1. Evolutionary Biology
  2. Genetics and Genomics
Download icon

Single nucleus sequencing reveals evidence of inter-nucleus recombination in arbuscular mycorrhizal fungi

  1. Eric CH Chen
  2. Stephanie Mathieu
  3. Anne Hoffrichter
  4. Kinga Sedzielewska-Toro
  5. Max Peart
  6. Adrian Pelin
  7. Steve Ndikumana
  8. Jeanne Ropars
  9. Steven Dreissig
  10. Jorg Fuchs
  11. Andreas Brachmann
  12. Nicolas Corradi  Is a corresponding author
  1. University of Ottawa, Canada
  2. Ludwig Maximilian University of Munich, Germany
  3. Leibniz Institute of Plant Genetics and Crop Plant Research, Germany
Research Article
  • Cited 0
  • Views 992
  • Annotations
Cite this article as: eLife 2018;7:e39813 doi: 10.7554/eLife.39813

Abstract

Eukaryotes thought to have evolved clonally for millions of years are referred to as ancient asexuals. The oldest group among these are the arbuscular mycorrhizal fungi (AMF), which are plant symbionts harboring hundreds of nuclei within one continuous cytoplasm. Some AMF strains (dikaryons) harbor two co-existing nucleotypes but there is no direct evidence that such nuclei recombine in this life-stage, as is expected for sexual fungi. Here, we show that AMF nuclei with distinct genotypes can undergo recombination. Inter-nuclear genetic exchange varies in frequency among strains, and despite recombination all nuclear genomes have an average similarity of at least 99.8%. The present study demonstrates that AMF can generate genetic diversity via meiotic-like processes in the absence of observable mating. The AMF dikaryotic life-stage is a primary source of nuclear variability in these organisms, highlighting its potential for strain enhancement of these symbionts.

https://doi.org/10.7554/eLife.39813.001

Introduction

Mating between compatible partners allows organisms to create genetic variation within populations and provides them with opportunities for adaptation to environmental change. In contrast, long-term clonal evolution should lead to the accumulation of deleterious mutations within lineages and, ultimately, to extinction (Kondrashov, 1988; Agrawal, 2001; Kaiser and Charlesworth, 2009; Keightley and Eyre-Walker, 2000; McDonald et al., 2016). Some eukaryotic groups have seemingly challenged this theory by evolving for millions of years without observable sexual reproduction. Such organisms have been referred to as ancient asexual scandals, and prominent examples of these evolutionary oddities have included the bdelloid rotifers and the arbuscular mycorrhizal fungi (AMF) (Normark et al., 2003; Judson and Normark, 1996; Sanders, 1999). In rotifers, genome diversity can be generated by reshaping chromosomes through horizontal gene transfers (Flot et al., 2013; Debortoli et al., 2016; Signorovitch et al., 2016). Other asexual mechanisms that increase genetic variability in clonal taxa include transpositions, DNA replication or inter- or intra-chromosomal rearrangements (Faino et al., 2016; Croll, 2014; Seidl and Thomma, 2014). Furthermore, the absence of observable mating does not necessarily indicate that sex is absent as, for example, several organisms long thought to be clonal were eventually found to harbor many genomic signatures of sexual reproduction within their genomes. Evidence of such cryptic sexuality includes the presence of meiosis-specific genes, mating-type loci, sex pheromones, and evidence of inter-strain recombination (Taylor et al., 2015).

AMF are ecologically successful organisms that have long been characterized as having lacked sexual reproduction for hundreds of millions of years (Sanders and Croll, 2010). These fungi form obligate symbioses with the roots of most land plants, and their presence in the soil can improve plant fitness and biodiversity (James et al., 1998; van der Heijden et al., 1998; van der Heijden et al., 2015). Their capacity to colonize early plant lineages such as liverworts and/or hornworts, and the presence of an AMF symbiotic signaling pathway in algal ancestors of land plants, are some indications that AMF-plant associations may have been established when plants colonized land (Bonfante and Selosse, 2010; Delaux et al., 2015). The AMF mycelium is unique among eukaryotes as it harbors hundreds to thousands of nuclei within one continuous cytoplasm, and there is no known life stage with one or two nuclei. Although sexual reproduction is not formally observed in AMF, these fungi contain many genomic regions with homology to those involved in mating in distant relatives (Halary et al., 2011; Riley and Corradi, 2013; Riley et al., 2014), suggesting that they are able to undergo sexual reproduction. The cellular mechanisms that could trigger sexual processes in AMF have however long been elusive.

