Healthy plants host taxonomically diverse microbial communities (Trivedi et al., 2020). Terrestrial plants are colonized by a huge variety of microorganisms that affect plant health and growth in beneficial, harmful, or neutral ways. Previous studies have confirmed that the host-microbial communities are not generally random assemblages, but show defined phylogenetic structures, and only a few major groups constitute the plant microbial communities with high abundance, including Proteobacteria, Actinobacteria and Bacteroidetes (Müller et al., 2016).

Many studies aimed to identify the core microbiome by phylogenetically distinct plant species or compartments. Yeoh et al. (2016) determined the core root microbes of sugarcane, including Proteobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, and Firmicutes, which were found to overlap with those of Arabidopsis thaliana, suggesting that some bacterial families may have a long-term association with plants (Yeoh et al., 2016). The study of core phyllosphere microbiomes of Arabidopsis, soybean and clover found that a high consistency of communities in leaves of these different plants (Delmotte et al., 2009). Proteobacteria, Actinobacteria and Bacteroidetes were the core phyllosphere microbiota in these three plants (Delmotte et al., 2009). Kembel et al. (2014) tested the relationships between the phyllosphere bacterial biodiversity and the functional traits, taxonomy and phylogeny of their plant host, and confirmed that the structure of phyllosphere bacterial community was highly correlated with the evolutionary relatedness of hosts and a series of plant functional traits related to host ecological strategies (Kembel et al., 2014). Similarly, Leff et al. (2015) observed the bacterial community composition is highly variable, but this variability is predictable and dependent on plant compartments (Leff et al., 2015). This work highlights the importance of considering plant spatial structure when studying plant-associated microbial communities and their effects on plant hosts.

The plant microbiome interacts with its host in a variety of manners and affect host’s growth. However, relatively little information is available on the composition and diversity of bacterial communities in different aboveground plant organs, especially those associated with flowers. Zarraonaindia et al. (2015) revealed some unique plant-associated microbiota by comparing the bacterial communities in grapevine leaves and flowers; the flower-associated microbiota almost entirely composed of Proteobacteria (Zarraonaindia et al., 2015). Junker and Keller (2015) determined that the leaves microbiome of Metrosideros polymorpha hosts a unique indicator community composed of relatively abundant bacterial groups; these indicator communities are accompanied by a large number of ubiquitous or rare bacteria with lower abundances.

Microorganisms colonizing the host plant benefit from nutrients provided by the plant and form taxonomically consistent community patterns (Müller et al., 2016). This mechanism consists of two aspects: on the one hand, plants provide unoccupied niches for invading microorganisms, which are able to take advantage of the provided niche or nutrients, resulting in stochastic colonization events; on the other hand, plant-microbe coevolution might provide the basis for plant-driven selection processes, leading to key species that actively recruit microbial members or at least provide functions to plant hosts, which contribute to the shaping of the ultimate microbial community during plant development (Müller et al., 2016).

Plants act as environmental filters to shape the microbial communities, and there are many studies about the effects of microbial communities on their hosts (Adair and Douglas, 2017). However, the study of the mechanisms underlying the assembly of the plant microbiome is still in its infancy. Based on microbial communities and genetic evidence, Reinhold-Hurek et al. (2015) proposed a enrichment model of root-related microbiomes, explaining the reasons for the decreased diversity and increased specificity of microbial communities from the bulk soil to the root interior (Reinhold-Hurek et al., 2015). Soils provide a highly diverse microbiome that is influenced by soil physicochemical factors, environmental factors, and vegetation types. This process is mainly divided into three steps. In the first step, in the rhizosphere, the community may be refined due to the influence of roots, such as carbon source, oxygen, pH or nutrient consumption. In the second step, the community shows more obvious refinement in close contact with the host on the root surface, and plant genotype has a significant influence. Biofilm formation or specific adhesion mechanisms may be involved in this enrichment step. Third, in the least complex community, Proteobacteria is a fast-growing bacterium with high metabolic activity, which dominates the endosphere bacterial community. For endophytic bacteria, specific bacterial traits are required to colonize in plant tissues, such as community sensing, plant polysaccharide degradation or resistance to reactive oxygen species. Compared with other plant compartments, host genotype may have the greatest influence on community structure in this habitat.

The establishment of plant microbiome is a dynamic process, reflecting the changes of community composition over time in response to environmental changes. Among them, environmental conditions, root exudates, microbe-microbe and plant-microbe interactions, and various stages of plant development all affect the assembly of microbial communities (Huang et al., 2014). It is not clear whether bacteria establish a founding community early in plant development, and then constantly colonize in the new habitat by clonal propagation (Müller et al., 2016). Therefore, it is very difficult to understand a system as complex as the plant microbiome and to distinguish between co-evolving interactions and stochastic processes.

