Wolbachia Affects Reproduction and Population Dynamics of the Coffee Berry Borer (Hypothenemus hampei): Implications for Biological Control

Wolbachia are widely distributed endosymbiotic bacteria that influence the reproduction and fitness of their hosts. In recent years the manipulation of Wolbachia infection has been considered as a potential tool for biological control. The coffee berry borer (CBB), Hypothenemus hampei, is the most devastating coffee pest worldwide. Wolbachia infection in the CBB has been reported, but until now the role of Wolbachia in CBB reproduction and fitness has not been tested. To address this issue we reared the CBB in artificial diets with and without tetracycline (0.1% w/v) for ten generations. Tetracycline reduced significantly the relative proportion of Wolbachia in the CBB microbiota from 0.49% to 0.04%. This reduction affected CBB reproduction: females fed with tetracycline had significantly fewer progeny, lower fecundity, and fewer eggs per female. Tetracycline also reduced the population growth rate (λ), net reproductive rate (R0), and mean generation time (T) in CBB; the reduction in population growth was mostly due to variation in fertility, according to life time response experiments (LTREs) analysis. Our results suggest that Wolbachia contribute to the reproductive success of the CBB and their manipulation represents a possible approach to CBB biocontrol mediated by microbiome management.


Introduction
Wolbachia is the most-widely studied genus of endosymbiotic bacteria that infect insects. Wolbachia species are cytoplasmically inherited. They are known as reproductive parasites; since they manipulate the reproduction of their hosts in order to enhance their own transmission [1][2][3]. Reproductive manipulations include four mechanisms; cytoplasmic incompatibility, parthenogenesis, feminization, and male killing [4][5][6][7]. However, the relation of Wolbachia and their insect hosts also can be mutualistic; several studies have shown that infection with Wolbachia may be necessary for the normal reproduction of their hosts [8][9][10][11][12][13][14]. For example, in some species of beetles (Family Curculionidae) the elimination of Wolbachia with antibiotics significantly reduced fecundity, and in some cases eggs laid by cured females were not viable [11,13].
The potential role of Wolbachia in biological control has received attention in recent years. Infection with Wolbachia can be manipulated to increase populations of beneficial insects such as parasitoids or to reduce populations of pests [2,15].
The coffee berry borer (CBB) Hypothenemus hampei (Ferrari), (Coleoptera: Curculionidae) is the most devastating pest of coffee worldwide [16][17][18][19]. CBB females tunnel through the fruit until they reach the endosperm, where they oviposit [16,19]. The entire life cycle occurs inside the fruit and the only stage susceptible to CBB management is when the fertilized females leave the fruit to search for another fruit to colonize. This cryptic life cycle and resistance to insecticides such as endosulfan make CBB control extremely difficult and expensive.
Infection of CBB with Wolbachia has been reported in adult females from several countries [20]. It is thought to induce cytoplasmic incompatibility and to distort the CBB's sex ratio towards females [20]. However, the effect of elimination of Wolbachia on sex ratio and other reproductive traits has not been tested in CBB. The objectives of this study were to determine the infection status of Wolbachia in females from Puerto Rico, and test (by treatment with antibiotics) the effect of removal of Wolbachia on CBB reproduction, sex ratio, and population dynamics.

Insect Rearing
Coffee berry borers were reared for ten consecutive generations on the Cenibroca artificial diet [21]. Approximately 10 mL of diet was placed in sterile, clear plastic vials (23 mm × 92 mm, with a plug of bonded dense-weave cellulose acetate, Genessee Scientific, San Diego, CA, USA). Insects were reared in the dark in a growth chamber (Model 818, Thermo Scientific, Dubuque, IA, USA) at 25 • C (±1 • C) and 80%-96% relative humidity. Colonies were maintained in the Center for Excellence in Quarantine and Invasive Species at University of Puerto Rico Agricultural Experimental Station in Río Piedras, PR, USA.
Adult CBB females were collected from infested fruits from arabica coffee in Adjuntas, Puerto Rico (18 • 10 42.4 N, 66 • 44 36 W, 527 m above sea level (masl)). Fruits were carefully dissected with aid of a stereoscope, and females were removed and surface-disinfected by submersion in 1% sodium hypochlorite for 1 min, followed by two rinses with sterile distilled water. For all generations three females were placed in each vial. After each generation adult females were removed from the diet and mixed (randomized) before disinfection and transferal to fresh diet to begin the next generation.

