Ectomycorrhizal fungal communities coinvading with Pinaceae host plants in Argentina: Gringos bajo el bosque

Transcription

1 Research Ectomycorrhizal fungal communities coinvading with Pinaceae host plants in Argentina: Gringos bajo el bosque Jeremy Hayward 1, Thomas R. Horton 1 and Martin A. Nu~nez 2 1 Department of Environmental and Forest Biology, State University of New York College of Environmental Science and Forestry, Syracuse, NY, USA; 2 Laboratorio Ecotono, INIBIOMA, CONICET, Universidad Nacional del Comahue, Bariloche, Argentina Author for correspondence: Martin A. Nu~nez Tel: nunezm@gmail.com Received: 12 January 2015 Accepted: 7 April 2015 doi: /nph Key words: biological invasions, coinvasion, ectomycorrhizas (ECMs), Pinaceae, Suillus. Summary Coinvasive ectomycorrhizal (ECM) fungi allow Pinaceae species to invade regions otherwise lacking compatible symbionts, but ECM fungal communities permitting Pinaceae invasions are poorly understood. In the context of Pinaceae invasions on Isla Victoria, Nahuel Huapi National Park, Argentina, we asked: what ECM fungi are coinvading with Pinaceae hosts on Isla Victoria; are some ECM fungal species or genera more prone to invade than others; and are all ECM fungal species that associate with Northern Hemisphere hosts also nonnative, or are some native fungi compatible with nonnative plants? We sampled ECMs from 226 Pinaceae host plant individuals, both planted individuals and recruits, growing inside and invading from plantations. We used molecular techniques to examine ECM fungal communities associating with these trees. A distinctive subset of the ECM fungal community predominated far from plantations, indicating differences between highly invasive and less invasive ECM fungi. Some fungal invaders reported here have been detected in other locations around the world, suggesting strong invasion potential. Fungi that were frequently detected far from plantations are often found in early-successional sites in the native range, while fungi identified as late-successional species in the native range are rarely found far from plantations, suggesting a means for predicting potential fungal coinvaders. Introduction The edaphic, climactic and biotic factors that cause some invasions to succeed and others to fail are complex (Binggeli, 1996; Richardson & Rejmanek, 2004; Dodet & Collet, 2012). The effects of interactions between species on the process of invasion remain particularly elusive: while they are frequently critically important (Richardson et al., 2000a; Keane & Crawley, 2002; Callaway et al., 2004), their context-dependency and intrinsic complexity make predicting outcomes difficult. Obligate mutualisms are an important class of interactions in biological invasions (Richardson et al., 2000a). In the case of obligate mutualisms, exotic plants must either form novel associations with native partners, or must be cointroduced with exotic partners (Richardson et al., 2000a; Dickie et al., 2010; Nu~nez & Dickie, 2014). Pinaceae species are obligately ectomycorrhizal (ECM): while plants can be grown to sexual maturity with fertilizer additions, under nonagroforestry conditions, growth and survivorship of nonmycorrhizal Pinaceae seedlings are extremely low (Hatch, 1936; Mikola, 1970; Nu~nez et al., 2009). Pinaceae species also have little or no compatibility with native ECM fungi found on isolated oceanic islands outside the native range of Pinaceae species, as well as those found in the southern hemisphere (Mikola, 1970; Tedersoo et al., 2007; Nu~nez et al., 2009; Dickie et al., 2010; Walbert et al., 2010; Hynson et al., 2013). Pinaceae species introduced into these areas are thus usually cointroduced, either deliberately or accidentally, with compatible ECM fungi (Mikola, 1970). Pinaceae species invading into native vegetation in these regions therefore coinvade (defined here as two or more organisms which invade together through a symbiosis maintained during invasion; Nu~nez & Dickie, 2014) with their ECM symbionts (Vellinga et al., 2009; Dickie et al., 2010; Nu~nez et al., 2013; Hayward et al., in press). Isla Victoria, part of Nahuel Huapi National Park in Neuquen Province, Argentina, has been heavily invaded by Pinaceae species, as well as by a variety of other gymnosperm and angiosperm families (Simberloff et al., 2002). We have previously reported that Pinaceae species invading into native vegetation far from their point of introduction on Isla Victoria may be critically limited by the availability of compatible ECM fungal inoculum (Nu~nez et al., 2009). Nevertheless, if we define species as invasive when they spread without human intervention in a zone to which they have been introduced (Richardson et al., 2000b), several Pinaceae species are invasive on Isla Victoria, as throughout much of Patagonia (Richardson et al., 1994; Simberloff et al., 2010; Nu~nez et al., 2013). The complex mosaic of plant 497

2 498 Research New Phytologist invasions into native Patagonian forests on Isla Victoria was investigated by Simberloff et al. (2002), in a paper entitled Gringos en el bosque: introduced tree invasion in a native Nothofagus/ Austrocedrus forest. Here, we investigate the below-ground component of the same system of invasions, focusing on invasive Pinaceae species, but also opportunistically sampling other nonnative ECM host plant species. Our title is intended both as an homage to Simberloff et al. (2002) and as a reminder that not all biological invasions are visible to the naked eye. We ask three questions. First, what nonnative ECM fungi are coinvading with Pinaceae host plants on Isla Victoria? Second, are some ECM fungal genera or species more prone to invade than others, both on Isla Victoria and on a global scale? Finally, are all ECM fungal groups that associate with Northern Hemisphere hosts also nonnative, or are some native ECM fungi compatible with nonnative plants? We investigated these questions through below-ground sampling of the ECM fungal communities associating with two Pinaceae genera, as well as opportunistic sampling of ECM angiosperm species, on Isla Victoria. Materials and Methods Field site We sampled ECM root tips from host trees growing on Isla Victoria, Neuquen Province, Argentina. Isla Victoria is an island of c.40km 2 in Lake Nahuel Huapi. The island is entirely contained within Nahuel Huapi National Park, and has been a protected area since The native plant community is dominated by the ECM Nothofagus dombeyi (Nothofagaceae) and the arbuscular mycorrhizal Austrocedrus chilensis (Cupressaceae); one other native ECM plant species, Nothofagus antarctica, occurs patchily. The island is heavily invaded by exotic vertebrates, including silver pheasants (Lophura nycthemera), fallow deer (Dama dama), red deer (Cervus elaphus) and wild boar (Sus scrofa; Simberloff et al., 2003; Nu~nez et al., 2013). Between 1900 and 1934, the