Final Report Summary - PHENOGENYEAST (Exploring the phenotypic landscape of nectar yeasts in relation to their genetic background)
4.1.1 State-of-the-art and setting of the research
One of the most widespread interactions between plant and animal species consists of plant–pollinator interactions. In those binary systems, nectar is the most common form of reward that plants offer to their animal counterparts in return for pollination. Although it has generally been assumed that nectar properties represent intrinsic plant features that are stable in time, it has recently been shown that nectar is commonly infested with microorganisms, most often yeasts belonging to the Metschnikowia genus. The genus Metschnikowia is characteristically attached to the flower-insect interface. Once they are vectored by insects to floral nectar, they are able to grow at very fast rates, reaching high abundances in individual nectaries in a short period of time and deterring a great part of other potential nectar colonizers, either related to insects, phylloplane, pollen or air. Previous research has shown that this genus possesses specific physiological properties that may explain their preponderance in nectar, including low dependence on nitrogen, fast glucose fermentation, resistance to ethanol and plant secondary compounds, and high sugar concentration tolerance. Although they normally reproduce clonally under natural conditions, a high genotypic diversity has been documented in both M. reukaufii and M. gruessii, which are the two most important species inhabiting floral nectar in Europe. Because yeasts are transported between flowers of different plant species and because nectar characteristics can differ substantially between plant species, they inevitably have to cope with contrasting chemical environments. Therefore, diversifying selection has been postulated as a key mechanism maintaining high genotypic diversity in these organisms. However, at present little is known about how nectar yeast species adapt to chemical features, implying that there is very little knowledge that supports this “diversifying selection” hypothesis.
4.1.2 Aims
In order to better understand yeast-mediated plant-pollinator interactions, the physiological characteristics of nectar-inhabiting yeasts need to be better characterized. Moreover, since nectar yeast populations have to cope with extreme environments, studying their physiological profile will undoubtedly contribute to better insights into their ecology. The overall objective of this project was to study interspecific and intraspecific variation in phenotypic traits in nectar-dwelling yeast species in relation to their origin, nectar features and their genotypic background. First, we investigated whether nectar yeast communities are organized according to physiological features of the two main yeast species / strains involved. Later, we tested the hypothesis that phenotypic clusters can be predicted by different environmental parameters related to the origin of these strains (e.g. plant host species, elevation, geographic origin, or collection date) or to the nectar chemical environment (nectar pH, sugar concentration, or nectar sugar composition). Third, we assessed the relative importance of the intraspecific component of phenotypic variation among nectar yeast strains, and the possible mechanisms contributing to keep this diversity. We asked whether genotypic data for these strains can be matched with their physiological profile and whether diversifying selection is taking place in nectar-inhabiting yeasts, thus giving rise to a wide genetic variance and a corresponding phenotypic one, as a result of a genotype x environment interactions.
More specifically, we aimed at:
1) testing the prediction that niche differentiation and trait divergence explain the co-occurrence of two yeast species (Metschnikowia reukaufii and M. gruessii) in floral nectar.
2) studying the phenotypic landscape explored by the nectar specialists M. reukaufii and M. gruessii, and on the role of nectar chemistry characteristics in determining phenotypic clustering of nectar yeasts.
3) testing the hypothesis that differential adaptation of yeast genotypes to the specific nectar conditions imposed by different host plant species or different individuals of the same host species contributes to creating and maintaining intraspecific phenotypic diversity, was tested by examining possible mechanisms (strain genotype, genotype x environment interaction) that lead to the formation of phenotypic clusters within a single yeast species.
4) providing a thorough review on nectar inhabiting microorganisms in floral nectar, and their effect on plant fitness and pollinator attraction.
