Media made from brown-rotted elm and pine wood for rearing Reticulitermes termites

Termites are not only decomposers in forest and savanna ecosystems, but also major pests of wooden buildings worldwide. Therefore, termites have long been of interest to researchers, and there has been a need for media for maintaining termites in the laboratory. Recently, a "brown-rotted pinewood mixed cellulose" (BPC) medium using the Japanese red pine (Pinus densiflora) was developed as a medium for long-term stable rearing of the subterranean termite Reticulitermes speratus. It is unclear whether the method of preparing the BPC medium can be applied to other termite species using different types of wood. Here, we developed two media for rearing two American termite species (Reticulitermes flavipes and R. virginicus): "brown-rotted elm wood mixed cellulose" (BEC) medium using winged elm (Ulmus alata) and BPC medium using loblolly pine (P. taeda). In colony foundation by R. flavipes dealates, about 80% of the incipient colonies on BEC and BPC media successfully produced larvae and workers, whereas less than 21% of the colonies on cellulose media did so. Long-term rearing of R. flavipes workers showed that the food consumption rate was higher on the BEC and BPC media than on the cellulose medium. On the other hand, long-term rearing of R. virginicus workers showed that the apparent digestibility coefficient was higher on the BPC medium than on the other media. These results indicate that the BEC and BPC media can be used for both colony foundation by dealates and rearing workers in R. flavipes, and that the BPC medium is more suitable for rearing R. virginicus workers. These new media and others made using the same process will contribute to the maintenance of termite colonies in the laboratory for a variety of experiments.


Introduction
Termites are important decomposers in forest, savanna, and desert ecosystems (Lo and Eggleton 2011), and some species, especially subterranean termites such as Reticulitermes species, are structural pests due to their high wood consumption ability (Rust and Su 2012). Therefore, termites have long been of interest to researchers in terms of genetics, ecology, physiology, ethology, and pest management (Rust and Su 2012;Arab et al. 2021), and there has been a need for media for rearing termite colonies in the laboratory for long periods of time. However, termite rearing methods vary widely from researcher to researcher, with many using natural wood and soil as food and nesting material (Becker 1969). Because using natural wood and soil may carry the risk of pathogen contamination (Rosengaus et al. 2011;Aguero et al. 2021), it is necessary to develop a medium for each termite species that is standardized in its manufacturing process, has a low risk of pathogen contamination, and contains sufficient nutrients for termite growth to serve as food and nesting material.
Recently, a medium consisting of brown-rotted pinewood powder and cellulose powder, called "brown-rotted pinewood mixed cellulose" (BPC) medium, was developed as a standard medium for rearing the Japanese subterranean termite Reticulitermes speratus (Mitaka et al. 2023). This medium is a modified version of mixed sawdust bait (Matsuura and Nishida 2001), which is made from logs or blighted trees of brown-rotted pinewood (Pinus densiflora) that has been crushed, autoclaved, and then yeast fermented for about a month. The BPC medium shows high success rate in the initial stage of colony establishment by dealate pairs (90% of dealate pairs nested in the 1 3 medium within a week) and high survival rate of worker populations (more than 80% of workers survived after the first month of rearing) in comparison with the other media (Mitaka et al. 2023). When colonies are established by dealates using natural rotted wood or soil, only 50% or fewer pairs of dealates survived after the first week of rearing (Eyer and Vargo 2022). Thus, it appears that a mixture of fermented brown-rotted wood powder and cellulose powder, such as the BPC medium, is a more suitable food and nesting material than untreated natural rotted wood and soil.
In this study, we set out to determine whether a medium made using the same process as the BPC medium, but from decaying wood of species other than P. densiflora could be adapted for the rearing of other termite species. We developed two types of media for rearing of two American termite species (Reticulitermes flavipes and Reticulitermes virginicus): "brown-rotted elm wood mixed cellulose" (BEC) medium using winged elm wood (Ulmus alata) and the BPC medium using loblolly pine (Pinus taeda). Also, the effects of these two media on the survival and growth of R. flavipes and R. virginicus individuals were tested.

