Demethylation Inhibitor Fungicides Have a Significantly Detrimental Impact on Population Growth and Composition of Nectar Microbial Communities

Abstract

Demethylation inhibitor (DMI) fungicides are a mainstay of modern agriculture due to their widespread use for crop protection against plant-pathogenic fungi. However, DMI residues can disperse and persist in the environment, potentially affecting non-target fungi. Previous research has demonstrated that DMIs and other fungicides inhibit yeast growth in floral nectar microbial communities and decrease fungal richness and diversity of exposed flowers with no apparent effect on bacteria. Nevertheless, the effect of DMIs on the population growth of different species of nectar inhabitants and the dynamics of these microbial communities remains understudied. To address these issues, in this study we created synthetic microbial communities including yeasts (Metschnikowia reukaufii and Metschnikowia pulcherrima) and bacteria (Rosenbergiella epipactidis and Comamonas sp.) and propagated them in culture media containing different DMIs (imazalil, propiconazole, and prothioconazole) at different doses or no fungicide. Our results showed that DMIs have a significant impact on some of the most common microbial inhabitants of floral nectar by favoring the growth of bacteria over yeasts. Furthermore, habitat generalists such as M. pulcherrima and Comamonas sp. were more impacted by the presence of fungicides than the nectar specialists M. reukaufii and R. epipactidis, especially upon dispersal across habitat patches. Future research should determine if the patterns observed in the present study hold true for other species of nectar microbes and explore the interaction between growth limitation due to fungicide presence, dispersal limitation, and other mechanisms involved in community assembly in floral nectar.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00248-024-02477-x.

Introduction

Fungi are ubiquitous in nature and key determinants of terrestrial ecosystem functioning [1]. In contrast to the global ecological importance of beneficial fungi, fungal phytopathogens stand out because of the diseases they cause in crop production systems, leading to a widespread and increasing use of fungicides for the protection of food resources [2, 3]. Commercial fungicides belong to several classes and include 14α-demethylation inhibitors (DMIs), which have been increasingly used in agriculture since the 1990s and currently have about a 16% share of the global fungicide volume market [4]. Although DMIs are chemically diverse, all these compounds inhibit fungal growth by interacting with the C14-demethylase (ERG11/Cyp51) involved in a key demethylation step within fungal sterol biosynthesis [4, 5]. Typically, DMIs have a broad spectrum of activity against a range of economically important fungal phytopathogens [4, 5]. Additionally, clinical DMIs represent the largest and most widely used class of antifungal drugs in human and animal medicine [6]. As a result of these applications, significant amounts of DMI residues can disperse and persist in the environment, potentially affecting non-target fungi and other (micro)organisms, what might alter community composition and have detrimental consequences for ecosystem functioning [7–10].

The floral nectar of angiosperms is the natural habitat of diverse microorganisms, mostly yeasts and bacteria, that can alter nectar’s chemical properties and have different effects on plant–pollinator interactions [11–14]. The growth of microorganisms in floral nectar depends on their capacity to use the nutrients of this plant secretion and tolerate challenging conditions, including high osmotic pressures, scarcity of nitrogen sources, and diverse toxins of plant origin [15–19]. Moreover, flowers are ephemeral, island-like habitats in which rapid growth is needed for population persistence [20, 21].

Previous research has demonstrated that agricultural and medical fungicides inhibit the growth of nectar yeasts [22, 23]. Additionally, spraying of fungicides on wild and crop plants significantly decreased fungal richness and diversity of exposed flowers but had no apparent effect on their bacterial communities [23, 24]. Orchard management practices including the frequent application of fungicides seem to have similar detrimental effects on the fungal communities of flowers [25]. However, these previous studies on the effects of fungicides on nectar- and flower-inhabiting microbes focused on culture-based or culture-independent methods that do not assess microbial growth and/or analyze the effects on a single microbial species at a time, thus precluding a deeper understanding of how fungicides affect the population growth of different microbial species and the dynamics of these communities. Using in vitro competition experiments including nectar specialist (i.e., species that are mostly found in nectar and flower visitors and show tolerance to the growth limiting factors of nectar [12, 26]) and generalist (i.e., species widely present in other habitats that are less tolerant to the growth-limiting factors of nectar) yeasts and bacteria, in this study we tested the following hypotheses: (i) that DMI fungicides significantly impact nectar microbial populations by favoring the growth of bacteria over yeasts; (ii) that generalist species will be more impacted by the presence of fungicides than nectar specialists; and (iii) that the effects of DMI fungicides on population growth and species composition of nectar microbial communities will be amplified when these communities are dispersed across habitat patches.

