Temperature Affects the Host Range of Rhabdochlamydia porcellionis

ABSTRACT The Rhabdochlamydiaceae family is a recent addition to the Chlamydiae phylum. Its members were discovered in cockroaches and woodlice, but recent metagenomics surveys demonstrated the widespread distribution of this family in the environment. It was, moreover, estimated to be the largest family of the Chlamydiae phylum based on the diversity of its 16S rRNA encoding gene. Unlike most Chlamydia-like organisms, no Rhabdochlamydiaceae member could be cultivated in amoebae, and its host range remains unknown. We tested the permissivity of various mammalian and arthropod cell lines to determine the host range of Rhabdochlamydia porcellionis, the only cultured representative of this family. While growth could initially be obtained only in the Sf9 cell line, lowering the incubation temperature of the mammalian cells from 37°C to 28°C allowed the growth of R. porcellionis. Furthermore, a 6-h exposure to 37°C was sufficient to irreversibly block the replication of R. porcellionis, suggesting that this bacterium either lost or never acquired the ability to grow at 37°C. We next sought to determine if temperature would also affect the infectivity of elementary bodies. Although we could not purify enough bacteria to reach a conclusive result for R. porcellionis, our experiment showed that the elementary bodies of Chlamydia trachomatis and Waddlia chondrophila lose their infectivity faster at 37°C than at room temperature. Our results demonstrate that members of the Chlamydiae phylum adapt to the temperature of their host organism and that this adaptation can in turn restrict their host range. IMPORTANCE The Rhabdochlamydiaceae family is part of the Chlamydiae, a phylum of bacteria that includes obligate intracellular bacteria sharing the same biphasic developmental cycle. This family has been shown to be highly prevalent in the environment, particularly in freshwater and soil, and despite being estimated to be the largest family in the Chlamydiae phylum is only poorly studied. Members of the Rhabdochlamydiaceae have been detected in various arthropods like ticks, spiders, cockroaches, and woodlice, but the full host range of this family is currently unknown. In this study, we showed that R. porcellionis, the only cultured representative of the Rhabdochlamydiaceae family, cannot grow at 37°C and is quickly inactivated at this temperature. A similar temperature sensitivity was also observed for elementary bodies of chlamydial species adapted to mammals. Our work demonstrates that chlamydiae adapt to the temperature of their reservoir, making a jump between species with different body temperatures unlikely.

visible inclusion ( Fig. 2A to C). Interestingly, unlike S. negevensis, R. porcellionis could not grow in mammalian cells and, more surprisingly, failed to grow in all arthropod cell lines except Sf9 cells (Fig. 1A). While bacteria were internalized in all cells tested (Fig. 1B), they either failed to initiate replication or formed aberrant bodies measuring more than 5 mm (in C6/36 cells).
R. porcellionis is unable to replicate at 37°C. Rhabdochlamydiaceae have been detected in various arthropods such as ticks (34,41), spiders (42), cockroaches (28), and woodlice (29). As those organisms are poikilothermic and have a lower body temperature than mammals (43,44), we reasoned that the Rhabdochlamydiaceae family might have adapted to the lower temperature of its host organisms and either lost or never acquired the ability to grow at 37°C. To test this hypothesis, we infected Sf9 cells with R. porcellionis and incubated them at 20, 28, 33, and 37°C. R. porcellionis grew at 20 and 28°C but not at 33 or 37°C (Fig. 3A). The lack of growth at 33 or 37°C could, however, be due to the loss of permissivity of Sf9 cells at those temperatures. As a control for this, we infected Sf9 cells Applied and Environmental Microbiology with W. chondrophila, another chlamydia-like organism known to have a wide host range and to grow at both 28 and 37°C (21,26,27), and incubated them at 28 and 37°C. Unlike R. porcellionis, W. chondrophila replicated at both temperatures (see Fig. S1 in the supplemental material), suggesting that Sf9 cells retain their permissivity at 37°C. This, however, does not exclude the possibility of Sf9 cells selectively losing their permissivity to R. porcellionis at 37°C. To test if the absence of growth in mammalian cells was also an effect of temperature, we infected A549, Ishikawa, and McCoy cells with R. porcellionis and lowered the incubation temperature to 28°C. This change indeed allowed the growth of R. porcellionis (Fig. 3B) in mammalian cells, although the doubling time appeared to be longer and more variable than that in Sf9 cells (Fig. 3C). The long doubling time (Fig. 3C), the absence of intermediate bodies, and the distorted appearance of the bacteria in McCoy cells (Fig. S2) indicate that those cells might be too different from the natural host of R. porcellionis to allow efficient growth.