island was substantially altered by logging and ranching operations; however, these disturbances were effectively ended in 1934 by the establishment of the national park (Nu~nez et al., 2011). The continuing impacts of these disturbances are not known, but they probably vary across the island. By the end of the last century, most of the island had reverted to Nothofagus Austrocedrus forest with a very low density of nonnative plants (Simberloff et al., 2002). However, before 1934, a large number of nonnative ECM tree species were planted on Isla Victoria in one of the first attempts of forestry with nonnative conifers in South America (Simberloff et al., 2002, 2010). While some species are no longer found on the island (perhaps as a result of grazing, population stochasticity, or other factors), at least 68 species are still present; some are invading into native Nothofagus stands (Simberloff et al., 2002; Nu~nez et al., 2009). Of these nonnative trees, the most important invaders are Pseudotsuga menziesii, Pinus ponderosa, Pinus monticola, Pinus sylvestris and Pinus contorta; P. menziesii outnumbers other invaders by an order of magnitude (Simberloff et al., 2002; Nu~nez et al., 2011). These invaders typically form dense invasion fronts close to plantations, with additional scattered individuals at distances up to c. 1 km from plantations (Simberloff et al., 2002; Nu~nez et al., 2011). The center of the plantations is found near S, W. The plantations follow the island s roughly north south orientation; the northernmost plantation s northern edge is near S, and the southernmost plantation s southern edge is near S. Field methods We attempted to sample ECM root tips from 10 host trees from each of two genera, from each of four distance classes and from each of three age classes ( = 240 trees). The two genera from which we sampled trees were Pinus and Pseudotsuga; the four distance classes were inside plantations, 0 50 m from plantations, m from plantations, and > 750 m from plantations; the age classes were trees 0 2 yr old, > 2 yr but not yet sexually mature; and sexually mature individuals. We adopted this sampling scheme for the following reasons: because ECM fungal associations can change with the age of host plants (Twieg et al., 2007), we sampled trees from each genus belonging to each of three age classes to improve our coverage of the full range of fungi associating with each host plant genus in each distance class. We sampled two genera because host specificity can occur at the genus level in Pinaceae and our goal was to maximize our recovery of Pinaceae-associating ECM fungal species (Molina et al., 1992; Kretzer et al., 1996; Ishida et al., 2007). Finally, we hoped to capture changes in ECM fungal communities with distance from plantations, and so chose a wide range of distances from plantation edges. We chose these distance classes to reflect a gradient of invasion density, from dense invasion fronts < 50 m from plantations, to extremely scarce individual trees > 750 m from plantations (Simberloff et al., 2002). Our determination of age classes followed previous work on Isla Victoria in which individual trees were cut at ground level, sanded, and aged by counting growth rings. This allowed us to determine the age classes of trees in this study by sight, as follows: sexually mature individuals were large trees (generally > 15 cm diameter at ground level) with evidence of past reproduction such as fallen or attached cones (male or female). Juvenile trees (0 2yr old) were identified as those < 10 cm in height, with minimal deposition of rigid woody material in the stem. For many of these juvenile trees, we confirmed our aging by clipping the stem at ground level and inspecting the stem for rings. Intermediate trees (> 2 yr old but not yet sexually mature) were those not identified as belonging to either of the two previous age classes. We did not identify trees below the genus level. All sampled trees outside plantations were established without human intervention (i.e. these trees were recruits, not planted individuals). Plantations were composed of Pinus spp., including Pinus ponderosa, Pinus monticola, Pinus sylvestris, and Pinus contorta, as well as Pinus menziesii (Simberloff et al., 2002). All sampled trees were separated by at least 10 m to minimize spatial autocorrelation (Lilleskov et al., 2004). In our sampling, we were able to locate only 46 Pinaceae individuals in appropriate age classes farther than 750 m from

3 New Phytologist Research 499 plantations. In this distance class it was only possible to find four sexually mature and seven sexually immature but older than 2 yr P. menziesii, and five Pinus < 2 yr old. All other distance classes were represented by 60 trees, yielding a total sampling effort of 226 trees. To determine whether other nonnative ECM hosts on Isla Victoria harbor additional ECM fungal species, we also harvested root tips opportunistically from four species of nonnative angiosperm ECM hosts: Quercus robur (14 trees sampled in February 2011; 10 trees sampled in June 2012), Castanea sativa (six trees sampled in February 2011; three trees sampled in June 2012), Carpinus betulus (two trees sampled in February 2011; two trees sampled in June 2012) and Fagus sylvatica (one tree sampled in June 2012). These trees are of low abundance and are apparently not found outside the area of initial introduction; records indicating whether the individuals sampled are recruits or planted specimens are not available (Simberloff et al., 2002). We harvested root tips by tracing roots away from the boles of individual trees to fine root clusters. We collected c. 15 cm of fine root material from a single fine root cluster per tree and stored the sample immediately in 2 ml 29 CTAB buffer. Within 2 wk, we rinsed samples in tap water and re-stored them in fresh 29 CTAB buffer until analysis. We sampled inside plantations only after determining that community differences between distance classes existed: samples outside plantations were taken in February 2011 (Austral summer); samples inside plantations were taken in June 2012 (Austral winter). Molecular methods We sorted ECM root tips in each sample to unique morphological types (morphotypes) using the characters of Agerer (1987). We analyzed two to three specimens of each morphotype using molecular methods, resulting in a total of 817 mycorrhizal root tips processed for DNA. We sorted by morphotypes only within samples, without lumping between samples. We extracted DNA from mycorrhizal root tips following the protocol of Nu~nez et al. (2013). We amplified the internal transcribed spacer (ITS) region of fungi in DNA extracts using combinations of the forward primers ITS1 (White et al., 1990), ITS1f (Gardes & Bruns, 1993), and NSA3 (Martin & Rygiewicz, 2005), and the reverse primers NLB4 (Martin & Rygiewicz, 2005) and ITS4. PCR