To reach these goals, we focused on two important yeast species that have been frequently found in floral nectar, i.e. M. gruessii and M. reukaufii. To that end, we assembled a vast strain collection of over 1000 living Metschnikowia isolates which were isolated in 2008 and 2013 in SE Spain. Environmental data related to the origin of the yeast isolates were also collected, including elevation, sampling site, collection date, and chemical properties of pristine nectar samples from the host plant species (sugar composition by HPLC analyses, nectar pH, total sugar concentration). A battery of 47 tests that are specifically relevant for the ecology of nectar yeasts was used to assess phenotypic variation among isolates and species. Phenotypic variation was assessed by measuring the growth on parameter plates after 2 days of incubation, thus mimicking the average lifespan of a flower in our sampling area. To increase reproducibility and accuracy we used a high-throughput screening platform. Data on phenotypic characteristics were grouped into 5 categories (carbon sources, nitrogen sources, hydrolysis, low water activity, and inhibitors) for further phenotypic characterization of the isolates. In order to obtain the genotype of the isolates we used Amplified Fragment Length Polymorphism (AFLP), a genome-wide scan method that was optimized for these species.
4.1.3 Major results
1. Interspecific variation: M. gruessii vs M. reukaufii, nectar specialized yeasts isolated in sympatry in SE Spain.
Identifying the ecological processes that underlie the distribution and abundance of species in microbial communities is a central issue in microbial ecology and evolution. Classical trade-off based niche theories of resource competition predict that co-occurrence in microbial communities is more likely when the residing species show trait divergence and complementary resource use. In this paper, we tested the prediction that niche differentiation and trait divergence explain the co-occurrence of two yeast species (Metschnikowia reukaufii and M. gruessii) in floral nectar. Assessment of the phenotypic landscape showed that both species displayed a significantly different physiological profile. Comparison of utilization profiles in single vs mixed cultures indicated that M. reukaufii and M. gruessii do not compete for most carbon and nitrogen sources. In mixed cultures, M. reukaufii grew better in sucrose solutions and in the presence of the antimicrobial compound digitonin than when grown as pure culture. M. gruessi on the other hand grew better in mixed cultures in glucose and fructose solutions (Fig. 1). Overall, these results indicate that niche differentiation and resource partitioning are important mechanisms contributing to species co-occurrence in nectar yeast communities and suggest that facilitation may occur within this intriguing consortium of two co-occurring Metschnikowia species. The results of this research were submitted for publication to Environmental Microbiology.
Fig. 1 Final growth after 2 days, measured as cells per µL, in nectar-mimicking parameters measured in M. gruessii and M. reukaufii in single culture vs. co-culture of the two species. Each point is the mean value obtained for a species, involving 2 strains per species and 2 replicas per strain, and vertical bars represent standard errors. The letters denote comparisons within each species across different conditions (mixed or single) and comparisons within single and mixed conditions across species. Means with the same letter were not significantly different at P < 0.05.
2. Intra-specific variation in nectarivorous yeasts
Floral nectars become easily colonized by microbes, most often species of the ascomycetous yeast genus Metschnikowia. Although it is known that nectar composition can vary tremendously among plant species, most likely corresponding to the nutritional requirements of their main pollinators, far less is known about how variation in nectar chemistry affects intraspecific variation in nectarivorous yeasts. Because variation in nectar traits likely affects growth and abundance of nectar yeasts, nectar yeasts can be expected to display large phenotypic variation in order to cope with varying nectar conditions. To test this hypothesis, we related variation in the phenotypic landscape of a vast collection of nectar-living yeast isolates from two Metschnikowia species (M. reukaufii and M. gruessii) to nectar chemical traits using non-linear canonical correspondence analyses. Nectar yeasts were collected from a total of 19 plant species from different plant families to include as much variation in nectar chemical traits as possible. For 19 species from the 25 plant species from which the strains were isolated, nectar samples and information on the nectar properties was newly collected. To ensure no yeast contamination, plants were bagged and nectar was plated, examined microscopically and filtered. Chemical features measured included pH, sugar composition (sucrose, glucose, and fructose) and sugar concentration.
