Microorganisms occur in the environment in either beneficial, hazardous, or neutral relationships. The nature of the microbe–microbe relationship or the host–microbe relationship is dependent on the ability of the microbe to withstand the external pressures exerted by the host or the environment it resides. As these microbes co-exist within a community, it is extremely difficult to decipher the complex ecological interactions that exist between them (Cruz et al., 2022; Geller and Levy, 2023). The plant ecosystem is one such example where microorganisms co-exist and contribute to plant health and productivity. Niche specificity and core abundance are major factors that determine the stability of a microbe in the plant ecosystem (Qiao et al., 2024). As microbes do not exist in isolation, their co-existence is highly dependent on the niche type (Barber et al., 2022). The law of competitive exclusion formulated by Gause states that two microbial species with the same niche exclude each other (Faust and Raes, 2012).

The cotton crop is devastated by the lethal cotton leaf curl virus (CLCuV), which is transmitted by the whitefly. Insect pests account for 37% of cotton yield losses, whereas the whitefly Bemisia tabaci is responsible for 50% of the total loss in cotton production (Oerke, 2006; Razaq et al., 2013). The genus Gossypium comprises 52 species, while only 4 are cultivated around the world, including Gossypium hirsutum, Gossypium barbadense, Gossypium arboreum, and Gossypium herbaceum (Ashraf et al., 2018; Hussain et al., 2020). Gossypium hirsutum accounts for 90% of worldwide cotton production, but it is susceptible to cotton leaf curl disease (CLCuD) (Hu et al., 2019). Gossypium arboreum is completely tolerant to CLCuD, but it is cultivated in less than 1% of cotton-growing areas worldwide due to its short fiber length (Edde, 2021). Conventional breeding strategies and transgenic approaches have not been able to mitigate the effects of this deadly virus. The microbiome approach has been proven successful in targeting fungal and bacterial pathogens. Biotrophic pathogens, such as the one causing CLCuD, are known to increase salicylic acid (SA) levels in infected plants as this phytohormone is found to be essential for gene-for-gene resistance, systemic acquired resistance (SAR), and reduction of disease development (Nawrath et al., 2006). Beneficial microbes from the phyllosphere can also switch on plant defense responses. Thus, plant immunity-boosting non-pathogenic microbiota is the new tool for conferring disease resistance in host plants (Legein et al., 2020).

Integrative metagenomics provides insights into the microbial community networks and ecological processes involved in biogeochemical cycles. It is still unknown how, under viral pathogen attack, the complex microbial communities interact with one another in the plant microbiome. The fields of genomics and ecology are brought together by network inference strategies, which aid in deciphering the relationships between more than two nodes involved in a particular network based on abundance data (Veiga et al., 2010). Furthermore, network topologies are useful for identifying the keystone species, i.e., those that play a pivotal role and, if perturbed, lead to maximum disruption in that network. In general, the keystone species were identified using a network property called the hubness score (Layeghifard et al., 2019), which seemed to correlate better with the properties of the ecosystem under study than looking at either the abundant or the prevalent species. A recent advancement in terms of incorporating local, semi-local, and global centrality measures, under the framework of integrated value of influence (IVI) (Salavaty et al., 2020) that also implicitly incorporates the hubness score, is shown to produce promising properties of the network while reducing the analytical biases. Therefore, we have incorporated the network statistics to explore how disease susceptibility correlates with some of these properties. We have employed the 16S rRNA gene amplification to unravel the spatial distribution patterns and complexities of the cotton microbiome in the CLCuD susceptible, partially tolerant, and resistant cotton varieties infected with CLCuV.

The term “microbiome influencer” holds prime importance in microbial ecology (Paul et al., 2022). It is the most influential node in an ecosystem that contributes to crop vigor and resilience as opposed to its most abundant species. The core community does hold primordial significance (Toju et al., 2018), but the stability of the microbial influencers is a major factor in the plant’s response to biotic or abiotic stressors (Qiao et al., 2024). Viruses are a major threat to plants worldwide, and the cotton leaf curl virus (CLCuV) is one of the deadliest viruses affecting the cotton crop (Mahmood-ur-Rahman et al., 2012). Crop phenology is majorly dependent on the microbiomes and mycobiomes that inhabit the internal and surrounding environment (Sharma et al., 2019; Ginnan, 2020). Rather than existing in isolation, microorganisms co-occur in ecological networks that determine the stability of the entire ecosystem and thus the plant’s response to biotic/abiotic stresses (Afridi et al., 2022). The drivers of these networks are the taxa termed the “most influential nodes” (Mukhtar et al., 2023).

