Argan fruit microbiomes: influence of biogeographic and soil driven environmental factors

Abstract

Background

Argan tree (Sideroxylon spinosum L.), also known as Argania spinosa, is an endemic woodland species in Morocco. Its fruit has multiple uses for food, cosmetics, and traditional medicine. Despite their importance, argan fruit microbiomes remain unexplored. In this study, we investigated the diversity and biogeographical patterns of the microbial communities associated with argan fruits.

Results

Bacterial and fungal communities were characterized from 36 argan trees across four locations, Agadir, Essaouira, Rabat and Berkane, along a 1,000 km transect, with altitudes ranging from 66 to 794 m above sea level. Our finding revealed significant variation in both fresh (p = 0.0108) and dry (p = 0.0191) fruit weights across locations, suggesting a geographic influence on fruit biomass. Alpha diversity analysis showed significantly higher bacterial richness in Agadir (p < 0.05), while fungal diversity did not differ significantly across sites (p > 0.05). Pantoea was the dominant bacterial genus (up to 89.0%), and Hanseniaspora was the dominant fungal genus (up to 82.4%). We also found correlations between microbial genera and environmental factors, such as altitude, soil pH, nutrient levels and calcium carbonate content. Pantoea abundance was positively correlated with clay content (p = 0.045). Alternaria abundance was positively associated with soil pH (p = 0.0092) and negatively associated with electric conductivity (p = 0.0026) and nitrate concentration (p = 0.039).

Conclusion

No significant differences in microbial community composition were detected among locations. This study reveals distinct biogeographic patterns in the Argan fruit-associated microbiome, shaped by soil properties and geographic location. Core taxa, including Pantoea and Hanseniaspora, were consistently dominant, alongside site-specific microbial signatures. These findings provide a foundation for microbiome-informed strategies to support Argan tree conservation and sustainable cultivation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08168-8.

Introduction

Argan (Sideroxylon spinosum L.) is an endemic, xerothermophile tree species belonging to the Sapotaceae family within the order Ericales and is considered a keystone species in Morocco due to its ecological, cultural and economic importance [1, 2]. It is the sole representative of the monotypic genus Argania. Phylogenetically, its closest relatives belong to the genus Sideroxylon, which is primarily distributed in tropical and subtropical regions, particularly in Africa and Central America. S. spinosum is regarded as a relic species from the Tertiary period. Its recently completed genome assembly, spanning approximately 655 Mb and consisting of 11 fully resolved telomere-to-telomere chromosomes, has provided valuable insights into its adaptation to extreme environments. Notably, the genome reveals an expansion of gene families associated with cuticle biosynthesis, drought tolerance, and lipid metabolism. The genomic features highlight the species’ evolutionary divergence from other members of Sapotaceae and support its unique status and ecological significance in North Africa [1].

Currently, argan trees are predominantly found in southwestern Morocco, particularly in the greater Agadir and Essaouira regions, where over 3,000,000 hectares of forests were declared a UNESCO biosphere reserve in 1998 [3, 4]. In addition to these core areas, isolated argan populations have also been reported in Arganate (in the Grou and Cherrat River valleys, south of Rabat) and in the Beni-Snassen hills near Berkane [1, 5–7]. The argan tree thrives in harsh arid and semi-arid environments, tolerating annual rainfall as low as 150–400 mm, temperatures reaching up to 45 °C, and elevations up to 1,500 m above sea level [4, 8]. Owing to its cultural and ecological significance, argan tree has been internationally recognized by UNESCO in 2014 and FAO in 2018 [9, 10].

Argan fruits produce a highly valued oil, which is in high demand both locally and globally, particularly in the food, cosmetics, and pharmaceutical industries [3, 11]. However, the sustainability of argan ecosystems is increasingly threatened by overgrazing, overexploitation, and habitat degradation, despite considerable research efforts [1]. Like all plants, different parts of the argan tree, fruits, roots, shoots, and leaves, host diverse microbial communities shaped by complex interactions between the plant and its environment [12–14]. Among these, the fruit associated microbiome plays a crucial role in plant health, post-harvest quality, resistance against pathogen attacks, and resilience to environmental stress. For example, microbiomes associated with citrus fruits enhance pathogen resistance and influence flavor [15, 16]; in grapevines, fruit-associated microbiota impact fruit quality and wine fermentation [17]; and in apples, the fruit microbiome affects post-harvest disease susceptibility [18]. Fruits and their seeds provide distinct microhabitats that support specialized assemblages [19]. Recent studies have begun to uncover the microbial composition of fruits and seeds [20–23], as well as the role of environmental factors such as soil proprieties and biogeography in shaping fruit-associated microbiomes [20, 23–26]. However, it remains unknown whether similar factors influence the microbial communities associated with argan fruits.

