Mycorrhiza‐mediated interference between cover crop and weed in organic winter cereal agroecosystems: The mycorrhizal colonization intensity indicator

Abstract The mycorrhizal fungi are symbiotic organisms able to provide many benefits to crop production by supplying a set of ecosystem functions. A recent ecological approach based on the ability of the fungi community to influence plant–plant interactions by extraradical mycelium development may be applied to diversified, herbaceous agroecosystems. Our hypothesis is that the introduction of a winter cereal cover crop (CC) as arbuscular mycorrhizal fungi (AMF)–host plant in an organic rotation can boosts the AMF colonization of the other plants, influencing crop–weed interference. In a 4‐years organic rotation, the effect of two winter cereal CC, rye and spelt, on weed density and AMF colonization was evaluated. The AMF extraradical mycelium on CC and weeds roots was observed by scanning electron microscopy analysis. By joining data of plant density and mycorrhization, we built the mycorrhizal colonization intensity of the Agroecosystem indicator (MA%). Both the CC were colonized by soil AMF, being the mycorrhizal colonization intensity (M%) affected by environmental conditions. Under CC, the weed density was reduced, due to the increase of the reciprocal competition in favor of CC, which benefited from mycorrhizal colonization and promoted the development of AMF extraradical mycelium. Even though non‐host plants, some weed species showed an increased mycorrhizal colonization in presence of CC respect to the control. Under intense rainfall, the MA% was less sensitive to the CC introduction. On the opposite, under highly competitive conditions, both the CC boosted significantly the mycorrhization of coexistent plants in the agroecosystem. The proposed indicator measured the agroecological service provided by the considered CCs in promoting or inhibiting the overall AMF colonization of the studied agroecosystems, as affected by weed selection and growth: It informs about agroecosystem resilience and may be profitably applied to indicate the extent of the linkage of specific crop traits to agroecosystem services, contributing to further develop the functional biodiversity theory.


| INTRODUC TI ON
In soil, the mycorrhizal fungi are key functional group, able to support the ecosystem services by activating beneficial symbiotic associations with many plant species (Jeffries, Gianinazzi, Perretto, Tournau, & Barea, 2003). At the same time, the high accessibility of the mycorrhizal fungi to soil resources through external hyphae development allows them to benefit of the photosynthetic carbon produced by the host plants through the fungal symbiosis (Smith & Read, 2008).
Referring to natural systems, a "social" role was attributed to the mycorrhizal mycelium network in facilitating and influencing plant organisms interactions, by affecting seedling establishment, altering plant-plant interactions, supplying and recycling nutrients (van der Heijden & Horton, 2009;Simard & Durall, 2004): This network development is mainly associated with the ectomycorrhizal fungi colonization in natural ecosystems (Simard & Durall, 2004).
The AMF, as endomycorrhizal fungi, predominantly colonize the host plant roots at first, by forming internal structures such as arbuscules, vesicles (Figure 1a), intra-and interhyphae, and then growing out to develop a complex, ramified extraradical hyphal network into the surrounding soil (Figure 1b), which can reach up to 30 m of fungal hyphae per gram of soil (Dai et al., 2013;Gianinazzi et al., 2010).
The influence of this AMF extraradical mycelia on the underground biogeochemical cycling and the composition of plant communities, as well as its role as ecological service provider, are widely recognized, particularly in organic cropping systems: root morphology modification, increasing mineral nutrient and water uptake, buffering effect against abiotic stress, protecting against root pathogens (Gianinazzi et al., 2010;Gosling, Hodge, Goodlass, & Bending, 2006;Leake et al., 2004).
Recently, by exploring the function of root exudates in modulating the development of plants, an interesting theory was formulated on the role of mycorrhizal network in transferring plant allelochemicals. These compounds influence the germination, growth, survival, and reproduction among plants (Putnam, 1988;Singh, Batish, & Kohli, 2010) from donors to target plants through extension of a bioactive zone, defined as "fungal fast lane" (Allen, 2007;Barto et al., 2011;Giovannetti, Sbrana, Avio, & Strani, 2004). This privileged transfer, promoted by water diffusion through mycorrhizal hyphae surface, or by the fungal hyphae through internal cytoplasmic flow, leads to accumulation of allelochemicals within the crop-weed rhizosphere at levels that could not be reached by the mere diffusion through the bulk soil (Achatz, Morris, Müller, Hilker, & Rillig, 2014).
At ecological level, this fascinating theory gives a key role to AMF extraradical mycelium, as a means of potential long-distance communication among coexistent plants for reaching space, water, and nutrient resources, influencing their reciprocal inhibition or growth (Barto et al., 2011).
Other recent in field studies demonstrated that, under organic management, the intercropped CC increased the mycorrhizal colonization of the cash crop, because of positive rhizosphere interactions among diversified plants in reduced volume of soil (Trinchera et al., 2016).
In addition, the ability of the rye to increase the AMF colonization in herbaceous system was already observed (Lehman et al., 2012;White & Wei, 2009). However, the relationship between AMF colonization and interference phenomena is not completely clarified and further studies in field are needed to evaluate the effect of CCs introduction on mycorrhizal colonization in conservative agroecosystems (Javaid, 2008;Javaid & Riaz, 2008;Jung et al., 2012;Khanh, Chung, Xuan, & Tawata, 2005;Lehman et al., 2012;Veiga et al., 2011).
On the assumption that soil-fungi-plant relationship is boosted by the organic management, we hypothesized that, where rye and spelt are introduced as winter cereal CCs in an organic rotation, the overall mycorrhizal colonization of the agroecosystem increases by extraradical mycelium development, thus affecting weed emergence, density, and species selection. To verify our hypothesis, a new indicator was developed, linearly correlated to the density of coexistent plant species and their mycorrhization (Duelli & Obrist, 2003), to assess the ecological service provided by the CC on the belowground functional biodiversity under organic management. In a 4-year organic rotation, our 2 years experiment consisted in a randomized block design (RBD), with three adjacent blocks (linear gradient). Two autumn-winter cereals, spelt (T. dicoccum L.) and rye (S. cereale L.) were used as CCs, compared to a control treatment without CCs (CNT). Each treatment (CNT, spelt and rye) was randomly replicated once in the three blocks (in total: nine plots). The same experimental design was repeated in 2014 and 2015 in two adjacent fields within the MOVELTE site.

