A screen for bacterial endosymbionts in the model organisms Tribolium castaneum, T. confusum, Callosobruchus maculatus, and related species

Reproductive parasites such as Wolbachia are extremely widespread amongst the arthropods and can have a large influence over the reproduction and fitness of their hosts. Undetected infections could thus confound the results of a wide range of studies that focus on aspects of host behavior, reproduction, fitness, and degrees of reproductive isolation. This potential problem has already been underlined by work investigating the incidence of Wolbachia infections in stocks of the model system Drosophila melanogaster. Here we survey a range of lab stocks of further commonly used model arthropods, focusing especially on the flour beetles Tribolium castaneum and Tribolium confusum, the cowpea weevil Callosobruchus maculatus and related species (Coleoptera: Tenebrionidae and Bruchidae). These species are widespread stored product pests so knowledge of infections with symbionts further has potential use in informing biocontrol measures. Beetles were assessed for infection with 3 known microbial reproductive parasites: Wolbachia, Rickettsia, Spiroplasma. Infections with some of these microbes were found in some of the lab stocks studied, although overall infections were relatively rare. The consequences of finding infections in these or other species and the type of previous studies likely to be affected most are discussed.

Maternally inherited microbes such as Rickettsia and Wolbachia can manipulate host reproduction in various ways in order to favor their own transmission (Charlat et al., 2003;Goodacre & Martin, 2012). Horizontal transfer has been achieved through laboratory manipulations (Riegler et al., 2004) and is assumed to also occur in nature, for example via transfer between a host and a parasite (Heath et al., 1999). However, the predominant route of transmission of these bacteria is vertical, hence there can be a selective advantage to the microbe favoring a bias toward infected females in the population. Such a bias can be achieved via distorting the offspring sex ratio in favor of females via male-killing (e.g., Hackett et al., 1986), feminizing male embryos (e.g., Kageyama et al., 2002) or by inducing parthenogenesis (e.g., Arakaki et al., 2000). Wolbachia has also been shown to cause cytoplasmic incompatibility (CI) altering reproduction in a range of ways (Yen & Barr, 1971;Werren, 1997;Stouthamer et al., 1999;Duron et al., 2008a). CI may greatly reduce fertility and even cause sterility, with effects being either unidirectional (i.e., between infected and uninfected individuals), or bidirectional (i.e., between individuals infected with different Wolbachia strains). Such incompatibilities, especially when bidirectional, could limit gene flow amongst populations of a species and can be proposed to be influential in longer term evolutionary processes such as the development of reproductive isolation and, ultimately, speciation (Wade & Stevens, 1985;Breeuwer & Werren, 1990;Telschow et al., 2005). Finally, general effects on reproduction and fitness have also been documented (see Table 1 for an overview). These effects need not be negative, indeed Wolbachia infections are shown to increase resistance to particular viruses in Drosophila (Hedges et al., 2008;Osborne et al., 2012) and Aedes aegypti (Bian et al., 2010).
It has been suggested that the finding of the intracellular bacterium Wolbachia in ca. 30% of commonly used Drosophila stocks (housed at the Bloomington Drosophila Stock Center) might call into question the results of many evolutionary studies (Clark et al., 2005). The widespread occurrence of Wolbachia in such a ubiquitously used model organism is potentially alarming because it raises the possibility that differences in reproduc-tive and/or fitness traits or compatibilities between populations might have a microbial basis rather than solely be caused by other proposed mechanisms.
It should be noted that the situation is more complicated than merely considering whether or not populations harbor Wolbachia (or any other individual reproductive parasite). Seemingly "uninfected" stocks could well harbor other bacteria that can similarly affect their hosts (e.g., Cardinium, Flavobacteria, Rickettsia, Spiroplasma, Arsenophonus). A survey of stocks of different Drosophila species indeed finds that several species also harbor Spiroplasma (Tucson Drosophila Species Stock Center: Mateos et al., 2006). Similarly extensive surveys have assessed infections with various reproductive parasites in further dipteran species belonging to the superfamily Muscoidea (ca. 70 species; Martin et al., 2012), and the Dolichopodidae and other Empidoidea (ca. 240 species; Martin et al., 2013a,b). Although Wolbachia infected flies more commonly, infections with Spiroplasma, Rickettsia, and Cardinium were also found (Martin et al., , 2013a. There is extensive evidence for both Wolbachia and Spiroplasma causing differences in host reproduction, including in Diptera (Duron et al., 2008a). Nevertheless, it is unclear how problematic the widespread infections in Drosophila stocks (Clark et al., 2005) actually are to the evolutionary studies carried out on them. How robust are conclusions drawn from previous experiments where the bacterium might-or might not-have been present? Here we review the type of experiments or traits under study that are likely to be most susceptible.
