Abstract
The role of microbial immigration in the veinal colonization pattern of Aureobasidium pullulans on the adaxial surface of apple leaves was investigated in two experiments at two periods (early and late seasons) in 2004 by applying green fluorescent protein (GFP)-tagged blastospores to the foliage of orchard trees. Individual leaves were resampled by a semidestructive method immediately after inoculation (t0) and about 1 (t1), 2 (t2), and 3 (t3) weeks later. At t0, there were no significant (P ≤ 0.05) differences in densities (cells/mm2) on veinal (excluding midvein) sites and those on interveinal sites, but at all points thereafter, densities were significantly higher on veins. GFP-tagged A. pullulans cells remained primarily as singletons on interveinal regions (≥90% at all points), while ≥20% of cells over veins at t3 were in colonies of ≥4 cells. The colonies that developed from single cells placed on midveins and other veins were significantly larger than those that developed on interveinal regions of detached field and seedling leaves incubated under controlled conditions. Colonies primarily developed linearly along veins, reaching average colony sizes (72 h) of 24.4 ± 12.7 (mean ± standard deviation) cells. In contrast, colonies on interveinal regions tended to average only 2.9 ± 1.3 cells, with less linearity. To examine the potential role of A. pullulans growth-inhibiting factors associated with interveinal features, single GFP-tagged A. pullulans cells in droplets previously incubated on interveinal sites were placed on midveins and compared to midvein colonies derived from cells in a water-only suspension. No differences in colony size resulted. Our results indicate that immigration limitation and growth-inhibiting factors are not the primary factors responsible for A. pullulans veinal colonization patterns in the field. Rather, indirect evidence suggests that growth-promoting substances occur locally in the veinal areas.
Microbial colonization patterns on the phylloplane have been described for numerous bacterial and fungal species (5, 8, 14, 29). Cells tend to be found in aggregates, clustered near veins, in crevices between epidermal cells, and at the bases of trichomes (5, 8, 14). However, it is still unclear why these populations are aggregated and how such patterns emerge. Possibilities include variation in nutrients (19), availability of water (7, 30), avoidance or tolerance of environmental stresses (15, 29), and differential arrival of microbial immigrants or erosion of cells at leaf microsites (21, 27). Knowledge of how microbial colonization patterns develop would advance understanding of population growth processes in nature, which could also enhance the success of foliar biocontrol efforts.
Most of the studies of phylloplane microbial colonization have been conducted with detached field or growth chamber (phytotron) leaves (14, 19, 29, 30). Obviously, phytotron and field conditions are different. In the field, microbes on the phylloplane are exposed to fluctuations in sunlight, moisture, temperature, and wind (9, 12). The cuticle erodes, which can alter leaf topography, the wettability of the surface, exudation of nutrients, and retention of microbes (5, 18, 24). The microbial species composition also changes over time, and some inhabitants may influence other colonists by producing inhibitory compounds (2, 13, 16). Colonizers may also be forced to compete for sites on the leaf surface (2, 14, 16, 30). In aggregate, these conditions cannot be closely duplicated in a phytotron.
Here we examine the role of microbial immigration in establishing Aureobasidium pullulans colonization patterns both in the field and in the growth chamber. A. pullulans is found to be associated predominantly with veins on the adaxial surface of apple leaves virtually throughout the growing season (17), but the factors responsible for establishing and sustaining the pattern are unknown. Our logical framework for investigating the factors that contribute to this pattern was as follows: if differential immigration is solely responsible for site-specific colonization differences, then this factor should be eliminated by providing immigrant cells in approximately equivalent numbers to all microfeatures. Conversely, if immigration is relatively unimportant compared to community and habitat features in colonization, then population densities should become different among features even when they receive similar numbers of initial immigrants. In the field, immigration can be controlled by release of a marked population by the investigator. Further, the influence of the existing microbial community on the introduced population can be explored by conducting the experiment on surface-disinfested or water-treated leaves and at different times of the year when leaves are relatively poorly or well colonized. In the laboratory, immigration can be controlled by positioning individual “immigrant” cells at predetermined features of interest.
We provide evidence here that differences in colonization of veinal and interveinal sites arise quickly and persist for both colonized and uncolonized (surface-disinfested) leaves when immigration is controlled. We find no evidence that growth-inhibiting factors limit colony development on interveinal regions; conversely, growth-promoting factors seem to be present over veinal areas.
MATERIALS AND METHODS
Preparation of inocula.
GFP-transformed A. pullulans (26) blastospores stored in 15% glycerol at −80°C (1 × 106 cells in 200 μl of glycerol solution) were added to 50 ml potato dextrose broth (PDB) and incubated on a platform shaker at 150 rpm. After 24 h, 1 ml was removed and incubated in yeast-nitrogen (YN) medium and incubated for a further 24 h to carbon starve cells (6). Aliquots of the suspension then were removed, washed by repeated centrifugation at about 16,000 × g in a microcentrifuge, and diluted with sterile water to appropriate concentrations (below).
Field treatments and leaf processing.
Experiments were conducted in 2004 with apple leaves from the terminal shoots (Malus domestica cv. Liberty scions grafted to M26 rootstock planted in 1998) of trees untreated with fungicide at the West Madison Agricultural Research Station in Madison, WI. Leaves on specific trees were chosen by assigning all trees a number and randomly generating numbers with replacement. Healthy, average-sized leaves were tagged on the east side of the chosen trees. Leaves were spaced on nonadjacent branches at least 1 m apart and approximately 1 to 2 m high. Random, untreated leaves adjacent to (on the same branches) and between (on different branches) GFP-treated leaves were periodically sampled and found to be free of GFP-tagged cells.