Preparing media for termite rearing
We collected logs of brown-rotted winged elm wood (U. alata) and loblolly pine wood (P. taeda), which were not infested with termites, to make standard media for rearing R. flavipes and R. virginicus. Logs of brown-rotted elm wood were collected from a forest in Lick Creek Park (College Station, Texas, USA) in August 2022. Logs of brown-rotted pine wood were collected in a forest at the Sam Houston State University Center for Biological Field Studies (Huntsville, Texas, USA) in September 2022.
In each tree species, logs were cut using a saw and were broken into pieces using a hatchet. Very hard parts of the logs (e.g., the heartwood) were removed, and the remaining parts were used to make the media, according to a previous study (Mitaka et al. 2023). First, wood fragments were soaked in water for a day, and then the soaked fragments were minced in a high-performance blender (Vitamix Professional Series 200,Vitamix Corporation,Cleveland,Ohio,USA). After water was filtered out through a finemesh laundry net, about 3 L residual powdered wood was autoclaved at 121 °C for 15 min and then dried for 5 min (STERIS Amsco LG-250 Laboratory System Sterilizer, Ohio, USA). The autoclaved powdered wood was soaked in about 6 L water with about 0.2 g dry yeast powder (Safinstant GOLD, Lesaffre Yeast Corporation, France) in a plastic container (295 × 330 × 160 mm). The container was covered with a lid and kept at about 25 °C for fermenting. Twice per week, we removed suspended matter with supernatant, added a quantity of fresh water equal to the amount of supernatant removed, and stirred the contents of the container. This fermenting process was continued for at least a month to completely remove suspended matter. After that, water was filtered out through the fine-mesh laundry net, and the remaining powder was dried at room temperature. Finally, the BEC medium (or BPC medium) was completed by mixing the fermented elm (or pine) powder with pure α-cellulose powder (Cellulose, microcrystalline, Thermo Fisher Scientific, Massachusetts, USA) at a 1:1 volume ratio.

Termite collection
Five stock colonies (colonies A to E) of R. flavipes were collected from forests in Bryan and College Station, Texas, USA, and two stock colonies (colonies F and G) of R. virginicus were collected from forests in Huntsville and College Station, Texas, USA (Table 1). Colonies A, B and C were used for collecting alates, and the other colonies were used for collecting workers. Termites were identified based on morphological characters (Scheffrahn and Su 1994).

Colony foundation
To determine which food media were suitable for colony establishment by dealates of R. flavipes, we compared the survival of female-male dealate pairs and emergence rates of their first brood using three different food sources: BEC, BPC, and pure α-cellulose media (Fig. 1A). Alates were collected from three colonies (A, B, and C). Because alates had already emerged in the nest wood when they were collected in the field, we brought the nest wood back to our laboratory, placed them in a bucket, kept them at about 25 °C, and waited for the alates to appear on the surface of the nest wood before collecting them. After collecting the alates, the wings were removed naturally by the alates, or with the assistance of forceps by lifting the wings directly over the thorax and head of the alates. The dealates were separated by sex, determined based on external abdominal characteristics. Individual pairs of dealated females and males were randomly chosen from each colony and placed in a 40-mm plastic dish lined with one of the three food media moistened with distilled water to a height of about 5 mm (equivalent to about 6280 mm 3 ); 50 (colonies A and C) or 20 (colony B) nestmate pairs were selected for each colony for each food medium. After 1, 2, 4, 8, and 16 weeks, we recorded the proportion of the established colonies in which both female and male dealates survived and the proportion of the colonies that produced eggs, larvae (1st and 2nd instars), and/or workers (3rd instar or older). Comparisons of these proportions (binary variable: present or absent) among the food media in each week were performed using a generalized linear mixed model (GLMM) with a binomial distribution. In these models, each proportion was a response variable, food was an explanatory variable, and colony was considered a random effect. Likelihood ratio tests (LRTs) were conducted to test the significance of the explanatory variable, and Tukey's HSD tests were conducted to test for differences between levels of a significant fixed effect. All statistical analyses were performed in R software ver. 4.0.0 (R Core Team 2016).