Considering how widespread such bacterial endosymbionts are among arthropods (Goodacre et al., 2006;Duron et al., 2008a;Hilgenboecker et al., 2008), and that Wolbachia is not the only microbe known to have such effects, we include in our study data from a range of model systems where we establish the presence of Wolbachia and of other microbes that are similarly implicated in altering the biology of their hosts. Conceivably, the presence of such parasites will be most relevant in model systems used extensively for studies on reproduction. Beyond Drosophila melanogaster, other lab organisms, which have (among other things) been used frequently for investigating reproductive biology, are the flour beetles belonging to the genus Tribolium (Tenebrionidae) and Callosobruchus seed beetles (Bruchidae), which we have chosen to target specifically here, including related species.
In summary, the aims of this study were twofold: (i) to survey a broad selection of commonly used laboratory strains of Tribolium and Callosobruchus beetles and related species for infection with 3 microbial reproductive parasites (Rickettsia, Spiroplasma, and Wolbachia) and (ii) to assess consequences of finding such infections in Infection increases tolerance to heat shock Rickettsia: Bemisia tabaci (Brumin et al., 2011) these or other species and discuss the type of previous evolutionary study and data most likely to be at risk from the confounding effects of endosymbiont infections.

Tenebrionidae
The red flour beetle Tribolium castaneum is a widespread pest and has become a major model system for the study of pre-and postcopulatory sexual selection (Fedina & Lewis, 2008;Pai & Bernasconi, 2008;Michalczyk et al., 2010;Sbilordo et al., 2011;Grazer & Martin, 2012), and sexual conflict (Michalczyk et al., 2011a). This promiscuous species has also been used to assess the causes and consequences of polyandry, with recent examples focusing on the roles played by inbreeding (Michalczyk et al., 2011b) and environmental change (Grazer & Martin, 2012). T. castaneum is also an important model in the study of host-parasite conflicts and immunity (e.g., Blaser & Schmid-Hempel, 2005;Zou et al., 2007;Wegner et al., 2008Wegner et al., , 2009Bérénos et al., 2009;Hangartner et al., 2013;Kerstes et al., 2013). Similar to D. melanogaster, one of the attractions of this system is the access to molecular tools such as the sequence of the entire genome of T. castaneum . Stocks of the related confused flour beetle Tribolium confusum are already known to commonly harbor Wolbachia, with the microbe causing CI in this species (Fialho & Stevens, 1996). Interestingly, this CI-inducing Wolbachia strain is genetically indistinguishable (based upon sequences from 4 gene regions) from that infecting the congeneric species Tribolium madens where it causes male killing (Fialho & Stevens, 2000).

Bruchidae
Callosobruchus (Bruchidae) seed beetles are pests of stored legumes and can be easily reared in the lab. These species have also been the subject of intense study in the context of postcopulatory sexual selection (e.g., Wilson et al., 1997;Eady et al., 2004;Maklakov & Fricke, 2009), sexual conflict (Crudgington & Siva Jothy, 2000;Arnqvist et al., 2005;Rönn et al., 2007Rönn et al., , 2011, and reproductive isolation (Fricke & Arnqvist, 2004), including targeted experimental evolution studies (e.g., Fricke & Arnqvist, 2007;Gay et al., 2009;Maklakov et al., , 2010. Callosobruchus has also been the focus of detailed studies of the fitness consequences of ageing and inbreeding (Fox et al., 2004(Fox et al., , 2006(Fox et al., , 2011aBilde et al., 2009) including relationships with the environmental context (Messina & Fry, 2003;Fox et al., 2011b). Finally, many of the above representatives of the bruchid and tenebrionid beetles and closely related species are widespread pests of stored products. This adds an applied perspective, as Wolbachia has been discussed intensively as a potentially useful ally in the fight against pests and vectors of disease, for instance, of mosquitos (see Laven, 1967;reviewed in Iturbe-Ormaetxe et al., 2011) or medflies (Zabalou et al., 2009). More generally, greater consideration of impacts of symbionts on insect pests and vectors has been argued to be critical in assessing risks and effectiveness of biocontrol measures (Zindel et al., 2011).