| Site description and experimental design
Cover crop was sown at the same rate of 250 kg/ha in both the first (2013) and second (2014) year. During the CC cycles, no weeding was carried out in plot treatments. In both the years of the trial, CCs were terminated in May, before the next vegetable crop of the rotation (Campanelli & Canali, 2012).

| Weeds and cover crops density
The five most abundant and representative weed species among the treatments were considered as reference weeds of the studied agro- In addition, the weed relative abundance was calculated as the ratio between the number of considered weeds and the total weeds, per surface unit (RA%).

| Mycorrhizal colonization intensity (M%)
To quantify the root mycorrhizal colonization intensity (M%) of each plant species, in both 2014 and 2015, at 175 DAS, the root apparatus of CC and weeds was sampled from field by using stainless steel cylinders of 6 cm diameter and 20 cm length (Trinchera et al., 2016). Per each CC and weed species, three root subsamples per plot were collected, then pooled to obtain n. 1 root sample × n. 3 treatments × n.
3 blocks (nine records/plant species). Roots were immediately separated from the soil by washing in distilled water in a sieve of 0.5 mm mesh and then ordinated into first-, second-, and third-order lateral roots for further analyses.
At random, from each pooled root sample of CC and weed, a total of 10 × 1 cm root pieces per plant (third-order lateral roots, diameter <2 mm) were cut from 5 to 15 mm from the root tip by a razor blade.
The root fragments were stained using a solution of 0.05% w/v methyl blue in lacto-glycerol (1:1:1 lactic acid, glycerol, and water) for 1 min and destained by distilled water for 1 min more (Grace & Stribely, 1991). Then, the root fragments were placed on grinded slides, mounted in a drop of glycerol, and observed under a light microscope (Nikon E100). The mycorrhizal colonization intensity of CC (M CCi %) and weed (M WEEDi %) roots was assessed by applying the method of Trouvelot, Kouch, and Gianinazzi-Pearson (1986), based on the observation of the root fragments occupied by AMF structures. For each "i" plant species and treatment, the total number of observed fragments was three plants × three blocks × 10 = 90. M i % was calculated attributing to each root fragment increasing scores from 0 to 5: 0 = no AMF structures within the root segment; 1 = structures occupy <1% of the root segment; 2 = structures occupy <10% of the root segment; 3 = structures occupy <50% of the root segment; 4 = structures occupy more than 50% of the root segment; 5 = structures occupy more than 90% of the root segment.
The M i (as %) for each "i" plant species was calculated as follows: where n5 is the number of fragments rated 5, n4 is the number of fragments rated 4, and so on.
Quantitative data on the CC and weeds AMF arbuscular richness (here not reported) were collected to verify if the observed external mycelium on roots was due to AMF colonization, in particular in those weeds usually recognized as non-host plants (as in R. crispus, S. media and P. aviculare).