We finally compared the infection efficiencies of R. porcellionis in A549, McCoy, Ishikawa, and Sf9 cells by measuring the inclusion-forming unit (IFU) count at 6 days postinfection. As shown in Fig. 3D, the IFU count was significantly lower in Ishikawa and A549 cells than in Sf9 cells. Surprisingly, there was no difference between McCoy and Sf9 cells. This might, however, be due to an underestimation of the IFU count in the latter, as infected Sf9 cells tend to detach from glass coverslips. Transient exposure to 37°C irremediably blocks the replication of R. porcellionis. Members of the Chlamydiae are known to enter a third nonreplicative stage, the aberrant body, when exposed to stresses such as antibiotic exposure (45), nutrient deprivation (46), or heat shock (47). Aberrant bodies are typically described as nonreplicating enlarged cells able to resume their regular cycle once the stress disappears, although their morphology was shown to vary as a function of the stresses (48,49). We thus wondered whether R. porcellionis could similarly recover and resume its growth cycle after an exposure to 37°C. To determine this, Sf9 cells at 2 days postinfection were incubated at 37°C for various durations. The infected cells were then further incubated at 28°C for four additional days to check for bacterial growth after stress removal. The growth was assessed by measuring the number of genome copies at the end of the incubation at 37°C and after 4 days of recovery at 28°C (Fig. 4A).
Our experiment shows that a transient exposure to 37°C as short as 6 h irreversibly blocks the replication of R. porcellionis (Fig. 4B). The effect of the temperature shift seems to be more deleterious for longer incubation at 37°C, as expected. However, the complete experiment lasted from 6 days, for the samples subjected to the 6-h shift at 37°C, to 8 days, for the samples subjected to the 48-h shift at 37°C. This difference of duration could also explain the more pronounced effect of longer incubations. Indeed, the number of genome copies at the end of the shift to 37°C was not affected by the duration of the exposure (P value = 0.126, one-way analysis of variance [ANOVA]), showing that it takes at least 48 h for temperature to affect the number of genome copies. It is therefore possible that the apparent lack of difference between 96 h and 0 h for the 6-and 12-h time points is due to the experiment not being long enough for the effect of temperature to manifest in terms of genome copies. The delay might be due to DNA being slowly released-and degraded by cytosolic DNases-from inactivated bacteria.
Temperature affects the infectivity of elementary bodies. Given the effect of temperature on the replication of R. porcellionis, we wondered whether it would also affect the infectivity of elementary bodies. To determine this, we incubated elementary bodies of R. porcellionis at 20°C (room temperature [RT]) or 37°C and measured the number of IFU every day for 4 days. To be closer to natural conditions, we did not perform the freeze-and-thaw cycle and directly filtered the supernatant from infected Sf9 cells. In addition, as Sf9 cells tend to detach after several days of infection, we used McCoy cells grown at 28°C for IFU quantification of R. porcellionis to avoid any bias due to cell detachment. We used a random intercepts mixed-effects linear model to predict the log-transformed count of IFUs based on an interaction variable between the duration of incubation in days and the temperature. When none of the counted cells was infected, we conservatively assumed an IFU count corresponding to one infected cell in 100.
