We used 10 strains of R. irregularis (Schenck and Smith). Specifically six homokaryotic strains: Cuba8 (DAOM 984909), 330 (DAOM229455), 66 (DAOM240720), 197198 (DAOM197198), 101 (DAOM240448), and 98 (DAOM240446) and four dikaryotic strains: A4 (DAOM664343), A5 (DAOM664344), SL1 (DAOM240409), and G1 (DAOM970895), which were recently analyzed for nuclear dynamics (Kokkoris et al., 2021). The strains were obtained from the Canadian Collection of Arbuscular Mycorrhizal Fungi (CCAMF), previously known as Glomeromycota In vitro Collection—(GINCO) and from the dikaryotic AMF collection—Corradi Lab at University of Ottawa. All the strains were grown and maintained in vitro with Ri T-DNA-transformed roots of Daucus carota cv. P68, growing in M-medium (Bécard and Fortin, 1988) solidified with 3% Phytagel (Sigma-Aldrich, St. Louis, MO, United States) in dual compartment Petri dishes. These strains were selected based on their MAT-loci in order to account for the MAT variation encountered in the dikaryons and their phylogenetic differences—i.e., they cluster in three distinct clades as reported in Ropars et al. (2016); Savary et al. (2018), and Kokkoris et al. (2021) (Supplementary Table 1).

Spores were extracted from the monoxenic in vitro cultures after dissolving pieces of the spore-containing medium, using sodium citrate buffer (pH 6.0; at 30°C) (Doner et al., 1991). Spores were then stored at 4°C for 20 days to break the spore dormancy and reduce spore mortality (Juge et al., 2002) before initiating the experiments.

Experiment 1: Asymbiotic trait variation between homokaryotic and dikaryotic strains of R. irregularis.

We followed a similar approach as used in Kokkoris et al. (2019a). Briefly, we chose 40 healthy-looking spores per strain (n = 400), with intact spore walls and no discoloration. Spores were imbedded individually in the center of 60 mm × 15 mm Petri dishes filled with M-medium (Bécard and Fortin, 1988) solidified with 3% Phytagel (Sigma-Aldrich, St. Louis, MO, United States). The Petri dishes were sealed with parafilm and incubated in a growth chamber (Precision, Thermo Fisher Scientific, Waltham, MA, United States) at 26°C in the dark.

We examined multiple LHTs during the AMF asymbiotic growth stage (Table 1). Initially, we used a dissecting microscope and the software Infinity Capture (Lumenera^® software, Ottawa, ON, Canada) to take pictures of all plated spores. The spore area and the spore diameter were evaluated in each image by using the oval selection and brush selection tools, and the straight-line tool in FIJI—ImageJ v. 1.53c (Schindelin et al., 2012).

Each Petri dish was examined every 2 days for germination using a dissecting microscope ZEISS Stemi 305 (Toronto, ON, Canada). With germination as a starting point, each spore was left to grow for 30 additional days, and the number of germination tubes, hyphal reach (HR; μm), total hyphal growth (μm), number of septae, and number of tips were quantified at the end of this period using microphotography and image processing. In detail, the number of germination tubes and number of septae were manually counted using a Nikon Eclipse 800 microscope/Nikon digital sight microscopy camera (Minato City, Tokyo, Japan) (DS-Ri2) and the software NIS-Elements BR version 4.60. We used the “add trace” tool (Neuron J Plugin in FIJI—ImageJ v. 1.53c) (Schindelin et al., 2012) on the acquired images, which fits a line to the hyphae semiautomatically, to measure the total hyphal length. We used the “add line tool” in FIJI—ImageJ v. 1.53c to measure the HR. Images were stitched together in Adobe Photoshop version: 22.1.1 (San Jose, CA, United States) when necessary to quantify the total hyphal length and the HR. To standardize the number of septae between the strains, we divided the total number of septae by the total hyphal length to calculate the number of septae per hyphal length. Finally, the branching intensity was calculated by dividing the number of hyphal tips by the total hyphal length.