Despite the ecological and economic importance of S. spinosum, the microbial communities associated with its fruits remain largely unexplored. To date, no study has assessed how environmental factors, such as soil physicochemical properties and geographical variation, shape the diversity and structure of the argan fruit microbiome. Addressing this knowledge gap, our study aimed to characterize the bacterial and fungal communities associated with argan fruits across diverse agroecological zones in Morocco, and to evaluate the influence of environmental gradients on these microbial assemblages. To achieve this, we employed metabarcoding approach targeting the 16 S rRNA gene for bacteria and ITS regions for fungi. Samples were collected from 36 trees across four agroecologically and environmentally distinct locations, covering a broad range of altitudes, soil types, and climatic conditions. We hypothesized that: We hypothesized that soil physicochemical properties would play a greater role than geographical location in shaping the diversity and composition of the argan fruit microbiome. These factors are known to influence microbial activity and plant-microbe interactions, and Moroccan soils vary widely along these gradients. We further expected that environmental variables, rather than spatial distance alone, would better explain microbial community shifts across sites. This study provides the first comprehensive insight into the argan fruit microbiome, offering a crucial baseline for understanding plant–microbe interactions and informing conservation strategies for this ecologically and culturally significant tree.

Materials and methods

Study areas and sampling

Samples were collected between May and June 2024 from four naturally growing populations of argan trees located in distinct regions of Morocco [6, 27, 28]. The plant material used in this study was identified by one of the co-authors of the manuscript, Professor Mohamed El Mderssa (Ph.D., Ecology of Natural Resources and Environment, Sultan Moulay Sliman University, Beni Mellal, Morocco). A voucher specimen was not deposited, as the collection took place during the fructification stage, which provided sufficient diagnostic features for unambiguous identification. High-resolution photographic documentation of the specimens in their natural habitat is available. The sampling areas differ in longitude, latitude, altitude (ranging from 66 to 792 m above sea level), and proximity of the Atlantic Ocean (8–115 km) (Table S1). The sampling sites were distributed across four locations representing the main distribution range of the argan tree: Agadir and Taroudant Provinces (3 sites), Essaouira Province (4 sites), Arganat (Oued Grou) (1 site), and Douar Mahjouba in Berkane Province (1 site) (Fig. 1). A total of nine sites were selected to capture the environmental heterogeneity of argan habitats. The location of Agadir, Taroudant and Essaouira province representing the core and most extensive portion of the species’ natural range, multiple sites were selected to encompass variation in altitude, distance from the coast, and soil characteristics. In contrast, only one site was sampled per argan population in the northern regions due to geographic isolation and limited accessibility (Table S1). The sampling map was created using QGIS version 3.42.1 (Free and Open-Source GIS software) [29].

Fig. 1.

Geographical distribution and sampling sites of Argan trees across four Moroccan regions. A Map of Morocco showing the locations of the four sampling regions (Location 1 to 4). B-H Satellite images of the nine distinct sampling sites across these regions: (B-D) AGADIR, (E-F) Essaouira, (G) Arganat (ouad grou) in Rabat, and (H) Douar Mahjouba in Berkane. Red markers indicate individual Argan trees sampled

Fruits collection and processing

A total of 36 Argan trees were sampled across nine sites. At each site, four trees were randomly selected as biological replicates (Fig. S1). For each tree, 10 fruits at the same stages of maturity were randomly collected, placed in plastic Ziploc bags and stored in a cool box filled with dry ice. Samples were transported to the laboratory and stored at -20 °C until further processing (Fig. S2).

Fruits were rinsed with sterile distilled water, dried with sterile paper towels, weighed, height and width were measured, and then dissected. The pulp and kernels of four fruits from each tree were pooled and ground in liquid nitrogen. Ground samples were immediately used for DNA extraction. Additionally, five fruits per tree were weighed fresh and then oven-dried after at 72 °C until constant weight confirmed by repeated weighing at 24-hour intervals to determine moisture content and dry weight.

Soil physicochemical analyses

From each site, a composite soil sample was collected from five random points (0–20 cm depth) and pooled into plastic Ziploc bags. Samples were transported under cool conditions and stored at 4 °C. Each composite sample was split into two portions: one was used for microbial DNA extraction, and other was sent to the Agricultural innovation and technology transfer center (AITTC) (Benguerir, Morocco) for physicochemical analyses. Parameters measured included: pH, electrical conductivity (EC), soil texture (clay, silt, sand), total nitrogen (TN), total phosphorus (TP), total potassium (TK), organic matter (OM), total calcium carbonate (TCaCO₃), nitrate (NO₃⁻) ammonium (NH₄⁺), chloride (Cl⁻), exchangeable oxides (K₂O, Na₂O, CaO, MgO), and trace elements (Cu, Mn, Fe, Zn, B). Analytical methods used are detailed in Table S2.