Worker survival, growth, and food digestion
To compare the long-term effects on the survival rate, body weight gain, food consumption rate, and apparent digestibility coefficient (ADC) of workers of R. flavipes and R. virginicus among the BEC, BPC, and pure cellulose media, we performed a feeding test for 8 weeks. We placed 50 workers in a small dish (40-mm diameter) filled with one of the above three media to a height of 5 mm (equivalent to about 6280 mm 3 ). To calculate the water content of the BEC and BPC media, we measured the wet weight of each medium in a small dish, dried them at 125℉ (equal to about 51.67 °C) for 24 h, and then measured the dry weight. The dish was covered with a lid and kept at about 25 °C under a 14L:8D photoperiod. Each treatment had ten replicates (5 replicates/colony × 2 colonies/species) using two R. flavipes colonies (D and E) and two R. virginicus colonies (F and G). We recorded the number of surviving individuals and their total weight every 4 weeks to calculate body weight per individual. Survival rate was compared using a GLMM. In this model, survival rate was treated as a response variable assuming a binomial distribution. Food types, weeks measured from when the experiment started, and interactions between food types and weeks were treated as explanatory variables. Colonies were treated as random effects. Body weights of survivors were compared using a linear mixed model (LMM). In this model, body weight was treated as a response variable. Food types, weeks measured from when the experiment started and interactions between food types and weeks were treated as explanatory variables. Colonies were treated as random effects. Multiple pairwise comparisons by Chi-square test were performed for post hoc comparisons of the interaction between food types and weeks in each of survival rate and body weights of survivor, and a sequential Bonferroni correction was applied to evaluate pairwise significance.
After recording survival rate and body weight of surviving workers 8 weeks after the start of the experiment, we separated the uneaten medium (wet powdery lump) and fecal material (wet paste-like material) in the small dish, dried them at 125 ℉ (equal to about 51.67 °C) for 24 h, and then measured their dry weights. Subsequently, we compared food consumption rate, and apparent digestibility coefficient (ADC) (Robbins 1983) among the three different foods. Food consumption rate was calculated as 100 × (W start -W finish )/ W start , where W start and W finish are the dry weights of the food at the start and finish of the experiment, respectively. ADC was calculated as 100 × (W start -W finish -W feces )/ (W start -W finish ), where W feces is the dry weight of feces at the end of the experiment. Food consumption rate and ADC were compared using an LMM. In these models, food consumption rate or ADC was a response variable, food was an explanatory variable, and the colony was considered a random effect. LRTs were conducted to test the significance of each explanatory variable, and Tukey's HSD tests were conducted to test for differences between levels of a significant fixed effect. All statistical analyses were performed in R software ver. 4.0.0 (R Core Team 2016).

Survival and growth of incipient colonies in R. flavipes
To confirm whether the BEC and BPC media are more suitable for colony establishment of R. flavipes dealates than the pure cellulose medium, we reared the established colonies for 16 weeks to determine the survival of the dealates that became the primary king and queen, and the timing of egg, larval, and worker emergence. The proportion of the established colonies in which both female and male dealates survived did not differ among all media types for each week (each week: LRT, df = 2, χ 2 ≤ 0.910, P > 0.05, Fig. 1B) and were more than 86% for the entire 16 weeks on all media (BEC: 88%, BPC: 88%, cellulose: 86%).