Two experiments were conducted beginning on 21 June (designated the early season) and 2 September (late season), respectively, of 2004. Each experiment consisted of 44 leaves disinfested with 15% hydrogen peroxide and 44 leaves treated with sterile water (control). Applications were made in the late afternoon by dipping each leaf in liquid (either peroxide or sterile water) for 1 minute with gentle swirling (11). Leaves were allowed to dry. Then, approximately 1 ml of GFP-tagged A. pullulans suspension containing 105 cells was sprayed onto the adaxial surface of each leaf uniformly to runoff with a 710-ml capacity Prosafe Sprayer (Contico International, St. Louis, MO). Before treating each leaf, the spray bottle was agitated. Leaves were allowed to dry before the first sampling (t0). Forty-two of the 44 leaves were processed as follows (allowing two extra leaves in reserve for contingencies that ultimately were not needed). Ten peroxide-treated and 10 water-treated leaves per experiment were imprinted onto potato dextrose agar amended with 0.4 μg/ml chloramphenicol to determine if peroxide treatment was effective in eliminating wild-type A. pullulans and other microbes (11). In the first experiment, ten leaf imprints yielded 0.6 ± 0.8 (mean ± standard deviation [SD]) wild-type A. pullulans colonies per leaf on peroxide-treated leaves compared to 27.5 ± 4.6 wild-type colonies on untreated leaves (in addition to wild-type A. pullulans, many other fungi remained on water-treated leaves only). The second experiment yielded similar results.
Twenty peroxide-treated and 20 water-treated leaves per experiment were sampled semidestructively (31) as follows. Prior to the t0 sampling, the leaf tip (ca. 1.0 to 1.5 cm) was removed from each leaf with scissors and discarded. The next-most-distal segment, of about 1 cm wide, was removed (t0) and placed in a sterile Whirl-Pak (Nasco, Fort Atkinson, WI) bag on ice for transport to the lab. At each subsequent sampling time, a 1-mm segment was removed (and discarded to eliminate the wounded tissue), and the adjoining 1-cm segment was cut from the leaf and processed. For the first experiment, segment 1 was sampled on 21 June (t0), and segments 2 to 4 were sampled on 29 June, 6 July, and 13 July, respectively. For the second experiment, segment 1 was sampled on 2 September, and segments 2 to 4 were sampled on 9, 18, and 23 September, respectively. Segments from 10 peroxide-treated and 10 water-treated leaves were examined by epifluorescence microscopy at each sampling date (discussed in the following section). Segments from the remainder of leaves were imprinted onto potato dextrose agar amended with 0.4 μg/ml chloramphenicol. Plates were viewed after 2 to 3 days, and GFP-tagged A. pullulans colony counts and locations were recorded.
At each sampling point in both experiments, controls on spatial variation were conducted as follows. Three intact peroxide-treated and three water-treated leaves were collected, and four segments from each of these leaves were removed as described above. Each segment was analyzed by epifluorescence microscopy. GFP-tagged A. pullulans cells/mm2 and GFP-tagged A. pullulans colonies of 2 to 3, 4 to 9, or ≥10 cells/mm2 were enumerated to determine whether the cells/mm2 and colonies/mm2 on midveins differed significantly among leaf segments at any (collection) point in time for each leaf examined. The same analyses were also conducted to determine whether cells/mm2 and colonies/mm2 on other veins and interveinal sites significantly differed.
Determination of A. pullulans counts on leaves.
At each of the four sampling dates per experiment, the adaxial surfaces of segments from 10 peroxide-treated and 10 water-treated leaves were viewed by epifluorescence microscopy with a 40× long-working-distance lens objective (area of field = 0.196 mm2) (in experiment 1, two peroxide-treated leaves were lost after early sampling). Three nonadjacent transects per segment were viewed. One transect was at the center of the segment; the other two were spaced equally between the center and the two edges of the segment. GFP-tagged A. pullulans counts with respect to feature (midvein, other veins, interveinal regions) were recorded. A field containing the midvein was designated a midvein field, and all GFP-tagged A. pullulans counts in the field were considered to be occupying a midvein site. Fields containing smaller veins (secondary, tertiary, or smaller veins) were designated veinal, and all cells within that field were considered to be occupying veinal regions. Finally, fields with neither smaller veins nor the midvein visible were considered interveinal regions, and GFP-tagged A. pullulans counts in these fields were considered to represent occupants of interveinal areas. Unoccupied sites were also recorded as either midvein, other vein, or interveinal, so the cells per square millimeter of leaf area containing the feature of interest could be recorded. GFP-tagged A. pullulans cells were designated as blastospores, SCC (swollen cells and/or chlamydospores), or hyphae, and the sizes of colonies, if present, were also recorded. On leaves used to determine the validity of the semidestructive sampling method, four segments per leaf (spaced equally apart) and two transects per segment were assessed.
Colony formation on seedling leaf microfeatures under controlled conditions.
GFP-tagged A. pullulans was prepared as previously described and diluted to concentrations of about two to three cells per μl. The third noncotyledonary leaves from phytotron-grown apple seedlings (female parent, cv. McIntosh) of similar ages, heights, and conditions were detached. Individual leaves (five per trial) were placed on a filter paper (abaxial surface on the paper) moistened with sterile water in the center of an 8.9- by 6.4- by 0.24-cm silicon rectangular frame affixed to a 12.7- by 10.2-cm glass microscope slide (whole-leaf chambers). The midvein site, a secondary vein (other vein) site, and an interveinal site were inoculated with 2 μl of the GFP-tagged A. pullulans suspension at the upper, middle, and lower thirds of each leaf (nine total sites/leaf). The leaves were allowed to dry, and a coverslip was immediately placed over the silicon frame such that it did not come in contact with the leaf. Single carbon-starved GFP-tagged blastospores (as above) were located at each site by epifluorescence microscopy, and the stage coordinates were recorded. Slides were stored in a humidity box placed in a phytotron (12-h-light/12-h-dark conditions maintained at a 22°C day temperature and a 24°C night temperature). Colony size was evaluated at 24, 48, and 72 h. Three trials were conducted. On additional seedling leaves, colony development from single GFP-tagged blastospores on midveins and interveinal regions was observed hourly for about 48 h.
