3.1 Abundance of AMF in roots of barley
At physiological maturity (GS90), AM fungal root colonization and percentage of root length cointaining arbuscules were significantly affected by the interaction among year (Y), genotype (G) and AM fungal inoculation (Inoc) (Table S3). In 2020, Inoc increased the AM fungal root colonization of Atomo and Concerto up to an average of 53% (Fig.3a). Compared with not-inoculated plants (-M), the increases were 27% in Atomo and 10% in Concerto, whereas in Atlante there was a slightly, but not significant promotion. Moreover, in 2020, the variety Atomo showed the highest rate of arbuscules in roots (14%) and compared with the other treatments the increase was 12% (Fig. 3a,b). In 2021, a different pattern was reported, AM fungal root colonization and arbuscules values were generally higher than in 2020 (76% vs 41% and 64% vs 4%) and only Concerto showed significant increases of both AM fungal traits in comparison to control (+17 and +25, respetively) (Fig. 3a,b). Thus, the pattern of response of Concerto was consistent across the years of cultivation.
The response of barley in term of AM fungal abundance was similar to values previously recorded in controlled sterile conditions with several spring cultivars and with many types of AM fungal inoculants (average AM fungal root colonization: 54%) (Baon et al. 1993; Coccina et al. 2019; Gray et al. 1991; Jensen 1983; Thirkell et al. 2019; Vierheilig et al. 2000; Watts-Williams et al. 2020). The responses in pot culture were recorded under a range of soil nutrient availability and water stress. However, similar to our results, also in controlled conditions, the cultivar affected AM fungal root colonization with a large range of response, from ca. 18% to 80%. Noteworthy, under water stress, the response in AM fungal colonization to inoculation was higher than in well-watered conditions (Jerbi et al. 2022). Positive AM fungal colonization responses following inoculation were consistenly observed under low soil P availability (Brown 2013; Jensen et al. 1984).
Air temperature also played a significant role in AMF-barley interaction and low temperatures reduced the rate of AM fungal colonization in spring and winter cultivars compared to control temperatures (Hajiboland et al. 2019). These evidences can explain the different results we obtained in the two years. Indeed, in 2020, when soil was clay loam with very low P availability and no drought stress conditions, two over three varities of barley (i.e. Atomo and Concerto) were responsive to inoculation in term of AM fungal traits. By contrast, in 2021, with silty clay loam soil with low P availability and drough stress, only the variety Concerto was responsive to inoculation. Under field conditions, few experiments were carried out on barley and for more than one year, and no one tested different genotypes (Beslemes et al. 2023; Clarke and Mosse 1981; Heydari et al. 2023; Khaliq and Sanders 2000; Powell et al. 1980). Beslemes et al. (2023), in contrast with our results, did not find interactions between AM fungal inoculation and year of cultivation. Indeed, they found consistent increases between years of cultivation in AM fungal colonization in comparion with the control (+M vs -M: 55% vs 49%).
Furthernore, our data highlighted an opposite pattern between years of cultivation in term of occurrence of vesicles in roots: in 2020 when AM fungal colonization and arbuscules were low, AMF invested their resources in vesicles, whereas in 2021 when AM fungal colonization and arbuscules were high, low percentages of vesicles were recored (Fig. 3c).
3.2 Effectiveness of AMF on barley grain yield
Barley varieties responded differently to AM fungal inoculation in the two years of cultivation in terms of grain yield and nutrient concentration (Table S3). In 2020, according to the recorded positive response of AM fungal root colonization and arbuscules, grain yield was significantly promoted by 64% and 37% in the inoculated varieties Atomo and Concerto, respectively. Furthermore in line with no changes in AM fungal root traits, a slight but not significant enhancement was reported in Atlante (7%) (Fig. 4b). In 2021, Concerto showed a consistent positive response in grain yield (+78%) (Fig. 3a,b; Fig. 4a). This was also supported by the recorded promotion of root colonization and arbuscules. Moreover, while Atlante showed strong increases in grain yield under inoculation (+134%), Atomo did not show grain yield increases. Thus, the variety Concerto had a more stable response to inoculation with the indigenous AM fungal consortium and can be a good candidate for AM fungal inoculation, irrespective to soil nutrient availability and drought stress.
