High temperatures augment inhibition of parasites by a honey bee gut symbiont

ABSTRACT

Temperature affects growth, metabolism, and interspecific interactions in microbial communities. Within animal hosts, gut bacterial symbionts can provide resistance to parasitic infections. Both infection and populations of symbionts can be shaped by the host body temperature. However, the effects of temperature on the antiparasitic activities of gut symbionts have seldom been explored. The Lactobacillus-rich gut microbiota of facultatively endothermic honey bees is subject to seasonal and ontogenetic changes in host temperature that could alter the effects of symbionts against parasites. We used cell cultures of a Lactobacillus symbiont and an important trypanosomatid gut parasite of honey bees to test the potential for temperature to shape parasite-symbiont interactions. We found that symbionts showed greater heat tolerance than parasites and chemically inhibited parasite growth via production of acids. Acceleration of symbiont growth and acid production at high temperatures resulted in progressively stronger antiparasitic effects across a temperature range typical of bee colonies. Consequently, the presence of symbionts reduced both the peak growth rate and heat tolerance of parasites. Substantial changes in parasite-symbiont interactions were evident over a temperature breadth that parallels changes in diverse animals exhibiting infection-related fevers and the amplitude of circadian temperature variation typical of endothermic birds and mammals, implying the frequent potential for temperature to alter symbiont-mediated resistance to parasites in endo- and ectothermic hosts. Results suggest that the endothermic behavior of honey bees could enhance the impacts of gut symbionts on parasites, implicating thermoregulation as a reinforcer of core symbioses and possibly microbiome-mediated antiparasitic defense.

IMPORTANCE

Two factors that shape the resistance of animals to infection are body temperature and gut microbiota. However, temperature can also alter interactions among microbes, raising the question of whether and how temperature changes the antiparasitic effects of gut microbiota. Honey bees are agriculturally important hosts of diverse parasites and infection-mitigating gut microbes. They can also socially regulate their body temperatures to an extent unusual for an insect. We show that high temperatures found in honey bee colonies augment the ability of a gut bacterial symbiont to inhibit the growth of a common bee parasite, reducing the parasite’s ability to grow at high temperatures. This suggests that fluctuations in colony and body temperatures across life stages and seasons could alter the protective value of bees’ gut microbiota against parasites, and that temperature-driven changes in gut microbiota could be an underappreciated mechanism by which temperature—including endothermy and fever—alters animal infection.

INTRODUCTION

Temperature has a fundamental effect on biological rates that influence organismal physiology and interspecific interactions (1 – 4), including infection of hosts by parasites (5 – 8). The influence of temperature on infection outcome reflects not only the direct effects of temperature on parasite performance but also parasite performance relative to that of the host and its immune system (6, 8). In many hosts, attaining high internal temperatures can reduce the intensity and effects of infection (9 – 13), whether due to the direct inactivation of parasites or the potentiation of host immunity (10, 14 – 16). Such temperature-mediated increases in host resistance to infection could favor the evolution of energetically expensive, endothermic life history strategies, often characterized by sustained high body temperatures (17).

In addition to the defenses of the host itself, parasites that establish in the gut are influenced by cooccurring species that form the host gut microbiota and are likewise influenced by host temperature. The community of non-pathogenic microorganisms (hereafter referred to as “symbionts,” although this term sensu stricto includes all host-associated species) in the host gut can substantially influence infection outcomes via physical, chemical, and host immunity-mediated effects (18). A growing number of studies have indicated that temperature influences the composition of the gut symbiont community in ways that are qualitatively similar across distantly related hosts (19). However, despite the known effects of temperature on microbial community composition, biochemistry, and the microbiome and parasitic infection of hosts, the ways in which temperature alters the effects of gut symbionts on parasites have received little attention (19). A more complete knowledge of this area would elucidate how climate and host thermoregulation, including the constitutively high body temperatures of endotherms and infection-associated elevations in body temperature known as “fever,” could affect parasitism via symbiont-driven effects. Although variation in the growth rates and effects of symbionts over the wide range of temperatures experienced by ectothermic organisms in temperate climates seems likely, the effects of temperature changes experienced by more strongly homeothermic hosts require further investigation.

Honey bees and their associated microbiota offer an excellent opportunity to study the joint effects of temperature and gut symbionts on parasites. Honey bees are facultatively endothermic, capable of sustaining body and hive temperatures that are similar to those of mammals. Like other insects, however, they are subject to wide temperature variations during outdoor foraging and while overwintering (20, 21). Honey bees also possess a well-conserved, culturable gut bacterial community (22, 23) that can improve host resistance to parasites (24), which can threaten colony health (25 – 27). The most abundant of these bacterial symbionts are the Lactobacillus “Firm-5” species, which are among the major members of the ileum and account for the majority of bacteria in the rectum (28, 29). Colonization by Firm-5 reproduces many of the metabolomic changes found in bees with an intact microbial community (30), and can support host carbohydrate metabolism, sequester potentially toxic metals, and improve immune function and resistance to opportunistic pathogens (31). More generally, many Lactobacillus species can inhibit the growth of pathogens (32), and some—including L. delbrueckii, a phylogenetic relative of Firm-5 (33)—thrive at high (>40°C) temperatures (34), making this genus well-suited to study of how endothermy-enabled temperatures alter the antiparasitic activity of symbionts.