Materials and Methods

Isolates

The in vitro competition experiments carried out in this study were performed using two yeast species, namely Metschnikowia reukaufii and Metschnikowia pulcherrima, and two bacterial species, Rosenbergiella epipactidis and Comamonas sp. Both M. reukaufii and R. epipactidis are regarded as nectar specialists that are commonly found in the floral nectar of diverse plants worldwide and seem to be highly adapted to survive in nectar [20, 27–30]. In contrast, M. pulcherrima and Comamonas sp. represent generalist species that are less commonly found in floral nectar but thrive in other floral and non-floral habitats [30–33]. The four isolates used in this study were been obtained in May 2024 from the floral nectar of Asphodelus ramosus (Asphodelaceae), Echium sp. (Boraginaceae), and Lavandula stoechas (Lamiaceae) plants collected in Robledondo (Madrid, Spain), and they are easy to differentiate from each other according to their morphology and other phenotypic traits (Table 1) (see details of the microbiological procedures in the Supplementary Methods). Permission to carry out the field work to obtain the plant samples yielding these isolates was granted by Consejería de Medio Ambiente, Agricultura e Interior, Comunidad de Madrid (permit no. 447/2024). Species-level identification of isolates was achieved by Sanger sequencing the D1 and D2 domains of the 26S rRNA gene (for yeasts) or the 16S rRNA gene (for bacteria), as explained in the Supplementary Methods. All isolates were stored at − 80 °C as 25% v/v glycerol (Sigma-Aldrich, Madrid, Spain) stocks until being used in the experiments detailed below.

Table 1.

Overview of the isolates used in this study

Microbial type	Isolate	Species	Origin	Plant host	Sampling location (GPS coordinates)	Genebank accession no
Yeast	SQC01	Metschnikowia reukaufii	Floral nectar	Asphodelus ramosus	Robledondo, Madrid, Spain (40.581907, − 4.211349)	PQ399657
	SQC03	Metschnikowia pulcherrima	Floral nectar	Asphodelus ramosus	Robledondo, Madrid, Spain (40.581907, − 4.211349)	PQ399658
Bacterium	SQC02	Rosenbergiella epipactidis	Floral nectar	Echium sp.	Robledondo, Madrid, Spain (40.586305, − 4.214537)	PQ399667
	SQC06	Comamonas sp.	Floral nectar	Lavandula stoechas	Robledondo, Madrid, Spain (40.583881, − 4.212011)	PQ399668

Inoculum Preparation

The starting microbial inocula used in the competition experiments (see below) were created by growing each isolate from the frozen glycerol stocks on yeast malt agar (YMA; Sigma-Aldrich) or trypticase soy agar (TSA; Sigma-Aldrich), for yeast or bacteria, respectively, and preparing cell suspensions from plate cultures of these on 50% v/v glycerol. Viable cell counts (measured as colony forming units (CFU) per mL) were determined by preparing dilutions of the glycerol stocks (typically 10⁻¹, 10⁻², 10⁻³, and 10⁻⁴) in saline solution (0.85% w/v NaCl; Sigma-Aldrich), which were then spread by onto YMA or TSA plates and incubated at 25 °C for 72 h before proceeding to CFU counting. Multiple aliquots of each glycerol stock were stored at − 80 °C until being used in the competition experiments (see below). The results of previous studies have demonstrated that short-term (< 1 month in the present study) storage of glycerol stocks of different species of nectar yeasts and bacteria (including Metschnikowia spp. and Rosenbergiella spp.) has no detrimental effect on their viability (Álvarez-Pérez et al., unpublished results; see also Refs. [34, 35]).