Colony formation on orchard leaf microfeatures under controlled conditions.
The same experimental procedure was repeated with detached leaves taken from randomly chosen trees at the West Madison Orchard collected on 6 June, 30 June, and 30 September. At each date, five leaves were placed in individual sterile Whirl-Pak bags on ice for transport to the laboratory as described previously. In the laboratory, three midvein, three other-vein (not done on 6 June), and three interveinal regions on each detached leaf were inoculated with single cells of carbon-starved GFP-tagged A. pullulans (by use of the experimental setup and procedures previously described), and colony size was evaluated at 24, 48, and 72 h, with the exception of leaves sampled on 6 June (colony size evaluated only at 72 h). When not being viewed, leaves were incubated in the conditions previously described.
Presence or absence of a growth inhibitor on seedling leaf microfeatures.
The third noncotyledonary leaves from phytotron-grown seedlings of similar ages, heights, and conditions were detached. Five 2-μl droplets of sterile water were placed at intervals of about 0.75 cm on the midvein and interveinal sites, and the leaves were incubated for 5 min in a petri dish. The interveinal drops were then withdrawn with a pipette and pooled in an Eppendorf tube (tube A). The same procedure was followed for drops on the midvein (tube B). The process was then repeated on another leaf, except an incubation period of 30 min was used (tube C was for drops from interveinal locations incubated 30 min, and tube D was for drops from midveins incubated 30 min). Sterile water put directly into Eppendorf tubes (control solution, tube E) served as a control. Carbon-starved GFP-tagged A. pullulans blastospores (as above) were then diluted with sterile water to about two cells per microliter. Equal volumes of cell suspension and potential inhibitor or control solution were combined and gently vortexed. Cell suspensions (2 μl) containing droplets from tube A (droplets on interveinal regions for 5 min) were placed on five midvein features (evenly spaced), and suspensions containing droplets from tube B (midveins, 5 min) were placed on five interveinal locations (positions as described above) on leaves different from those used for the initial incubation. Three leaves were inoculated in this manner. The process was repeated with the 30-min suspension and cell mixture (tubes C and D). Two control leaves were inoculated as described above with cell suspensions from tube E (control) on five midvein and five interveinal locations per leaf. Coverslips were placed on all leaf chambers, and the coordinates of single cells were recorded via epifluorescence microscopy. If cells were spaced at a distance where colonies could potentially merge, the site was not used, and new leaves were inoculated. Whole-leaf chambers were randomized and stored in two humidity boxes in the phytotron. Colony size was recorded at 72 h. The experiment was done twice.
Presence or absence of a growth-inhibiting substance on field leaves.
The above-described procedure was repeated with field leaves (terminal) collected on 27 June, 15 July, and 24 September from random Liberty trees (as previously described).
Statistical methods.
GFP-tagged A. pullulans population sizes were log10 (x + 1) transformed to achieve normality prior to statistical analysis. To determine if differences in cells/mm2 on midveins, other veins, and interveinal areas existed, analysis of variance tests were performed with the PROC MIXED procedure in SAS version 8.2 (SAS Institute, Cary, NC). Because multiple segments of the same leaves were repeatedly measured, segments nested within leaves, and sites nested within segments, with an AR(1) structure between segments were used in the analysis of variance. Additionally, Tukey multiple comparison adjustments were made, allowing regions to be compared to each other (midveins to other veins, midveins to interveinal regions, and other veins to interveinal regions). The percentages of individual GFP-tagged A. pullulans cells distributed as single cells and as colonies of 2 or 3, 4 to 9, or ≥10 cells were compared by use of Student's t test (assuming unequal variances) on untransformed data, conducted with Sigma Plot v. 5.0 (SPSS Inc., Richmond, CA). Colonies of the same size category (or single cells) obtained at corresponding sampling points from the individual sites were compared (Student's t test). Student's t tests were also used to compare mean colony sizes on detached leaves after single cells were placed on different leaf features.
RESULTS
Validation of semidestructive sampling method with an applied population.
No significant differences (P ≤ 0.05) in counts of GFP-tagged A. pullulans cells/mm2 on individual leaf features were found among the four segments at any sampling point. Thus, based on these data (also see reference 31), GFP-tagged A. pullulans cells/mm2 on midveins, other veins, and interveinal features of a given sampled leaf segment were taken as representative of the GFP-tagged A. pullulans cells/mm2 on the remainder of the leaf at that point in time. Likewise, there were no differences (P ≤ 0.05) in numbers of GFP-tagged A. pullulans colonies of any size category on individual features among segments.
Comparison of peroxide-treated versus water-treated leaves.
Counts of GFP-tagged A. pullulans cells/mm2 on peroxide- and water-treated leaves were compared for experiment 1 and experiment 2 for the morphotypes SCC only, blastospores only, and both cell categories combined. In both experiments, no differences (P ≤ 0.05) were found between treatments when treatments alone, treatments and regions, treatments and segments, and treatments and segments and regions were compared. Therefore, no further comparisons were made between peroxide- and water-treated leaves; counts on segments from both treatment procedures were pooled to increase sample sizes, and analyses were conducted with pooled data.