Previously, in a meta-analysis on the effect of AM fungal inoculation on cereal grain yields, no effect was reported on barley both in controlled and field conditions (Zhang et al. 2019). However, they pointed out the lack of experiments on barley and the importance to study the effect of breeding and environment to validate the pattern. Under low soil nutrient availability, yield grain increases were reported in field inoculation with a mixture of AMF (Beslemes et al. 2023; Masrahi et al. 2023). In accordance with the general relationship found in crop plants between ∆AM and MRyield (Lekberg et al. 2005; Pellegrino et al. 2015), in 2020 we found a significant relationship (R2=0.71; P=0.004). By contrast, no relationship was found in 2021 (R2=0.024; P=0.691). This is in contrast with our expectations that higher AM fungal colonization would have greater driven grain yield under drought stress. However, our results support that soil nutrient availability are key drivers of the interaction AM fungal inoculation and crop yield (Zhang et al. 2019).
3.3 Effectiveness of AMF on barley nutrient uptake
Inoculation determined in both years a consistent increase of P in grain of the three studied varieties (Fig. 4d; Table S3). However, the relative increase in 2021 (42%), under low soil P availability and drought stress, was larger than in 2020 (24%), characterized by a very low soil P availability and no drought stress. Thus, we can assert that under low and very low soil P availabilities, AM fungal inoculation in field conditions strongly promote the concentration of P in barley grains. Moreoverwe can assert that the symbiosis is more efficient under drought conditions. Nevertheless, the ∆AM and mycorrhizal P response ratio were not related between each others in both years (R2=194; P=0.236; R2=0.361; P=0.087). Furthermore, the positive and significant relationship we found between grain yield and P concentration (Fig. 4g) does not support a diluition effect mediated by AMF.
By contrast to our results, a meta-analysis on the field inoculation did not find significant changes of P concentration in wheat grains (Pellegrino et al. 2015). However, the variability we observed between years is in accordance with Porcel and Ruiz-Lozano (2004) who reported that AMF enable host plants to grow and uptake P more efficiently under drought stress, through plant osmotic adjustment (Harrier 2001).
As regard N concentration in grain, at both years of cultivation, Atomo and Atlante were reported to be positively affected by AM fungal inoculation (+6% and +8%, respectively), while Concerto was not affected in 2020 and negatively affected in 2021 (Fig. 5d). This result can not be explained by significant relationships between ∆AM and mycorrhizal N response ratio (data not shown). Moreover, grain yield and N concentration were not significant related (Fig. 4g). Thus, as previously observed in some genotypes of wheat, the consistent increases in Atomo and Atlante could be explained by a change of AM fungal communities in roots induced by inoculation (Marrassini et al. 2023). In several experiments, no changes in N grain concentration were reported in wheat (Pellegrino et al. 2015). However, the number of meta-analized field trials was low pointing out that field response of cereals to AMF deserves more study. Indeed, Glomus sp. and Gigaspora sp. and a multiple-species AM fungal field inoculum promoted the uptake of N in grain of two cultivars of barley (Beslemes et al. 2023; Masrahi et al. 2023).
In 2021, the concentration of P was also promoted in the shoots of Concerto sampled at the four-leaves unfolded stage (GS14) (+18%) (Fig. S1a). This pattern of response to the AM fungal inoculation can support the success of the inoculation and the high responsiveness of this variety. Actually, this reponse is in line with the increase of AM fungal abundance in the roots of Concerto and with its grain yield response. In addition, averaged over genotypes, N shoot concentration was increased in inoculated treatments (Fig. S1b), further supporting the success of inoculation at early plant growth stages.
The response in K uptake of barley genotypes to AM fungal inoculation was consistent among years (Fig. 5h). Atomo was the sole responsive variety (+12%). Overall, considering the two years, no relationship was observed between the ∆AM and mycorrhizal K response ratio (R2=0.000; P=0.939), and no relationship was underlighted between grain yield and K concentration (Fig. 4g). The increases of K observed in Atomo was similar to the results previously recorded on barley shoots and grain under field inoculation (Powell et al. 1980; Mahrahi et al. 2023). In addition, irrespective of inoculation, the three varieties showed higher grain K concentration in 2021 than 2020 (Fig. 5i). This is probably linked to differences in soil fertility.
Moreover, we did not observe any effect on Mg concentration in grain due to AM fungal inoculation, with the exception of the promotion in 2020 with Atomo and Atlante (+39% and +17%, respectively). Grain Mg concentration was significantly and positively related to yield (Fig. 4g). Therefore, we can support an indirect effect of AMF. By contrast, not significant relationships were found between the ∆AM and mycorrhizal Mg response ratio (2020: R2=0.432, P=0.054; 2021: R2=0.091, P=0.430). Taking into consideration that K/Mg ratios is important for the nutritional status of plants(Xie et al. 2021), the fact that Mg and K were promoted in Atomo, could be an indicator of efficient physiological processes.
