The microbiota of avocado floral nectar inhibits pathogens and improves plant fitness

Abstract

Floral nectar-living microbes contribute to flower protection and pollinator health and are primarily determined by nectar chemical composition. Microbial communities in non-hexose-rich nectars and their ability to inhibit pathogens and modulate plant development have been poorly explored. We used metabarcoding to examine the richness and relative abundance of bacteria and fungi from avocado, a globally important crop with a unique nectar chemical composition, whose production is severely affected by diseases and low pollination rates. We also explored the antagonistic activity of the nectar microbial culturable fraction and its volatile organic compounds (VOCs) against avocado pathogens Phytophthora cinnamomi and Colletotrichum gloeosporioides, and against the most devastating honeybee pathogens Ascosphaera apis and Paenibacillus larvae. Furthermore, we experimentally analyzed the effects of microbial isolates and their VOCs on plant growth and the activation of jasmonic acid (JA) defense responses in Arabidopsis thaliana. Pseudomonas, Acinetobacter, Protomyces, and Vishniacozyma were the dominant microbial genera inhabiting avocado nectar. From 43 evaluated isolates, 17 bacteria and three yeasts inhibited the plant and honeybee pathogens, promoted the growth of A. thaliana seedlings, and induced JA signaling. Microbial VOCs emitted by all tested isolates promoted lateral root formation and increased plant biomass. Collectively, our findings highlight the selectivity of avocado nectar over its microbiota, which could directly impact plant fitness and contribute to the health of its pollinators.

The floral nectar in avocado hosts microbes with anti-pathogenic activity and probiotic properties in a model plant species.

Introduction

Pollination in angiosperms is a crucial event for the conservation of terrestrial ecosystems that involves a multipartite ecological interaction between flowers, pollinators, and their associated microorganisms (Vannette, 2020). Nectar-consuming insects shape the composition of microbial communities in flowers, which could be neutral, beneficial, or pathogenic for plants (Mueller et al., 2023). In turn, the floral microbiota alters nectar chemistry and influence pollinator and plant health (Cullen et al., 2021; Martin et al., 2022). Nectar, a substance present in flowers, mainly contains sugars and secondary metabolites, and can be considered a niche where only microorganisms tolerant to osmotic stress and low nitrogen availability can develop (Warren et al., 2020). Nectar microbial communities tend to be species poor and dominated by a few bacteria and yeast taxa (Mueller et al., 2023). Microbial isolates present in nectar can range from plant, pollinator, and even human pathogens, to commensals and attractors of pollinators, and natural enemies of pest insects and phytopathogens (Sasu et al., 2010; Adler et al., 2021; Álvarez-Pérez et al., 2024), positioning nectar as a potential source of beneficial microorganisms for the plant-pollinator interactions and inhibition of plant pathogens. Although some microorganisms isolated from nectar from different flowering crops such as apple, pear, blueberry, and almond trees have been described to inhibit plant pathogens or tolerate different management practices (Smessaert et al., 2019; Rering et al., 2023; Schaeffer et al., 2023), data on entomopathogenic and/or growth-promoting microorganisms from nectar are still scarce.

Persea americana Mill. (avocado) (Lauraceae), an ancient angiosperm native of Mexico and Central America, has gained economic relevance worldwide due to its nutritional value (Rendón-Anaya et al., 2019; Bhore et al., 2021). The market research and consulting firm Grand View Research valued the global avocado market at US$15.83 billion in 2023 (Grand View Research, 2023). However, the agricultural management of this crop is highly input-intensive due to soil impoverishment, incidence of diseases, a lengthy period of vegetative growth required to reach the productive stage, and low pollination rates (Pérez-Jiménez, 2008; Twizeyimana et al., 2013; Santini et al., 2022). Root rot caused by the soil-borne oomycete Phytophthora cinnamomi is the most serious disease, with a decrease in production of between 45% and 90% in affected areas (Pérez-Jiménez, 2008), while anthracnose caused by the opportunistic phytopathogen Colletotrichum gloeosporioides is the main cause of post-harvest losses (Xoca-Orozco et al., 2017). Over the past 5 years, major avocado-producing countries have reported phytosanitary alerts due to C. gloeosporioides incidence in orchards (Ayvar-Serna et al., 2021; Bustamante et al., 2022), affecting up to 80% of production with significant economic losses (Fuentes-Aragón et al., 2020; Bernal et al., 2025). Although anthracnose is a disease that manifests in fruits, infection usually occurs at the flowering stage. As flowers may host bacteria and fungi that inhibit the growth of several Colletotrichum spp. pathogens, as described in blueberries (Rering et al., 2023), the potential of nectar microorganisms to act as biocontrol agents of C. gloeosporioides in avocado should be investigated. On the other hand, previous studies have shown that certain volatile organic compounds (VOCs) produced by microorganisms associated with avocado antagonize the growth of P. cinnamomi (Guevara-Avendaño et al., 2019), which highlights the relevance of investigating the ability of VOCs produced by avocado nectar-living microorganisms to suppress the growth of this oomycete.

The small, greenish avocado flowers produce a peculiar nectar with a high sucrose concentration (79–80%), a small fraction of fructose (1%), and traces of K⁺, stachyose, and perseitol (Liu et al., 1995). Flowers open twice; first in the pistillate female phase and then in the male staminate phase (Bender, 2002; Can-Alonzo et al., 2005). This dichogamous flowering pattern, combined with the low pollen release and highly specific chemical composition of avocado nectar, leads to <1% of the flowers reaching fruit set (Bender, 2002), the remainder falling to the ground and decomposing, probably incorporating their microbiota into the soil. One of the strategies implemented to remedy the low pollination rate in avocado has been the introduction of Apis mellifera, positively impacting the production in commercial orchards (Peña and Carabalí, 2018). Unfortunately, the management of this generalist pollinator has caused the emergence of pathogens such as the bacterium Paenibacillus larvae (Fünfhaus et al., 2018) and the fungus Ascosphaera apis (Chaimanee et al., 2017), leading to bee colony losses (Genersch, 2010; Goulson et al., 2015). In this context, it is important to determine whether P. larvae or A. apis is part of the avocado nectar microbiota or whether nectar microorganisms could antagonize the growth of these bee pathogens.