Trypanosomatid parasites of the honey bee hindgut, namely Crithidia mellificae (35) and the currently dominant Lotmaria passim (36), are prevalent, widespread, and in some cases associated with shortened bee lifespan and colony losses (26, 37 – 43). Both Crithidia and Lotmaria are extracellular parasites that colonize the gut epithelium of the ileum and rectum, forming layers of adherent cells that carpet the gut wall but do not result in histologically evident damage to the host (35, 36, 44). These honey bee-associated species are close relatives of the bumble bee parasite Crithidia bombi. Crithidia bombi infection is profoundly affected by the bumble bee gut microbiota (45, 46), which is taxonomically similar to that of honey bees (47). This includes negative correlations with the abundance of Lactobacillus Firm-5 clade members (46, 48). Members of this clade colonize the hindgut of honey bees as well (28, 30, 47, 49), where they are one of the major groups found in the ileum and dominate the community of the rectum (29, 30, 50), and are therefore likely to interact directly with trypanosomatid parasites in these gut regions. Bumble bee-associated Lactobacillus symbionts exerted temperature-dependent, acidity-driven effects on C. bombi growth in cell cultures (51, 52), and could contribute to the reduction of infection seen in bees reared at high temperatures [37°C (53)]. The gut microbiota of other insects can also influence infection with trypanosomatids, including species of medical and veterinary importance (54). However, the effects of the honey bee microbiome, including Lactobacilli, on trypanosomatid infection remain unclear. The only work thus far that tested the effects of gut microbial manipulations on infection found that pre-treatment of worker bees with another symbiont, Snodgrassella alvi, resulted in greater intensities of L. passim infection (55).

To evaluate how temperature affects the growth rates of honey bee Lactobacillus symbionts relative to trypanosomatid parasites and the effects of symbionts on parasite growth, we conducted three sets of experiments to: (i) compare the temperature dependence of parasite and symbiont growth rates (ii); test for effects of symbionts and the acids they produce on parasite growth; and (iii) determine how the effects of symbionts vary with temperature and alter the thermal niche of parasites. Our results show that the high colony temperatures produced by the endothermic behavior of bees can selectively favor the growth of symbionts and their inhibition of parasites, such that the presence of symbionts effectively reduces parasite heat tolerance. The effects of symbionts changed dramatically over a narrow, biologically significant temperature window, indicating that small changes in temperature similar to those shown by fever-expressing hosts can have large effects on the growth rates and parasite-inhibiting potential of symbionts.

MATERIALS AND METHODS

Cell cultures

Two strains of Lactobacillus symbionts, Lactobacillus helsingborgensis (strain WkB8) and Lactobacillus kullabergensis [strain WkB10, taxonomic classification inferred from reference (56)], kindly provided by N.J. Moran and J.E. Powell, were grown in De Man, Rogosa and Sharpe (MRS) media supplemented with 0.05% cysteine in screw-cap microcentrifuge tubes at 37°C (22), which supported growth without supplemental CO₂. These isolates are hereafter referred to by full genus and strain names, for consistency with their original description and to differentiate their genus from Lotmaria. The honey bee trypanosomatid parasites C. mellificae [ATCC 30254 (35)] and L. passim [strain BRL (36)] were obtained from the American Type Culture Collection and R.S. Schwarz. Trypanosomatids were grown in “FPFB” medium including 10% heat-inactivated fetal bovine serum [pH 5.9–6.0 (57)] at 28°C in vented cell culture flasks. Cultures were transferred to fresh media every 2 days.

Effects of temperature on symbiont versus parasite growth rate

Dense cultures [net optical density (OD) ~0.8] were diluted to a starting OD of 0.010 in fresh media and incubated in microcentrifuge tubes at 3°C increments between 25°C and 46°C. Although slightly higher absolute growth rates might be achieved for cells taken directly from the logarithmic phase of growth, dense cultures were necessary to enable sufficient dilution for all replicates across the range of incubation temperatures to originate from the same source tube, and to avoid rapid growth during the setup of the experiments. We ran six experimental blocks, each of which used parallel incubations at seven different temperatures (three blocks in 3°C increments from 25°C to 43°C and three blocks from 28°C to 46°C). One replicate tube per strain and temperature was destructively sampled at each of two time points (18 and 24 h). Growth rates were calculated as the slope of the curve of ln(OD) versus time using the 18 h time point only, at which point we could be confident that cultures were still in the logarithmic phase of growth. For comparison, we used previously reported growth rates of the two honey bee trypanosomatids (58), which included two replicate runs per temperature at 20°C and in 2°C increments from 23°C to 41°C, plus a third replicate for each of four temperatures (25°C, 31°C, 33°C, and 35°C).

