Vine y ar d manag ement syst ems influence arbuscular m ycor r hizal fungi recruitment by grapevine rootstocks in New Zealand

Aims: Arbuscular m y corrhizal fungi (AMF) can perf orm significant functions within sustainable agricultural ecosy stems, including vine y ards. Increased AMF diversity can be beneficial in promoting plant growth and increasing resilience to environmental changes. To effectively utilize AMF communities and their benefits in vine y ard ecosy stems, a better understanding of how management systems influence AMF community composition is needed. Moreo v er, it is unknown whether AMF communities in organically managed vineyards are distinct from those in con v entionally managed vine y ards. Methods and Results: In this study, vine y ards w ere surv e y ed across the Marlborough region, New Zealand to identify the AMF communities colonizing the roots of different rootstocks grafted with Sauvignon Blanc and Pinot Noir in both con v entional and organic sy stems. T he AMF communities were identified based on spores isolated from trap cultures established with the collected grapevine roots, and by next-generation sequencing technologies (Illumina MiSeq). The identified AMF species/genera belonged to Glomeraceae, Entrophosporaceae, and Div ersisporaceae. T he results re v ealed a significant difference in AMF community composition between rootstocks and in their interaction with management systems. Conclusions: These outcomes indicated that vineyard management systems influence AMF recruitment by rootstocks and some rootstocks ma y theref ore be more suited to organic systems due to the AMF communities the y support. T his could pro vide an increased benefit to organic sy stems b y supporting higher biodiv ersity.


Introduction
Grapevine ( Vitis vinifera L.) is considered one of the most important fruit crops in the world based on area of cultivation and economic value (Torregrosa et al. 2015 ).Despite its economic importance, agricultural practices such as plant protection product application and intensive tillage, can negatively impact biodiversity, soil quality, and soil biology and are unsustainable (Barros-Rodriguez et al. 2021 )).The conventional system involves a high number of inputs and intensive use of agricultural land and has a negative impact on the environment, resulting in the degradation and a reduction of ecosystem biodiversity (Giffard et al. 2022 ).In contrast, an organic system is considered more sustainable with a focus on reducing fertilizer, plant protection product inputs by using biological and biodynamic management systems (Altieri 2002 ).It has also been reported that organic management systems positively affect the diversity of several groups of organisms, including arbuscular mycorrhizal fungi (Leksono 2017 ).
Arbuscular mycorrhizal fungi (AMF) are obligate biotrophs that belong to the subphylum Glomeromycotina (Spatafora et al. 2016 ) and colonize > 80% of vascular plants (Smith and Read 2008 ).In grapevine, AMF can promote plant growth, enhance nutrient uptake, and increase plant resistance against pathogenic fungi and nematodes (Linderman and Davis 2001, Schreiner et al. 2007, Gianinazzi et al. 2010, Moukarzel et al. 2022, 2023, Schreiner et al. 2023 ).Most studies on AMF have focused on natural ecosystems, revealing that roots of individual plants are infected by multiple AMF species, often with 10-20 taxa per root sample (Stukenbrock andRosendahl 2005 , Ji et al. 2013 ).Other studies showed that AMF communities were more diverse within some agricultural ecosystems than previously hypothesized (Hijri et al. 2006 ).
The development of new molecular techniques, such as nextgeneration sequencing (NGS), allows researchers to obtain a greater depth of sequencing.However, there is a debate on whether operational taxonomic units (OTUs) identified with NGS represent separate species or whether these methods overestimate AMF richness and diversity (Dumbrell et al. 2010, Gorzelak et al. 2012, House et al. 2016 ).Recent studies have showed that the Illumina MiSeq platform was effective for characterizing the diversity of AMF communities in both the roots and the rhizosphere soils in many agricultural ecosystems, including vineyards ( Van der Heijden et al. 2015, Bouffaud et al. 2016, Cui et al. 2016, Zhu et al. 2016, Berruti et al. 2017, Zeng et al. 2019 ).
To manage AMF communities and their benefits in vineyard ecosystems, a better understanding of how management systems influence AMF diversity and composition is needed.As shown by Oehl et al. ( 2003 ), the intensity of agricultural management has a large influence on AMF community diversity.Soil and foliar fertilization, especially with phosphorus (P), reduces AMF root colonization in many ecosystems, including vineyards (Karagiannidis and Nikalaou 1999 ).Furthermore, the application of herbicides for weed control, which is common in vineyards, has a negative impact on AMF communities and weed/cover crops can promote beneficial soil microbes in the soil, such as AMF (Baumgartner et al. 2004 ).However, research indicated that the diversity of mycorrhizal hosts within the vineyard did not have any significant influence on AMF root colonization and community diversity in grapevines (Baumgartner et al. 2010 ).Another study suggested that weeds such as Plantago lanceolata L. and Tanacetum cinerariifolium could provide a wider spectrum of AMF for colonizing grapevine roots (Radi ć et al. 2012 ).In some small-scale studies, organic management showed to enhance AMF richness (Oehl et al. 2004, 2005, Hijri et al. 2006 ) and colonization levels (Bending et al. 2004 ), but in a more recent work specifically on grapevines, there was no clear evidence that management system affected the diversity of AMF in vineyard soils (Bouffaud et al. 2016 ).The difference in results can be explained either by the limited number of field sites sampled or the use of varying methods to identify AMF.Therefore, it is still unclear whether organic or conventional management differ in their ability to stimulate or suppress AMF species diversity, especially in grapevines.In New Zealand, the arbuscular mycorrhizal fungal communities were mostly driven by the rootstock cultivar in the surveyed vineyards (Moukarzel et al. 2021 ).However, there have been no studies in New Zealand on whether AMF communities found in organically managed vineyards are distinct from those found under conventional vineyards.Thus, the objectives of this study were to (i) determine the AMF community diversity colonizing the roots of different rootstocks grafted with Sauvignon Blanc and Pinot Noir, and (ii) investigate the impact of management systems and the rootstock on the AMF community.

