The bacterial community structure dynamics in Meloidogyne incognita infected roots and its role in worm-microbiome interactions

Background Plant parasitic nematodes such as Meloidogyne incognita have a complex life cycle, occurring sequentially in various niches of the root and rhizosphere. They are known to form a range of interactions with bacteria and other microorganisms, that can affect their densities and virulence. High throughput sequencing can reveal these interactions in high temporal, and geographic resolutions, although thus far we have only scratched the surface. We have carried out a longitudinal sampling scheme, repeatedly collecting rhizosphere soil, roots, galls and second stage juveniles from 20 plants to provide a high resolution view of bacterial succession in these niches, using 16S rRNA metabarcoding. Results We find that a structured community develops in the root, in which gall communities diverge from root segments lacking a gall, and that this structure is maintained throughout the crop season. We detail the successional process leading toward this structure, which is driven by interactions with the nematode and later by an increase in bacteria often found in hypoxic and anaerobic environments. We show evidence that this structure may play a role in the nematode’s chemotaxis towards uninfected root segments. Finally, we describe the J2 epibiotic microenvironment as ecologically deterministic, in part, due to active bacterial attraction of second stage juveniles. Conclusions High density sampling, both temporally and across adjacent microniches, coupled with the power and relative low cost of metabarcoding, has provided us with a high resolution description of our study system. Such an approach can advance our understanding of holobiont ecology. Meloidogyne spp., with their relatively low genetic diversity, large geographic range and the simplified agricultural ecosystems they occupy, can serve as a model organism. Additionally, the perspective this approach provides could promote the efforts toward biological control efficacy.


Introduction
Root knot nematodes (RKN; genus Meloidogyne ) are among the world's most devastating plant pathogens, causing substantial yield losses in nearly all major agricultural crops [1] . M. incognita and closely related species are found in all regions that have mild winter temperatures [2] , and are regarded as one of the most serious threats to agriculture as climate change progresses [3] . In their life-cycle, M.
incognita hatch in the soil, then invade a root.
Thus, the nematodes are exposed to the soil microbiome, rhizobacteria, root epiphytes and endophytes. Once inside the roots, the females modify the cells in order to establish a feeding site and form the characteristic knots for which they are named. Each knot contains at least one nematode feeding from a unique cell-type (the giant cells), surrounded by a gall of dividing cortical cells [4][5][6][7] . Throughout its life cycle stages, M. incognita are known to interact with microbes, such as cellulase-secreting bacteria and plant effector secreting bacteria, or bacterial and fungal antagonists [8][9][10][11][12][13] . Consequently, it appears that the geographic or temporal variation in the rhizosphere and root bacterial and fungal communities, can partly explain why such a near isogenic group of nematodes [14, 15] would display variable infestation success [8][9][10][11][16][17][18][19] .
The interaction of M. incognita virulence and microbial taxa in the various niches they occupy, have been studied in the context of biological control, revealing complex relationships, which efficacy diminishes with the transfer from lab to field [20] . Common themes in this line of research include the isolation of Meloidogyne pathogens from the cuticle of second stage juveniles (J2) [21][22][23] , or the identification of soil microbes and bacterial volatile compounds with antagonistic effects [24,25] , key taxa including Rhizobia [26] , Trichoderma and Pseudomonas [27,28] , Pasteuria [29] , Pochonia [30,31] and some mycorrhiza [31][32][33] .
Despite the importance of this plant parasite, and its close ties with its cohabiting microbiome, microbial ecology studies utilizing deep sequencing approaches are a handful. Only a few studies have attempted to characterize the taxonomic and functional core microbiota [34,35] , or tie the microbial community composition in the soil or plant with RKN suppressiveness [36][37][38][39] . In such studies, temporal dynamics of the microbiome in each of the various niches the nematode occupy at its different life stages, or along the crop season, is not often considered.
In this study we aimed to describe the

Results
To study the temporal dynamics of the bacterial community in the rhizosphere, the roots, galls and J2s in infected eggplants, we sampled these four niches from 20

Alpha diversity
To study the temporal changes in alpha-diversity during the vegetation season we calculated the total observed ASV, Pielou's evenness [40] , Shannon's diversity [41] and Faith's phylogenetic