Fig. 2 nMDS ordination plots of M. reukaufii (left panel) and M. gruessii (right panel) isolates. Symbol shapes indicate the plant family from which each isolate was recovered. The gradient of maximum total sugar concentration is fitted into the ordination diagrams using thinplate splines and is indicated by the green contour lines. Upper panel. All tests. The fit statistics were: R-sq.(adj) = 0. 35 (left) and 0.31 (right), p < 0.0001. Middle panel. Carbon sources. R-sq. (adj) = 0.23 and 0.23 p < 0.0001 for left and right figure, respectively. Bottom panel. Inhibitors. Fit statistics were R-sq. (adj) = 0.30 and 0.15 p < 0.0001 and p = 0.001 for left and right figure, respectively.
As expected, the two nectar yeasts displayed large variation in phenotypic traits (Fig. 2), particularly in traits related to growth performance in carbon sources and inhibitors, which was significantly related to the host plant from which they were isolated. Total sugar concentration and relative fructose content significantly explained the observed variation in the phenotypic profile of the investigated yeast species, indicating that sugar concentration and composition are the key traits that affect phenotypic variation in nectarivorous yeasts. The results of this study have recently been published in FEMS Microbiology Ecology (see 4.2).
3. Genotype x phenotype interaction
An overall aim in biology is to understand how selection for certain traits in the context of an organism´s ecology favors genetic diversity and the establishment of distinct variants within populations. Widely distributed clonal organisms that are present in heterogeneous habitats, such as some yeast species, present a particularly interesting system to address the mechanistic base of genetic diversity. Moreover, the systematic study of natural genetic variation in yeasts in natural environments helps us to unravel complex interactions between ecology, selective pressures, traits and molecular mechanisms. Besides, it is still unclear to what extent heritable genetic variation finally ends up into measurable and predictable ecological consequences, i.e. as affecting phenotypic traits. Such an outcome depends, firstly, on the existence of strong genotype-phenotype links.
To that end, we have carried out the first attempt to associate genotypic and phenotypic variation in natural populations of nectar-living yeasts. In our study, we have used allelic information on 76 and 118 AFLP loci for two nectar-living yeast species and results from 45 phenotypic tests. We compared the phenotypic profile of 29 M. reukaufii and 50 M. gruessii strains, respectively, with their genotype profile (AFLP) by using new statistical approaches that permitted the calculation of correlations between two matrices that strongly differ in ranks (growth measured by area and presence/absence data in AFLP profile). We found a 20% of correlation between phenotype and AFLP dissimilarities matrices for M. gruessii nectar-living strains (Fig. 3), and nearly zero in the particular case of M. reukaufii strains. This kind of correlation studies has been performed in the genus Saccharomyces, in order to find more efficient ways to pick strains with interesting traits from an industrial perspective, as winery or brewery. Although the extensive knowledge on Saccharomyces genome enables the design of more adequate methods to define the genotype of the isolates, the correlation found for M. gruessii is not far from the level of correlation found in these studies (30%), suggesting that correlations between genotypic and phenotypic properties are widespread in yeasts.
Moreover, our results showed that environmental differences were much more relevant for genetic and phenotypic variation than geographic distance among populations. The results of this study are currently being written down in a manuscript that will be submitted to Molecular Ecology.
Fig. 3 Metschnikowia gruessii phenotype vs. genetic relatedness. Clustering of the strains based on AFLP genotype was established using a combination of the Dice coefficient (for generating the distance matrix) and the Unweighted Pair Group Method with Arithmetic Mean (UPGMA) clustering algorithm. Phenotypic maps depict growth for strains in different test conditions, as indicated in the top bar. Font colour in plant species column correspond to the different sites where samples where retrieved.
4. Review on the effect of nectar microbes on plant-pollinator interactions.
Finally, we wrote a thorough review on the effects of nectar microbes on plant-pollinator interactions. More specifically, we reviewed the latest investigations that were performed on nectar chemistry and microbial-mediated changes on nectar chemistry and pollination attraction and compared these results with previous studies. Besides, we reviewed nectar community assembly, and we described the high intraspecific variation in the Metschnikowia genus and its mechanistic base with the data gathered during this project.
