata and identified several interkingdom signals between P. mirabilis and blow flies that influence blow fly attraction (Tomberlin et al. 2012, Liu et al. 2016).

Research during the early to mid-twentieth century, as well as more recently, determined blow flies respond to compounds associated with decomposition, such as indole (Hobson 1936, Hobson 1938, Dethier 1947). Grabbe and Turner (1973) extracted and fractionated bovine blood that had been inoculated and incubated; compounds isolated from the various fractions included phenol, p-cresol, indole, skatole, and ethanethiol. They found that a dilute aqueous mixture of indole, skatole, phenol, and p-cresol was attractive to screwworm flies in the laboratory. Similarly, Chaudhury et al. (2012) found that an aqueous slurry of media remaining after larval development to be attractive to gravid screwworm flies. The volatiles collected from these waste media using solid phase microextraction method yielded five electrophysiologically-active chemicals: dimethyl disulfide, dimethyl trisulfide, phenol, p-cresol, and indole (Chaudhury et al. 2014). A synthetic blend of these compounds was attractive to females of both primary and secondary screwworms (Chaudhury et al. 2014) as well as to female L. sericata (Chaudhury et al. 2015). Further research is necessary to identify and relate the many VOCs to the species of bacteria that produce them, and determine their specific roles as attractants, arrestants, and stimulants in host location, feeding, mating, oviposition, and other life processes of blow flies. However, it is again worth noting the similarities in attraction between Cochliomyia sp. and Lucilia sp. to similar classes of molecules. In some instances, these bacterial cues/signals are clearly important to numerous organisms (see Davis et al. 2013), indicating that the blow fly system is a useful model for dissecting insect and animal interactions with microbes.

At the time the aforementioned studies were conducted, the ecological relevance of the VOCs was not known. However, within the past decade it was determined that these volatiles are associated with bacterial activity, specifically bacterial communication and decision-making (i.e., quorum sensing (Lee et al. 2007)). More recent research is beginning to work toward bridging the nutritional ecology of bacteria with blow fly behavioral ecology, as many of these volatiles are by-products of the break-down of essential amino acids (Liu et al. 2016). Specific examples of VOCs include dimethyl disulfide, which is produced by the breakdown of methionine (Hayward et al. 1977), is a recognized by-product of vertebrate carrion decomposition (Vass et al. 2002, Paczkowski et al. 2015), and a regulator of blow fly attraction to such resources (Urech et al. 2004). Other by-products include isobutyl amine and phenylacetic acid, which are produced by the degradation of valine (Richardson 1966) and phenylalanine (Erdmann and Khalil 1986), respectively. Indole is produced when tryptophan is broken down (Sasaki-Imamura et al. 2010). In many cases, the response of blow flies to these volatiles is sex and ovarian-status specific. Males responded to dimethyl disulfide, which is produced late in the decomposition process (Paczkowski et al. 2015), a time when virgin females are attracted to the resource (Mohr and Tomberlin 2014, 2015), potentially for a protein meal necessary for oogenesis. In contrast, gravid flies respond to phenylacetic acid, which could indicate the presence of beneficial bacteria for resulting offspring (Liu et al. 2016). Interestingly, males do not respond to phenylacetic acid. However, this could be biologically relevant, as the ability to secure a mate would be highly unlikely as virgin females would be absent. Additionally, female blow flies typically mate only once. Females were highly attracted by phenylacetic acid and isobutyl amine, while males offered no response or were repelled (Liu et al. 2016). These results are significant, as they indicate the presence of beneficial bacteria (see discussion of Proteus mirabilis below), and that blow flies are potentially responding to microbial activity in relation to nutritional value of the resource they are attempting to access.

Role of Bacteria in Growth and Development

Many bacteria have a mutually-beneficial relationship with their hosts. Flies have been shown to be dependent on bacteria and their metabolic products for growth and development, as the immature stages of many fly species fail to develop in the absence of bacteria (Schmidtmann and Martin 1992, Zurek et al. 2000). In most studies, flies show the best survival rates in unsterilized or mixed bacterial environments, but some bacteria species enhance the development and survival rates of different fly species: Escherichia coli and Lactobacillus plantarum in the face fly, Musca autumnalis DeGeer (Diptera: Muscidae) (Hollis et al. 1985); Streptococcus sanguis, Staphylococcus sp., and E. coli in the house fly, Musca domestica L. (Diptera: Muscidae) (Schmidtmann and Martin 1992, Watson et al. 1993, Zurek et al. 2000); Acinetobacter sp., Flavobacterium odoratum, Citrobacter freundii and Serratia fanticola in the stable fly, Stomoxys calcitrans L. (Diptera: Muscidae) (Lysyk et al. 1999, Romero et al. 2006); and bacteria from families of Pseudomonadaceae, Corynebacteriaceae, Micrococcaceae, and Bacillaceae in the horn fly, Haematobia irritans L. (Diptera: Muscidae) (Perotti et al. 2001).

Perhaps the increased survival rates seen with mixed bacterial environments are because different bacteria are more advantageous at different stages of growth dependent on the metabolites they produce and the nutritional needs of the arthropod at that stage. Zurek et al. (2000) collected third-instars of house fly larvae from two common sources found at animal production facilities. They found that Bacillus coagulans, Bacillus sp., Clavibacter michiganese, Corynebacterium aquaticum, Lactococcus garviae, Microbacterium esteraromaticum, Microbacterium lacticum, Microbacterium liquefaciens, Ochrobacter anthropic, Sphingobacterium spiritivorum, Sphinomonas capsulataa, Staphylococcus epidermidis, Staphylococcus lentus, Streptococcus sanguis, Xanthobacter flavus, and Yersinia pseudotuberculosis, all supported larval growth to some extent to the pupal stage, whereas Corynebacterium seminale, Gordona amarae, Microbacterium barkei, Morganella morganii, Providencia rettgeri, Providencia stuartii, and Serratia marcescens did not. Streptococcus sanguis and Sphingobacterium spiritivorum supported larval development through eclosion as adult flies, whereas Bacillus sp. and Staphylococcus epidermidis did not.