Comparison of GFP-tagged A. pullulans densities on veinal and interveinal sites.
Counts of GFP-tagged A. pullulans cells/mm2 on veins (excluding the midvein) and on the interveinal region at the start of both experiments (segment 1) were not significantly different, allowing comparisons between features to be made at the later sampling times. By 29 June (second sampling point) during experiment 1, GFP-tagged A. pullulans total cells/mm2 and SCC/mm2 on other veinal regions became greater (P ≤ 0.05) (1.2 ± 0.57 and 0.83 ± 0.31, respectively) than the corresponding cells/mm2 on interveinal regions (0.41 ± 0.42 and 0.18 ± 0.19, respectively), and these differences remained thereafter (Fig. 1A to C) [here and elsewhere, data are log10 (x + 1) transformed and given as means ± SD unless otherwise stated]. In general, SCC counts on other veins increased (P ≤ 0.05) after application (0.0 ± 0.0 cells/mm2) to the second sampling point on 29 June (0.83 ± 0.31) and did not increase or decrease after this point (P ≥ 0.05). On interveinal regions, SCC/mm2 increased slightly between the first and second sampling dates (0.004 ± 0.02 and 0.18 ± 0.19, respectively) and remained fairly static. Blastospore densities did not become different among regions at any point and tended to decrease after application. SCC/mm2 values were lower (P ≤ 0.05) than blastospore densities on individual regions (midveins, other veins, and interveinal regions) at application and were higher (P ≤ 0.05) than blastospore densities at each sampling point thereafter, with the exception of interveinal regions, where there were no significant differences after the first sampling date. Similar significant differences were found in the second experiment, but initial cells/mm2 and subsequent changes differed (Fig. 1D to F).
FIG. 1.
Experiment (Expt) 1 (A to C) and experiment 2 (D to F) showing the mean log10 (x + 1) GFP-tagged A. pullulans (GFP-Ap) total cells/mm2 (A and D), SCC/mm2 (B and E), and blastospores/mm2 (C and F) (± SD) on midveins, other veins, and interveinal regions for each experiment (n = 20 for each experiment). Leaves were repeatedly sampled over time in each experiment by use of semidestructive sampling.
At the start of the first experiment, total GFP-tagged A. pullulans counts were significantly higher (P ≤ 0.05) on midveins than on other veins and interveinal regions, which violated our premise of equivalent numbers at all sites at the outset of the experiment. Thus, only general comparisons (without statistical analysis) between the midvein and other sites were made. The changes in cells/mm2 (in both experiments) mirrored those seen for other veins, with densities remaining higher on the midvein throughout both experiments. Counts for GFP-tagged A. pullulans cells/mm2 for all cell types on midveins and on other veins were not significantly different at the start of experiment 2, but by the second sampling point and all remaining points, counts on the midvein were significantly higher (Fig. 1D and E).
Comparison of GFP-tagged A. pullulans densities on individual leaves.
Results from peroxide- and water-treated leaves after both experiments were similar, so only results from peroxide-treated leaves during the first experiment are presented. Overall, increases and/or decreases in GFP-tagged A. pullulans densities on individual leaves (Fig. 2) differed substantially from averages shown in Fig. 1. On midvein sites, counts dropped precipitously on two leaves at the final sampling date (13 July), while three leaves showed large increases (the remainder being intermediate). Similar differences were found on other vein regions, though decreases were not as large as on midvein sites. On interveinal regions there were only slight differences among leaves. However, counts for all but one leaf decreased by the final sampling date (13 July).
FIG. 2.
The mean log10 (x + 1) GFP-tagged A. pullulans (GFP-Ap) cells/mm2 (± SD) on leaf segments from peroxide-treated leaves on midveins (A), other veins (B), and interveinal regions (C) over time (experiment 1). The GFP-tagged cells were applied on 21 June 2004, and by use of semidestructive sampling the same leaves were followed over time (leaves 3 and 10 were lost after 29 June). The initial cells/mm2 on each leaf were adjusted to zero to make comparisons easier among leaves on midveins (D), other veins (E), and interveinal sites (F). Results from experiment 2 were similar.
Formation of colonies over time.
In both experiments, GFP-tagged A. pullulans was distributed after application largely as single cells (Fig. 3). GFP-tagged A. pullulans remained primarily as singletons over interveinal features throughout both experiments. However, colonies developed over midveins and other veins. At the final two sampling dates in both experiments, there were more colonies (P ≤ 0.05) of two or three cells on other veins and midveins than on interveinal regions (data not shown). During the first experiment, there were more colonies (P ≤ 0.05) of four to nine cells on other veins and midveins than on interveinal regions at the 6 July sampling date (on 13 July, colonies of four to nine cells were more numerous on other veins and midveins than on interveinal sites, but differences were not significant) (Fig. 4A). In the second experiment, there were more (P ≤ 0.05) colonies of four to nine cells on midveins and other veins on 9, 18, and 23 September. Colony development in experiment 1 may have been influenced by a large rain event on 4 July (over 32 mm of rain fell in a 24-h period). By the final sampling point, larger colonies were apparent on midveins and other veins.
FIG. 3.
The percentages of GFP-tagged A. pullulans (GFP-Ap) distributed as single cells (
) and in colonies of 2 or 3 cells (
, 4 to 9 cells (
), or ≥10 cells (
) on midveins (A), other veins (B), and interveinal regions (C) during two experiments (Expt.) conducted during the 2004 growing season. GFP-tagged cells were applied on 21 June and 2 September 2004. The same leaves were sampled during each trial by semidestructive sampling. Colony size was defined as the number of GFP-tagged A. pullulans cells in direct contact with each other. Each field was designated as midvein, other vein, or interveinal field, but multiple colonies and/or single cells could be present in each field. Segment collection dates (month-day) are indicated.