Antibiotics Treatment
For direct and continuous feeding of larvae and adults, tetracycline hydrochloride and penicillin G (Fisher Bioreagents, Pittsburgh, PA, USA) 0.1% (w/v) were added to the Cenibroca diet. Tetracycline is efficient at removing Wolbachia [14,22,23]; penicillin is not active against Wolbachia [3,24,25] but sometimes is used to control bacterial growth in insect diets; it was used as a parallel control. An additional control group received no antibiotics.

Detection of Wolbachia
DNA was extracted from the entire bodies of adult females obtained from infested fruits in Adjuntas, Puerto Rico, and from laboratory colonies reared on Cenibroca diet with and without antibiotics. Insects were conserved dry and stored in the freezer at −20 • C; total DNA was extracted using the DNeasy Plant Mini Kit (Qiagen Sciences, Hilden, Germany).
The presence of Wolbachia in the CBB females was determined by amplification of part of the wsp gene (primers wsp-F: 5 GGGTCCAATAAGTGATGAAGAAAC 3 and wsp-R: 5 TTAAAA CGCTACTCCAGCTTCTGC 3 [26] with annealing at 58 • C. Positive controls for PCRs were total DNA of wild Drosophila melanogaster (Meigen) (known to carry Wolbachia) and negative controls were dH 2 O. PCR products were cleaned on QIAquick columns (Qiagen, Inc., Hilden, Germany) and sequenced in both directions in the University of Puerto Rico Sequencing and Genotyping Facility (UPR-SGF). The resulting sequences were assembled, manually examined for errors using CodonCode Aligner (version 5.1.5; CodonCode Crop., Centerville, SD, USA), and BLASTed in GenBank to compare the CBB sequences with other Wolbachia sequences.

Phylogenetic Classification in Wolbachia Supergroups
Phylogenetic analysis was used to assign sequences to Wolbachia supergroups. Three Wolbachia wsp sequences from this study were compared to GenBank reference sequences from 41 species of insects, including CBB [20]. These 41 sequences were selected from Wolbachia supergroups A and B (Table A1) and the sequences from this study were deposited in GenBank (KX436087, KX436088, and KX436089). Sequences were manually aligned with ClustalW in Mesquite [27]. A phylogenetic tree was constructed with Mega 6.0 [28] using neighbor joining, a maximum composite likelihood model, and bootstrap analysis with 1000 replications.

Relative Proportion of Wolbachia in the CBB Microbiota
Relative proportions of Wolbachia in the microbiota of adult females from control and tetracycline diets (after eleven generations) were determined using Illumina-based 16S typing. Briefly, pooled DNA from five females, with four replicates per treatment, was extracted using DNeasy Plant Mini Kit (Qiagen Sciences). Bacterial 16S rRNA gene was amplified using the hypervariable region V4 with primers 515-F and 806-R [29] with the addition of barcoding sequences and Illumina (Illumina Inc., San Diego, CA, USA) adapters. Sequencing was done on the Illumina MiSeq platform at UPR-SGF. Sequences were assigned to Wolbachia using OTU (Operational Taxonomic Units) tables generated with QIIME (Quantitative Insights Into Microbial Ecology) [30]. This experiment is described in detail by Mariño [31].

Effect of Antibiotic Treatments on CBB Reproduction and Sex Ratio
For each generation eggs, larvae, pupae, juveniles, and adults in 20 vials per treatment were counted by carefully dissecting the diet from each tube. The sex ratio was calculated as the number of CBB females divided by the number of males. Sex was determined for adult and juveniles by visual inspection of the body and wing shape and size; males are smaller than females and have reduced degenerative wings [16].
To determine the effect of the elimination or reduction of Wolbachia on CBB fecundity, females from the fifth generation of the three treatments were transferred individually to vials with artificial diet (n = 50 per treatment); oviposition and number of eggs per female were registered daily for two months.