We modeled the temperature dependence of growth rate for each Lactobacillus and trypanosomatid strain using a Sharpe-Schoolfield equation modified for high temperatures (15, 59, 60).

$${\displaystyle rate(T)=\frac{r_{T_{ref}\cdot }{e}^{\frac{-E}{k}\left(\frac{1}{T}-\frac{1}{T_{ref}}\right)}}{1+e^{\frac{E_h}{k}\left(\frac{1}{T_h}-\frac{1}{T}\right)}}}$$ (1)

In Equation (1), rate refers to the maximum specific growth rate [in (h⁻¹)]; r_Tref is the growth rate [in (h⁻¹)] at the calibration temperature T_ref (293K, i.e., 20°C, chosen to be well below the temperature of peak rate); E is the activation energy (in eV), which primarily affects the upward-sloping portion of the thermal performance curve (i.e., responsiveness of growth to temperature) at suboptimal temperatures; k is Boltzmann’s constant (8.62∙10⁻⁵ eV∙K⁻¹); E_h is the deactivation energy (in eV), which determines how rapidly the thermal performance curve decreases at temperatures above the temperature of peak growth T_pk (in K); T_h is the high temperature (in K) at which growth rate is reduced by 50% (relative to the value predicted by the Arrhenius equation—which assumes a monotonic, temperature-dependent increase) (60); and T is the experimental incubation temperature (in K). Four models were fit, one for each strain of Lactobacillus and each trypanosomatid (58).

Models were optimized using nonlinear least squares, implemented with R packages rTPC and nls.multstart (61). Confidence intervals (CIs) on parameter values and predicted growth rates were obtained by bootstrap resampling of the residuals [10,000 model iterations, R package “car” (62)]. Expected ratios between growth rates of symbionts and parasites were calculated by dividing the average of the predicted growth rates for the two Lactobacillus strains by the predicted rates for each of the trypanosomatid species at the corresponding temperature. We also used the bootstrap model predictions to estimate the following traits not explicitly included in the model: peak growth rate; temperatures of peak growth rate (T_pk ), and 50% inhibition relative to the peak value due to low and high temperatures (referred to as “cold tolerance” and “heat tolerance” in figures); and thermal niche breadth (i.e., the number of degrees between the low and high temperatures of 50% rate inhibition). The 0.025 and 0.975 quantiles for parameter estimates, predicted growth rates at each temperature, and traits derived from bootstrap predictions were used to define 95% confidence intervals. Estimates from different models were considered significantly different from each other when their 95% confidence intervals did not overlap. This and all subsequent analyses were conducted using R for Windows v4.0.3 (63).

Effects of symbiont spent medium and acidity on parasite growth

To generate spent media, each of the two Lactobacillus strains was grown in 15 mL conical centrifuge tubes for 2 days at 37°C from a 100× dilution of stationary phase cultures. The resulting supernatant (pH 4.6) was sterile-filtered, and an aliquot was neutralized with 1 M NaOH to the original pH of the MRS media (pH 5.6). Meanwhile, an aliquot of fresh MRS media was acidified with glacial acetic acid [the main acid in the honey bee gut (64) and of isolated Lactobacillus strains WkB8 and WkB10 under aerobic conditions (64)] to the pH of the spent media (pH 4.6). Each of the two trypanosomatids was then grown in a mixed growth medium consisting of equal volumes of fresh trypanosomatid-specific “FPFB” media and one of six Lactobacillus-specific MRS media-based treatments, originating from a single batch of spent media for each experiment: spent or neutralized media from each of the two Lactobacillus strains, acidified MRS media, or a fresh MRS media control. The initial pH of treatments after the addition of an equal volume of fresh FPFB media was 4.7–4.8 for the Lactobacillus spent media, as opposed to pH 5.7–5.8 for the neutralized spent media and fresh MRS media controls. The spent media treatments tested for symbiont-mediated inhibition of parasite growth; the acidified fresh media and neutralized spent media treatments indicated the extent to which this inhibition was dependent on pH.

To assess parasite growth, cell cultures of each trypanosomatid were diluted to a net OD of 0.040 in the appropriate mixed media. Growth of n = 6 replicate wells per treatment was measured in 96-well plates incubated in a plate reader spectrophotometer at 31°C. Plates were covered with low-evaporation lids to minimize inhibition due to evaporative concentration of the media. We included cell-free negative controls for each of the MRS media treatments, to control for changes in OD of media independent of parasite growth. OD values were corrected by subtraction of the OD of the corresponding cell-free media treatment and time point. Growth was estimated by measurement of OD at 600 nm in 5 min intervals for 24 h, with a 30-s shake before each read. Growth rate was calculated as the maximum slope of ln(OD) versus time over a rolling 4 h window, after exclusion of the first 3 h to allow OD to stabilize (15, 58, 65). Samples for which the r-squared value for the growth rate fell below 0.9 (vs >0.98 for all non-acidified samples) were considered to have a negligible growth rate and were assigned a rate of 0. Visual inspection of growth curves indicated that none of these samples exceeded a net OD of 0.050.

