Species Richness of Yeast Communities in Floral Nectar of Southern Spanish Plants
|
|  Yüklə 0.72 Mb.
|
| Species Richness of Yeast Communities in Floral Nectar of Southern Spanish Plants
María I. Pozo & Carlos M. Herrera & Pilar Bazaga Abstract Floral nectar of insect-pollinated plants often contains dense yeast populations, yet little quantitative information exists on patterns and magnitude of species richness of nectar-dwelling yeasts in natural plant commu- nities. This study evaluates yeast species richness at both the plant community and plant species levels in a montane forest area in southern Spain, and also explores possible correlations between the incidence of different yeast species in nectar and their reported tolerance to high sugar concentrations, and between yeast diversity and pollinator composition. Yeast species occurring in a total of 128 field- collected nectar samples from 24 plant species were identified by sequencing the D1/D2 domain of the large subunit rDNA, and rarefaction-based analyses were used to estimate yeast species richness at the plant community and plant species levels, using nectar drops as elemental sampling units. Individual nectar samples were generally characterized by very low species richness (1.2 yeast species/sample, on average), with the ascomycetous Metschnikowia reukaufii and Metschnikowia gruessii ac- counting altogether for 84.7% of the 216 isolates identified. Other yeasts recorded included species in the genera Aureobasidium, Rhodotorula, Cryptococcus, Sporobolomy- ces, and Lecythophora. The shapes and slopes of observed richness accumulation curves were quite similar for the nectar drop and plant species approaches, but the two approaches yielded different expected richness estimates. Expected richness was higher for plant species-based than for nectar drop-based analyses, showing that the coverage  M. I. Pozo (*) : C. M. Herrera : P. Bazaga
Estación Biológica de Doñana, Consejo Superior de Investigaciones Científicas (CSIC), Avenida Américo Vespucio s/n, Isla de La Cartuja 41092 Seville, Spain e-mail: maribelpozo@ebd.csic.es of nectar yeast species occurring in the region would be improved by sampling additional host plant species. A significant correlation was found between incidence of yeast species in nectar and their reported ability to grow in a medium containing 50% glucose. Neither diversity nor incidence of yeasts was correlated with pollinator compo- sition across plant species. Introduction
Microbial diversity differs widely across habitats. In marine microbial communities, for example, published estimates range from hundreds of operational taxonomic units (OTUs) per milliliter in open water [42] to thousands in sediments [21]. Apart from microbial surveys carried out in soil and aquatic environments, other terrestrial environ- ments such as plant-associated microhabitats have been studied less frequently [10, 39]. The diversity of yeast communities, and more specifically those that inhabit floral nectar, still remains a comparatively unexplored subsystem of the microbial communities associated with plants. Although it has been long known that yeasts are frequent and can reach high densities in floral nectar [4, 18, and references therein], relatively little is known on their diversity levels and the factors that influence it [17]. It may be hypothesized that nectar yeast communities are shaped by the interplay between two conflicting forces. On one side, floral nectars, owing to its high energetic content [36], could be considered highly favorable media for microbial growth, but on the other, their high osmotic pressure and frequent presence of secondary compounds [1, 25, 33] could frequently limit the number of yeast species that finally constitute nectar communities [17 and refer- ences therein]. In addition to its physical and chemical properties, floral nectar has another particularity. Yeasts need vectors to colonize floral nectar, hence insect visitation history and pollinator identity can be two additional factors influencing the diversity and composition of nectar yeast communities [5, 20, 29]. The main objective of this study was to assess quantitatively the species richness of yeast communities in the floral nectar of a large sample of southern Spanish insect-pollinated plants. This study will also test if observed differences between yeast species in their frequency of occurrence in floral nectar are related to gross physiological differences, by relating the incidence of individual species with their reported tolerance to high osmolarity conditions. In addition, possible corre- lations across plant species between yeast species richness and pollinator composition will also be explored. Materials and Methods
24 plant species belonging to nine different families were cultured for yeast identification. A complete list of the species surveyed and their familial affiliations is shown in the “Appendix.” Fabaceae (21% of species), Lamiaceae, Planta- ginaceae, Iridaceae (17% of species each), Caprifoliaceae, and Ranunculaceae (8% of species each) were the families contributing most species to our sample. Brassicaceae, Oleaceae, and Solanaceae each contributed a single species to the sample. The species studied here are a subset of those examined microscopically for occurrence of nectar- inhabiting yeasts in the same region by Herrera et al. [18]. Flowering branches, inflorescences, or individual flow- ers of study species were collected in the field and preserved in a portable cooler inside plastic bags or glass jars for a few hours until taken indoors, and then kept at ambient temperature. Extractions of 1 μl nectar samples, using calibrated microcapillaries, were conducted around 12–24 h after field collection. Ten nectar drops per plant species were streaked individually onto yeast malt agar plates (1.0% glucose, 0.5% peptone, 0.3% malt extract, 0.3% yeast extract, and 2.0% agar) with 0.01% chloram- phenicol and incubated at 25°C. Isolates were obtained from the resulting colonies following standard morpholog- ical criteria described in Yarrow [50]. The D1/D2 domain of the 26S subunit ribosomal DNA was two-way sequenced for all the isolates as described in [27]. Gblocks [7] was used to trim the resulting alignment so that the beginning and ends of the consensus sequences were all the same. Although physiological characteristics are as important as molecular characterization for the identification of yeast species, for practical reasons, we relied exclusively on DNA sequences for identification, which