Learn more: PMC Disclaimer | PMC Copyright Notice
Methylotrophic Metabolism Is Advantageous for Methylobacterium extorquens during Colonization of Medicago truncatula under Competitive Conditions
Abstract
Facultative methylotrophic bacteria of the genus Methylobacterium are commonly found in association with plants. Inoculation experiments were performed to study the importance of methylotrophic metabolism for colonization of the model legume Medicago truncatula. Competition experiments with Methylobacterium extorquens wild-type strain AM1 and methylotrophy mutants revealed that the ability to use methanol as a carbon and energy source provides a selective advantage during colonization of M. truncatula. Differences in the fitness of mutants defective in different stages of methylotrophic metabolism were found; whereas approximately 25% of the mutant incapable of oxidizing methanol to formaldehyde (deficient in methanol dehydrogenase) was recovered, 10% or less of the mutants incapable of oxidizing formaldehyde to CO2 (defective in biosynthesis of the cofactor tetrahydromethanopterin) was recovered. Interestingly, impaired fitness of the mutant strains compared with the wild type was found on leaves and roots. Single-inoculation experiments showed, however, that mutants with defects in methylotrophy were capable of plant colonization at the wild-type level, indicating that methanol is not the only carbon source that is accessible to Methylobacterium while it is associated with plants. Fluorescence microscopy with a green fluorescent protein-labeled derivative of M. extorquens AM1 revealed that the majority of the bacterial cells on leaves were on the surface and that the cells were most abundant on the lower, abaxial side. However, bacterial cells were also found in the intercellular spaces inside the leaves, especially in the epidermal cell layer and immediately underneath this layer.
Bacteria of the genus Methylobacterium are facultative methylotrophs that are capable of growth on methanol and methylamine, as well as C2, C3, and C4 compounds (28). They belong to the Alphaproteobacteria and are sometimes referred to as pink-pigmented facultative methylotrophs. Methylobacterium strains are widespread in the environment and have been isolated from soils, dust, and lake sediments. They have also been found in association with plants, specifically with leaf surfaces, and they have been hypothesized to potentially dominate the phyllosphere bacterial population (6, 15).
Some Methylobacterium strains possess nitrogen-fixing and nodulation capabilities, which they use in symbioses with Crotalaria and Lotononis plant species (20, 21, 53). Recently published data suggest that the degrees of plant-Methylobacterium association vary from very strong, as exemplified by symbioses, to semitight, as exemplified by endophytic association (7, 24, 41, 56), to loose, as exemplified by epiphytic association on plant surfaces (6, 15, 40). In the case of symbiosis, the benefit for the plant is evident, in contrast to the looser forms of association between methylotrophs and plants. However, beneficial effects have been suggested for the latter associations due to the production of plant hormones, such as cytokinins and indole acetic acid by methylotrophs (17, 19, 23, 42, 54). In addition, the beneficial effects proposed for Methylobacterium involve production of vitamin B12, which has been suggested to stimulate plant development (1, 17, 18). Also, Methylobacterium strains may be associated with plant nitrogen metabolism by the means of bacterial urease (16).
It has been suggested that the consistent success of Methylobacterium strains in colonization of the phyllosphere is due to their ability to utilize methanol as a carbon and energy source (6). The release of methanol by plant leaves is well documented (10, 30). It has been speculated that methanol is produced mainly as a by-product of pectin metabolism during cell wall synthesis (39). The precursors of pectin contain numerous galacturonate methyl esters, presumably to facilitate transport through the cell wall. These methyl esters are demethylated by pectin methylesterases (34), resulting in methanol production. There is experimental evidence that most of the methanol is produced inside leaves and is emitted primarily through stomata (30, 37).