Study sites
Twelve vineyards were chosen for this study and are in the Marlborough district, situated in the South Island of New Zealand.This region was chosen as it represents 71.0% (29 654 hectares) of the total grape production and is the centre of the New Zealand wine industry ( https://www.marlboroughwinenz.com/climate ).Half of the selected sites are managed organically and the other half conventionally.The vineyards follow the standard practices outlined by Sustainable Winegrowers New Zealand ( https://www.nzwine.com/ en/ sustainability/ swnz ).For most selected organic vineyard sites, conventional vineyard sites in proximity (1-3 km apart) planted with the same scion variety, rootstock (if possible), and of relatively similar vine age were selected (Table 1 , Fig. 1 ).The paired organic and conventional vineyards were selected to ensure that the geo-climatic conditions and soil types were similar.The soil physio-chemical properties for each site are presented in the Supplementary Table S1 .

Root sampling and processing
From each vineyard, root samples were directly collected from the soil from each rootstock variety in two sampling directions.Five grapevines per rootstock were randomly selected from separate bays, avoiding the end bay (potential edge effect).From each grapevine, a root sample was collected interrow at 20 cm from the grapevine trunk and a sample within row at 10 cm from the trunk, both at a soil depth of 15 cm.Half of the hand-collected fine grapevine roots (0.5-1 g) from the field were used for setting up trap cultures for AMF spore recovery.The other half of the collected roots (0.5 g) were used for DNA extraction.

AMF trapping cultures and mycorrhizal colonization assessment
Root trap cultures were established for each of the root samples collected from both the inter and within row sites from the grapevine rootstocks at each of the vineyard sampling sites.Trap cultures were set up using a grapevine rootstock of the same cultivar the sample was collected from as bait plant following the method described in Moukarzel et al. ( 2021 ).The plants grew in the greenhouse for 2 years to multiply AMF spores in the pots and to provide sufficient time for the communities to colonize the roots of the baiting plants.The trap cultures were then harvested and four root replicates per treatment were cleared, stained in trypan blue using the optimized staining method of Moukarzel et al. ( 2020 ).The roots were then assessed for percentage of mycorrhizal root length using the magnified gridline intersection method described by Brundrett et al. ( 1996 ).