Inoculation with AMF increased the concentration of Zn in grain of Atlante and Atomo, irrespective of the year of cultivation (+10% and +9%, respectively), while in Concerto no effect was observed (Fig. 5b). This response was not mediated by the increase of grain yield (Fig. 4g). Our results confirmed the general pattern of Zn response of crops, including wheat, corn and rice, to AMF (Lehmann et al. 2014; Pellegrino et al. 2015), altough no relationship was observed between the ∆AM and mycorrhizal Zn response ratio (R2=0.035; P=0.457). Nevertheless, our data did not confirm that the response is modulated by the availability of Zn in soil, since it did not vary between years of cultivations.
A positive effect of R. irregularis was reported on shoot and grain Zn concentrations of barley when soil was fertilised with Zn (Al Mutairi et al. 2020; Cardini et al. 2021; Watts-Williams and Cavagnaro 2018). However, in accordance with our results, the AMF-mediated effect depended on barley genotypes (Al Mutairi et al. 2020). We also highlighted a significant interaction between barley genotype and year of cultivation (Table S3). Overall grain Zn concentration was higher in 2021 when soil had an adequate level of Zn availability than in 2020 when soil had very low Zn availability (Fig. 5c). However, the relative increase of Zn uptake into grain varied among barley varieties, and Concerto that in general uptakes less Zn in grain showed a relative lower increase (2021 vs 2020) respect to the other genotypes (Fig. 5b). Our data are in agreement with the results of a large study on the variability of grain Zn concentration in many wheat cultivars (Fan et al. 2008).
The variety Concerto showed also a consistent increase of Fe in grain in 2020 when soil had an optimum level of Fe and in 2021 when soil had a low Fe availability (+24% and +130%, respetively) (Fig. 5a). Moreover, in 2020, Atomo showed, similarly to root colonization and yield responses, an increase in grain Fe concentration following inoculation (+140%). By contrast, in 2021, Atomo did not show changes in grain Fe concentration, similarly to the responses in root colonization and yield. Finally, no changes in grain Fe uptake were recorded in Atlante following inoculation, according to root colonization in both years (Fig. 5a). The positive mycorrhizal effect observed on barley Fe uptake is in line with the significant and positive relationship between the ∆AM and the Fe mycorrhizal response ratio found in 2020 and 2021 (R2=0.449, P=0.048; R2=0.590, P=0.016). Moreover, the increase of Fe concentration in grain were not determined by yield reduction (Fig. 4g). Overall, our results are in line with the positive effect reported by Lehmann and Rillig (2015) on crops (e.g., grasses) in lab and field conditions and with the large variability observed among wheat genotypes (Pellegrino et al. 2020).
Arbuscular mycorrhizal fungal inoculation did not determine any changes on Mn uptake in barley grain or determined significant decreases (i.e., in 2020: -19% and -34% in Atomo and Concerto, respectively) (Fig. 5j). This is in accordance with the significant and negative relationship observed between the ∆AM and the mycorrhizal Mn response ratio (R2=0.464; P=0.043). This is in accordance with the overall decrease (-4%) under AM fungal inoculation (Lehmann and Rillig 2015). Explanations could be the fact that AMF increases soil pH and the availability of Mn in soil with subsequent leaching, and the reduction of Mn-reducing microbes or the promotion of Mn-oxidizing microbes.
The concentration of Cu in grain was not affected by AMF, with the exception of Atlante that showed significant increases irrespective to year (+8) (Fig. 5k). This is in agreement with the not significant relationship observed between the ∆AM and the mycorrhizal Cu response ratio (R2=0.292; P=0.133). Moreover, our results are partially in agreement with the positive AM-fungal mediated effect found on crop Cu uptake (+29%) (Lehmann and Rillig 2015). However, since the Cu response to AMF was reported to be modulated by the availability of Cu in soil, with reductions at high availabilities, we can suppose that the availability of Cu in our soil was (very) low, and thus the response in Atomo and Concerto was undetectable. Moreover, since the year of cultivation variably affected the response of barley genotype in grain Cu uptake, irrespective of AM fungal inoculation, we can state that genotype in interaction with soil Cu availability could have determined these changes (Fig. 5l).