Prior to testing, all the tenebrionid beetle stocks assayed in the present study had been maintained at large population sizes and housed on organic flour (with 10% brewer's yeast) in dark climate chambers at a constant 30°C (as standard for the stocks used, see Grazer & Martin, 2012). Although higher rearing temperatures are frequently used for tenebrionids, temperatures of above ca. 36°C are known to impact on endosymbiont infections, thus stocks that had been kept under these conditions in the past were avoided in our study (see e.g., Sakamoto et al., 2008). Bruchid beetle stocks were held in climate chambers at constant conditions of 27°C and 45% (±10%) relative humidity under a 12 : 12 h light : dark cycle. Beetles were held in 1 L glass jars and maintained at large population sizes of 250-300 beetles per generation and provided with excess amounts of black-eyed beans (Vigna unguiculata).
Sample beetles for PCR testing were removed from their stock containers and subsequently kept in 70% ethanol until DNA extractions. DNA was extracted from abdominal tissue using QIAGEN DNEasy kits and eluted in 100 μL distilled water. The success of DNA extraction was established by polymerase chain reaction (PCR) using host-specific primers designed to amplify a fragment of the mitochondrial cytochrome oxidase I (COI) gene (tenebrionid beetles) and a nuclear microsatellite dimer repeat (Callosobruchus beetles) respectively (COI primers: Co12309 5 -TTT ATG CTA TAG TTG GAA TTG G-3 and Co12776 5 -GGA TAA TCA GAA TAT CGT CGA GG-3 , as described in Hedin & Maddison, 2001; Callosobruchus microsatellite primers: 5 -ATG GCG ATT GCT ATT CTG TTG-3 and 5 -AAA TAA CAG GCA TCA AAA CAA CAT-3 ; Fricke et al., unpublished). Amplification of host DNA was obtained from all our samples indicating that DNA extraction had been successful. Samples were subsequently tested by PCR for Wolbachia, Rickettsia spp. and Spiroplasma spp. using previously described methods which were as follows: (i) A section of the Wolbachia cell surface protein gene wsp was amplified using WSP-F (5 -TGGTCCAATAAGTGATGAAGAAACTAGCTA-3 ) and WSP-R (5 -AAAAATTAAACGCTACTCCAGC-TTCTGCAC-3 ; Jeyaprakash & Hoy, 2000). (ii) A section of the citrate gene in Rickettsia spp was amplified using RICS741F (5 -CATCCGGAGCTAATGGTTTTGC-3 ) and RCIT1197R (5 -CATTTCTTTCCATTGTGCCATC-3 ; Davis et al., 1998). (iii) A section of the intergenic ribosomal spacer of the Spiroplasma ixodetis group was amplified using Spits-J04 (5 -GCCAGAAGTCAGTGTCCTAACCG-3 ) and Spits-N55 (5 -ATTCCAAGGCATCCACCATACG-3 ; Majerus et al., 1999). All PCRs were carried out in an MJ cycler in a total volume of 25 μL containing 1 unit of Taq, 2.5 mmol/L MgCl 2, 0.5 mmol/L of each dNTP, 400 nmol/L of each primer and 1 μL of DNA solution, in a buffer of 10 mmol/L Tris-HCl, 50 mmol/L KCl pH 8.3 (20°C). An initial denaturation at 94°C for 1 min was followed by 35 cycles of 94°C for 30 sec, 55°C (endosymbiont genes) or 50°C (COI gene) or 53°C (Callosobruchus microsatellite) for 20 sec and 72°C for 30 sec. Bands were visualized by gel electrophoresis on a 1.5% agarose gel stained with ethidium bromide.