Experiment 2: Symbiotic trait variation between homokaryotic and dikaryotic strains of R. irregularis.

To eliminate external confounding effects (e.g., microbial interactions, local abiotic variation, etc.) commonly affecting field and in-planta studies, and thus to maximize detection probability of genetic effects, all AMF strains investigated here were propagated under stable growth conditions with receiver operating characteristics (ROCs). Despite its artificial nature (Kokkoris and Hart, 2019a), AMF culturing on ROC is the best available system to obtain unbiased information on LHTs of AMF—since hyphal growth and network complexities are otherwise hidden and concealed within soil environments and cannot be easily observed and measured (Koch et al., 2004; Azcon-Aguilar et al., 2017). Furthermore, the sterility of the system limits unwanted interactions from exogenous factors, such as bacteria in sandwich systems in Petri plates filled with sterile grit (e.g., see Pepe et al., 2017; Sbrana et al., 2020).

We used ROCs of three plant species that have been shown in the past to trigger responses in the nucleotype ratios of the dikaryotic strains (Kokkoris et al., 2021). The species used were Daucus carota (Carrot) cv. P68, Cichorium intybus (Chicory), and Nicotiana benthamiana (Nicotiana) cv. benthamiana.

Each experimental unit consisted of a dual-compartment Petri dish with the “root compartment” containing an individual ROC and the AMF spores, and the “fungal compartment” purposed to hold only the growing fungal tissue (hyphae and spores). The two compartments were united with a sterile filter paper bridge to facilitate hyphal crossing to the hyphal compartment. Seven replicates were plated for each of the 27 plant host × AMF strain combinations (n = 189, i.e., three ROCs and nine R. irregularis strains).

Both split-plate compartments were filled with M-medium (Bécard and Fortin, 1988) solidified with 3% Phytagel (Sigma-Aldrich, St. Louis, MO, United States). The root compartment was supplemented with sucrose (10 g/L) to allow for root growth (Bécard and Fortin, 1988), whereas the fungal compartment was left sucrose free to promote sporulation. We standardized the propagule density to 10 viable spores per strain based on the spore germination ability of each strain as evaluated from our first experiment (see result section for Experiment 1 and Supplementary Table 1). Although the germination conditions for Experiment 2 were not the same as in our first experiment (presence of ROC host), and while the presence of a host has been demonstrated to stimulate spore germination and to promote hyphal growth (Bécard and Piché, 1989; Bécard et al., 1992), the low number of propagules per strain (10 viable propagules) used in our experiment aimed to dilute any potential host effect on germination. In accordance with the first experiment, spores with intact spore walls and no discoloration were chosen. The initial ROC fragments were standardized based on length and age, and we specifically used young root tips of 3 cm in length for all experimental units to reduce any interactive effects (Powell, 1976; Bécard and Fortin, 1988). Spores and roots were plated at the same time and positioned at the far end of the root compartment, furthest away from the paper bridge between the compartments. Similar to our first experiment, the plates were incubated in a growth chamber (Precision, Thermo Fisher Scientific) at 26°C in the dark. The plant roots were periodically redirected to prevent them from growing into the fungal compartment. Contaminated plates or plates where the AMF failed to germinate or colonize the host were discarded.

We examined multiple LHTs during the symbiotic growth stage (Table 1). The number of days until emergence of the extraradical mycelium (ERM) was recorded manually after daily observation of the Petri plates using a dissecting microscope (ZEISS Stemi 305). Time to sporulation (in days) was also assessed through daily culture observation.

To assess the hyphal exploration speed, semicircles (1 cm intervals) were drawn on the hyphal compartment side of each Petri plate lid. Starting from the time point, the hyphae crossed the filter paper bridge to the fungal compartment, we recorded the time (in days) required for the hyphae to reach each of the three semicircles. In the end, we calculated the average time it took for the hyphae to reach as far as 1 cm (see Supplementary Figure 1 for method details).