DNA extraction and PCR amplification

Approximately 100 mg of freeze-dried fruit tissues was used for DNA extraction using the DNeasy Plant Pro kit (Qiagen, Germany), following manufacturer’s protocol with a final elution volume was 50 µl. DNA concentration was measured using a BioSpectrometer and µCuvette G1.0. (Eppendorf BioSpectrometer^® fluorescence, Eppendorf, Hamburg, Germany), yielding concentrations ranging from 17.0 to 57.4 ng/µl. DNA extracts were stored at − 20 °C until use. To minimize amplification of chloroplast DNA [30], bacterial 16 S rRNA genes were amplified targeting the V5 − V6 regions using the primers with Custom Sequence (CS) adapters: CS1 − 799 F (ACACTGACGACATGGTTCTACA-AACMGGATTAGATACCCKG) and CS2 − 1115R (TACGGTAGCAGAGACTTGGTCT-AGGGTTGCGCTCGTTG). For fungi, the ITS2 region was amplified using primers with CS adapters: ITS3 (TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAHCGATGAAGAACRYAG), and ITS4 (GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGTCCTCCGCTTATTGATAT GC). PCR reactions were performed in a final volume of 25 µL containing 12.5 µL Platinum™ Direct PCR Universal Master Mix (ThermoFisher, Temara, Morocco). The thermal cycling conditions were initial denaturation of 95 °C for 3 min, followed by 34 cycles of 95 °C for 30 s, 55 °C for 30 s and 72 °C for 1 min, with a final extention at 72 °C for 5 min. PCR amplification was run on a Mastercycler X50s (Eppendorf, Genrmany). Amplicons were verified on 1% agarose gels visualized with iBright FL1500 Imaging System (ThermoFisher, Temara, Morocco).

Library preparation and sequencing

Amplicon from the 16 S rRNA gene and ITS2 regions were purified using Agencourt AMPure XP beads (Beckman Coulter, USA) with two ethanol washes and resuspension in 10 mM Tris (pH 8.5). A second PCR was used to attach Illumina sequencing adapters and index tags. This reaction included 5 µL of purified PCR product, 2.5 µL of Fluidigm Access Array Barcode 384, and 1X KAPA HiFi HotStart ReadyMix (Roche Sequencing Solutions), in total volume of 50 µl. The indexing PCR conditions were: 95 °C for 3 min, 8 cycles of 95 °C for 30 s, 55 °C for 30 s, 72 °C for 30 s; and a final extension at 72 °C for 5 min. Indexed amplicons were cleaned with AMpure XP beads and quantified using a Qubit a fluorometer and the DNA HS assay kit (Thermo Fisher, Temara, Morocco). Libraries were normalized and pooled following Illumina’s protocol. Sequencing was performed on an Illumina MiSeq platform (Illumina, MegaFelx, Casablanca, Morocco) using a MiSeq Reagent Kit V3 (2 × 300 bp paired end reads) [31].

Bioinformatics and statistical analysis

Fresh weight, dry weight, and aspect ratio values were compared using one-way analysis of variance (ANOVA). When significant differences were detected, Tukey’s Honest Significant Difference (HSD) post-hoc test was applied to identify pairwise group contrasts. Statistical analyses and graphical outputs were generated using R (version 4.4.2) [32], with visualization performed using the ggplot2 (3.5.1) package. Raw reads were processed using the DADA2 pipeline [33] in R (version 4.4.2). For the 16 S rRNA reads, forward and reverse reads were truncated at 81 and 93 nucleotides, respectively; ITS2 reads were truncated at 80 and 90 nucleotides. Low-quality reads were removed (Table S3). Amplicon sequence variants (ASVs) were inferred and taxonomically classified using BLAST against the SILVA database (Release 138.2) for bacteria, and the UNITE database for fungi [34, 35], using the default settings of the assign Taxonomy function [31]. Reads corresponding to plant DNA were removed.