Increases in Ca uptake in grain were observed, irrespective to year of cultivation, in Atlante and Concerto (+21% and 95%) (Fig. 5e). A direct effect of AMF could be hypothesized, since the uptake was not determined by yield increases (Fig. 4g). However, we can not directly link the AM abundance in roots with Ca uptake, since no significant relationship was observed between the ∆AM and the mycorrhizal Ca response ratio (R2=0.028; P=0.508). Calcium changes under AM fungal inoculation was scarcely investigated, and Watts-Williams and Gilbert (2021) found no changes in barley inoculated with R. irregularis. Furthermore, we did not observe any changes among barley varieties in 2021, whereas in 2020 Atomo and Concerto showed a higher uptake of Ca (Fig. 5f). In 2020, soil with a higher CEC and lower Ca availability could explain the interaction. Overall, our results support the fact that AMF can contribute to food nutrition.
Finally, PERMANOVA allowed to summarize the pattern of plant and AM fungal parameters and to highlight significant interactions between Y, G and Inoc (Table S4). The PCO biplot (Fig. 6) showed that all genotypes differently responded to inoculation in the two year of cultivation. Although the year of cultivation was very strongly affecting the pattern of reponse, with an explained variance of 65%, the third-order interaction explained 7%. Looking at the PCO biplot, it showed a clear difference between years. In 2021 the agronomic reponse was less variable among replicate plots, and this encouraged a further investigation on root AM fungal communities. Our results are in contrast with our first hypothesis that genotype exerts a greater control over the response of barley to AM fungal inoculation than the environment. Indeed, our hypothesis was based on recent findings that wheat genotype inoculated in the field with the same indigenous consortium was a major driver of the agronomic response (Marrassini et al. 2024). This supports a less variable response of wheat.
The PERMANOVA pairwise comparisons, utilised to dissect this interaction, highlighted consistent differences among the inoculated and not inoculated group of varieties and in 2020 and in 2021 (Table S5). Concerto was consistently and positively affected by AM fungal inoculation in both years, while Atomo and Atlante only in 2020 and 2021, respectively. This supports a robust and stable response of Concerto to inoculation across years and demostrates its less susceptibility to pedo-climatic variability. Therefore, the modern crossbreed variety Concerto developed in UK, showed a high mycorrhizal responsiveness to the indigenous AMF and this highlight the a good local adaptation of the crop to soil and AMF, regardless their origin. Furthermore, the year was consistently found to significantly affect the pattern of plant and AM fungal parameters in both inoculated and not-inoculated groups of barley varieties, underlining the major role of environment in shaping plant response under the same agronomic practice. Finally, PERMDISP showed significant difference among barley genotypes (Table S4). A higher variable pattern of response was observed in Atomo, whereas less variable patterns in Concerto and Atlante (Fig. 6). Therefore, irrespective to inoculation and year of cultivation, the response of Concerto and Atlante are more stable.
3.3. Diversity of AMF in the roots of barley
To test our hypothesis that host plant preference in the AMF colonizing plant genotypes inoculated with indigenous AMF is driven by changes of community structure and not by changes in composition of AMF, we investigated the AM fungal community diversity in roots of the three barley genotypes. The study allowed also to investigate host plant preference under no inoculation. Moreover, we investigated the role of plant growth stage on modulating host plant preference. We focused the investigation on root samples collected in 2021.
3.3.1. Illumina sequencing output, AM fungal richness and alpha-diversity in barley roots
After curation of the AM fungal sequences, 54,214 reads, ranging from 535 to 2,507 reads per sample, were retrieved and assigned to 26 VTXs (Fig. S2; Table S6). The 26 AMF VTXs belonged to three orders (i.e., Diversisporales, Entrophosporales, Glomerales) and four families (i.e., Diversisporacea, Gigasporaceae, Entrophosporaceae, Glomeraceae). The accumulation curves and rarefaction analyses of AMF confirmed that the Illumina sequencing effort was sufficient for the analysis (data not shown). Barley genotype and inoculation differently affected AM fungal richness (S) and alpha-diversity (H’ and λ) in the two growth stages (G12 and GS90) (Table S7). Atlante showed a higher diversity (S and H’) than the other genotypes (+51% and +22%, respectively) (Table S8).