Metabolic resources are known to be key determinants of microbial colonization of plants, and microbes compete for utilization of organic compounds released by plants (29, 48). Therefore, it appears likely that methylotrophic bacteria profit from their ability to utilize methanol and that methylotrophy provides a selective advantage upon phyllosphere colonization. In this study, we aimed at obtaining direct experimental evidence for this hypothesis. To do this, we used Methylobacterium extorquens AM1, which is the most well-studied representative of the pink-pigmented facultative methylotrophs. Core enzymes involved in methylotrophy in this organism are well known (Fig. (Fig.1),1), and they have been purified and characterized in detail (59). In addition, a whole set of mutants is available and has been described over the past few years (3). Upon growth on methanol as the sole source of carbon and energy, methanol is first oxidized to formaldehyde by a periplasmic methanol dehydrogenase that is composed of two subunits, MxaF and MxaI (13). The formaldehyde is then utilized in the cytoplasm; it is oxidized to CO2 for energy generation in the pathway dependent on the cofactor tetrahydromethanopterin, or it is assimilated into cell biomass via the serine cycle.
One-carbonmetabolism of M. extorquens. Formaldehyde is produced in the periplasm of the cell from methanol and is transferred into the cytoplasm. Part of the formaldehyde is oxidized to CO2, and part is assimilated via the serine cycle. mxaF,I, genes encoding large and small subunits of methanol dehydrogenase (13); fae, gene for tetrahydromethanopterin (H4MPT)-dependent formaldehyde-activating enzyme (58); mtdA, gene for NADP-dependent methylene tetrahydromethanopterin/tetrahydrofolate dehydrogenase (8, 57); mtdB, gene for NAD(P)-dependent methylene tetrahydromethanopterin dehydrogenase (14); mch, gene for methenyl-tetrahydromethanopterin cyclohydrolase (43); fhc, formyltransferase/hydrolase complex (44, 45); fdh, gene for formate dehydrogenase (4, 25).
Mutants with lesions in primary methanol oxidation to formaldehyde or in formaldehyde oxidation via the tetrahydromethanopterin-linked pathway (Fig. (Fig.1)1) are unable to grow in the presence of methanol as a sole carbon and energy source. However, the two types of mutants differ from each other with respect to sensitivity to formaldehyde or methanol. Whereas mutants defective in the tetrahydromethanopterin-dependent pathway are poisoned by formaldehyde-generating C1 compounds, mutants defective in primary methanol oxidation, such as a ΔmxaF mutant, are not poisoned by these compounds (31). The latter mutants are capable of growth on multicarbon compounds in the presence of methanol, as the toxic intermediate of methylotrophy, formaldehyde, is not generated in these mutants (Fig. (Fig.1).1). A ΔmxaF mutant was thus an ideal choice to test the importance of methanol conversion in Methylobacterium upon plant colonization. In addition, we included a mutant with a defect in tetrahydromethanopterin biosynthesis (i.e., ΔmptG) (46, 49) that has pronounced sensitivity to formaldehyde or methanol (the MIC of the latter compound is 1 μM) (31). Medicago truncatula, a well-studied legume that can establish a symbiotic association with the soil bacterium Sinorhizobium meliloti, was used as the plant model in this study. Sterilization of M. truncatula seeds can be achieved readily, so a sterile system was used to study the colonization of this model plant by M. extorquens. Bacterial inoculation was performed at the seed level throughout this study, and seedlings and plants were kept under sterile conditions. We also included a microscopic analysis with a gfp-expressing strain to localize the bacteria on the plants.
MATERIALS AND METHODS
Bacterial strains and growth conditions.
Derivatives of M. extorquens AM1 used in this study are listed in Table Table11 and were kindly provided by C. J. Marx and M. E. Lidstrom, University of Washington, Seattle. To generate gfp-expressing strains with lesions in mxaF and mptG, strain CM174.1 (32) was used as a recipient, and mutagenesis was performed with the allelic exchange vectors pCM194.1 and pCM253.1 as described previously (31).
TABLE 1.