So far, little is known about the source of microbial communities and how plants shape their specific microbial communities. Although the initial bacterial communities are similar to their respective seed banks, they become more specific and more diverse as plants grow and develop (Chaparro et al., 2014; Müller et al., 2016). Hardoim et al. (2012) found that about 45% of the bacterial communities in rice progeny seeds were similar to the parental plant, indicating that the rice microbiome could be directly obtained from seeds and transmitted in the form of endophyte (Hardoim et al., 2012). The diversity and dynamics of seed microbiota represent the culmination of a complex process of microbial interactions mediated by the plant throughout its life cycle (Nelson, 2018). Microbes associated with endosperm are more likely to vertically spread, while those associated with seed coat are more likely to horizontally transmit (Barret et al., 2016).

Soil represents an extremely rich microbial pool on Earth, serving as a major seed bank for the microbiota of the rhizosphere and root, and a driving force for community formation (Müller et al., 2016). In many studies, it has been confirmed that soil type has a significant impact on rhizosphere microbial communities (Gottel et al., 2011; Lundberg et al., 2012; Edwards et al., 2015). In addition, under the same environmental and soil conditions, plant species and genotype are one of the driving factors for the structural and functional diversity of plant microbiomes (Matthews et al., 2019). The functional diversity of the microbiome is also affected by plant varieties or cultivars. Even the sister lines of the same hybrid had different root-associated diazotrophs in rice (Knauth et al., 2005). Moreover, the influence of plants on microbial communities seems to be heritable, reflected in the finding that the gene expression of the microbiota in interspecific hybrid rice showed the intermediate level between the parental species (Knauth et al., 2005). These studies suggest that plants act as filters for their microbiome. However, quantitative analysis of microbial community composition between ecotypes or plant varieties showed that differences in microbial diversity related to genotypes were relatively small compared with environmental factors (Edwards et al., 2015).

Finally, microorganisms themselves can also influence the establishment of microbial communities, which can be mediated via either directly at the microbial-microbial interaction or indirectly through host-microbial interactions. This feature has been confirmed in many studies. For example, Chapelle et al. (2016) analyzed the rhizosphere microbiome of sugar beet seedlings, and found that invasive pathogenic fungi could induce stress responses in rhizobacterial communities, leading to shifts in microbial community composition (Chapelle et al., 2016). Han et al. (2020) determined that some rhizosphere microbes in nodules were related to the composition of rhizobia, and confirmed that bacterial communities played a crucial role in the interaction between soybean rhizobia and host (Han et al., 2020). Niu et al. (2017) assembled a simple and representative synthetic bacterial model and studied the community assembly process of maize seedlings. The abundance of representative strains was monitored by selective culture-dependent method, and the role of each strain in community assembly was studied. When Enterobacter cloacae was removed, the community structure was completely lossed, indicating that E. cloacae plays the role of key species in this model ecosystem (Niu et al., 2017). In summary, current studies indicate that microbe-microbe-host interactions are one of the factors affecting the assembly of the plant microbiome.

Many agricultural soils are of poor quality due to deficiency of one or more plant nutrients, so plant growth in this condition is suboptimal. In order to solve this problem and achieve higher crop yields, cultivators increasingly rely heavily on inorganic chemical-based fertilizers of nitrogen and phosphorus. The production of chemical fertilizers is not only costly, but also depletes non-renewable resources, such as oil and gas, and causes harmful effects to humans and the environment (Shirokov and Tikhnenko, 2021). It is clearly beneficial and sustainable if effective biological means of providing nitrogen and phosphorus to plants could be used to replace part of the chemical nitrogen and phosphorus usage (Glick, 2012). PGPB can help host plants to uptake nutrients, including nitrogen, iron, and phosphorus, which is considered a direct mechanism of plant growth promotion (Figure 2).

The physiological activities of most plants are regulated by one or more plant hormones. In addition to plants, many beneficial microbes can also synthesize plant hormones (Figure 2) (Khan et al., 2020). PGPB can promote nutrient uptake and metabolism of host plants by producing plant hormones (Stegelmeier et al., 2022). Studies have found that PGPB can secrete plant hormones, resulting in facilitated seed germination, accelerated root growth, changed root morphology, and increased root biomass (Olanrewaju et al., 2017). At present, many studies found that PGPB Azospirillum, Aeromonas, Azotobacter, Bacillus, Paenibacillus, Burkholderia, Enterobacter, Pantoea, Pseudomonas and Rhizobium are able to secrete plant hormones (Khabbaz et al., 2019). Plant hormones generally include abscisic acid, cytokinin, ethylene, gibberellin and indole-3-acetic acid (IAA); among them, IAA and ethylene are the two classical phytohormones in the plant-bacteria interaction (Stegelmeier et al., 2022).