All PCRs were run in the presence of both positive and negative controls. The list of stocks tested for presence of Rickettsia, Spiroplasma, and Wolbachia can be found in Table 2 and associated footnotes.

Tenebrionidae
The results of our PCR survey for infections with the 3 endosymbionts are displayed in Table 2. Results confirm the presence, as expected, of Wolbachia in T. confusum, where it has previously been shown to cause CI (Wade & Stevens, 1985). In all T. confusum strains except HP70 both males and females were positive for Wolbachia. Previous studies have indicated that separate stocks may harbor identical (or at least compatible) Wolbachia strains (Fialho & Stevens, 1996). Preliminary crosses between infected and uninfected stocks appear to confirm this result (Martin, unpublished data). In contrast with T. confusum, individuals from the large number of T. castaneum strains tested were all apparently devoid of Wolbachia infections. Whereas in the former 6 of 8 stocks tested positive for Wolbachia, in contrast none of the ca. 40 T. castaneum strains tested appeared to be infected, although 3 of these were found to carry Spiroplasma, and 1 harbored Rickettsia (for details see Table 2). The closely related species Tribolium freemani also appeared to be free of Wolbachia infection as was the single strain of T. madens tested in this survey. Others have shown that T. madens can be infected with Wolbachia strains genetically indistinguishable from that infecting T. confusum and that the bacterium distorts sex ratio by causing malekilling (Fialho & Stevens, 2000). In accordance with the lack of infection in this study, no bias in sex ratio was apparent in the stock tested here (Martin, personal observation). Similarly, no Wolbachia infections were found in the remaining congeneric species (Tribolium anaphe, T. audax, T. brevicornis, or T. destructor) or any of the other tenebrionid species tested (Gnatocerus cornutus, Latheticus oryzae, or Palorus ratzeburgii) although we note that the number of samples tested for these species was very small (only a single individual in some cases) and thus our power to detect endosymbionts that are at anything less than 100% prevalence was low. Tests for Rickettsia and Spiroplasma detected neither of these types of bacteria in any of the individuals tested.

Bruchidae
Results are presented in Table 3. Testing 16 different populations of Callosobruchus maculatus and 2 related species C. rhodesianus and C. analis shows generally very low infection rates. Spiroplasma could not be detected from any of the samples while Wolbachia was only found in 1 individual out of 4 tested in C. rhodesianus. All 3 species C. maculatus, C. rhodesianus, and C. analis show single infections with Rickettsia.

Discussion
Tests for endosymbiotic bacteria in the tenebrionid and bruchid beetles in this study appear to indicate that symbionts may be less common in these groups than in the insects assessed previously (Hilgenboeker et al., 2008). The overwhelming majority of currently available data are from studies on the interaction of insect hosts with Wolbachia, with far less being known about effects of other endosymbionts (examples in Table 1). Of the 4 classic phenotypes (CI, male-killing, feminization, and parthenogenesis), all have been documented in a range of host species for Wolbachia and a few of these have also been shown to be caused by infections with other known endosymbionts. In Tribolium spp. specifically, research has focused solely on Wolbachia, with evidence to indicate that this symbiont causes CI in T. confusum and male-killing in T. madens (Fialho & Stevens, 2000). Further impacts on nonreproductive traits are also possible as evidenced by recent work suggesting a negative effect of Rickettsia infection on long-distance dispersal behavior in a spider .
