In early July 2017, 2-year-old potted seedlings (6 ± 1 mm in root collar diameter, 37 ± 3 cm in shoot length) of M. communis (originated from seed) raised in the greenhouse of Faculty of Agriculture, Tarbiat Modares University, Iran (35°44′ N, 51°10′ E, and 1,215 m a. s. l.), were transferred to 5-L plastic pots containing a mixed soil (agricultural soil, green manure + sand + coco peat in a ratio of 1: 1: 1: 2) with physico-chemical properties indicated in Table 1. The green manure consisted of decomposed plant litter originating from various local forest trees, mainly oak and hawthorn species (Quercus spp., Crataegus spp.) and pistachio (Pistacia vera).

The experiment was carried out as a 3 × 7 factorial (water regime × microorganism inoculation) in a completely randomized design with three replicates each comprising four seedlings. The following microorganism inoculation treatments were applied: (1) control (no inoculation), (2) AMF-inoculation with Funneliformis mosseae, (3) AMF-inoculation with Rhizophagus irregularis, (4) combined inoculation of F. mosseae + R. irregularis, (5) PGPR-inoculation with Pseudomonas fluorescens, (6) PGPR-inoculation with Pseudomonas putida, and (7) combined inoculation of P. fluorescens + P. putida. For mycorrhizal fungi 40 g of inoculum (100 propagules g⁻¹ of carrier material) and for Pseudomonas bacteria 15 ml of inoculum containing 10⁷ ml⁻¹ live bacterial cell were used. For the dual AMF inoculation, 40 g of each fungus was used and for the dual PGPR treatment, 15 ml of each bacterium were applied. Inocula were prepared from the Microbial Bank of the Microbiology Department of the Soil and Water Research Institute of Tehran, Iran. Water deficit stress included 100% field capacity (FC100, no stress), 60% FC (FC60, mild stress), and 30% FC (FC30, severe stress). Drought stress was applied via the watering regime following Zarik et al. (2016) for 180 days.

The percentage of AMF-colonized roots was assessed in a complementary study (Azizi et al., 2021). Mycorrhizal colonization in control plants was highest in the FC100 treatment (9.4%) and dropped by nearly 80% under FC30 conditions. By contrast, mycorrhization in singly and dually inoculated plants ranged from 51 to 72.7% under FC100 conditions and only halved in response to increasing water limitation.

For biomass determination, one seedling of each replicate was removed from the soil and after rinsing off the soil attached to the roots, seedlings were separated into their main organs and dried at 70°C for 48 h. Then, a digital scale with an accuracy of 0.0001 g, was used to measure leaf biomass (LB), stem biomass (SB), and root biomass (RB).

At the end of the experiment, the intracellular CO₂ concentration (C_i) was measured using a portable gas exchange device LI-6400 (LiCor Inc., Lincoln, NE, United States). For this purpose, four fully developed, healthy leaves were selected from the top of the crown from each of three seedlings per treatment. Measurements were taken between 9 and 12 am on a sunny day with a light intensity of 1,400 μmol m⁻² s⁻¹ (Dong et al., 2017). To measure the relative water content (RWC), three healthy, fully developed leaves from the top of the crown from each of three seedlings per treatment were sampled. After determination of fresh weight (FW), the leaf samples were placed in distilled water in the dark for 24 h to maximize absorption for turgor weight (TW) determination. Then turgid leaves were placed in an oven at 70°C for 48 h. Leaf weight was measured after drying (DW) and RWC was calculated according to Equation (1) (Yang et al., 2007).

Following the results of photosynthesis and transpiration in our previous study (Azizi et al., 2021), mesophyll conductance (g_m) and water use efficiency (WUE) were determined according to Equations (2, 3) (Moradi-ghahderijani et al., 2017; Biswas et al., 2019).

where, A = photosynthesis, C_i = intracellular CO₂ concentration, and E = transpiration.

At the end of the experiment, a representative seedling per replicate (average height and diameter) was chosen and its root and leaves were washed and placed in an oven (at 70°C) for 48 h until dry weight was reached. For root and foliar nitrogen determination by the Kjeldahl method, 0.5 g powdered sample was weighed and transferred to digestion tubes. One catalyst tablet was added with 10 ml of concentrated sulfuric acid. The tubes were placed in a digestion furnace for 3–4 h at 40°C (until the color of the samples turned blue). After cooling, 10 ml of distilled water was added to each of the tubes prior to titration (Chapman and Pratt, 1962).