The high selectivity of floral nectar as a microenvironment and the resulting competitiveness of its microbiota (Warren et al., 2020; Mueller et al., 2023), combined with the close contact between nectar-living microorganisms and plant tissues (Sasu et al., 2010), led us to hypothesize that (i) avocado nectar, with its particular chemical composition, recruits a different microbial community from that described in other plant models; (ii) the nectar microbiota performs probiotic functions in their plant hosts and pollinators as antagonists of pathogens; and (iii) the microbiota possesses phytostimulating properties influencing endogenous plant signaling. Our goal was to find broad-spectrum antagonistic strains, active against both phytopathogens and entomopathogens, capable of activating molecular signaling in host plants. We used metabarcoding sequencing to explore the diversity and composition of the avocado nectar microbial community and a culture-dependent approach to characterize the antagonistic activity of avocado nectar microbiota against the avocado pathogens C. gloeosporioides and P. cinnamomi and the honeybee pathogens A. apis and P. larvae. Moreover, we tested the ability of nectar-inhabiting microorganisms to promote plant development and defense in the model plant Arabidopsis thaliana. Collectively, this information could help to eventually design sustainable strategies to mitigate the impact of diseases that affect avocado production, whilst contributing to understand multipartite interactions between pollinators, plants, and their associated microbiota.

Materials and methods

Study site and sampling

The study site was located at Acuitzio del Canje, state of Michoacán, Mexico (19°30′N; 101°20′W) at 2080 m a.s.l., in the main avocado-producing area worldwide (FAOSTAT, 2016). The neighboring vegetation corresponds to oak forests. Rainfall ranges between 800 mm and 1300 mm year⁻¹, with an average annual temperature of 24 °C, a maximum temperature of 32 °C, and a minimum of 10 °C. Avocado nectar samples were collected with sterile 5 µl capillaries by random sampling of at least 10 female-phase flowers from five different avocado trees without symptoms of any disease in March 2019. Samples were transported to the laboratory for microbiological or molecular processing.

Metabarcoding and bioinformatic analyses

To characterize the nectar microbial communities, we extracted genomic DNA from nectar samples using the DNeasy Blood and Tissue Kit (QIAGEN), according to the manufacturer’s instructions. Briefly, the nectar was collected from 45 female-stage flowers on five healthy trees. Subsequently, nectar from 15 flowers was pooled to construct three libraries for 16S rDNA and internal transcribed spacer 2 (ITS2) metabarcoding.

Library preparation was performed at the Genomics Lab of the National Laboratory of Ecological and Synthesis (LANASE) following Illumina’s procedure with some modifications. A 5 µl aliquot of each nectar pool was used to characterize the bacterial community by amplifying the V4 region of the 16S rDNA gene, using primers 515-YF (Parada et al., 2016) and 806R (Apprill et al., 2015) (Supplementary Table S1). For the fungal community, the same amount of nectar from the three different pools was used to amplify the ITS2 region, with a mixture of four forward primers and the degenerate reverse primer ITS4NGS (Tedersoo et al., 2014) (Supplementary Table S1). All PCRs had a total volume of 50 μl, containing 5 μl of nectar-extracted DNA, 0.2 μM of each primer, and 25 μl of 1× Qiagen Multiplex PCR Master Mix. The PCR cycling parameters were: 95 °C for 15 min; 25 cycles at 94 °C for 30 s, 55 °C for 30 s, 72 °C during 30 s; and a final extension for 5 min at 72 °C. The amplicons were indexed with the Nextera XT kit v2 and purified with the ProNex® Size Selective Purification System, following the Purification PCR amplicon protocol. Each library was analyzed by capillary electrophoresis with a Qiaxcel system to estimate their final size (450 bp for the 16S rDNA gene and from 500 bp to 620 bp for the ITS2 metabarcode, including indexes and adapters). Paired-end sequencing (2×250 bp) was performed on a MiSeq instrument (Illumina, USA) at the Sequencing Unit of the Mexican National Institute of Genomic Medicine (INMEGEN). Each library was sequenced in triplicate. The generated sequences were demultiplexed and the primers removed. Bioinformatics analysis was conducted in QIIME2-2023.7, using DADA2 (Callahan et al., 2016) for trimming and truncating low-quality reads, removing chimeras, and generating amplicon sequence variants (ASVs). Taxonomy assignments were performed with the q2-feature-classifer sklearn with self-trained classifiers (Bokulich et al., 2018). For the 16S trained model, the ‘SILVA 138 at 99%’ database (Robeson et al., 2021) was restricted based on the primers (Werner et al., 2012), and default features were used. The ITS model was trained with full sequences without restriction of the UNITE ‘sh_qiime_release_s_04.04.2024’ database (Abarenkov et al., 2024), and a k-mer length of [6,6] (Bokulich et al., 2018) and a confidence limitation of the 98% assignments were used. Plastid, mitochondrial, and archaeal reads were removed from 16S assignment, and non-assigned reads at the phylum level were removed for both datasets. Data derived from bacterial and fungal sequencing were deposited in the Sequence Read Archive of NCBI under accession PRJNA1236352 (https://www.ncbi.nlm.nih.gov/sra/PRJNA1236352).

Isolation of the nectar culturable microbial fraction

Nectar samples from 50 female-stage flowers on five healthy trees were immediately mixed with 100 µl of 0.9% sodium chloride upon arrival at the laboratory. Suspensions were spread on the surface of Petri dishes with solid R2A medium supplemented with 20% sucrose and incubated at 28 °C. Isolated colonies were re-streaked on Luria–Bertani (LB) solid medium and potato dextrose agar (PDA) supplemented with 100 mg l⁻¹ ampicillin until pure cultures were obtained. The 97 resulting microbial isolates were grouped into 43 morphotypes, according to morphological criteria and Gram staining as described in Guevara-Avendaño et al. (2019). All isolates were preserved at −20 °C in LB or potato dextrose broth supplemented with 20% glycerol.