The total hyphal length was quantified 30 days after the hyphae entered the hyphal compartment. We used a microscope Zeiss AxioZoom.V16 (Toronto, ON, Canada) and the function “panorama” in the ZEN 2.3 software (Carl Zeiss MicroImaging, Göttingen, Germany) to acquire complete stitched images of the entire fungal compartment. The hyphal length in the entire fungal compartment was traced semi-automatically using the “add trace” tool (Neuron J Plugin in FIJI—ImageJ v. 1.53c) (Schindelin et al., 2012).

The total number of spores was quantified manually in the entire hyphal compartment using the function “cell counter” in FIJI—ImageJ v. 1.53c (Schindelin et al., 2012). The spores that were produced from the strains growing with tobacco host plant were quantified in the root compartment since none of the strains produced spores in the fungal compartment.

To quantify branching and hyphal network density, we used the software RhizoVision Analyzer (version 2.0.0 beta© 2018–2020 Noble Research Institute, LLC, Ardmore, OK, United States) (Seethepalli et al., 2021). For the fungal compartment, we defined three regions of interest (1.5 cm × 1.5 cm)—on the left, center, and right side of the plate—with the parameters “invert image” and “filter noisy components of the image” selected, with “Root type = disconnected root” and by adjusting the imaging threshold level manually to provide the most accurate read. This analysis provides the hyphal branching value and hyphal network area. The “Network Area” of the image is determined by counting the total number of hyphal pixels in the defined region (higher value = denser hyphal network).

To visualize entire mycelia, partial plate images were stitched together after tracing all hyphae in FIJI (NeuronJ) in the fungal compartment. We used Adobe Photoshop version 22.1.1. to extract the hyphal pattern and then we used BioRender (Toronto, ON, Canada) to create figures that demonstrate the actual hyphal growth patterns (Figures 1, 2 and also see Supplementary Figure 2 for the method details and Supplementary Figure 3 for additional original full plate images). We did not create such figures for tobacco since almost no hyphae propagated into the fungal compartment (none for the homokaryons and some hyphae for the dikaryons SL1 and A5).

The AMF intraspecific variability in phenotypic traits is known to exist both at the asymbiotic (Kokkoris et al., 2019a) and symbiotic stages (Hart and Reader, 2002; Koch et al., 2004, 2006; Munkvold et al., 2004; Ehinger et al., 2009). Here, we showed that some of this variability is linked to the nuclear status of R. irregularis, as we found contrasting fitness-related LHTs between homokaryons and dikaryons.

In the asymbiotic stage, homokaryotic spores germinate faster and more frequently compared to dikaryotic spores and thus may get a priority effect in symbiotic associations (Werner and Kiers, 2015). This could be especially important when plants are abundant, and roots are easily encountered. However, dikaryotic strains appear to grow significantly faster once their spores have germinated, producing a wider and more branched hyphal network, ensuring a more efficient exploration of the surrounding area. Therefore, the priority advantage of AMF homokaryons must rapidly decline as dikaryons eventually catch up in colonizing roots.

Dikaryons also outgrew homokaryons during the symbiotic stage. We found that dikaryons establish a larger hyphal network. A wider hyphal network implies a more efficient nutrient uptake for the host since the ERM acquires resources (Jakobsen et al., 1992). The ability to explore larger areas, and subsequently reach more nutrients, suggests that dikaryons could be more beneficial to plant partners than homokaryotic relatives—i.e., they should provide plants a larger access to nutrients. Future studies based on nutrient tracing should investigate whether hosts do actually obtain more nutrients from dikaryotic strains of R. irregularis, and whether they preferentially form symbioses with dikaryons rather than homokaryons when both strains coexist (and particularly when nutrients are scarce).

The slower germination rates and faster post-germination growth and increased exploration capabilities of dikaryons are difficult to interpret. Perhaps, these result from initial difficulties in coordinating inter-nucleus genetic interactions in newly germinating spores. Alternatively, maintaining two nuclei instead of one may initially result in higher metabolic costs for dikaryotic spores (Kokkoris et al., 2021). However, once interactions are faster, owing to fine-tuning, hyphal growth could then proceed at full speed, thanks to complementary proteins produced by the coexisting parental genotypes. Although speculative, this hypothesis is supported by similar mechanisms seen in other heterokaryotic fungi (e.g., Neurospora crassa and Agaricus bisporus), where each nucleotype expresses unique proteins leading to increased fitness for the heterokaryotic strains (Gehrmann et al., 2018).