To measure phosphorus content a calorimetric method (yellow color of molybdate and vanadate) was utilized. After digestion of 1 g of powdered leaf and root sample, 5 ml of extract was poured into a 25 ml balloon and 5 ml of ammonium molybdate vanadate solution was added. Then, absorbance at 480 nm was read using a spectrophotometer (Lambda 45-UV/Visible, PerkinElmer, Waltham, MA, United States; Chapman and Pratt, 1962).

To measure the concentration of K, Ca, and Mg of roots and leaves, the samples were dried in an oven at 48°C for 48 h (Ribeiro et al., 2002). Then, 1 g of dried and powdered sample material was mixed with 10 ml of concentrated nitric acid and after staying under the hood for 12 h, it was exposed to 80°C for 2 h. The prepared solution after mixing with 3 ml of concentrated perchloric acid was kept at 160°C for 5 h. After cooling, the prepared samples were smoothed with filter paper and reached the 25 ml volume. Then, the content of the above elements was assessed using an atomic absorption spectrometer (Ribeiro et al., 2002).

The data were analyzed using SPSS statistical software version 23 (SPSS Inc., Chicago, IL, United States), and graphs were drawn in Excel software version 2016 (Microsoft Office, 2016). Kolmogorov–Smirnov and Levene tests were used to evaluate the normality and variance homogeneity of the data, respectively. Two-way ANOVA was used to determine the overall significance of the treatments, and Duncan’s new multiple range test was used as a post hoc test to compare differences between group means.

The detrimental effects of drought were evident in dry mass of all organs (Table 2). However, single AMF and PGPR inoculations (partly) compensated and dual inoculations sometimes even overcompensated the negative effects of drought stress compared to the well-watered, non-inoculated control (Figures 1A–C). For instance, in the FC60 treatment, dual inoculation almost always resulted in significantly larger biomass accrual than both the FC60 and the FC100 non-inoculated control. The strongest inoculation effects occurred in the FC30 treatment of leaf biomass, where dual AMF and PGPR more than doubled the values seen in the control (Figure 1A). The stimulation under severe drought was less pronounced in roots, where dual AMF inoculation produced 73% higher biomass than the control, but this was not statistically different from the magnitude seen in single AMF inoculations. Similarly, the dual PGPR inoculation stimulated root biomass significantly by 56%, but this was not significantly different from the increase related to single PGPR inoculations (Figure 1B). Under severe drought, only the single R. irregularis inoculation, the dual AMF and the dual PGPR inoculation caused significant increases in stem biomass compared to the control. The stem biomass values linked to the remaining inoculation treatments were in between the values of the control and the three effective inoculation treatments and therefore did not differ significantly from either (Figure 1C).

There was neither a statistically significant two-way interaction nor a microbial inoculation effect on seedling WUE (Table 2; Figure 2A). However, the severe drought treatment (FC30) produced significantly higher WUE compared to the remaining watering regimes, which showed similar values (Figure 2A). We did not detect a significant water regime × microbial inoculation interaction but both main effects were statistically significant (Table 2). Irrespective of the inoculation treatment, the intracellular CO₂ concentration (C_i) increased significantly with increasing drought stress (Figure 2B). Regardless of the water regime, the dual AMF inoculation induced significantly lower C_i values than the control and single inoculations (Figure 2B). In the FC30 treatment, the dual AMF treatment caused an 11% decrease in C_i compared to non-inoculated control seedlings (Figure 2B). A similar pattern was seen for the dual PGPR inoculation treatment across drought conditions, except for the FC30 group, where seedlings dually inoculated with PGPR showed C_i values that were significantly lower than the control but similar to all singly inoculated seedlings (Figure 2B).

Mesophyll conductance (g_m) declined with increasing soil water deficit, but this effect varied with inoculation treatment resulting in a significant two-way interaction (Table 2; Figure 3A). Dual AMF and PGPR inoculation caused the highest g_m values in well-irrigated (+86% relative to the control) and mildly drought-stressed seedlings (+63%–66% relative to the control; Figure 3A). Under severe water deficit, dual AMF and PGPR inoculation increased g_m by 64 and 57%, respectively, compared to the non-inoculated control (Figure 3A).

Water regime and microbial inoculation also had a significant interactive effect on RWC (Table 2). Dual AMF and PGPR inoculation significantly promoted leaf RWC in well-irrigated seedlings while single inoculations had no effect (Figure 3B). Under FC60 conditions, all AMF inoculations equally improved RWC relative to the control (>20%) and a similar increase was seen in the dual PGPR treatment (Figure 3B). The effects of the single PGPR inoculations were slightly less pronounced but still represented a significant improvement compared to the control (Figure 3B). Under severe drought (FC30), all microbial inoculations significantly increased foliar RWC between 30 and 43% without significant differences among inoculation treatments (Figure 3B).