Effect of Antibiotic Treatment on Life Table Parameters of the CBB
Five cohorts of 20 eggs each were established for each type of diet. Eggs were maintained in the dark at 25 • C (±1 • C) and a relative humidity of 80%-96%. Each cohort was evaluated daily for 50 days at 24-h intervals to determine CBB survival and the transitions to the next stages of development. Also, the development time from egg to adult was determined.
For population dynamics analysis, the basic life cycle for the CBB was defined as egg, larva, pre-pupa, pupa, juvenile, and adult. The probabilities of the projection matrix were calculated at intervals of six days: (Gi) probability of growth or molt between stages, (Pi) probability of surviving and remaining in the same stage, and (Fi) fertility of adult females; these values were based on the data obtained in this study for oviposition of fifty females. The probability of adult survival was estimated as 0.99, because observed mortality of adults was very rare once they emerged; several studies also reported low natural mortality rates in field and laboratory experiments [32][33][34].
Based on the projection matrix (Equation (1)) for each cohort in the three types of diet, we calculated: the population growth rate (λ), the net reproductive value (R 0 ), and mean generation time (T) [35].
A life table response experiment (LTRE) analysis was done to determine the contribution of each life cycle transition in the observed reduction of the population growth rate (λ) in CBB populations treated with tetracycline (0.1% w/v) compared with the other two diets (control and penicillin).

Data Analysis
The abundance of individuals in various stages of development of the coffee berry borer and the number of eggs laid per female were estimated using a log-linear model (GLM) with a Poisson distribution; type of diet was defined as an independent variable. Chi-square analysis was used to compare differences in the numbers of females that oviposited among types of diet and to determine differences in the number of reads of Wolbachia found in adult females between tetracycline vs. control diets. One way ANOVAs followed by post-hoc Tukey tests (significance level p = 0.05) were used to test differences among types of diet on sex ratio, life table population parameters, developmental time from egg to adult, and the duration of each stage of development. The package popbio (version 2.4.3) was used to calculate all life table parameters and perform the LTRE analysis [36]. All analyses were performed in R [37].

Detection of Wolbachia
A ≈ 750-bp fragment was amplified and sequenced from the Wolbachia wsp gene from CBB field-collected and females reared in the laboratory on the Cenibroca diet. The positive control (Drosophila) amplified a band of similar size; no product was amplified in any of the negative controls. Sequences of seven wsp fragments exhibited 100% sequence similarity to Wolbachia from Drosophila melanogaster (FJ403332).
The wsp sequences from H. hampei from Puerto Rico grouped with those of Wolbachia supergroup A. The sequences from CBB formed a separate subclade within supergroup A, closely related to one from a symbiont of the parasitoid Tachinaephagus zealandicus (Ashmead) (DQ380884) [38] (Figure 1). In contrast, a wsp consensus sequence from H. hampei from India and Brazil (AF389084) grouped with supergroup B [20], and with others from the adzuki bean beetle Callasobruchus chinensis (Linnaeus) (AB038339) [26].

Effect of Antibiotic Treatment on Wolbachia Infection and CBB Reproduction
After eleven generations on diet with tetracycline Wolbachia was still detected in CBB. However, the relative proportion of Wolbachia in the CBB microbiota was significantly lower in females from the diet with tetracycline than from the control diet (0.04% vs. 0.49%, χ 2 = 6265.9, df = 1, p < 0.001).
The log-linear model gave a good fit for the effects of diet on CBB total population (χ 7 2 = 225, p < 0.0001). In general, the estimated population was significantly higher in control and penicillin diets than in tetracycline (Table A2). The estimated population per generation increased independently of type of diet (Table A2). After the F 3 generation females reared with tetracycline produced significantly fewer progeny, and in the F 10 generation less than half as many individuals were produced compared with control and penicillin diets ( Figure 2A).The estimated abundance of individuals at all stages of development varied significantly among diets (GLM, Poisson errors, eggs Z = 232.2, p < 0.0001; larvae Z = 194.7, p < 0.0001, pupae Z = 107.9, juveniles Z = 52.20, p < 0.0001, and adults Z = 295.4, p < 0.0001).
In general, the number of individuals in all stages of development was significantly lower in diets with tetracycline than in controls.
The proportion of females that oviposited was significantly lower in diet with tetracycline (38% in tetracycline, 68% in control, and 66% in penicillin, χ 2 = 11.50, df = 2, p = 0.003). Also, the number of eggs per female was significantly lower in diets with tetracycline: on average, each female laid 3.45 eggs compared with 18.83 in control and 19.58 in penicillin (GLM, Poisson errors, eggs Z = 86.79, p < 0.0001) ( Figure 2B). Significant differences were found in sex ratio between diets (F = 6.48, df = 2, p = 0.002). The sex ratio was less skewed towards females in tetracycline diet than in control and penicillin diets (19.4:1 for tetracycline, 24.1:1 for control, and 26.5:1 for penicillin). These differences in sex ratio were significant from the F3 generation and all subsequent generations; no differences were observed in F1 and F2 (Table 1). However, no differences were observed in the number of males among diets (control vs. penicillin: Z = 1.32, p = 0.25; control vs. tetracycline: Z = −0.52, p = 0.59). from the F3 generation and all subsequent generations; no differences were observed in F1 and F2 (Table 1). However, no differences were observed in the number of males among diets (control vs. penicillin: Z = 1.32, p = 0.25; control vs. tetracycline: Z = −0.52, p = 0.59).