Strains used in this study
| Strain | Description | Reference |
|---|---|---|
| M. extorquens AM1 | Rifr derivative | 38 |
| CM194.1 | ΔmxaF, mutation in the large subunit of methanol dehydrogenasea | 31 |
| CM253.1 | ΔmptG (Δorf4), mutation in the first enzyme of tetrahydromethanopterin biosynthesis, β-ribofuranosylaminobenzene 5′-phosphate synthaseb | 31 |
| CM174.1 | katA::(loxP-trrnB-PmxaF-gfp-tT7) | 32 |
| CM194.1_gfpK | katA::(loxP-trrnB-PmxaF-gfp-tT7), ΔmxaF::kan | This study |
| CM253.1_gfpK | katA::(loxP-trrnB-PmxaF-gfp-tT7), ΔmptG (Δorf4)::kan | This study |
The strains were grown at 30°C on a minimal medium (MM) (38) containing 18.5 mM sodium succinate. Antibiotics were added to the following final concentrations: 50 μg of rifamycin/ml, 50 μg of kanamycin/ml, and 10 μg of tetracycline/ml. For seed inoculation experiments, bacterial suspensions were prepared from fresh cultures grown for 24 h.
Plant growth conditions.
Seeds of M. truncatula Jemalong (line A17) were surface sterilized in H2SO4 for 10 min. After five rinses with sterile distilled water, seeds were placed on water agar plates (0.5%) and kept in the dark at 20°C for 48 h for germination. Subsequently, the seedlings were grown on Fahraeus medium (9) supplemented with 0.33 g (NH4)2SO4 per liter in square petri dishes (102 by 120 by 17 mm; Greiner Bio-one, France) in a growth chamber at 25°C with a photoperiod consisting of 16 h of light and 8 h of darkness.
Seed inoculation and sampling.
Bacterial cultures were diluted approximately 1:50 with sterile MM containing 500 μM succinate, and the optical density at 600 nm was adjusted to 0.05, which corresponded to about 5 × 106 CFU per ml. The resulting suspensions were used for inoculation experiments with individual strains; for competition experiments, the adjusted suspensions were mixed 1:1 (vol/vol). Sterilized seeds of M. truncatula were incubated together with the bacterial suspensions (8 ml) for 4 h in 15-ml plastic cones with gentle shaking. To determine that the seeds were sterilized successfully, a negative control was included in every experiment. For this control, seeds were incubated in sterile MM containing 500 μM succinate and treated like the inoculated seeds.
To determine the size of the bacterial population after inoculation, seeds were placed individually in 1 ml of MM and sonicated for 4 min in an ultrasonication bath (Transsonic T275; Prolabo, France). Eight seeds were analyzed for each bacterial inoculum. Cell suspensions were then serially diluted and plated onto MM using the drop plate method (5 μl per drop). The plates were incubated at 30°C for 4 days before the colonies were counted. To do this, we used a dilution that contained a minimum of 5 and a maximum of about 70 colonies per drop deposited. Colonies were counted with a Leica MZ FLIII fluorescence stereomicroscope (Leica Microscope Systems AG, Wetzlar, Germany) equipped with a GFP3 fluorescence filter set.
Bacteria were harvested from the seedlings and from young M. truncatula plants at different times. To do this, different parts of plants (roots, cotyledons, and leaves) were collected aseptically separately. The times at which plant parts were collected were 48 h after inoculation (seedling) and 9 and 16 days after inoculation, which corresponded to the first-leaf and first-trifoliate-leaf stages, respectively. At each sampling time, roots, cotyledons, and leaves were collected and sonicated as described above to remove the bacteria. Cell suspensions were subsequently serially diluted and plated (5 μl/drop) onto MM. As described above, eight seedlings or plants were analyzed for each strain or strain mixture. To obtain the total bacterial count for one plant, the sum of the counts for the different plant parts was determined.
In vitro competition experiments.