Data was imported into phyloseq package version 1.50.0 [33] for downstream analysis. Alpha diversity (Observed, Shannon, and Simpson indices) was calculated using the estimate_richness command. Pairwise Wilcoxon tests with adjusted p-values were used to assess differences between locations. Beta-diversity was calculated using the Bray-Curtis dissimilarity and visualized via Principal Coordinate Analysis (PCoA) using vegan package version 2.6.10 [36]. Permutational multivariate analysis of variance (PERMANOVA) was conducted with the adonis function. Taxonomic composition was visualized using ggplot2, RColorBrewer, and ggthemes [37–39]. Core and unique taxa were identified using ggvenn [40]. Spearman correlations between microbial genera and environmental variables were calculated using the stats package version 4.4.2 and visualized as annotated heatmaps with pheatmap [41]. Redundancy Analysis (RDA) was performed using vegan to explore the relationships between microbial communities cand environmental parameters. Mantel tests were conducted using the linkET package version 0.0.7.4 to assess correlations between microbial community structure and environmental parameters.

Results

Soil’s physicochemical properties

We observed substantial variations in soil physicochemical properties across sampling sites (Table 1). All soils were generally alkaline, with pH values ranging from 7.78 to 9.14 and exhibited low electrical conductivity, ranging from 0.05 to 0.14 mS/cm. Soil texture varied widely, with clay, silt and sand contents ranging from 8 to 34%, 10–64%, and 18–82%, respectively. Macronutrient levels also varied considerably across sites: total nitrogen ranged from 0.01 to 0.20%, total phosphorus ranged from 1.13 to 6.70%, and total potassium rom 0.06–0.38%. Calcium carbonate content spanned a wide range, from 0.0% to 58.3%, indicating a gradient from non-calcareous to highly calcareous conditions soils Micronutrients profiles differed across locations, with potassium oxide (93.72–503 mg/kg), magnesium oxide (9–1091 mg/kg), and calcium oxide (16523 mg/kg) showing pronounced variation. Elevated magnesium oxide and potassium oxide concentrations were observed in sites S6 and S7, while calcium oxide levels were highest in S1 and S2. Additionally, we observed site-specific differences in iron, zinc and manganese concentrations (Table 1).

Table 1.

Soil physicochemical properties

	pH	EC	Clay	Slit	Sand	TN	TP	TK	OM	TCaCO3	NNO3	NH4+	Cl	K2O	Na2O	CaO	MgO	Cu	Mn	Fe	Zn	B
		ms/cm	%
Site1	8.68	0.09	32	42	26	0.2	2.98	0.32	3.85	35	4.25	<0.5	30.46	445.68	56	16523	363	0.61	4.24	2.64	0.25	0.32
Site2	8.66	0.1	30	38	32	0.17	3.10	0.19	2.38	58.3	12.05	1.29	23.29	240.04	52	16151	527	0.32	2.22	2.54	0.21	0.18
Site3	8.62	0.09	34	40	26	0.14	4.46	0.24	1.71	28.8	9.07	<0.5	21.4	265.12	44	15290	344	0.56	2.25	1.95	0.43	0.20
Site4	8.07	0.05	8	10	82	0.13	1.13	0.06	1.61	0.2	8.91	<0.5	31.69	93.72	17	2323	91	0.12	1.74	2.93	0.22	0.14
Site5	7.78	0.11	18	14	68	0.17	6.70	0.1	2.34	0	43.24	0.57	54	126.89	30	3426	328	0.73	21.34	6.68	2.34	0.26
Site6	8.61	0.14	30	40	30	0.18	1.47	0.27	2.87	7.7	33.99	0.22	129.17	233.00	76	15367	1091	0.64	2.91	2.66	0.27	0.37
Site7	8.65	0.14	18	64	18	0.18	4.36	0.38	2.75	8.4	26.74	1.23	66.84	503.00	92	14911	869	2.65	4.43	2.85	0.68	0.93
Site8	8.72	0.06	16	34	50	0.09	3.68	0.23	1.02	0.1	7.65	0.99	25.01	286.00	24	2788	166	1.12	3.35	2.37	0.39	0.86
Site9	9.14	0.05	10	16	74	0.01	1.35	0.07	0.16	1.9	3.22	1.21	28.38	64.00	14	7141	113	0.3	2.01	1.53	0.21	0.11

Fruit weight differences and morphological variation among locations and environmental parameters

Significant differences in average fruit weight were observed across locations (F = 4.40, p = 0.011; Fig. 2). Fresh fruit weight ranged from 2.0 to 14.5 g, and dry weight from 0.3 to 5.7 g (mean of five fruits per tree). The highest median fruit weights were recorded in Berkane, whereas fruits from Essaouira and Agadir exhibited lower weights. In addition to weight, we observed marked differences in fruit morphology, including size, shape, and maturity (Fig. S2). Fruit aspect ratio revealed highly significant differences among samples (ANOVA, p < 0.001; Fig. S3). Aspect ratio values ranged from 0.9 to 2.0, showing a broad range of fruit shapes from round to elongated. Field observations of significant morphological variation, such as fusiform, oval, and round fruit shapes, are supported by these quantitative results.