Effect of Antibiotic Treatment on Life Table Parameters of the CBB
The duration of the juvenile stage was significantly longer in tetracycline than in control and penicillin. However, the duration of the remaining stages did not differ between diets; also, the egg-to-adult developmental time did not differ significantly between diets, but was slightly higher for individuals reared with tetracycline ( Table 2).

Effect of Antibiotic Treatment on Life Table Parameters of the CBB
The duration of the juvenile stage was significantly longer in tetracycline than in control and penicillin. However, the duration of the remaining stages did not differ between diets; also, the egg-to-adult developmental time did not differ significantly between diets, but was slightly higher for individuals reared with tetracycline ( Table 2). Life table parameters for the three diets are presented in Table 3. Net reproductive value R 0 differed significantly among diets ( Table 3). The net reproductive value (R 0 ) was 23 females/female for tetracycline, 51 for control, and 45 for penicillin. In other words, a female on these diets produced on average 23, 51, and 45 new females respectively in each generation. For all diets the population growth rate was >1, but populations treated with tetracycline had a slightly lower λ; the LTRE showed that reduction in fecundity of adult females feeding with tetracycline made the largest contribution to the reduction in λ (Figure 3). Although differences in the mean generation time (T) were not significant, tetracycline treatment increased the time by 6.2 days, which means that females reared with tetracycline will produce fewer generations per year (Table 3).  Life table parameters for the three diets are presented in Table 3. Net reproductive value R0 differed significantly among diets ( Table 3). The net reproductive value (R0) was 23 females/female for tetracycline, 51 for control, and 45 for penicillin. In other words, a female on these diets produced on average 23, 51, and 45 new females respectively in each generation. For all diets the population growth rate was >1, but populations treated with tetracycline had a slightly lower λ; the LTRE showed that reduction in fecundity of adult females feeding with tetracycline made the largest contribution to the reduction in λ (Figure 3). Although differences in the mean generation time (T) were not significant, tetracycline treatment increased the time by 6.2 days, which means that females reared with tetracycline will produce fewer generations per year (Table 3).
Transitions P 1 G 1 P 2 G 2 P 3 G 3 P 4 G 4 P 5 G 5 P 6 F i

Formatting of Mathematical Components
For the projection matrix, variations in the size and age structure of a population N through time t and t+1 were computed from the equation: where A is a population matrix and N is a vector describing the age of structured population. Thus, the projection matrix was:

Wolbachia in Insects and Status in the CBB
Wolbachia are widespread and common insect endosymbionts and have been detected in more than 66% of insect species tested [7,39]. In many species, infection rate is variable; infected individuals within a species ranges from <3% to 100% [40,41]. Also, the titer of Wolbachia can be as low as <0.01 or as high as >1200 cells per insect cell [42,43] (though Wolbachia are not uniformly distributed throughout the insect but concentrated in special organs). Furthermore, there are populations with high infection frequency of Wolbachia, but titer per individual is low. This was demonstrated in the bark beetle Pytogenes chalcographus (Linneaus), which had infection frequencies of 85%-100%, but the titer per cell was <0.5 [44]. It is not clear if there is a threshold titer for Wolbachia to affect insects.
In the case of the CBB, there are no published studies that reported the infection rate or titer. Vega et al. [20] detected Wolbachia in ten of fifteen samples, which could suggest a high infection frequency, but they did not report how many individuals were tested per sample. The data reported here are averages from pooled samples, so there is no way of detecting variation in titer among individuals. In this study the success rate for amplifying Wolbachia using conventional PCR was <10%. 16S sequencing showed a low relative proportion of Wolbachia in the CBB microbiota of adult females reared in artificial diets and collected from the field, 0.49% and 0.16% respectively [31]. A low titer of Wolbachia in the CBB may explain the difficulty of amplification, as Vega et al. [20] suggested. Yet the data presented here suggest that even a low titer of Wolbachia can have a significant effect on insect reproductive biology. Furthermore, the difference between treatments with 0.49% relative proportion of Wolbachia in females from control diet and 0.04% in females from tetracycline suggests the possibility of a dosage-dependent effect.