Mutant strains CM194.1 and 253.1, as well as the wild-type strain and green fluorescent protein (GFP)-labeled strain CM174.1, were grown as described above, and the optical density at 600 nm was adjusted to 0.1. Mixtures of bacteria that were used in competition experiments were diluted 1:100, and the suspensions were spread on minimal agar plates containing succinate as a multicarbon substrate at concentrations of 18.5 mM and 1 mM. In parallel, the mixtures were spread on the same medium that contained methanol at different concentrations in addition to succinate. After 3 days of incubation, cells were harvested from the agar surfaces, and the numbers of bacterial cells were determined by using dilution series as described above.
Statistical analysis.
All single-inoculation and competition experiments with M. extorquens wild-type strain AM1 and mutants were repeated five times, and each data set consisted of data from eight different plants that were harvested at four times during plant development. For each time and plant part, the data from the five trials were combined and analyzed for consistency by an analysis of variance (no significant interaction, P > 0.05), while significant differences in development between different strains were proven by using the same analysis of variance in combination with post hoc pairwise t tests with the Bonferroni correction using SYSTAT, version 11 (SPSS Inc., Richmond, Calif.). For this analysis, absolute cell numbers were log transformed, while percentages were arcsin transformed.
Microscopic visualization of M. extorquens on M. truncatula plants.
Colonized leaves and roots were observed with a Zeiss Axiophot II fluorescence microscope (Carl Zeiss AG, Oberkochen, Germany) and a Leica TCS SP2 confocal microscope. Confocal optical sections were obtained with a distance of 1 μm between sections. Image projections were made with the Leica confocal software, and these projections were processed with Image-Pro Plus (Media Cybernetics, Silver Spring, Md.). For observations within leaves the plant material was stained with propidium iodide.
RESULTS
Methylotrophy-deficient mutants are less competitive than wild-type M. extorquens AM1 during colonization of M. truncatula.
To test the importance of the methylotrophic metabolism of M. extorquens for plant colonization, mutants that were not able to grow on methanol as a sole carbon and energy source were tested in competition with the wild-type strain. The following strains were included in this study: wild-type M. extorquens AM1, a GFP-labeled strain of M. extorquens (CM174.1), and two mutant strains, a ΔmxaF mutant (CM194.1) and a ΔmptG mutant (CM253.1) (Table (Table11 and Fig. Fig.1).1). Although the last two strains are not able to grow on methanol, they grow at similar rates in the presence of an alternative growth substrate (e.g., succinate) (31).
To distinguish between wild-type and mutant strains, a GFP label was used. GFP expression apparently did not affect the fitness of the strains in our experiments. This was evident from competition experiments performed with wild-type cells and cells of GFP-labeled strain CM174.1, which revealed equal fitness throughout plant development (Fig. (Fig.2).2). Furthermore, the results described below were independent of whether the mutant or the wild type was tagged with GFP. Thus, the GFP-labeled strain CM174.1 was considered to be like the wild type in competition with the ΔmxaF and ΔmptG mutants, and the GFP label was used to distinguish between the strains in competition experiments. Five independent experiments were performed, and they produced consistent results. The experiments revealed that the ΔmxaF strain (CM194.1) was less competitive than the wild type. The effect was clear even at the tegument after seed germination, at which the percentage of the ΔmxaF (CM194.1) cells was significantly (P < 0.05) lower than the percentage of the wild-type cells. After 1 week of plant growth, the initial percentage (50%) had changed to 30% for the aerial part of the plants and to 20% for the roots. This tendency persisted over the 2-week experiment, and the mutant cells finally accounted for about 20% of all cells on all plant parts tested (Fig. (Fig.2).2). The ΔmptG mutant (CM253.1) exhibited a more severe decrease in fitness than the CM174.1 strain and was affected more than the ΔmxaF strain (CM194.1) in competition experiments. After seed germination, the fitness of CM253.1 was significantly (P < 0.05) impaired on the tegument, as well as on the emerging roots, and this became more pronounced after 1 and 2 weeks of plant growth, respectively, when this strain accounted for a little over 10% of the cells for the leaves and for less than 10% of the cells for the roots (Fig. (Fig.22).