| Scanning electron microscopy (SEM) of extraradical hyphae on mycorrhizal roots
To obtain a visual evidence of the development of AMF extraradical hy-

| Mycorrhizal colonization intensity of the agroecosystem (MA%)
The MA% indicator is defined as the mycorrhizal colonization intensity of the agroecosystem. When compared to the treatment without CCs (here, the CNT), it measures the agroecological service supplied by the CC introduction, in terms of increased mycorrhization of coexistent plants in the agroecosystem, thus affecting the plant-plant interference. It was obtained by weighting the contribution of each plant species, whose density derives from the field interference, to its mycorrhization. In-field and in-lab data were aggregated, namely the total density of all plant species (CC and weeds), the specific density of the five most representative weed species, and the root mycorrhizal colonization intensity of CC and weeds, as result of the plant-plant interspecies rhizosphere interaction within each designed agroecosystems.
The contribution of each plant species (CC or weed) to the mycorrhization of the agroecosystem was calculated as (M i × D i ), assuming 100% the sum of contributions of considered plant species.
The M i × D i was used as an endpoint: The higher it was for the "i" plant species, the higher was the "i" plant contribution to the AMF colonization of the agroecosystem.
As a quantitative aggregate function, the MA, expressed as percentage, weights the contribution of each plant species to the mycorrhizal colonization of the whole agroecosystem. This agroecological indicator is referred to the surface soil unit (i.e., per m 2 ), calculated in accordance to the below-reported function: where: M WEEDi is the mycorrhizal colonization intensity of the "i" weed, in %; D WEEDi is the specific "i" weed density (N plant /m 2 ); M CCi is the mycorrhizal colonization intensity of the "i" CC, in %; D CCi is the density of the "i" CC (N plant /m 2 ); D WEED-TOT is the total weed density (N plant /m 2 ).
Evidently, the higher is the number of considered plant species, the more the MA indicator can accurately describe the mycorrhization of the agroecosystem.

| Weeds and cover crops sampling and density
In  (Table 1).
In 2014, the RA% under rye was significantly lower respect to that in CNT and under spelt, while in 2015, no significant differences among the treatments were recorded (Table 1).
Although at different order of magnitude, in 2014 and 2015 the specific D WEEDi , calculated for each weed species, was affected by the CC (

| Mycorrhizal colonization intensity of CC and weed (M%)
In 2014, the M% of the spelt was significantly higher than that of the rye, while in 2015 no significant differences were recorded between the CC (Figure 2).
The M weed % was considered only when intraradical structures, such as interhyphae, coils, and vesicles, were recognized, since the P. aviculare, the R. crispus, and the S. media are known as non-host endomycorrhizal plant species: These structures were rarely observed in P. aviculare, while they were found more frequently in S. media and R. crispus roots (Figure 3).
In both the years, the M i % was different among considered weed species, being significantly affected by the CC (Table 3).