As suggested by optimal foraging theory, an inferior competitor (potentially blow fly larvae in the case of carrion) could avoid such challenges simply by eliminating the competitor (potentially bacteria). Greenberg (1968) previously determined a five log reduction of Gram-positive and Gram-negative bacteria exposed to conditions experienced in the midgut of larval Calliphora vicina Robineau-Desvoidy (Diptera: Calliphoridae). Mumcuoglu et al. (2001) observed similar interactions between L. sericata and E. coli. They determined only 18% of E. coli ingested by L. sericata larvae was present in the hindgut. Reducing bacterial counts could free up nutrients for larval use as well as decrease the likelihood of becoming infected and killed. In fact, larvae of C. macellaria developed significantly faster (7.8 and 7.7 d) and had greater survivorship (88% and 94%) on sterile blood agar or a modified Harris rearing media, respectively, than those provided the same diets inoculated with bacteria, such as E. coli O157:H7 (9.1 d and 35.7%), Providencia sp. (9.7 d and 17.3%), E. faecalis (8.2 d and 75.5%), Ochrobactrum sp. (8.4 d and 81%), or a mixture (9.3 d and 27.5%; Ahmad et al. 2006).

While some bacteria are harmful to fly larvae, one in particular, Proteus mirabilis, has been considered beneficial. Blow fly colonization of vertebrate carrion results in increased pH, which is hypothesized to reduce competing or pathogenic bacteria, due to P. mirabilis activity (Barnes et al. 2010). This bacterium is able to persist in the digestive tract of L. sericata through immature development to the adult stage (Wei et al. 2014a). Additionally, the bacterial mortality recorded by Greenberg (1968), which was previously discussed, was suspected to be due to pH-specific activities of by-products, or "mirabilicides", produced by P. mirabilis (Barnes et al. 2010). Furthermore, what is most interesting is that beetle species that colonize carrion after blow flies, decrease the pH and their excretion/secretions (ES) kill P. mirabilis (Barnes et al. 2010). One could speculate that by doing so, such modifications would reduce beetle competition with fly larvae utilizing the same resource. These observations paint a picture of microbial warfare among insects competing for a resource, with different microbial tolerances as a mechanism by which different organisms can come to dominate or be excluded from ephemeral carrion resources (Janzen 1977, Burkepile et al. 2006). In this instance, P. mirabilis and L. sericata seem to be in alliance against other insects (like carrion beetles) and bacteria (those killed by "mirabilicides").

Since this initial work, a tremendous amount of research exploring the antimicrobial properties of the ES produced by bacteria associated with blow fly larvae and the larvae themselves has been conducted (Cazander et al. 2009a, Cazander et al. 2009b, Cazander et al. 2010, Barnes and Gennard 2011, Cazander et al. 2012, Barnes and Gennard 2013, Cazander et al. 2013). Two of these antimicrobials, phenyl acid acid (noted above) and phenylacetaldehyde, were identified from P. mirabilis isolated from larvae of C. hominivorax (Erdmann and Khalil 1986). Genetic and microscopic evidence indicating that P. mirabilis can reside in L. sericata salivary glands (Singh et al. 2015, Blenkiron et al. 2015), which are relatively microbe free, supports the concept that these two species coevolved to kill other microbes. However, it should be noted that while many bacteria are suppressed by ES, some were able to survive pupation of C. macellaria, indicating some level of "resistance" to being digested by the larvae (Ahmad et al. 2006). This ability to persist in the alimentary canal of the insect through development is cause for concern as these flies could serve as mechanical vectors for these pathogens, as well as creating an environment in which antibiotic resistance is developed and amplified in some circumstances (Wei et al. 2014a,b), but also provides further support for the concept that certain flies and microbes act as mutualists to compete against others.

Molecular Techniques for Evaluating Bacterial–Blow Fly Interactions

Like all other organisms, development and survival of blow flies also depends on their associated microbiome. Hence, to better understand how bacteria associated with blow flies regulate physiology and behavior of blow flies, the first step is to understand the microbiome associated with the fly species that colonize carrions. However, our current understanding of bacteria associated with blow fly species is limited (Thompson et al. 2013, Wei et al. 2014a, Singh et al. 2015, Pechal and Benbow 2016). This is mainly because culture-based methods used in the past were not able to discover the majority of bacteria associated with insects (Thompson et al. 2013). New culture-independent sequencing technologies (i.e. next-generation sequencing technologies) are capable of performing comprehensive surveys of bacteria (both rare and abundant) in a cost effective and timely manner. To do this, the first step is the selection of DNA extraction methods that give relatively unbiased estimates of the presence of both Gram-positive and Gram-negative bacteria. Bacterial DNA from different life stages of blow flies can be extracted either by using organic extraction methods (Zheng et al. 2013, Singh et al. 2015) or by using commercial kits (Iancu et al. 2016).

Extracted DNA can be used either for amplification and sequencing of targeted marker loci (e.g., 16S ribosomal DNA (16s rDNA) for prokaryotes, 18S ribosomal DNA (18S rDNA) for eukaryotes) or for direct whole genome shotgun sequencing (i.e. metagenomic sequencing). In marker gene sequencing approaches, different variable regions of targeted loci are amplified using barcoded universal primers and then barcoded-amplified products are pooled and sequenced using next-generation sequencing platform of choice (e.g., Hiseq/Miseq (Illumina Inc. USA), Ion PGM System (Thermo Fisher Scientific, USA)). On the other hand, in the metagenomic sequencing approach, DNA from the whole community is sheared and then directly sequenced using next-generation sequencing platform of choice (e.g., Hiseq/Miseq (Illumina Inc. USA)). Both approaches have some advantages. Marker gene-based sequencing approaches are comparatively cheap, provide information on both rare and abundant microbial taxa, and are easy to analyze (as user friendly bioinformatics pipelines are freely available (e.g., QIIME (Caporaso et al. 2010), mothur (Schloss et al. 2009), RDPipeline (Cole et al. 2014)). Metagenomic sequencing approaches provide information not only on microbial community structure but also on microbial function, and avoid PCR-based biases associated with marker gene approaches (Lee et al. 2012). Recovery of rare taxa requires high sequencing depth in metagenomic sequencing approaches, which ultimately increases sequencing cost, and complicates already complicated metagenomic data analysis pipelines (Sharpton 2014). One method for balancing the advantages of both methods is to use single-molecule information to extrapolate the metagenomic gene content by using prior knowledge of sequenced bacterial genomes (Langille et al. 2013). Given the importance of blow fly interactions with microbes, we predict many more studies of microbial communities in blow flies in the coming years.