Fig. 2.

Distribution of fresh and dry weights of Argan fruits across locations (Agadir, Berkane, Essaouira, and Rabat) . Boxplots represent the fresh (green) and dry (brown) weights of argan fruits. Each data point corresponds to the average weight of five fruits sampled from a single tree. Statistical analysis was performed using one-way ANOVA followed by Tukey’s HSD post hoc test to evaluate differences among locations. Significant variation was observed for both fresh weight (p = 0.0108) and dry weight (p = 0.0191)

Microbiome structure and richness across locations

We obtained 193,536 high-quality bacterial reads and 219,699 fungal reads, resulting in 139 and 288 ASVs, respectively. Rarefaction curves indicated sufficient sequencing depth for both bacterial and fungal communities (Fig. S4).

Bacterial richness varied significantly by location (p < 0.05) (Table 2; Fig. 3). Agadir fruits exhibited the highest bacterial diversity, significantly exceeding that of Essaouira (p = 0.0468) and Rabat (p = 0.0091), while other pairwise comparisons were not significant (p > 0.05; Fig. 3A; Table 2). Shannon, and Simpson indices revealed no statistically significant differences, although Agadir consistently showed higher values (Fig. S4A–B, Table 2). For Fungal communities, none of the diversity indices revealed significant variation between locations (Fig. 3B; Fig. S4C–D).

Table 2.

Alpha-diversity adjusted p-values between locations using pairwise Wilcox-test

	Bacteria			Fungi
	p-value
	Shannon	Simpson	Observed Richness	Shannon	Simpson	Observed Richness
AGADIR vs. ESSAOUIRA	0.37	0.14	0.047*	0.46	0.25	0.96
AGADIR vs. RABAT	1.00	0.63	0.009**	1.00	0.86	1.00
AGADIR vs. BERKANE	0.94	0.89	0.56	0.95	1.00	0.94
ESSAOUIRA vs. RABAT	1.00	0.64	1.00	0.44	0.23	0.59
ESSAOUIRA vs. BERKANE	0.47	0.47	0.49	0.68	0.36	0.95
RABAT vs. BERKANE	0.72	0.72	0.72	0.86	0.63	0.86

Bold values indicate statistically significant differences (*p < 0.05; **p < 0.01)

Fig. 3.

Alpha diversity of microbial communities associated with ARGAN fruit: (A) Observed ASV richness based on 16S rRNA gene sequencing across locations: Agadir, Essaouira, Rabat, And Berkane. B Observed ASV richness based on ITS2 region sequencing at the same locations. Pairwise comparisons were performed using the Wilcoxon test. Significance levels from Wilcoxon tests are indicated as follows: (*): p < 0.05, (**): p < 0.01

PCoA based on Bray-Curtis analysis dissimilarity, revealed no significant clustering by location for either bacteria (Fig. 4A) or fungi (Fig. 4B). PERMANOVA confirmed these results, showing no significant location effect on bacterial (p = 0.454) or fungal (p = 0.112) community structure.

Fig. 4.

PCoA analysis of community composition based on Bray–Curtis dissimilarity of bacterial communities (16S rRNA gene) (A); PERMANOVA (Adonis test, p = 0.454). And fungal communities (ITS2 region) (B) PERMANOVA (Adonis test, p = 0.112)

Microbial community composition and core taxa locations

Bacterial communities were dominated by the phylum Pseudomonadota, comprising over 70% of all samples (Fig. S6-A). In Rabat, the community consisted exclusively of Pseudomonadota. Other phyla, such as Halobacteriota and Chlamydiota, appeared only in Essaouira, while Patescibacteria was the second most abundant in Berkane (10.93%). At the family level, Erwiniaceae was predominant across sites (Fig. S6-B). At the genus level (Fig. 5-A), Pantoea was the dominant genus across all locations: Essaouira (89.04%), Berkane (66.95%), Rabat (65.50%), and Agadir (45.05%). Agadir exhibited the greatest bacterial diversity, harboring 22 unique genera (41.5% of all detected), whereas Rabat had only one unique genus (Delftia) among eight detected. Only Pantoea and Rheinheimera were shared across all four locations (Fig. 6-A).

Fig. 5.

Taxonomic Composition and relative abundance of the 20 most prevalent taxa across Locations. A Relative abundance of bacterial community composition at genus level based on 16S rRNA gene sequencing. B Relative abundance of fungal community composition at the genus level based on ITS region sequencing. (Taxa not listed among the top 20 most common are grouped under “Others”). Detailed statistical value of relative abundance of each taxon per location are in the excel sheet of the supplementary file

Fig. 6.