One of the most widespread interactions between plant and animal species consists of plant–pollinator interactions. In those binary systems, nectar is the most common form of reward that plants offer to their animal counterparts in return for pollination. Although it has generally been assumed that nectar properties represent intrinsic plant features that are stable in time, it has recently been shown that nectar is commonly infested with microorganisms, most often yeasts belonging to the Metschnikowia genus. The genus Metschnikowia is characteristically attached to the flower-insect interface. Once they are vectored by insects to floral nectar, they are able to grow at very fast rates, reaching high abundances in individual nectaries in a short period of time and deterring a great part of other potential nectar colonizers, either related to insects, phylloplane, pollen or air. Previous research has shown that this genus possesses specific physiological properties that may explain their preponderance in nectar, including low dependence on nitrogen, fast glucose fermentation, resistance to ethanol and plant secondary compounds, and high sugar concentration tolerance. Although they normally reproduce clonally under natural conditions, a high genotypic diversity has been documented in both M. reukaufii and M. gruessii, which are the two most important species inhabiting floral nectar in Europe. Because yeasts are transported between flowers of different plant species and because nectar characteristics can differ substantially between plant species, they inevitably have to cope with contrasting chemical environments. Therefore, diversifying selection has been postulated as a key mechanism maintaining high genotypic diversity in these organisms. However, at present little is known about how nectar yeast species adapt to chemical features, implying that there is very little knowledge that supports this “diversifying selection” hypothesis.
4.1.2 Aims
In order to better understand yeast-mediated plant-pollinator interactions, the physiological characteristics of nectar-inhabiting yeasts need to be better characterized. Moreover, since nectar yeast populations have to cope with extreme environments, studying their physiological profile will undoubtedly contribute to better insights into their ecology. The overall objective of this project was to study interspecific and intraspecific variation in phenotypic traits in nectar-dwelling yeast species in relation to their origin, nectar features and their genotypic background. First, we investigated whether nectar yeast communities are organized according to physiological features of the two main yeast species / strains involved. Later, we tested the hypothesis that phenotypic clusters can be predicted by different environmental parameters related to the origin of these strains (e.g. plant host species, elevation, geographic origin, or collection date) or to the nectar chemical environment (nectar pH, sugar concentration, or nectar sugar composition). Third, we assessed the relative importance of the intraspecific component of phenotypic variation among nectar yeast strains, and the possible mechanisms contributing to keep this diversity. We asked whether genotypic data for these strains can be matched with their physiological profile and whether diversifying selection is taking place in nectar-inhabiting yeasts, thus giving rise to a wide genetic variance and a corresponding phenotypic one, as a result of a genotype x environment interactions.
More specifically, we aimed at:
1) testing the prediction that niche differentiation and trait divergence explain the co-occurrence of two yeast species (Metschnikowia reukaufii and M. gruessii) in floral nectar.
2) studying the phenotypic landscape explored by the nectar specialists M. reukaufii and M. gruessii, and on the role of nectar chemistry characteristics in determining phenotypic clustering of nectar yeasts.
3) testing the hypothesis that differential adaptation of yeast genotypes to the specific nectar conditions imposed by different host plant species or different individuals of the same host species contributes to creating and maintaining intraspecific phenotypic diversity, was tested by examining possible mechanisms (strain genotype, genotype x environment interaction) that lead to the formation of phenotypic clusters within a single yeast species.
4) providing a thorough review on nectar inhabiting microorganisms in floral nectar, and their effect on plant fitness and pollinator attraction.