Beta diversity
Beta diversity analyses were performed in order to study temporal and niche effects on the bacterial community composition. Weighted and unweighted pairwise UniFrac distances [43] were computed to account for changes in relative abundances or in the presence and absence of ASVs, respectively. Principal coordinate analysis (PCoA) [44,45] and biplots [45] were then used to visualize the relationships among the different data classes, and the key ASVs that explain them. ASVs were referred to by both their taxonomic assignment and the first six digits of their MD5 digests, referring to the full digests, as they appear in the representative-sequences fasta file and biom table.  The larger the deficit in core ASVs compared to the core taxa they are assigned to, the lower the ratio and the larger the drift. We based this approach on the notion that ecological drift can increase the genetic diversity among samples that were obtained from one niche [49] . We formulated the difference as the ratio of core ASV count to core taxa count (ASV to taxa ratio - With this notion in mind, among all the niches, only J2 had R > 1 on average, in their 100% core microbiome ( Fig. 5A ). The R value in J2, was significantly different from R in other niches

Bacterial succession
In addition to the temporal dynamics of alpha and beta diversity, we investigated the temporal change of discrete features (ASVs or taxa) to characterise the bacterial succession in various niches. We focused our investigation on features that we have identified as "important" or "dynamic" (see Methods section), based on a features volatility analysis [50] . We also investigated the two most abundant ASVs belonging to the included taxa, were they not already considered. The mean relative abundance of features is presented as a heatmap ( Fig. 6 ), organised according to niche   ( Fig. 6D ). Within the galls ( Fig. 6A ), an "early" bacterial community, including 14 taxa Another important aspect of the gall bacterial succession is the origin of taxa. Only a few taxa in the root clearly originated from the naive roots.
In the early community, these include

Attraction assay
In 2017, gall bearing eggplant roots were sliced and washed with PBS in order to attempt the isolation of RKN related bacteria (supplementary file S1). As RKN pathogens often actively attract J2s, we carried out an attraction assay, testing the attraction of J2s to each of two isolates given the isolate and fresh root as options, or the sterile medium and a fresh root as control. One isolate in particular presented a 10-fold larger attraction of J2s than the root (supplementary file S1). Upon sanger sequencing (supplementary file S1), this isolate had an identical V3-V4 region sequence as Pseudomonas 108751f2 ( Fig. 4 ). Microbacterium [61] .
A/N/P-Rhizobium, another taxon which occurred both in the gall and J2 samples, showed a covarying pattern of relative abundance between the two niches ( Fig. S2 ). It has been shown in the past that bacteria belonging to this group were able to successfully invade plant hosts by using nematodes as a transicion vector [62] .
Therefore, this observed covarying pattern of relative abundance may be non-incidental.
Both ASVs and taxa were analysed, assuming that in some cases discrete ASVs that represent functionally and taxonomically similar bacteria are ecologically interchangeable. In such cases, patterns that would be observed at the species or genus level, might be lost at the ASV level, and vise-versa . We grouped ASVs into taxa based on their lowest available taxonomic level.
One parameter with which we identified key ASVs and taxa in the system was "importance".
Importance is defined as the euclidean distance of the relative abundance vector of a given taxon or ASV from a null vector of the same length [50] .  ( Fig. 7 ).
The attractants we introduced with the pipette tips were prepared as follows. The isolate was incubated in aquaus beef-extract peptone medium (beef extract 3 g L −1 ; peptone 10 g L −1 ; NaCl 5 gL −1 ; [129] ) for 48 hours at 37°C. The culture was filtered using a 0.45 µm syringe filter to obtain the bacterial extracts (whole filtrate).
The size fractions were then separated using Amicon Ultra-15 centrifugal filter units with Ultracel-PL membrane, following manufacturer's instructions. M. incognita eggs were sieved from infected tomato roots using a set of mesh #200 sieve on top of mesh #500 sieve (Tyler.S.W), and J2 were hatched with a Baermann tray, as described in [112] . Fig. S1: Observed-ASV alpha-rarefaction curves for each niche.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
The datasets generated and analysed during the current study are available in the National Center