Competition Experiments

The in vitro competition experiments designed in this study consisted of a set of 0.2-mL PCR tubes containing growth media (sucrose + amino acids solutions) with different doses of DMIs or without any fungicide. Briefly, on the day when the propagation experiments were initiated (day 0), the glycerol stocks corresponding to the two yeasts (SQC01 and SQC03; Table 1) and the two bacterial isolates (SQC02 and SQC06) were thawed and combined to yield 1 mL of cell suspension containing ~ 10⁴ cells/μL of each species. Then, 5-µL aliquots of this four-species synthetic microbial community were inoculated into PCR tubes containing 95 µL of sucrose + amino acids solution (15% w/v of sucrose (Sigma-Aldrich) and 0.632 mL of a 1 g/mL stock of casamino acids (Gibco, Detroit, MI, USA) in 100 mL of PCR-grade water (Sigma-Aldrich)) with different concentrations of imazalil (IZL; 0.5, 0.125, 0.031, and 0.008 µg/mL), propiconazole (PPZ; 0.064, 0.016, 0.004, and 0.001 µg/mL), and prothioconazole (PTZ; 8, 2, 0.5, and 0.125 µg/mL) (Sigma-Aldrich). Fungicide test concentrations were selected according to the minimum inhibitory concentrations determined in previous studies for Metschnikowia yeasts [22, 23]. All fungicides were diluted in dimethyl sulfoxide (DMSO; Sigma-Aldrich), and further diluted (1/100) to test concentrations in sucrose + amino acids solution. Positive and negative controls containing no fungicide but a 1/100 dilution of DMSO were inoculated with a 5-µL aliquot of the synthetic microbial community and 5 µL of 50% w/v glycerol, respectively. All tubes were covered with a breathable membrane (Breathe-Easy; Diversified Biotech, Boston, MA, USA) to allow gas exchange and incubated with no agitation for 24 h at 25 °C. On the next day (day 1), the content of the tubes was mixed up by pipetting up and down several times, and 5-µL aliquots of the evolved cultures were transferred into a new PCR containing 95 µL of sucrose + amino acids solution with the same fungicide (IZL, PPZ, PTZ, or control) at the same concentration as on day 0. This propagation of the synthetic communities mimics the transfer of nectar microbial communities from flower to flower carried out by pollinators [36]. Inoculated tubes were incubated under the same conditions as in the previous step and the remaining cell suspensions were diluted as explained above and plated by duplicate on YMA supplemented with 0.1 g/L of chloramphenicol (Sigma-Aldrich) or TSA supplemented with 0.1 g/L of cycloheximide (Sigma-Aldrich) to obtain CFU counts of yeasts and bacteria, respectively. This operation was repeated two more times (days 2 and 3 of the propagation experiment), to simulate the average flower lifespan of non-winter plant species in Mediterranean ecosystems [37]. Four biological replicates of each test condition starting from different initial communities were run on different weeks.

Data Analysis

All statistical analyses were conducted in R v.4.4.1 [38] using Rstudio v. 2024.09.0 + 375 [39]. The normal distribution of CFU density values was assessed by the Shapiro–Wilk test, using the shapiro.test() function of R package “stats” [38]. Correlation between CFU densities, expressed as log₁₀(CFU + 1), across treatments and propagation steps of different microbial groups and species was analyzed by non-parametric Spearman’s rank correlation tests, using the R package “Hmisc” v.5.1–3 [40].

The effect of different factors—namely, microbial type (yeasts vs. bacteria, nectar specialists vs. habitat generalists, or species-level identification), treatment (one of the 12 DMI-containing media or control nectar without fungicides), and propagation step (1, 2, or 3 days)– and their interactions on log₁₀(CFU + 1) values was analyzed by repeated-measures factorial analysis of variance (ANOVA) of aligned rank transformed (ART) data, as implemented in the R package ARTool v.0.11.1 [41, 42]. Moreover, the adequacy of different experimental conditions (combination of fungicide treatment, dose, and propagation step) to support the growth of the four species included in the synthetic microbial communities was assessed by nonmetric multidimensional scaling (NMDS) of CFU densities using the metaMDS() function of “vegan” v.2.6–8 [43] and considering Euclidean distances, two dimensions, a maximum of 250 random starts in search of stable solution, and a maximum of 999 iterations.