Except for P, there was no significant interaction between water regime and microbial inoculation for any of the tested root nutrients but both main effects were statistically significant (Table 2). With increasing drought stress, root N, P, K, Ca, and Mg content decreased, but the addition of AMF and PGPR, especially the dual inoculations, significantly improved the nutrient status of seedling roots (Figures 4A–E). Across inoculation treatments, root N declined by 32% in mildly drought-stressed seedlings and by 72% under severe drought compared to well-watered control conditions (Figure 4A). Across water regimes, we observed the same inoculation-induced stimulatory pattern: all single AMF and PGPR inoculations caused similar increases in root N relative to the non-inoculated control (FC100: ca. 16%, FC60: ca. 30%, and FC30: ca. 45%), and the dual inoculations had even larger effects of equal magnitude among AMF and PGPR (FC100: ca. 34%, FC60: ca. 62%, and FC30: ca. 145%; Figure 4A).

Root P also showed a general decline with increasing drought intensity, but remarkably, single AMF and PGPR inoculations failed to improve root P status over the non-inoculated control in the FC100 and FC30 treatments (Figure 4B). Only under mild drought stress (FC60), single microbial inoculations significantly promoted root P uptake by a similar magnitude in comparison to the control. However, across water regimes, dual inoculations resulted in the largest increase in root P, significantly exceeding values seen in singly inoculated and control seedlings (Figure 4B). Root K, Ca, and Mg decreased by more than half with increasing drought intensity (Figures 4C–E). Within water regimes, single AMF and PGPR inoculations mostly stimulated K, Ca, and Mg content significantly, but the dual AMF and PGPR inoculations always had a far greater effect of equal magnitude (Figures 4C–E).

In the most severe drought regime (FC30), the dual AMF or PGPR inoculation increased root N by 151%–138%, P by 176%–181%, K by 114%–112%, Ca by 126%–136%, and Mg by 130%–140% compared to non-inoculated seedlings (Figure 4).

We found no evidence of a statistically significant interaction between water regime and microbial inoculation for any of the measured leaf nutrients apart from Mg (Table 2). However, as main effects, both water regime and microbial inoculation had a significant influence on foliar nutrient concentration. Regardless of the inoculation treatment, increasing water deficit reduced foliar N, P, K Ca, and Mg by at least about 50% (Figures 5A–E). In most cases, single AMF or PGPR inoculation caused a significant increase of similar magnitude in foliar nutrient status compared to non-inoculated controls (Figures 5A–E). However, peak foliar nutrient concentrations were invariably associated with dual AMF and PGPR inoculations (both with the same effect size) and the differences to singly inoculated seedlings and the non-inoculated control were almost always statistically significant (except for leaf N in the FC60 treatment and Mg in the FC100 treatment; Figures 5A–E). In the FC30 treatment, dual AMF and PGPR inoculations promoted N concentration by 102%–107%, P by 143%–149%, K by 83%–84%, Ca by 10%–98%, and Mg by 102%–106% compared to the non-inoculated control (Figures 5A–E).

Foliage, stem, and root biomass decreased with increasing drought stress, but the inoculation with microorganisms, especially the dual AMF and PGPR inoculations fully or at least partially compensated for the negative effects of drought stress when compared to the uninoculated, well-watered control (Figure 1). A reduction in leaf area and thus leaf biomass provides an effective means to curtail water use and is a common adaptation to drought in many plant species (Larcher, 2003). Our companion study showed drought-induced reductions in foliar gas-exchange of the myrtle seedlings that were fully or partially compensated by dual AMF or PGPR inoculation in the FC60 and FC30 treatments, respectively (Azizi et al., 2021). Growth processes have been shown to respond more sensitively to abiotic stresses (such as drought) than photosynthesis and it is therefore likely that the growth reductions seen in the stems and roots of our myrtle seedlings are due to carbon sink limitation rather than a limited supply of photo-assimilates (Palacio et al., 2013). Moreover, lack of soil water at any plant growth stage reduces the absorption, transport, and metabolization of nutrients, impairing carbon storage and dry matter (Hu and Schmidhalter, 2005). A disruption of the root-soil interface, which can already occur under mild drought stress, results in substantial losses in root conductivity triggering stomatal closure prior to the onset of cavitation (Kumar et al., 2010; Rodriguez-Dominguez and Brodribb, 2020). However, AMF-induced changes in root morphology and the vast hyphal-driven expansion of the root system strongly increase root uptake and nutrient transport (Orfanoudakis et al., 2010). PGPR also effectively improve seed germination, accelerate growth in the early stages, induce root formation, and increase the number of root hairs (Heinonsalo et al., 2004). Some rhizobacteria contain the enzyme 1-aminocyclopropane-1-carboxylate (ACC) deaminase that reduces the production of ethylene, which in turn regulates auxin synthesis and allocation, ultimately leading to stimulation of root growth (Xie et al., 1996; Růžička et al., 2007). Both Pseudomonas species used in our study have been shown to produce ACC (Saravanakumar and Samiyappan, 2007; Gamalero et al., 2008) suggesting that the increase in root growth observed in our myrtle seedlings is related to ACC activity. Increasing biomass and growth of various crops (e.g., rice, foxtail millet, and peppermint) resulting from AMF and PGPR soil inoculation during drought conditions have also been reported in agreement with our findings (Ruíz-Sánchez et al., 2011; Gong et al., 2015; Chiappero et al., 2019).