conditions were as follows: 3 min at 94 C, followed by 35 cycles of 35 s at 94 C, 55 s at 53 C, and 45 s at 72 C, adding 2 s per cycle to the extension time, with a final extension period of 10 min at 72 C. We amplified products in 25 ll reactions using 12.5 ll DNA extract diluted with PCR-grade water at ratios varying from 1 : 1 to 1 : 25; all other reagents followed manufacturer s guidelines. We generated restriction fragment length polymorphism (RFLP) profiles from PCR products using the restriction enzymes HinfI andhaeiii (New England Biolabs, Ipswich, MA, USA) following the manufacturer s protocols. We compared RFLP types only with types generated using the same primer set. We sequenced at least five exemplars of each RFLP type, except when fewer than five exemplars were present, in which case all exemplars were sequenced. Sequencing was performed automatically on an ABI3750XL sequencer at the laboratories of Operon, inc. (Eurofins MWG Operon, Huntsville, AL, USA) using standard chemistry. In one genus (Suillus) we found RFLP patterns difficult to interpret, as a result of fragments of similar molecular weight (which do not separate well in gel electrophoresis) shared among several operational taxonomic units (OTUs). To ensure that we were not aggregating similar genotypes, we reamplified all Suillus-like RFLP types and used a third restriction enzyme, DpnII, to resolve ambiguities, particularly between Suillus luteus and Suillus lakei. Bioinformatics and statistics To group and identify sequences, we followed a barcode gap approach (Schoch et al., 2012). First, we classified recovered sequences to approximate family based on BLAST comparison to the NCBI GenBank database. For each approximate family-level group, we created an alignment in SeaView 4 (Gouy et al., 2010) using MUSCLE (Edgar, 2004), for a total of 17 alignments with an average of approx. eight sequences per alignment. We used these alignments to assign OTUs in Mothur 1.27 (Schloss et al., 2009) grouped by the nearest-neighbor method using a threshold value of 3% dissimilarity in ITS sequence with gaps penalized as a single change and end gaps not penalized. Assignment to species level was made with ITS dissimilarity < 3% and no conflicting assignments in the top 25 BLAST hits; assignment to genus was made with ITS dissimilarity < 10% and no conflicting assignments in the top 25 BLAST hits; assignment to family was made at < 20% ITS dissimilarity and no conflicting assignments in the top 25 BLAST hits. In all cases, we required BLAST coverage of > 75% of our query sequence for assignment. All specific, generic and familial names used here, including those for specimens found as fruit-bodies, refer to the results of sequence-based identification. We uploaded sequences for each OTU to Gen- Bank under accession numbers KM KM (Supporting Information Table S1). We performed all statistical analyses statistics in R ( using the packages vegan (Oksanen et al., 2015) and rich (Rossi, 2011). Richness and diversity were calculated using the diversity and specpool functions in vegan. Because samples from inside plantations were not sampled simultaneously with all other samples, and are therefore not strictly comparable, we do not include samples from inside plantations in most statistics. However, we do include them in a constrained correspondence plot (Fig. 1), where only OTUs shared between two or more distance classes are included. To compare OTUs detected here with OTUs previously detected associating with Pinaceae host plants far from the native range, we included sequences generated in studies elsewhere in the world which used similar methodologies (Walbert et al., 2010; Hynson et al., 2013) in alignment and OTU assignments, as earlier. To assist in determining whether certain fungal OTUs detected associating with Pinaceae host plants might be native to the southern hemisphere, we used phylogeographic analyses. We created a list of candidate OTUs potentially native to the southern hemisphere based on BLAST results, as follows: no BLAST results > 97% identified as originating in the northern

4 500 Research New Phytologist Fig. 1 Constrained correspondence analysis (CCA) plot with ectomycorrhizal fungal community data constrained by host plant distance class and conditioned with host plant age class and genus, using biplot scaling optimized for species scores. LC scores are shown. CCA axis 1, eigenvalue = and proportion explained, 0.612; axis 2, eigenvalue and proportion explained, 0.305, resulting in proportion explained by the first two axes, Ellipses represent 95% confidence intervals for centroid positions. Only operational taxonomic units (OTUs) that are shared between two or more distance classes are included. hemisphere; at least one BLAST result in the top 20 originating in the southern hemisphere. For these candidates, we downloaded the top 20 BLAST hits from GenBank. We aligned these sequences to those sequences representing our OTUs using MUSCLE, as earlier. We used PhyML 3.0 (Guindon et al., 2010) using the GTR nucleotide model (as selected using jmodeltest 2; Darriba et al., 2012), optimized invariable sites, and the SPR tree-searching algorithm with five random starts. Branch support scores were calculated using alrt. An OTU accumulation curve for our sampling effort of Pinaceae-associating ECM fungi is shown in Fig. 2. Pairwise Monte Carlo tests (function c2cv; Rossi, 2011) suggest that observed OTU richnesses do not significantly differ between any of the distance classes we sampled (see also Tables 1, 2). However, pairwise Monte Carlo tests of fungal communities rarefied to 10 Results From 817 Pinaceae root tips and 93 angiosperm root tips analyzed using molecular methods, we recovered 544 successful amplifications in which only a single PCR product was evident. These 544 amplifications yielded a total of 57 fungal OTUs grouped at 3% ITS dissimilarity; all RFLP types were assigned unambiguously to OTUs. Of these 57 OTUs, eight displayed similarity only to groups containing fungi of endophytic ecology, four were not assignable to a known ecology, and the remaining 45 could be assigned to ECM lineages. We exclude fungi of endophytic or unknown ecology from our analyses. We detected 28 ECM fungal OTUs associating with Pinaceae host plant species in this study and 17 with angiosperm host plant species. Including fungi detected only as sporocarps, and fungi detected in Nu~nez et al. (2009, 2013), we have detected 48 putatively ECM fungal OTUs on Isla Victoria, of which 33 were detected associating with Pinaceae species (Table S1). Fig. 2 Species accumulation curve for the overall community of ectomycorrhizal fungi sampled on Isla Victoria. The curve was generated using the exact method of Ugland et al. (2003) as implemented in the vegan package in R (Oksanen et al., 2015); the shaded region represents a 95% confidence interval.