Endophytic bacteria indirectly promote the growth of host plants by inhibiting plant pathogens and pests with the production of antagonistic compounds such as antibiotics, toxins, siderophores, lytic enzymes, and antimicrobial volatile organic compounds (Figure 2) (Elnahal et al., 2022). Among them, Actinobacteria, Bacillus, Enterobacter, Paenibacillus, Pseudomonas and other bacteria are common genera with antibacterial activity against plant pathogens (Afzal et al., 2019). Some of the enzymes produced by PGPB, including chitinase, cellulase, glucanase, protease and lipase, can lyse a portion of the cell walls of pathogens (Ghosh et al., 2021). PGPB that secrete one or more of these enzymes have been shown to have biocontrol activities against a wide range of competitor pathogens (Selari et al., 2023). Naama-Amar et al. (2022) reported that a candidate biocontrol agent Frateuria defendens could inhibit Spiroplasma melliferum growth by secreting antimicrobial metabolites (Naama-Amar et al., 2022). A plant growth-promoting endophytic bacterium Pseudomonas protegens MP12 can produce important antifungal compounds (2, 4-diacetylphloroglucinol, pyoluteorin and pyrrolnitrin), and exhibits inhibitory effects on the mycelial growth of phytopathogens such as Botrytis cinerea, Alternaria alternata, Aspergillus niger, Penicillium expansum and Neofusicoccum parvum (Andreolli et al., 2019). Bacillus produced antibiotic lipopeptides, such as iturin, bacillomycin, bacilysin, fengycin, surfactin, and zwittermycin, which have broad spectrum of action against pathogens (Saravanakumar et al., 2019).

PGPB can also activate plant-induced systemic resistance (ISR) against a broad spectrum of pathogens (Datta et al., 2022). ISR is phenotypically similar to systemic acquired resistance (SAR) that occurs when plants activate their defense mechanisms in response to pathogen infection (Meena et al., 2020). The prime function of ISR is to activate plant defense mechanisms and protect unexposed parts of plants against pathogenic microbe and herbivorous insect invasions (Sardar et al., 2021). PGPB-mediated ISR depones on jasmonic acid and ethylene signaling pathways in plants (Abid and Karim, 2021). These hormones stimulate the host plant’s defense responses to a range of pathogens without the need for PGPB to interact directly with pathogens. In addition to involvement in jasmonate and ethylene signaling pathways, ISR also plays a crucial role in host resistance by enhancing the activity of pattern recognition receptors through cellular or hormonal defenses (Hossain and Chung, 2019).

PGPB can also emit volatile organic compounds (VOCs) as biocontrol factors or deterrents against pathogens (Choub et al., 2022). Recently, many studies have reported the production of VOCs secreted by PGPB that disrupted cell membrane integrity, spore germination and mycelial growth of plant pathogenic fungi and other competing bacteria species (Chatterjee and Niinemets, 2022). Recent advances in analytical science such as the headspace solid-phase microextraction/gas chromatography–mass spectrometry (HS-SPME/GC–MS), have been used for extraction, identification, and characterization of VOCs emitted by PGPB and other bacterial species (Drabińska et al., 2022; Tabbal et al., 2022; Wu et al., 2022). Additionally, some VOCs can repress the expression of virulence traits involved in host colonization, such as motility, root colonization, and biofilm formation, thus, for example, effectively controlling tomato wilt (Raza et al., 2016). In summary, VOCs produced by PGPB help plants cope with abiotic and biotic stresses by inducing systemic resistance or inhibiting the growth of a wide range of pathogen.

In addition to producing substances that harm or inhibit plant pathogens, some PGPB may compete for nutrients or niches with pathogens (Selari et al., 2023). In fact, it is widely believed that this competitiveness of PGPB works together with other biocontrol mechanisms to repress the growth of plant pathogens (Olanrewaju et al., 2017). It is important to note that the effectiveness of PGPB depends on their versatility and adaptation to new niches as well as their ability to colonize and compete with other members of plant microbiome (Rilling et al., 2019). Root exudates contain chemo-attractants like amino acids, organic acids and specific sugars that are good carbon sources for PGPB activity and competitive colonization (Feng et al., 2021). Additionally, the mutual signal exchange between plants and microsymbionts, along with flagellar motility, further enhances the affinity of PGPB in root surface (Raina et al., 2019).

In addition to beneficial traits such as IAA production, nitrogen fixation and phosphate dissolution, PGPB can accumulate osmotic adjustment substances under stressful conditions (Figure 2) (Feng et al., 2023). Under drought and salt stress conditions, plant cells can improve the water activity of cells and maintain normal metabolic activities by accumulating compatible solutes (soluble sugars, polyols, glycine betaine, proline, free amino acids) (Ali et al., 2022). Under abiotic stress, one of the most significant changes is the accumulation of plant proline that can improve the activity of various enzymes, stabilize intracellular pH and remove reactive oxygen species, thereby maintaining antioxidant activity (Abdelaal et al., 2021). Many PGPB can induce plant proline synthesis under stress conditions, which helps to maintain cell osmotic pressure and improve plant salt tolerance (Khabbaz et al., 2019).