A Venn diagram representing the distribution of bacterial genera (16S rRNA dataset) across four geographical locations: Agadir, Essaouira, Rabat, and Berkane. B Venn diagram showing the distribution of fungal genera (ITS2 dataset) across the same locations. Each number indicates genera number shared or unique to specific combinations of locations. % Represent the proportion relative to the total detected genera

Fungal communities consisted of only two phyla: Ascomycota (dominant, ~ 95%) and Basidiomycota (Fig. S6-C). At the family level, Agadir showed the greatest diversity, while Berkane had the least. Saccharomycodaceae was dominant in Berkane (82.4%) and Essaouira (43.6%), while Agadir’s community was enriched in Cladosporiaceae (25.3%) and other diverse families including Trichocomaceae, Dothideomycetes, and Neoschizotheciaceae (Fig. S6-D). At the genus level (Fig. 5-B), Cladosporium dominated Agadir (25.3%), followed by Hanseniaspora, Aureobasidium, and Alternaria. Hanseniaspora was dominant in Berkane (82.4%) and Essaouira (43.6%), indicating lower diversity. In Rabat, Aureobasidium was the most abundant (26.3%). Rare genera such as Knufia, Morinia, and Elsinoe appeared only in Rabat and Essaouira. Agadir and Essaouira harbored the most unique fungal genera, 11 (35.5%) and 8 (25.8%), respectively. Only Hanseniaspora and Aureobasidium were common to all locations (Fig. 6–B). Pantoea was the only bacterial genus found in over 50% of samples, qualifying as core taxon. No fungal ASVs met core criteria across all samples. However, Hanseniaspora (asv18) was identified as a location-specific core taxon in Agadir and Rabat.

Fruit microbial divergence along edaphic gradients

Spearman correlation analyses revealed significant associations between soil variables and microbial taxa (Fig. 7). In bacteria, Pantoea, Alcaligenes, and Raoultella positively correlated with organic matter, K₂O, and EC, while Nocardioides and Rhizobium showed negative correlations. Pantoea abundance was positively associated with latitude (p = 0.031) and clay content (p = 0.045), and negatively with soil NH₄⁺ (p = 0.0013). TM7a was positively correlated with EC (p = 0.0084), B (p = 0.0146), and Cu (p = 0.0285). Thermus was positively associated with altitude (p = 0.0499), but negatively with Fe (p = 0.0157). Acinetobacter was positively associated with altitude and distance from the coast, but negatively with K₂O (p = 0.0373).

Fig. 7.

The Spearman correlation matrix shows the correlations between environmental factors and genera (A) for bacterial genera (16S rRNA dataset) and the same environmental factors (B) for fungal genera (ITS dataset). The correlation coefficient's strength, which ranges from -0.4 (strong negative) to +0.4 (strong positive), is reflected in color intensity. Genera and environmental factors are grouped by dendrograms according to how similar their correlation profiles are. Statistically significant relationships (p < 0.05) are shown by asterisks (*)

For fungi, several genera showed noteworthy correlations with soil chemical and spatial variables. Aureobasidium and Cladosporium were positively associated with chloride and altitude, suggesting that these genera may be adapted to more saline or higher-elevation environments. Alternaria (p = 0.0092), Hanseniaspora (p = 0.026), and Cladophoma (p = 0.023) exhibited significant positive correlations with soil pH, indicating a preference for less acidic soil conditions. These same genera showed negative correlations with soil Fe and NO₃⁻ concentrations, suggesting potential sensitivity to elevated iron and nitrate levels. Overall, the fungal correlations reveal that soil salinity, elevation, pH, and nutrient availability are key factors influencing fungal community structure across the study locations.

Environmental drivers of Argan fruit microbial composition

Redundancy analysis (RDA) (Fig. 8) was performed after removing highly collinear variables (Fig. S6) and revealed a clear association between bacteriak taxa and environmental parameters. For bacteria, the main contributing factors were pH, altitude, EC, and soil nutrient availability (Total N, Total P, and Total K). Pantoea showed a negative association with pH, while genera such as Acinetobacter and Raoultella aligned more closely with variations in soil nutrients and EC. For fungi, the RDA indicated that altitude, sand content, and nutrient availability (Total N, Total P, and Total K) were the strongest explanatory variables. Hanseniaspora was negatively associated with altitude, whereas species such as Stemphylium and Pleiochaeta aligned with gradients of soil nutrients. Overall, while the resulting model explained a modest proportion of variation, it revealed coherent ecological patterns linking specific microbial genera to soil chemistry and geographical distribution. Mantel tests supported the patterns observed in the RDA. Bacterial communities correlated more strongly with environmental variables than fungal communities. The strongest correlation for bacteria was with CaCO₃ (r = 0.335, p = 0.008), and for fungi with TN (r = 0.325, p = 0.022) (Fig. 9). These findings emphasize the role of soil nutrients and spatial factors in shaping fruit microbiota.