When necessary, P-values were adjusted for the number of simultaneous comparisons using Bonferroni’s correction. In all cases, P-values lower than 0.05 were considered statistically significant. Different functions of “ggpubr” v.0.6.0 [44], “ggordiplots” v.0.4.3 [45], and “ggplot2” v.3.5.1. [46] were used for data visualization.

Results

CFU density values obtained in the competition experiments for all strains considered independently, for the combination of yeast strains (i.e., M. reukaufii + M. pulcherrima) and of bacterial strains (R. epipactidis + Comamonas sp.), and for nectar specialist microbes (M. reukaufii + R. epipactidis) vs. habitat generalists (M. pulcherrima + Comamonas sp.) departed from a normal distribution (P-values < 0.001, in all cases). A significantly negative correlation was found between the log₁₀(CFU + 1) values obtained across all treatments and propagation steps of the competition experiment for yeasts and bacteria (Spearman’s ρ =  − 0.695, P < 0.001) and for nectar specialists and habitat generalists (ρ =  − 0.558, P < 0.001; Fig. S1). Similarly, negative correlation between log₁₀(CFU + 1) results was found for all pairwise comparisons of R. epipactidis with the other species tested, i.e., M. reukaufii (ρ =  − 0.639, P < 0.001), M. pulcherrima (ρ =  − 0.568, P < 0.001), and Comamonas sp. (ρ =  − 0.515, P < 0.001) (Fig. S2). In contrast, significantly positive correlation between CFU density values was obtained for all pairwise comparisons of M. reukaufii, M. pulcherrima, and Comamonas sp. (Fig. S2).

In general, higher CFU density values were obtained in all fungicide treatments for nectar specialists than for habitat generalists, and in most treatments for bacteria than for yeasts (Figs. S3 and S4); the only exception was the lowest concentration of imazalil (i.e., 0.008 µg/mL) on day 1, when the average ± standard deviation (S.D.) log₁₀(CFU + 1) values for yeasts and bacteria were 5.94 ± 0.17 and 5.79 ± 0.21, respectively. Such differences in CFU densities of specialists vs. generalists and bacteria vs. yeasts were amplified on days 2 and 3 of the propagation experiment, particularly at the highest imazalil and propiconazole concentrations (0.5 and 0.125 µg/mL, and 0.064 and 0.016 µg/mL, respectively; Figs. S3 and S4). Bacteria also displayed higher log₁₀(CFU + 1) values than yeasts in the positive control on days 2 and 3 (6.31 ± 0.22 vs. 5.76 ± 0.26 and 6.54 ± 0.05 vs. 5.29 ± 0.54, respectively), but not on day 1 (5.92 ± 0.33 vs. 5.98 ± 0.36). In contrast, nectar specialists reached higher log₁₀(CFU + 1) than habitat generalists in the positive control throughout all the experiment (6.30 ± 0.31 vs. 3.89 ± 0.47, respectively, on day 1; 6.46 ± 0.14 vs. 2.80 ± 0.52 on day 2; and 6.58 ± 0.08 vs. 1.54 ± 0.90 on day 3).

At the species level, the highest CFU densities were obtained in most treatments for R. epipactidis (Fig. 2); again, the only exception was the lowest concentration of imazalil on day 1, when the log₁₀(CFU + 1) was slightly higher for M. reukaufii than for R. epipactidis (5.94 ± 0.17 vs. 5.78 ± 0.21, respectively). In contrast, the log₁₀(CFU + 1) obtained for M. pulcherrima and Comamonas sp. were in all cases much lower than those obtained for R. epipactidis, especially on days 2 and 3 (Fig. 1). Moreover, at all prothioconazole concentrations and the lowest concentrations of imazalil and propiconazole (0.031 and 0.008 µg/mL, and 0.001 µg/mL, respectively), M. reukaufii also outperformed M. pulcherrima and Comamonas sp. on day 1, whereas the highest concentrations of these two latter DMI fungicides had a clear negative impact on the CFU counts of M. reukaufii. The detrimental effect of fungicides on the population growth of M. reukaufii (only at the highest concentrations of imazalil and propiconazole), and M. pulcherrima and Comamonas sp. (in all fungicide treatments) were amplified in subsequent propagation steps of the community (i.e., days 2 and 3; Fig. 1). Similarly, the two nectar specialists (i.e., M. reukaufii and R. epipactidis) displayed higher CFU densities than the two habitat generalists (i.e., M. pulcherrima and Comamonas sp.) on day 1, an effect that was also amplified in subsequent propagation steps.