Placement of CBB Symbionts in Wolbachia Supergroups
For Wolbachia classification the wsp gene was preferred by Zhuo et al. [5]; wsp is evolving more rapidly than 16S rDNA and ftsZ genes. According to phylogenetic studies using these three genes, the genus Wolbachia is divided into eight supergroups (A-H) [45][46][47][48]. The supergroups A and B are exclusively found in arthropods and are the most common; C and D are found in nematodes. The remaining supergroups (E-H) have been proposed more recently and may be less common [46,47,49]. However, the diversity and the number of supergroups in Wolbachia are increasing as different genes are explored: for example, eleven supergroups were recognized when the additional genes gltA and groEL were included [50].
The term supergroup is used to define a clade of phylogenetically related strains [5,49]; the Wolbachia in CBB females from Puerto Rico belonged to supergroup A, while the 'consensus sequence' from India and Brazil belonged to supergroup B [20] (Figure 1). The subdivision of Wolbachia into supergroups was done in response to its high genetic diversity compared with other endosymbionts [49]. However, to our knowledge there are no studies demonstrating functional differences between Wolbachia supergroups A and B. More studies should be conducted to determine the diversity and distribution of Wolbachia strains in the CBB, and to elucidate the significance of this diversity in CBB biology.

Antibiotic Treatments and Their Effect on CBB Reproduction and Fitness
The effect of tetracycline on insect reproduction and fitness of Wolbachia-infected insects can result from curing or reducing Wolbachia infection, but could also reflect direct toxicity on host physiology [14]. It is also possible that tetracycline caused other changes in the microbiota unrelated to Wolbachia. These possible explanations are explored further in Mariño [31]. To reduce the possibility that the observed effects on CBB were a result of direct antibiotic toxicity or other changes in the microbiota, we included a parallel treatment with penicillin G; penicillin G is not effective against Wolbachia [3,24,25]. Tetracycline and penicillin differ in their modes of action and ranges of activity; however, both antibiotics can affect directly the fecundity and survival of insects [51][52][53]. Moreover, in this study penicillin did not affect reproduction in the CBB, the reverse of what would occur if the observed effects were due to toxicity. It also reduces (but no eliminates) the possibility that the results reflect the effect of antibiotics on nontarget groups of bacteria, because penicillin has a narrower spectrum of activity than tetracycline.
Our results showed that females reared on penicillin and control diets produced a similar number of individuals in all stages of development, and there were no significant effects of penicillin on reproduction and life table parameters (Figure 2 and Tables 1-3). Pseudomonas fulva (Iizuka and Komagona) is involved in caffeine degradation in the CBB [54]. When P. fulva was eliminated from adult females fed with a mix of three antibiotics including tetracycline, there was a significant reduction in the number of eggs in females fed with the antibiotics [54]. However, reinoculation with P. fulva did not restore the normal reproduction of females. This suggests that P. fulva was not responsible for the reduced egg production, and Wolbachia may have been responsible. This reduced reproduction with tetracycline may be due to the effect of reduction of Wolbachia (though this possibility was not mentioned by Ceja-Navarro et al. [54]; Rickettsia, which includes other endosymbiotic bacteria that affect insect reproduction, was recently reported in the CBB [54], but, we did not detect it in CBB from Puerto Rico [31]. Treatment with tetracycline is a well-known method to eliminate Wolbachia from hosts [11,14,55] and test its effect on the reproduction of its hosts. However, our treatment with tetracycline 0.1% (w/v) did not eradicate it, even after eleven generations of continuous exposure. Tetracycline 0.1% or (even less) was enough to remove Wolbachia from other hosts [22,[55][56][57]. Even though the infection was not completely cured, our results strongly suggest a relationship between Wolbachia and the successful reproduction of CBB females. After three generations, the reproduction of females treated with tetracycline started to decrease significantly; by the tenth generation their progeny was less than half compared to females reared in penicillin and control diets (Figure 2A). Also, females treated with tetracycline produced significantly fewer eggs ( Figure 2B).