5 New Phytologist Research 501 Table 1 Observed richness, estimated richnesses (Chao and the abundance coverage estimator, ACE), Simpson s diversity and sample size for fungal communities associating with Pinaceae host plants in each of the four distance classes in this study, calculated using the vegan package (Oksanen et al., 2015) in R 2.15; richnesses are shown for each distance class or site as a whole, not for individual trees Distance class Observed richness Chao estimator ( SE) ACE estimator ( SE) Simpson s diversity (1 D) n (trees) Plantations m m m Overall Table 2 The number of plants in each distance class with which we detected each ectomycorrhizal fungal operational taxonomic unit (OTU) found in this study OTU Inside plantations 0 50 m m 750 m+ Rhizopogon Amphinema Lactarius quieticolor Suillus luteus Cortinarius Suillus lakei Tomentella Cortinarius Pseudotomentella tristis Leotiomycetes Hebeloma Hebeloma Pyronemataceae Rhizopogon Melanogaster Inocybe Inocybe Boletus edulis Sebacinaceae Tricholoma Thelephora terrestris Inocybe Cortinarius Cortinarius Helotiales Tomentella Russula samples revealed that while the richness of the 0 50 m distance class was not significantly different from the richness of the m distance class (5000 bootstrap replicates; P > 0.4), both were significantly richer than the furthest distance class (5000 bootstrap replicates; P < 0.025; Bonferroni-corrected P-values < 0.05). Considering all sites outside plantations, the estimated richness is (Chao estimator SE) or (second-order jackknife estimator). Permutational ANOVA tests using Chao community indices suggest that ECM fungal communities vary significantly with distance class of host tree outside plantations (ADONIS stratified by host plant age class and genus; permutations; df = 2; R 2 = 0.065; P < 0.002), with genus of host tree (ADONIS stratified by host plant age class and distance class; permutations; df = 1; R 2 = 0.064; P < 0.001) and with age class of host tree (ADONIS stratified by distance class and genus; permutations; df = 2; R 2 = 0.04; P < 0.029); however, the very substantial residual variances in each case limit predictive value. Constrained correspondence analysis (CCA; Legendre & Legendre, 2012; Oksanen et al., 2015) revealed that a model relating ECM fungal community structure to host plant age class is not significant, whether or not the effects of host plant distance class and host plant genus are partialled out (permutational ANOVA function anova.cca; 500 permutations; F = 1.15; P = 0.15). The differences between these results and the ADO- NIS test, earlier, are attributable to the stratified design of the latter, which gives greater weight to the relatively less intensively sampled mature age class. To determine whether some part of the ECM fungal community is more strongly affected by host plant distance class than expected by chance, we used CCA to relate community structure to the sole explanatory variable distance class, with the effects of host plant age class and genus partialled out; only OTUs shared between two or more distance classes were included. A permutational ANOVA suggested that the constraint is significant (anova.cca; df = 3; 200 permutations; F = 1.6; P = 0.015), but explains relatively little of total ECM fungal community structure (constrained inertia, 0.39; total inertia, 7.7). Suillus luteus was the only OTU with predictive value outside plantations in indicator species analysis (ISA); it was significantly associated with the furthest host plant distance class (indicator value 0.09; P < 0.005). We categorized nine OTUs as potentially native to the southern hemisphere based on BLAST hits: Inocybe OTUs 1, 2, 4, 5, 6, and 7, Tomentella OTUs 4 and 6, and Sebacinaceae OTU1. The BLAST hits for these sequences all included sequences deposited by Nouhra et al. (2013), who carried out substantial below-ground sampling in a pristine Nothofagus forest < 200 km from our sampling site. Top GenBank blast hits included sufficient alignable sequences for informative placement only of the Inocybe OTUs; the phylogeny generated for these OTUs is included as Fig. S1. Discussion The coinvasive community Overall, we have detected 42 ECM OTUs on Isla Victoria compatible with Pinaceae host plants, including 26 OTUs

6 502 Research New Phytologist associating with Pinaceae host plants outside plantations. While not all of these are certain to represent coinvasion events, and at least two are unlikely to be northern hemisphere invaders (see the Associations with native fungi subsection later), the majority of these OTUs are probably not native to Patagonia. While this community is substantially poorer than virtually all native-range communities sampled thus far (Tedersoo et al., 2012), it is by far the richest below-ground accumulation of exotic Pinaceaeassociating ECM fungi detected to date. Dickie et al. (2010), Walbert et al. (2010) and Hynson et al. (2013) all worked with a single genus of nonnative trees and reported 14, 19 and 24 OTUs respectively; each worked with a single genus of nonnative trees). Simberloff et al. (2002) report seven invasive plant species on Isla Victoria. In native forests in the northern and southern hemispheres, below-ground communities are frequently richer than above-ground plant communities; these results suggest that this pattern holds true for at least some invasions, too. This high richness of previously unobserved and cryptic invaders highlights the underappreciated role of fungi in invasion biology (Desprez-Loustau et al., 2007). Northern hemisphere angiosperm ECM hosts harbored 12 ECM fungal OTUs not detected with Pinaceae host plants, out of a total of 17 ECM fungal OTUs associating with these angiosperm hosts. Some of these probably represent native fungi (see the Associations with native fungi subsection later), but others, including Scleroderma bovista, have been described previously from the northern hemisphere, and probably represent introduced species. The reasons for the apparent restriction of these introduced fungi to angiosperm hosts (and their consequent failure to coinvade) are unclear, and we note that we did not saturate the species-accumulation curve for Pinaceae-associating ECM