Fig. 2.

Nonmetric multidimensional scaling (NMDS) analysis of the adequacy of different experimental conditions to support the growth of the microorganisms included in the synthetic communities generated in this study. The growth gradients of Metschnikowia reukaufii (A), Metschnikowia pulcherrima (B), Rosenbergiella epipactidis (C), and Comamonas sp. (D), expressed as log₁₀(CFU + 1) values, were fitted into the ordination diagrams using the gg_ordisurf() function of the ‘ggordiplots’ v.0.4.3 [45]. In all panels, the shape of the points denotes the fungicide compound (IZL, imazalil; PPZ, propiconazole; PTHZ, prothioconazole) or the growth control (CRTL), whereas their color indicates the fungicide concentration (from highest (1) to lowest (3), see details in Materials and Methods and the legend of Fig. 1), and the size corresponds to the day after the onset of the experiment

Fig. 1.

Boxplots showing the results, expressed as log₁₀(CFU + 1) values, obtained in the in vitro competition experiments testing for the impact of different fungicide treatments on the growth of yeasts (Metschnikowia reukaufii and Metschnikowia pulcherrima) and bacteria (Rosenbergiella epipactidis and Comamonas sp.). Panels show the results obtained on different days (1, 2, and 3) after the onset of the experiment for different treatments: imazalil (A), propiconazole (B), prothioconazole (C), and growth control containing no fungicide (D)

ART-based ANOVA confirmed that CFU density significantly depended on the type of microorganism (yeast vs. bacteria, nectar specialist vs. habitat generalist, and identity at the species level), the fungicide treatment (i.e., combination of DMI compound and dose), the propagation step, and all the two-way and three-way interactions between these factors (Table 2). Partial eta squared (η_p²) values were in all cases ≥ 0.092, indicating that most independent variables and their combinations had a large effect on the response variable [47].

Table 2.

Results of the non-parametric factorial analysis of variance (ANOVA) using aligned rank transformed (ART) data of the growth of nectar microorganisms in presence of different demethylation inhibitor (DMI) fungicides

Factors	Analysis by microbial typesa					Bacteria vs. yeastsa					Specialists vs. generalistsa
	df	df res	F	P	ηp2	df	df res	F	P	ηp2	df	df res	F	P	ηp2
Microbial species/groups (M)b	3	465	986.36	< 0.001*	0.864	1	231	739.33	< 0.001*	0.762	1	231	754.14	< 0.001*	0.766
Fungicide treatment (F)c	12	465	27.80	< 0.001*	0.418	12	231	24.45	< 0.001*	0.560	12	231	4.62	< 0.001*	0.193
Propagation step (P)d	2	465	437.82	< 0.001*	0.653	2	231	45.00	< 0.001*	0.280	2	231	285.15	< 0.001*	0.711
M × F	36	465	17.90	< 0.001*	0.581	12	231	62.46	< 0.001*	0.764	12	231	6.25	< 0.001*	0.244
M × P	6	465	131.20	< 0.001*	0.629	2	231	156.81	< 0.001*	0.576	2	231	486.41	< 0.001*	0.808
F × P	24	465	1.97	0.004*	0.092	24	231	3.80	< 0.001*	0.283	24	231	1.59	0.044*	0.142
M × F × P	72	465	1.35	0.039*	0.172	24	231	3.43	< 0.001*	0.263	24	231	2.07	0.003*	0.177

^aART ANOVA results: df, degrees of freedom; df res, degrees of freedom of residuals; F, F statistics; P, P-values (statistically significant values are denoted by an asterisk); η_p², partial eta squared