As shown in Fig. 5A, R. porcellionis elementary bodies (EBs) incubated for 24 h at 37°C were less infectious than their counterpart incubated at 20°C, although the incubation temperature did not significantly predict the IFU count (R 2 = 0.77, beta for the interaction term = 20.55, P value = 0.06). The interaction term implies that for every day of incubation at 37°C, the IFU count will be reduced by a factor of 3.55 (10 0.55 ) compared to the same incubation at room temperature. Due to the low initial quantity of infectious particles (Fig. 5B), no infected cell could be observed after the first time point. Using a freeze-and-thaw cycle would only marginally improve the IFU count ( Fig. 3D) and introduce an experimental bias, while resorting to bead beating to increase elementary body yield was shown to be counterproductive for R. porcellionis (40). In the absence of alternatives to obtain results for later time points, we did not repeat the experiment with these bacteria. We instead resorted to more tractable chlamydiae from two different families, W. chondrophila and C. trachomatis, to check whether exposure to different temperatures also affected their infectivity. As both bacteria infect mammals, we expected their EBs to be adapted to the temperature of their host and the number of IFU to decrease faster at 20°C This highlights the deleterious effect of an incubation at 37°C for the EBs of all three species. Interestingly, the inclusions formed by EBs incubated at 37°C also tended to be smaller. The difference between the well-formed inclusions for C. trachomatis and W. chondrophila and the dissemination of reticulate bodies in the host cytoplasm for R. porcellionis is also striking. Cells were fixed at 1 day (for C. trachomatis and W. chondrophila) or 6 days (for R. porcellionis) postinfection and stained with concanavalin A (red), DAPI (blue), and antibodies against the different bacteria (green). Bar, 10 mm. Incub., incubation duration. than at 37°C. This proved to be false, as the IFU count decreased faster at 37°C than at 20°C for both W. chondrophila (R 2 = 0.97, beta = 20.68, P value = 3.27 Â 10 211 ) and C. trachomatis (R 2 = 0.94, beta = 21.06, P value = 2.08 Â 10 210 ). Compared to an incubation at room temperature, the IFU count of C. trachomatis and W. chondrophila is thus predicted to be reduced by a factor of 11.48 and 4.79 for every day of incubation at 37°C, respectively. Altogether, these results suggest that while the reticulate bodies of the Chlamydiae have different thermal preferences, the elementary bodies of distant families share the same temperature sensitivity.

DISCUSSION
Several factors are known to influence the host range of the Chlamydiae. The ability to inhibit apoptosis has, for instance, been suggested to be a hallmark of chlamydiae infecting multicellular organisms (20,23), while chlamydiae adapted to multicellular organisms appear to have lost the ability to grow in amoebae (3,19). The adaptation to specific temperatures has already been demonstrated to be an important determinant for the host range of the members of the Parachlamydiaceae family (22,50,51). In the present study, we demonstrated that this extends to other families of the Chlamydiae phylum by showing that R. porcellionis is likely specialized for the temperature ranges encountered in Porcellio scaber (43,52). As a consequence of this specialization, R. porcellionis is quickly inactivated if exposed to the body temperature of mammals, making it unlikely for this bacterium to be pathogenic for mammals and restricting its host range to organisms with a temperature lower than 33°C (Fig. 3A). The use of different incubation temperatures should thus be considered in addition to the inclusion of diverse host cells (40) when attempting to isolate environmental chlamydiae. Indeed, independently of its ability to grow in amoebae, a chlamydial symbiont sharing the same temperature sensitivity as R. porcellionis would fail to grow in the Acanthamoeba castellanii subculture system frequently used for the isolation of chlamydiae from environmental samples (6,12,53).