Temperature-dependent effects of symbionts on parasite growth in parasite-symbiont cocultures

To evaluate how temperature modulates the ability of symbionts to inhibit parasites, we grew L. passim in the presence and absence of Lactobacillus strain wkB10 at temperatures between 20.7°C and 38°C, which spans the range of temperatures commonly observed in worker bees within the winter cluster of colonies in a temperate climate (21). We focused on L. passim for these experiments because it is considered the most prevalent honey bee trypanosomatid globally (36, 40, 66) and showed an average growth rate comparable to that of the Lactobacillus strains over this range of colony temperatures. A dense culture of L. passim, grown 2 days at 28°C from a 25-fold dilution, was diluted to a net OD of 0.020 in fresh FPFB media. A dense culture of Lactobacillus strain wkB10, grown 2 days at 37°C from a 100-fold dilution, was diluted to a net OD of 0.040 in fresh MRS media. The L. passim cell suspension was then mixed with an equal volume of either the Lactobacillus cell suspension (“Lactobacillus present” treatment, initial net OD of 0.010 for the parasite and 0.020 for the symbiont) or fresh MRS without cells (Lactobacillus absent). These initial densities were chosen such that L. passim would remain in log-phase growth throughout the experiment, while Lactobacillus would approach the stationary phase at higher temperatures.

Two replicate microcentrifuge tubes, each containing 1,400 µL of media, were incubated in parallel for 23 h at each of seven different temperatures; two additional tubes were refrigerated at ~0°C immediately following setup for measurement of initial cell densities and media pH. Following the incubation, parasite densities were quantified by microscopic cell counts using a Neubauer hemocytometer at 400× magnification, using a 100-µL aliquot of each tube diluted with an equal volume of 50% glycerol to slow the movement of the parasite cells. Parasite growth rates were estimated based on the ratio of parasite cell densities at 23 h versus 0 h of incubation. The remaining sample volume was used for measurement of pH, determined to the nearest 0.01 pH units by immersing the probe of a calibrated Mettler-Toledo “Seven Easy” meter (Mettler-Toledo, Columbus, OH, USA) into the media following transfer into a 15-mL conical vial. The experiment was repeated three times for a total of six replicate samples per temperature and Lactobacillus treatment.

To quantify the temperature dependence of parasite growth rate, separate Sharpe-Schoolfield models were fit to the data for parasite growth rate in the presence and absence of Lactobacillus, using the approach described above (Effects of temperature on symbiont versus parasite growth rate section). To compute the proportion of parasite growth inhibition by Lactobacillus, we averaged the growth rates of the two Lactobacillus-containing and Lactobacillus-free replicates within each temperature and experimental block. Percent inhibition was then computed from the ratios of growth rates in the presence versus the absence of Lactobacillus for each temperature-block combination.

We modeled the correlations between temperature and parasite inhibition, temperature and pH, and pH and parasite inhibition using generalized linear models (67). Each model included experiment block as a random effect. In addition, the models for the effects of temperature and of pH on parasite inhibition included quadratic predictor terms [i.e., (temperature)², (pH)²] to better describe the curvilinear relationship between these variables that was apparent from visual inspection of the data. Models were fit using R package glmmTMB (67), with predictions generated with R package emmeans (68). Graphs were produced using R packages ggplot2 and cowplot (69, 70).

RESULTS

Effects of temperature on growth rates

Relative to the trypanosomatid parasites measured in our previous experiments (58), growth of both strains of Lactobacillus symbionts increased more strongly with temperature and peaked at higher temperatures, resulting in progressively higher projected growth rates of the symbionts relative to the parasites—particularly L. passim—as temperatures increased (Fig. 1). Lactobacillus growth rates increased by six- to sevenfold over the 15°C temperature window between 25°C and 40°C, whereas growth rates of trypanosomatids varied by <2-fold over a similar 15°C interval from 20°C to 35°C. These differences in temperature responsiveness were reflected in the estimates for activation energy (model parameter E), which was more than two times as high in both strains of Lactobacillus than in either species of trypanosomatid (Fig. 2; see Fig. S1 for additional parameters). In addition, Lactobacillus growth rates peaked at temperatures more than 4°C higher than those of C. mellificae and more than 6°C higher than those of L. passim (Fig. 2). The more steeply angled, high temperature-shifted thermal performance curves of the two Lactobacilli were likewise reflected by their higher temperatures of cold- and heat-related 50% growth inhibition and narrower thermal breadth (i.e., width of the thermal interval over which growth rates exceeded 50% of the peak value, Fig. S2).