was accomplished by BLAST-querying the GenBank database (last accessed 17 December 2008). Although BLAST searches generally achieved very high similarity scores (usually between 90% and 100%, 91% of the sequences being above 98% similarity score), possible biases in richness estimates caused by the presence of undescribed species were also considered. To this end, this study evaluated OTUs defined on the basis of similarity of DNA sequences [22, 30, 43]. Determination of the number of distinct molecular OTUs occurring in a set of DNA sequences, and assignment of sequences to OTUs, was done with the program DOTUR (Distance-Based OTU and Richness [42]). A PHYLIP (http://evolution.genetics.washington.edu/phylip.html)- generated molecular distance matrix was used as input to DOTUR, which assigned sequences to OTUs based on a predetermined distance threshold. Pairs of isolates with molecular distances smaller than the chosen threshold were considered as belonging to the same OTU. A DNA dissimilarity cut-off of 3% was used in the analyses reported below. Although this value is larger than the 1% threshold suggested for species-level rDNA differentiation in yeasts [24; but see 28], this value was chosen because it has been commonly used to distinguish “molecular” fungal
species in environmental studies [37], which would facilitate comparisons with other studies. In any case, the main conclusions of the DOTUR analyses were robust to variations in the OTU discrimination threshold in the range 1–3% (results not shown). Statistical Analyses Densities of up to 105 yeast cells/µl of nectar are often found in the wild [18], hence direct assessment of true yeast diversity in our samples was impractical, and we had to rely on sampling [22]. Sample-based rarefaction methods, applied to species presence-absence data, were used to assess overall yeast species richness at the plant community level following procedures described by Colwell [9] and Gotelli and Colwell [16]. Nectar drops will provide the elemental “samples” for all analyses, but two different approaches will be adopted to evaluate yeast species richness. In the first one, species occurrence data from all nectar samples will be analyzed together, irrespective of the plant species of origin. This procedure will provide “drop- based” rarefaction curves that assess overall species richness of nectar yeasts at the particular multispecific set of plant species which was sampled. Differences between plant species in life style or floral characteristics could influence their nectar yeast communities, so the second approach considered separately the different host plant species from which nectar samples had been collected. In these analyses, data from all individual nectar samples from the same plant species were combined into a single sample, and “plant species-based” rarefaction curves were obtained. Drop-based and species-based average rarefaction curves were computed with the EstimateS 8.0 program [9], using 50 randomizations and sampling without replacement. Additionally, in order to estimate yeast species richness expected in nectar, the nonparametric estimators ICE and Chao2 were used, because our taxa richness data are based on incidence [8]. Rarefaction generates the expected number of species in a small collection of n samples drawn at random from the large pool of N samples [44]. In contrast, richness estimators predict the total richness of a community from samples. This study compared nonpara- metric estimators with the rarefaction analyses in order to evaluate the accuracy of richness estimates. To test whether the frequency of occurrence of individ- ual yeast species in floral nectar was related to their osmotolerance, yeasts species were classified into two categories according to their reported growth response in 50% glucose tests, namely, “osmophilic” (positive growth response) and “non-osmophilic” (variable or negative growth response). Physiological data were obtained from Barnett et al. [3] and the on-line CBS yeast database available at http://www.cbs.knaw.nl/yeast/BioloMICS.aspx, except for Aureobasidium pullulans and Metschnikowia unknown sp., for which we determined experimentally their responses to variable glucose concentration. Frequencies of occurrence of osmophilic vs. non-osmophilic species in both nectar drops and plant species were compared with the Wilcoxon rank-sum test, using the NPAR1WAY procedure in SAS and exact P value estimation (SAS Institute, Cary, NC, USA). Correlations across plant species between pollinator composition and yeast incidence were explored using detailed data on pollinator composition obtained in the study region for 12 of the 24 plant species sampled (18, and C. M. Herrera, unpublished data). Only for this analysis, we refer to “diversity” instead of “species richness,” as both species richness and relative species abundances are incorporated into the Shannon–Wiener diversity indices. A principal components analysis was conducted on the variance– covariance matrix of the proportion of flower visits contributed by bumblebees, solitary bees, Lepidoptera, Coleoptera, and Diptera in order to reduce the number of variables describing pollinator composition of each plant species. The fist component, which was mainly correlated with the frequency of flower visits by bumblebees and solitary bees (correlations = 0.6 and 0.4, respectively), accounted for the 73.2% of total variance. Relationships between nectar yeast incidence and diversity, and pollinator composition, were explored by correlating the frequency of occurrence, yeast abundance, and yeast diversity (Shannon– Wiener index) of yeast cells in nectar, on one side, with the pollinator composition in each plant represented by scores on the first principal components analysis component, on the other. The yeast-related variables mentioned above are also correlated with pollinator diversity calculated by Shannon– Wiener index. For statistical analyses, nonparametric tests were applied (Spearman rank correlation) as implemented in SAS 9.1 statistical package (SAS Institute, Cary, NC, USA) CORR procedure. Results
 Table 1 Plant species from which nectar yeast isolates were achieved, number of DNA isolates, number of DNA isolates containing different yeast taxa
Plant species DNA sequences, N Metschnikowia reukaufii, N Metschnikowia gruessii, N Aureobasidium pullulans, N Cryptococcus spp.a, N Rhodotorula spp.b, N Sporobolomyces roseus, N Lecythophora hoffmannii, N Metschnikowia sp., N   
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||