fungi. Nevertheless, Ishida et al. (2007) suggests that many fungi in the native range are not shared between angiosperm hosts and Pinaceae hosts; the failure of some fungi to associate with Pinaceae may reflect patterns of mycorrhizal specificity. Mycorrhizal specificity may reflect innate or genetic factors, or the influence of the environment (e.g. edaphic, climactic, or hydrological) (Molina et al., 1992). If some ECM fungi are capable of associating only with noninvasive plant hosts such as the angiosperms present on Isla Victoria, their ability to invade may be critically limited by their partner s invasion potential. These findings highlight the ability of biotic interactions, including mutualisms, to influence the invasion process. We have previously reported that Pinaceae invasions are limited on Isla Victoria by availability of ECM inoculum (Nu~nez et al., 2009). In the light of this finding, the richness of putatively exotic ECM fungal communities at distances from plantations > 50 m may be surprising. While some ECM species (particularly in the family Rhizopogonaceae) are capable of establishing longlived spore banks (Ashkannejhad & Horton, 2006; Bruns et al., 2008), it is likely that propagules of many of the exotic fungi detected here have low survivability in the absence of a compatible plant host, as evidenced by the failure of congeneric or conspecific fungi to colonize compatible host plants after a rest period (Baar et al., 1999; Taylor & Bruns, 1999; Nguyen et al., 2012). In a study from the native range of Pinaceae species, we have reported that seedlings recruiting far from sources of ECM inoculum are likely to establish with resistant propagule communities (Ashkannejhad & Horton, 2006). After the establishment of host trees, additional fungal species may colonize, even without resistant propagules (Visser, 1995; Collier & Bidartondo, 2009; Hayward et al., in press). We have previously reported eight exotic ECM fungi found on Isla Victoria capable of colonizing seedlings far from plantations, presumably largely from resistant propagules, at least in part (Nu~nez et al., 2009). We hypothesize that some or many of the additional ECM fungal OTUs detected in this study far from plantations represent secondary colonization, and therefore may not be capable of enabling invasions (sensu Hayward et al., in press) to the same extent as fungi possessing resistant propagules. The apparently higher richness of ECM fungi found outside plantations here than found by Nu~nez et al. (2009) may reflect local range expansions by some ECM fungi during the course of invasion, or may simply reflect the methodological differences between that study and this. The spatial complexity of the plantations on Isla Victoria, the large population of exotic mammals (S. scrofa, D. dama and C. elaphus, all of which vector ECM inoculum; Nu~nez et al., 2013), and the complicated land-use history of the site (potentially including movement of soil or other materials; Koutche, 1942) may all contribute to the dispersal of ECM fungi far from plantations. Differences in invasion potential Outside plantations, we found differences in ECM fungal communities with differing distances from plantation edge, with rarified pairwise Monte Carlo tests suggesting that the furthest host plant distance class contained fewer ECM fungal OTUs than nearer distance classes. Because samples from plantations were not taken contemporaneously with those from invading plants, we exercise caution in making comparisons between these sampling units. However, evidence for seasonal patterns in ECM fungal communities associating with evergreen trees such as Pinaceae species is equivocal (Taylor & Bruns, 1999; Koide et al., 2007). In particular, several of the patterns we observed are probably the result of differences in invasion potential, and are unlikely to derive from seasonal confounds. By far the greatest number of OTUs not found in any other distance class were found inside plantations (Table 2), consistent with at least some of these OTUs being poor dispersers or invaders. Cortinarius, the most frequent genus found only inside plantations, has long been considered a late-stage ECM fungal genus (Deacon et al., 1983; Bruns, 1995; Redecker et al., 2001) with minimally resistant propagules (Torres & Honrubia, 1994; Ishida et al., 2008). Latestage fungi (sensu Peay et al., 2011) are defined by a characteristic suite of co-occurring functional traits, including strong competitive abilities in root-dense environments and poor dispersal abilities, among others. These fungi are not known to be affected differently by seasonal variations in temperature or precipitation, but are known to associate preferentially with older trees (Koide et al., 2007; Walker et al., 2008; Jumpponen et al., 2010; Peay et al., 2011; Peay et al., 2012). Tedersoo et al. (2007) and

7 New Phytologist Research 503 Hynson et al. (2013) also report Pinaceae-associating Cortinarius species close to or inside forestry plantations in the Seychelles and Hawaii, respectively, suggesting that the pattern we observed may be true elsewhere, too. Of the four Cortinarius OTUs detected here, only one was ever detected > 50 m from plantations. The most frequently encountered OTU inside plantations was Amphinema OTU1; many Amphinema species share several lifehistory traits with Cortinarius species, including the formation of dense fringes of rhizomorphs (Agerer, 2001) and propagules that are apparently not resistant (Kipfer et al., 2010). These functional traits may imply an increased reliance on vegetative spread at the expense of spread via spores (Newton, 1992), which could contribute to the reduced invasion potential of these genera. From an applied perspective, the identification of fungi that do not invade can be key for sustainable forestry with nonnative species (Nu~nez & Dickie, 2014), as inoculation with these fungi alone may help prevent the large-scale spread of Pinaceae species. While inoculation of seedlings with typical late-stage fungi such as Cortinarius spp. may not be practicable, and care must be taken to avoid introducing potential novel invaders (Brundrett et al., 1996; Ishida et al., 2008), it may nevertheless be