The inability to formally exclude that our observations are due to the effect of temperature on the host cells alone is a key limitation of this study. Our experiments, however, make this hypothesis unlikely, as an effect of the temperature exclusively on the host cells would imply the rapid (Fig. 4B) loss of permissivity of Sf9 cells for R. porcellionis at 37°C but not for W. chondrophila (Fig. 3A and see Fig. S1 in the supplemental material), as well as the gain of permissivity of mammalian cells at 28°C (Fig. 3B). In addition, a previous work showed that insect cell lines do not lose their permissivity to intracellular bacteria and can notably be used for the culture of Mycoplasma species when grown at 37°C (54). Incubations at 37°C also do not seem to affect protein synthesis in Sf9 cells, and heat-tolerant cells growing at 37°C could even be obtained (55). The tolerance of Sf9 cells to high temperatures was hypothesized to be due to Spodoptera frugiperda being adapted to warm climates (55). Along with previous evidence showing a similar temperature sensitivity in species of the Parachlamydiaceae family (50,51), an effect of the temperature on the bacteria themselves thus appears to be the a more parsimonious explanation. Conversely, the effect of cold shock on mammalian cells remains poorly understood and has been studied essentially for its use in the production of recombinant proteins (56,57). The possibility that the growth of R. porcellionis in mammalian cells is due in part to an effect of the temperature on the host cell therefore cannot be excluded.
The attempt to assess whether elementary bodies share the temperature sensitivity of reticulate bodies was impaired by the difficulty of purifying enough R. porcellionis EBs. The unequivocal results obtained for two chlamydial species from different families and the trend observed for R. porcellionis (Fig. 5A), however, suggest that the elementary bodies of this bacterium also lose their infectivity faster at 37°C. Elementary bodies of different chlamydial species may therefore be more similar in terms of thermal preferences than their respective reticulate bodies (50,51), although this trend would need to be confirmed in other chlamydiae such as Simkania negevensis or in species of the Parachlamydiaceae family. Of note, our results also suggest that the better extracellular survival of Chlamydia-like organisms than of C. trachomatis observed in a previous study (58) may be due to differences in incubation temperatures rather than the metabolic capacity of those organisms.
The ability to grow in both insect and mammalian cells ( Fig. 1A and Fig. 3B) that we observed for R. porcellionis has also been reported for various Chlamydiaceae (59, 60) species and suggests that despite the long evolutionary distances between those organisms, their cells are still similar enough to allow the growth of bacteria from both families. The long doubling times and frequent aberrant bodies observed in mammalian cells may be due either to differences in the host cell physiology or to the repercussion of a lower fitness of the host cell on the bacteria. In contrast to the previous observation that R. porcellionis can grow in different insect cells (40), we could not observe growth in either the C6/36 or the IRE/CTVM19 cell line. The lack of growth in the latter cell line might be due to the specialization of R. porcellionis for isopods at the cost of the ability to grow in cells originating from the more distantly related arachnids. This explanation, however, fits poorly with the observation that R. porcellionis grows in mammalian cells. The permissivity of cell lines might therefore be more related to the cell type than to the host species, as suggested by the restricted tissue distribution of R. porcellionis and Rhabdochlamydia crassificans in their respective hosts (28,61).
It is unclear how the thermal preferences of R. porcellionis generalize to the other members of the Rhabdochlamydiaceae. The high predicted diversity of the Rhabdochlamydiaceae family (5, 42) and the demonstration that species of the same chlamydial family can have different thermal preferences (50,51) indeed imply the possibility that other rhabdochlamydiae could have adapted to various ranges of temperatures. However, the similar thermal preferences exhibited by ixodid ticks (44,62) and woodlice (43,52) suggest the possibility that the same evolutionary mechanism that drove the adaptation of R. porcellionis toward lower temperatures could have had the same effect in tick-borne rhabdochlamydiae (30,34,41). This will, however, remain purely speculative until the thermal preferences of additional rhabdochlamydial species can be assessed. Interestingly, a similar temperature sensitivity has also been reported for Wolbachia (63, 64), another arthropod symbiont whose host range broadly overlaps that of the Rhabdochlamydiaceae (65). Of note, Wolbachia appears to influence the thermal preferences of its host to favor its replication (66). A similar effect might also be found in R. porcellionis, as it could prevent the accidental cure of the host of the bacterial infection by a short exposure to temperatures higher than 33°C (Fig. 4B).