To reach these goals, we focused on two important yeast species that have been frequently found in floral nectar, i.e. M. gruessii and M. reukaufii. To that end, we assembled a vast strain collection of over 1000 living Metschnikowia isolates which were isolated in 2008 and 2013 in SE Spain. Environmental data related to the origin of the yeast isolates were also collected, including elevation, sampling site, collection date, and chemical properties of pristine nectar samples from the host plant species (sugar composition by HPLC analyses, nectar pH, total sugar concentration). A battery of 47 tests that are specifically relevant for the ecology of nectar yeasts was used to assess phenotypic variation among isolates and species. Phenotypic variation was assessed by measuring the growth on parameter plates after 2 days of incubation, thus mimicking the average lifespan of a flower in our sampling area. To increase reproducibility and accuracy we used a high-throughput screening platform. Data on phenotypic characteristics were grouped into 5 categories (carbon sources, nitrogen sources, hydrolysis, low water activity, and inhibitors) for further phenotypic characterization of the isolates. In order to obtain the genotype of the isolates we used Amplified Fragment Length Polymorphism (AFLP), a genome-wide scan method that was optimized for these species.
4.1.3 Major results
1. Interspecific variation: M. gruessii vs M. reukaufii, nectar specialized yeasts isolated in sympatry in SE Spain.
Identifying the ecological processes that underlie the distribution and abundance of species in microbial communities is a central issue in microbial ecology and evolution. Classical trade-off based niche theories of resource competition predict that co-occurrence in microbial communities is more likely when the residing species show trait divergence and complementary resource use. In this paper, we tested the prediction that niche differentiation and trait divergence explain the co-occurrence of two yeast species (Metschnikowia reukaufii and M. gruessii) in floral nectar. Assessment of the phenotypic landscape showed that both species displayed a significantly different physiological profile. Comparison of utilization profiles in single vs mixed cultures indicated that M. reukaufii and M. gruessii do not compete for most carbon and nitrogen sources. In mixed cultures, M. reukaufii grew better in sucrose solutions and in the presence of the antimicrobial compound digitonin than when grown as pure culture. M. gruessi on the other hand grew better in mixed cultures in glucose and fructose solutions (Fig. 1). Overall, these results indicate that niche differentiation and resource partitioning are important mechanisms contributing to species co-occurrence in nectar yeast communities and suggest that facilitation may occur within this intriguing consortium of two co-occurring Metschnikowia species. The results of this research were submitted for publication to Environmental Microbiology.
Fig. 1 Final growth after 2 days, measured as cells per µL, in nectar-mimicking parameters measured in M. gruessii and M. reukaufii in single culture vs. co-culture of the two species. Each point is the mean value obtained for a species, involving 2 strains per species and 2 replicas per strain, and vertical bars represent standard errors. The letters denote comparisons within each species across different conditions (mixed or single) and comparisons within single and mixed conditions across species. Means with the same letter were not significantly different at P < 0.05.
2. Intra-specific variation in nectarivorous yeasts
Floral nectars become easily colonized by microbes, most often species of the ascomycetous yeast genus Metschnikowia. Although it is known that nectar composition can vary tremendously among plant species, most likely corresponding to the nutritional requirements of their main pollinators, far less is known about how variation in nectar chemistry affects intraspecific variation in nectarivorous yeasts. Because variation in nectar traits likely affects growth and abundance of nectar yeasts, nectar yeasts can be expected to display large phenotypic variation in order to cope with varying nectar conditions. To test this hypothesis, we related variation in the phenotypic landscape of a vast collection of nectar-living yeast isolates from two Metschnikowia species (M. reukaufii and M. gruessii) to nectar chemical traits using non-linear canonical correspondence analyses. Nectar yeasts were collected from a total of 19 plant species from different plant families to include as much variation in nectar chemical traits as possible. For 19 species from the 25 plant species from which the strains were isolated, nectar samples and information on the nectar properties was newly collected. To ensure no yeast contamination, plants were bagged and nectar was plated, examined microscopically and filtered. Chemical features measured included pH, sugar composition (sucrose, glucose, and fructose) and sugar concentration.