Spore isolation and identification from trap cultures
AMF spores were recovered using the wet sieving method (Daniels and Skipper 1982 ).For morphological identification to genus level, two to four spores from each morphotype were added to a drop of polyvinyl-lactic acid glycerine (PVLG) mixed 1:1 (v/v) with Melzer's reagent (Brundrett et al. 1994 ) and visualized under the microscope at δ400 × magnification.The identification of AMF was based on spore morphology description in Blaszkowski ( 2012 ).
DNA was extracted from single AMF spores following the method described in Moukarzel et al. ( 2021 ).The AMF spores were identified by sequencing five replicates per spore morphotype (across all the vineyards) using primer set NS31/AML2.PCR was conducted using 0.2 mM dNTPs (0.5 μL), 10 μmol (1 μL) of each of the primers NS31 (TTG-GA GGGCAA GTCTGGTGCC), and AML2 (GAA CCCAAA-CACTTTGGTTTCC), 5 U of Taq DNA polymerase (0.25 μL), 10 × reaction buffer (2.5 μL) (Roche Diagnostics, Mannheim, Germany), and 1 μL of genomic DNA extract.The volume was adjusted to a total volume of 25 μL using 19.75 μL of RNA-free PCR H 2 O.The PCR was run as follows: initial denaturation at 94 • C for 3 min, 30 cycles at 94 • C for 45 s, 63.1 • C for 60 s, and 72 • C for 60 s, followed by a final extension period at 72 • C for 10 min.The PCR products were checked on a 1.5% agarose gel containing 2 μL ethidium bromide (EtBr, 10 mg/mL) for the expected band size of 550 bp.The determination of band sizes was facilitated through the utilization of a 1 kb plus DNA ladder (New England Biolab, #N3232S).Amplified fragments were cleaned using magnetic beads (Beckman Coulter, USA) before sending for sequencing reactions.The PCR products were then sequenced with both primers (forward and reverse) at the Lincoln University Sequencing Facility using Sanger dideoxy sequencing technology (Applied Biosystems, HITACHI, 3500xL, Genetic Analyzer, New Zealand).A consensus DNA sequence was generated using BioEdit software (Hall 1999 ), used to query the MaarjAM ( https:// maarjam.ut.ee/ ) and NCBI databases using the basic local alignment search tool (BLAST), and identified to genus level based on similarity with sequences present in the database.

Root DNA extraction
In-row collected root samples were selected for DNA extraction as they produced more spores in the trap cultures than the intra-row samples.Root samples were washed several times with sterile water, then placed into a mortar and crushed into a powder using liquid nitrogen with a pestle.The powdered tissue (30-50 mg) was then loaded into a 1.5 mL microcentrifuge power bead tube (DNeasy PowerPlant Pro Kit, Qiagen Laboratories, Hilden, Germany) and suspended in 400 μL Qiagen lysis buffer.The tubes were vortexed at maximum speed for 10 min.Total DNA was then extracted following the manufacturer's instructions with a final elution volume of 30 μL.

Libr ary prepar ation and PCR amplification
Libraries of small subunit ribosomal RNA (18S region) fragment amplicons were prepared for each sample.A single PCR was performed using primers NS31 and AML2.This primer set was modified for usage within the Illumina MiSeq platform according to the protocols of Caporaso et al. ( 2012) and Morgan and Egerton-Warburton ( 2017 ) as follows: NS31 (forward primer) forward Illumina adapter P5 (AA TGA T A CGGCGA CCA CCGA GA TCT A CA C) + forward pad (T A TGGT AA TT) + forward linker (CT) + NS31 forward primer (TTGGA GGGCAA GTCTGGTGCC); AML2 (reverse primer): forward Illumina adapter P7 (C AAGC A-GAA GA CGGCA T A CGA GAT) + Golay barcodes (example: GCTGT ACGGA TT) + reverse pad (AGTC AGTC AG) + forward linker (AC) + AML2 reverse primer (GAACCCAAA-CACTTTGGTTTCC). Reverse primer constructs were modified to include a 12-base Golay index to enable demultiplexing during data processing ( Supplementary Table S2 ).
For each sample, PCR was carried out using 0.2 mM dNTPs (0.5 μL), 10 μmol (1 μL) of forward primer NS31 iL, 10 μmol (1 μL) of uniquely barcoded reverse primer AML2 iL, 5 U of Taq DNA polymerase (0.25 μL), 10 μg/ μL bovine serum albumin (0.25 μL BSA), 10 × reaction buffer (2.5 μL) (Roche Diagnostics, Mannheim, Germany), and 1 μL of genomic DNA extract.The volume was adjusted using 18.5 μL of RNA-free PCR H 2 O to a total volume of 25 μL.The PCR conditions were as described previously and were performed in duplicate for each sample, checked for the correct band size on a 1.5% Agarose gel, and pooled following sample normalization and PCR clean-up using UltraClean PCR Clean-up kit (QIAGEN-MO BIO).The samples were sent to the Auckland genomics facility (University of Auckland, New Zealand) and sequenced on an Illumina MiSeq using version 2, 2 × 250-bp paired-end chemistry.