Precisely to what degree endosymbiont infections could confound results obtained from lab populations will depend on how the microbe affects the host. For example, if CI-causing bacteria remain undetected in particular insect stock populations, this could compromise studies involving interpopulation crosses. Furthermore, if the stock populations in question are not uniformly infected, it could also explain differential reproductive successes across studies of single populations. Temporal changes in reproductive success of single populations might also occur if the natural rate of bacterial transmission from mother to offspring is altered under laboratory conditions, such that populations experience rapid changes in the frequency of endosymbiont infections after only a few generations in the lab. Such issues could be especially problematic when assessing reproductive isolation using postzygotic measures, as is often the case in studies directed toward understanding processes such as genetic isolation and speciation. Prezygotic measures could also be confounded if infection status affects mate preferences (see e.g., Markov et al., 2009) or the frequency of mating ; see also Table 1). Table 2 Overview of results from PCR screens for microbial reproductive parasites in 11 tenebrionid species. Beetle stocks were screened for infection with the 3 endosymbionts Wolbachia, Rickettsia, and Spiroplasma using PCR: "+" indicates positive infection status and "-" no infection. Samples include individuals from numerous strains of the extensively used sexual selection model system Tribolium castaneum, T. confusum and related species. Stock names and locations are provided where known.  It seems perhaps less likely that phenotypes involving sex ratio skew, such as parthenogenesis, feminization or male killing could "silently" affect experimental populations. A strong bias toward females might appear likely to be picked up during routine work, although actual protocols used would need to be evaluated to assess possible risks of missing skewed sex ratios. More general and less drastic negative (or positive) effects, for example on fitness are perhaps less likely to be an issue. Here it is unclear whether one could argue that Table 3 Overview of results from PCR screens for microbial reproductive parasites in 3 bruchid beetle species. Callosbruchus beetle stocks were screened for infection with the 3 endosymbionts Wolbachia, Rickettsia, and Spiroplasma using PCR: "+" indicates positive infection status and "-" no infection. N = 1 individual per sample. Location names (where known) indicate where beetles were sampled. F = female, M = male.

Location
Sex Notes: The Brazil and South India strains were split and held in different laboratories (by G. Keeney [USA] and R. Smith [Leicester])details of their history can be found in Fricke and Arnqvist (2004). The origins of the Nigeria mix and Poly B 4 strains can be found in Fricke and Arnqvist (2007) (for other stocks and further details see: Giga & Smith, 1991;Rönn et al., 2007Rönn et al., , 2011Rankin & Arnqvist, 2008).
patterns would be majorly influenced by undetected endosymbionts, unless populations used are not uniformly infected. Laboratory populations will also be, or have been, affected by a large range of other intrinsic and extrinsic factors. These remain for the most part equally silent, and may for example include nematodes, mites, other pathogens or parasites, or selfish genetic elements such as Medea in T. castaneum (Lorenzen et al., 2008). In this respect, reproductive parasites are probably not truly a greater challenge than any other of these unknowns, which already have to be taken into account. Artificial transfer experiment protocols exist for Tribolium beetles (Chang & Wade, 1996), potentially offering a controlled way to assess effects on existing (or novel) hosts experimentally. Indeed, the effects on reproduction of the various symbionts remain largely unre-solved for many populations (or species). Reproductive parasites can specifically impact on reproductive traits (see Table 1), so beyond obvious involvement in conflict between host and symbiont, they can impinge on sexual conflict between males and females (see Martin & Gage, 2007). A promising and targeted means of illuminating the separate and combined action of these (interspecific and intraspecific) evolutionary conflicts would be to use a combined experimental evolution approach akin to previous experiments focusing on either sexual conflict (e.g., Martin & Hosken, 2003Fricke & Arnqvist, 2007;Gay et al., 2009;Hosken et al., 2009;Maklakov et al., 2010;Michalczyk et al., 2011a) or host-parasite conflict (e.g., Bérénos et al., 2009;Kerstes et al., 2013;reviewed in Kerstes & Martin, 2014). Findings of experimental evolution studies in Tribolium and Callosobruchus (e.g., Gay et al., 2009;Maklakov et al., , 2010Michalczyk et al., 2011a) coupled with detailed knowledge of reproduction in these study systems could provide a solid base for understanding interactions between hosts and their reproductive parasites.
One means of assessing symbiont effects has been to treat animals with antibiotics to cure them of their infections. However, treatment with this antibiotic also has the potential to influence other fitness traits and likely eliminates other known or unknown bacteria with unpredictable consequences. Furthermore, there are potentially other confounding effects, such as persistent effects on metabolism, after curing with Tetracycline (see e.g., Ballard & Melvin, 2007).
