The aerial parts of living plants including leaves, stems, buds, flowers and fruits provide a habitat for micro-organisms termed the phyllosphere. Bacteria are considered to be the dominant microbial inhabitants of the phyllosphere, although archaea, filamentous fungi and yeasts may also be important. These microbes can be found both as epiphytes on the plant surface and as endophytes within plant tissues (Arnold et al. 2000; Inacio et al. 2002; Lindow and Brandl 2003; Stapleton and Simmons 2006). The global surface area of the phyllosphere has been estimated to total over 4 × 108 km2, supporting bacterial populations in the region of 1026 cells (Morris and Kinkel 2002). Furthermore, recent estimates of the diversity of phyllosphere bacteria in the 20 000 vascular plants inhabiting the Brazillian Atlantic forest, suggests the possible occurrence of 2–13 million phyllosphere bacterial species in this habitat alone (Lambais et al. 2006).
The phyllosphere represents a niche with great agricultural and environmental significance. There is growing evidence for important interactions of phyllosphere microbial inhabitants that may affect the fitness of natural plant populations and the quality and productivity of agricultural crops. Phyllosphere bacteria can promote plant growth and both suppress and stimulate the colonization and infection of tissues by plant pathogens (Lindow and Brandl 2003; Rasche et al. 2006a). Similarly, fungal endophytes of leaves may deter herbivores, protect against pathogens and increase drought tolerance (Arnold et al. 2003; Schweitzer et al. 2006). Furthermore, interactions in the phyllosphere zone determine the extent to which human pathogens are able to colonize and survive on plant tissues, an area of increasing importance with the rise in cases of human disease associated with consumption of fresh salad, fruit and vegetable produce (Whipps et al. 2008).
There is evidence for functional roles within the phyllosphere microbial community which given the size of the habitat could have global significance. The best studied of these is nitrogen fixation. Measured rates of bacterial nitrogen fixation in the phyllosphere vary widely, but in the phyllosphere of trees in some tropical habitats rates of over 60 kg N ha−1 have been reported, although amounts fixed in the phyllosphere of temperate trees is generally considerably lower (Freiberg 1998). Furthermore, nitrogen fixation or the presence of nitrogen-fixing bacteria has been reported in the phyllosphere of many crop plants (e.g. Murty 1983; Miyamoto et al. 2004). Other environmentally important microbial processes for which there is evidence in the phyllosphere include methanol degradation (Corpe and Rheem 1989; Van Aken et al. 2004) and nitrification (Papen et al. 2002), although the rates of these processes and their ubiquity within the phyllosphere remains to be elucidated.
Most knowledge of the structure and activities of phyllosphere microbial communities has been established using culture-dependant methods. However, these are recognized to significantly underestimate diversity, with only 0·1–3% of environmental bacteria considered culturable (Wagner et al. 1993). Data gathered using these methods therefore relate only to culturable members of the community and provide no information on the vast majority of microbes present in samples. As in other areas of environmental microbiology, the recent application of culture-independent methods based on the characterization of small subunit rRNA gene sequences for microbial community analysis is providing new insights into the complexity of phyllosphere microbial communities and their interactions with plants and the wider environment.
In the current paper, we review the extent to which the use of culture-independent approaches has changed our understanding of the structure and diversity of phyllosphere communities. A variety of plant, microbial and environmental factors control establishment of microbial communities in the phyllosphere, but recently there has been recognition of the role that plant genotype plays in selecting phyllosphere communities. The evidence for the different factors that regulate the structure of phyllosphere communities is discussed with special reference to the role of plant genotype. Finally, we suggest approaches and priority areas for future research on phyllosphere microbiology. Although much of the phyllosphere literature is concerned with interactions between plant and plant pathogens (bacteria and fungi that cause diseases in plants), in the current review emphasis is placed on studies of micro-organisms that live in the phyllosphere without causing obvious damage to the plant, as absence of disease is the normal situation in nature.