Bioinformatics analysis
Primers were first trimmed from the raw reads using Cutadapt v2.8 (Martin 2011 ), followed by processing the reads using USEARCH v8.0 (Edgar 2010 ).Briefly, reads < 200 nucleotides, with average Phred quality score < 25 or with internal ambiguous bases were discarded.Read sequences with a maximum expected error rate of 1.0 were also removed.The remaining high-quality filtered reads were then merged and de-replicated, followed by removal of chimeras and singletons using USEARCH.Reads were clustered into operational taxonomic units (OTUs) at 97% sequence similarity using UP-ARSE (Edgar 2013 ).OTUs were queried by BLASTn searches against the NCBI and MaarjAM databases using cutoffs of query coverage at 85%, and 97-99% identities for genus and species level taxonomic classifications.During data processing, it was found that most sequenced amplicons ( ∼530 bp) were too long to allow for overlap with the Illumina MiSeq version 2, 2 × 250-bp, and thus could not be aligned and assembled.Therefore, an OTU table was generated for both the forward and reverse sequences separately using the methods described above.The AMF identity data from the two databases were compiled into one file by comparing the identity information generated by the two databases (data not shown).The forward sequences were used for the analysis as they generated better-quality sequences that gave a higher identity compared to the reverse sequences, as also suggested by Davison et al. ( 2012 ).Samples were rarefied to the minimum library size (9997) before further statistical analysis ( Supplementary Fig. S1 ).

Statistical analysis
Statistical analysis of mycorrhizal colonization was done using analysis of variance followed by a Tukey's test to examine differences between management systems.To compare the AMF communities between scion varieties, rootstocks, management systems, and their interactions, a further analysis using MicrobiomeAnalyst ( https://www.microbiomeanalyst.ca/) was carried out as suggested by Sergaki et al. ( 2018 ).A phyloseq file was generated by combining the OTU, taxonomy (phylum to species level), and metadata (sample variables) files together.Alpha diversity (Shannon and Chao1) analysis was performed using the phyloseq package (McMurdie and Holmes 2013 ).The results were plotted across samples and viewed as box plots for each group or experimental factor.Beta diversity analysis was performed using the Bray-Curtis dissimilarity index and was visualized by principal coordinate analysis (PCoA) to show AMF differences between samples and groups (McMurdie and Holmes 2013 ).Statistical significance of the clustering pattern in ordination plots was evaluated using permutational analysis of variance (PERMANOVA) test.A permutation-based multivariate analysis of variance was utilized to assess a matrix of pairwise distances, partitioning both inter-group and intragroup distances (Massa et al. 2020 ).Furthermore, a nonparametric factorial Kruskal-Wallis sum-rank test was employed for LEfSe (linear discriminatory analysis effect size).LEfSe analysis employs a non-parametric factorial Kruskal-Wallis sum-rank test to identify taxa with significant differential abundance across groups.Subsequently, it validates biological significance through pairwise unpaired Wilcoxon rank-sum tests and estimates effect sizes using LDA analysis for each differentially abundant taxon.Features were deemed statistically significant if they met the criteria of an adjusted P -value < .05 and an LDA score of at least 1.0 (Segata et al. 2011 ; http:// huttenhower.sph.harvard.edu/lefse/ ).

AMF colonization confirmation and assessment
Staining the roots with trypan blue confirmed colonization by AMF, with typical AMF structures such as hyphae, arbuscules, and vesicles present in all roots sampled from the trap cultures (Fig. 2 a).Roots directly sampled from the field that were checked prior to trap culture initiation were also shown to be colonized by AMF (Fig. 2 b).The percentage of mycorrhizal colonization varied between 60% in root samples collected from organic Pinot Noir with Riparia Gloire rootstock and 87% for SO4 rootstock inoculated with root samples collected from conventional Sauvignon Blanc with SO4 rootstock (Fig. 3 a).The ANOVA test showed that the roots from the trap cultures set up for root material from the different vineyards differ significantly ( P = .002)in their percentage AMF colonization ( Supplementary Table S3 ).The AMF colonization in rootstock SO4 in trap cultures inoculated with root samples collected from conventional Sauvignon Blanc grafted to rootstock SO4 was significantly higher compared to root samples from SO4 rootstock inoculated with roots from the organic Pinot Noir and Sauvignon Blanc grafted to SO4 rootstock (by 24-26%, respectively), rootstock 5C inoculated with root samples collected from organic Sauvignon Blanc grafted to rootstock 5C (by 22%), and Riparia Gloire rootstock inoculated with roots from the organic Sauvignon Blanc grafted to rootstock Riparia Gloire (by 22%).Vines grown with roots derived from conventional vineyards had significantly ( P < .001)more mycorrhizal colonization than vines grown with roots derived from organic vineyards (Fig. 3 b), with colonization of roots from conventional vineyards being 10% higher than those from organic vineyards ( Supplementary Table S4 ).