Fig 1.

Growth of symbiotic Lactobacillus strains WkB8 and WkB10 responds more strongly to temperature and persists at higher temperatures than growth of the trypanosomatid parasites Crithidia mellificae and Lotmaria passim. (A) Growth rates of Lactobacillus strains wkB8 (solid lines) and wkB10 (dashed lines) between 25°C and 46°C. Points show observed rates. Lines and shaded bands show Sharpe-Schoolfield model predictions and 95% bootstrap confidence intervals. (B) Sharpe-Schoolfield models for temperature-dependent growth of C. mellificae (solid lines) and L. passim (dashed lines), as reported previously (58). (C) Projected growth rates of Lactobacillus (average of strains wkB8 and wkB10) relative to C. mellificae (solid lines) and L. passim (dashed lines) between 25°C and 41°C. Yellow shaded region represents temperature range at center of brood-rearing honey bee colony [33.8–37°C (20)]. Models are based on 41 rate measurements per species of Lactobacillus and 26 measurements per species of trypanosomatid.

Fig 2.

Growth rates of symbiotic Lactobacilli (upper panels) were more responsive to temperature and peaked at higher temperatures than did growth of trypanosomatid gut parasites (lower panels). Left panels show Sharpe-Schoolfield model parameter E, a measure of how strongly growth rate increases with temperature at temperatures below the temperature of peak growth (right panels). Points and error bars show bootstrap medians and 95% confidence intervals.

Based on model predictions, the expected growth rates of symbionts relative to trypanosomatids increased with temperature across the 25–37°C interval that is most common in honey bee workers (21), changing by 4.67-fold for Lactobacillus (averaged across the two strains) relative to C. mellificae and 9.25-fold relative to L. passim (Fig. 1). Even over the narrow 33.8–37°C temperature range of the central honey bee brood cluster (20), growth rates of Lactobacillus increased by 65% relative to C. mellificae and by more than threefold relative to L. passim (Fig. 1).

Acidity-mediated inhibitory effects of Lactobacillus spent media

Spent media from each of the two Lactobacillus strains resulted in full inhibition of both C. mellificae and L. passim (Fig. 3). Acidification of fresh media to an equivalent pH with acetic acid likewise extinguished growth, whereas neutralization of the spent media restored 89–90% of growth in C. mellificae and 81–82% in L. passim, indicating that acidity was necessary and sufficient to account for most of the observed inhibition by spent media (Fig. 3).

Fig 3.

Acidity-mediated inhibition of Crithidia mellificae and Lotmaria passim growth by Lactobacillus spent media. Growth rates of C. mellificae (upper panels) and L. passim (lower panels) in spent media (A–C) of Lactobacillus strains wkB8 (left) and wkB10 (center) before and after neutralization from pH 4.6 to 5.6 with 1 M NaOH, and in fresh Lactobacillus media (right) with and without acidification from pH 5.6 to 4.6 with acetic acid. Boxplots show medians and interquartile range for n = 6 samples per group. Points show individual observations, horizontally offset to mitigate overplotting.

Temperature-dependent effects of Lactobacillus on Lotmaria passim growth in parasite-symbiont cocultures

In parasite-symbiont cocultures, the presence of Lactobacillus altered the thermal niche of L. passim by inhibiting the parasite’s growth in a temperature-dependent fashion (Fig. 4). The 95% confidence intervals for model-predicted parasite growth rates in the presence and absence of Lactobacillus overlapped below 23°C, but the symbiont’s presence had increasingly pronounced inhibitory effects as temperature increased (Fig. 4). These effects resulted in a 24% lower predicted peak parasite growth rate and a 1.9°C reduction in the temperature at which parasite growth declined to less than 50% of peak rate, from 34.68°C in the absence of Lactobacillus (95% CI: 34.16–35.55°C) to 32.76°C in its presence (95% CI: 32.06–33.46°C) (Fig. 5). Confidence intervals for other parameters and traits overlapped in the two Lactobacillus treatments (Fig. S3 and S4).

Fig 4.

Coculture with Lactobacillus affects Lotmaria passim growth in a temperature-dependent manner, with stronger effects of the symbiont at higher temperatures. Lines and shaded bands show Sharpe-Schoolfield model fits and 95% confidence intervals for L. passim growth rate in the presence (solid lines) and absence (dashed lines) of Lactobacillus. Points show individual observations; circles and triangles indicate the presence and absence of Lactobacillus, respectively. Models are based on 42 rate measurements per Lactobacillus treatment.

Fig 5.

The presence of Lactobacillus reduced peak Lotmaria passim growth rate (left) and lowered the temperature at which growth rate declined to <50% of peak value (right). Points and error bars show estimates and 95% bootstrap confidence intervals for traits derived from Sharpe-Schoolfield model fits shown in Fig. 4.