FIG. 4.
The percentages of GFP-tagged A. pullulans (GFP-Ap) distributed in colonies of four to nine cells during experiment 1 (A) and experiment 2 (B), both conducted in the 2004 growing season. GFP-tagged cells were applied on 21 June (experiment 1) and 2 September 2004 (experiment 2). The same leaves were sampled during each experiment by semidestructive sampling. Colony size was defined as the number of GFP-tagged A. pullulans cells in direct contact with each other. Each field was designated as midvein (•), other vein (○), or interveinal field (▾), but multiple colonies could be present in each field.
Colonies of ≥10 cells were present on midveins and other veins during both experiments, but in general there were no significant differences between these sites.
Growth from single cells on seedling leaves under controlled conditions.
Results from all three trials were similar, so only results from the first are presented. Mean colony sizes over midveins and other veins were significantly higher than those over interveinal regions at 24, 48, and 72 h (Fig. 5). At 72 h, the mean colony sizes (nontransformed data ± SD hereafter) on midveins, other veins, and interveinal regions were of 36.3 ± 13.6, 30.9 ± 10.7, and 3.5 ± 1.5 cells, respectively. The mean colony size on midveins was not different from that on other veins (P ≤ 0.05) (except in one trial, where midvein colony size was greater). At 72 h, colonies on the veins (data for midveins and other veins were combined) of seedling leaves (data from all three trials were combined) had reached mean colony sizes of 24.4 ± 12.7 cells, and colonies on interveinal sites had reached mean colony sizes of 2.9 ± 1.3 cells.
FIG. 5.
Growth from single GFP-tagged A. pullulans (GFP-Ap) cells on midveins, other veins, and interveinal sites on apple seedling leaves under controlled conditions. Five leaves were inoculated with single cells on three midvein, other-vein, and interveinal sites per leaf. The number of cells in each colony was determined every 24 h for 72 h. These results are representative of three trials conducted with field leaves and of three with seedling leaves.
We also followed the development of colonies from single cells. The small sample size and the methodology precluded formal statistical comparisons, but the following general observations were made. When budding was observed on midvein regions, cells budded from poles (Fig. 6A). The initial mother cell became swollen and budded numerous daughters (Fig. 6A). Daughter cells then tended to bud one or two cells from the poles, and the colony continued to develop linearly (Fig. 6B). New cells tended to be produced at the periphery of the developing colony, and the colonies were confined to the midvein. In contrast, the initial cell on interveinal regions tended to bud one or two cells if budding occurred at all. Occasionally, a daughter cell would form one bud. Colonies tended to overlie epidermal cell junctions and showed less linearity on the rare occasions that colonies of ≥4 cells developed (Fig. 6C).
FIG. 6.
Colony development on leaf features from single GFP-tagged blastospores. (A) Initial cells tended to become swollen, and by 24 h, most budded at least a single cell on interveinal sites and multiple daughter cells on veinal sites. (B) At 72 h, colony development on veinal sites was often linear. (C) On interveinal sites, large colonies showed less linearity, though they were rare (Fig. 3C).
Growth from single cells on field leaves under controlled conditions.
After 72 h, the average GFP-tagged A. pullulans colony size (nontransformed data) on leaves harvested on 6 June was greater (P ≤ 0.05) on midvein sites than on interveinal areas (9.9 ± 3.9 cells on midveins and 1.5 ± 0.5 cells on interveinal regions). When the experiment was repeated with leaves harvested on 30 June, the average GFP-tagged A. pullulans colony sizes were significantly greater at 24, 48, and 72 h on midveins and other veins than on interveinal sites. Results for the midveins and other veins did not significantly differ from each other at any time point. Similar results were found with leaves harvested on 30 September.
Colony sizes on midveins, other veins, and interveinal regions were also compared among sampling dates. Mean midvein colony size at 72 h was significantly larger on older leaves (30 June and 30 September) than on leaves from 6 June (colony sizes were 9.9 ± 3.9, 21.2 ± 13.5, and 29.0 ± 10.1 cells on leaves from 6 June, 30 June, and 30 September, respectively). There were no significant differences between the leaf samples of 30 June and those of 30 September. The same results were found for interveinal regions, with significantly larger colonies on older leaves (1.5 ± 0.5, 3.1 ± 1.9, and 3.0 ± 1.6 cells on leaves from 6 June, 30 June, and 30 September, respectively). Mean colony sizes on other veins did not significantly differ between the 30 June leaves and the 30 September leaves (other veins were not evaluated for the 6 June leaves).
Presence or absence of a growth-inhibiting factor.
Colonies on midveins derived from single GFP-tagged A. pullulans cells plus droplets incubated on interveinal regions for 5 min or 30 min were compared to colonies on midveins derived from single cells plus sterile water droplets. There were no differences (P ≤ 0.05) among treatments in any trial that involved seedling or field leaves (Table 1). When colonies on midveins (derived from single cells plus droplets incubated for 5 or 30 min on interveinal regions and in water alone) were compared to colonies on interveinal sites (derived from cells plus droplets incubated for 5 or 30 min on midveins and in water alone), colony size was always larger (P ≤ 0.05) on midveins (all treatments). These comparisons show no evidence for the presence of an inhibitor on interveinal sites. However, when colonies on interveinal regions were compared among treatments, some differences were present. Mean colony size (at 72 h) from single GFP-tagged A. pullulans cells plus droplets incubated on midvein regions for 30 min and added to interveinal sites were larger (P ≤ 0.05) than colonies on interveinal regions developed from single GFP-tagged A. pullulans cells plus sterile water in one of two trials with seedling leaves and in two of three trials with field leaves (leaves from 15 July and 24 September). There were no significant differences when the same comparisons were made with droplets incubated for 5 min.