Microbial diversity in the phyllosphere
The microbial communities of the phyllosphere are diverse, supporting numerous genera of bacteria, filamentous fungi, yeasts, algae and in some situations protozoans and nematodes (Morris et al. 2002; Lindow and Brandl 2003). Bacteria are the most numerous and diverse colonists of leaves, with culturable counts ranging between 102 and 1012 cells per gram of leaf (Thompson et al. 1993; Inacio et al. 2002). Culture-based studies of sugar beet over the whole of the growing season have found more than 78 bacterial species representing 37 known bacterial genera (Thompson et al. 1993). Similar studies in wheat have revealed 88 bacterial species representing 37 known bacterial genera (Legard et al. 1994).
Recent studies have demonstrated that profiling of phyllosphere communities based on culture-dependent methods is likely to be inaccurate and to underestimate diversity (Rasche et al. 2006b). In the case of the phyllosphere, use of culture-independent approaches has shown that although assumptions regarding the dominant inhabitants are largely correct, the diversity of phyllosphere communities is far greater than previously recognized. Analysis of 16S rDNA cloned directly from leaf samples has demonstrated that proteobacteria are the dominant group found on leaves (Table 1), confirming data obtained using culture-dependant methods (e.g. Thompson et al. 1993). α- and γ-proteobacteria are generally the dominant bacterial inhabitants of the phyllosphere, with bacteroidetes being usually important. β-proteobacteria and firmicutes can also form a large part of the bacterial community in some situations, with acidobacteria, actinobacteria and cyanobacteria occurring infrequently (Kadivar and Stapleton 2006; Idris et al. 2004; Lambais et al. 2006; Rasche et al. 2006b,c).
In a study of phyllosphere bacterial communities in a tropical Brazillian forest, 97% of the bacterial sequences were from previously undescribed species with phyllospheres of different plant species supporting from 95 to 671 bacterial species (Lambais et al. 2006). The extent to which such diversity occurs in other plant species is unclear. Those sequences showing 95% or less homology to known bacterial species database entries comprised 15·2% of the sequences obtained from Thlapsi geosingense (Idris et al. 2004), and 7·9, 2·3, 3·5 and 1·2% of those sequences obtained from Crocus, potato, pepper and maize, respectively (Kadivar and Stapleton 2003; Rasche et al. 2006a,b; Reiter and Sessitsch 2006). However, in a study of a range of temperate agricultural crop species, 5 of 17 bands cut from 16S rRNA denaturant gradient gel electrophoresis gels had less than 90% similarity to database entries, suggesting that in some situations phyllospheres of crop plants may support large numbers of novel bacteria (Yang et al. 2001). The number of sequences investigated in the culture-independent studies conducted to date has been limited, so that only dominant members of the community are likely to have been detected, and the true extent of bacterial diversity in the phyllosphere therefore remains to be determined.
Yeasts are the major epiphytic fungal group in the phyllosphere with filamentous fungi largely occurring as dormant spores rather than active mycelia except on older leaves (Andrews and Harris 2000; de Jager et al. 2001). Culturable yeast populations can range between 10 and 1010 CFU g−1 leaf (Thompson et al. 1993; Inacio et al. 2002). The diversity of culturable yeasts appears to be mostly limited to the genera Cryptococcus, Sporobolomyces and Rhodotorula, although total species number can reach over 40, with multiple species of each coexisting in the phyllosphere, together with a number of other genera that occur less frequently (Thompson et al. 1993; Inacio et al. 2002; Glushakova and Chernov 2004).
Filamentous fungus population sizes can range between 102 and 108 CFU g−1 leaf. Cladosporium and Alternaria are usually considered the most abundant fungi found on leaves, although several other genera, including Penicillium, Acremonium, Mucor and Aspergillus are also found (Thompson et al. 1993; Inacio et al. 2002). Filamentous fungi appear to occur ubiquitously as endophytes, the diversity of which may be substantial, particularly in long-lived tropical leaves. Using culture-dependant approaches, over 340 genetically distinct taxa were recovered from individuals of two tropical forest understory plant species at two sites. Furthermore, there was evidence for host preference within the endophyte community (Arnold et al. 2000). Culture-independent approaches have not yet been used to characterize fungal diversity in the phyllosphere.