Using genome analyses, it was recently shown that isolates of the model AMF Rhizophagus irregularis are either homokaryotic – that is co-existing nuclei harbor one putative mating-type (MAT)-locus - or dikaryotic – two populations with different nuclear genotypes coexist in the mycelium, each with a divergent MAT-locus (Ropars et al., 2016). The MAT-locus is the genomic region that gives sexual identity to fungal isolates, and isolates of the AMF genus Rhizophagus harbor a genomic location with significant similarities to the MAT-locus of basidiomycetes (Ropars et al., 2016). AMF dikaryons were proposed to originate from hyphal fusion and nuclear exchange (plasmogamy) between compatible homokaryons (Ropars et al., 2016). In dikaryotic fungi, compatible nuclei fuse (karyogamy) and undergo recombination, but there is no direct evidence yet that these sexual processes occur in AMF dikaryons. In particular, although inter-isolate recombination has been inferred to exist based on analyses of single loci (Riley et al., 2014; Croll and Sanders, 2009; Gandolfi et al., 2003; den Bakker et al., 2010), direct evidence of genetic exchange between nuclei of opposite MAT-loci co-existing within the same mycelium, as is expected for fungal mating processes, has never been reported.

Results

Here, we tested the hypothesis that karyogamy and inter-nuclear recombination occur in the AMF mycelium using a single-nucleus sequencing approach. Sequenced nuclei were isolated from seven isolates of the genus Rhizophagus, including three known Rhizophagus irregularis dikaryons (A4, A5; SL1 (Ropars et al., 2016)), two R. irregularis homokaryons (A1, C2 (Ropars et al., 2016)), and two closely related species (Rhizophagus diaphanus MUCL-43196, Rhizophagus cerebriforme DAOM-227022). Genome data from R. irregularis SL1 (also known as DAOM-240409), R. diaphanus and R. cerebriforme represent new additions to public databases from representatives of the Glomeromycotina, and all sequencing data were assembled using SPAdes (Bankevich et al., 2012) to facilitate comparisons with published genome data (Ropars et al., 2016).

As expected for an AMF dikaryon, the SL1 assembly contains two putative MAT-loci, with a genome architecture identical to that of known R. irregularis dikaryons (Figure 1A). The isolate also harbors the high allele frequency peak at 0.5 characteristic of AMF dikaryons (Table 1, Figure 1B) (Ropars et al., 2016; Corradi and Brachmann, 2017). Despite these similarities, the SL1 assembly is more fragmented than those of all other Rhizophagus relatives (Table 1). Higher fragmentation is also seen in other known dikaryons (Table 1; Ropars et al., 2016), a feature that probably stems from a higher intra-mycelial genetic diversity in these isolates. Importantly, BUSCO, K-mer graphs and transposable elements (TE) analyses all show that all isolates analysed in this study share highly similar gene repertoires, genome sizes and TE counts (Table 2, Supplementary file 1). Thus, the higher assembly fragmentation of SL1 may result from other factors such as, for example, the presence of higher heterokaryosis and/or inter-nuclear recombination. In contrast to SL1, the R. diaphanus and R. cerebriforme assemblies harbor only one MAT-locus and show a conventional homokaryotic allele frequency pattern (Figure 1B).

A.Predicted MAT-locus of R. irregularis dikaryon isolate SL1.

(B) Genome-wide allele frequency of homokaryons and dikaryons. R. irregularis SL1 shows the hallmark 50:50 ratio reference to alternate allele frequency of dikaryotic isolates. The R. irregularis plots for A1, A4, A5 and C2 represent new analyses of publicly available genome sequence data obtained by (Ropars et al., 2016). The homokaryotic isolates R. irregularis A1 and C2, and R. cerebriforme and R. diaphanus are also shown as comparison. Blue dashed line highlights the 0.5 allele frequency.

https://doi.org/10.7554/eLife.39813.002
Table 1
Summary statistics of genomes and nuclei analyzed in this study. * Values from (Ropars et al., 2016).
https://doi.org/10.7554/eLife.39813.003
Rhizophagus irregularisRhizophagus cerebriformeRhizophagus diaphanus
SL1A1A4A5C2
A. Genome Assembly Statistics
Assembly Coverage136x68x95x76x96x150x120x
Number of Scaffolds29,27911,301*11,380*14,626*10,857*15,08715,496
Assembly Size (Kb)211,501125,869*138,327*131,461*122,873*171,896170,781
Assembly SNP/Kb0.500.250.740.790.350.410.23
B. Single Nucleus Statistics
Number of Nuclei Analyzed16.012.014.08.09.015.012.0
Average Assembly Coverage14.08%57.30%46.84%58.40%61.42%22.39%11.45%
Average Position Depth11.818.122.122.722.915.412.3
Number of SNPs Against Reference
-- Total SNPs346,382241,294423,166291,025251,648262,030101,681
-- Average SNPs Per Nuclei - basic filtering35,47351,43573,34976,05558,15830,53410,554
-- Average Divergence with reference - basic filtering0.12%0.07%0.12%0.10%0.08%0.08%0.06%
Average - Inter-Nucleus Divergence - basic filtering0.38%0.13%0.24%0.16%0.14%0.24%0.21%
Table 2
BUSCO and K-mer assembly size estimation
https://doi.org/10.7554/eLife.39813.004
BUSCO analysisK-mer estimation
Genes foundCompletenessPredicted genome size
A. R. irregularis
A126892.41%131.7 Mb
C226390.69%148.6 Mb
A426290.34%143.5 Mb
A526792.07%130.1 Mb
SL126591.38%146.5 Mb
B. R. cerebriforme26792.07%119.7 Mb
C. R. diaphanus26792.07%121.1 Mb