Percentages of M. extorquens wild type (open bars), ΔmxaF mutant strain CM194.1 (light gray bars), and ΔmptG mutant strain CM253.1 (dark gray bars) in competition with GFP-labeled strain CM174.1 during colonization of M. truncatula plants. The bars indicate means, and the error bars indicate one standard error of the mean for five independent trials. For absolute numbers of bacteria see Fig. Fig.33.
M. extorquens is able to use carbon substrates other than methanol during colonization of M. truncatula.
In parallel with the competition experiments described above, plant seeds were inoculated with the different M. extorquens strains individually. We did not observe significant differences for any of the strains for the total number of bacterial cells obtained from the plants at each time (Fig. (Fig.3).3). The ΔmxaF mutant CM194.1 and the ΔmptG mutant CM253.1 were both able to colonize sterile M. truncatula plants at the wild-type level in the absence of bacterial competition. These results indicated that methanol is not the only carbon and energy source that Methylobacterium can benefit from. These results were especially surprising for the ΔmptG mutant, which is known to be inhibited in vitro by trace amounts of methanol due to its nonoperational tetrahydromethanopterin-dependent formaldehyde oxidation pathway (31).
Totalnumbers of cells for M. extorquens wild type and mutants upon single inoculation and in competition during colonization of M. truncatula plants. The bars indicate means for five independent trials, and the error bars indicate one standard error of the mean for the independent trials. (B, C, and D) Bacterial counts on leaves, teguments and cotyledons, and roots, respectively. (A) Total bacterial counts for the different plant parts. For single-inoculation experiments, strains are indicated as follows: open bars, M. extorquens wild type; cross-hatched bars, GFP-labeled strain CM174.1; light gray bars, ΔmxaF strain CM194.1; and dark gray bars, ΔmptG strain CM253.1. For competition experiments, the total bacterial counts for the wild-type, CM194.1, and CM253.1 strains and GFP-labeled strain CM174.1 are indicated by the lightest hatched bars, intermediate gray hatched bars, and darkest hatched bars, respectively.
The total number of bacteria harvested in the competition experiments was not significantly higher than the number recovered after individual inoculation, indicating that there was a maximal number of M. extorquens cells that developed in the times observed (Fig. (Fig.33).
For entire plants, the bacterial concentration increased from 105 CFU per seedling (mean for all trials and strains) that were attached to the seeds to 2.5 × 106 CFU per seedling within 2 days after germination. The bacterial concentration reached 2 × 107 CFU per plant at 1 week after germination and 9 × 108 CFU per plant at 2 weeks after germination (Fig. (Fig.3A).3A). Bacterial development was analyzed separately on the different plant parts (Fig. 3B to D). First, we distinguished between the teguments and the emerging roots of the seedlings. On the teguments 2.3 × 106 CFU per plant was found (Fig. (Fig.3C),3C), and 105 CFU per young root was found (Fig. (Fig.3D).3D). During the first week after seed germination, the cotyledons developed, as did the first leaf, whereas the root elongated. The numbers of bacterial cells found on the cotyledons after 1 week were the same order of magnitude as the numbers of bacterial cells found on the teguments. The numbers increased only marginally from the first week to the second week of growth of M. truncatula, from 2.8 × 106 CFU per plant to 3.4 × 106 CFU per plant (Fig. (Fig.3C).3C). On day 9, 9.3 × 104 CFU per first leaf was found (Fig. (Fig.3B).3B). One week later, when a trifoliate leaf had developed, the average number of bacteria on the leaves was 4.7 × 105 CFU per plant. The increase in cell numbers was most pronounced on the roots for total bacterial growth. The concentration of bacteria increased from 105 CFU per root after germination to 1.6 × 107 CFU per root after 1 week and then to 8.5 × 108 CFU per root after 2 weeks (Fig. (Fig.3D3D).