There are various developing technologies that show promise to significantly increase throughput of analysis to provide a finer resolution of understanding about the diversity and structure of phyllosphere communities and to link diversity with functioning. Culture-independent analysis using phylogenetic specific primers represents a powerful method to investigate the dynamics and distribution of specific bacterial groups of interest (Sessitsch et al. 2002; Miyamoto et al. 2004). Additionally multiplex terminal restriction fragment length polymorphism (TRFLP), in which several phylogentic groups or functional genes can be analysed at the same time, provides an opportunity to improve throughput of samples in a cost-effective manner (Singh et al. 2006). However, these techniques remain time-consuming, and future developments will depend on high throughput methods. Phylogenetic microarrays clearly provide a way forward, allowing the presence and amount of thousands of micro-organisms to be determined simultaneously, and could also be used to detect novel members of the phylogenetic groups. Similarly, functional gene arrays provide a means of characterizing activity of the phyllosphere community, and when used with phylogenetic microarrays, for linking community structure to function (Sessitsch et al. 2006).
In order to understand and predict the diversity and structure of phyllopshere communities, it is necessary to understand the biological and environmental factors that control the establishment and dynamics of microbial communities on the leaf surface. This is the focus of the remainder of the review.
Sources of microbes colonizing the phyllosphere
The sources of micro-organisms on the phyllosphere can be manifold. Epiphytic filamentous fungi, yeasts and bacteria may arrive on the leaf surface through insect-, atmosphere-, seed- or even animal-borne sources. Tree buds, seeds of annual plants and the debris from previous crops are likely to be the most important sources for the colonization of new plants and leaves as they are a major source of bacteria already adapted to the phyllosphere (Manceau and Kasempour 2002).
Those micro-organisms that show no or limited multiplication in the phyllosphere are considered transient epiphytes whereas those with the capacity for multiplication in the absence of wounds are known as residual epiphytes (Suslow 2002). Microbial populations can vary in size among and within plant species over short periods of time (Hirano and Upper 1989) as well as over the growing season (e.g. Thompson et al. 1993; Legard et al. 1994; Inacio et al. 2002), with few epiphytic bacteria present on leaves shortly after emergence from buds or seeds, but increasing in quantity subsequently (Hirano and Upper 1993). There is a general succession of microbial populations on leaves over the growing season with bacteria dominating initially, followed by yeasts and finally filamentous fungi (Kinkel 1997).
The atmospheric microflora can vary in composition and concentration diurnally and seasonally as well as in response to environmental events such as rainfall and high wind (Kinkel 1997; Zak 2002), directly influencing the immigration of micro-organisms to the phyllosphere. Local vegetation, and in areas of crop production, agricultural practices such as harvesting and cultivation, also influence atmospheric microbiology and colonization of nearby plants (Lindemann et al. 1982; Lacey 1996; Lighthart 1997). Immigration of micro-organisms to leaves from the atmosphere can take place through impaction onto the leaf surface, sedimentation or rain splash as well as from contamination with soil (Venette and Kennedy 1975; Lacey 1996).
There is increasing evidence that micro-organisms on seeds or roots can become endophytic in the roots, enter the vascular system and be transferred internally to the aerial parts of plants where they establish as phyllosphere endophytes (Lamb et al. 1996; Wulff et al. 2003). Endophytes can also arise from ingression into the internal leaf spaces following colonization by epiphytes, suggesting that epiphytes and endophytes are really part of a continuum in the phyllosphere (Beattie and Lindow 1999; Wilson et al. 1999).
Micro-organisms that have arrived in the phyllosphere have to become established and colonize the leaf to become a residual epiphyte. The pattern of distribution of micro-organisms on leaves is not even. The most common sites of bacterial colonization are in the epidermal cell wall junctions (Blakeman 1985; Davis and Brlansky 1991) especially in protected sites in grooves along the veins (Mansvelt and Hattingh 1987; Leben 1988; Mariano and McCarter 1993), at stomata (Mew and Vera Cruz 1986; Mariano and McCarter 1993) and at the base of trichomes (Mew and Vera Cruz 1986; Mansvelt and Hattingh 1987; Mariano and McCarter 1993). They are also found under the cuticle (Corpe and Rheem 1989), in depressions in the cuticle (Mansvelt and Hattingh 1987), near hydrathodes (Mew et al. 1984) and in specific sites that only occur on particular plants such as stomatal pits in oleander and pectate hairs in olive (Surico 1993). In general, greater numbers of bacteria are found on lower than upper leaf surfaces (Leben 1988; Surico 1993) possibly because of the lower leaf surface having a greater density of stomata or trichomes, or a thinner cuticular layer (Beattie and Lindow 1999).