Fig. 8.

A Redundancy Analysis (RDA) plot illustrating the relationship between bacterial genera and environmental variables. B RDA plot showing the relationship between fungal genera and the environmental parameters

Fig. 9.

Integrated visualization of Pearson correlation coefficients (heatmap) and Mantel test results (curved lines) between environmental parameters and microbial community composition (Bacteria and Fungi)

Discussion

This study provides the first comprehensive characterization of the Argan fruit-associated microbiome across its natural distribution range in Morocco. It was designed to test three related hypotheses: (i) that the diversity and composition of fruit-associated microbial communities vary across geographical regions; (ii) that soil physicochemical properties exert a stronger influence on fruit-associated microbial community assembly through edaphic filtering than geographical location alone; and (iii) that among soil variables, pH and phosphorus availability are the principal determinants of bacterial and fungal community structure. By integrating amplicon-based microbial profiling with analyses of fruit traits, spatial gradients, and soil physicochemical characteristics, this study elucidates the relative contributions of biogeography and soil-driven environmental filtering to the assembly of Argan fruit-associated microbial communities.

Geographical variation in Argan fruit traits across sampling locations

We first evaluated the influence of geographical location on Argan fruit traits and microbial patterns. Fruit weight and morphology varied significantly among the four locations (ANOVA, p = 0.011), indicating strong location-specific phenotypic differentiation likely driven by combined effects of climate, altitude, and soil conditions. This is consistent with previous reports of morphometric heterogeneity in argan populations across environmental gradients [42, 43]. Despite these marked phenotypic differences among locations, microbial community structure did not show significant geographic clustering, as evidenced by PCoA and PERMANOVA analyses. This indicates that geographical location alone is insufficient to explain variation in fruit-associated microbial communities, even though it strongly influences fruit morphology. Contrary to the hypothesis of uniform microbial diversity, our results showed significant spatial variation in microbial diversity indices across sites, suggesting that microbiome assembly is not uniform even within a single plant species.

Soil physicochemical drivers of microbial assembly

Having established that geographical location alone did not significantly structure microbial communities, we thereafter evaluated the role of soil physicochemical properties. The significant differences in soil physicochemical properties across sites (Table 1) suggested that edaphic filtering plays a key role in microbial community assembly. Our study highlighted that microbiome diversity is influenced by both biogeographic patterns and local environmental conditions. This supports our hypothesis that geographical location would influence microbiome composition but have a weaker effect than key soil physicochemical properties. Although fruit weight and morphology varied significantly among locations, PERMANOVA and PCoA analyses showed that location did not significantly explain variation in bacterial or fungal community structure. Instead, correlations, RDA, and Mantel tests consistently indicated that pH, altitude, TN, and CaCO₃ explained a larger portion of the variance in microbial composition than geographic coordinates or distance from the coast. This finding was particularly clear for bacteria, where pH and altitude emerged as a strong predictor, whereas CaCO₃ showed a minor impact. For fungi, TN remained the most influential nutrient-related factor, confirming its importance in structuring fungal assemblages. Regarding the third hypothesis, that pH and TP would be the most influential individual parameters, the analysis partly supports this expectation. pH shows major associations with shifts in both bacterial and fungal community structures, while TP displayed a minor contribution. Instead, other nutrient variables, notably TN for fungi and TotalK with organic matter for bacteria contributed more substantially to the constrained variance. These results indicate that while pH remains a robust predictor across microbial groups, phosphorus is not the primary driver in this system. Rather, nitrogen-related variables and other soil chemical properties also play important roles, and the dominant environmental drivers differ between bacterial and fungal communities.

Core and persistent taxa of the Argan fruit associated Microbiome

We have evaluated the microbial taxa and their ecological persistence. The argan fruit-associated microbiome was dominated by a few key taxa. Pantoea was the most dominant bacterial genus in three of the four sites (Rabat, Essaouira, and Berkane), accounting for ~ 65–89% of the bacterial community. Notably, Pantoea and Rheinheimera were shared across all locations, highlighting the ecological persistence of Pantoea in the argan fruit habitat. Previous studies have reported Pantoea as a common epiphytic or endophytic associate of various fruits. For example, Pantoea comprised 21.66% of the olive fruit microbiome [44] and up to 46–48% of the apple fruit microbiome, where it was identified as part of the genotype-specific core [25]. While some Pantoea species are known pathogens, many are mutualists or commensals with plant-beneficial traits [45–47]. Their known ability to colonize plant surfaces and tolerate environmental stress [48] may explain their ecological flexibility and dominance in the Argan fruit microbiome.