The proportion of the established colonies that produced eggs was the highest on the BEC medium and the lowest  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Week Proportion of colonies with eggs n.s. n.s.  4 5 6 7 8 9 10 11 12 13 14 15 Fig. 1 Experimental scheme of colony foundation test A and dynamics of the proportions of the founded colonies with surviving dealates B, eggs C, larvae D, and workers E. The values denote the mean ± standard error of the mean (SEM) of 120 replicates (colonies A and C: 50 replicates, colony B: 20 replicates). Different letters indicate significant differences (GLMM followed by Tukey's HSD test, P < 0.05). NA indicates "not applicable" because there were no larval or worker counts that week. After 16 weeks, in the BEC and BPC media, all eggs hatched in more than 90% of the founded colonies, some of the hatched larvae developed into workers, and the dealates stopped laying eggs for a time. In the cellulose medium, however, 64% of the founded colonies still had eggs and only 17% of the colonies had hatched larvae on the cellulose medium at 1 week post-rearing (LRT, df = 2, χ 2 = 10.641, P < 0.01, Tukey's HSD test, P < 0.05, Fig. 1C), but at 2 weeks post-rearing, conversely, the proportion was the highest on the cellulose medium (LRT, df = 2, χ 2 = 22.815, P < 0.001, Tukey's HSD test, P < 0.001, Fig. 1C). At 4 and 8 weeks post-rearing, there were no differences in the proportion among the media (4 weeks later: LRT, df = 2, χ 2 = 2.811, P > 0.05; 8 weeks later: LRT, df = 2, χ 2 = 0.304, P > 0.05), and eggs were found in more than 65% of the established colonies on all media types (Fig. 1C). At 16 weeks post-rearing, the proportion dropped significantly in the BEC medium (10%) and BPC medium (8%), but remained high in the cellulose medium (64%) (LRT, df = 2, χ 2 = 123.100, P < 0.001, Tukey's HSD test, P < 0.001, Fig. 1C).
Proportion of the established colonies that produced larvae was compared among the media only at 8 and 16 weeks after rearing, because larvae did not emerge until 8 weeks after rearing. At 8 weeks post-rearing, larvae hatched in more than 65% of the established colonies on the BEC and BPC media, but in only 7% of the established colonies on the cellulose medium (LRT, df = 2, χ 2 = 145.230, P < 0.001, Tukey's HSD test, P < 0.001, Fig. 1D). This trend continued until 16 weeks after the start of rearing, when the final proportions were 78% on BEC medium, 80% on BPC medium, and 17% on cellulose medium (LRT, df = 2, χ 2 = 135.780, P < 0.001, Tukey's HSD test, P < 0.001, Fig. 1D).
The proportion of the established colonies that produced workers was compared among the media only at 16 weeks after rearing because no workers molted from the larvae until 16 weeks after rearing began. As a result, workers appeared in more than 79% of the established colonies on the BEC and BPC media, but in only 21% of the colonies on the cellulose medium (LRT, df = 2, χ 2 = 118.840, P < 0.001, Tukey's HSD test, P < 0.001, Fig. 1E).
In addition, after 8 weeks of colony foundation, only in the cellulose medium, dozens of dishes had brown, black, or red stains developed on the surface of the cellulose medium, whereas no such growth was seen in the BEC or BPC medium.