Previously, the AM fungal diversity was only studied in single varieties of barley utilizing the taxonomic-based assessment of spores in soil and the molecular characterization in roots (Aguilera et al., 2017; Kaidzu et al. 2020). Therefore, our study highlights for the first time a host plant preference in AMF across barley genotypes. This is in accordance with the results obtained among genotypes of other crops (Kavadia et al. 2020; Mao et al. 2014; Parvin et al. 2021). Indeed, root traits variability among varieties of cereals, including barley, can be large (Nakhforrosh et al., 2014; Sendek et al. 2019; Robinson et al. 2018). Therefore, differences in root architecture could have determined the observed variable pattern of AM fungal diversity.
However, Inoc increased AM fungal richness and H’ only in Atlante at GS14 (+133% and +75%, respectively), while λ was promoted in Atlante and Atomo only at GS14 (+54% and +68%). In addition, AM fungal richness was higher at GS90 than GS14 (+23%). Since AM fungal diversity in roots promote plant productivity by the promotion of the abundance of productivity-promoting AMF (van der Heijden et al. 1998; Vogelsang et al. 2006), when inoculation determines the promotion of the diversity traits (i.e., Atlante and Atomo), a benefit of productivity can be expected.
3.3.2. Community composition and structure of AMF in barley roots
Sixty-five percentage of AM fungal taxa (VTXs) (65%) were shared among genotypes, irrespective to AM fungal inoculation and plant growth stage (GS). Similarly, barley genotypes shared 52% and 50% VTXs in inoculated conditions and in control, irrespective to GS. Moreover, at GS14 the percentage of VTXs shared in inoculated conditions was similar to the one in the controls (32% and in 37%), while at GS90 it reached 65% in inoculated plants and 40% in controls. These results support our hypothesis of a good host preference in AMF across genotypes of barley in both inoculation treatments. They are in accordance with the variability in AM fungal composition observed in field across three wheat genotypes cultivated in a similar climatic area (Marassini et al., 2024; Pellegrino et al. 2020). Thus, also in barley there is a core composition of AMF that can be considered generalists, but there are several specialist taxa uniquely retrieved in the roots of each genotype. Moreover, since inoculation with indigenous AMF did not modify the rate of shared taxa at early stage, we can state that locally-sourced inocula did not change the AM fungal recruitment of the crop. By contrast, the increase of shared VTXs observed at GS90 due to inoculation, allows to highlight that indigenous AM fungal inoculants can reduce the variability in host preference among genotypes. Similarly, inoculating the same AM fungal consortium in wheat, the percentange of shared VTXs at maturity was higher than in the controls (46% vs 35%) (Marrassini et al. 2024). This supports the idea that indigenous taxa contained in the inoculum, during crop growth, colonize a larger number of plant genoypes and that our consortium was composed by high-compatible taxa.
Focusing on the pattern of AM fungal composition in inoculated and not-inoculated barley genotypes, the average percentage of shared VTXs was 51% and 65% at GS14 and GS90, while the percentage of VTXs uniquelly retrieved in inoculated plants was 23% and 18%, respectively. Thus, we can not fully confirm our hypothesis that inoculation with a local AM fungal consortium did not modify the composition of the AM fungal communities. However, we can assume that when indigenous inoculants are applied they start to compete with closely related AM fungal species present in the soil at very low abundance, and thus not detectable by the molecular tools. Indeed, when individuals are rare in a microbial community, Illumina sequencing has some limitations in the accurancy of detection (Cheng et al. 2023; Egan et al. 2018). Moreover, the VTXs uniquelly found in the inoculated barley genotypes had a very low abundance (0.6%).
Our findings are in accordance with previous results obtained on bread wheat genotypes and in sunflower, inoculated with the same AM fungal consortium in similar climatic conditions (Arcidiacono et al. 2024; Marrassini et al. 2024). Therefore, our data supports the low environmental and ecological impact of local AM fungal inoculants. Thus, biofertilization with indigenous AMF should be considered a suitable practice for the management of cropping systems.
Looking at AM fungal community structures in the roots of barley genotypes in the PCO plot and in the shade plot (Fig. 7a,b), we can observe the significant interaction between G, Inoc and GS. This is supported by the results of PERMANOVA that highlighted a high percentage of explained variance of the third order interaction (24%) (Table S9). Overall, at GS14, no differences were recorded among not-inoculated barley genotypes, while significant differences were recorded between inoculated Atlante and Atomo/Concerto (Table S10). At GS90, no differences were recorded among not-inoculated barley genotypes and neither among inoculated genotypes. Therefore, at early and late plant development and under no inoculation, host preference in AMF is more discriminated by the pattern of composition than by the pattern of the community structure. The same occured at plant maturity in inoculated genotypes. By contrast, at early plant development, in inoculated conditions, host preference is discriminated by both composition and community structure. These findings highlight that to study host preference in managed field crops it is important to take into account these two aspects during the growing season.
