Tropical pitcher plants (Nepenthes) act as ecological filters by altering properties of their fluid microenvironments

Characteristics of host species can alter how other, interacting species assemble into communities by acting as ecological filters. Pitchers of tropical pitcher plants (Nepenthes) host diverse communities of aquatic arthropods and microbes in nature. This plant genus exhibits considerable interspecific diversity in morphology and physiology; for example, different species can actively control the pH of their pitcher fluids and some species produce viscoelastic fluids. Our study investigated the extent to which Nepenthes species differentially regulate pitcher fluid traits under common garden conditions, and the effects that these trait differences had on their associated communities. Sixteen species of Nepenthes were reared together in the controlled environment of a glasshouse using commonly-sourced pH 6.5 water. We analyzed their bacterial and eukaryotic communities using metabarcoding techniques, and found that different plant species differentially altered fluid pH, viscosity, and color, and these had strong effects on the community structure of their microbiota. Nepenthes species can therefore act as ecological filters, cultivating distinctive microbial communities despite similar external conditions, and blurring the conceptual line between biotic and abiotic filters.


Supplemental Discussion: Taxonomic composition of experimental pitchers Bacteria
Previous studies indicate a characteristic set of bacterial taxa associate with pitchers of Nepenthes species, as well as with the convergently evolved pitchers of New World pitcher plants in the genus Sarracenia . Some of the common bacterial taxa in our samples match what appear to be common associates of other pitcher plant taxa, especially the orders Acetobacterales (formerly under Rhodospirillales), Actinomycetales, and Rhizobiales, but also Sphingobacteriales (including the family Chitinophagaceae), Burkholderiales, Enterobacteriales, and Xanthomonadales (Bittleston 2018). While not dominant in terms of relative abundance, Acidobacteria and Caulobacterales occur fairly frequently across multiple samples, and these taxa have also been found to associate with Nepenthes in nature (Sickel et al. 2016;Bittleston 2018). Some of the common taxa in our study may also be common inhabitants of plant surfaces in general, such as Sphingomonadales, Pseudomonadales, and Xanthomonadales that have all been found in earlier studies of phyllosphere bacteria in other plant species (Vorholt 2012;Vacher et al. 2016). Other abundant or frequent taxa in our study include bacteria that do not appear to be frequent associates of pitchers, and these could represent environmental bacteria that are a consequence of the experimental glasshouse setting: Chlamydiae, Rickettsiales, TM6, and Verrucomicrobia. Most samples also possess a small fraction of OTUs that cannot be assigned at the phylum level. Overall, while the bacterial communities in our experimental plants may not be typical of those found in wild Nepenthes, we nevertheless recovered many commonalities in the diversity and abundance of certain groups.
Acetobacteriaceae, particularly the genus Acidocella (and to a lesser extent, Acidisoma) show up as frequent associates of Nepenthes pitchers in most microbiome studies so far, from natural or seminatural settings (Chou et al. 2014;Kanokratana et al. 2016;Sickel et al. 2016;Bittleston et al. 2018). Considering Acidocella sp.'s apparent relationship with Nepenthes and given that it is not known to be a common environmental bacterium, previously isolated from only a few extreme habitats (Kishimoto et al. 1995;Belova et al. 2009;Kimoto et al. 2010;Jones et al. 2013), it would be worthwhile to probe the function of taxa from this genus in relation to its host.
The possibility of vertical transmission cannot be completely ruled out regarding important Nepenthes symbionts. Vertical transmission has been seen in other phyllosphere systems (Vorholt 2012), and although it was initially established that pitchers are sterile prior to opening (Buch et al. 2012), more recent microbiome studies were able to find bacterial DNA in unopened pitchers (Chou et al. 2014;Takeuchi et al. 2015;Kanokratana et al. 2016), including Acidocella (Chou et al. 2014;Kanokratana et al. 2016). In each case, only a minority of unopened pitchers examined yielded detectable amounts of DNA, so the possible occurrence of bacteria in unopened pitchers is unresolved. In known cases of vertical transmission in plants, seed-associated bacteria can spread systemically throughout the developing plant (Vorholt 2012), so it is interesting to note that Sickel et al. (2016) found that bacterial composition did not differ significantly between pitcher fluid, pitcher external surfaces, and leaf lamina in their study. As evidence against vertical transmission of Acidocella sp., a previous greenhouse Nepenthes microbiome study did not find Acidocella as a prominent taxon (Takeuchi et al. 2015). However, the Takeuchi et al. (2015) study took place in temperate Germany, and it seems likely that key associates of pitchers such as Acidocella sp. are themselves geographically restricted, which could explain its presence both in the wild and in the Singapore greenhouse, without needing to invoke vertical transmission.