Antagonism assays against avocado pathogens

Avocado pathogens were the fruit fungal pathogen C. gloeosporioides and the soil-borne oomycete P. cinnamomi. To implement the dual antagonism assays, 10 µl of a standardized suspension (0.5 of absorbance at OD₆₀₀) from each nectar microbial isolate were inoculated at 2.5 cm from a mycelium disc (5 mm diameter) placed in the center of PDA plates. Three isolates were evaluated per plate. As a control, 10 µl of LB medium were inoculated, following the design described in Guevara-Avendaño et al. (2018). To evaluate the antagonistic activity of strains by volatile emission, the two-sealed-base-plates method described in Guevara-Avendaño et al. (2019) was employed. Briefly, the strains were pre-grown for 24 h as described above, but lids were replaced by another base plate containing a disc of 5 mm diameter of fungal or oomycete mycelium on PDA. The plates were sealed on top of each other with Parafilm® and incubated at 28 °C. Three assays were set up only with growth medium as controls. After 5 d of incubation, the fungal growth inhibition percentage was determined by measuring the mycelial diameter, with the formula described in Guevara-Avendaño et al. (2018). Each isolate was tested in triplicate. To assess the effect of the antagonistic isolates on the hyphal morphology, 10 representative mycelium samples of each fungal/oomycete phytopathogen grown under control conditions and in the presence of the two isolates with the highest antifungal activity were collected, processed, and analyzed by SEM in a JEOL JSM-IT300 microscope according to Méndez-Bravo et al. (2018).

The in vivo antifungal activity of the two most active isolates against C. gloeosporioides was evaluated by implementing wound and drip assays in anthracnose-free avocado fruits as described by Trinidad-Ángel et al. (2017) with modifications. Twenty ripe and healthy Hass fruits were washed and disinfected with 1% sodium hypochlorite for 5 min and rinsed three times with sterile distilled water, followed by immersion in 70% ethanol, then air-dried in a laminar flow hood. After disinfection, four fruits were designated as the control; the other 16 were wounded on two opposite sides in the middle part with a cork borer to generate cavities of 5 mm diameter and 3 mm depth. Wounded fruits were divided into groups of four for the different treatments: (i) 20 µl of sterile water on both sides (mock, negative control); (ii) 20 µl of a pathogen spore solution at a concentration of 1×10⁵ spores ml⁻¹ applied on both sides (positive control); (iii) 20 µl of bacterial suspension at a concentration of 1×10⁶ CFU ml⁻¹on both sides (biocontrol agent); and (iv) 20 µl of bacterial suspension plus 20 µl of the pathogen spore solution on both sides (pathogen+biocontrol agent). All fruits were placed in plastic containers and incubated for 10 d in a plant growth chamber (Lumistell ICP-55, Mexico) with a photoperiod of 16 h of light, 8 h of darkness, light intensity of 200 μmol m² s⁻¹. After the incubation period, fruits were cut in half and photographed to determine the size of the lesions caused by C. gloeosporioides using the ImageJ2/Fiji program (Wayne Rasband National Institutes of Health, USA).

Molecular identification of nectar isolates

Nectar isolates with ≥20% inhibition against C. gloeosporioides and P. cinnamomi were cultured in liquid medium for 48 h at room temperature and then centrifuged at 3500 g for 15 min. The pellets were resuspended in phosphate-buffered saline (PBS), adjusting the biomass to ∼100 mg FW, and DNA was extracted using the Dneasy Blood and Tissue kit (Qiagen, The Netherlands). The genomic DNA concentration was adjusted to 25 ng µl⁻¹. To identify bacterial isolates, primers 27F and 1492R were used to amplify the 16S rRNA gene, while for yeasts, primers ITS-1 and ITS-4 were used to amplify the ITS region (Supplementary Table S1). Amplifications were performed in a SureCycler 8800 (Agilent Tech., CA, USA), using 50 µl of reaction mix containing 25 ng of template DNA, 200 µM of each dNTP, 1.25 mM of MgCl₂, 1× Taq buffer, 0.4 µM of each primer, and 0.5 U of Taq DNA polymerase (Qiagen, Germany). For 16S rRNA amplification, the PCR conditions werte as follows: initial step at 95 °C for 4.5 min; 40 cycles at 95 °C for 1 min, 53 °C for 1 min, and at 72 °C for 2 min; with a final step at 72 °C during 5 min. For amplification of the ITS region, the PCR cycling parameters were 95 °C for 3 min, followed by 35 cycles at 95 °C during 45 s, 56.2 °C for 35 s, 72 °C for 45 s, and a final single cycle at 72 °C for 5 min. Amplicons were purified using the Wizard® SV Gel and PCR Clean-Up System (Promega, USA) and adjusted to 20 ng µl⁻¹ for sequencing on a Hitachi 3500 Genetic Analyzer (Applied Biosystems) at ENES-UNAM Morelia. The obtained sequences were deposited in GenBank (accession nos PV366450–PV366466 and PV366592–PV366594). Sequences were manually checked in BioEdit 7.2.5. and aligned using MEGA 7, utilizing the multiple alignment program MUSCLE. The edited sequences and their best matches in the GenBank nucleotide database (www.ncbi.nlm.nih.gov) were used to construct the alignment.

Antagonism assays against bee pathogens

The strains of A. apis and P. larvae were acquired at the Mexican National Collection of Microbial Strains and Cell Cultures from the Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional (CINVESTAV).

Antagonism assays against A. apis were carried out as described above for the avocado pathogens, using maltose yeast extract (MYM) plates as medium instead of PDA. To assess the antifungal activity of the isolates against P. larvae, 200 µl of a standardized suspension (0.5 absorbance at OD₆₀₀) of P. larvae were spread on the surface of Petri dishes with Brain Heart Infusion (BHI) agar. Four 5 mm paper discs moistened with 10 µl of standardized suspensions of three nectar microbial isolates, and 10 µl of LB medium as a control, were placed at the cardinal points of each plate. After 5 d of incubation at 28 °C, the halo of inhibition was measured. Each combination of isolates was tested in triplicate. The growth on MYM and BHI agar medium of each isolate was previously corroborated.

Effect of nectar isolates on plant growth and development

Arabidopsis thaliana (L., Heynh) ecotype Columbia 0 (Col-0) and reporter line JAZ1/TIFY10A::GFP:uidA (Grunewald et al., 2009) were used to evaluate the effect of the selected isolates on plant growth and jasmonic acid (JA) response. Arabidopsis seeds were surface-disinfected with 96% ethanol for 5 min and 20% sodium hypochlorite for 7 min. After five washes with sterile distilled water, seeds were stored at 4 °C during 48 h for stratification. Growth promotion induced by volatiles was tested in vitro sowing five seeds per plate, and eight seeds were sown in undivided Petri dishes to evaluate the effect of the isolates in dual culture assays. The plates, which contained solid 0.2× MS medium (Murashige and Skoog basal salts mixture, Cat. M5524; PhytoTechnology), were placed vertically into a plant growth chamber (Lumistell ICP-55, Mexico) with a photoperiod of 16 h of light, 8 h of darkness, light intensity of 200 μmol m² s⁻¹, and temperature of 22 °C.

