Abstract
The rate at which entire microbial genomes are being sequenced has accelerated rapidly over the past two years, promising to revolutionise our understanding of microbial molecular biology and genetics. The Bacillus subtilis genome sequence is the first complete genome of a free-living soil and rhizosphere bacterium. Data derived from the genome sequence and the systematic functional analysis programme, together with the wealth of knowledge already available for this organism, open up new opportunities to study the behaviour and ecology of this soil and plant growth-promoting rhizobacterium at the molecular level. In this review we examine the Bacillus subtilis 168 genome sequence in the light of clues it might provide for the role of this species in natural environments and discuss suitable methods for applying the available data and resources to the study of this and related organisms in natural systems.
1 Introduction
The recent publication of the genome sequence of the Gram-positive bacterium Bacillus subtilis[1] is an important landmark in the study of microbial ecology since it represents the first published genome sequence for a soil-living bacterium. The combination of its genetic blueprint, intensively studied biochemistry and physiology and extreme genetic amenability makes B. subtilis an unrivalled bacterium in which to investigate molecular responses to life in the soil in general and the rhizosphere in particular. In this review we use B. subtilis as a model to illustrate the value of genome sequence data for advancing studies in microbial ecology.
2 The taxonomy and distribution of Bacillus spp.
The genus Bacillus, of which B. subtilis is the type strain, is both taxonomically and metabolically diverse. The primary habitat is the soil and associated plants, rivers and estuarine waters, although some species are pathogenic for mammals (e.g. B. anthracis) and insects (e.g. B. sphaericus, B. thuringiensis). An important common characteristic is their ability to form endospores that allow them to survive for extended periods under adverse environmental conditions. The ability to sporulate and their metabolic diversity are significant factors that have led to their successful colonisation of a wide variety of environments.
Six taxonomic groups, named after their most prominent member species, have been described for Bacillus spp. [2]. Members of the B. polymyxa group, recently renamed Paenibacillus, are auxotrophs, most commonly associated with rotting plant materials, composts and the rhizosphere. Some members of this group are able to fix nitrogen [3] and may contribute significantly to the uptake of nitrogen by crops such as Canadian wheat. There is also evidence that members of this group produce plant hormones such as gibberellin [4]. Members of the B. brevis group, renamed Brevibacillus, are found in both soil and water habitats. The B. sphaericus group are most noted as insect pathogens and are found in the sediments of pools, lakes and drainage ditches where insect larvae thrive. Two groups of thermophiles, which include members of the genus Alicyclobacillus and B. stearothermophilus, are found predominantly in soils from a variety of thermal and non-thermal sites.
The B. subtilis group includes the most intensively studied of the bacilli, including species such as B. subtilis itself, B. licheniformis and B. amyloliquefaciens which are of industrial importance [5]. However, these intensive in vitro laboratory investigations have rarely been extended to in vivo environmental studies. Consequently, our knowledge of the ecology of B. subtilis is far from complete. For example, it is not even clear at which sites in the environment active growth occurs. Even though B. subtilis is isolated from a wide variety of habitats, in many cases this is likely to be due to the presence of spores, rather than a reflection of active growth.
Evidence is emerging that the vegetative form of B. subtilis is prevalent in nutrient rich environments such as the rhizosphere. There are numerous reports of the isolation of B. subtilis from the rhizosphere of a range of plant species [6–9] at concentrations as high as 107 per gram of rhizosphere soil [8]. Bacillus species have been investigated for their plant growth enhancing effects since the early 1960s [10] and a relationship between B. subtilis and plants has now been established. In 1989 B. subtilis A13 was categorised as a plant growth-promoting rhizobacterium (PGPR) by Kloepper and co-workers [11] who described it as being moderately competitive in the rhizosphere. Strains of B. subtilis have been shown to synthesise antifungal peptides [12] and plant growth promoting substances, including gibberellin and indole-acetic acid [4,13]. This has led to the development of this bacterium as a biocontrol agent. Seed inoculants containing B. subtilis have been shown to increase crop yields, although it has not been established whether this is due to its plant growth promoting or disease protecting activities. A commercial seed treatment (KODIAK, Gustafson, Texas), which uses spores of B. subtilis, represents the first description of a PGPR being used commercially and seed treatments containing B. subtilis are currently being applied to a wide range of crops over 4 million hectares.