TABLE 1.
Assay for site-specific presence of growth inhibitors on either detached seedling or field leaves
| Leaf origin (trial no. or harvest date [mo/day]) | Droplet incubation time (min) | Mean (± SD) size (no. of cells) of A. pullulans colonies developed on:
|
|||
|---|---|---|---|---|---|
| Midveins (with indicated droplet incubation site)b
|
Interveinal regions (with indicated droplet incubation site)c
|
||||
| Interveinal regions | None (control) | Midveins | None (control) | ||
| Phytotron (trial 1) | 0 | NAa | 39.1 ± 20.9 | NA | 5.4 ± 1.5 |
| 5 | 46.8 ± 16.9 | NA | 6.3 ± 4.4 | NA | |
| 30 | 43.9 ± 20.8 | NA | 8.7 ± 5.5 | NA | |
| Phytotron (trial 2) | 0 | NA | 30.9 ± 5.8 | NA | 2.7 ± 0.9 |
| 5 | 28.1 ± 8.0 | NA | 3.1 ± 1.2 | NA | |
| 30 | 30.1 ± 8.2 | NA | 4.9 ± 1.8 | NA | |
| Field (6/27d) | 0 | NA | 29.4 ± 9.4 | NA | 2.6 ± 1.1 |
| 5 | 32.6 ± 7.3 | NA | 3.1 ± 1.4 | NA | |
| 30 | 29.9 ± 12.7 | NA | 2.2 ± 0.6 | NA | |
| Field (7/15) | 0 | NA | 33.0 ± 11.4 | NA | 2.4 ± 1.2 |
| 5 | 35.7 ± 10.8 | NA | 3.1 ± 1.2 | NA | |
| 30 | 35.0 ± 9.8 | NA | 3.6 ± 1.4 | NA | |
| Field (9/24) | 0 | NA | 29.2 ± 9.5 | NA | 3.9 ± 1.4 |
| 5 | 28.4 ± 8.2 | NA | 3.7 ± 1.8 | NA | |
| 30 | 31.5 ± 7.2 | NA | 5.4 ± 2.0 | NA | |
NA, not applicable.
Midvein colonies originated from single GFP-tagged A. pullulans cells added to droplets previously incubated on interveinal regions for 0 (control [never placed on leaves]), 5, or 30 min; these inocula were then placed on midveins for 72 h. Leaves were inoculated on five midvein sites (three leaves/trial with the exception of controls, which had two leaves/trial).
Interveinal colonies originated from single GFP-tagged A. pullulans cells added to droplets previously incubated on midveins for 0, 5, or 30 min; these inocula were then placed on interveinal regions for 72 h. Leaves were inoculated on five intervein sites (three leaves/trial with the exception of controls, which had two leaves/trial).
Dates reflect times of harvest for field leaves. Leaves were detached and evaluated under controlled environmental conditions in the laboratory.
DISCUSSION
Spatial variation in colonization by A. pullulans in the field has been documented previously (5, 17, 21) and, at least since 1971, differences in microbial immigration have been mentioned as a possible cause for this heterogeneity (21). We hypothesized that limited or differential immigration is solely responsible for observed colonization patterns. We tested this by applying GFP-tagged A. pullulans in approximately equivalent numbers to veinal and interveinal regions on leaves in an orchard (t0). The field data did not support our hypothesis, because differences in distribution and colonization quickly developed (t1), with veinal sites being more extensively colonized at least as early as the first week after application. Further strong evidence for rejecting our hypothesis is that colonies developing under controlled environmental conditions from single cells placed on midveins and other veins were significantly larger than those placed on interveinal regions. This finding was consistent for both detached seedling and field leaves harvested on multiple dates.
We could not strictly test the immigration hypothesis in the field with respect to midveins because at the outset they had significantly higher numbers of GFP-tagged A. pullulans cells than other regions did. This apparently resulted from deficiencies in our application method to establish controlled immigration. It may reflect the unique topography or other features of midveins, such as abundant trichomes, which may trap more applied or naturally immigrating inoculum. Throughout most of the 2003 and 2004 growing seasons, densities of wild-type A. pullulans found on midveins were significantly higher than those found on other veins and interveinal regions (17). In general, however, mean colony sizes (derived from single cells) on midveins and other veinal regions at 72 h did not differ significantly either on detached seedling or on field leaves, suggesting that midveins may not be local sites inherently more hospitable for growth than are other veins. Thus, differential immigration could account for the colonization differences that develop between midveins and other veins in the field.
Examination of developing colonies on seedling apple leaves under controlled conditions suggests that topography affects colony morphology and perhaps colony size. Cells budded primarily from the poles, and colonies on veinal sites expanded linearly along veins, remaining confined primarily to the parallel channels above the anticlinal walls of the epidermal cells overlying the veins. In contrast, colonies developing on interveinal sites developed slowly, and when larger colonies occurred (rarely), they tended to expand along the irregular epidermal cell junctions. At a microbial scale, the recessed contours of epidermal cell junctions and undulating veins may be analogous to valleys surrounded by mountains that cannot be scaled. We are currently using inanimate leaf replicas to explore the relationships among colony development, topography, and nutrients.