Four days after germination, seedlings in undivided plates were inoculated with 100 µl (10⁸ CFU ml⁻¹) of microbial suspension, placed at a distance of 2.5 cm from root tips, to assess the plant growth-promoting effect of nectar isolates. To evaluate the effect of volatiles emitted by nectar isolates, 100 µl (10⁸ CFU ml⁻¹) of bacterial suspension were inoculated in the plant-less compartment of the divided Petri plates. Control plants were grown without microbial inoculant. All treatments were performed in triplicate (three different dishes) in three independent experiments (three replicates). Root system architecture (primary root length and lateral root number), biomass accumulation, and JAZ1 expression were analyzed at 7 d after inoculation. The histochemical analysis of transgenic seedlings to evaluate the expression of JAZ1-GUS was performed with the β-glucuronidase Reporter Gene Staining Kit (Sigma-Aldrich), according to the manufacturer’s instructions. Then, plant tissue was clarified according to the protocol of Malamy and Benfey (1997), and seedlings were placed on glass slides, covered with coverslips, and sealed with commercial nail varnish to analyze the JAZ1 expression with an optical microscope (Zeiss AXIO Zoom V16 with an Axiocam 305 color camera, Germany).

Statistical analyses

A Shapiro–Wilk test was used to test whether data fitted a normal distribution. A multiple one-way ANOVA with a Tukey’s post-hoc test was conducted using STATISTICA v.10 software to determine statistical differences between treatments, using a P-value ≤0.05. Additionally, the treatments in antagonism assays were compared with the control using a post-hoc Dunnett test.

Results

Microbial community composition in the avocado floral nectar

We employed a metabarcoding approach to explore the complexity of the nectar microbial community to assess the bacterial and fungal richness and abundance. A total of 634 bacterial ASVs and 249 fungal ASVs, with a minimal size of 234 bp after filtering with DADA2, were recorded. The rarefaction analysis revealed that the abundance of ASVs reached a plateau, indicating that the saturation of sequencing depth for the samples was reached, except for the triplicate 3A in the 16S sequencing. The dominant bacterial order was Pseudomonadales, represented by the Acinetobacter and Pseudomonas genera, reaching a relative abundance ranging from 40% to 94%, followed by Massilia and Polaromonas associated with Burkholderiales, and by Sphingomonas (Fig. 1A; Supplementary Fig. S1). The most abundant fungal ASVs were Protomyces (23–89%), the phyllosphere-inhabiting Vishniacozyma (0.6–27%), and Filobasidium (1.5–19.5%) (Fig. 1B; Supplementary Fig. S2). The low diversity and large dominance recorded for the bacterial and fungal communities from nectar samples provide evidence of the high selectivity that avocado nectar exerts over the microbial community.

Fig. 1.

Structure of the microbial community associated with the avocado floral nectar. Relative abundance of (A) bacterial and (B) fungal genera obtained by metabarcoding sequencing. A total of 634 bacterial ASVs with a 99% confidence limitation and 243 fungal ASVs with a 98% confidence limitation were assigned.

Strains isolated from avocado floral nectar differentially inhibit the growth of C. gloeosporioides and P. cinnamomi

To explore the potential function of bacteria and fungi recorded by the metabarcode, we obtained a culturable fraction isolated from the avocado floral nectar. A total of 50 flowers from five avocado trees were sampled in the morning immediately after anthesis, avoiding the arrival of nectar-consuming insects. From the nectar of these 50 flowers, 97 isolates were obtained, which were classified into 43 morphotypes (Supplementary Table S2); 51% of them were Gram-positive bacteria, 39% Gram-negative bacteria, and 9% yeasts. Randomly selected representative strains of each morphotype were used to perform antagonism assays against C. gloeosporioides and P. cinnamomi by volatile emission or dual culture assays. Two isolates (C-2.5 and H-19.2) showed significant inhibition of the mycelial growth of C. gloeosporioides through volatile emission (Fig. 2A), while 17 isolates inhibited radial growth in dual culture assays (Fig. 2A). Isolates I-59.2, B-1.8, and I-70 showed the highest antagonistic activity in dual culture assays and, although their inhibitory effect through the emission of volatiles was not significant, B-1.8 and I-70 caused a lower mycelial growth density by indirect interaction (Fig. 2B, C). These two isolates were then selected to evaluate their effect on hyphae morphology and to perform an avocado fruit rot inhibition assay. SEM micrographs of C. gloeosporioides mycelium grown in the vicinity of B-1.8 showed that this strain decreased the thickness of hyphae, causing their collapse and abnormal branching, while I-70 caused hyphal collapse (Fig. 2D). The confrontation of B-1.8 and I-70 against C. gloeosporioides by wound and drip in the exocarp of avocado fruits demonstrated that these isolates reduced the damage caused by anthracnose by ∼50% and 94%, respectively (Fig. 2E; Supplementary Fig. S3).

Fig. 2.

Antagonistic effect of selected microbial isolates from avocado floral nectar against Colletotrichum gloeosporioides. Antifungal activity of volatile and diffusible compounds produced by 43 representative morphotypes was evaluated under in vitro conditions, and the most active of them were tested in avocado fruit rot inhibition assays. (A) Percentage of inhibition of mycelial growth by emission of volatile and diffusible compounds. (B) Representative images of the mycelial growth inhibition of plugs of C. gloeosporioides exposed to volatiles emitted by the most active microbial isolates. (C) Images of mycelium in dual culture inhibition assays with the most active isolates. (D) SEM images of the morphological alterations in the hyphae structure induced by the isolates. (E) Representative images of the avocado fruit rot inhibition assays. The values represent the mean ±SE (n=3) and asterisks indicate a significant difference compared with control (Dunnett test, P≤0.05).