Discussion
Reticulitermes termites nest primarily in logs of decaying pine trees, but may also nest in hardwood logs (Smythe and Carter 1970;Smythe et al. 1971). In fact, in the forests where we collected termites in this study, Reticulitermes termites often nest in decaying wood of loblolly pine (P. taeda) or in decaying hardwoods including winged elm (U. alata). Therefore, we attempted to prepare media not only from pine wood but also elm wood and determine whether these media are suitable for rearing two termite species, R. flavipes and R. virginicus. Colony foundation tests using dealates of R. flavipes revealed that the proportion of the established colonies with larvae and workers was much higher in the BEC and BPC media than in the cellulose medium after 16 weeks post-foundation. Also, at 16 weeks after colony foundation, the proportion of founded colonies with eggs was low on the BEC and BPC media because most eggs hatched into larvae, while it remained high in the cellulose medium where many eggs did not progress to the larval stage (BEC: 10%, BPC: 8%, cellulose: 64%, Fig. 1C). These results indicate that nesting and egg laying occur on all the media by eight weeks after colony foundation, but larval development is delayed or impaired in the celluloseonly environment. It is known that yeast can produce various organic acids, including acetic acid, in the process of fermenting wood (Sirén et al. 2015), and that acetic acid acts as both a nutrient (Brune 2014) and an antimicrobial agent (Inagaki 2018) for termites. Therefore, the organic acids including acetic acid might be a good nourishment for pairs of R. flavipes dealates and sterilize the inside of the BEC and BPC media. Because Reticulitermes dealates slightly increase their own body weight and the amount of protein and lipid in their bodies during the first week or two after colony establishment (Inagaki et al. 2020), they are inferred to use the nutrients that they ingested from food during this period to produce the first brood or substances necessary for maintaining the hygienic environment in the nest. If the dealates are unable to ingest wood-derived components other than cellulose during this period, they might not to be able to provide the eggs with sufficient nutrition for development. Also, pine wood contains various kinds of antimicrobial terpenes (Gershenzon and Dudareva 2007), so that the BPC medium made from pine wood is presumed to be less prone to the growth of pathogens. In addition, the cellulose medium does not contain any antimicrobial substances, and in fact, there were stains on the surface of this medium only in dozens of dishes, while there were no such stains in either of the media containing wood. Therefore, it is The values denote the mean ± standard error of the mean (SEM) of ten replicates (5 replicates/colony × 2 colonies). Different letters indicate significant differences. A, B GLMM, pairwise comparisons of the interaction between food types and weeks by log rank test, followed by Bonferroni correction, P < 0.05; C, D LMM followed by Tukey's HSD test, P < 0.05. Food consumption rate was lower in cellulose medium than BEC and BPC media, although there were no significant differences in survival rate, body weight per individual, and ADC among media possible that the dealates devoted more nutritional resources to the production of antimicrobial substances to maintain nest hygiene in the cellulose medium than in the BEC and BPC media, and thus the dealates did not devote sufficient nutrition to and/or care for their eggs. Overall, it can be said that nutrients and antimicrobial compounds derived from brown-rotted wood are essential for the healthy development of incipient colonies of R. flavipes, and both the BEC and BPC media are suitable for rearing incipient colonies. During long-term rearing of R. flavipes workers, there were no differences in survival rate, body weight per worker, and ADC among the BEC, BPC, and cellulose media ( Fig. 2A, B and D), but food consumption rate was higher on the BEC and BPC media than on the cellulose medium (Fig. 2C). The reason why there was no significant difference in ADC among the media is that the workers on the cellulose medium produced less fecal material than on the other media (see Dataset in the Supplementary information). In the long-term rearing test of R. virginicus workers, there were no significant differences in survival rate and body weight per worker among media ( Fig. 3A and B), but ADC was higher in the BPC medium than in the other media (Fig. 3D). This is because food consumption rate was not significantly different among the media (Fig. 3C), while fecal material output was lower on the BPC medium. As mentioned earlier, the BPC medium made from pine wood also contains terpenes with antimicrobial activity (Gershenzon and Dudareva 2007), so that workers do not need to coat the inner walls of nests with large amounts of fecal material, which is considered to have antimicrobial activity (Rosengaus et al. 1998), while nesting on the BPC medium. However, the different ADC values between workers of R. food types and weeks by log rank test followed by Bonferroni correction, P < 0.05; C, D LMM followed by