In line with the results reported for Juglans regia and Eucaplytus camaldulensis, WUE remained unaffected by soil microbial inoculation (Liu et al., 2019; Mateus et al., 2021), but increased with increasing drought intensity in our myrtle seedlings (Figure 2). Water deficit affects photosynthesis and transpiration to different degrees, resulting in significant differences in WUE (El Hafid et al., 1998). WUE usually rises under drought stress, but this increase is not associated with enhanced production because the increase is driven by reduced transpiration rather than an increase in photosynthetic carbon assimilation (Blum, 2009). In the present study, g_m decreased with increasing water deficit stress, probably linked to reduced leaf water status and thus tissue hydration (Azizi et al., 2021; Figure 3). Lower rates of photosynthetic CO₂ assimilation in the presence of higher levels of C_i indicate low g_m and impaired carbon metabolism of mesophyll cells (Ratnayaka and Kincaid, 2005). Microbial soil inoculation, particularly dual AMF and PGPR applications, mitigated drought-related reductions in g_m, which was mirrored by lower C_i values indicating greater utilization of absorbed CO₂. Similar findings in terms of C_i and g_m were obtained with Azospirillum brasilense and Bacillus sp. in seedlings of the tropical canopy tree Cariniana estrellensis (Tiepo et al., 2018), and with Rhizophagus irregularis in seedlings of Robinia pseudoacacia (He et al., 2016) under drought conditions.

The RWC of M. communis seedlings decreased with increasing drought stress, but soil inoculation with AMF or PGPR (especially the dual inoculations) largely or fully compensated for these declines in the FC60 treatment (also compared to the well-watered and uninoculated control) and still had a positive, albeit less pronounced, effect in the FC30 treatment (Figure 3). A similar pattern was observed for leaf water potential of these myrtle seedlings published recently as part of the overarching study (Azizi et al., 2021). Cupressus atlantica and tropical tree seedlings also showed increases in leaf RWC in response to AMF or PGPR treatments (Zarik et al., 2016; Tiepo et al., 2018). AMF hyphal networks greatly expand the root surface thus giving plants access to a larger soil volume and because of their small diameter they can grow into the smallest pores and crevices that fine roots could not reach. The symbiosis with AMF allows plants to take up more water and thus maintain favorable water relations over a wider range of environmental conditions and under drought, AMF have been shown to be able to increase hyphal moisture uptake (Augé et al., 2015). PGPR have been demonstrated to improve the water relations of seedlings by producing phytohormones that favorably affect plant water relations such as abscisic acid and auxin (Kothari et al., 1990; Boiero et al., 2007).

Root and leaf N concentration decreased under water deficit conditions, but the soil inoculation treatments (especially dual AMF or PGPR inoculations) partially offset the negative drought effects on plant nutrient dynamics compared the non-inoculated control seedlings (Figure 5). These results are consistent with the studies of Abbaspour et al. (2012) on Pistacia vera seedlings and with those of Armada et al. (2015) on Zea mays. Rhizosphere bacteria increase the rate of nitrate translocation from root to shoot by increasing the amount of cytokinin in the host plant (Flores et al., 2005).

Mycorrhizal fungi also activate glutamine synthetase, arginase, and urease by affecting root physiology and thus increase the N uptake and utilization efficiency of host plants. Arginase and urease are key enzymes in the transfer of N from the mycelium to the roots of the host plant. Several ammonium and nitrate transporters have been identified in the mycelium of AMF (Tisserant et al., 2012) enabling mycelial N uptake in the form of nitrate or ammonium and subsequent conversion to organic compounds by glutamine synthetase (Bago et al., 2001).