In the inhibition tests against P. cinnamomi, six isolates inhibited mycelium growth by the emission of volatile compounds and 25 of them showed a percentage of inhibition >20% by direct contact (Fig. 3A). The volatile compounds from five isolated strains, again including the one identified as I-70, reduced the aerial mycelium density of the oomycete. Although strain B-1.8 did not inhibit this phytopathogen through volatile compounds on this occasion, it was the most effective of 17 bioactive isolates against P. cinnamomi in dual culture assays (Fig. 3A, B). SEM micrographs showed a drastic reduction in hyphal thickness caused by B-1.8, whilst I.70 induced hyphal thinning and distortion. Both isolates caused shriveling of the hyphal surface (Fig. 3D).

Fig. 3.

Inhibition of Phytophthora cinnamomi growth by strains isolated from avocado floral nectar. (A) Inhibition percentage of mycelial radial growth by emission of volatile and diffusible compounds of the microbial strains. (B) Representative images of inhibition assays by volatile emission or (C) in dual culture assays with the most active strains. (D) SEM images of morphological alterations induced by the isolates with the highest percentage of inhibition. The values shown represent the mean ±SE (n=3) and asterisks indicate significant differences with respect to control (Dunnett test, P≤0.05).

Molecular identification of avocado nectar strains antagonistic to C. gloeosporioides and P. cinnamomi

The 20 isolates (17 bacteria and 3 yeasts) that were active against the phytopathogens were tentatively identified by 16S rRNA and rDNA–ITS1–ITS4 gene sequencing. The closest sequence matches based on the BLAST similarity analysis are shown in Table 1. The most represented bacterial genus was Pseudomonas with 10 isolates that showed 99–100% identity, being closely related to P. flavescens (A-1.9 and A-2.7), P. aylmerensis (A-2.3, B-1.8 and F-2.5), P. protegens (B-1.9), P. cerasi (B-1.10 and B-2.5), and P. viridiflava (B-1.1 and G-2.1). Seven of the remaining isolates were closely related to Paenibacillus favisporus (B-2.1), Erwinia aphidicola (B-2.8), Curtobacterium ammoniigenes (C-1.3), Nocardioides endophyticus (I-57), Klebsiella quasipneumoniae (G-1.5), Streptomyces eurythermus (I-70), and Dietzia timorensis (I-59.2). The best match for the ITS1 and ITS4 sequences classified the three yeast isolates into the Filobasidium (C-1.4 and G-2.3) and Aureobasiduim (D-1.5) genera (Table 1). The taxonomic identity of the active culturable fraction matches the dominance of the genera recorded with the analysis of the microbial communities.

Table 1.

Molecular identification of selected isolates from avocado nectar

ID	GenBank accession number	NCBI best match (accession number)	Identity (%)
A-1.9	PV366466	Pseudomonas flavescens (NR_114195.1)	98.02
A-2.3	PV366450	Pseudomonas aylmerensis (NR_169460.1)	100
A-2.7	PV366465	Pseudomonas flavescens (NR_114195.1)	96.02
B-1.8	PV366451	Pseudomonas aylmerensis (NR_169460.1)	100
B-1.9	PV366461	Pseudomonas protegens (NR_114749.1)	100
B-1.10	PV366457	Pseudomonas cerasi (NR_146827.1)	100
B-1.11	PV366463	Pseudomonas viridiflava (NR_114482.1)	100
B-2.1	PV366455	Paenibacillus favisporus (NR_029071.1)	99.7
B-2.5	PV366459	Pseudomonas cerasi (NR_146827.1)	100
B-2.8	PV366453	Erwinia aphidicola (NR_117000.1)	100
C-1.3	PV366452	Curtobacterium ammoniigenes (NR_041495.1)	100
C-1.4	PV366593	Filobasidium chernovii (NR_073223.1)	98.9
D-1.5	PV366594	Aureobasiduim melanogenum (NR_159598.1)	97
F-2.5	PV366454	Pseudomonas aylmerensis (NR_169460.1)	100
G-1.5	PV366458	Klebsiella quasipneumoniae (NR_132596.1)	100
G-2.1	PV366455	Pseudomonas viridiflava (NR_114482.1)	99.7
G-2.3	PV366592	Filobasidium chernovii (NR_073223.1)	98.9
I-59.2	PV366462	Dietzia timorensis (NR_112775.1)	100
I-57	PV366460	Nocardiodes endophyticus (NR_135731.1)	100
I-70	PV366464	Streptomyces eurythermus (NR_025869.2)	99.6

Effects of the phytopathogen-antagonistic strains against Apis mellifera pathogens

With the aim of exploring how broad the spectrum of the nectar strains to inhibit pathogens would be, not only against phytopathogens but also against entomopathogens, the 20 identified isolates were screened for their potential antagonistic activity against the fungus A. apis and the bacterium P. larvae. Five bacterial isolates and one yeast (C-1.4) showed moderate but significant activity against A. apis (P≤0.05) by volatile emission; notably, 13 of the tested isolates displayed antifungal activity by direct contact in the dual culture assays (Fig. 4A). From these 13 active strains, Pseudomonas spp. B-2.5 and B-1.8, Streptomyces sp. I-70, Paenibacillus sp. B-2.1, and Erwinia sp. B-2.8 inhibited the pathogen mycelial growth by >20% (Fig. 4A–C). In particular, the isolates I-70, B-2.1, and B-1.8 caused hyphal wrinkling and shriveling on A. apis (Fig. 4D).

Fig. 4.

Effect of the identified isolates on the growth of honeybee pathogens. (A) Percentage of inhibition of Ascosphaera apis mycelial growth by the emission of volatile and diffusible compounds by the selected microbial strains. (B) Representative images of mycelial growth inhibition of A. apis exposed to the volatiles released by the most active strains. (C) Dual culture inhibition assays of A. apis with the most active strains. (D) SEM images of the morphological alterations in the hyphal structure of A. apis induced by the selected strains. (E) Antibacterial activity of the selected strains co-cultured with Paenibacillus larvae. (F) Images of the inhibition assays after 5 d. The values shown represent the mean ±SE (n=3) and the asterisk indicates a significant difference relative to control (Dunnett test, P≤0.05).