In addition to their use for agricultural crop enhancement, Bacillus species are potential hosts for the delivery of genetically engineered products into the soil and the rhizosphere. However, before this potential is realisable, an improved understanding of the behaviour of Bacillus species in these environments is required, both in terms of their interactions with plants and with the indigenous microbial population. This information will also be required to assess the likely risks and benefits of releasing genetically modified bacilli into the environment.
At the most fundamental level, the mechanisms responsible for the success of B. subtilis in the soil and rhizosphere are encoded by its genome. Analyses of its genome and proteome [14], its patterns of gene expression in vivo and the function of novel genes, are likely to reveal information on the strategies that enable B. subtilis and related bacteria to colonise these environments.
3 The Bacillus subtilis genome
The genome sequence of B. subtilis 168 was completed in June 1997. B. subtilis 168 is a tryptophan auxotrophic mutant [15,16] derived from the original B. subtilis ATCC6051 which was isolated from boiled hay infusion by Cohn in 1875. The genome is 4.2 Mbp in size and comprises 4100 protein-coding genes.
The B. subtilis genome has evolved with the specific function of enabling the growth and survival of this micro-organism within its natural environment. A significant proportion of the genome has already been shown to have functions specifically related to this purpose and it is anticipated that this will be extended by the inclusion of genes currently of unknown function. Although 168 is a laboratory strain, examination of the B. subtilis genome sequence from the perspective of a soil micro-organism is a first step in gaining new insights into processes underlying the behaviour of this taxon in its natural habitat.
3.1 Survival and stress responses
Under adverse conditions, B. subtilis initiates a series of transitional responses that are designed to maintain or restore growth. These include the induction of macromolecular hydrolases (e.g. proteases and polysaccharidases), chemotaxis and motility, and competence (i.e. the ability to take up DNA from the environment). If these responses fail to re-establish growth, sporulation is induced. The ability of Bacillus species to form highly resistant endospores imparts an enormous competitive advantage in environments such as soil, where long periods of drought and nutrient deprivation are common. At least 4% of the B. subtilis genome is dedicated to the processes of sporulation, germination and outgrowth.
The B. subtilis genome encodes more than forty temperature-shock and general stress proteins which presumably contribute to this organism's ability to survive shorter periods of adversity. One of the most commonly encountered is likely to be osmotic stress resulting from frequent wetting and drying of its habitat. B. subtilis protects itself from osmotic stress by the accumulation of osmoprotectants from exogenous sources via three distinct operons encoding high affinity transport systems, opuA, opuB and opuC[17]. In addition, the osmoprotectant glycine betaine can be synthesised from choline or glycine via the products of the gbsA and gbsB genes [18].
Bacteria in a temperate soil must be able to adapt rapidly to changes in temperature. B. subtilis encodes numerous ‘heat shock inducible’ genes, including chaperones and proteases [19]. Proteins induced in response to lower temperatures include cold acclimatisation proteins, for those synthesised continuously at low temperature, and cold shock proteins (Csp), for those synthesised in response to a sudden temperature downshift. The B. subtilis CspABCD proteins are essential for cold acclimatisation under laboratory conditions [20], CspC being identified directly as a result of the genome sequencing project. These proteins bind to RNA in a co-operative manner and may function as RNA chaperones to facilitate the initiation of translation under low temperatures [20]. Other proteins such as trigger factor (tig), a peptidyl prolyl isomerase, may also play a role in protein folding at lower temperatures [21]. Interestingly, Pseudomonas fluorescens homologues of the glycine betaine uptake pathway and of trigger factor have been shown to be induced during the colonisation of the sugar beet rhizosphere [22].
3.2 Substrate utilisation
B. subtilis can utilise a wide range of substrates and analysis of the genome reveals 77 putative ABC transporters, 18 amino acid permeases and at least 16 PTS sugar transporters. It is also able to utilise sugars available in the rhizosphere in the form of complex polysaccharides. Indeed, the ability of Bacillus spp. to secrete a wide variety of extracellular macromolecular depolymerases is a major factor contributing to their colonisation of soil. Genes encoding secreted amylases, arabinases, chitonases, mannanases, cellulases and xylanases are evident in the genome sequence. Proteases are also frequently encountered, both intracellular and extracellular, the latter allowing proteins to provide sources of both carbon and nitrogen.