Bacteria Associated With Blow Flies

Blow flies and other filth flies (e.g., Sarcophagidae and Muscidae) use decaying organic matter for nutrition and larval habitat and thus often interact with and transmit pathogens of human disease. The US Federal Drug Administration identified 21 filth fly species as significant threats to human health (Olsen 1998). These flies are synanthropic and thus have preferential associations with human populations, increasing the risk for pathogenic bacterial transfer (Polvony 1971, Greenberg 1973). Many bacteria, pathogens included, can exist in biofilms (i.e., complex structures composed of components that include proteins, glycoproteins, DNA, and carbohydrates that are constructed by microbes as a habitat (Costerton 2007)) in facilities such as hospitals that flies can easily interact with just by contact, and can lead to large numbers of human bacterial infections (Barraud et al. 2009). Additionally, these flies can harbor and transfer antibiotic resistant bacteria, some with multidrug resistance (Graham et al. 2009). Despite the importance of the blow fly to carrion decomposition and to pathogen dispersal, very little has been done to investigate their microbiomes. In addition to studies noted above, Caballero et al. (1996) documented by culture technique the bacterial content of C. hominivorax during sheep myiasis. The most prominent genera were Escherichia, Proteus, Providencia, Staphylococcus, and Streptococcus. Singh et al. (2015) compared the bacterial community structure associated with different stages of L. sericata and L. cuprina from colony raised flies reared on beef liver, using metgenomic analyses. They determined that the majority of bacteria came from the phyla Proteobacteria, Firmicutes and Bacteriodetes. Proteobacteria is a Gram-negative phylum that contains many pathogens, such as Escherichia coli, Wolbachia, Bordetella, and Salmonella. Bacteria from this phylum encompass a wide variety of metabolic capabilities, including chemoautotrophs that utilize hydrogen gas, ammonia, and methane during the decomposition of organic matter. The Firmicutes are a group of Gram-positive bacteria with low –G + C content. Many are very resistant to environmental desiccation and often produce endospores to survive extreme conditions. This phylum contains pathogens such as Clostridia and Bacilli. Bacteriodetes are Gram-negative rod shaped bacteria that are also widely distributed in the environment. While they can be opportunistic pathogens but most are not pathogenic.

Most bacterial genera identified in the Singh study were shared amongst the two fly species (L. sericata and L. cuprina) investigated (Singh et al. 2015). Providencia, Ignatzchineria, and Lactobacillus constituted several of the most dominant populations shared between the two species. Providencia, already noted above as a key component of the blow fly microbiome, is a member of the family Enterobacteriaceae. It can be an opportunistic pathogen living in soil, water, and sewage, causing diarrhea with fever and tachycardia, and leading to low blood pressure. The other Proteobacterium, Ignatzchineria was first isolated as the dominant species in the anterior portion of the digestive tract in larval Wohlfahrtia magnifica Schiner (Diptera: Sarcophagidae) flesh flies (Toth et al. 2006, Gupta et al. 2011). This species has been associated with myiasis, but was not generally associated with severe human disease until reports in which it was isolated in cases of bacteremia and urinary tract infection (Maurin et al. 2007, Roudiere et al. 2007, Barker et al. 2014, Le Brun et al. 2015). Neither the mechanism nor the epidemiology of an Ignatzchineria infection has been defined and the cases were all associated with a corresponding fly larvae infestation where Ignatzchineria was isolated concurrent with other bacteria, such as Enterococcus and Providencia (Le Brun et al. 2015). Lactobacillus is a Firmicute that is a member of the lactic acid bacterial group. These bacteria constitute a major portion of the microbiota of the gut and other body sites. They produce alcohol and lactic acid from sugars which lowers pH and that controlled fermentation is exploited by industry in the production of items such as yogurts, beer, sourdough bread, cider, other fermented foods and animal feeds. It is worth noting that this genus was also found in high relative abundance in the remarkably microbe-depauperate L. sericata salivary gland, along with Proteus (Singh et al. 2015). In combination, the metabolic by-products of these two groups would be expected to neutralize one another, which may be important to salivary gland function. Lactobacillus can also inhibit the activity of other microorganisms by producing bacteriocins with antimicrobial and antifungal properties (Inglin et al. 2015).

In contrast to their role in transmission of bacteria, blow flies may also actively participate in the eradication of pathogens. Early on, the ES of blow flies were determined to contain bactericidal substances (Greenberg et al. 1970). Lucilia sericata larvae are used to debride wounds and contain substances with antimicrobial properties (Blueman and Bousfield 2012), such as a DNAse capable of degrading genomic bacterial DNA (Brown et al. 2012). Flies also produce insect defensins (i.e. lucifensin; Cerovsky et al. 2010) with antimicrobial properties, and some bacteria that flies carry are capable of killing other bacteria, such as Proteus sp., which produce mirabilicides (i.e. phenylacetic acid and phenylacetaldehyde; Erdmann and Khalil 1986). Additionally, larvae express lysozymes in their midgut, which kill bacteria during their passage through the midgut (Valachova et al. 2014). Some of these compounds appear to have selective capabilities; the ES of L. sericata was more effective against Gram-positive bacteria, like Staphylococcus aureus, than against Gram-negative bacteria, like Proteus sp. and Pseudomonas sp. (Jaklic et al. 2008). Many questions remain about how these compounds are deployed. For example, do flies use ES compounds to selectively control the microbial community structure in their environment to their own advantage, such as to deter predators and competitors or attract mates?