In the FDH-228 network, more nodes have an IVI of 100, whereas in the PFV-2 network, only Chloroflexi has an IVI of 100. Members of the phylum Chloroflexi possess anaerobic fermentation potential and have been discovered in diverse habitats such as hot springs, sediments, and anaerobic sludge digesters (Hug et al., 2013; Xia et al., 2016; Petriglieri et al., 2018; Bovio et al., 2019). If the influential node stands prominent against all other nodes, i.e., if the difference is huge, then any perturbation in that influential node should propagate and disrupt the network far more, which is evident for PFV-2. On the contrary, in the FDH-228 resistant variety, on average, there are many influential nodes, with authority predominantly shared among firmicutes. We hypothesize that the community is robust against perturbations and should manage challenges better as a result of the decentralization of authority.

From GLLVM, it is quite apparent that the influential nodes all have consistently similar signs, whether all positive or all negative, which are associated with a particular compartment as compared to the reference. This seems to suggest that these species form part of a cohort that is increasing or decreasing globally as a cohort rather than exhibiting local changes in a few members. Compartment-wise network analysis revealed that drought stress disrupted the microbial network in the root endosphere, which contained the most influential nodes, compared to the phyllosphere and rhizosphere (Gao et al., 2022).

Microbial communities inhabiting different plant compartments exhibit varying capabilities owing to selection pressures associated with the compartment type. As this research aims to identify bacteria that could serve as biocontrol agents, it was necessary to screen all plant compartments for microbes that may possess disease-suppressing abilities. The recruitment of root-associated microbes relies heavily on the root exudates secreted by the plant roots (Massalha et al., 2017) and also the environmental factors such as soil pH, salinity, soil type, soil structure, soil moisture, and soil organic matter (Sindhu et al., 2022). Conversely, leaf-associated bacteria exhibit a low species richness as the phyllosphere is relatively nutrient-poor compared to the highly fertile rhizosphere. The leaf surface constitutes an inhospitable environment that is characterized by fluctuations in temperature, moisture level, and nutrient availability (Thapa and Prasanna, 2018). Phyllospheric bacteria, therefore, possess the ability to maintain environmental homeostasis by producing secondary metabolites or exogenous polysaccharides to aid in the survival of the host plant (Jackson et al., 2015).

It is interesting to note that in the partially tolerant PFV-1 network, the most influential nodes are root and leaf endophytic bacteria. It can be observed that the microbial influencers are in abundance in the network of the highly susceptible variety’s rhizosphere. The rhizosphere is known to harbor commensals and recruits them from the surrounding environment (Bulgarelli et al., 2013; Müller et al., 2016). In a previous study, rhizospheric microbial taxa and influential nodes were enriched in plants infected with the soil-borne yellow mosaic virus. The presence of beneficial taxa, including nitrogen fixers, such as members of Bradyrhizobiaceae, Xanthomonadaceae, Sphingomonadaceae, and Comamonadaceae, in the co-occurrence networks of infected wheat plants reveals that the pathogen is interdependent on the beneficial microbes that have maintained the ecological niche in the presence of disease (Wu et al., 2021). Azospirillum, one of the top influential nodes in the FDH-228 co-occurrence network, is a gram-negative nitrogen fixer capable of IAA, CK, and GA₃ production (Cassán and Diaz-Zorita, 2016). Georgenia, one of the influencers in the susceptible PFV-2 network, is a heterotroph capable of aerobic denitrification and has been previously isolated from deep-sea sediments and forest soils (Li et al., 2007; Wang et al., 2015; Rajta et al., 2022). Acetivibrio is an obligate anaerobe (Charoensuk et al., 2019) and Nitrosospira is an ammonia oxidizer (Koike et al., 2022), and both belong to the PFV-2 co-occurrence network.

In the resistant variety, the phyllosphere is home to the most influential nodes present in the network. The leaf epiphytic region is dominated by the top influencers, which is in accordance with our previous findings, where the SA-producing bacteria isolated from the phyllosphere of the FDH-228 variety conferred disease resistance against CLCuD in the susceptible variety (Aqueel et al., 2023). The phyllosphere microbiota are the most selected microbes and partake in nutrient cycling due to their specialized adaptations to climate change (Dorokhov et al., 2018; Cavicchioli et al., 2019; Koskella, 2020). They are also crucial for immune priming and pathogen elimination (Bell et al., 2019; Chen et al., 2020).

Our co-occurrence network analyses of the CLCuD-infected cotton plants with varying levels of susceptibility have revealed that the microbiome influencers show a consistent response in different compartments. The network of the resistant Gossypium arboreum possesses many influential taxa from the phylum Firmicutes. The revelation of these networks can help us understand the crosstalk between the plant genotype and microorganisms inhabiting various plant compartments under pathogenic attack. This will aid in the utilization of these ‘influential ecological drivers’ for viral disease suppression in cotton.

The datasets presented in this study can be found in online repositories (European Nucleotide Archive under the project accession number PRJEB67645). The meta data associated with the samples can be found in the Supplementary material.

RA: Formal analysis, Writing – original draft, Conceptualization, Data curation, Investigation, Visualization. AB: Formal analysis, Data curation, Investigation, Visualization, Writing – review & editing. UZI: Formal analysis, Funding acquisition, Methodology, Software, Supervision, Writing – original draft. KAM: Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing.