Mycorrhizal networks: common goods of plants shared under unequal terms of trade.

Plants commonly live in a symbiotic association with arbuscular mycorrhizal fungi (AMF). They invest photosynthetic products to feed their fungal partners, which, in return, provide mineral nutrients foraged in the soil by their intricate hyphal networks. Intriguingly, AMF can link neighboring plants, forming common mycorrhizal networks (CMNs). What are the terms of trade in such CMNs between plants and their shared fungal partners? To address this question, we set up microcosms containing a pair of test plants, interlinked by a CMN of Glomus intraradices or Glomus mosseae. The plants were flax (Linum usitatissimum; a C(3) plant) and sorghum (Sorghum bicolor; a C(4) plant), which display distinctly different (13)C/(12)C isotope compositions. This allowed us to differentially assess the carbon investment of the two plants into the CMN through stable isotope tracing. In parallel, we determined the plants' "return of investment" (i.e. the acquisition of nutrients via CMN) using (15)N and (33)P as tracers. Depending on the AMF species, we found a strong asymmetry in the terms of trade: flax invested little carbon but gained up to 94% of the nitrogen and phosphorus provided by the CMN, which highly facilitated growth, whereas the neighboring sorghum invested massive amounts of carbon with little return but was barely affected in growth. Overall biomass production in the mixed culture surpassed the mean of the two monocultures. Thus, CMNs may contribute to interplant facilitation and the productivity boosts often found with intercropping compared with conventional monocropping.


INTRODUCTION
Arbuscular mycorrhizal fungi (AMF) inhabit the soils of virtually all terrestrial ecosystems, forming symbiotic associations with most plants (Parniske, 2008;Smith and Read, 2008). The host plants incur substantial carbon costs to sustain this symbiosis (Jakobsen and Rosendahl, 1990), but in return, they obtain multiple benefits from the fungal partners, above all, the provision of mineral nutrients. AMF may supply up to 90% of the host plant's nitrogen and phosphorus requirements (Smith and Read, 2008). Moreover, AMF are important determinants of plant community structure and ecosystem productivity (Grime et al., 1987;van der Heijden et al., 1998), and they represent a crucial asset for sustainable agriculture (Rooney et al., 2009). Typically, AMF exhibit little host-specificity; a single individual may form a common mycorrhizal network (CMN) between several co-existing plant individuals, even from different species (Whitfield, 2007;Smith and Read, 2008;Bever et al., 2010). Such CMNs may be enlarged through hyphal fusion of conspecific AMF (Giovannetti et al., 2004). The The potential role and importance of CMNs is most apparent in the case of mycoheterotrophic plants. These plants connect themselves to an existing CMN to receive both carbon and mineral nutrients (Bidartondo et al., 2002;Courty et al., 2011). There is an ongoing debate whether carbon transfer through CMNs may also occur among autotrophic plants (Bever et al., 2010;Hodge et al., 2010). This is of a certain academic interest, but it may obscure a more general and obvious question arising from recent literature (Hodge et al., 2010;Hammer et al., 2011;Kiers et al., 2011;Smith and Smith, 2011;Fellbaum et al., 2012) What are the terms of trade between plants and their shared fungal partners? Put another way, what is the "investment" of a given plant into a CMN (in the currency of assimilated carbon), and what is the "return of investment" in terms of mineral nutrients provided by the CMN? Indeed, different co-cultivated plants benefit differently from their CMN, depending on the AMF species involved, and these differences significantly affect plant co-existence (Zabinski et al., 2002;van der Heijden et al., 2003;Wagg et al., 2011). However, up to now, the relationship between carbon investment and nutritional benefit of different plants engaged in a CMN has never been assessed.
To address the terms of trade in a CMN experimentally, we established a model system consisting of two plant individuals growing side by side in compartmented microcosms ( Fig. 1). The roots of the plants were confined to their respective "root hyphal compartments" (RHC). In experiments with AMF inoculation, however, the plants were able to connect through CMN in the "hyphal compartment" (HC) or in the "label-hyphal compartment" (LHC). We assessed the carbon investments of the single plants into the CMN through stable isotope tracing. We chose the C 3 -plant flax and the C 4 -plant sorghum for our experiments. Due to the different isotope fractionation during C 3 versus C 4 carbon fixation, these two species display distinctly different carbon isotope ratios (δ 13 C ~33‰ for flax and ~14‰ for sorghum). This difference in 13 C signature of C 3 and C 4 plants has been widely used to track C flows in mycorrhizal symbioses (e.g.: Allen and Allen, 1990;Fitter et al., 1998). The plants were grown either in "monocultures", as pair of identical plant species, or in a "mixed culture", as pair of different plant species. We used two different AMF species in the experiments for inoculation, Glomus intraradices and Glomus mosseae (recently re-named Glomus irregulare (Rhizophagus irregularis) and Funelliformis mosseae, respectively (Schüssler and Walker, 2010)). The chosen experimental setup allowed us to harvest the bulk of the CMN in the hyphal compartment (HC; Fig. 1), and to estimate the respective carbon investment of the two plants into the CMN through the analysis of the δ 13 C of isolated AMF hyphae or, with higher precision, of the AMF-specific fatty acid C16:1ω5 (Olsson and Johnson, 2005).
We estimated the return of investment with respect to nitrogen and phosphorus for each of the two plants, using 15 N and 33 P as tracers added to the label-hyphal compartment (LHC; Fig 1). As a control, we also grew two monocultures and a mixed culture without any AMF inoculation.

Impact of Common Mycorrhizal Networks on Monocultures and Mixed Culture
A first experiment in the compartmented microcosms (Fig. 1) demonstrated that in mixed culture with sorghum, flax grew poorly in the absence of AMF. Its growth was significantly enhanced (almost by a factor of three), however, in the presence of a CMN formed by Glomus intraradices (Fig. 2, compare center top and bottom). Growth of sorghum, in contrast, was not significantly affected by the presence or absence of a CMN (Fig. 2). Comparing the growth performance of mono-versus mixed culture of flax and sorghum in a CMN, respectively was equally striking (bottom part of Fig. 2): Flax profited substantially (+46% more biomass) from a neighboring sorghum, while sorghum was only marginally, negatively (-7%) affected by the mixed culture growth with flax as neighbor. Thus, the biomass increase of flax did not happen at relevant expense of the neighboring sorghum. Apparently, the two plants had different terms of trade with the CMN of G. intraradices, resulting in an overall higher productivity of the mixed culture, compared to the mean of the two monocultures of flax and sorghum (5.97 g ± 0.18 SEM versus 5.36 g ± 0.14 SEM, respectively, p-value = 0.039; amounting to 11% overall biomass increase by mixed culturing).
The carbon investment of the two plants into the CMN was quantified through the analysis the carbon isotope composition (δ 13 C; for definition see materials and methods) of extracted AMF hyphae (Supplemental Fig. S1A). The hyphal material obtained from the flax monoculture had a δ 13 C value of ~27‰, i.e., slightly higher to the δ 13 C of the host plant (~33‰). Hyphae from the sorghum monoculture displayed a δ 13 C of ~13‰, very close to the value of sorghum plants (δ 13 C = ~14‰). Interestingly, the δ 13 C of the hyphal material from the mixed culture was also very close to the one of the sorghum monoculture, providing first indication that around 80% of the carbon invested into the CMN originated from sorghum (Supplemental Fig. S1A).
In the mixed culture, the return of investment in terms of nutrient uptake by the plants, measured as relative uptake of 33 P and 15 N from the label-hyphal compartment (LHC; Fig. 1), was similarly unbalanced, but in the opposite sense. In the mixed culture, flax obtained the lion's share of both nutrients, i.e. in the range of ~80%, compared to about ~20% for sorghum (Supplemental Fig. S1, B and C).
To confirm and extend these findings, we conducted a second experiment with two different AMF (G. intraradices and G. mosseae). As in the first experiment, biomass accumulation and P and N content of the flax plants were higher when grown in a CMN together with a neighboring sorghum, irrespective of the fungal species (Fig. 3, A, C and E). In contrast, the growth performance of sorghum was not affected (Fig. 3B). As a consequence, the overall productivity in the mixed culture was, again, higher than in the combined monocultures. Despite the absence of any significant growth effect, sorghum also seemed to benefit from the AMF, as indicated by the significant increase in P and N (except N with G.i. in mixed culture) contents (Fig. 3, D and F). Growth limiting factors at the time of labeling, e.g. constraints of rooting space, might have lead to a surplus or "luxury" carbon as well as a reduced sink strength for soil nutrients.

An AMF-specific Fatty Acid as Biomarker for the Plants' Carbon Investment
In order to quantify the carbon investments into the CMN more precisely, we selectively analyzed the carbon isotopic composition of the AMF specific fatty acid (FA C16:1ω5) in the lipid fraction obtained from the hyphal compartment (HC, Fig. 1). This way, potential contamination of the hyphal material by non-symbiotic fungi or other microorganisms, can be excluded. Indeed, confirming its use as a marker for AMF, we found FA C16:1ω5 exclusively in the microcosms inoculated with AMF. As expected, the FA C16:1ω5 in the hyphal compartment (HC, Fig. 1 discrimination during carbon transfer from the plants to the lipids of the arbuscular mycorrhizal fungi (Fig. 4). Remarkably, in the mixed culture (F:S), the δ 13 C values for the extraradical mycelium of both G. intraradices and G. mosseae were much closer to δ 13 C of sorghum than to that of flax in monoculture, roughly confirming our initial finding that the carbon invested into the CMN of the mixed culture derived to ~70% from sorghum and only to ~30% from flax, independent of the fungi involved (Fig. 4).

Nutritional Benefit Gained via Common Mycorrhizal Networks
The non-mycorrhizal systems did not take up any 33 P and relatively little 15 N ( When comparing the acquisition of the isotopically labeled nutrients through the CMN in monocultures and mixed culture, we observed marked differences depending on the fungal species involved. With a CMN formed by G. intraradices, flax received more than twice as much 33 P, and also a little more 15 N, in mixed culture than in monoculture ( plant pair via the CMN was secured by flax. In contrast, with a CMN formed by G. mosseae, flax did not benefit significantly from a neighboring sorghum (Fig. 5, A and C, F:S, black columns). At the same time, sorghum did not suffer intensively from the neighboring flax, although there was still a significantly reduced uptake of nutrients in mixed culture compared to monoculture (Fig.5, B and D, F:S, black columns). In terms of relative uptake, there was no significant difference between flax and sorghum in the mixed culture (Supplemental Fig. S2). Thus, the respective nutrient return to flax and sorghum strongly differed between the two AMF: Flax was much more efficient than sorghum in exploiting the CMN of G. intraradices, whereas in symbiosis with G. mosseae the two plants exploited the CMN on equal terms, though sorghum invested much more carbon than flax (Fig.4).

Mycorrhizal Root Colonization and Hyphal Length Density
Both flax and sorghum were well colonized by AMF, with total colonization ranging between 40 and 62 % and arbuscular colonization ranging between 31 and 43% (Supplemental Table S2). For flax, colonization was similar in monoculture and mixed culture with both AMF species. As for sorghum, both root colonization and arbuscular colonization were significantly reduced in mixed culture, compared to the monoculture, indicating that sorghum plants were less engaged in the symbiosis in the presence of neighboring flax plants. The extraradical hyphal-length density in the hyphal compartment was slightly higher in sorghum monoculture than in the flax monoculture, although statistically significant only in the case of G. mosseae (Supplemental Fig. S3).
In the mixed culture, hyphal-length density was intermediate.

Uneven Terms of Trade in a Common Mycorrhizal Network
Our results (for a graphical synopsis see Fig. 6) emphasize the importance of the terms of trade within a CMN as a driver for co-existence of mycorrhizal plants in ecosystems.
In our mixed-culture experiments, sorghum, as the plant with the higher biomass, consistently provided the bulk of carbon to both tested fungal partners, investing at least twice as much into the CMN as flax. However, the nutritional benefit to the two host plants strongly depended on the fungus involved: In the case of G. intraradices, flax might be viewed as a "cheater" on sorghum, acquiring 80-90% of the total labeled N and P provided by the CMN, whereas the acquisition of labeled N and P was more balanced in the case of G. mosseae (Fig. 6). Obviously, in our experiments, carbon investment and nutritional benefit were not tightly linked. This stands in contrast to recent findings where the resource exchange in the symbiosis of plants with AMF appeared to rely on reciprocal "fair-trade" (Javot et al., 2007;Pietikainen and Kytoviita, 2007;Kiers et al., 2011;Fellbaum et al., 2012). At least, with lower levels of root colonization sorghum did express a negative response to the diminished nutritional benefit in mixed culture with flax, so that certain reciprocity of investment and nutritional benefit became also apparent in our system. It has been proposed that the symbiosis between plants and AMF is based on the exchange of "luxury goods" (Kiers and van der Heijden, 2006). Hence, CMNs can exist without causing significant additional costs to either partner, especially when the carbon cost is negligible for the main carbon donor. This appeared to be the case for sorghum, which dominated (~60% by biomass weight) in our mixed cultures, or more obviously, for large trees supporting small mycoheterotrophic plants (Courty et al., 2011).
In natural plant communities, the demand for "AMF services" such as soil nutrient acquisition and, vice versa, the availability of "luxury goods" such as a surplus of carbon, is expected to dynamically change for the different plants, depending on their strategies to respond to environmental cues and their specific life-history traits with consecutive phases of vegetative growth, maturation, senescence etc. Thus, CMNs supposedly function as dynamic "marketplaces" in biodiverse ecosystems, where the symbionts involved and apparently organized in networks of plant-AMF assemblages

Sharing Luxury Goods Maximizes Productivity
Our experimental data clearly demonstrate that an unbalanced use of the CMN not only can increase the growth of an individual plant such as flax, but it also can increase the productivity of our two-plant model ecosystem by sharing the benefit of a luxury good

Organisms and Microcosms
The two host plants used were flax (Linum usitatissimum cv. Agatha) and sorghum (Sorghum bicolor cv. Pant Chari-5). The two fungal partners, both of the genus Glomus Balzer, Wetter-Amönau, Germany (http://www.labor-balzer.de).The RHCs were inoculated with a 2 g (approx. 100 spores/ compartment) inoculum of one of the Glomus strains, or with 2 g sterilized (120° C, 20 min) inocula as non-mycorrhizal control. In addition, the RHCs received 10 ml of a microbial wash to equalize microbial communities (Koide and Elliott, 1989). This wash was prepared by wet sieving 100 g of each inoculum through a 32 μ m sieve and a paper filter (FS 14 ½,Schleicher & Schuell) yielding a final volume of 1 L. The LHC (Fig. 1)  RHCs and HC, and thereby adjusted to equal soil water content of 90% field capacity by weighing. In addition, every week during the first eight weeks of cultivation, the RHC was amended with 8 ml of a P-free Hoagland Solution (Gamborg and Wetter, 1975;Zabinski et al., 2002).

Experimental Design
In each of the two RHCs, one single plant was grown, yielding microcosms with "monocultures" (a pair of identical plants) or a "mixed culture" (one flax and one sorghum plant). In the preliminary experiment, the plants were inoculated either with G. intraradices (Gi) or with the sterilized control inoculum (NM). In the second experiment, growth experiments were also conducted with G. mosseae (Gm). The microcosms were harvested after 12 weeks of growth.

Plant Growth Performance and Symbiotic Interaction
Roots were washed thoroughly, excess moisture was removed, and fresh weight was determined. Two subsamples were weighed, one of which was then used for determination of root dry weight. The other aliquot was cleared using a 10% KOH solution, and stained in Trypan Blue for mycorrhizal-structure identification inside the root (Phillips and Hayman, 1970). The percentage of root length occupied by hyphae, arbuscules and vesicles was estimated for each sub-sample by a modified line intersection method (McGonigle et al., 1990). A minimum of 50 line intersection per root sample were scored for AMF. Shoot and root samples were dried for 24h at 105 ºC and weighed separately; the sum corresponds to the "total biomass" indicated in the figures.
Dried shoots and roots were ground at 30 Hz in a mixer mill (MM2224, Retsch, Haan, Germany). Aliquots of 2 mg were weighed in for elemental analyses. Nitrogen and carbon concentrations were determined using an ANCA elemental analyzer/mass spectrometer (Europa Scientific Ltd., Crewe, UK). P concentration of shoots and roots was measured using the molybdate blue method on a Shimadzu UV-160 spectrophotometer (Shimadzu Biotech, Duisburg, Germany) after acid digestion www.plantphysiol.org on August 29, 2017 -Published by Downloaded from Copyright © 2012 American Society of Plant Biologists. All rights reserved. (Murphy and Riley, 1962). The substrate of the HC was stored at -20 ºC. A subsample of 50 g was used for hyphal length density measurements, determined by the grid-lineintersection method (Jakobsen et al., 1992).

Nutrient Gain of the Mycorrhizal Network
Plant 33 P contents were measured using a Packard 2000 liquid scintillation counter (Hewlett-Packard, Waldbronn, Germany). The 15 N content of plants was analyzed with an ANCA mass spectrometer (Europe Scientific Ltd., Crewe, UK). Relative 33 P and 15 N uptake were calculated by dividing uptake of individual plants by the total uptake of both plants of the microcosm.

Carbon Contribution to Mycorrhizal Network
The carbon isotope composition of plant shoot and roots and of hyphal biomass was determined using an ANCA IRMS. Extraradical hyphae were extracted from the HC by a wet sieving method (Johansen et al., 1996). The recovered hyphae were dried in a DNA-Speed Vac (Savant) prior to bulk mass spectrometric analysis. For compoundspecific analyses of the AMF-specific fatty acid (FA) C16:1ω5, lipid extraction was carried out according to previously described methods (Elvert et al., 2003;Niemann et al., 2005). Briefly, total lipid extracts were obtained by suspending and sonicating 25 g of freeze-dried substrate of the HC in organic solvents of decreasing polarity. Internal standards (n-nonadecanol and n-nonadecanoic acid) of known concentration and carbon isotopic composition were added prior to extraction. Total lipid extracts were saponified with a methanolic KOH-solution (6%). After extraction of the neutral fraction from this mixture, FAs were methylated using a boron triflouride solution (14% BF3 in

Statistical Analysis
Experiment 1 was set up in a randomized block design where each treatment was replicated four times. Mean comparison among treatments were performed by independent paired t-tests for dry weight and relative uptake of 33 P and 15 N of the two individual plants.
Experiment 2 was set up in a randomized block design including two temporal blocks with a time lag of 4 weeks. Each block contained three replicates, with a resulting total of 6 replicates per treatment. An analysis of variance (ANOVA) was performed on the total biomass, on the P and N content, and on the total and arbuscular colonization for each plant species separately, where the two latter parameters were arcsinetransformed to fit the assumption of normal distribution. The ANOVA was based on the three factors culture system (with two levels), AMF (with three levels) and block (with two levels). Pairwise comparisons between the treatments were done with planned contrast analysis. Independent paired t-tests were performed to analyze whether means of relative uptake of 33 P and 15 N of the two individual plants differed significantly from each other. An ANOVA with the factors treatment (9 levels) and block (two levels) was executed on the fungal parameter hyphal length density. A probability of P  Conversely, growth of sorghum was only marginally influenced by the culture system.    invested more than twice as much than flax in terms of carbon. The return, in form of the nutrients P and N, is illustrated by the yellow and orange arrows, respectively. In the CMN formed by G. intraradices, the return was extremely uneven; flax obtained 80-94% of the nutrients delivered by the CMN, and sorghum, the main investor, only 6-20%. In the CMN formed by G. mosseae, both flax and sorghum received an approximately equal share of the nutrients delivered by the CMN, but since flax invested less than half as much carbon compared to sorghum, it still benefited from its neighbor. sorghum invested more than twice as much than flax in terms of carbon. The return, in form of the nutrients P and N, is illustrated by the yellow and orange arrows, respectively. In the CMN formed by G. intraradices, the return was extremely uneven; flax obtained 80-94% of the nutrients delivered by the CMN, and sorghum, the main investor, only 6-20%. In the CMN formed by G.
mosseae, both flax and sorghum received an approximately equal share of the nutrients delivered by the CMN, but since flax invested less than half as much carbon compared to sorghum, it still benefited from its neighbor.