Horizontal and Vertical Transmission

Microbes have limited ability for dispersal by self-propelled motility (e.g. flagella, axial filament, or gliding); and instead rely on other means for spatial and temporal transmission. Horizontal transmission of microbes is the passage of the symbiont from one host to another unrelated host, and vertical transmission of microbes is the passage of the symbiont from parents to offspring (Fine 1975). A well-studied alpha-proteobacteria that is vertically transmitted among a variety of arthropods, including some blow flies, is Wolbachia (Stouthamer et al. 1999, Baudry et al. 2003, Mingchay et al. 2014). Studies suggest that a benefit of vertically-transmitted infection is the prevention of more virulent infection, by horizontally transmitted organisms (Lively et al. 2005). Wolbachia also has the capability to cause reproductive isolation between infected and uninfected flies, although no such evidence for Wolbachia-induced reproductive isolation has been documented in blow flies (Baudry et al. 2003, Mingchay et al. 2014).

Horizontal transmission of pathogenic bacteria by the exterior surfaces and mouthparts, of flies, along with internal transfer via feces and vomit has long been known to occur in many fly species (Förster et al. 2007, 2009; Pava-Ripoll et al. 2012). Blaak et al. (2014) demonstrated the horizontal acquisition of Escherichia coli with extended-spectrum antibiotic resistance to most beta-lactam antibiotics by blow flies at a poultry facility. Horizontal transmission of various bacteria in blow flies was recently assessed in a metagenomic study of L. sericata and L. cuprina, that characterized and compared the bacteria on adult flies and the fresh liver before they oviposited on it, followed by the resulting larvae and the aged liver after larval development (Singh et al. 2015). Of the 138 genera found, adult flies had 23 unique genera, larvae had 6, fresh liver had 14, and aged liver had 40. There were 15 genera shared by all, of which Proteus, Enterococcus, and Lactobacillus (all described above) were the most dominant. Thirteen bacterial genera were found on the aged liver that were not present on the fresh liver, but were present on the adult flies; thus representing horizontal transfer from the adult flies to the liver; the dominant bacteria of these was Staphylococcus. Additionally, four bacteria were found associated with the larva that were present on the liver, but not on adult flies, representing horizontal transfer from the liver to the larvae; the dominant bacteria of these were Vagococcus and Lactococcus. No bacteria were found shared between only adults and larvae and not liver, so while vertical transmission could have occurred it is not possible to isolate that from horizontal transmission in this experimental design. Interestingly there were many bacteria (88 genera) present on the liver that were not transferred to the larvae. Perhaps larval ES or environmental conditions (i.e. temperatures within the larval mass) prevented the transfer of more genera to the larvae.

Dispersal strategies for bacteria include many routes, such as the avoidance of elimination during insect development discussed above, the most prominent being physical interaction with a contaminated host; the exchange through contact with airborne droplets or light weight particles containing bacteria that can remain airborne for long periods, such as fungal spores; and fecal to oral, wound or mucosal surface transmission. There is also indirect contact with an environmental substance or surface, or a host, such as a fly that retains the microbe after it was acquired by contact with a contaminated host or other surface (Barro et al. 2006). Ebert (2013) reviews the many symbionts, microbes among them, that utilize multiple methods of transmission (mixed-mode transmission), not just exclusively one (single-mode transmission). Combinations of these strategies can enhance the likelihood of persistence depending on the ecological conditions and plays a role in increased genetic drift and the evolution of virulence and genome architecture (Moran and Baumann 2000, Ebert 2013). Symbionts can persist within a population under a wider range of ecological conditions if they have mixed versus single modes of transmission (Lipsitch et al. 1995a, Lipsitch et al. 1995b).

The interdependence between insects and bacteria is more amalgamated than simply for growth augmentation. Blow fly oviposition is also induced by metabolic products of bacteria, some of which were noted above (Hammack et al. 1987, Chaudhury et al. 2010). In an interesting study by Pechal (2012), the microbiome associated with decomposing swine carcasses that were accessible to flies or in which the flies were excluded was characterized over five days. They demonstrated significant changes in the bacterial community during decomposition between insect access and exclusion carcasses. When flies were not present, Proteobacteria was a dominant taxon throughout the 5-d sampling period, whereas relative abundance of Firmicutes decreased as decomposition progressed. However, when flies were present, the opposite occurred, as Proteobacteria decreased over time and Firmicutes became the dominant taxon by the fifth day of decomposition. At the genus level, Psychrobacter and Moraxella were dominant for both exclusion and access carcasses. But other bacterial succession patterns differed as decomposition progressed. Without insect access, Aeromonas and Shewanellaceae were detected only on the first day, Peptostretpococcus was detected only on the fifth day and Proteus transitioned to the dominant taxon by the third and fifth day. When insects were present, Providencia was dominant on the first and third day, but Bacillales was dominant by the fifth day. Proteus and Corynebacterium were present on the third day of decomposition, but by the fifth day, Psychrobacillus and Ignatzschineria were the dominant taxa and Clostridium sensu stricto (10%) was also detected. Additionally, Pechal and Benbow (2016) investigated the internal microbiome between Calliphora terraenovae Macquart (Diptera: Calliphoridae) utilizing streams where decomposing salmon carrion was a readily available resource or streams without carcasses. Bacteroidetes, Firmicutes, and Proteobacteria were the dominant phyla found increased on the C. terraenovae larval samples. They shared microbial communities with their food resource (salmon carrion); for example, Firmicutes, dominated by Vagococcus, Clostridium, and the candidate family Tissierellaceae increased due to this interaction.

We know from previous studies that both Proteus and Providencia are important players in the attraction of blow flies (Hammack et al. 1987, Chaudhury et al. 2010, Ma et al. 2012, Tomberlin et al. 2012), but what we do not know is the extent to which these bacteria are transported by the flies and seed the carcass as a result of the fly interaction or the interkingdom communication with the bacteria already associated with carcasses. While the microbiome of the fly was not analyzed in the Pechal (2012) study, it is interesting to note that an insect-associated genus, Ignatzschineria, was not a dominant member of the bacterial community on exclusion carcasses, but was in the later stages of the accessible carcasses, eluding to the likelihood that the insects played a role in the presence of this genus on the carcass. On the other hand, Proteus was dominant in the later stages of the exclusion carcass, indicating the bacteria were already present on the carcass or came from the environment and that the blow fly species that secondarily colonize the remains may be responding to its olfactory cues to select oviposition sites on the resource. This observation is interesting, given the non-intuitive response of Proteus to the presence of insect colonization, considering the empirical evidence to suggest a mutualism between Proteus and some blow flies. This result would suggest that such mutualism may be more complex than the simple laboratory experiments would suggest. Alternatively, it may be the case that other members of the carrion insect community are capable of eliminating Proteus, even in the presence of their hypothesized mutualists.