The inhibitory effects of Lactobacillus increased quadratically with temperature (temperature: coefficient −25.71 ± 8.82 SE, Z = −2.91, P = 0.004; Temperature²: coefficient 0.51 ± 0.15 SE, Z = 3.39, P < 0.001); inhibition exceeded 50% at 32.9°C (95% CI: 30.26–34.82°C, Fig. 6A). The final pH of the coculture decreased by 0.098 ± 0.0060 SE pH units per 1°C increase in temperature (Z = −16.38, P < 0.001, Fig. 6B). Across all temperatures, stronger inhibitory effects of Lactobacillus were quadratically correlated with lower final pH (pH: coefficient −707.75 ± 203.32 SE, Z = −3.48, P < 0.001; pH²: coefficient 65.60 ± 20.05 SE, Z = 3.27, P = 0.001), with 50% reduction in growth at a final pH of 4.69 (95% CI: 4.52–4.96, Fig. 6C).

Fig 6.

Temperature-dependent, pH-associated effects of Lactobacillus on growth of Lotmaria passim in symbiont-parasite cocultures. (A) Proportional inhibitory effects of Lactobacillus on L. passim increased with temperature. (B) Higher temperatures resulted in lower pH of the growth media. (C) Lower pH was correlated with greater proportional inhibition of L. passim growth. Lines and shaded bands show estimated marginal means and standard errors of linear model predictions. Points show individual observations. Shapes indicate different experimental blocks. The correlation between temperature and coculture pH (B) is based on 42 observations. Models for percent inhibition (A–C) are based on 21 observations due to averaging across the two replicates in each experimental block.

DISCUSSION

Due to the fundamental effects of temperature on biological rates, including parasite and symbiont growth and metabolism, temperature has the potential to alter the effects of symbionts on parasite establishment, such that both symbiont and parasite physiology shape the temperature dependence of infection. We found that high temperatures augmented growth rates of symbionts relative to parasites, that symbionts inhibited parasite growth via production of acids, and that high temperatures amplified this inhibitory effect over a narrow temperature window relevant to thermoregulation in honey bees and classically homeothermic organisms.

Symbionts were more heat-tolerant than parasites

Our finding of greater heat tolerance in the honey bee symbiont than in the trypanosomatids is consistent with differences in conventional culturing temperatures and prior investigations of bumble and honey bee-associated microbes. Most trypanosomatids are cultured at 25–28°C (71), whereas honey bee gut bacteria thrive at 35–37°C (22). The ~6°C lower peak growth temperature of L. passim relative to the Lactobacillus symbionts closely resembles the difference in peak temperatures for C. bombi (~34°C) and Lactobacillus bombicola (~40°C) from bumble bees (51). Other key symbionts of honey bees are also heat-tolerant and able to grow above 40°C (72). In our experiments, the Lactobacillus symbiont’s greater heat tolerance and temperature responsiveness led to rapid changes in the relative growth rates of symbionts versus parasites, especially the less heat-tolerant L. passim, over the temperature range found in colony-dwelling honey bees [20–37°C (21)]. This suggests that honey bees’ endothermic behavior could select for the growth of coevolved symbionts at the expense of opportunistic parasites. Although both trypanosomatids were still able to grow above 34°C, growth rates of Lactobacillus relative to the parasites (especially L. passim) increased particularly sharply over the 33.8–37°C range of the colony’s core during the brood-rearing season (20). This indicates that the warm temperatures at the center of the colony would favor the formation of a core symbiont-dominated microbiome during the microbiome’s formative period in the 2–4 days after adult bees emerge from pupation, including the primacy of Lactobacillus Firm-5 species in the rectum (28).

Symbiont-mediated reduction in pH inhibited parasite growth

Although growth rates depicted the suitability of the colony habitat for trypanosomatids relative to symbionts, the growth of symbionts is only directly relevant to the proliferation of parasites if there is some interaction between the two. Both tested Lactobacillus strains, as well as other Firm-5 isolates, produce acids from hexoses (64, 73). We found that Lactobacillus-spent medium inhibited the growth of both honey bee parasites in an acidity-dependent manner within the gut pH range observed in bees. Environmental pH is a strong driver of gut microbiome composition (74), where single-unit changes in pH alter the relative growth rates of cooccurring species and community-level production of short-chain fatty acids (75). Production of acids by fermentative bacteria, including Lactobacilli, contributes strongly to their chemical inhibition of gut parasites (32, 76). In humans, diets that acidify stool pH from 6.5 to 5.5 are associated with suppression of opportunistically pathogenic Enterobacteriaceae species, such as E. coli, that grow poorly in acidic environments (77). Similarly, supplementation of infants with milk sugar-fermenting Bifidobacteria reduces stool pH from 5.97 to 5.15 and reduces levels of Enterobacteriaceae, Clostridiaceae, and endotoxins (78), whereas reduced levels of Bifidobacteria have been associated with a 1.5-unit increase in fecal pH and increased abundance of Enterobacteriaceae, Clostridiaceae, and other dysbiosis-associated bacteria over the past century (79).