Mutant lacking methanol dehydrogenase is less competitive in vitro than the wild type in the presence of methanol when an alternative carbon source is present at a growth-limiting concentration.
As shown by the experiments described above, carbon sources other than methanol are available to Methylobacterium during colonization of M. truncatula. We tried to mimic these conditions in vitro. We compared the fitness of the ΔmxaF methanol dehydrogenase mutant with the fitness of the wild type. If succinate was added at a concentration of 18.5 mM, the presence of an additional 1 mM methanol or 100 mM methanol did not influence the ratio of the ΔmxaF CM194.1 mutant to the GFP-labeled strain CM174.1. However, if succinate was added at a growth-limiting concentration, 1 mM, addition of methanol negatively affected the abundance of the ΔmxaF mutant compared to that of the CM174.1 strain; addition of 1 mM methanol resulted in a 30% decrease in the number of mutant cells, and addition of 100 mM methanol resulted in a 60% decrease in the number of mutant cells (Fig. (Fig.4).4). These experiments indicated that Methylobacterium benefits from the presence of methanol when a multicarbon source is present under growth-limiting conditions. Furthermore, the change in levels observed for the ΔmxaF mutant in competition with the wild type reflects the results observed when the bacteria were harvested from M. truncatula plants quite well. As expected, when ΔmptG mutant CM253.1 was in competition with the GFP-labeled CM174.1 strain, there was a drastic decrease in the level of the mutant in the presence of 1 and 100 mM external methanol, and the mutant accounted for less than 3% of the cells; in contrast, when the wild type was in competition with the GFP-labeled CM174.1 strain, there were no differences.
Ratios of M. extorquens ΔmxaF mutant strain CM194.1 to GFP-labeled strain CM174.1 before (suspension) and after in vitro competition in the absence and presence of methanol. Competition experiments were performed in the presence of 18.5 mM (open bars) and 1 mM (gray bars) succinate. The values are based on mean bacterial counts from two independent dilution series.
M. extorquens colonizes the surface and intercellular spaces of M. truncatula leaves.
Using epifluorescence and confocal laser scanning microscopy, bacterial sites of colonization were visualized for different parts of M. truncatula. The gfp-expressing strain M. extorquens CM174.1 was used for these observations. Bacteria were inoculated at the seed level to ensure full bacterial establishment on the leaves and roots. The majority of the bacteria were found on the outside of leaves, and the bacteria were most abundant on the lower, abaxial side. They predominantly occupied niches such as the base of trichomes and sites along the veins and crevices between adjacent epidermal cells (Fig. 5A to C). These sites are the sites that were found to be predominantly colonized by other bacteria (35, 36). Confocal scanning microcopy revealed that the bacteria also entered the leaves and were present in the intercellular spaces, especially between epidermal cells (Fig. (Fig.5D)5D) but also between mesophyll cells just underneath the epidermal cell layer (Fig. 5G to I). Occasionally, we observed stomata packed with bacteria (Fig. 5E and F), suggesting that these organs might be the sites of bacterial entrance in the apoplast. Endophytic occurrence of Methylobacterium was not observed in a previous study by Omer et al., who used clover plants (40); however, this might have been due to the spray inoculation of bacteria used in this study, followed by 3 days of incubation.
Confocal microscope images of M. truncatula colonized by gfp-labeled M. extorquens. (A to C) Surfaces of the lower part of leaves. In panel C, bacteria are visible at the base of trichomes. (D to I) Interior of the leaves. (D) Epidermal cell layer. (E and F) Epidermal cell layer with stomata. (G and H) Mesophyll. Chloroplasts are visible due to their red autofluorescence. (I) Mesophyll. Cell walls and nuclei are stained red by propidium iodide. Bars = 20 μm.