Beyond Pantoea, other bacterial taxa such as TM7a, Thermus, Raoultella, and Acinetobacter were significantly associated with the Argan fruit microbiome. The positive correlation of TM7a with electrical conductivity (EC) and trace elements like boron and copper suggests a preference for nutrient-enriched or moderately saline microenvironments, like findings in grapefruits [49]. The presence of Thermus, typically associated with thermophilic and oligotrophic conditions, indicating adaptation to iron-deficient upland soils. This is consistent with studies in montane coffee ecosystems where Thermus abundance increased with elevation [50]. Acinetobacter was more abundant at higher altitudes and inland locations and negatively correlated with potassium oxide (K₂O), further supporting the role of edaphic gradients in shaping community composition.

Fungal communities were dominated by Hanseniaspora, a yeast genus commonly associated with sugar-rich fruits and early colonization [51]. It was the most abundant genus across all sites, particularly in Berkane (82%) and Essaouira (44%). Its positive correlation with pH and negative correlation with iron and nitrate aligns with patterns observed in grape and fig microbiomes, where acidic, iron-rich conditions suppressed its prevalence [49, 52]. In contrast, genera such as Cladosporium, Aureobasidium, and Alternaria were more abundant in Agadir and Rabat. For instance, Cladosporium was positively correlated with organic matter and total potassium, indicating a preference for nutrient-rich environments, consistent with previous observations in apple and olive fruit microbiomes [53].

While this study provides robust correlative evidence linking soil properties to Argan fruit microbiomes, it is limited by its cross-sectional design and lack of functional resolution. Future studies incorporating temporal sampling, host genotypic data, and functional omics approaches will be essential to disentangle causal mechanisms and assess the ecological consequences of fruit-associated microbial variation for Argan reproduction and resilience.

Conclusion

This study provides the first comparative analysis of Argan fruit-associated bacterial and fungal communities across different Moroccan soil environments and demonstrated that soil physicochemical properties are the primary drivers of fruit microbiome composition. Despite distinct differences in fruit morphology and weight among locations, microbial community structure showed limited geographic differentiation, indicating that spatial location alone does not strongly constrain fruit-associated microbiota. Instead, soil pH, nutrient availability, electrical conductivity, and carbonate content appeared as key determinants of microbial diversity and taxonomic composition. Bacterial communities were more strongly structured by edaphic factors than fungal communities, with CaCO₃, total nitrogen, total P, and organic matter showing the strongest associations. In contrast, total P demonstrated comparatively weaker effects. The dominance and persistence of Pantoea across sites, together with location-specific enrichment of fungal genera, for instance, Hanseniaspora and Cladosporium, further emphasized the selective influence of soil conditions on fruit-associated microbiota. Together, these findings demonstrated that soil pH and nutrient status employ a stronger influence on Argan fruit-associated microbial communities than geographic location, highlighting the importance of edaphic gradients in shaping plant-microbiome interactions in semi-arid ecosystems. This soil-driven structuring of the fruit microbiome provides a clearer ecological framework for understanding microbial assembly in Argan trees and emphasizes the need to consider soil properties when interpreting fruit-associated microbial variation across landscapes.

Abbreviations

ASV

Amplicon Sequence Variant

B

Boron

CaCO₃

Calcium Carbonate

CaO

Calcium Oxide

Cu

Copper

EC

Electrical Conductivity

Fe

Iron

K₂O

Potassium Oxide

MgO

Magnesium Oxide

Mn

Manganese

Na₂O

Sodium Oxide

PCoA

Principal Coordinates Analysis

RDA

Redundancy Analysis

TK

Total Potassium

TN

Total Nitrogen

TP

Total Phosphorus

Zn

Zinc

Authors’ contributions

FZA: experimentation, analysis of sequencing data, and draft the manuscript. JI & MEM: designed the sampling strategy and collected the data; MH: conceptualization, manuscript review and editing. BA: conceptualization, manuscript drafting, review and supervision. All authors contributed to the article and approved the submitted version.

Funding

This research project was funded by OCP Group, Project Number AS-77 and Seed Grant − 151STPR07-4.

Data availability

The datasets used in this study are available online with the accession numbers **PRJNA1307012** and **PRJNA1307026** in the NCBI SRA database.

Declarations

Ethics approval and consent to participate

This research does not include or report the results of a clinical trial; therefore, trial registration and ethical approval were not required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.