This work demonstrates the adaptation of R. porcellionis to the range of temperatures encountered in its host, at the cost of the ability to successfully infect other species with higher body temperatures. In particular, the temperature sensitivity of R. porcellionis precludes its transmission to mammals and excludes a pathogenic role of this bacterium for humans (36)(37)(38)67). It finally highlights the importance of testing different incubation temperatures when attempting to recover chlamydiae from environmental samples. Bacterial strains. The R. porcellionis strain was acquired from the DSMZ collection (DSM 27522) and cultivated in Sf9 cells. The infected cells were passaged once a week, and fresh cells were added approximately every four passages to compensate for host cell death due to the presence of the bacteria. Waddlia chondrophila strain WSU 86-1044 (ATCC VR-1470) was cocultivated with A. castellanii in PYG broth at 32°C. Suspensions of EBs were collected at 7 days postinfection, diluted 10 times, and used to infect fresh A. castellanii. Chlamydia trachomatis (ATCC VR-902B) was cultivated in McCoy cells incubated at 37°C in DMEM supplemented with 10% FCS and 1 mg mL 21 cycloheximide.

MATERIALS AND METHODS
Infection procedure. Cells were seeded in a 24-well plate (Corning) at a density of 1 Â 10 5 or 3 Â 10 5 cells per well 2 h before infection. The infection procedures were performed as described in previous publications (26,40).
For R. porcellionis, suspensions of infected Sf9 cells were subjected to a freeze-thaw cycle to disrupt the cells, followed by a filtration through a 5-mm-pore filter to remove the debris. Plated cells were then infected with the filtrate at a multiplicity of infection (MOI) of ;0.1 to 1 and centrifuged for 15 min at 130 Â g at room temperature, followed by an incubation of 30 min at 28°C. The medium was then replaced to remove noninternalized bacteria.
W. chondrophila EBs were collected from the supernatant of A. castellanii at 5 days postinfection. The supernatant was then filtered through a 5-mm-pore filter to remove cell debris. Plated cells were infected with the filtrate at an MOI of ;0.1 to 1 and centrifuged at 1,790 Â g for 10 min at room temperature. After 30 min of incubation at either 28°C (for Sf9 cells) or 37°C (for mammalian cells), the medium was replaced to remove noninternalized bacteria.
C. trachomatis EBs were collected from the supernatant of infected McCoy cells at 3 days postinfection. The supernatant was then filtered with a 5-mm-pore filter. Plated cells were infected with the filtrate and centrifuged at 900 Â g for 15 min at room temperature. The infected cells were then incubated for 30 min at 37°C, and the medium was replaced to remove noninternalized bacteria.
After the infection, the cells were incubated at their usual growth temperature, unless specified otherwise. The samples were collected at various time points for quantification by quantitative PCR (qPCR) and immunofluorescence staining. The MOI was estimated by measuring the proportion of infected cells with 10-fold serial dilutions of EB filtrate in immunofluorescence.
Inclusion-forming unit (IFU) quantification. Cells were plated in a 24-well plate at a density of 3 Â 10 5 cells per well 2 h before the infection and were then infected with serial 10-fold dilutions of EB suspensions. After the initial 30 min of incubation, the medium was replaced with fresh medium supplemented with 1 mg mL 21 of cycloheximide (69). The cells infected with W. chondrophila or C. trachomatis were incubated for 24 h at 37°C with 5% CO 2 , while the cells infected with R. porcellionis were incubated for 6 days at 28°C, with 5% CO 2 for mammalian cells. The cells were then fixed and stained for immunofluorescence, and the proportion of infected cells was determined using an epifluorescence microscope. At least 100 cells were counted for each condition.