Infections with Rickettsia and Spiroplasma are found across a wide range of arthropods so were hence also tested for in this study in addition to Wolbachia. In fact, multiple infections within species or groups of species are not uncommon (e.g., Weeks et al., 2001;Goodacre et al., 2006). In this study we only found very few infected individuals and only one multiply infected female (C. rhodesianus infected with both Wolbachia and Rickettsia, in contrast with Kondo et al., 1999; see Table 3). More generally, though, further complications could arise if different infections interact with one another. Such intermicrobial interactions may be a promising area of future research (see e.g., Engelstädter et al., 2008).
Clearly, evolutionary biologists need to be aware of the complex relationship between a study organism and its associated symbionts or parasites. Studies such as this or the large-scale work already undertaken on Drosophila (Clark et al., 2005;Mateos et al., 2006) and other Diptera  can only be informative. Researchers should be grateful rather than alarmed that leading lab "work-horses" such as Drosophila, Tribolium, or Callosobruchus are not impervious to the range of microbial diversity commonly found in the wild. For a start, the majority of arthropod species are likely to have evolved in contact with Wolbachia, so study organisms infected with this parasite are probably more representative of the situation in the wild. Moreover, this should really be seen as a valuable opportunity to address pressing questions in a burgeoning area of research, using the well-understood systems that model lab organisms such as Tribolium provide. Here one can draw not only upon a wealth of extensive and highly relevant information on host reproduction but also access the full array of genetic tools available for these species.
To conclude, we provide data on infections with 3 common reproductive parasites in stock populations of the popular model systems T. castaneum and C. maculatus and a range of related species. We confirm an emerging pattern where Wolbachia infections are widespread in T. confusum stocks, yet the same types of bacteria (i.e., those that are sensitive to our detected methods) appear to be conspicuously absent in other Tenebrionidae assessed (see also Chang & Wade, 1996;Kageyama et al., 2010). Additionally, our results confirm a lack of Wolbachia infections in C. maculatus matching previous surveys (Kondo et al., 1999;Kageyama et al., 2010). In contrast, Wolbachia has previously been documented in C. analis and C. chinensis (Kageyama et al., 2010). However, symbionts other than Wolbachia were not assessed in previous surveys where tenebrionid or bruchid host species were included (e.g., Kageyama et al., 2010). Here, C. maculatus is found to harbor infections with Rickettsia, illustrating the point that assessing several symbionts is worthwhile (this also holds for T. castaneum, see Table 2).
It is important to emphasize that our failure to detect bacterial DNA in particular species/stocks included in this study does not imply that these are necessarily free of endosymbionts. It only indicates that the individuals that we tested do not carry bacterial strains that we can detect. We further note that the number of individuals that we have tested in our study is small. Low prevalence of endosymbionts, such as male killers (which may have a lower prevalence within a population than their CIinducing counterparts), within a population or very low bacterial titers would make it less likely that they would be detected in our study. Furthermore, divergent bacterial strains can remain undiagnosed even if present at high prevalence and/or high titer if they are not detected by our PCR methods (e.g., as demonstrated by Simões et al., 2011). The use of next generation sequencing technology to sequence all those bacteria found within in combination with more comprehensive sampling may be a useful step forward in the study of endosymbionts in model lab organisms, as has been applied for other invertebrate groups (e.g., Kautz et al., 2013.) Overall, we suggest that the widespread distribution of reproductive parasites in lab stocks is not by itself a basis for universal concern. Clearly, however, earlier interpretations should always be open to additional scrutiny or re-evaluation if necessary, that is, if stocks are infected. As a case in point, we find that the T. castaneum source population used in several recent studies (Morrow et al., 2003;Michalczyk et al., 2010Michalczyk et al., , 2011aSbilordo et al., 2011;Grazer & Martin, 2012;Hangartner et al., 2013) is free of infection with the symbionts assessed. We further propose that valuable new insights could be gained by considering new data on bacterial infections including all known reproductive parasites in further hosts. This may be particularly useful in model systems for sexual selection and related themes such as the genera Tribolium and Callosobruchus where extensive knowledge of reproduction is already available. Finally, more detailed knowledge accrued concerning infections can help build strong foundations for mounting biocontrol measures against target taxa (see e.g., Xi et al., 2005).