^bMicrobial species: Metschnikowia reukaufii, Metschnikowia pulcherrima, Rosenbergiella epipactidis, or Comamonas sp. (see details in Table 1). Microbial groups: yeasts vs. bacteria, or nectar specialists vs. habitat generalists

^cImazalil (0.5, 0.125, 0.031, or 0.008 µg/mL), propiconazole (0.064, 0.016, 0.004, or 0.001 µg/mL), prothioconazole (8, 2, 0.5, and 0.125 µg/mL), or growth control (no fungicide)

^dFirst, second, or third propagation step of the competition experiment (i.e., 24, 48, or 72 h, respectively)

The stress values of the NMDS ordination based on log₁₀(CFU + 1) data of the four species included in the competition experiments was 0.066, meaning that the plot with reduced dimensions provided a good representation of the data set (see stress plot in Fig. S5). Graphical representation of the NMDS ordination of the different experimental conditions with contour lines representing growth values confirmed the importance of the fungicide treatment (compound and dose) and propagation step (days after the onset of the experiment) in explaining growth variation of the different microbial species included in the competition experiments (Fig. 2). However, contour lines representing the growth of M. reukaufii, M. pulcherrima, and Comamonas sp. were closer to each other and covered broader log₁₀(CFU + 1) ranges than the contour lines obtained for R. epipactidis, thus indicating that experimental conditions had a lower impact on the growth of R. epipactidis than on the other species. Additionally, the contour plots confirmed that the effect of DMIs on the growth of the different microbial populations and community composition (i.e., increased log₁₀(CFU + 1) values for R. epipactidis and decreased values for the other species) was amplified on days 2 and 3 of the competition experiment (Fig. 2). Finally, the different direction of the growth gradient observed for yeasts (increasing log₁₀(CFU + 1) values from the lower right to the upper left corners of the NMDS plot for M. reukaufii and from the upper right to the lower left for M. pulcherrima), highlighted the higher susceptibility of M. reukaufii than M. pulcherrima to the presence of fungicides, especially to the highest concentrations of imazalil and propiconazole.

Discussion

DMI residues can persist in soils, water bodies, and other habitats even for several months or years after their use in agriculture and other applications [48–50]. Accordingly, the characterization and mitigation of the possible pernicious effects of DMIs for non-target fungi and other organisms is becoming a cause of concern for the scientific community, policy makers, and other stakeholders demanding more sustainable agricultural practices [4, 51, 52].

Given the frequent use of DMIs in crop protection against fungal diseases, nectar microbial communities might be often exposed to these compounds, especially in locations that are close to agricultural fields [22–25]. Furthermore, nectar yeasts and bacteria are transferred from flower to flower by insects and other floral visitors over the flowering season and may even overwinter in the gastrointestinal tract of the pollinators that are also frequently exposed to azoles [53, 54]. Analyzing the impact of DMIs on the relative growth of different microbial species might be a first step to understanding the actual threat that these chemicals pose to wild nectar microbial communities.

Using in vitro competition experiments including nectar specialist yeast and bacteria and habitat generalists, in this study we have demonstrated that the detrimental effects of DMI on microbial growth significantly depend on the identity of the microorganisms, the fungicide compound and dose, the exposure time, and the interactions between all these factors. Furthermore, as predicted, DMI were particularly pernicious for yeasts, so that nectar bacteria had a competitive advantage over Metschnikowia yeasts to reach higher population growth and dominate in the community. However, R. epipactidis was able to outcompete Comamonas sp. both in presence and in absence of DMIs, which agrees with their consideration as nectar specialist and habitat generalist, respectively. In contrast, M. reukaufii reached lower CFU densities than M. pulcherrima at the highest concentrations of imazalil and propiconazole, thus indicating that exposure to DMIs can give a chance to habitat generalists displaying lower susceptibility to these fungicides to avoid the competition for nectar resources and persist in this habitat.