Carbohydrate substrates are taken up into the cell by a variety of transport systems, including PTS and ABC transporter systems. Recent evidence suggests that the regulation of some uptake pathways, notably the arabinose and xylose pathways, may be controlled by a quorum sensing mechanism involving an oligopeptide quorum factor, in which induction is dependent of high cell densities [23]. Seven phr genes, encoding regulatory peptides, have been identified in the genome downstream of aspartate phosphatases that dephosphorylate transcriptional regulators and at least two oligopeptide transport systems are evident in the genome sequence.
Ammonium is assimilated via glutamine synthetase, with glutamine serving as a universal donor of amino and amide groups. Many free amino acids are able to provide a source of nitrogen. Genes for urease production (ureABC) are induced in the presence of urea or when ammonium levels are low [24]. The nar operon encodes a respiratory nitrate reductase which enables B. subtilis to grow under anaerobic conditions using nitrate as an electron acceptor in place of oxygen [25].
Opines are produced by plants in response to signalling factors from rhizobacteria such as Rhizobium and Agrobacterium. Interestingly, a number of genes on the B. subtilis genome point to the existence of pathways for the utilisation of opines as carbon and nitrogen sources. The product of the gene ytmM shows similarity to OccM, an octopine transport permease of Agrobacterium and to a nopaline transport permease, NocQ, as well as to permeases from other amino-acid transporters. The yurP gene product shows similarity over most of its length to MocD, required for the catabolism of mannityl opines in A. tumefaciens[26]. The yurL gene is located in the same transcript as yurP and its product shows homology to MocE, a kinase required for opine catabolism. YurR shows homology to the agaE gene product of A. tumefaciens, required for the utilisation of agropinic and mannopinic acid (accession number AF035413). The yrbE gene product shows homology to the rhizopine catabolism protein MocA, but also to a streptomycin biosynthesis gene.
The availability of inorganic phosphate in the environment is one of the major growth limiting factors for plants and bacteria [27,28]. B. subtilis has evolved complex regulatory systems for conserving phosphate, and for recovering and utilising phosphate from organic sources. The secretion of multiple alkaline phosphatases may contribute in a mutualistic way to its interaction with plants in the rhizosphere by providing sources of inorganic phosphate [28].
3.3 Chemotaxis and motility
Motility and chemotaxis involves a highly co-ordinated signalling pathway. The structural and regulatory genes which control the pathway for chemotaxis and motility are referred to as the flagellar-chemotaxis-motility genes (fla-che-mot). Sigma factor D is critical for controlling the timing and induction of the chemotaxis and motility pathway. The process of chemotaxis and motility has been studied under laboratory conditions where it has been shown that all 20 common amino acids act as attractants [29]. The importance of chemotaxis and motility in vivo in the environment, particularly in relation to plant/Bacillus interactions, remains to be determined.
3.4 Surface adhesion
The ability of micro-organisms to adhere to plants roots and to soil particles is an important factor in determining their ability to colonise the soil and rhizosphere, although to date there have been no studies on the molecular basis for this in B. subtilis. A variety of compounds have been implicated in binding in other micro-organisms, including surface proteins such as ricadhesins of Rhizobium sp. [30] and Agrobacterium sp. [31], and related proteins in Pseudomonas putida[32]. Polysaccharides have also been implicated in promoting the adhesion to plant cells of Rhizobium, Agrobacterium and Azospirillum. In Azospirillum, attachment to wheat roots is a two-stage process initially involving a surface protein component but requiring a polysaccharide component for irreversible adhesion. Analysis of the B. subtilis genome has not revealed proteins showing sequence similarities to ricadhesins, although the products of two genes of unknown function, ycdH and yfiQ, show homology to proteins involved in attachment to solid surfaces in other micro-organisms. The product of ycdH is a putative lipoprotein with homology to proteins involved in adhesion, including FimA of Escherichia coli. However, it also shows similarity to a zinc binding permease from Streptococcus pneumoniae[33]. YfiQ shows similarity to IcaC of Staphylococcus epidermidis, thought to be required for the secretion of a polysaccharide biofilm adhesin [34].