Ten of the nectar isolates co-cultured with P. larvae generated growth inhibition halos ranging from 1 mm to 5 mm. Pseudomonas spp. A-2.7, B-1.8, and A-2.3 showed the highest inhibitory activity, followed by the yeast Aureobasidium sp. D-1.5 (Fig. 4E, F).

Plant growth-promoting properties of avocado nectar isolates

As a first approach to explore the potential capacity of the 20 bioactive isolates to establish multipartite interactions, beyond pathogen inhibition, we performed in vitro interaction assays with A. thaliana to evaluate the microbial effect on plant development and the activation of the induced systemic response. Remarkably, the volatiles emitted by all the microbial isolates increased the primary root length, promoted lateral root organogenesis and growth, and stimulated biomass accumulation of A. thaliana seedlings (Fig. 5A–D). When the isolates were co-cultured with the seedlings by direct contact, most of them modified the root architecture; isolates Pseudomonas spp. A-1.9, A-2.3, and A-2.7, Nocardioides sp. I-57, and Streptomyces sp. I-59.2 stimulated root branching or biomass accumulation without any inhibitory effect on primary root length (Fig. 6A–D), while Pseudomonas spp. B-1.8 and F-2.5, Erwinia sp. B-2.8, Curtobacterium sp. C-1.3, and Klebsiella sp. G-1.5 decreased the primary root length but stimulated lateral root formation and increased the FW accumulation (Fig. 6A–D). These results indicate that the avocado nectar microorganisms possess plant growth-promoting activity and could differentially impact plant signaling pathways.

Fig. 5.

Effects of volatile compounds produced by the isolates from avocado nectar on the root system architecture and biomass accumulation in Arabidopsis. Four-day-old seedlings were exposed for 7 d to the volatiles released by the selected isolates to determine their effect on plant development. (A) Representative images of Arabidopsis seedlings grown in compartmentalized plates in interaction with volatiles produced by the isolates. (B) Quantitative analysis of primary root length, (C) lateral root number, and (D) total FW of seedlings grown on the surface of agar 0.2× MS-containing plates. Values shown represent the mean ±SE (n=45 seedlings) and asterisks indicate significant differences between treatments (Tukey test, P≤0.05).

Fig. 6.

Effects of the co-cultivation of the microbial isolates with Arabidopsis seedlings on root growth and development and biomass accumulation. Four-day-old seedlings were transferred to 0.2× MS medium inoculated with the identified isolates to analyze root architecture and biomass accumulation 7 d after interaction. (A) Representative images of Arabidopsis seedlings interacting with the identified isolates. (B) Quantitative analysis of primary root length, (C) lateral root number, and (D) total FW. Values shown represent the mean ±SE (n=72 seedlings) and asterisks indicate statistically significant differences (Tukey test, P≤0.05).

Induction of defense responses in A. thaliana by the plant growth-promoting isolates

JA is one of the main phytohormones orchestrating the plant defense response against biotic stressors with developmental programs (Raya-González et al., 2012; Carvalhais et al., 2017). Sensing of plant growth-promoting (PGP) microorganisms may result in triggering JA-mediated induced systemic resistance (ISR) and the recruitment of beneficial microbes through JA signaling (Carvalhais et al., 2017). Therefore, we analyzed the expression of the JA-responsive Jasmonate-zim domain protein 1 (JAZ1) gene as an indicator of pathway activation in JAZ1/TIFY10A::GFP:uidA A. thaliana transgenic seedlings, directly or indirectly interacting with the 20 strains selected for their activity against necrotrophic phytopathogens and PGP properties. Histochemical GUS activity assays in the co-cultured seedlings indicated that Pseudomonas spp. A-1.9, A-2.3, B-1.8, B-1.11, and F-2.5, Paenibacillus sp. B-2.1, and Streptomyces I-70 induced an exacerbated gene expression in shoot tissues and in the apical region of primary roots, from the xylem to the external cell layers, which coincides with developmental shifts such as root hair differentiation, shortness of the meristematic zone, and thickening of roots (Fig. 7). The bacterial isolates Pseudomonas spp. A-2.3, A-2.7, and B-2.5 and Curtobacterium sp. C-1.3, and the yeasts Filobasidium spp. C-1.4 and G-2.3, and Aerobasidium sp. D-1.5 induced the expression of the JA-responsive gene only in the shoot system (Fig. 7), probably indicating the activation of an ISR response. The isolates Erwinia sp. B-2.8, Klebsiella sp. G-1.5, Dietzia sp. I-57, Nocardioides sp. I-59.2, and Streptomyces I-70 induced the systemic activation of the JA-responsive pathway with a moderated root expression and shoot-induced signal of the gene marker (Fig. 7), which could be differentially related to a PGP or to pathogenic activity. The volatiles emitted by Pseudomonas spp. A-1.9, A-2.3, A-2.7, B-1.8, B-1.10, B-2.5, and G-2.1, Paenibacillus sp. B-2.1, Curtobacterium sp. C-1.3, Klebsiella sp. G-1.5, and Dietzia sp. I-59.2 increased JAZ1 gene expression (Supplementary Fig. S3), indicating a possible role as airborne bacterial signals in interactions with plants.

Fig. 7.

Effects of selected isolates on defense-related jasmonic acid gene expression in Arabidopsis. Four-day-old transgenic seedlings expressing the jasmonic acid-inducible JAZ1/TIFY10A::GFP:uidA gene marker were transferred to 0.2× MS medium inoculated with the different isolates to analyze activation of the JA signaling pathway 7 d after the interaction (at least 10 seedlings per treatment were analyzed).

Discussion

In angiosperms, floral structures constitute the first interphase in plant–pollinator interactions, where nectar represents the main nutritional reward (Vannette, 2020). Floral nectar harbors specialized microbiota that could impact both plant and pollinator health (Vannette et al., 2013; Martin et al., 2022; Álvarez-Pérez et al., 2024). However, beneficial traits of the nectar microbiota to antagonize plant and pollinator pathogens and promote plant fitness at early developmental stages have been scarcely explored. Here, we reported for the first time the composition of the nectar-inhabiting bacterial and fungal communities, and some functions of their culturable fraction, in a crop with agronomical, ecological, and evolutionary importance, the avocado P. americana Mill.