DISCUSSION
Methylotrophic bacteria are commonly found in association with plants (5, 15). These microorganisms colonize plant surfaces but have also been found as endophytes (7, 24, 41, 56), while some representatives exhibit a true symbiosis with Crotalaria and Lotononis plants (20, 21, 53). The overall prosperity of this association has been attributed to the unique ability of these bacteria to grow at the expense of methanol, a plant cell wall product (6). In this work we found that methanol is indeed consumed during colonization by Methylobacterium. This was apparent from the development of higher numbers of cells of the wild type in competition with the ΔmxaF methanol dehydrogenase mutant having a defect in methanol oxidation to formaldehyde. After colonization of plants for 2 weeks in the competition experiment with the ΔmxaF strain and the wild-type strain of M. extorquens, the mutant accounted for about 20% of the cells, compared with 50% at the beginning of the experiment (Fig. (Fig.2).2). However, this value also suggested that the bacteria can use carbon sources other than methanol during colonization. This was clearly observed after separate inoculation with the ΔmxaF mutant; upon colonization of M. truncatula the numbers of ΔmxaF mutant cells were comparable to the numbers of wild-type cells under these conditions (Fig. (Fig.3).3). We concluded from these experiments that the bacteria can use alternative carbon sources during plant colonization and might in fact cometabolize methanol with an alternative carbon source. We mimicked in vitro the availability of two alternative growth substrates, succinate and methanol. The results showed that methylotrophy provides an advantage only when the C4 carbon source is present at growth-limiting concentrations (Fig. (Fig.4).4). Methanol conversion in the presence of a multicarbon substrate has not been investigated previously in Methylobacterium. However, it is known that enzymes specifically involved in methylotrophy (e.g., methanol dehydrogenase and tetrahydromethanopterin-dependent enzymes) are also synthesized and active in the absence of methanol, but at lower levels than the levels under methanol growth conditions (2, 26). In this respect it is interesting that Corpe and Basile (5) observed that isolates of methylotrophic bacteria from plant sources were facultative methylotrophs rather than obligate methylotrophs. Facultative methylotrophic metabolism might therefore be an important trait for methylotrophs during colonization. The availability of methanol might change drastically with time and with the localization of the bacteria. It has been assumed that methanol is emitted by the stomata (30, 37). If this is true, the bacteria located inside plant leaves should be exposed to a higher methanol concentration, whereas for bacteria on the outside the availability of methanol might be low. Therefore, it would be interesting to study the spatial distribution of the methanol dehydrogenase mutant without competition and in competition with the wild-type strain.
The nature of the alternative carbon source(s) consumed by Methylobacterium during plant colonization is unknown. On leaf surfaces, leakage of small amounts of carbohydrates, amino acids, and organic acids has been detected (11, 55). The availability of these carbon substrates may be responsible for the unequal distribution of the bacteria on the leaf surfaces that we found for Methylobacterium and that has also been described in detail for other bacteria (35, 36). Sugars have been found to be carbon substrates for other epiphytes, such as Erwinia herbicola (27) and Pseudomonas fluorescens (33); however, they are not utilized by M. extorquens (28). Amino acids are also not common as growth substrates. In contrast, organic acids are excellent carbon substrates, and these compounds may be used in the phyllosphere by Methylobacterium; however, this remains to be demonstrated. In nature, the availability of alternative carbon sources for Methylobacterium obviously depends on bacterial competition as well, both within and outside the facultative methylotroph group. Thus, specialization of metabolism must play an important role in the coexistence of epiphytic bacteria (61).
Our experiments indicate that there are growth limitation effects on Methylobacterium upon colonization of M. truncatula. If bacteria are limited by a carbon source in the first place, we would have expected higher cell numbers for the wild type than for methylotrophy mutants due to utilization of methanol by the former. However, we found no significant differences in the total number of Methylobacterium cells after single inoculation and in competition experiments, independent of which strain was used. Bacterial growth might be limited by the availability of nitrogen or iron, for example.
We found that the percentages of the methanol dehydrogenase mutant of M. extorquens in competition with the wild-type cells were similar on the leaves and on the roots (Fig. (Fig.2).2). This suggests that methanol generation also occurs in the roots. To our knowledge, methanol emission by roots has not been measured; however, it could be predicted based on the activity of pectin methylesterases (47, 51, 60).