Bacterial populations in the phyllosphere can differ in distribution over very small scales, as little as 0·1 mm2 (Kinkel et al. 1995) and are often well-described by a log-normal distribution (Hirano et al. 1982; Ishimaru et al. 1991) whereas yeasts and filamentous fungi may be better described by a normal distribution (Kinkel et al. 1989). Micro-organisms may occur individually on the leaf surface but frequently, they occur as aggregates or biofilm-like structures containing bacteria (Kinkel et al. 1995; Morris et al. 1997, 1998; Jacques et al. 2005), yeasts (Last 1955) and filamentous fungi (Bernstein and Carroll 1977).
Clearly, not all micro-organisms that arrive in the phyllosphere are able to colonize and grow. To some extent this reflects processes of emigration through dispersal mechanisms such as rain splash, wash-off, bounce-off, water movement or removal by insects (Kinkel 1997). Ability to survive and grow are dependent on the environmental, physicochemical and genetic features of the plant and specific properties exhibited by the phyllosphere micro-organisms, which together determine the structure and diversity of the microbial community. Evidence for such selection is supported by the findings of Miyamoto et al. (2004), in which 16S rRNA-TRFLP with Clostridia-specific primers was used to show the presence of diverse Clostridia populations within Miscanthus sinensis, which were shown to be distinct to those Clostridia populations inhabiting soil around plants. Furthermore, as a substantial proportion of those bacteria inhabiting the phyllosphere appear to be novel to this habitat there have been suggestions that some may be unique or specialists to this habitat (Yang et al. 2001; Lambais et al. 2006).
There are a number of areas relating to the colonization of phyllosphere that require more complete understanding. The transmission of micro-organisms from roots to aerial parts of plants appears to have been a neglected area of research and the importance of this environmentally protected phyllosphere colonization route needs to be elucidated. This could be particularly important for soils contaminated with human pathogens.
Leaf characteristics and environmental factors controlling phyllosphere microbiology
Following arrival of microbial cells or propagules on the leaf surface, a variety of factors determine whether cells are able to colonize the leaf, and where cells become localized. Establishment is determined by interaction between leaf and environmental characteristics that interact to control conditions prevailing in the phyllosphere habitat. The first point of contact of microbial cells immigrating to the phyllosphere is the cuticle (Beattie 2002). This waxy surface, often microcrystalline in nature, serves several functions: a diffusion barrier, reducing water and solute loss and aqueous pollution ingress; as a reflectant to minimize temperature fluctuations; conferring water repellence; and providing protection from pathogens (Beattie 2002). The water repellence is particularly important in preventing immigration of micro-organisms to the leaf surface. This is so especially on young leaves where the cuticle is intact relative to older leaves because as the cuticle erodes, wettability increases (Beattie 2002). In addition, cuticle-mediated limitation of nutrient loss from the leaf is particularly important in supporting epiphytic microbial populations. Use of whole cell bacterial biosensors for sucrose, fructose and glucose have revealed that these sugars are present only in discrete localized sites on the leaf (Leveau and Lindow 2001; Miller et al. 2001). This, and recent microscopical evidence (Monier and Lindow 2005), suggests that most microbial immigrants to the phyllosphere are exposed to nutrient poor environments and that only few cells randomly land in zones of relatively abundant nutrients that support growth. Other discrete sites of nutrient loss such as wounds or glandular trichomes (Monier and Lindow 2005), or sites of nutrient enrichment including pollen (Diem 1974) or honeydew (Dik and Vanpelt 1992) also provide sites for microbial growth. Other nutrients such as nitrogen sources or iron are not considered as growth limiting to micro-organisms on the phyllosphere as the carbon sources (Lindow and Brandl 2003). Interestingly, when the resurrection fern, Polypodium polypodioides, is exposed to rainfall after a period of desiccation, the complex phyllosphere community undergoes changes in overall structure and activity, reflecting use of labile organic substrates in the form of an enrichment culture (Jackson et al. 2006). Whether this occurs with other plants is unknown.