Nectar microbial metapopulations and metacommunities consist of thousands of flowers representing ephemeral, island-like habitats that are linked by pollinators and other floral visitors acting as dispersal agents for microorganisms [21, 55, 56], which in the present study was mimicked by pipetting an aliquot of the evolved communities into new PCR microtubes (i.e., artificial flowers) containing fresh medium with or without fungicides. The persistence of individual microbial species in this system depends, among other factors, on both a high population growth and a high number of dispersed cells across habitat patches [21]. Higher growth rates and dispersal ability (at least when associated to some floral visitors, such as thrips [57]) for nectar bacteria than for yeasts might confer a dispersal advantage to bacteria in this system, especially in the presence of DMIs and other fungicides inhibiting yeast growth. Our observation that, at least in some treatments, yeast growth was critically reduced after two propagation steps of the competition experiment and, accordingly, community composition was significantly enriched in bacteria seems to confirm this latter hypothesis.

A limitation of the present study is that, for the sake of simplification, we used a single strain (i.e., genotype) of each bacterial and yeast species in the competition experiments. However, both M. reukaufii and M. pulcherrima show intra-specific variability in their susceptibility to DMI fungicides [22, 23], which might be relevant for population dynamics and yeast-bacterium interactions. Furthermore, the four microbial species tested were introduced to synthetic communities at similar cell densities and at the same time. Nevertheless, it is well known that priority effects (i.e., phenomena in which the effects of species on one another depend on their arrival order into a local site and initial abundance [58]) have a relevant role in the assembly of nectar microbial communities [59–62]. In particular, Chappell et al. [59] demonstrated that the nectar specialist bacterium Acinetobacter nectaris exerts a strongly negative priority effect against M. reukaufii by reducing nectar pH. In return, M. reukaufii populations might rapidly develop resistance against such negative priority effect if constantly exposed to bacterial-induced reduction of nectar pH [59]. DMIs and other fungicides might alter the dynamics of priority effects in floral nectar (e.g. by acting synergistically with pH reduction by bacteria, or any other mechanisms that have not been identified yet, to inhibit yeast growth) but, to our knowledge, such a possibility remains to be studied. Finally, our study design did not account for possible differences between the four species tested in their lag time and/or growth rate in the sucrose + amino acids solution used as growth medium, and which might account, at least in part, for the observed dominance of nectar specialists over habitat generalists.

Conclusions

In conclusion, the results of this study confirm that DMI fungicides have a significant impact on some of the most common microbial inhabitants of floral nectar by favoring the growth of bacteria over yeasts. Moreover, we have demonstrated that, in general, habitat generalists such as M. pulcherrima and Comamonas sp. are more impacted by the presence of fungicides than the nectar specialists M. reukaufii and R. epipactidis, especially upon dispersal across habitat patches. Future research should determine if the patterns observed in the present study hold true for other species of nectar microbes and explore the interaction between growth limitation due to fungicide presence, dispersal across habitat patches, and other mechanisms determining community assembly in this habitat, including competition for nutrient resources, niche modification, and priority effects.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contribution

Conceptualization and resources: SQ-C and SA-P.

Investigation, formal analysis, and data curation: AR, SQ-C, and SA-P.

Writing—original draft preparation: SQ-C and SA-P.

Writing—review and editing: all authors.

Supervision: SA-P.

Funding: SA-P.

Funding

This work was supported by grants RYC2018-023847-I and CNS2022-135237 from MCIN/AEI/ 10.13039/501100011033 and “ESF Investing in your future” awarded to SA-P. SQ-C acknowledges grant PIPF-2023/ECO-29442 from Consejería de Educación, Ciencia y Universidades, Comunidad de Madrid.

The funders had no role in the preparation of the manuscript or decision to publish.

Data Availability

The nucleotide sequences determined in this work have been deposited in the GenBank/ENA/DDBJ databases (accession numbers: PQ399657, PQ399658, PQ399667, and PQ399668). All relevant results are included in this paper or available as supplementary materials at the journal’s website.

Declarations

Ethics Approval

Ethics approval was not required. All yeast and bacterial isolates used in this study were collected following local regulations and the specifications of the Nagoya Protocol on Access and Benefit Sharing (ABS).

Consent to Participate

Not applicable.

Consent for Publication

Not applicable.

Competing Interests

The authors declare no competing interests.