Trehalose, a non-reducing disaccharide, can act as a protective agent to alleviate a variety of environmental stresses, including drought, salinity and extreme temperatures (Greco et al., 2021). When cells are dehydrated, trehalose can form a gel phase to bind with surrounding macromolecules and membranes thereby replacing water loss and decreasing the damage to cells (del Carmen Orozco-Mosqueda et al., 2022). In addition, trehalose prevents the degradation, denaturation and aggregation of proteins under high and low temperature stress (Vinciguerra et al., 2022). Under stress conditions, trehalose and ACC deaminase synergistically protect crops from stress by stabilizing biological structures and lowering ethylene levels, respectively (del Carmen Orozco-Mosqueda et al., 2022). Compared with wild type Rhizobium etli, the number of nodules, nitrogenase activity and biomass of Phaseolus vulgaris increased when inoculated with R. etli overexpressing trehalose synthase, and the host plant inoculated with mutants recovered more strongly under drought stress (Suárez et al., 2008). Although it is also possible to overexpress trehalose genes in transgenic plants, it is much simpler to achieve the same goal through co-inoculation with PGPB. Moreover, since most strains have no specific host, therefore, the same strain can effectively help different plants mitigate stress.

Biofilms are structurally complex microbial communities attached to living or abiotic surfaces and surrounded by complex extracellular polymers. Microbial colonization of plant roots can be promoted by the formation of biofilms (Niu et al., 2020). Gallique et al. (2017) reported that mutations in three T6SS-related genes (hcp1, hcp2, or hcp3) of Pseudomonas fluorescens MFE01 did not reduce biofilm formation, but these three Hcp proteins are essential for the formation of mature biofilm structures. However, in another similar study, these three T6SS-related gene mutants of P. aeruginosa PAO1 had no impact on biofilm formation or environmental adaptation ability, but swarming motility was reduced, which is another important aspect of biofilm formation ability, suggesting that T6SS is associated with biofilm formation but not environmental adaptation (Chen et al., 2020). The T6SS gene cluster present in the Acidovorax citrulli was confirmed to be involved in multiple biological processes, including colonization, competition and biofilm formation (Fei et al., 2022).

Evidence is also emerging that T6SSs could contribute to inter-bacterial competition. Interestingly, many PGPB harbor one or more T6SS, however, the function of T6SS in PGPB is poorly understand. In a previous study, it was found that the T6SS of PGPB Azospirillum brasilense provides antibacterial activities against a number of plant pathogens in vitro, and might confer T6SS-dependent bio-control protection to microalgae and plants against bacterial pathogens (Cassan et al., 2021). However, Lin et al. (2018) found instead of being an antihost or antibacterial weapon of the bacterium, the T6SS in Azorhizobium caulinodans ORS571 seems to participate specifically in symbiosis by increasing its symbiotic competitiveness.

Most organisms with T6SS are not pathogenic and are found in diverse environments as symbionts. Although the secretory system was thought to be involved in the interaction between PGPB and plants, only recently research has been functionally dissected in nitrogen-fixing Azoarcus olearius, in which T6SS positively affected plant colonization efficiency (Jiang et al., 2019). The T6SS effector Hcp1 up-regulated in rice inoculated with a well-known growth-promoting rhizobacteria Herbaspirillum rubrisubalbicans, suggesting that T6SS plays a role in rice colonization (Valdameri et al., 2017). Similarly, the T6SS mutant of PGPB Enterobacter sp. J49 showed a significant decrease in the epiphytic and endophytic colonization, indicating that although T6SS is not essential, it may participate in bacterial colonization (Lucero et al., 2022). Salinero-Lanzarote et al. (2019) described the growth-promoting trait of T6SS in Rhizobium etli Mim1, which can effectively increase nodule size in legumes, suggesting that T6SS plays an active role in the Rhizobium-legumes symbiosis. The knockout mutant of T6SS gene cluster in rice endophyte Kosakonia showed a decreased ability to colonize the rhizoplane and endosphere, suggesting that T6SS is involved in the colonization process of plant-bacteria interaction (Mosquito et al., 2020). Another plant-associated microbe, P. taiwanensis, has been shown to use its T6SS to mediate colonization resistance against bacterial plant pathogens through T6SS-mediated secretion of pyoverdine, an iron chelator (Chen et al., 2016). These results suggest that there may be a relationship between T6SS and plant growth promoting properties.