Ubiquitous microbes reach nectar via wind, rain, and animal transport. However, only a few taxa are able to survive UV radiation, the presence of antimicrobial compounds, low nitrogen availability, and osmotic stress that characterize nectar as a highly selective ecological niche (Pozo et al., 2011). A low microbial diversity in nectar community composition suggests that dominant taxa could be particularly important for plant health (Good et al., 2014; Jacquemyn et al., 2021). In hexose-rich nectars, specialist yeasts of the genus Metschnikowia predominate, followed by the generalist Aureobasidium and Cryptococcus, while the most common bacteria belong to the Enterobacteriaceae family and to the genera Acinetobacter and Bacillus (Pozo et al., 2011; Vannette, 2020; Mueller et al., 2023; Álvarez-Pérez et al., 2024). Here, the high relative abundance of Acinetobacter, Pseudomonas, and Vishniacozyma in avocado nectar was consistent with that reported in other nectars, but Bacillus, Metschnikowia, and Cryptococcus were not included among the most abundant taxa in the analyzed samples, while this was the first report of Protomyces as the dominant taxon in nectar (Fig. 1). These compositional differences in nectar microbial communities are probably due to the chemical composition of avocado nectar and the species that pollinate its flowers (de Vega et al., 2009). Perseitol content in avocado nectar could be inhibiting the germination of Bacillus spores (Rizk et al., 1989), while having a nectar rich in sucrose with low fructose and glucose content is not optimal for the development of an alcohol-fermenting yeast incapable of hydrolyzing sucrose, such as Metschnikowia (Vicente et al., 2020; Torrellas et al., 2023). The case of Protomyces, a little-researched phytopathogen in floral nectar with the ability to assimilate sucrose and which causes hyperplasia and formation of galls in the veins of leaves, petioles, peduncles, stems, and fruits (Čadež et al., 2021; Leharwan et al., 2021), is interesting, since there are no reports of this genus as a causal agent of diseases in avocado (González-Santarosa et al., 2014), suggesting that avocado nectar could act as a reservoir for this phytopathogen. Finally, even the most abundant genera are likely to benefit from each other, since not all can use sucrose as a carbon source. For example, Acinetobacter spp. and Filobasidium spp. are both non-saccharolytic. Whilst the bacteria have n-alkanes, the yeast degrades higher fatty acid esters and terpenoids to promote the aroma and flavor in fruits (Wang et al., 2023; Wu et al., 2023). They both probably thrive on avocado nectar at the expense of n-alkanes, VOCs, and polyphenols that are synthesized to mediate interactions with insects and other microbes and provide protection against UV radiation, as reported in other plant species (Bartwal et al., 2013). On the other hand, Aureobasidium spp. and Pseudomonas spp. can assimilate a wide range of compounds, but the latter prefer organic acids and amino acids as a carbon source, rather than sugars (Rojo, 2010; Chi et al., 2022). Pseudomonas spp. could therefore benefit from the tricarboxylic acids produced by other microorganisms that inhabit the nectar (Anderson and Guerra, 1985).

Our data indicated that 20 out of 43 tested isolates were active against C. gloeosporioides, P. cinnamomi, and entomopathogens in dual culture, whilst their emitted volatiles promoted plant development (Figs 2–4). This reinforces the theory that the few microbial taxa present in nectar play important roles in plant fitness (Good et al., 2014) and demonstrates a high potential of nectar for as a source of beneficial microorganisms for agriculture. Recently, the antagonistic activity against Colletotrichum pathogens of several nectar microbes, isolated from blueberries, was described in in vitro assays, showing that the pathogen inhibition did not affect nectar consumption by honeybees (Rering et al., 2023). Here, most of tested isolates displayed a greater pathogen inhibition through direct contact than through volatile emission, which suggests that the inhibition of mycelial growth observed in dual culture assays could be due to the production of diffusible compounds and cell wall-degrading enzymes. Pseudomonas spp., Paenibacillus spp., Klebsiella spp., and Aureobasidium spp. produce phenazines, exophilins, lyamocins, and lytic enzymes such as proteases and chitinases that damage the fungal cell wall (Merín et al., 2014; Raza et al., 2015; Di Francesco et al., 2017; Jisha et al., 2018; Hsu et al., 2021; Zhang et al., 2022; Iqbal et al., 2023). This is supported by the microscopy images showing damage of the mycelium of the different pathogens when exposed to nectar isolates. Future analysis of the metabolomic and volatile profiles from our antagonistic microbial isolates, exposed to different artificial nectar formulations, should shed some light on these antifungal chemical compounds.

Molecular identification of the selected nectar isolates largely coincided with the genera that reached the highest relative abundances in the metabarcoding analysis (Fig. 1; Table 1). Pseudomonas was one of the most dominant genera of the community, with a relative abundance of 32.8%, and 50% of the bioactive microbial isolates belonging to this genus. Similarly, Erwinia, Klebsiella, and Filobasidium were among the taxa with the highest relative abundance in the nectar microbial community, and were also retrieved in the culturable fraction. Exceptions included Paenibacillus sp. B-2.1, Curtobacterium sp. C-1.3, Dietzia I-57, Nocardioides I-59.2, and Streptomyces I-70, which were not detected among the most abundant taxa in avocado nectar, and Acinetobacter, the bacterial genus with the highest relative abundance in the community, although none of the 20 isolates selected for molecular identification corresponded to this taxon. This could be because Acinetobacter spp. from avocado nectar do not present antagonistic activity against C. gloeosporioides and P. cinnamomi, which was our main selection criterion. However, since antagonistic activity of Acinetobacter spp. strains has been reported against several Colletotrichum and Phytophthora species, this warrants future efforts to retrieve Acinetobacter isolates and investigate their potential bioactivity (Syed-Ab-Rahman et al., 2018; Patel et al., 2019; Matos et al., 2024).

Regarding the effect of nectar microorganisms on plant fitness, microbial volatile and direct contact interactions with our model plant showed that volatile emission by all the evaluated strains promoted root branching and biomass accumulation, while the accumulation of some diffusible microbial metabolites led to activation of the JA defense pathway (Figs 5–7). Volatile and diffusible compounds from Pseudomonas spp., Klebsiella spp., Dietzia spp., Streptomyces spp., and Aureobasidium spp. have been explored for their role in promoting plant development. Benzyl alcohol, 2-phenylethyl alcohol, 3-(methyl-thio)-1-propanol, trimethylamine, 3-hydroxy-2-butanone (acetoin), 2,3-butanediol, and 2-pentylfuranare are some examples of VOCs produced by these genera that promote nutrient availability and root development (Camarena-Pozos et al., 2019; Jones et al., 2019; Ravelo-Ortega et al., 2023), and should be further studied for their effects on avocado development.