Plant leaves also release a wide range of volatile organic compounds into the boundary layer around leaves. These can include small molecules such as CO2 and acetone, medium-sized molecules including terpenoids and a number of aldehydes and alcohols, as well as large molecules such as long-chain hydrocarbons and sesquiterpenoids; sulfides and nitrogen-containing compounds also occur. It is unclear whether these could be nutrient sources directly, but there is evidence that some of these compounds can be inhibitory or toxic to some fungi (Mechaber 2002). Similarly, some proteins secreted by glandular trichomes can inhibit some pathogens (Shepherd et al. 2005). There are also data to suggest that plants can release a number of compounds in response to damage that not only promote microbial development but can selectively inhibit microbial growth as well (Dingman 2000).
Characteristics of the plant species themselves may also influence the microbial carrying capacity of the leaf. The total number of culturable bacteria from broad leaf succulent herbaceous plants such as cucumber, lettuce and bean can be significantly higher than that from grasses or waxy broad-leaved plants such as cabbage and citrus (O’Brien and Lindow 1989; Lindow and Andersen 1996; Kinkel et al. 2000). Culture-independent approaches have demonstrated that community structure on leaves from individuals of the same species is similar, but varies significantly between species (Yang et al. 2001). Lambais et al. (2006) showed that just 0·5% of the bacterial species recorded in tropical tree canopies were common to all tree species. Furthermore, both bacterial and fungal population size on leaves has been correlated with leaf position, plant architecture and height in the canopy (Wildman and Parkinson 1979; Oliveira et al. 1991; Jacques et al. 1995). We would suggest that microbial diversity and community structure are also influenced by these factors, although this remains to be shown.
Information is needed to characterize the arrangement and dynamics of communities in time and space, especially at the landscape scale. In particular the relative importance of environmental factors, location and plant species in determining the composition and dynamics of phyllosphere communities needs to be addressed. Biogeographical approaches to the analysis of microbial communities show potential to elucidate these fundamental relationships (Ramette and Tiedje 2007).
Microbial adaptations to the phyllosphere habitat
In addition to plant and environmental factors, properties of the microbial colonists themselves determine the extent to which they are able to establish on the leaf surface. For some micro-organisms, this reflects their inherent ability to survive in the existing habitat whereas others are capable of modifying the environment to ameliorate the levels of stress they are exposed to.
Culture-independent analyses have indicated that tolerance to ultraviolet (UV) radiation is likely to be an important selection pressure for survival and growth in this habitat (Kadivar and Stapleton 2006; Stapleton and Simmons 2006), and most isolated phyllosphere micro-organisms are capable of withstanding high UV radiation levels on the leaf surface (Sundin 2002). In fungi, dark melanin-type pigments are thought to play a key role as protective pigments along with UV-B-induced hyphal-wall thickening, the latter protecting lower levels of the fungal colony (Fourtouni et al. 1998; Sundin 2002). Interestingly, the most UV-B-tolerant strains of bacteria from the peanut phyllosphere were those that produced pink or orange pigments (Sundin and Jacobs 1999) and so multiple UV-B protectant mechanisms may be exhibited by phyllosphere micro-organisms.
A low level of water availability and nutrients are key limiting factors for microbial growth in the phyllosphere so epiphytes display a variety of mechanisms to overcome these limitations. For example, some epiphytic Pseudomonas spp. can release surfactants that increase the wettability of leaf surfaces making it easier for micro-organisms to use water and increasing solubilization and diffusion of nutrients, thereby increasing substrate availability to epiphytic bacteria (Bunster et al. 1989). A number of phyllosphere bacteria have recently been shown to increase permeability of the cuticle enhancing water and nutrient availability in the phyllosphere (Schreiber et al. 2005). Another, potentially related, mechanism to increase nutrient availability may relate to the ability to produce toxins that affect ion transport across plant cell plasma membranes (Quigley and Gross 1994; Hutchison et al. 1995). Plant pathogenic Pseudomonas syringae pv. syringae secrete the toxin syringomycin that eventually leads to cell lysis. Nevertheless, low levels are produced by nonpathogenic epiphytic strains of P. syringae pv. syringae such that necrosis and disease do not occur although release of plant nutrients is still stimulated (Hutchison et al. 1995). Interestingly, syringomycin also acts as a surfactant providing two possible mechanisms to enhance nutrient availability in the phyllosphere.