As dikaryons contain significantly more nuclei than homokaryons (Kokkoris et al., 2021), the presence of nucleophagy—i.e., targeted degradation of nuclei (Kokkoris et al., 2020a)—may also explain the faster growth rate of dikaryons. Indeed, in other fungi, this process is known to create a rapid flux of nutrients (phosphate and nitrate) when needed by the organism (presumably for growth) (Shoji et al., 2010), and it is thus possible that a similar mechanism may provide an advantage to dikaryons of AMF as well.

It was recently shown that AMF dikaryons can change the ratio of parental nuclei in response to hosts, possibly as a result of adaptive genetic interaction with the plant partner (Kokkoris et al., 2021). In support of this, we found that AMF dikaryons consistently produced more spores and more hyphae and a denser hyphal network irrespective of the host plant identity. A higher spore production rate was previously reported for strain A4 (Koch et al., 2004), and higher hyphal density was previously reported for strain C3 (Ehinger et al., 2009), both of which are now known to be dikaryons.

Hyphal density is key for establishment and functioning of the mycorrhizal symbiosis. Thus, our findings indicate that dikaryons may have an advantage over homokaryons when it comes to establishing mycorrhizal connections with different host plants, possibly as a result of their higher genetic diversity (two nucleotypes) and their ability to regulate these nuclear states in response to their plant host (Kokkoris et al., 2021). Within this context, the ability of AMF dikaryons to interact more efficiently across a wider host range suggests dikaryons are being more generalist than homokaryons and may thus be encountered in ecosystems that harbor more diverse plant communities.

The ERM of a single AMF strain can interconnect with numerous distinct plants and form a common mycorrhizal network (CMN) (Giovannetti et al., 2004). CMNs enable resource redistribution among coexisting plants, resulting in higher plant survival, fitness and diversity (Hart et al., 2003), and community composition (Hart et al., 2001; Zobel and Öpik, 2014). The present study suggests that AMF nuclear status may be relevant for hyphal network establishment. Future studies should now examine this critical question by investigating the potential beneficial effects of AMF dikaryons for plant communities, and determine their distribution and frequency in the field and across distinct environmental conditions. Accounting for AMF nuclear status in ecological models could be key for answering questions of mycorrhizal ecology (Kokkoris et al., 2020b).

Our findings also have implications for the fungal bio-inoculum industry. The most commonly used strain for inoculants (R. irregularis DAOM197198—homokaryon) grows fast in carrot ROC and produces large quantities of spores (Douds, 2002; Rosikiewicz et al., 2017). However, there were cases in which it reduced host nutritional benefits (Kokkoris and Hart, 2019a, b; Kokkoris et al., 2019a) and failed to establish in large-scale inoculation trials (Farmer et al., 2007; Kokkoris et al., 2019b; Thomsen et al., 2021). In this study, DAOM197198 grew well with carrot as host plant and produced the most spores among all homokaryons, but surprisingly ranked last in terms of growth and spore production with the other two host plants, indicating host preference. In contrast, the observed ability of the R. irregularis dikaryons to grow better across a wider host range, a trait that could also translate into higher benefits to host plants, suggests that AMF dikaryons are more reliable inocula.

Besides DAOM197198, to our knowledge, all the R. irregularis strains used in inoculants, and the majority of the R. irregularis strains growing in AMF collections (with the exception of the five known dikaryotic R. irregularis strains), are homokaryotic (Kokkoris et al., 2021). With single spore in vitro inoculations being a necessary step for the establishment of a pure strain culture, it is possible that the observed dominance of homokaryons over dikaryons in vitro is due to the increased germination ability of homokaryons.