Moreover, inoculation determined at GS14 changes in AM fungal community structure only in Atlante, while at GS90 in both Atlante and Atomo. This results support the high variability in AM fungal community structure among wheat genotypes (Mao et al. 2014; Marrassini et al., 2024; Pellegrino et al. 2020) and gave new insights in the mechanisms by which diversity within AM fungal populations is enhance or maintain in roots. Indeed, a genotype like Concerto whose AM fungal community structure was similar in inoculated and not-inoculated treatments and AM fungal abundace and yield were promoted, can be considered a suitable variety in biofertilization programs. Previously, other factors, such as soil phosphate availability and host species were identified as main determinant of the root AM fungal structure (Cavagnaro et al., 2005; Ehinger et al., 2009; Eom et al. 2000; Helgason et al. 2002; Vanderkoornhuyse et al. 2003). Finally, the fact that two barley genotypes, Atomo and Atlante, in not-inoculated and inoculated conditions, respectively, showed significant differences in the AM fungal community structure between growth stages, is in accordance with a previous work reporting differences along the plant cycle in bread wheat, oat and barley, inoculated and not with a AMF (Aguilera et al. 2021). The importance of the phenological stage in shaping AM fungal community structure in roots of wheat was also reported in field conditions (Marrassini et al. 2024) and and perennial and annual grasses and non-grasses species (Lingfei et al. 2005). The role of barley phenological stage was underlined also by PERMDISP (Table S9). The higher variability at the early development of barley can be attributed to mechanisms of competition among AMF within a community (Chagnon et al., 2013; Jansa et al. 2008;).
3.4. Modelling barley productivity against AM fungal abundance and structure in roots
Our results showed that there was a significant relationship between the traits of AM fungal abundance in roots and barley productivity (RELATE: ρ=0.587; P=0.001) (Fig. 8) and the BEST analysis allowed us to highlight arbuscules as the best predictor of barley productivity. Therefore, as we hypothesized, under low soil fertility AM fungal colonization traits, such as the percentage of arbuscules, are key and consistent determinants of the productivity of barley. Similarly, in field conditions, arbuscules were positively and strongly correlated with all functional traits of sunflower grown at low and high soil fertility (Arcidiacono et al. 2024). The association between increased AM fungal root colonization and increased barley productivity is consistent with the results of meta-analytic works (Lekberg and Koide 2005; McGonigle 1988; Pellegrino et al., 2015; Treseder, 2013).
Moreover, the agronomic response of barley was driven also by changes in the structure of the AM fungal community, as highlighted in Fig. S3 and by the significance of RELATE analysis (ρ=0.514; P=0.005) (Fig. 8c), and not by changes of composition (ρ=0.084; P=0.167; data not shown). This relationship is clear comparing the pattern of AM fungal community structure in roots of barley genotypes at GS90 displayed by the nMDS plot (Fig. S3a) and the pattern of yield and nutrient uptake displayed by the PCO biplot (Fig S. 3b). Therefore, we confirmed our hypothesis that changes in AM fungal community structures and not in composition determine barley productivity. Moreover, the BEST analysis highlighted the best predictors of barley productivity (ρ=0.44, P=0.046) that was Glomus sp. VTX00342, considering one descriptor (r=0.333), or the taxa VTX00342 with Septoglomus sp. VTX00064, considering two descriptors (r=0.406) (Fig. 8d). These BEST predictors were also highlighted in the nMDS plot (Fig. S3a). Indeed, Glomus sp. VTX00342 and Septoglomus sp. VTX00064, putative members of the indigenous AM fungal inoculum, were more present in inoculated than not-inoculated plots (+46% and +56%, respectively). Previously, changes in AM fungal community structure and not in composition were identified as main driver of crop productivity under inoculation with indigenous AMF (Arcidiacono et al., 2024; Marrassini et al., 2024; Pellegrino et al. 2022). Indeed, the inoculated indigenous strains, highly compatible with all barley genotypes, as well as with the local environmental conditions, are preferentially recruited by the roots.