In addition, examination of the B. subtilis 168 genome sequence reveals three open reading frames, ywtB, ywsC and ywtA which show a high level of similarity (49%, 67% and 78% respectively) to the genes, capA, capB and capC required for γ-polyglutamate (γ-PGA) synthesis in B. anthracis. γ-PGA is an extracellular protein polymer produced by a number of bacilli including B. subtilis (natto) [35]. However, to our knowledge, there have been no reports of the function of this polymer or of the production of γ-PGA in B. subtilis 168. Furthermore, there are at least 12 putative genes on the genome which could be responsible for the production of polysaccharide polymers, many showing similarity to the cap family of genes involved in polysaccharide synthesis in Staphylococcus aureus. The function and characteristics of these polysaccharides also remains to be investigated.
3.5 Specific interactions with plants and the indigenous bacterial population
About 4% of the genome of B. subtilis is devoted to the production of antibiotics and antibiotic-like substances. These include three regions coding for multi-functional polyketide synthetases for the production of peptides such as surfactin and the antifungal antibiotic, fengycin [36].
Bacillus species are thought to enhance plant growth via the synthesis of plant growth hormones. Strains of B. subtilis are reported to synthesise giberellic acid [4] although the genetic basis for the synthesis of any plant growth regulating compound has not been reported. Evidence in the genome sequence for their production by B. subtilis strain 168 is inconclusive. A single gene, ysnE, shows homology to acetyl-transferases and is homologous to a gene required for the synthesis of indoleacetic acid by Azospirillum sp.
3.6 Signalling pathways
The relationship between Bacillus species and plant roots was previously thought to be non-specific, however, there is increasing evidence of signalling between Bacillus, plant roots and the indigenous population. Bacillus spp. are able to alter the dynamics of the legume-Rhizobium symbiosis implicating them in direct interactions with the plant and the Rhizobium at a molecular level [37]. It is possible that B. subtilis produces specific signalling molecules to interact with a plant roots in a similar fashion to Rhizobium and Agrobacterium. Rhizobium species are able to secrete lipochito-oligosaccharides (LCOs) called Nod factors which promote nodulation by inducing root hair deformation [38]. Nod factors consists of a tetramers or pentamers of N-acetylglucosamine with an acyl chain attached to the non-reducing sugar. Analysis of the B. subtilis genome sequence reveals at least two open reading frames (ORFs) which show similarities to the genes of Rhizobium involved in the synthesis of chito-oligosaccharides. The products of yjeA and ybaN show significant similarity to NodB, a chito-oligosaccharide deacetylase from Rhizobium meliloti, although YjeA shows a stronger similarity to XylD, an endo-1,4-β-xylanase D of Cellulomonas fimi. Polysaccharides have also been shown to be involved in inducing plant tissue differentiation and to provide irreversible anchoring in the adhesion to roots of Gram-negative rhizobacteria such as Azospirillum, Pseudomonads and Rhizobium. It may therefore be relevant that the B. subtilis genome includes genes that show similarity to genes in other organisms involved in capsular and exopolysaccharide synthesis, as indicated above.
3.7 Analysis of open reading frames
Functions can be ascribed to ORFs experimentally or by comparison with previously identified proteins. In the latter case, bio-informatics tools such as BLAST [39] are used to interrogate existing sequence databases for homology. Protein sequences showing identities of 25% or more are likely to have evolved from a common ancestor and are therefore likely to perform similar functions. Identities below this level may still provide valuable pointers to function, but need to be interpreted with care. In some cases high levels of identity can define a general function (e.g. ABC transporter, two-component regulator), without necessarily identifying the specific role of a protein. Finally, it should be remembered that three-dimensional structure as well as the primary sequence is important for the functional activity of a protein, and consequently tools for predicting and comparing proteins structures are likely to play an increasingly important role in the identification of ORFs.
Although analysis of a complete microbial genome sequence reveals many new ORFs, it is not possible to ascribe functions to all of the detected ORFs on the basis of currently available sequence databases and bio-informatics tools. ORFs without identifiable functions have been referred to as URFs (unidentified reading frames) [40]. Even in well characterised species such as E. coli and B. subtilis, URFs represent about 40% of the detected ORFs, while in the less characterised Methanococcus jannaschii they represent more than 60% of ORFs. Even in the minimalist genome of Mycoplasma genitalium (470 ORFs) URFs account for more than half of the ORFs. Given the intensive in vitro studies on bacteria, it is likely that the function of a significant proportion of the URFs will relate to growth and survival in natural habitats.