Identification of arbuscular mycorrhizal fungi spores produced in trap cultures
Five spore morphotypes were isolated from the trap cultures (Table 2 ) generated from roots sampled from the different  grapevine rootstocks and varieties in both conventional and organic vineyards.Sampling position, inter or intra row, did not affect ( P = .263)the diversity of AMF communities (data not shown).However, there were differences in AMF community diversity between the 12 vineyards based on observations of spore morphotypes and abundance in the trap cultures (

Next generation sequencing analysis of AMF communities
The total number of OTUs obtained was 3822 of which 3321 OTUs belonged to AMF, 105 OTUs to bacteria, 366 OTUs to other fungi, and 30 OTUs to Vitis sp.Data filtration was performed to remove all bacteria, plant, fungi, and other unassigned OTUs that were not AMF.The unclassified, undetermined, and unassigned sequences were also removed from the analysis.

Community structure and composition
The identified AMF community in grapevine roots across the two management systems belonged to three main families: Glomeraceae, Entrophosporaceae, and Diversisporaceae.The most abundant AMF genera in organic vineyards and conventional were in the same order: Glomus , followed by Entrophospora , Funneliformis , Rhizophagus , and Div er sispora (Fig. 4 a; Table 3 ).The relative abundance of Div er sispora and Rhizophagus was significantly higher ( P = .034,P = .048,respectively) in organic vineyards compared to conventional vineyards.There was no significant difference ( P > .05) in relative abundance of other genera between the management systems (Fig. 4 a).The relative abundance was significantly different with Funneliformis ( P = .035)between the scion varieties.Its relative abundance was higher in Pinot Noir (4%) compared with Sauvignon Blanc (0.04%) (Fig. 4 b).

Alpha di ver sity
Shannon and Chao1 diversity indices showed no significant differences between the management systems ( P = .846and P = .291,respectively) (Fig. 5 a and b).There was a significant difference in the Shannon diversity index for rootstock alone ( P = .004and P = .008,respectively) and interaction between management system and the rootstock ( P = .019and P = .037,respectively) (Fig. 5 c and e) ( Supplementary Table S6 ).The results of the post-hoc pairwise comparison ( Supplementary Tables S7 -S8 ) showed that rootstock 101-14 was significantly higher in species richness and abundance than rootstocks Schwarzmann ( P = .039)and Riparia Gloire ( P = .014).In addition, SO4 was significantly higher in species richness and abundance than Riparia Gloire ( P = .021).Conventional vineyards with rootstock 101-14 had significantly higher species diversity compared to organic vineyards with Schwarzmann ( P = .028)and conventional vineyards with Riparia Gloire and 3309C ( P = .034and P = .038,respectively) (Fig. 5 c).Moreover, organic vineyards with Schwarzmann had significantly higher species diversity compared to conventional vineyards with rootstock SO4 ( P = .037).No significant differences were shown using the Chao1 diversity index for the management system ( P = .264),rootstock ( P = .729),and their interactions ( P = .587).
There was no significant difference in the Shannon and Chao1 diversity indices between Pinot Noir and Sauvignon Blanc ( P = .324and P = .425,respectively), or in the interaction between management practice and scion variety ( P = .241and P = .516,respectively) ( Supplementary Table S6 ).

Beta di ver sity
There was a significant difference in the beta diversity between samples from organic to conventional vineyards (Fig. 6 a; P = .047)and between the two scion varieties (Fig. 6 d; P = .047)but not in the interaction between management system and scion variety (Fig. 6 e; P = .067).In addition, no significant difference in beta diversity was observed between the rootstocks alone (Fig. 6 b; P = .667)or the interaction between the rootstock and management system (Fig. 6 c

Soil analysis
The PERMANOVA analysis of soil physio-chemical properties ( Supplementary Table S9 ) showed no significant ( P > .05)difference in these properties across the vineyards.The only significant effect was observed with soil moisture content ( P = .011).

Discussion
This study is the first to compare AMF community composition and structure in conventional and organic vineyards in New Zealand using complementary techniques (trap culture and Illumina MiSeq).These two techniques allowed comprehensive identification of AMF communities colonizing grapevines.The most abundant AMF species amplified from grapevine root samples were those within the Glomeraceae.The dominance of this family has also been observed in other vineyard studies (Balestrini et al. 2010, Lumini et al. 2010, Holland et al. 2013, Trouvelot et al. 2015, Massa et al. 2020 ) and other agricultural systems (Cesaro et al. 2008, Berruti et al. 2016 ).AMF communities were also dominated by Entrophosporaceae, followed by less abundant AMF species belonging to Diversisporaceae.Other studies have also detected species associated with these families in vineyards.Sequences of phylotypes assigned to Entrophosporaceae were observed in Oregon vineyards in the USA (Schreiner and Mihara 2009 ).Balestrini et al. ( 2010 ) reported that members of the Diversisporaceae family were mostly found in sandy vineyard soils, but in this study, they were also found in vineyards with clayloamy soil texture with sand composition varying from 16 to 46%.In the study of Schreiner and Mihara ( 2009 ), Paraglomeraceae and Archaeosporaceae were detected using the primer set ARCH131/ITS4 that specifically target these two families.However, in the present study, these two families were detected using the barcoded primer set NS31/AML2, which suggested that this primer set is efficient in detecting the AMF families previously reported by other studies (Davison et al. 2012 , Morgan andEgerton-Warburton 2017 ).Some studies have highlighted the dominance of members of the Acaulosporaceae family in vine roots (Schubert andCravero 1985 , Oehl et al. 2004 ), while in other studies (Hempel et al. 2007 , Schreiner andMihara 2009 ), Acaulosporaceae were not found in grapevine, suggesting that the presence of this AMF family is variable in vineyards or that the method used did not amplify Acaulosporaceae from grapevine roots.Similar results were observed in a related study of the AMF communities associated with grapevine rootstocks, where no Acaulospora species were identified from the sequenced DGGE bands   ( Moukarzel et al. 2021 ).Schreiner and Mihara ( 2009 ) showed that species within the genus Acaulospora were found as spores in soil, but no Acaulospora spp.were amplified from roots.This could indicate that species within this family are not frequent colonizers of grapevine roots in Marlborough vineyards.
Grapevine roots sampled both in the field and from the trap cultures were colonized by AMF, as observed by the presence of hyphae, vesicles, and arbuscules in the microscopic assessment of the root samples.Mycorrhizal colonization of roots from the trap cultures showed that the colonization of roots derived from the conventional vineyard trap cultures was 10% higher than the root derived from the organic vineyard trap cultures.This result contrasts with the generally accepted opinion that AMF colonization is enhanced in plant roots under organic farming systems.In a recent study, it was found that colonization intensity was higher on conventional farms, as observed in this study (Chen et al. 2022 ).However, the different results in this study could be due to the variation in the quantity of initial root inoculum in the pots as this varied between 0.1 and 1.0 g between samples (Moukarzel et al. 2021 ).It was more difficult to collect roots in organic vineyards as they were not as close to the soil surface as in conventional vineyards.However, root mycorrhizal colonization patterns, including degree of colonization and presence of different structures, are known to be AMF species/genus dependent, with some species/genera colonizing the internal root tissue more intensively than others (Johansen et al. 2016 ).The difference seen could therefore be related to the difference in the species/genera observed in the different vineyards across the two management systems.Studies have shown that the propagation of AMF in trap cultures may be difficult because the exact natural conditions cannot be replicated, which causes a bias towards species that are more tolerant to greenhouse conditions, such as Glomus and Funneliformis species (Cuenca et al. 2004, Oliveira et al. 2010). Trejo-Aguilar et al. ( 2013 ) demonstrated that species belonging to Glomerales were the most persistent in trap cultures, which was also seen in this study .However, in this study , it was shown that Glomus spp.were also dominant when sequencing was performed on the roots collected from these vineyards.This indicated that the trap cultures were a true reflection of Glomus spp.dominance in grapevine roots from Marlborough vineyards.
The abundance of species from the Glomeraceae family in both organic and conventional vineyards indicated that a diverse number of species in this family are present and easily colonize grapevine roots.According to Johansen et al. ( 2016 ), Glomeraceae is also abundant in New Zealand ecosystems.Further, the dominance of AMF species belonging to this family seen in this study of a managed vineyard ecosystem could be explained by the relatively high growth rate and the rapid recovery of hyphal networks by members of this AMF family following disturbance caused by agricultural practices (Berruti et al. 2016 ).This is also supported by the observed higher abundance of Glomeraceae in the organic vineyards compared to conventional, where mechanical weeding is a practice.
The results of this study showed that the relative abundance and the presence/absence of AMF species differed between the rootstocks in both the trap culture and the high-throughput sequencing results.This is in agreement with an earlier study where different community compositions of AMF in rootstocks were demonstrated in trap culture and denaturing gradient gel electrophoresis (Moukarzel et al. 2021 ).A difference in alpha diversity richness was also observed between the rootstocks and in interaction with the management systems, based on Shannon index (Chong et al. 2020 ).This was observed with the Shannon diversity for rootstock SO4, where there was a higher diversity index under organic compared with conventional management, suggesting this rootstock supports colonization by a wider diversity of AMF than the other rootstocks assessed in this study.This could indicate that the management system in a vineyard influences the plant-microbe interactions that lead to colonization.Schmidt et al. ( 2019 ) stated that microbial communities are shaped by interactions between agricultural management and host selection processes.This was seen in this study, where rootstocks played an important role in the selection of AMF communities in a particular management system.This was demonstrated by some rootstocks being colonized by a higher diversity of AMF in conventional systems, whereas other rootstocks had higher AMF diversity in organic systems.This indicates that the maximum impact from AMF in any vineyard management system may depend to some degree on the rootstock that was planted.However, for some of the rootstock comparisons, the comparison analysis was within scion (Riparia Gloire and 5C-Schwarzmann and 3309C only 1 vineyard), but for 101-14 and SO4, the comparison included both Pinot Noir and Sauvignon Blanc, suggesting that scion may also impact the AMF-rootstock interaction.A study suggested that Citrus scion genotype had a greater influence on the AMF community structure than that of the rootstock, where the physical root-AMF association occurs (Song et al. 2015 ).Further studies using different scion genotypes may provide information on the contribution of the scion on shaping the AMF communities in vineyards.
The outcomes of this study showed a difference in AMF community composition (Beta diversity) between management systems and varieties.This indicated that the AMF communities associated with grapevines in the Marlborough vineyard region are affected by these factors.Montes-Borrego et al. ( 2014 ) reported that AMF community composition of apple trees was affected by the agricultural management (organic vs. integrated).The no differences in AMF community composition between rootstocks and their interaction with management systems means that it is the relative composition in each rootstock that is different.
The soil analysis results of this study showed no significant difference between soil properties across the studied vineyards.This indicated that the differences in AMF communities observed are not attributable to any differences in edaphic factors.The only significant effect was observed with soil moisture content, which varied between the vineyards from 6.5 to 18.76%.This variation in soil moisture could be related to the vines being irrigated or not prior to sampling the vineyards.Recent studies have shown that soil moisture is an important predictor of bacterial alpha diversity while spatial distances among samples were a better predictor of fungal beta diversity than soil properties (Li et al. 2020 ).
The AMF species recovered from the trap cultures belonged to Glomus , Funneliformis , Entrophospora , and Ambispora genera.The taxa identified by sequencing of the different AMF spore morphotypes corresponded to those determined from the field root samples using MiSeq.However, species that belong to Rhizophagus and Div er sispora genera were only detected by sequencing the field root samples and were not recovered in the trap cultures.The absence of these genera in the trap cultures could indicate that the conditions in the trap cultures were not favourable for spore formation by these genera.The trap culture results also showed that Entrophospora spp.were not present in organic vineyards but are present in conventional vineyards.This corresponded with the MiSeq results, where Entrophospora spp.were more abundant in conventional vineyards than in organic vineyards.This highlights that the use of multiple techniques (trap cultures and MiSeq) is required to gather a full picture of AMF genera.
As expected, the metabarcoding provided greater depth of AMF taxa identification to genus level and allowed detection of some AMF species with good taxonomic resolution.Studies demonstrated that MiSeq-Illumina sequencing is a robust technique that also allows the study of AMF communities in terms of relative abundances (Egan et al. 2018, Panneerselvam et al. 2020 ).However, there are some limitations to using MiSeq version 2 × 250 bp with barcoded samples as it does not allow the assembly of forward and reverse reads, resulting in the use of one-direction reads, which restricted the identification of taxa to species level.The 2 × 250-bp paired-end chemistry is the same method used in recent studies (Morgan and Egerton-Warburton 2017, Suzuki et al. 2020, Ramana et al. 2023 ).The use of pyrosequencing could have enabled a better resolution to species/family and to resolve more AMF taxa, as demonstrated by the study on Agathis australis in New Zealand using the same primer set NS31/AML2 (Padamsee et al. 2016 ).However, the pyrosequencing platform is obsolete.Therefore, future studies using sequencing platforms that allow longer sequences to be assembled will improve the ability to identify AMF to species level (House et al. 2016, Padamsee et al. 2016, Morgan and Egerton-Warburton 2017 ).
Several sequences in this study were not assigned to any species-level accession even when the results were compiled from the two databases (NCBI and MaarjAM).This indicated that they could be novel AMF species that are present in New Zealand.This was also seen in another study, where some AMF species within the samples from Agathis australis in New Zealand were suggested to be unique to Agathis australis (Padamsee et al. 2016 ).In another study by Morgan and Egerton-Warburton ( 2017 ), a large number of OTUs could not be assigned to any species-level accession in the Maar-jAM database, suggesting that novel AMF species might occur in the Yucatan Peninsula in Mexico.This result could also be due to the lack of a comprehensive reference database of AMF sequences from named AMF species, which limits the capacity to identify AMF to species level as suggested by House et al. ( 2016 ).In this study, it could also reflect the poor characterization of AMF species in New Zealand, which was also reported by Russell et al. ( 2002 ) and analysed in Dickie and Holdaway ( 2010 ).
In conclusion, this is the first study to carry out a comprehensive analysis of the community composition and structure of AMF associated with grapevine roots in New Zealand using high-throughput next generation sequencing (Illumina MiSeq).This research showed that rootstocks and their interaction with management systems are main drivers of the AMF community colonizing the roots.A rootstock will select a particular AMF community from the diverse community present at a site.This means that some rootstocks might be more suited to organic systems due to the AMF communities they support under this management system.This could provide an increased benefit under organic systems by supporting higher biodiversity.Therefore, it is important to ensure that the vineyard/nursery sites have a diverse community of AMF that the rootstock can select from to provide the most benefit.This could be done by minimizing agricultural practices that negatively affect these communities, such as mechanical weeding, fertilization, and tillage (review by Trouvelot et al. 2015 ), and adopting practices that could influence the formation and the diversity of AMF in the vineyards, such as a cover cropping system (Radic et al. 2012 ).Further work is also needed to elucidate and identify the benefits of different AMF communities to grapevine productivity.

Figure 1 .
Figure 1.The location and distance of the 12 vine y ards used in this study.

Figure 2 .
Figure 2. R epresentativ e micrographs confirming AMF colonization of trypan blue-stained grape vine roots.(a) Grape vine root from trap cultures sho wing typical AMF colonization.(b) Grapevine root sampled from the field showing typical AMF colonization.A: arbuscules, H: hyphae, and V: vesicle.The scale bar represents 100 μm.

Figure 3 .
Figure 3. (a) Percentage arbuscular m y corrhizal fungi colonization of stained grapevine roots from the trap cultures produced from roots of different rootstocks (1 0 1-14, 5C, SO4, Schwarzman, Riparia Gloire, and 3309C) inoculated with root samples collected from 12 different vine y ards with 2 management systems (conventional and organic), 2 varieties (Pinot Noir and Sauvignon Blanc), and the same rootstocks used for trap cultures.(b) Percentage arbuscular m y corrhizal fungi colonization of stained grapevine roots from the trap cultures between two management vineyard systems (con v entional and organic).Bars with different letters are significantly different ( P ≤ .05).Error bars show means ± SE.Mean of four samples.Sch: Schw arzmann, R G: Riparia Gloire, Con v: con v entional, and org: organic.

Figure 4 .
Figure 4. R elativ e abundance (%) of AMF taxa at genus le v el betw een (a) management sy stems, (b) v arieties, and (c) rootstocks sampled in vine y ards.

Figure 5 .
Figure 5. Alpha diversity measured by Shannon and Chao1 diversity indices plotted for (a-b) management systems, (c-d) rootstocks, and (e-f) management and rootstock.The line inside the box represents the median, while the whiskers represent the lowest and highest values within the 4.5 for Shannon and 20 0 0 for Chao1 interquartile range (IQR).The dots represent the samples and the black diamond dots represent the average of the samples.Alpha diversity analysis was performed using the phyloseq package of MicrobiomeAnalyst ( https://www.microbiomeanalyst.ca).

Table 1 .
Summary information of the 12 vine y ards used in this study, including variety, rootstock, year of planting, and management systems.

Table 3 .
AMF identification to genus and species le v el based on blast information, including accession number based on the MaarjAM and NCBI databases in both conventional and organic systems.