possible to locate ECM fungal species that are both useful in forestry and minimally invasive. The fungi most frequently detected (> five detections) in host plant distance classes > 50 m from plantations included S. luteus, Rhizopogon OTU3, Lactarius quieticolor, and S. lakei; of these OTUs, only S. luteus was detected more than five times at distances from plantations > 750 m. Suillus and Rhizopogon species are known to produce highly resistant propagules and are vectored by mammals (Baar et al., 1999; Ashkannejhad & Horton, 2006; Nguyen et al., 2012; Nu~nez et al., 2013). Both genera are frequently reported associating with Pinaceae species far from the native range, including in invasive contexts (Mikola, 1970; Van der Westhuizen & Eicker, 1987; Chu-Chou & Grace, 1988; Tedersoo et al., 2007; Dickie et al., 2010; Walbert et al., 2010; Hynson et al., 2013). Hayward et al. (in press) showed that S. luteus may drive Pinaceae invasions even in the absence of other coinvasive EM fungi, suggesting that suilloid fungi can play an outsized role in such invasions. The significant indicator score of S. luteus with the furthest distance class in ISA suggests that it is significantly less frequent in distance classes < 750 m, perhaps indicating that Suillus is replaced by other fungi as they colonize (Taylor & Bruns, 1999; Hynson et al., 2013). Leotiomycetes OTU1 was also detected more frequently in the furthest distance class than close to plantations; conversely, Rhizopogon OTU3 was most frequently detected near plantations, although it was also detected in other distance classes. In many respects, the patterns in ECM fungal communities observed in the invasive context of Isla Victoria are similar to those that have been observed in successional environments in the native range. Suilloid fungi are characteristic of the establishment of Pinaceae species in burned or previously uncolonized areas in the northern hemisphere, driven by mammalian vectoring and the production of resistant soil spore banks analogous to resistant seed banks (Cazares & Trappe, 1994; Visser, 1995; Baar et al., 1999; Ashkannejhad & Horton, 2006; Kipfer et al., 2011). A pattern of decreasing dominance of suilloid fungi in latersuccessional environments has also been observed in the native range (Taylor & Bruns, 1999). Gardes & Bruns (1996) and Peay et al. (2011) have suggested that there may be a competitive tradeoff between fungi spreading predominantly via vegetative growth and those spreading through spores; they report that suilloid fungi are less frequent inside forests than at their edges because of competitive interactions. Increasing dominance of rhizomorph-forming taxa and/or nonresistant-propagule forming species, such as Cortinarius, Amphinema and Russula in older-successional stands, has been reported in the native ranges by Nara et al. (2003), Horton et al. (2005), and Kipfer et al. (2010), among many others. We suggest that the dominance of suilloid fungi far from plantations and the dominance of so-called latestage fungi near and inside plantations are consistent with successional and competition-driven patterns in the native ranges, and probably reflect the same underlying ecological processes. If so, life-history traits of ECM fungi as observed in the native range may provide a means for predicting their coinvasive potential: we hypothesize that ECM fungi characteristically associated with late-successional forests will rarely, if ever, facilitate rapid coinvasions, while some early successional fungi may permit rapid expansion of invasion fronts. Common invaders By comparing the sequences recovered in this study with sequences recovered from other sites where Pinaceae species are invasive, we have determined that many of the OTUs detected as invaders on Isla Victoria have been reported associating with Pinaceae spp. elsewhere outside their native range (Table 3). Given that forests in the native range may contain hundreds or thousands of species (Molina et al., 1992;Bueeet al., 2009),the repeated occurrence of a small handful of fungi in geographically and climatically disparate environments strongly suggests that those fungi have exceptional invasive potential and should be carefully monitored (Nu~nez & Dickie, 2014). Notably, these fungi show strong overlap at the genus level with Pinaceae-associating fungi known to possess resistant propagules (Glassman et al., 2015), again highlighting the importance of this character for fungal invasion potential. Relatively few invasive species of any taxonomic affinity have shown the ability to invade both tropical and temperate systems, but those that can include some of the most damaging invasive species globally (Lowe et al., 2000). That Table 3 A list of operational taxonomic units (OTUs) detected in this study also detected with pines elsewhere in the world in studies using similar methodology: Hawaii, Hynson et al. (2013); New Zealand, Walbert et al. (2010) OTUs in New Zealand, Hawaii and Argentina OTUs in Argentina and Hawaii OTUs in Argentina and New Zealand Amphinema 1 Rhizopogon 1 Pseudotomentella 1 Thelephora terrestris Pyronemataceae 1 Tomentella 2 Hebeloma 2 Pyronemataceae 2 Suillus luteus Pseudotomentella tristis Tomentella 3 Amanita muscaria Rhizopogon 2

8 504 Research New Phytologist repeated invasion by a small group of species across several regions has previously gone undetected highlights the difficulty of identifying and tracking cryptic organisms when they invade. Associations with native fungi Despite incomplete coverage in the GenBank database of northern hemisphere fungi, we detected very few Pinaceae-associating ECM fungal OTUs on Isla Victoria not previously reported from the northern hemisphere (Table S1). Of these taxa, two are likely to represent southern hemisphere native fungi: Inocybe OTU2 (recovered from four ECM root tips from two trees: one P. menziesii and one Pinus sp.) and Sebacinaceae OTU1 (recovered from two root tips from a single P. menziesii in this study and previously from both P. menziesii and P. contorta in a glasshouse bioassay; Nu~nez et al., 2009). The ectomycorrhizas of both Sebacinaceae OTU1 and Inocybe OTU2 were morphologically normal, with the cortical Hartig net