The inability to ascribe functions to more than one third of the genes in a genome has required the development of new strategies for analysing micro-organisms. In the cases of Saccharomyces cerevisiae and B. subtilis, highly co-ordinated systematic functional analysis programmes have been established, involving groups with a wide range of technical expertise. In the latter case approximately 1000 unknown genes are being systematically disrupted with an integrational vector to generate a collection of isogenic mutants strains [41]. The mutant construction strategy includes the introduction of a lacZ transcriptional reporter for monitoring the expression of the target gene and a controllable promoter that facilitates the expression of any genes downstream in the same operon (Fig. 1).
a: The integrative vector pMutin2. b: Integrational inactivation of B. subtilis genes using pMutin2.
4 Functional analysis of microbial genes in natural environments
The availability of a large set of defined reporter gene mutants in B. subtilis provides an unparalleled resource for studying in vivo gene function in a soil bacterium. However, the use of lacZ as a reporter in natural soil microbial communities is limited by its widespread distributions in such environments. The ease with which gene replacement may be carried out in B. subtilis, makes it feasible to replace the lacZ gene in these mutants with a marker that is more suitable for in vivo studies (e.g. gfp, encoding the jellyfish-derived green fluorescent protein).
4.1 Molecular methods for studying gene expression in environment
Genes required for successful colonisation of the soil and rhizosphere are often co-ordinately regulated by specific environmental factors. A knowledge of the pattern of gene expression in the natural environment should be a high research priority since it is likely to lead to an improved understanding of the molecular basis of rhizosphere colonisation and will identify genes that should become the focus of intensive functional analysis studies.
Recently, a number of novel methods have been developed for the identification of genes expressed in vivo. Although originally developed to study pathogenesis, they should be capable of being adapted for studies on the colonisation of ecosystems. The in vivo expression of genes in an individual cell may be monitored by fluorescence and confocal microscopy or flow cytometry. For example, luxAB gene fusions have been used to monitor the expression of genes induced by root derived chemicals [42] and to examine phosphate limitation in the barley rhizosphere [43]. The green fluorescent protein has proved valuable for studying the expression of nitrogenase genes in Azoarcus sp. in the rhizosphere of rice [44].
Procedures based on promoter trapping techniques (e.g. in vivo expression technology or IVET) [45] have been developed to identify promoters that are active in vivo through their ability to drive the expression of genes that are essential for survival [22,46]. Although a powerful technique, IVET requires the analysis of large numbers of fusion constructs in order to identify the relevant promoter fragments.
Another promising method for identifying genes that are important in vivo is signature-tagged mutagenesis [47]. The technique involves the unique tagging and subsequent identification of individual mutant types. A defined population, comprising equal numbers of tagged mutants, is introduced into the test environment. After a suitable period the population is recovered and the survival and/or growth of individual mutants assessed by quantifying their signature tag, using a combination of PCR and DNA hybridisation. A reduction in the proportion of a particular mutant is indicative of its failure to colonise the environment as well as the wild-type. It has recently been applied to the functional analysis of yeast genes [48]. Tagged mutant strains were analysed under selective growth conditions and the level at which each strain survived was determined by hybridisation of the tags to high-density oligonucleotide arrays. The relative abundance of the tags at different times during the selective growth processes allowed the fitness of each mutant in the pool to be quantified.
5 Prospects for the future: How will knowledge of genome sequences improve matters?
The accumulation of genome sequences for soil-living micro-organisms not only holds out the prospect of a detailed understanding of their molecular biology, but will also facilitate the extension of existing techniques and the development of novel techniques for the study of molecular ecology. For example, it can be anticipated that, with the accumulation of data on gene-function relationships, the application of bio-informatics techniques will reveal much about the potential in vivo activity of a micro-organism. The utility of techniques such as IVET and signature tagged mutagenesis will be improved by the systematic generation of directed gene fusions or mutants and by rapid identification of in vivo active genes. Function analysis programmes, such as that being developed for B. subtilis, will provide in vitro data that are likely to reveal clues to in vivo activity.
The availability of genome sequences means that emerging technologies such as high-density oligonucleotide arrays, which will permit the rapid and simultaneous quantification of the expression of every gene on the genome [49], can be applied to the expression of genes in natural environments. It will also allow the construction of organisms with reduced or minimal genomes for in vivo studies [50].
Ultimately, genome sequence data will be combined with molecular biological, biochemical and physiological data to generate mechanistic and predictive models that will allow the behaviour of micro-organisms in a range of environments to be investigated in silico. We envisage such studies will be an aid to, rather than a replacement for, genuine ecological investigations.
References