Fig. 2 nMDS ordination plots of M. reukaufii (left panel) and M. gruessii (right panel) isolates. Symbol shapes indicate the plant family from which each isolate was recovered. The gradient of maximum total sugar concentration is fitted into the ordination diagrams using thinplate splines and is indicated by the green contour lines. Upper panel. All tests. The fit statistics were: R-sq.(adj) = 0. 35 (left) and 0.31 (right), p < 0.0001. Middle panel. Carbon sources. R-sq. (adj) = 0.23 and 0.23 p < 0.0001 for left and right figure, respectively. Bottom panel. Inhibitors. Fit statistics were R-sq. (adj) = 0.30 and 0.15 p < 0.0001 and p = 0.001 for left and right figure, respectively.
As expected, the two nectar yeasts displayed large variation in phenotypic traits (Fig. 2), particularly in traits related to growth performance in carbon sources and inhibitors, which was significantly related to the host plant from which they were isolated. Total sugar concentration and relative fructose content significantly explained the observed variation in the phenotypic profile of the investigated yeast species, indicating that sugar concentration and composition are the key traits that affect phenotypic variation in nectarivorous yeasts. The results of this study have recently been published in FEMS Microbiology Ecology (see 4.2).
3. Genotype x phenotype interaction
An overall aim in biology is to understand how selection for certain traits in the context of an organism´s ecology favors genetic diversity and the establishment of distinct variants within populations. Widely distributed clonal organisms that are present in heterogeneous habitats, such as some yeast species, present a particularly interesting system to address the mechanistic base of genetic diversity. Moreover, the systematic study of natural genetic variation in yeasts in natural environments helps us to unravel complex interactions between ecology, selective pressures, traits and molecular mechanisms. Besides, it is still unclear to what extent heritable genetic variation finally ends up into measurable and predictable ecological consequences, i.e. as affecting phenotypic traits. Such an outcome depends, firstly, on the existence of strong genotype-phenotype links.
To that end, we have carried out the first attempt to associate genotypic and phenotypic variation in natural populations of nectar-living yeasts. In our study, we have used allelic information on 76 and 118 AFLP loci for two nectar-living yeast species and results from 45 phenotypic tests. We compared the phenotypic profile of 29 M. reukaufii and 50 M. gruessii strains, respectively, with their genotype profile (AFLP) by using new statistical approaches that permitted the calculation of correlations between two matrices that strongly differ in ranks (growth measured by area and presence/absence data in AFLP profile). We found a 20% of correlation between phenotype and AFLP dissimilarities matrices for M. gruessii nectar-living strains (Fig. 3), and nearly zero in the particular case of M. reukaufii strains. This kind of correlation studies has been performed in the genus Saccharomyces, in order to find more efficient ways to pick strains with interesting traits from an industrial perspective, as winery or brewery. Although the extensive knowledge on Saccharomyces genome enables the design of more adequate methods to define the genotype of the isolates, the correlation found for M. gruessii is not far from the level of correlation found in these studies (30%), suggesting that correlations between genotypic and phenotypic properties are widespread in yeasts.
Moreover, our results showed that environmental differences were much more relevant for genetic and phenotypic variation than geographic distance among populations. The results of this study are currently being written down in a manuscript that will be submitted to Molecular Ecology.
Fig. 3 Metschnikowia gruessii phenotype vs. genetic relatedness. Clustering of the strains based on AFLP genotype was established using a combination of the Dice coefficient (for generating the distance matrix) and the Unweighted Pair Group Method with Arithmetic Mean (UPGMA) clustering algorithm. Phenotypic maps depict growth for strains in different test conditions, as indicated in the top bar. Font colour in plant species column correspond to the different sites where samples where retrieved.
4. Review on the effect of nectar microbes on plant-pollinator interactions.
Finally, we wrote a thorough review on the effects of nectar microbes on plant-pollinator interactions. More specifically, we reviewed the latest investigations that were performed on nectar chemistry and microbial-mediated changes on nectar chemistry and pollination attraction and compared these results with previous studies. Besides, we reviewed nectar community assembly, and we described the high intraspecific variation in the Metschnikowia genus and its mechanistic base with the data gathered during this project.