In agreement with the findings of Zhang et al. (2019) and Ortiz et al. (2015), root and leaf P content in our myrtle seedlings decreased with increasing drought stress but AMF and PGPR inoculation mitigated this effect (Figure 5). While both single and dual inoculations produced significant positive effects in leaf P, root P responded more strongly to the dual inoculation treatments. This finding implies that multiple AMF or PGPR acting in concert are required to significantly increase P availability and subsequent root uptake while the activity of single AMF or PGPR species seems to be sufficient to steer allocation toward increased P supply to the leaves. Phosphorous plays an important role in many plant physiological processes linked to energy storage and transfer, photosynthesis, regulation of enzyme activity, and carbohydrate transport and it also affects plant water relations (Waraich et al., 2011). Numerous studies have shown that following AMF inoculation, P uptake in plants increases under stress conditions (Garg and Manchanda, 2008; Bowles et al., 2016), which has been attributed to the secretion of organic acids and phosphatase enzymes solubilizing inorganic P from soil minerals and mineralizing organic P sources (Subramanian et al., 2006).

Water deficit stress negatively affected root and leaf K of M. communis seedlings, but the inoculation treatments, especially the dual AMF and PGPR inoculations increased K uptake in both organs (Figure 5). Potassium is crucial for turgor control and thus for cell expansion during growth and guard cell osmoregulation (stomatal control), not to mention its key role in maintaining plasma membrane potential. The amount of plant-available K depends on the K content in the soil solution and the level of exchangeable K (Haby et al., 1990). Both K sources may increase in the presence of AMF and PGPR through organic acid-mediated silicate weathering and mineral dissolution causing the release of K thus allowing plants to increase their uptake (Campanelli et al., 2013).

Calcium plays a vital role in the regulation of many physiological processes in plants, thereby affecting growth processes and responses to environmental stresses. For example, the movement of water and dissolved mineral salts is affected by Ca through its influence on membrane structure and stomatal function, cell division, and cell wall construction (Hu and Schmidhalter, 2005). However, under water deficit stress, reduced water uptake and curtailed transpiration result in a decrease in leaf Ca content (Maksimović et al., 2003). In line with the findings of Dominguez-Nuñez et al. (2014) on Pinus halepensis and Meddich et al. (2015) on Phoenix dactylifera, we observed AMF- and PGPR-induced increases in root and foliar Ca contents of myrtle seedlings (Figure 5). Apart from increases in Ca availability resulting from the exploitation of a larger soil volume through AMF hyphae and accelerated mineral weathering linked to PGPR secretions, the positive inoculation effect on Ca levels in myrtle seedlings may in part be due to increases in stomatal conductance and transpiration given the role of xylem sap flow in Ca mobility (Ruiz-Lozano et al., 1995; Wu and Xia, 2006; Garg and Bhandari, 2016; Azizi et al., 2021).

A very similar pattern of drought-induced decrease was observed for Mg contents in roots and leaves of Myrtus seedlings exposed to drought (Figure 5). Under water deficit stress, single and dual inoculations with AMF or PGPR increased the Mg content in seedling roots and foliage which is consistent with the results of Danielsen and Polle (2014) on Populus × canescens and Armada et al. (2015) on Zea mays. The increase in Mg uptake is probably due to the hyphal expansion of the root system and consequently enhanced uptake of this element by the plant in the case of AMF and likely results from microbial solubilization of Mg-bearing carbonates and minerals in the PGPR treatments (Smith and Read, 2008; Evelin et al., 2012).

Our study revealed that, depending on drought intensity, AMF or PGPR inoculation can largely or at least partially offset the detrimental effects of drought on biomass production, water relations, and nutrient dynamics of M. communis seedlings. Especially the dual inoculations proved to be very potent suggesting even greater benefits from inoculations including multiple AMF and PGPR species. These findings motivate further research testing the effects of combined AMF and PGPR inoculations and linked to this, determining the composition of the microbial consortium that optimally supports myrtle health and performance. Most importantly though, our results highlight soil inoculations with beneficial microorganisms as a cost-effective, easy-to-use tool to promote drought resistance of myrtle. Such readily applicable approaches are urgently needed in support of conservation initiatives geared toward restoring natural myrtle populations and habitats. At the same time, such methods can be used to refine operational nursery practices to help future-proof myrtle cultivation.