Microbial Genomes

While bacterial community studies are invaluable for providing information regarding common and important members of blow fly associated communities of bacteria, there can be additional valuable information gleaned from the whole genomes of specific species and strains isolated from blow flies and blow fly resources. One area where this endeavor is important relates to bacterial community studies. For example, algorithms can be used to predict community function from the community structure information provided by the 16S rDNA gene (Langille et al. 2013). This is done by extrapolating from the presence of specific taxa (via 16S sequences) to the gene compositions of entire genomes and communities using microbial genome databases to provide a prediction of metabolic potential estimated from the types and predicted functions of genes in the sequenced genomes associated with 16S sequences. This capability enables researchers to extend their interpretations of 16S data to the expected functionality of the community, facilitating the ability to ask questions like, does the community have the ability to catabolize a specific amino acid or produce a specific antibiotic? Such capability with blow fly bacteria communities may allow us to expand our understanding of the specific biological processes involved in fly–microbe interactions. However, the quality of a result from any application that relies on a database will depend greatly on the content of the database. Thus, it is imperative that genomes of additional blow fly-derived bacteria be sequenced and deposited into relevant databases to enable the best results from functional predictions. This task is most important when there are blow fly associated bacteria for which there is no known genome sequence, such as the Ignatzshineria commonly found with blow flies. Genome sequences for strains from this genus would enhance our ability to understand mechanisms by which these bacteria interact with flies.

Even in instances where a bacterial genome has been sequenced for a particular taxon, it may still be important to produce additional genome sequences for blow fly-derived strains. For example, the genus Proteus, and specifically P. mirabilis, has been noted numerous times in this review as important bacteria that interact with blow flies. Proteus sp. are also important to medicine and ecology (Drzewiecka 2016), and numerous strains of the genus, including P. mirabilis, have been sequenced (Pearson et al. 2008, Sullivan et al. 2013, Mac Aogáin et al. 2016). Recently, an L. sericata-derived strain of P. mirabilis was sequenced and assembled (Yuan et al. 2016), revealing some interesting features of that strain. First, there are two high quality reference genomes for the species and the blow fly derived strain is much more similar to the BB2000 strain used to study self-recognition during swarming of P. mirabilis (Gibbs et al. 2008, Cardarelli et al. 2015) than is was to the original reference genome obtained from a clinical strain of the species. Second, it is clear that the fly-derived strain contains lineage specific insertions and deletions. It remains to be seen if there are genes relevant to fly interactions in those regions of the genome, but such a hypothesis will be important to test. It is clear that P. mirabilis and other bacteria often have a core genome (composed of sequences shared by all members) as well as numerous auxiliary components (composed of sequences not shared by all members of the species and sometimes unique to one strain; Collins and Higgs 2012). Evolutionary comparisons, such as those described below, of fly-derived and non-fly-derived bacteria will help to determine if core genome components, auxiliary components, or both components of bacterial genomes are important to fly–bacterial interactions.

Microbe Evolution in Hosts

Evolutionary analyses of bacteria are a promising avenue for elucidating fly–microbe interactions. The approaches can take two different forms: studies of naturally-occurring microbes and experimental evolution of microbial populations in the laboratory. Both approaches have been under-utilized to understand fly–microbe interactions in general, and within blow flies in particular. We highlight areas where evolutionary approaches can be applied to improve our understanding of the effects of bacteria on blow fly behavior and physiology.

Comparisons of host-associated bacteria across blow fly species have tremendous potential to resolve cryptic variation and similarities in life history strategies across species. While the larval habitats of blow flies are easily characterized, the breadth of adult diet choice and greater adult mobility can expose the flies to bacteria whose origin is more difficult to determine. For example, a female bias in visits to oviposition sites could result in sexually dimorphic microbial communities within blow fly adults (Mohr and Tomberlin 2014), which may translate into variation in the bacterial communities of their offspring and in traits dependent on bacteria. Adult females will also use a variety of nutrient resources for oogenesis, ranging from pollen (Brodie et al. 2015) to carrion, which will differ considerably in their microbe contents (Fridman et al. 2012, Metcalf et al. 2013, Shade et al. 2013, Aleklett et al. 2014, Pechal et al. 2014). Differential resource use may also enable fly-associated bacteria to exchange genes with bacteria from a diverse pool of community members (Akhtar et al. 2009, Crippen and Poole 2009), which may alter bacterial phenotypes compared to more commonly studied "standard" strains of the same bacterial species. Thus, E. coli from a fly may not contain the same gene content as common laboratory strains that are well studied but raised continuously in monoculture, and cannot be expected to demonstrate the same traits as their "domesticated" conspecifics. This logic is also at play in P. mirabilis strains described above, where the clinical strain that was originally sequenced was not considered a sufficient reference genome to study self-recognition, as the strain used to study that system exhibited a different gene content compared to the clinical strain (Sullivan et al. 2013). Community analyses from a broad sampling of blow flies and their larval substrates could therefore identify specific aspects of fly behavior, diet, and life history that affect the microbial communities found across species. These approaches could also produce candidate bacterial species that affect blow fly behavior and physiology in a sex-, diet-, or life-stage-specific manner.