characteristic of Pinaceae ectomycorrhizas (data not shown). Inocybe OTU2 is nested within a clade that, based on ITS sequences, is apparently represented in GenBank only by species found in the southern hemisphere (Fig. S1). Matheny et al. (2009) report that temperate southern hemisphere Inocybe species are largely derived from northern temperate clades; a taxon nested within a southern temperate clade is thus unlikely to represent a northern hemisphere species derived from southern hemisphere ancestors. While we cannot rule out that Inocybe OTU2 represents a coinvasive northern hemisphere species, we consider it more likely that its association with Pinaceae species is novel. Sebacinaceae OTU1 has been recovered from root tips of Nothofagus alpina and Nothofagus obliqua in Lanin National Park, > 100 km from our field site, in locations > 10 km from any nonnative trees (E. Nouhra, pers. comm.; Nouhra et al., 2013); we have also detected it associating with N. dombeyi on Isla Victoria (J. Hayward, unpublished). Mikola (1970) and Dickie et al. (2010), among others, report that very few, if any, Gondwanan ECM fungi are compatible with Pinaceae species. We agree with Dickie et al. (2010) that ECM associations between fungi native to the southern hemisphere and Pinaceae species are rare: associations between Inocybe OTU2 or Sebacinaeae OTU1 and Pinaceae species make up < 2% of the mycorrhizal root tips we analyzed. However, we suggest that the associations of Inocybe OTU2, and Sebacinaceae OTU1 with Pinaceae species represent the exception to this rule, and are among the first reported instances of Pinaceae species associating with Gondwanan ECM taxa. By contrast, ECM angiosperm hosts native to the northern hemisphere growing on Isla Victoria appear to associate more freely with putative South American native fungi. Quercus robur, the angiosperm tree we sampled most frequently, associated with three additional Inocybe OTUs (4, 5 and 7) likely to represent Gondwanan taxa, and one we consider equivocally Gondwanan or northern hemisphere (OTU 6; Fig. S1). Fagus sylvatica, Q. robur and C. sativa were associated with two Tomentella OTUs (OTUs 4 and 6) for which the top BLAST hits represent putatively Argentinian-native species (Table S1). However, because the biogeography and species richness of Tomentella is poorly understood, we are less confident that these represent genuine novel pairings. The dominant native ECM plant in this study is Nothofagus, a tree more closely related to plants introduced from the northern hemisphere that are members of Fagaceae (Fagus, Quercus, Castanea) than those that are members of Pinaceae. ECM fungi are more likely to associate with plant species within a family or genus (Molina et al., 1992; Ishida et al., 2007), a pattern that may extend to higher levels of biological organization. The apparently higher compatibility of angiosperms with southern hemisphere native ECM fungi may thus be consistent with patterns of conserved association seen in the native region. For both Pinaceae and northern hemisphere angiosperms, the ecological consequences of associating with southern hemisphere native fungi are unclear. While Pinaceae seedlings apparently do not receive benefits from southern hemisphere native fungi (Nu~nez et al., 2009), no study has examined whether older trees are able to receive such benefits in native or invasive settings for most ECM fungi. Early-successional pine seedlings in the native range typically possess few root tips colonized by Inocybe species (Ashkannejhad & Horton, 2006; Kipfer et al., 2010). Both putatively native Pinaceae-associating Inocybe OTUs were detected only with trees > 2 yr in age (Table S2). All angiosperm hosts sampled were mature individuals, so it is not clear whether their associations with native fungi could aid in establishment; however, we have previously reported the association of Sebacinaceae OTU1 with seedlings (Nu~nez et al., 2009). Conclusion We report for the first time a number of ECM fungi with apparently low ability to coinvade, including several genera typically associated in the native region with late-successional environments. By contrast, our findings confirm previous reports that suilloid fungi possess notable potential to invade and to facilitate coinvasions by Pinaceae species. We suggest that the performance of ECM fungi in early-successional areas in the native region may provide valuable clues about their potential to invade and facilitate invasion. ECM fungi with poor potential to invade have potential applications in sustainable forestry, and should be investigated further. Finally, we report that while Pinaceae species apparently rarely associate with southern hemisphere-endemic ECM fungi, such associations do occur, with unknown ecological impact. By contrast, several ECM angiosperm species apparently have greater compatibility with southern hemisphere symbionts. We suggest that the ecological impacts of these associations, both for individual trees and for the invasional process, may prove fertile ground for future investigations. Acknowledgements This research was funded by National Science Foundation grants DEB and REU supplement to T.R.H. and DEB to Daniel Simberloff and M.A.N. Additional funding was provided by a SUNY-ESF seed grant to J.H. We thank Juan Karlanian and Max Reitmann for field help. We acknowledge the assistance of Espacio S.A., and the

