Taxonomic profiling of Nasutitermes takasagoensis microbiota to investigate the role of termites as vectors of bacteria linked to ironwood tree decline in Guam

Abstract

The ironwood tree (Casuarina equisetifolia, family Casuarinaceae), an indigenous agroforestry species in Guam, has been threatened by ironwood tree decline (IWTD) since 2002. Formation of bacterial ooze by the wilt pathogen from the Ralstonia solanacearum species complex and wetwood bacteria (primarily Klebsiella species) has been linked to IWTD. In addition, termite infestation of trees was statistically associated with IWTD. Termites are known carriers of a diverse microbiome. Therefore, we hypothesized that termites could be vectors of bacteria linked to IWTD. To investigate the potential role of termites as pathogen vectors, we employed next-generation 16S rRNA gene sequencing to describe the bacteria diversity of Nasutitermes takasagoensis (Family Termitidae) workers collected from 42 ironwood trees of different disease stages in Guam in association with tree-, plot-, and location-related factors. Nasutitermes takasagoensis workers account for the majority of termite infestations of ironwood trees. The bacterial phyla composition of N. takasagoensis workers was typical for wood-feeding higher termites consisting mainly of Spirochaetes and Fibrobacteres. However, Ralstonia species were not detected and Klebsiella species were rare even in termites collected from trees infected with Ralstonia and wetwood bacteria. Feeding experiments suggested that termites prefer to consume wood with low pathogen content over wood with high pathogen load. Termites were able to ingest Ralstonia but Ralstonia could not establish itself in healthy termite bodies. We concluded that N. takasagoensis workers are not vectors for Ralstonia spp. or the bacterial endophytes associated with wetwood (Klebsiella, Pantoea, Enterobacter, Citrobacter, and Erwinia) that were previously observed in IWTD-infested trees. The bacterial diversity in termite samples was significantly influenced by various factors, including Tree Health, Site Management, Plot Average Decline Severity, Proportion of Dead Trees in the Plot, Proportion of Trees with Termite Damage in the Plot, Presence of Ralstonia, and Altitude.

Fig 1. Bacterial diversity rarefaction in Nasutitermes takasagoensis worker samples.

a) Sequence-based rarefaction curves of bacteria diversity showing the number of ASVs, Faith’s phylogenetic distance and Shannon diversity indices in 42 samples of Nasutitermes takasagoensis workers plotted against sequencing depth. b) Sample-based rarefaction curves with effective bacterial diversity for different metrics plotted against the number of samples. c) Coverage-based rarefaction curves with effective diversity plotted against estimated sample coverage. Solid lines indicate intrapolation up to the actual sample size; dashed lines represent extrapolation to twice the sample size.

Fig 2. Relative abundance of bacterial phyla associated with 42 samples of N. takasagoensis workers collected from ironwood trees in Guam.

Phyla are shown in decreasing abundance from bottom to top. The sample name on the x-axis encodes the following seven factors: (1) presence (Positive) or absence (Negative) of Ralstonia (Rs), (2) Tree Health (Healthy or Sick), (3) Tree DS (symptomless (DS = 0), slightly damaged (DS = 1), distinctly damaged (DS = 2), heavily damaged (DS = 3), and nearly dead (DS = 4)), (5) level of Site Management (None, Moderate, High), (6) Altitude (High, Low) and (7) Parent Material (Sand, Lime, Tuff).

Fig 3. Tree-, plot and location-related factors with significant effects on different aspects of alpha diversity of termite bacteria communities.

Different letters indicate significant difference. a) Faith’s PD of bacteria communities of termites collected from healthy and sick ironwood trees. b) Pielou’s evenness and c) Shannon diversity index of bacterial communities of termites collected from highly, moderately or non-managed sites. d) Correlation between Pielou’s evenness of termite bacterial communities and Plot Average DS. e) Correlation between Shannon diversity of bacterial communities of termites and Proportion of Dead Trees in Plots. f) Correlation between Faith’s PD of bacterial communities of termites and Proportion of Trees with Termites in Plot.

Fig 4. Impact of R. solanacearum and wetwood bacterial loads of wood on consumption by N. takasagoensis workers.

Net consumption (g) of wood pieces which tested positive or negative for species from the R. solanacearum complex (Rs) and contained low or high amounts of wetwood bacteria by N. takasagoensis workers. Different letters indicate significant differences determined by Tukey’s Studentized Range Test for four-choice bioassay (S7 Table).

Fig 5. Dynamics of bacterial phyla abundance in N. takasagoensis workers: Impact of R. solanacearum concentrations and feeding durations.

Taxa barplots showing the shift in relative abundance of bacterial phyla of termites between the two phases of the experiment with different Ralstonia concentrations (No Ralstonia control, 10⁻⁸, 10⁻⁶, and 10⁻⁴) and durations of feeding on Ralstonia inoculated filter paper (Phase 1: Termites were fed for 2, 4, and 6 days with different concentrations of Ralstonia. Phase 2: Termites fed for 6 days with Ralstonia were fed for two additional days on filter paper only for a total duration of the experiment of 8 days). There were five replicates for each Ralstonia concentration, time point and colony (N1, N2, N3).

Fig 6. Temporal variation in alpha diversity metrics of bacterial communities in worker guts.

Box plot shows differences in alpha diversity (ASV richness, Faith’s PD, Pielou’s evenness, Shannon diversity) of bacterial communities in termites during Phase 1 (represented by red letters below the boxes) and for the overall duration of feeding including Phase 1 and Phase 2 (represented by blue letters above the boxes).