Another, perhaps more widespread mechanism, is the production and release of plant growth regulators. Production of indole-3-acetic acid (IAA) is common among bacterial epiphytes (Glickmann et al. 1998; Brandl et al. 2001) and is associated with enhanced nutrient leakage and microbial fitness (Brandl and Lindow 1998; Manulis et al. 1998). Lindow and Brandl (2003) have also made the suggestion that presence of a functional type III secretion pathway in Pseudomonas fluorescens and Pseudomonas putida (Preston et al. 2001), which provides the capacity to modify the local habitat, may be needed for growth and survival in the phyllosphere. Production of pili and flagellae may also be important in allowing bacterial attachment and colonization of the phyllosphere (Romantschuck et al. 2002). A whole range of genes and gene products that are important for phyllosphere colonization are now being identified using molecular techniques (Gal et al. 2003; Gourion et al. 2006) and may provide further insights into mechanisms involved in epiphytic growth.
As mentioned earlier, bacterial distribution on the leaf surface is not uniform and frequently, aggregates of cells occur (Morris and Monier 2003). The presence of these aggregates may provide the epiphytes with an ability to colonize and survive in the phyllosphere and modify the local environment. The production of extracellular polysaccharides (EPS) by bacteria may protect the bacteria from water stress and help anchor the cells to the leaf surface (Morris et al. 1997; Gal et al. 2003). By analogy with biofilms (Morris et al. 2002; Morris and Monier 2003), these aggregates may also protect from UVR, predation and bacteriocides, moderate pH and gas exchange, enhance genetic exchange particularly through plasmid transfer and allow cell density-dependent behaviour. The latter, often mediated by accumulation of diffusible molecules such as N-acyl homoserine lactones through quorum sensing may have numerous effects on microbial behaviour including EPS and antibiotic production as well as pathogenicity traits (Swift et al. 1994; Greenberg 1997). Interestingly, if signalling controls the formation and functioning of aggregates, it may be possible in future to manipulate the microbial populations on the phyllosphere if the molecular signals and receptors essential for aggregate behaviour can be identified (Morris et al. 2002).
Plant genotype and phyllosphere microbiology
It is clear that microbial populations in the phyllosphere can vary markedly in size and composition both spatially and temporally on the same plant, differ between different plants and parts of plants in the same place, and even differ on the same plant species in different places (Lindemann et al. 1984; Morris and Lucotte 1993; Lindow and Andersen 1996; Kinkel 1997). Much of the variability must reflect the environmental conditions prevailing at the time and place of sampling, thereby influencing the processes of microbial immigration, emigration, growth and death. However, the microbial population that does develop must relate to a large extent to the phenotypic characteristics exhibited by the plant, controlled ultimately by its genetic make-up. Certainly, there are ‘hot-spots’ of microbial growth on the leaf associated with specific sites and it would be expected that these would similarly be under the influence of plant genetic characteristics. We suggest that within plant species, genotype has a key role in determining colonization and establishment of microbial communities in the phyllosphere. However, few studies have addressed the relationship between plant genetic control of phenotypic characteristics and their concomitant effects on microbial populations in the phyllosphere, despite its potential importance.
Several studies have used culture-dependent approaches to investigate the impact of plant genotype on phyllosphere microbiology. Adams and Kloepper (2002) showed that endophytic bacteria population sizes and structure differed among nine cotton cultivars, and in pea 5 of 11 cultivars were found to contain endophytic bacteria with one showing a higher colonization level than the others (Elvira-Recuenco and van Vuurde 2000). In a gnotobiotic system with tomato, one cultivar out of four supported fewer Pseudomonas sp. on the shoot exterior following application of the bacterium to the seed (Pillay and Nowak 1997). Differences in ability to support populations of Pseudomonas syringae pv. syringae were also found between different cultivars of snap bean (Hirano et al. 1996; Upper et al. 2003). However, no differences in occurrence of native, epiphytic mycoparasites were observed between the three main coffee cultivars or between clones of the same group (ten Hoopen et al. 2003). Similarly, no differences were found between epiphytes on three cultivars of apple (Becker and Manning 1983) or in endophytes in three cultivars of wheat (Larran et al. 2002).