Effect of temperature on EB infectivity. C. trachomatis and W. chondrophila EBs were collected from the supernatant of infected cells as in a standard infection procedure. To be closer to a natural infection, EBs of R. porcellionis were collected from the supernatant of infected cells, without any prior lysis step. The supernatants were filtered with a 5-mm-pore filter. The filtrate was then diluted 1:1 in PYG for W. chondrophila, 1:1 in DMEM supplemented with 10% FCS for C. trachomatis, and 1:1 in Grace medium supplemented with 10% FCS for R. porcellionis. The suspensions of elementary bodies were then incubated at either 20°C or 37°C in 24-well plates. In the case of C. trachomatis, the plate was incubated in the presence of 5% CO 2 . IFUs were quantified using McCoy cells immediately after the dilution in fresh medium or after 1, 2, 3, or 4 days of incubation.
Effect of incubation temperature on growth. Sf9 cells were plated in a 24-well plate at a density of 10 5 cells per well, infected with R. porcellionis at an MOI of ;0.1 to 1, and incubated for 48 h at 28°C. The plates were then incubated at 37°C for 6, 12, 24, or 48 h before being switched back to 28°C for four additional days. Samples were taken for bacterial growth quantification by qPCR right after the switch to 28°C and after the subsequent 4 days of incubation.
Quantitative PCR. Genomic DNA was extracted using the Wizard SV genomic DNA purification kit (Promega, Dübendorf, Switzerland) following the manufacturer's protocol. Quantitative PCR for R. porcellionis (36)  Immunofluorescence staining. Infected cells grown on glass coverslips were fixed with ice-cold methanol for 5 min at different time points after the infection. Cells were then washed three times with phosphate-buffered saline (PBS) and incubated for at least 2 h in PBS with 0.1% saponin, 0.04% NaN 3 , and 10% FCS (blocking solution). The coverslips were then incubated at room temperature for 2 h in blocking solution with rabbit anti-Simkania negevensis antibodies (25) (dilution at 1:1,000), rabbit anti-Waddlia chondrophila antibodies (71) (dilution at 1:1,000), or goat antibodies targeting the major outer membrane protein of Chlamydia trachomatis (dilution at 1:1,000) (LSBio, Seattle, WA, USA). Anti-Simkania antibodies were used to detect R. porcellionis, as antibodies raised against a chlamydial species often cross-react with related species (40,72). After the incubation with the primary antibody, the coverslips were washed three times in PBS with 0.1% saponin and incubated for 1 h at room temperature in blocking solution with 1.6 mg mL 21  Electron microscopy imaging. Sf9 and McCoy cells were plated in T25 flasks at a density of 10 6 cells per flask and infected with R. porcellionis at an MOI of ;1. After the initial 30 min of incubation at 28°C, the old medium was replaced with fresh medium supplemented with 1 mg mL 21 cycloheximide. The cells were then incubated for 6 days at 28°C before collection. The cell suspension was centrifuged at 500 Â g for 10 min, and the pellet was resuspended in a solution of 4% paraformaldehyde (Electron Microscopy Sciences [EMS], Hatfield, PA, USA) and 2.5% glutaraldehyde (Fluka, Buchs, Switzerland) in a 0.1 mol L 21 phosphate buffer at pH 7.4 (PB buffer) and incubated at 4°C for 4 h. After an additional centrifugation at 500 Â g for 10 min, the cells were resuspended in a solution of 1% paraformaldehyde in PB buffer. They were then directly postfixed by a fresh mixture of 1% osmium tetroxide (EMS, Hatfield, PA, USA) with 1.5% potassium ferrocyanide (Sigma, St. Louis, MO, USA) in PB buffer for 1 h at room temperature (RT). The samples were then washed three times in distilled water and spun down in 2% lowmelting-temperature agarose in H 2 O (Sigma, St. Louis, MO, USA), allowed to solidify on ice, cut in 1-mm 3  Statistical analysis. The results of this study are given as means with standard deviations. The linear regression models with random effect were fitted using the lme4 package (73). Doubling times were calculated by dividing 48 h (the interval between time points used in this work) by the log 2 of the highest fold change observed between two consecutive time points. All statistics were performed with R (v4.2.0).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 1 MB.