| SEM analysis of extraradical hyphae on mycorrhizal roots
The

| D ISCUSS I ON
Mycorrhizal symbiosis is already recognized as an effective strategy played by AMF plants for overcoming numerous biological (soil microorganisms and fungi biodiversity, allelopathic interaction, etc.) and environmental factors (climatic conditions, competition for light, water, and nutrients, etc.), variable in time and space (Afzal et al., 2000;Baum et al., 2015;Gosling et al., 2006;Leake et al., 2004). In our experimental cover cropped agroecosystem, we verified that this strategy was profitably used by both CC and weeds to obtain a mutual ecological advantage, especially in unfavorable environmental conditions.
Given the effect of CC on plant density and mycorrhization data, the yearly temperatures and precipitations had a key role in modulating CC-weed interference. In 2014, under the sudden and heavy rainfall and in absence of CC (unweeded CNT), weeds had to sustain a reduced competition for water and nutrient, while in 2015 the lowest average rainfalls, with the most abundant precipitation recorded in February (just before the plant sampling date) favored the selection of V. persica and the A. arvensis (Craine & Dybzinski, 2013). It was noticed that, in both the years, the highest AMF colonization of the A. arvensis was positively correlated to its highest density in field: this suggests that A. arvensis has been benefited respect to the other weed species by the mycorrhizal colonization, a profitable strategy for optimizing water and nutrient uptake under competition (Allen, Swenson, Querejeta, Egerton-Warburton, & Treseder, 2003;Marschner & Dell, 1994).
Rye has been able to reduce weeds by homogeneously covering the soil, as confirmed by the highest D CC , probably exploiting also its recognized allelopathic properties (Barnes et al., 1987;Belz, 2007;Ciaccia et al., 2015;De Albuquerque et al., 2010). However, the lowest D WEED-TOT and RA% suggest that the rye was strongly active, but not selective in containing weed (Cheng & Cheng, 2015;Tabaglio, Marocco, & Schulz, 2013). By comparing each specific weed density to the corresponding mycorrhization intensity, the relationship between weed selection and mycorrhization emerged again under rye: in 2014 the S. media, and in 2015 the A. arvensis, were the most abundant weed species (25% of the total), with a corresponding increase of mycorrhization of +9.4% in S. media and +30% in A. arvensis respect to those recorded in the CNT. As a matter of fact, the Anagallis genus is characterized by strong allelopathic potential, particularly on gramineous plants, such as millet or wheat: this property, together with the highest weed mycorrhization recorded in field, explain its predominance in such competitive agroecosystem (Rebaz, Shaukat, & Siddiqui, 2001). The other species (i.e., S. media, P. aviculare and R. crispus), which are generally considered as non-host endomycorrhizal plants (Ronikier & Mleczko, 2006), showed in our systems intraradical structures usually formed by AMF, such as inter-radical hyphae, coils and vesicles, rarely in P. aviculare, sometimes in R. crispus and more frequently in S. media. The near absence of R. crispus in the rye treatment, regardless of environmental conditions, suggests that its lacking mycorrhization made it sensitive to the allelochemicals exuded by rye (La Hovary et al., 2016). A previous in vitro test on R. crispus seeds showed their high sensitiveness to the allelopathic activity of rye, which was not only able to reduce its rootlet elongation, but also to inhibit root fungi colonization at emergence (Trinchera, Testani, Ciaccia, Tittarelli, & Canali, 2015). The spelt treatment consistently contained weeds, due to its ability to advantageously compete for water, nutrient, and soil niches under rainy conditions (Gross et al., 2010). Given the highest RA% recorded under spelt, particularly in 2014, it was the more selective CC in inhibiting weed growth: Actually, the net prevalence of some weed species found in both the years (S. media and A. arvensis in 2014, while V. persica and A. arvensis in 2015) and the quite absence of the others (P. aviculare and R. crispus) allowed to infer that the spelt promoted specific spelt-weed association for optimizing the resource uptake (Allen & Allen, 1984). In both the years, the highest spelt M% corresponded to the higher mycorrhization of the coexistent weeds, compared to that recorded in the CNT: this is another evidence of the synergic, and not antagonistic, CC/weed interaction, mediated by root mycorrhization.  Indeed, the agroecological service provided by agroecosystem mycorrhization depends on the ability of the mycorrhizal plants, such as the winter cereal CCs, to support the diversity and abundance of agronomically beneficial AMF taxa, optimize the ecosystem resources, and mediate positively the weed population dynamics through the AMF-CC-weed interaction (Vatovec, Jordan, & Huard, 2005).
The proposed MA% indicator would quantify the above-described ecological service at the scale of the considered agroecosystems. In 2014, under less competitive environmental conditions, the CCs did not significantly affect the overall mycorrhizal colonization of the agroecosystem compared to the CNT. Conversely in 2015, due to increased mycorrhization of specific weeds, also offset by a corresponding increase of the CC contribution to AMF colonization, the MA% increased noticeably in both the cover cropped systems respect to the uncovered one.
To explain the sensitiveness of the system cover cropped with rye to environmental conditions, we can refer to its allelopathic properties. It was verified, in field and in vitro, that the presence of mycorrhizal mycelium on plant roots determined the increase of allelopathic compounds transfer, resulting in a reduced growth of target plants (Achatz et al., 2014). In our CC system, this preferential transport of allelochemicals through the AMF extraradical hyphae as belowground "highways" may be claimed as a strategy used by the rye to contain weeds (Barto et al., 2011). Our results demonstrated the eco-social role of root mycorrhizal colonization in mediating CC-weed interaction, although affected by the climatic conditions and the CC. Given the dependence of plant species density from its mycorrhization in field, the mycorrhization appears as an effective strategy played by the CC to contain weed, and by the weed to compete with CC. Finally, our study assessed the extent of the linkage of specific crop traits to agroecosystem services, thus providing potential guidance to exploit suitable biodiversity-based management options at field and farm scale, contributing to further develop the functional biodiversity theory (Bàrberi, 2015;Costanzo & Bàrberi, 2014).

ACK N OWLED G M ENTS
Special thanks are due to Marco Renzaglia, Fabio Fusari, and Andrea Marcucci for their support in sampling and analytical activities. Very special thanks are due to Francesco Riva for his support in divulgation and participative approach within the project. This study has been carried out in the frame of the RizoSem research project: "Study of rhizosphere interaction and interference among crop and weeds in organic horticultural systems" (financed by the "Organic farming" Office, DG-PQA V-Italian Ministry of Agriculture, Prot. Mipaaf N. 9944, 23/05/2013).

CO N FLI C T O F I NTE R E S T
None declared.

DATA ACCE SS I B I LIT Y
The data will be archived in dryad (