The association of facultative methylotrophs with the rhizosphere has been investigated less than the association of these organisms with the phyllosphere. On sterile plants, Methylobacterium developed very well on the root surfaces, and the numbers of cells were higher than the numbers of cells in the aerial parts. The fact that the roots develop earlier than the leaves and the fact that Medicago roots develop especially fast should be taken into consideration. Generally, methylotrophs might be less abundant in the rhizosphere, where root exudates nourish a large and diverse population of bacteria (22). There might be exceptions, for instance, for the root-nodule-forming Methylobacterium nodulans strains from certain Crotalaria species that are nonpigmented (53), since pigmentation is a typical trait of phyllosphere bacteria and bacteria that are abundant in the air (52).
Furthermore, we observed that the numbers of cells were much higher on cotyledons than on leaves. This is not surprising since we inoculated plants at the seed stage and since there is direct contact between the seed envelope after germination and the cotyledons, so that colonization of cotyledons should be easier than colonization of the leaves that develop later. Higher concentrations of carbon substrates at the surface of cotyledons that function as storage organs might also contribute to fast bacterial growth.
In this study we conducted fitness tests for colonization by a Methylobacterium mutant (ΔmptG mutant) that is not capable of synthesis of the cofactor tetrahydromethanopterin. As mentioned above, this mutant is unable to grow in the presence of methanol as a sole carbon and energy source, and it is also inhibited by methanol during growth on multicarbon compounds, such as succinate (31). Due to the formation of methanol by the plant, we expected that the ability of this mutant to colonize plants would be severely impaired. In the competition experiment we found that this mutant was indeed affected more than the mutant lacking methanol dehydrogenase (Fig. (Fig.2);2); however, the ΔmptG mutant was capable of colonizing M. truncatula at wild-type levels when it was inoculated alone (Fig. (Fig.3).3). It could be assumed that the methanol concentration emitted by the plant is below the level of toxicity for the mutant lacking tetrahydromethanopterin. If this were true, we would expect to obtain the same level of recovery for this mutant as for the mutant lacking methanol dehydrogenase in competition experiments; however, this was not the case. Alternatively, the finding that there was a more pronounced effect on the mutant lacking tetrahydromethanopterin than on the mutant lacking methanol dehydrogenase could be explained by utilization of other formaldehyde-generating compounds that could serve as growth substrates (e.g., methylamine). To our knowledge, no information concerning the availability of methylamine on plant surfaces exists. We favor the hypothesis that the inhibitory effect of formaldehyde on the mutant lacking tetrahydromethanopterin during plant colonization is less pronounced. The bacteria might induce an extraprotective formaldehyde oxidation enzyme system when they colonize the plants, such as more or less specific alcohol dehydrogenases that are predicted to exist based on the genome sequence of M. extorquens AM1 (http://www.integratedgenomics.com/genomereleases.html#list6). Alternatively, the plant might detoxify the formaldehyde produced in the bacterial periplasm. Activity of formaldehyde dehydrogenases has been measured in several plant species (e.g., soybean) and has been suggested to be involved in formaldehyde detoxification (12, 50).
More experiments are necessary to obtain further insight into the metabolic traits that enable Methylobacterium to colonize plants. The approach using mutants with lesions in specific C1 metabolic pathways, as described here, is a useful tool for verifying hypotheses, such as the consumption of methanol. Many questions remain; for instance, there are questions concerning the alternative carbon sources or the principal nitrogen source. In addition, it will be interesting to learn how the exact localization of the bacteria influences their metabolism and how metabolism is affected under natural competition conditions. Competition experiments and in vivo fluorescence labeling should help workers obtain further insight into the answers to these questions in the future.
Acknowledgments
This work was supported by the Centre National de la Recherche Scientifique and by the Max-Planck-Gesellschaft.