Culture-dependent and DNA sequencing approaches have revealed which bacteria taxa are associated with blow fly larvae and adults (see above), and genome sequencing has identified inter-strain differences between blow fly associated and other P. mirabilis strains (Yuan et al. 2016). However, we do not know which specific genetic changes in Proteus and other bacteria genomes are responsible for successful colonization of blow fly salivary glands and digestive tract. Laboratory experiments in which microbes evolve within hosts are especially revealing of the aspects of host physiology that have the greatest effects on bacterial survival (Brockhurst and Koskella 2013, Hoang et al. 2016). For example, E. coli populations in the guts of laboratory mice evolve higher growth rates and reduced motility (Giraud et al. 2008, Lee et al. 2010). Evolution experiments involving bacteria within blow fly larvae and adults therefore have tremendous potential to reveal specific bacterial traits that are adaptive to the fly digestive tract or salivary glands. Sequencing of bacteria strains that evolved within blow flies would allow for the identification of specific mutations that promote colonization, and the targeting of the products of those genes could reduce the ability of pathogenic bacteria to be vectored by blow flies. In addition, host-specific adaptation should result in an evolved bacteria strain that affects fly physiology differently than the ancestral strain. Aside from immune-related genes (see below), little is known about specific fly traits that regulate bacterial proliferation. It would thus be informative of fly traits involved in bacterial interactions to compare the phenotypes of blow flies fed ancestral and derived strains of host-evolved bacteria.

Comparative Genomics of Blow Flies to Study Ecological Interactions Between Bacteria and Hosts

Genome sequences of blow flies will also improve our understanding of interactions with bacteria and the specific aspects of fly physiology responsible for regulating bacteria. For example, comparisons of gene content and sequence evolution in the genomes of related fly species can reveal long-term fly–bacteria interactions. The fly innate immune system includes effector proteins that suppress bacteria growth or kill microbes, such as antimicrobial peptides (AMPs). These AMPs are promising candidates for treatment of bacteria infections, especially those that are resistant to existing antibiotics (Bexfield et al. 2004, Cerovsky et al. 2010, Poppel et al. 2015). The genes encoding effector proteins vary in copy number across Drosophila sp. and they have expanded along the house fly lineage (Sackton et al. 2007, Scott et al. 2014, Sackton et al. 2016). In addition, genes encoding immune system proteins involved in recognizing bacteria are more likely to evolve under positive Darwinian selection in Drosophila and have also experienced a copy number expansion in house flies (Sackton et al. 2007, Scott et al. 2014, Sackton et al. 2016). Both of these patterns suggest that the fly immune system evolves in an arms race to suppress bacterial infections, rapidly changing to keep up with a diverse community of rapidly evolving bacteria. These observations make blow flies an ideal system to study fly interactions with bacteria, given their associations with numerous bacteria taxa.

The growing number of genome sequences of blow flies and their close relatives will allow for similar comparative analyses of gene content and sequence evolution to elucidate specific aspects of blow fly physiology involved in host–microbe interactions. The expansion of immunity-related genes in the house fly genome suggests that house flies live in a more pathogen-rich environment than Drosophila, and require a more diverse immune repertoire to suppress infections (Scott et al. 2014). The selection pressures on the blow fly immune system are expected to differ depending on exposure to microbes, which will depend on larval habitat and adult behavior. Comparisons of immune-related genes across blow fly species and between blow flies and other filth flies could therefore reveal differences in the diversity and abundance of microbial pathogens and other bacteria encountered across the fly life cycle. Additionally, these analyses have the potential to identify cryptic differences in larval and adult habitats between species.

Functional Genomics as a Tool to Understand Blow Fly–Bacteria Interactions

Sequenced blow fly genomes also open up new possibilities for experiments that interrogate the effect of bacteria exposure on fly physiology. Numerous studies on blow flies and their relatives have demonstrated that the expression of genes encoding AMPs is up-regulated upon exposure to bacteria, but sequenced genomes now allow for the identification of novel recognition and effector components of the immune system and other genes involved in the physiological response to bacteria exposure through genome-wide analyses of gene expression (Joyner et al. 2013, Nayduch et al. 2013, Poppel et al. 2015, Sackton et al. 2016). For example, 47 expressed AMPs were identified in a high throughput RNA sequencing experiment in which L. sericata larvae were infected with P. aeruginosa and S. aureus (Poppel et al. 2015). The activities of 23 of these AMPs were tested against a panel of Gram-positive and -negative bacteria, and their effects varied across AMP–bacteria combinations and depended on which other AMPs were present in the treatment (Poppel et al. 2015). Additional experiments could determine which AMPs (or combinations of AMPs) are best suited to be used as antimicrobial agents for wound therapy or provide targets for pest control measures.

The power of genomic approaches to identify previously uncharacterized components of the blow fly response to bacteria can be increased by comparing multiple species or different strains within a species (Sackton and Clark 2009, Sackton et al. 2010), assaying gene expression in flies carrying mutations in genes involved in host–bacteria interactions (Broderick et al. 2014), or considering different diets, life stages, and sexes (Jacobs et al. 2016). These approaches will only become more feasible as the genomes of additional blow flies are sequenced and the tools for targeted mutagenesis become more tractable in nonmodel organisms (Sander and Joung 2014).

Our understanding of the bacterial perspective of fly–microbe interactions will also benefit from improvements in functional genomics. For example, the effects of specific bacterial gene products on blow fly physiology could be assessed through exposing flies to mutant bacteria or the metabolic products that differ between mutant and wild type bacteria (Ma et al. 2012, Tomberlin et al. 2012, Liu et al. 2016), and measuring blow fly gene expression genome-wide in various tissues. In addition, the molecular nature of phenotypic plasticity in bacteria strains could be interrogated through genome-wide expression analyses of bacteria grown in different conditions (including exposure to flies and even harvested from within flies) to identify novel genes whose expression is associated with variable response to environmental cues. These differentially expressed genes could be candidates for attracting or repelling blow flies. Likewise, expression of bacterial genes when fly-derived species are competing with those that are 1) the same species, but not fly derived, 2) controlled in maggot therapy, 3) differentially competitive within flies (like the Proteus – Salmonella competition in blow flies studied by Greenberg (1965) and Greenberg et al. (1970)), or 4) in blow fly habitats but are not commonly associated with blow flies, would be useful for understanding the ecology and molecular mechanisms underpinning microbial interactions in the wild.