In this context, the fact that all tested isolates emitted VOCs with plant growth-promoting activity (Fig. 5), despite belonging to different taxa, is noteworthy. However, different species related to the isolates described here have been shown to produce widely known plant growth-promoting volatiles; for example, Paenibacillus polymyxa emits acetoin and 2,3-butanediol, the best-known growth-promoting VOCs (Ryu et al., 2003; Lee et al., 2012). Pseudomonas spp., Erwinia amylovora, and Curtobacterium flaccumfacien also produce 2,3-butanediol (Cho et al., 2008; Santoro et al., 2011; Groenhagen et al., 2013; Parmagnani et al., 2023; Gaudioso et al., 2025), while trimethylamine produced by Streptomyces venezuelae has recently been shown to increase the rate of photosynthesis, thus contributing to improve plant health and growth (Li et al., 2025). On the other hand, yeasts such as Filobasidium and Aerobasidium often produce dimethyl disulfide, which promotes root system development (Groenhagen et al., 2013; Meldau et al., 2013). Although the VOCs released by Dietzia spp. and Nocardioides spp. are less explored, they produce ketones, alkanes, lipids, alcohols, and sulfides, compounds similar to those mentioned for their activity as promoters of plant development (Biçer et al., 2021; Widada et al., 2021). Since several reports provide evidence of how a single VOC, even at low concentrations, can modulate plant growth through different hormone-related pathways (Garnica-Vergara et al., 2016; Raya-González et al., 2017), further studies are needed to determine the molecular mechanisms underlying the plant growth promotion by avocado nectar isolates.

The activation of JA signaling by nectar microbial isolates is of particular interest, since it could, in addition to coordinating the defense against fungi, drive nectar secretion and the plant response against herbivores and their mechanical damage (Heil et al., 2012; Millán-Cañongo et al., 2014; Wang et al., 2021; López-García et al., 2023). These JA-inducing properties are especially striking in terms of ecological interactions, since some pathogenic bacteria can move into the vascular system of the host plants through the floral nectaries, for example Erwinia tracheiphila, the causative agent of wilt disease in cucumber (Sasu et al., 2010). A previous report described how the strawberry rhizosphere strain Streptomyces globisporus SP6C4 is able to translocate endophytically from roots to flowers and back to the roots, and to antagonize the gray mold pathogen Botrytis cinerea whilst reducing honeybee mortality caused by P. larvae and Serratia marcescens (Kim et al., 2019). These results support the idea that the rhizosphere, the anthosphere, and their associated microbiota impact the plant ecosphere, including the health of pollinators.

It is important to note that avocado nectar communities may also possess other functional attributes that are not explored here. According to several reports, some isolates belonging to the genera Dietzia and Nocardioides could protect plants against salt stress, heavy metals, and other air pollutants (Bharti et al., 2016; Ma et al., 2023; Shahid et al., 2024). Some species related to our identified isolates have traditionally been reported as phytopathogens, such as Curtobacterium flaccumfaciens (López-Ramírez et al., 2022; Tokmakova et al., 2024) and nectar specialists Erwinia (Pozo et al., 2011; Jacquemyn et al., 2013; Vannette, 2020). However, our data indicated that volatile and diffusible compounds of Erwinia sp. B-2.8 and Curtobacterium sp. C-1.3 promoted root system development in A. thaliana seedlings. Further genomic insights into the most promising plant growth-promoting strains will allow us to gain a deeper understating of the mechanisms underpinning their bioactivity.

Collectively, our results show that avocado nectar harbors a competitive specialized microbiota with bioactivity either as a phytostimulant, mostly through the emission of volatile compounds, or as contributors to plant health, by antagonizing pathogens and activating the JA defense pathway. Our findings suggest that the nectar microbiota could directly impact plant fitness and contribute to the health of its pollinators. Future work will focus on exploring the ecological significance of these nectar microbes in the avocado–pollinator interactions, by evaluating their attractiveness to floral visitors and their ability to disperse in the environment and colonize different ecological niches.

Supplementary data

The following supplementary data are available at JXB online.

Table S1. List of primers used to create the 16S and ITS2 gene libraries to characterize the nectar microbial communities

Table S2. Classification of isolates into morphotypes according to their growth in LB medium and their Gram staining.

Fig. S1. Heatmap showing the most abundant bacterial ASVs.

Fig. S2. Heatmap showing the abundance of the fungal ASVs.

Fig. S3. Quantitative analysis of anthracnose lesions in avocado fruits.

Author contributions

CML-G, IAR-G, YP-B, LAV-E, MMT, VP-C, LER-G, MG-M, OH-C, JL-M, and AM-B: methodology; CML-G, IAR-G, and AM-B: formal analysis; CML-G, IAR-G, YP-B, LAV-E, MMT, VP-C, LER-G, MG-M, OH-C, and JL-M: investigation; CML-G, IAR-G, and AM-B: writing—original draft; CML-G, AIR-G, MQ-A, FR, and KO: writing—review & editing; YP-B, LAV-E, MMT, VP-C, LER-G, MG-M, OH-C, and JL-M: data curation; MQ-A, FR, KO, and AM-B: funding acquisition; AM-B: conceptualization, resources, visualization, and supervision.

Funding

This research was funded by the Mexican Ministry of Science, Humanities, Technology, and Innovation (SECIHTI) grant no. CB-2014, 242999 and by the SECIHTI-PRONACES-PEE grant (project ‘PERSEA’, 322772). CML-G thanks the Dirección General de Asuntos del Personal Académico (DGAPA, UNAM) for her post-doc fellowship grant, and AM-B thanks SECIHTI (CVU 50123) for the fellowship grant (854328) received during his sabbatical leave.

Data availability

The strain sequences were deposited in GenBank (accession nos PV366450–PV366466 and PV366592–PV366594). Data derived from bacterial and fungal sequencing were deposited in the Sequence Read Archive of NCBI under accession PRJNA1236352 (https://www.ncbi.nlm.nih.gov/sra/PRJNA1236352).