Tukey's HSD test, P < 0.05. The workers that fed BPC medium showed the highest ADC value among media, although there were no significant differences in survival rate, body weight per individual, and food consumption rate among media 1 3 flavipes and R. virginicus when fed BEC and cellulose media (Figs. 2D and 3D) suggest that different termite species have different digestive efficiencies for each medium and different amounts of fecal material required to coat the inner walls of the nests. The need for sterilization in the nest may depend on the antimicrobial spectrum and producible amounts of antimicrobial substances possessed by each termite species and the type of microorganisms that can be introduced into the nest material. Selecting the type of medium that is compatible with each termite species is likely to be crucial for the long-term rearing of termites. Now, it can be said that R. flavipes workers can be reared on BEC, BPC, or cellulose media, but when rearing R. virginicus workers, BPC medium is more suitable than the BEC and cellulose media because they digest food more efficiently. We demonstrated that the BEC and BPC media are usable for establishing incipient colonies of R. flavipes, but to further increase the size of the incipient colonies, soil with high levels of nitrogen may be needed in addition to these wood-based media. This is because a previous study showed that organic soil is necessary to accelerate the growth of founding colonies of the termite Coptotermes formosanus, which is closely related to Reticulitermes termites (Mullins et al. 2021). Decaying wood inhabited by termites is generally considered to be poor in nitrogen (Scharf and Peterson 2021). Although a previous study reported that yeast fermentation increases crude protein content of brown-rotted wood powder by a small amount, the crude protein content of the BPC medium is only 2% (Mitaka et al. 2023), making it a rather poor source of nitrogen. Therefore, it is possible to continue rearing termites only on the BEC or BPC medium, but if we want to increase the efficiency of termites' intake of nitrogen sources and increase colony growth rate, a rearing method that provides termites access to both the medium and organic soil may be better. However, there is a risk that the introduction of organic soil may allow pathogens to intrude into the termite nest, thereby killing termites due to disease infection. Therefore, it may be safer to introduce soil only after the number of workers in the colony has increased sufficiently to allow the workers to devote effort to sterilize the nest. In addition to organic soil, adding uric acid as a nitrogen source to the BEC or BPC medium may also be effective in promoting egg production by termite queens. This is because it has been shown that adding uric acid to diet increases egg production when rearing incipient or mature colonies of termites (Shellman-Reeve 1990;Brent and Traniello 2002). In addition, a recent study revealed that only kings and queens in a subterranean termite colony express urate oxidase, which catalyzes the first step of nitrogen recycling from stored uric acid, and the resulting metabolites can be used for egg production by queens (Konishi et al. 2023). However, while termite urate oxidase is highly expressed in mature primary and secondary reproductives, it is rarely expressed in alates and in primary kings and queens of incipient colonies that have been established for about 6 months (Konishi et al. 2023). Together, these studies suggest that the addition of uric acid to the BEC or BPC media may be better for rearing colonies with a mature (or maturing) king and queen and many workers than for rearing young colonies with an immature primary king and queen and a few workers.
In this study we showed that the method for preparing the BPC medium for rearing R. speratus reported in a previous study (Mitaka et al. 2023) is more generally applicable to other tree species and other termite species. We then showed that both the BEC medium made from elm wood and the BPC medium made from pine wood can be used for rearing incipient colonies and workers of R. flavipes. Considering a previous study reporting that less than 50% of the pairs of dealates survived when natural rotted wood was used directly for colony establishment (Eyer and Vargo 2022), it is noteworthy that the survival rates of pairs of dealates are much higher when the incipient colonies are established on the BEC and BPC media (> 87%, Fig. 1B). Moreover, the BPC medium was shown to be more suitable for rearing R. virginicus workers than the other media tested. Since these two termite species can nest in decaying wood other than winged elm and loblolly pine wood (Smythe and Carter 1970), it is possible that there are other tree species that could be used as media for culturing these Reticulitermes species. However, at the very least, both brown-rotted pinewood and hardwood are likely to be available for rearing R. flavipes, and it may be more efficient to use pine wood than hardwood for rearing R. virginicus. This work will contribute to the stability and efficiency of termite breeding experiments.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.