For example, in mammals we know that compositional and structural shifts of gut microbes and their associated metabolites can affect behaviors, such as neurodevelopmental disorders (Hsiao et al. 2013), and that the nasal cavity, also a rich source of indigenous microbiota, can influence health and disease (Yan et al. 2013). But how much influence do the nasal and gut microbe community structures have on the olfactory capabilities and resulting behavior in animals? Francois et al. (2016) evaluated the development of olfactory epithelium in the nares and nasal cavities of germfree mice. In normal pathogen-free mice, although individual variability exists, they determined that the olfactory epithelium was primarily occupied by Bacteriodetes (15–60%; dominated by Bacteriodaceae) and Firmicutes (30–70%; dominated by Enterococaceae, Lachnospiraceae, and Ruminococaceae)) and less so by Proteobacteria (5–25%; dominated by Enterobacteriaceae) and Actinobacteria (<10% dominated by Bifidobacteriaceae). Interestingly, the absence of these gut microbes resulted in a thinner layer of olfactory cilia, a reduced epithelial cell turnover rate and a reduced expression of many genes associated with the olfactory signal transduction pathway. Thus, the microbial community structure may indeed influence the odor sensory capabilities of its host. Flies use the olfactory sensilla and receptors on their antenna and maxillary palps to smell odors and locate resources (Shanbhag et al. 1999, Wasserman and Itagaki 2003). We know that the area around mouthparts can be heavily contaminated with bacteria (Barro et al. 2006). Could the bacteria present around the antennae and fly mouthparts, as well as their gut microbiome, affect development of odor perception organs in the fly? Could raising flies under certain lab conditions or on specific resources, which affect the community structure of the gut microbiome, change adult olfactory capabilities and ultimately the behavior of the fly being investigated? What impact would this have on the olfactory capabilities of wild flies that exist in different ecoregions and are exposed to a differential range of microbes? How much are the host’s olfactory capabilities and behavior related to their gut microbial community structure? These are open questions that will be important to study.

Proteomic Responses in Hosts and Microbes

Transcriptomic analyses have potential to identify blow fly genes involved in clearance of bacteria (e.g., AMPs) because those genes are transcriptionally up-regulated upon exposure to microbes. However, genes encoding other components of the fly immune system do not experience changes in expression upon infection, and instead the gene products react to bacterial exposure through modifications to the proteins themselves. For example, information about bacterial infection in flies is communicated through the Toll, IMD, JAK/STAT, and JNK pathways via the cleavage, phosphorylation, and degradation of signaling proteins, which ultimately leads to the up-regulation of effector genes that directly respond to the bacteria (Lemaitre and Hoffmann 2007, Buchon et al. 2014). Proteomic approaches are therefore necessary to characterize the physiological response of blow flies to bacteria mediated by the immune system. Some of these proteomic techniques, including the combination of two dimensional electrophoresis with mass spectrometry (Marouga et al. 2005, Samyn et al. 2006) and iTraq (Dong et al. 2007, Evans et al. 2007), are greatly facilitated by a sequenced genome in order to determine the potential proteins present in the sample. The growing number of blow fly and bacteria genomes in this system will improve our ability to apply proteomic technologies to understand the blow fly response to bacteria as proteomic analyses are enhanced by prior genomic knowledge (Samyn et al. 2006). Applying this approach to different blow fly species exposed to various bacterial taxa and strains will allow for the identification of fly and bacteria species-specific interactions, which are candidate evolved differences across blow flies in response to the variation in substrates in which they develop and diets they consume (see above).

A similar logic applies to the microbial side of the equation, as protein studies have revealed critical aspects of several host–microbe interactions (Shao et al. 2002, Zhang et al. 2005, Chisholm et al. 2006, Lee et al. 2008, Schmidt and Völker 2011), including the role of pathogen proteins in evading immune detection, virulence, communication, and survival. As an example Yersinia and Pseudomonas virulence is regulated in part by proteases (Shao et al. 2002)—enzymes that cleave other proteins. Thus, while mRNA studies can provide valuable information regarding interactions between bacteria and hosts, there are also levels of regulation independent of RNA that are equally valuable. As the genomes of blow flies and their microbial associates are produced, these applications will become more feasible—opening up new areas of inquiry.

Metabolic Studies

Metabolic consequences of insect–microbe interactions are important for understanding the mechanisms underpinning them. As an example, this review has demonstrated the importance of indole to a number of blow fly–microbe interactions. This amino acid-derived molecule is also important to bacterial signaling and mammalian regulation of inflammation. In this era where we are learning through studies of bacteria communities that bacteria impact and are associated with a wide array of eukaryotic phenotypes, it will be necessary to step beyond bacterial community structures to the mechanisms underpinning the importance of certain community members on eukaryotic traits. Thus, it will be important to determine if indole is the exception or the rule when attempting to understand interactions among bacteria and eukaryotes (Lee et al. 2015). It will be important to answer the following (and related) questions: Are molecules like indole a common means used by many bacteria to regulate interactions with eukaryotes or is indole a special molecule with properties that are distinct from other metabolites important to the system? Are there many or few metabolites important to fly–microbe interactions? Are regulatory metabolites important to fly interactions distinct to certain clades of bacteria or are they common across taxa? How do flies alter the physiology of bacteria (and vice versa)? The answers to some of these questions can be elucidated, in part, through genomic studies of fly-derived bacteria described above. However, these studies will need to be paired with in vivo experimental manipulations to fully appreciate the greater picture of their interactions.

Repeating previous experiments, augmented with modern chemical and biological techniques, to enhance our metabolic understanding of blow fly interactions with bacteria would be helpful. For example, in the 1960s and 1970s Bernard Greenberg and his colleagues studied the competitive dynamics of bacteria associated with flies (Greenberg 1973). One particular focus was the competition between Proteus and Salmonella within filth flies (Greenberg 1965, 1968, 1969, Greenberg and Klowden 1972). In this system, the impact of specific bacteria on the health and success of the fly (and vice versa) becomes important. Therefore, they conducted gnotobiotic experiments on M. domestica and L. sericata to determine the impact of P. mirabilis (as compared to a mixed bacterial community) on the persistence of Salmonella in flies, finding that exclusion by P. mirabilis was more effective in L. sericata than in house flies (Greenberg et al. 1970). This strategy is critical in blow fly studies, as these flies are associated with numerous bacterial species (Singh et al. 2015). Therefore, it is necessary to break the impact of a fly associated community into its individual components in order to dissect the role of specific taxa in community functions. Recently, this strategy has also been used to dissect the impact of specific bacteria on blow fly life history (Crooks et al. 2016). This approach is feasible in blow flies as many of their bacterial associates can be cultured.