Eukaryotes
Symbiotic arthropod communities in pitcher plants, including several families of dipteran larvae and mites, have been well-characterized over decades of research (Beaver 1979). Being in an enclosed glasshouse, however, our experimental plants exhibited no evidence of being colonized by symbiotic arthropods. Additionally, microfauna such as nematodes and rotifers that may be expected to be fairly common members of wild pitcher communities (Quisado 2013;Bittleston et al. 2016;Bittleston 2018) were not common in our samples. Nevertheless, insect prey was available in the glasshouse, primarily fungus gnats (Sciaridae), which are common indoor plant pests that can be found in potting media. This prey DNA was detectable in our samples, enabling us to examine the impact of prey capture on microbiome dynamics.
Few studies have documented the composition of eukaryotic microbial communities in wild Nepenthes Bittleston 2018;Bittleston et al. 2018). Fungi such as Saccharomycetes, Agaricomycetes, and basal fungi (e.g. Mucoromycotina, Chytridiomycota) have been previously found in wild (Bittleston 2018;Bittleston et al. 2018), and these taxa also were present in our samples. However, while ascomycetous yeasts in Saccharomycetes appear to be the dominant fungi in wild communities (Bittleston 2018), our samples were dominated by Basidiomycota in Ustilaginomycotina and in Agaricomycotina (especially Tremellomycetes). Tremellomycetes can exist in single-celled yeast form, so these fungi could hypothetically occupy a similar ecological niche to the fungal yeasts in wild pitchers. The Ascomycota in our samples include Trichocomaceae (including the well-known genus Penicillium), which are common members of indoor microbial communities (Barberán et al. 2015); Chaetothyriales ("black yeasts" including the family Herpotrichiellae), which were previously found in wild pitchers (Bittleston 2018); as well as Sordariomycetes and Leotiomycetes, also members of Pezizomycotina. A single pitcher sample was dominated by Microsporidia, a basal fungus and obligate intracellular parasite of animals. So while certain fungal taxa have been found in wild pitchers, the overall taxonomic composition of fungi in our study appears to have had some major influence from the microbial pool of the greenhouse environment.
Pitcher plant microbiomes in this and previous studies also include Protista. Algae have been found to be common inhabitants of pitchers in the wild , Bittleston 2018, and we found a diverse community of algae throughout our samples, including members of Archaeplastida (Trebouxiophyceae and Chlorophyceae), Stramenopiles (Chrysophyceae and diatoms), and Discoba (Euglenozoa). Rhizaria and Amoebozoa are two abundant taxa in our samples that have also been found associated with pitchers in other studies , Bittleston 2018. Another common taxon across our samples was Alveolata, particularly Gregarinasina (gregarines). These are obligate arthropod parasites, especially of larval mosquitoes (Chen 1999;Tseng 2007) and they have been previously found in pitcher communities in association with the symbiotic larval dipterans . As there were no mosquitoes or dipteran larvae of any kind inhabiting our pitchers, the frequent occurrence of gregarines here is somewhat surprising. These gregarines may have been parasitizing the adult fungus gnats trapped by the pitchers; this is supported by ANCOM analysis showing that gregarine relative abundance is higher in pitchers with visible prey than those without visible prey (see figure) Perhaps more culture-independent surveys of eukaryotic diversity will reveal that gregarines are an even more important component of ecosystems than currently appreciated, perhaps infecting a wide assortment of arthropods in a variety of ecological contexts (Dabert and Dabert 2008;Criado-Fornelio et al. 2017).
Overall, the eukaryotic community composition appears to be somewhat uneven at the broad taxonomic level compared to the bacteria. Multiple phyla generally can be seen co-occurring within most samples for bacteria. However, for eukaryotes, most samples appear to be dominated by a single broad taxon, i.e. fungus-, metazoan-, or protist-dominated communities.