Culture-independent community profiling approaches have been particularly valuable for elucidating interactions between plant genotype and phyllosphere microbial community structure. Several studies have indicated that different cultivars of the same plant species exhibit different phyllosphere microbial populations. Phyllosphere populations of bacteria were found to differ between cultivars of sweet pepper and tomato (Rasche et al. 2006b; Correa et al. 2007) and both endophytes and epiphytes differed between varieties of potato (Sessitsch et al. 2002; Rasche et al. 2006a,c). Plant genotype differences may affect some microbial communities more than others. For example, phyllosphere bacterial community structure was shown to vary between wheat cultivars, although there was found to be no difference in archaeal communities (Stapleton and Simmons 2006). Recently, lettuce cultivar was shown to affect colonization of leaves by Salmonella enterica serovars (Klerks et al. 2007), with significant serovar–cultivar interactions demonstrated. Furthermore, diversity of endophyte bacterial populations varied between the three lettuce cultivars used, and data suggested that the degree to which S. enterica serovars were able to colonize plants endophytically was in part determined by competitive interactions with the natural endophyte bacterial community.
Culture-dependent analysis showed that genetic modification of potato with an antibacterial peptide, magainin, failed to influence the number or structure of phyllosphere bacterial or fungal populations although magainin-expressing potato tubers did exhibit lower total numbers of bacteria than unmodified plants (O’Callaghan et al. 2004). In contrast, genetic modification of potato with a gene producing antibacterial T4-lysozyme or attacin/cecropin, was shown to induce greater difference in phyllosphere microbial community structure to the parent line, relative to variations between three cultivars (Rasche et al. 2006a,c). However, field site and plant growth stage had greater effects on bacterial community structure than either cultivar or genetic modification.
Furthermore, microbial communities selected by different genotypes can show differing responses to environmental variables. Rasche et al. (2006b) showed that chilling sweet pepper plants altered endophyte bacterial community structure, with the extent of the effect differing between cultivars, and dependant on cultivar chilling tolerance. Similarly, in a study of wheat cultivars, it was shown that the response of phyllosphere bacterial communities to UV-B radiation depended on host genotype. However, it was not clear whether these differences reflected direct effects on the bacterial community or indirect effects associated with differences in the plant responses to UV-B (Stapleton and Simmons 2006). Furthermore, plant genotype can influence colonization and survival of microbial inoculants in the phyllosphere. Correa et al. (2007) showed that the survival of a plant growth promoting Azospirillum inoculant differed in the phyllosphere of contrasting tomato genotypes, and that the response of the phyllosphere bacterial community to inoculation varied between genotypes.
In the case of fungi, there is limited data on plant genotype-diversity relationships, although several studies have demonstrated differences in the nature of endophytes associated with contrasting host genotypes. For example, distinct communities of endophytes were shown to be associated with different Populus hybrids, with the percentage of condensed tannins in bark implicated in directing these differences (Schweitzer et al. 2006).
Although plant genotype appears to be an important factor determining the structure of phyllosphere microbial communities, the mechanisms controlling these interactions remain to be elucidated. Various plant science resources are available that show potential for examining plant genotype–phyllosphere microbiology interactions. In particular recombinant inbred mapping populations (Asins 2002) have the potential to identify plant genes controlling leaf microbiology.
Conclusions and future directions
Although culture-independent molecular analysis of microbial populations in the phyllosphere is still in its infancy, it is clear from recent studies that phyllosphere microbiology is greatly more complex than previously understood. Although progress has been made in elucidating the structure and distribution of microbial communities in the phyllosphere, very little is known of the functional consequences of the community or its composition for the fitness of individual plants, the quality and microbiological safety of fresh produce and wider environmental processes. Microbes reach the phyllosphere by atmospheric deposition from plant and soil sources, but may also colonize plants through the roots, and become transported to aerial parts. The relative importance of these mechanisms remains to be determined. Although microbial establishment and colonization has long been recognized to be the result of interplay between plant and environmental factors, and the physiological characteristics of microbial colonists, it has now been clearly demonstrated that within plant species, contrasting genotypes can support different microbial communities. This understanding provides opportunities to understand the molecular mechanisms by which plants control microbial populations in the phyllosphere. Such studies could provide methods to manipulate phyllosphere communities via plant genotype, providing exciting opportunities to manage applied aspects of phyllosphere microbiology, such as the survival of human pathogens or the activity of beneficial microbes.