9 New Phytologist Research 505 staff and rangers of Parque Nacional Nahuel Huapi. Finally, we gratefully acknowledge four anonymous reviewers for their helpful comments. References Agerer R Colour atlas of ectomycorrhizae. Schw abisch Gm und, Germany: Einhorn-Verlag. Agerer R Exploration types of ectomycorrhizae. Mycorrhiza 11: Ashkannejhad S, Horton TR Ectomycorrhizal ecology under primary succession on coastal sand dunes: interactions involving Pinus contorta, suilloid fungi and deer. New Phytologist 169: Baar J, Horton TR, Kretzer AM, Bruns TD Mycorrhizal colonization of Pinus muricata from resistant propagules after a stand-replacing wildfire. New Phytologist 143: Binggeli P A taxonomic, biogeographical and ecological overview of invasive woody plants. Journal of Vegetation Science 7: Brundrett M, Bougher N, Dell B, Grove T, Malajczuk N Working with mycorrhizas in forestry and agriculture. Canberra, Australia: Australian Centre for International Agricultural Research. Bruns TD Thoughts on the processes that maintain local species diversity of ectomycorrhizal fungi. Plant and Soil 170: Bruns TD, Peay KG, Boynton PJ, Grubisha LC, Hynson NA, Nguyen NH, Rosenstock NP Inoculum potential of Rhizopogon spores increases with time over the first 4 yr of a 99-yr spore burial experiment. New Phytologist 181: Buee M, Reich M, Murat C, Morin E, Nilsson RH, Uroz S, Martin F Pyrosequencing analyses of forest soils reveal an unexpectedly high fungal diversity. New Phytologist 184: Callaway RM, Thelen GC, Rodriguez A, Holben WE Soil biota and exotic plant invasion. Nature 427: Cazares E, Trappe JM Spore dispersal of ectomycorrhizal fungi on a glacier forefront by mammal mycophagy. Mycologia 86: Chu-Chou M, Grace LJ Mycorrhizal fungi of radiata pine in different forests of the north and south islands in New Zealand. Soil Biology and Biochemistry 20: Collier FA, Bidartondo MI Waiting for fungi: the ectomycorrhizal invasion of lowland heathlands. Journal of Ecology 97: Darriba D, Taboada GL, Doallo R, Posada D jmodeltest 2: more models, new heuristics and parallel computing. Nature Methods 9: 772. Deacon JW, Donaldson SJ, Last FT Sequences and interactions of mycorrhizal fungi on birch. Plant and Soil 71: Desprez-Loustau ML, Robin C, Buee M, Courtecuisse R, Garbaye J, Suffert F, Sache I, Rizzo DM The fungal dimension of biological invasions. Trends in Ecology and Evolution 22: Dickie IA, Bolstridge N, Cooper JA, Peltzer DA Co-invasion by Pinus and its mycorrhizal fungi. New Phytologist 187: Dodet M, Collet C When should exotic forest plantation tree species be considered as an invasive threat and how should we treat them? Biological Invasions 14: Edgar RC MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32: Gardes M, Bruns TD ITS primers with enhanced specificity for basidiomycetes-application to the identification of mycorrhizae and rusts. Molecular Ecology 2: Gardes M, Bruns T ITS-RFLP matching for identification of fungi. In: Clapp J, ed. Methods in molecular biology. Species diagnostics protocols. Clifton, NJ, USA: Humana Press, Glassman SI, Peay KG, Talbot JM, Smith DP, Chung JA, Taylor JW, Vilgalys R, Bruns TD A continental view of pine-associated ectomycorrhizal fungal spore banks: a quiescent functional guild with a strong biogeographic pattern. New Phytologist 205: Gouy M, Guindon S, Gascuel O SeaView version 4: a multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Molecular Biology and Evolution 27: 221. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Systematic Biology 59: Hatch AB The role of mycorrhizae in afforestation. Journal of Forestry 34: Hayward J, Horton TR, Pauchard A, Nu~nez MA. In press. A single ectomycorrhizal fungal species can enable a Pinus invasion. Ecology. doi: / Horton TR, Molina R, Hood K Douglas-fir ectomycorrhizae in 40- and 400-year-old stands: mycobiont availability to late successional western hemlock. Mycorrhiza 15: Hynson NA, Merckx VSFT, Perry BA, Treseder KK Identities and distributions of the co-invading ectomycorrhizal fungal symbionts of exotic pines in the Hawaiian Islands. Biological Invasions 15: Ishida TA, Nara K, Hogetsu T Host effects on ectomycorrhizal fungal communities: insight from eight host species in mixed conifer broadleaf forests. New Phytologist 174: Ishida TA, Nara K, Tanaka M, Kinoshita A, Hogetsu T Germination and infectivity of ectomycorrhizal fungal spores in relation to their ecological traits during primary succession. New Phytologist 180: Jumpponen A, Jones KL, David Mattox J, Yaege C Massively parallel 454-sequencing of fungal communities in Quercus spp. ectomycorrhizas indicates seasonal dynamics in urban and rural sites. Molecular Ecology 19: Keane RM, Crawley MJ Exotic plant invasions and the enemy release hypothesis. Trends in Ecology & Evolution 17: Kipfer T, Egli S, Ghazoul J, Moser B, Wohlgemuth T Susceptibility of ectomycorrhizal fungi to soil heating. Fungal Biology 114: Kipfer T, Moser B, Egli S, Wohlgemuth T, Ghazoul J Ectomycorrhiza succession patterns in Pinus sylvestris forests after stand-replacing fire in the Central Alps. Oecologia 167: Koide RT, Shumway DL, Xu B, Sharda JN On temporal partitioning of a community of ectomycorrhizal fungi. New Phytologist 174: Koutche V Estacion forestal de Puerto Achorena, Isla Victoria; suorganizacion y trabajos. Buenos Aires, Argentina: Ministerio de Agricultura, Direccion de Parques Nacionales, Argentina. Kretzer A, Li Y, Szaro TM, Bruns TD Internal transcribed spacer sequences from 38 recognized species of Suillus sensu lato: Phylogenetic and taxonomic implications. Mycologia 88: Legendre P, Legendre LF Numerical ecology. Waltham, MA, USA: Elsevier. Lilleskov EA, Bruns TD, Horton TR, Lee Taylor D, Grogan P Detection of forest stand-level spatial structure in ectomycorrhizal fungal communities. FEMS Microbiology Ecology 49: Lowe S, Browne M, Boudjelas S, De Poorter M of the world s worst invasive alien species: a selection from the global invasive species database. Auckland, New Zealand: The Invasive Species Specialist Group (ISSG). Martin K, Rygiewicz P Fungal-specific PCR primers developed for analysis of the ITS region of environmental DNA extracts. BMC Microbiology 5: 28. Matheny PB, Aime MC, Bougher NL, Buyck B, Desjardin DE, Horak E, Kropp BR, Lodge DJ, Soytong K, Trappe JM et al Out of the Palaeotropics? Historical biogeography and diversification of the cosmopolitan ectomycorrhizal mushroom family Inocybaceae. Journal of Biogeography 36: Mikola P Mycorrhizal inoculation in afforestation. International Review of Forestry Research 3: Molina R, Massicotte H, Trappe JM Specificity phenomena in mycorrhizal symbioses: community-ecological consequences and practical implications. In: Allen M, ed. Mycorrhizal functioning: an integrative plant fungal process. London, UK: Springer, Nara K, Nakaya H, Wu B, Zhou Z, Hogetsu T Underground primary succession of ectomycorrhizal fungi in a volcanic desert on Mount Fuji. New Phytologist 159: Newton AC Towards a functional classification of ectomycorrhizal fungi. Mycorrhiza 2:75 79.