Now it is possible to conduct these experiments and investigate molecular changes in bacteria and flies associated with gnotobiotic flies as compared to sterile and nonsterile flies at a level of detail that was not possible when Greenberg and his colleagues were conducting their experiments (Graf 2016). This can be done at the mRNA and proteomic levels described above. However, metabolic shifts can also be studied. It is possible to assay basic metabolic responses of flies and bacteria with any number of commercially available panels and tests, such as Biolog plates (Garland and Mills 1991, Garland 1997, Smalla et al. 1998, Choi and Dobbs 1999, Dobranic and Zak 1999, Classen et al. 2003, Bochner et al. 2011), which can be used to evaluate metabolic profiles of both bacteria and eukaryotes. Given the connections between bacterial signaling molecules and essential amino acids, the metabolic profiles of both bacteria and flies will be useful in dissecting their interactions. Some work has already been done with this tool in carrion ecology (Pechal et al. 2013), using plates that included molecules found to be of importance in previous blow fly studies (Ma et al. 2012, Tomberlin et al. 2012). However, individual blow flies and bacteria are small components of that complex system, making it difficult to assign metabolic responses to specific bacteria. Further, and more specific, dissection of blow fly interactions with bacteria using these and similar tools will be useful in determining and differentiating among hypotheses regarding the mechanisms regulating them.

In addition, mass spectrometry (MS) and other chemical analysis techniques can be used to investigate the specific impacts of the presence or absence of key bacteria noted in this review on metabolic responses of flies. Nano-scale ion MS has been used to evaluate metabolic roles of specific bacteria in vivo (Lechene et al. 2007, Li et al. 2008). In the blow fly system, it is also useful to observe advances in the study of the chemistry of decomposition. Forensic studies of the VOCs found during decomposition are commonly conducted, as signatures of human and animal decomposition are useful in death investigations (Vass et al. 2002, Statheropoulos et al. 2005). This literature demonstrates a complex array of products produced during decomposition, including known fly attractants—many of which are chemically similar to other components of the system and include many components produced by the numerous microbes associated with decomposing remains. In order to more fully dissect the chemical signatures of decomposing remains, two dimensional (through the use of double gas chromatography separation steps) MS has been used to enhance the distinctions among similar molecules (Stadler et al. 2012, Perrault et al. 2015). Such advanced chemical analysis approaches will also be useful in gnotobiotic dissection of the fly–bacteria interactions in relation to and independent of decomposition processes. As noted above the mutation of a single P. mirabilis gene impacts L. sericata biology, yet the chemical impact of that mutation alters the concentrations of numerous molecules produced by the bacterium, including several known and potentially novel fly attractants derived from essential amino acids (Ma et al. 2012, Tomberlin et al. 2012, Liu et al. 2016). The potential impact of such bacterial mutations on internal interactions with hosts could also be numerous. Therefore, to fully appreciate the impacts of mutations within a species or metabolic shifts due to presence of different bacteria on the complex metabolite repertoire of eukaryotes, it will be useful to employ modern chemical analyses to screen for novel components regulating blow fly–microbe interactions. Doing so will assist in breaking down the roles of particular bacteria in blow fly interactions. Expanding the approach to mixtures of key species and whole communities will allow for the determination of the degree to which single species versus communities are important to such interactions.

Future Endeavors

Technological and conceptual advances have set the stage for furthering our understanding of how blow flies interact with bacteria and other microbes. The ability to sequence and assemble genomes rapidly and cheaply has opened the door to developing a more comprehensive genetic, physiological, and metabolic framework for the types of studies described above that surpass the community sampling information provided by the small number of 16S rDNA sequencing studies described in the previous section. These proposed advances can be determined through genomic studies of both the microbes in this system and the flies. Pairing such studies with organismal studies of behavior and physiology will deepen our understanding of how blow flies interact with bacterial symbionts.

In conclusion, due to their close association with decomposing organic matter, blow flies also have close associations with numerous bacteria that thrive in these environments. Research has shown that bacteria are the initial colonizers of such material, and their degradation of these materials results in the production of VOCs that attract blow flies to colonize the material. Additionally, numerous bacteria species have been identified from the flies themselves, and laboratory studies indicate some bacteria are required for normal immature development, while other species enhance development. These findings hold true for other groups of flies such as mosquitoes (Diptera: Culicidae) and black soldier flies (Diptera: Stratiomyidae), suggesting that blow flies might be a good model system for studying such interactions. The bacterium P. mirabilis is commonly found throughout numerous systems, and a compound it produces, indole, is rapidly being identified as a compound that has enormous effects on the behavior of many higher animals.

From basic research aimed at understanding the underlying mechanisms by which these two groups of organisms interact, numerous areas of application can arise. These systems can be mimicked and manipulated to develop trapping mechanisms for filth flies in agricultural settings. The mechanisms by which quorum-sensing compounds such as indole are produced have implications in medicine, as these result in pathogenicity in many bacteria species that affect human health. The response of blow flies to bacterially-derived VOCs in terms of resource location can advance forensic entomology, allowing researchers to determine how, why, and when blow flies are attracted to a carrion resource. These applications represent only the beginning of this exciting area of research and the establishment of blow fly interactions with bacteria as a model system for studying the effects and impacts of microbe–insect interactions and interkingdom communication.

Acknowledgment

A.M.T. and J.K.T. would like to thank Texas A&M University and Texas A&M AgriLife Research for providing support to develop this manuscript. We thank Robert Alaniz at Texas A&M for helpful discussions regarding gnotobiotic research and Natalie Gasz at Deakin University for reviewing this manuscript. AMT is funded by the National Institute of Justice grant 2015-DN-BX-K020. JKT, TLC, AMT, and BS were also funded by the National Institute of Justice grant 2010-DN-BX-K243. Points of view in this document are those of the authors do not necessarily represent the official position or policies of the U.S. Department of Justice.

References Cited