There is no evidence from our data that poor colonization of interveinal areas results from growth-inhibiting factors. When single cells were applied to midveins in droplets previously incubated on interveinal sites, or in sterile water (control), the resulting colony sizes did not differ significantly. However, when single cells in sterile water, or in droplets incubated on midvein regions for 5 or 30 min, were placed on interveinal regions, subsequent colony size was significantly higher in the droplets incubated for 30 min than in the water controls (in three of five trials). This suggests that a growth-promoting substance was present at levels detectable by our methods on some midvein regions, where they may play a role in spatial variation. Contact with aqueous solutions induces leaching of substances through the cuticle (20, 23, 25). Presumably, the droplets on leaves for 5 min were not in contact with the surface long enough for an effect of leachates to be detected. Further studies are needed to determine what A. pullulans growth-promoting substances are present in veinal areas. Also, the possibility that growth-inhibiting substances may play a role cannot be ruled out completely. For example, our methods did not allow us to eliminate the possibility that nonsoluble inhibitors were present or that other organisms produce substances that inhibit A. pullulans growth on field leaves.
Notwithstanding the apparent presence of growth-promoting factors over veinal sites, colony expansion in the field was slow. While colonies of ≥10 cells increased in frequency, there were fewer than 5% of cells distributed into this size category. These results support previous studies of season-long colonization trends of wild-type A. pullulans, where colonies grew slowly (17). This may be due to limitation of nutrients and/or harsh environmental conditions. Most blastospores converted to SCC, a more adhesive and resistant phenotype (3, 10, 22, 28).
Surprisingly, there did not appear to be differences between peroxide-treated leaves and water-treated leaves in colonization dynamics. Because hydrogen peroxide effectively kills microbes, including A. pullulans, on the apple phylloplane (11) and was shown to do so here, the absence of an effect may indicate that other phylloplane inhabitants have little impact on A. pullulans colonization. This could indicate that A. pullulans occupies a specific leaf microhabitat and that other organisms have little effect because they do not occupy the same microsites. However, because our experimental protocol may have failed to remove dead microbial cells on peroxide-treated leaves, space could remain occupied as on water-treated leaves. Likewise, our surface disinfestation protocol does not address the possibility of growth-promoting substances as the primary factor in A. pullulans colonization. If immersion in peroxide or water resulted in similar removals (or leakages) of nutrients, then subsequent growth levels would not differ between treatments. Further studies of how other phylloplane colonists impact A. pullulans growth and colonization are needed.
This work further demonstrates the advantages of being able to repeatedly sample the same leaves over time (31), thereby uncovering potentially subtle leaf-specific changes that otherwise would be obscured in experimental noise. Similar patterns tended to develop among individual leaves, with densities becoming significantly higher on veinal sites than on interveinal sites. Nevertheless, individual leaves differed both in the amounts of colonization that they supported and in A. pullulans population dynamics on microsites. This may be due to differences in cuticle erosion, which could alter topography, surface wettability, and exudation of nutrients (2, 18, 24), or in other inherent properties. Leaf position could also influence exposure to abiotic factors (such as UV radiation), deposition of exogenous nutrients, or photosynthesis. Differences among leaves likely influence microbial populations in general, not just A. pullulans. We also wonder about the intriguing prospective role that phylogenetic differences in veins and veinal patterns may have in influencing the rates and degrees of phylloplane colonization of different plant species. Answers to these questions require an expanded design, which could include successive semidestructive sampling experiments over the course of the growing season.
Finally, our results may have important implications for biological antagonism, especially because A. pullulans is a potential biocontrol agent (1, 3, 4, 16). Even when A. pullulans was applied at relatively high levels, interveinal sites remained poorly colonized and large colonies were slow to develop, leaving extensive portions of the veins uncolonized. Furthermore, individual leaves differed in the amounts of colonization that they supported, which could severely limit the effectiveness of a foliar antagonist if survival is low on some leaves. Research is needed to determine how enhanced microbial biocontrol on the phylloplane can best be accomplished.
Acknowledgments
Support by NSF grant DEB-0075358 and USDA Hatch grants 142-P446 and 142-4873 is acknowledged with appreciation.
We thank Russell Spear for helping in many ways and Jack Whitney and Keyton Kuykendall for laboratory and field assistance.
Footnotes
Published ahead of print on 1 December 2006.
REFERENCES
- 1.Andrews, J. H., F. M. Berbee, and E. V. Nordheim. 1983. Microbial antagonism to the imperfect stage of the apple scab pathogen, Venturia inaequalis. Phytopathology 73:228-234. [Google Scholar]
- 2.Andrews, J. H., and R. F. Harris. 2000. The ecology and biogeography of microorganisms on plant surfaces. Annu. Rev. Phytopathol. 38:145-180. [DOI] [PubMed] [Google Scholar]
- 3.Andrews, J. H., R. F. Harris, R. N. Spear, G. W. Lau, and E. V. Nordheim. 1994. Morphogenesis and adhesion of Aureobasidium pullulans. Can. J. Microbiol. 40:6-17. [Google Scholar]
- 4.Andrews, J. H., and C. M. Kenerley. 1978. The effects of a pesticide program on non-target epiphytic microbial populations of apple leaves. Can. J. Microbiol. 24:1058-1072. [DOI] [PubMed] [Google Scholar]
- 5.Andrews, J. H., R. N. Spear, and E. V. Nordheim. 2002. Population biology of Aureobasidium pullulans on apple leaf surfaces. Can. J. Microbiol. 48:500-513. [DOI] [PubMed] [Google Scholar]
- 6.Anonymous. 1998. The Difco manual, p. 576-580. Difco Laboratories, Sparks, MD.