In the gut of honey bees, one effect of bacterial colonization is a nearly identical reduction in hindgut pH, from pH ~6 to pH ~5.2 (64). This is accompanied by the accumulation of organic acids that are absent from the guts of germ-free bees (64). Lactobacillus Firm-5 species are the most abundant members of the honey bee gut microbiome overall (29, 30) and the rectum specifically (29, 29, 80), and produce these acids when grown in culture (64). Hence, these symbionts are likely major drivers of acidification in the posterior ileum and rectum, where both C. mellificae and L. passim proliferate (35, 36), with L. passim specifically showing a particular affinity for the rectum (44).

Our findings of Lactobacillus acidity-mediated inhibition of honey bee trypanosomatids are consistent with pH-dependent inhibition of C. bombi by Lactobacillus bombicola from bumble bees (52), and match the negative correlations between C. bombi infection intensity and abundances of acid-producing Lactobacillus and Gilliamella in these hosts (46, 48). Our results also show a strong correlation between pH and L. passim growth in cocultures, with inhibition occurring over a physiologically relevant range. Given the similar effects of different sources of acidity, including both lactic and acetic acids, on the related C. bombi (52), it seems plausible that multiple acids in the honey bee gut could contribute to such pH-mediated inhibition. The rectal pH of adult bees in colonies ranged between 4.2 and 6.0 (81). Whereas pre-colonization of bees with S. alvi—which consumes organic acids (30)—increased L. passim infection (55), we found that an endpoint coculture pH below 5 was associated with sharply reduced L. passim growth rates, suggesting that microbial or climatic factors that acidify gut pH to the lower end of this observed range would improve resistance to infection. Such acidity-driven pressure on parasites could be one explanation for honey bee parasites’ high tolerance of acidity relative to trypanosomatids from insects with less acidic guts (58).

Besides their effects on pH, bacterial symbionts could also influence parasite establishment in the bee gut via other mechanisms. For example, symbionts reduce rectum concentrations of glucose by sixfold relative to germ-free bees (64). Sugar concentrations can be limiting for growth of C. bombi (82), and glucose is used preferentially as a carbon source by many insect-stage trypanosomatids (83). Symbionts capable of aerobic respiration, such as S. alvi (64), may additionally deplete gut oxygen, creating anoxic conditions under which trypanosomatids cannot survive (83). Furthermore, gut bacteria that metabolize dietary phytochemicals, such as flavonoids (30, 84), could alter the bioavailability and activity of these compounds, including those that inhibit trypanosomatid growth (65, 85, 86). Alternatively, biofilm-forming symbionts can physically compete with parasites for space along the gut wall (87), which could prevent the attachment of trypanosomatids to the gut epithelium (44). Finally, colonization by symbionts, including Lactobacillus Firm5 species, can also stimulate the immune responses of the host and improve the clearance of gut pathogens (31, 88).

High temperatures amplified the parasite-inhibiting effects of symbionts

Our coculture experiments showed that the inhibitory effects of symbionts on parasites increased with temperature. Both Lactobacillus strains grew faster and produced acids more rapidly as temperatures increased over the range found in bee colonies. This high-temperature inhibition resulted in lower peak parasite growth rates and reduced parasite heat tolerance relative to when symbionts were absent. Hence, social bees’ endothermic maintenance of high colony temperatures could exert antiparasitic effects not only through direct parasite inhibition, but also via potentiation of the parasite-inhibiting effects of symbionts. The bumble bee symbiont Lactobacillus bombicola resulted in similar temperature-dependent inhibition of the parasite C. bombi, lowering the temperature of peak parasite growth rate and resulting in a more rapid decline in parasite growth rate as temperatures increased (51). Using a parasite-symbiont system from a related host, our results build on these findings by investigating a broader range and increased number of temperatures, enabling formal comparisons between parasite thermal performance curves in the presence versus absence of symbionts.

The manipulability of both microbiome and body temperature in a host with a predominantly endothermic life history strategy makes honey bees a useful system to explore the value of endothermy as a reinforcer of symbioses and symbiont-mediated defense against parasites. Prior work has shown that high (35°C) temperatures increase the resistance of honey bees to colonization by heat-sensitive environmental yeasts relative to cooler (29°C) temperatures, consistent with the seasonal appearance of yeasts during colder weather in field-collected forager bees (89). On the other hand, high (35°C) temperatures enhanced the growth of and gut colonization by the core symbiont S. alvi in bumble bees relative to low (28–29°C) temperatures (72). These results implicate bee endothermy in both resistance to opportunistic gut microbes and establishment of key symbionts.