This fertility reduction agrees with data from other beetles in the family Curculionidae. In the date stone beetle (Coccotrypes dactyliperda Fabricius) mated females fed with 3% tetracycline produced significantly fewer eggs than control females [11]. Similarly, in the rice water weevil (Lissorhoptrus oryzophilus Kuschel) females treated with 0.25%-2.5% tetracycline produced fewer eggs and none of these eggs were viable [13]. The authors of both studies attributed the reduction in fecundity in these species to the elimination of endosymbiotic bacteria, mainly Wolbachia.
Wolbachia are commonly known as reproductive parasites, due their capacity to manipulate the reproduction of their hosts in different ways to enhance their own transmission [1]. However, some Wolbachia can be mutualists [6], involved in nutritional mechanisms reported in Drosophila melanogaster [58] and recently in Cimex lectularius (Linneaus) [59] and Aedes albopictus (Skuse) [60]. Our results suggest that Wolbachia plays an important role in CBB reproduction.
In the only previous study of Wolbachia in CBB, Vega et al. [20] suggested that Wolbachia plays a role in the female-biased sex ratio observed in the CBB. Our results showed that Wolbachia's effects on the sex ratio are at best indirect: females treated with tetracycline produced significantly fewer daughters than females from controls, but tetracycline did not affect significantly the number of males produced ( Table 1). The less skewed sex ratio observed with tetracycline was more a consequence of the reduction in the number of females than the production of more males.
Contrary to Vega et al. [20], we suggest that Wolbachia is not involved in the sex determination of CBB; a significant increase in the number of males in progenies from females treated with tetracycline would be the expected result if Wolbachia were involved in sex determination. Other mechanisms, mainly the sex determination systems haplodiploidy and functional haplodiploidy, are known to induce highly female-biased sex ratios [22,70]. Functional haplodiploidy is the sex determination system described for the CBB [71]; in this system, the sex ratio distortion is a consequence of high male mortality, which occurs due to the loss of paternal chromosomes [9,12].
However, this result needs to be confirmed through the production of Wolbachia-cured lines. The best method to assess the impact of Wolbachia on sex determination in its hosts is to compare an infected line with a Wolbachia-cured line [72]. It is also important to determine the minimum dose of tetracycline needed to completely cure Wolbachia infections; 3% (w/v) tetracycline eliminated endosymbiotic bacteria, including Wolbachia, from Coccotrypes dactyliperda (Coleoptera: Curculionidae) [11]. However, such high concentrations also exacerbate toxicity issues. Alternatively, another antibiotic like rifampicin could be used; rifampicin is also effective against Wolbachia [14,24].
Relatively few studies have measured the effect of Wolbachia in arthropod population dynamics. Demographic studies allow comparing the growth potential for a population under different conditions; life time response experiments (LTREs) deconstruct differences on population growth into contributions from each life cycle transition [73]. In the case of insect pests like the CBB, this type of analysis could provide insight into which specific life cycle transition could be targeted for management.
Wolbachia-free populations of Liposcelis tricolor (Badonnel), (Psocoptera: Liposcelididae) had lower performance in all life parameters evaluated: intrinsic rate of increase rm, net reproductive value R 0 , and mean generation time. Similarly, our data showed that tetracycline reduced the values of all life table parameters evaluated in the CBB. The LTRE analysis identified the fecundity of adult females as the stage contributing most to the reduction of λ (Figure 3). Taken together, these data suggest that Wolbachia are important for successful reproduction of the CBB.

Conclusions
In this study we detected Wolbachia supergroup A in adult CBB females from Puerto Rico. The relative proportion of Wolbachia in the CBB microbiome is very low, less than 1%.
Wolbachia appears to contribute to reproductive success in the CBB: antibiotic treatment significantly reduced the proportion of Wolbachia, and had negative effects on reproduction and population dynamic of CBB. This result opens the door to the possibility of CBB biocontrol mediated by manipulating Wolbachia infection, either by reducing Wolbachia infection or encouraging growth of competing microorganisms in the CBB microbiota. The effects and mechanisms of Wolbachia on CBB appear to be complex, but our results show they have pronounced effects on the reproductive biology of this important pest. Considering the CBB's importance, the possibility of biological control approaches that exploit this new information deserves further investigation.   Table A2. Results of generalized linear model evaluating the relationship between CBB total population and type of diet (control, penicillin 0.1%, and tetracycline 0.1%) and between generations for CBB on diet with tetracycline. *** p < 0.001; ** p < 0.01; * p < 0.05; p < 0.1.