- 7.Axtell, C. A., and G. A. Beattie. 2002. Construction and characterization of a proU-gfp transcriptional fusion that measures water availability in a microbial habitat. Appl. Environ. Microbiol. 68:4604-4612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Brodie, I. D. S., and J. P. Blakeman. 1975. Competition for carbon compounds by a leaf surface bacterium and conidia of Botrytis cinerea. Physiol. Plant Pathol. 6:125-135. [Google Scholar]
- 9.Jacobs, J. L., T. L. Carroll, and G. W. Sundin. 2005. The role of pigmentation, ultraviolet radiation tolerance, and leaf colonization strategies in the epiphytic survival of phyllosphere bacteria. Microb. Ecol. 49:104-113. [DOI] [PubMed] [Google Scholar]
- 10.Jacobs, J. L., and G. W. Sundin. 2001. Effect of solar UV-B radiation on a phyllosphere bacterial community. Appl. Environ. Microbiol. 67:5488-5496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kinkel, L. L., and J. H. Andrews. 1988. Disinfestation of living leaves by hydrogen peroxide. Trans. Br. Mycol. Soc. 9:523-528. [Google Scholar]
- 12.Kinkel, L. L., M. Wilson, and S. E. Lindow. 2000. Plant species and plant incubation conditions influence variability in epiphytic bacterial population size. Microb. Ecol. 39:1-11. [DOI] [PubMed] [Google Scholar]
- 13.Koitabashi, M., M. Iwano, and S. Tsushima. 2002. Aromatic substances inhibiting wheat powdery mildew produced by a fungus detected with a new screening method for phylloplane fungi. J. Gen. Plant Pathol. 68:183-188. [Google Scholar]
- 14.Leveau, J. H. J., and S. E. Lindow. 2001. Appetite of an epiphyte: quantitative monitoring of bacterial sugar consumption in the phyllosphere. Proc. Natl. Acad. Sci. USA 98:3446-3453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lindow, S. E. 1991. Determinants of epiphytic fitness in bacteria, p. 295-314. In J. H. Andrews and S. S. Hirano (ed.), Microbial ecology of leaves. Springer-Verlag, New York, NY.
- 16.McCormack, P. J., H. G. Wildman, and P. Jeffries. 1994. Production of antibacterial compounds by phylloplane inhabiting yeasts and yeastlike fungi. Appl. Environ. Microbiol. 60:927-931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.McGrath, M. J., and J. H. Andrews. 2006. Temporal changes in microscale colonization of the phylloplane by Aureobasidium pullulans. Appl. Environ. Microbiol. 72:6234-6241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Mechaber, W. L., D. B. Marshall, R. A. Mechaber, R. T. Jobe, and F. S. Chew. 1996. Mapping leaf surface landscapes. Proc. Natl. Acad. Sci. USA 93:4600-4603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mercier, J., and S. E. Lindow. 2000. Role of leaf surface sugars in colonization of plants by bacterial epiphytes. Appl. Environ. Microbiol. 66:369-374. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Potter, C. S., H. L. Ragsdale, and W. T. Swank. 1991. Atmospheric deposition and foliar leaching in regenerating southern Appalachian forest canopy. J. Ecology 79:97-115. [Google Scholar]
- 21.Pugh, G. J. F., and N. G. Buckley. 1971. The leaf surface as a substrate for colonization by fungi, p. 431-445. In T. F. Preece and C. H. Dickinson (ed.), Ecology of leaf surface micro-organisms. Academic Press, New York, NY.
- 22.Ramos, S., and A. I. Garcia. 1975. A vegetative cycle of Pullularia pullulans. Trans. Br. Mycol. Soc. 64:129-135. [Google Scholar]
- 23.Rea, A. W., S. E. Lindberg, and G. J. Keeler. 2000. Assessment of dry deposition and foliar leaching of mercury and selected trace elements based on washed foliar and surrogate surfaces. Environ. Sci. Technol. 34:2418-2425. [Google Scholar]
- 24.Schreiber, L., M. Skrabs, K. D. Hartmann, P. Diamantopoulos, E. Simanova, and J. Santrucek. 2001. Effect of humidity on cuticular water permeability of isolated cuticular membranes and leaf disks. Planta 214:274-282. [DOI] [PubMed] [Google Scholar]
- 25.Tukey, H. B., Jr. 1970. Leaching of substances from plants. Annu. Rev. Phytopathol. 21:305-324. [Google Scholar]
- 26.Vanden Wymelenberg, A. J., D. Cullen, R. N. Spear, and J. H. Andrews. 1997. Expression of green fluorescent protein in Aureobasidium pullulans and quantification of the fungus on leaf surfaces. BioTechniques 23:686-690. [DOI] [PubMed] [Google Scholar]
- 27.Wilson, D., and G. C. Carroll. 1994. Infection studies of Discula quercina, an endophyte of Quercus garryana. Mycologia 86:635-647. [Google Scholar]
- 28.Wilson, H. A., V. G. Lilly, and J. G. Leach. 1965. Bacterial polysaccharides. IV. Longevity of Xanthomonas phaseoli and Serratia marcescens in bacterial exudates. Phytopathology 55:1135-1138. [PubMed] [Google Scholar]
- 29.Wilson, M., S. S. Hirano, and S. E. Lindow. 1999. Location and survival of leaf-associated bacteria in relation to pathogenicity and potential for growth within the leaf. Appl. Environ. Microbiol. 65:1435-1443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wilson, M., and S. E. Lindow. 1994. Inoculum density-dependent mortality and colonization of the phyllosphere by Pseudomonas syringae. Appl. Environ. Microbiol. 60:2232-2237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Woody, S. T., R. N. Spear, E. V. Nordheim, A. R. Ives, and J. H. Andrews. 2003. Single-leaf resolution of the temporal population dynamics of Aureobasidium pullulans on apple leaves. Appl. Environ. Microbiol. 69:4892-4900. [DOI] [PMC free article] [PubMed] [Google Scholar]