Our results indicate that higher temperatures not only promote symbiont growth, but also augment symbionts’ value as a defense against gut parasites. Consistent with this idea, honey bees collected from cooler winter colonies exhibit gut colonization by non-core microbial taxa in addition to—rather than in place of—their pre-existing core symbionts (49). This observation suggests that the core symbiont community might require high temperatures to resist invasion even when its major members, including Lactobacillus Firm-5, are abundant (49), although differences in bee age and diet likely contribute to the pattern as well. Temperature-mediated inhibition of trypanosomatids specifically is of practical importance for honey bees, where infection intensities were highest mid-winter in one longitudinal survey (43) and correlated with winter colony collapse in another (26). Many aspects of bee biology—including bee diet, age, and foraging activity—differ between summer and winter, and to our knowledge, the effects of temperature on trypanosomatid infection have yet to be tested experimentally. Nevertheless, our results suggest that environmental factors that perturb thermoregulation, such as exposure to pesticides, agricultural landscapes, and food shortage (21, 90 – 92), could exacerbate this seasonal susceptibility. Our results indicate that parasite-symbiont interactions can change substantially even over the narrow (~4°C) temperature range within which developing honey bees are incubated (20), a variation that can nevertheless influence bee resistance to other pathogens (93). We found a threefold increase in the growth rates of symbionts relative to parasites over this range, accompanied by a doubling of antiparasitic activity, implying that the thermal environment under which bees emerge could determine the ability of parasites to colonize the gut during formation of the symbiont community (80).

Beyond honey bees, our results show how parasite-symbiont interactions can shift even over a relatively narrow span of temperatures that parallels changes seen in diverse fever-expressing hosts. Parasite-exposed insects that exhibited behavioral fever showed 2–6°C increases in preferred body temperature (11). Virus-challenged zebrafish showed similar 3°C changes in preferred water temperature (16). Among endotherms, bacteria-challenged rabbits experienced 3°C increases, similar to those found in goldfish and iguanas (94). Even in the absence of febrile behavior, a 3–5°C range in temperature was sufficient to explain intraspecific variation in resistance to chytrid infection in frogs (95). Moreover, temperature changes of this magnitude are within the amplitude of circadian temperature oscillations found in many birds and mammals. Fluctuations of 3°C occur even among strongly homeothermic species (96, 97), with changes of <5°C considered indistinguishable from normal diel variation (98) and drops of 20–30°C frequent in torpor-expressing species (99). Hence, temperature-mediated alteration of symbiont-driven resistance to parasites has the potential to occur frequently in both endo- and ectothermic hosts. The presence of the appropriate gut symbionts, including Lactobacillus species, could amplify the effects of these thermal disturbances and reduce the critical temperatures needed to inhibit parasite growth.

Conclusions

The presence of parasites can shift the thermal niche of hosts towards temperatures that enhance escape from or suppression of infection (15). High body temperatures often mitigate infection of warm-adapted hosts, including honey bees and other insects (10, 93, 100 – 102), with less heat-adapted parasites (8). Conversely, the host’s responses to temperature can alter the optimal temperature for parasite reproduction (5, 8). Our findings complement these results, showing how the presence of gut symbionts can limit both the peak growth rate and heat tolerance of a globally prevalent honey bee gut parasite. The extent of symbiont-mediated inhibition varied substantially over the range of temperatures and gut pH levels previously reported in adult bees, indicating the potential for endothermic behavior and core gut symbionts of honey bees to synergistically inhibit parasite growth. Our finding that high temperatures promote symbiont-mediated parasite inhibition shows how changes in host temperature could augment or diminish the protective effects of the microbiome against infection. This illustrates an interaction between the well-established effects of temperature and gut microbial communities on resistance to parasites, with implications for understanding the effects of climate on infection of ectotherms and the evolution of endothermy and fever. The ability of symbionts to shift the thermal niche of parasites highlights the value of investigating parasite thermotolerance in the context of the host microbiota.

DATA AVAILABILITY

All data are supplied in the Supplementary Information, Data S1.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01023-23.

Data Set S1. Growth rates. aem.01023-23-s0001.csv.

Growth rates of Lactobacillus and trypanosomatids in isolation.

Data Set S2. Spent media data. aem.01023-23-s0002.csv.

Growth of trypanosomatids with Lactobacillus spent medium.

Data Set S3. Coculture data. aem.01023-23-s0003.csv.

Growth of L. passim in cocultures.

Data Set S4. Model summaries. aem.01023-23-s0004.csv.

Sharpe-Schoolfield models for Lactobacillus and trypanosomatids in isolation.

Data Set S5. Coculture model summaries. aem.01023-23-s0005.csv.

Sharpe-Schoolfield models for Lactobacillus and trypanosomatid cocultures.

Figures S1 to S4. aem.01023-23-s0006.pdf.

Supplementary figures and description of data files.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.