To find direct evidence of inter-nuclear recombination in AMF, we obtained and compared partial genome sequences from 86 single nuclei – that is single haploid genotypes – isolated from R. irregularis dikaryons (SL1, n = 15; A4, n = 14; A5, n = 8), R. irregularis homokaryons (A1, n = 12; C2, n = 9), and from the species R. diaphanus (n = 12) and R. cerebriforme (n = 15). Nuclei were isolated and analysed using a combination of fluorescence-activated cell sorting (FACS), whole genome amplification (WGA) (Ropars et al., 2016), and Illumina sequencing. Each nucleus was genotyped by mapping paired-end Illumina reads obtained from WGA-based DNA against their respective genome reference. The combination of PCR-based WGA and Illumina sequencing methods results in variation in average depth position and reference coverage among nuclei and isolates (Table 1). Using basic filtering methodologies (see Material and Methods), we find that nuclear genomes diverge from their respective reference from 0.06% for R. diaphanus, to a maximum of 0.12% for SL1 and A4 (Table 1). Using this method, pairwise nuclear genome comparisons show a high average inter-nuclear similarity ranging from 99.62% in SL1 to a maximum of 99.87% in A1 (Supplementary file 2). When present, variability is scattered across the genome (Supplementary file 3).

It is noteworthy that most attempts to validate SNP scored using basic filtering methods resulted in the identification of many false positives. Specifically, only 4% (6 out of 148 tested) of SNPs identified using the abovementioned methods were confirmed using PCR and Sanger sequencing procedures performed on the original DNA extracts. In order to maximize the detection of true positives in our study, and thus ensure the detection of bone-fide recombinants, we focused our searches for inter-nuclear diversity on the 100 largest scaffolds of each isolate, and scored SNPs only within regions devoid of repeats, manually confirmed to be single copy using reciprocal BLAST procedures, and with frequencies between 0.26–0.74. This conservative approach has likely removed some true positives from our analysis - for example this filtering method resulted in an average 94.2% decrease in total SNP counts (Supplementary file 4) - but we believe that this caveat is largely offset by a much higher SNP validation rate, that is 84%, or 30 out of 36 SNPs tested, were validated using Sanger sequencing procedures.

We selected this stringent methodology to genotype all 86 sequenced nuclei and search for evidence of inter-nucleus recombination in all seven isolates. Using this approach, we find that the total number of SNPs detected is, on average, over 30 times larger in dikaryons compared to homokaryotic relatives, with homokaryons harboring only between 9 (R. diaphanus) to 131 SNPs (R. cerebriforme) along the 100 largest scaffolds (Supplementary file 4). Removing the filtering based on allele frequencies resulted in a slightly higher SNP count, but did not alter the overall results – that is dikaryons are still far more variable than homokaryons (Supplementary file 5). Consistent with previous findings based on much shorter nuclear regions (Ropars et al., 2016), our pairwise sequence comparisons show that nuclei isolated from AMF dikaryons always segregate into two main clades, depending on their genotype (Figure 2), and that each nucleus always harbors one of two co-existing MAT-loci.

Similarity matrix of nuclei isolated from AMF dikaryons.

The heat map reflects the level of similarity sequence between nuclei based on the SNPs detected along single copy regions. The number ID of the nuclei with a MAT-locus and genotype verified using PCR and Sanger sequencing are shown in coloured boxes. In A4 and A5, clustering of single nuclei correlates well with the MAT-locus identity of each nucleus (A, B). In R. irregularis SL1, this correlation is not evident, as reflected by the mosaic genetic pattern of single nuclei (C). R. irregularis SL1 also shows evidence of inter-nucleus recombination, as evidenced by the high sequence similarity between nuclei with opposite mating-type, for example nucleus 7 (C). Variable regions that could not be assigned to either MAT-locus are left with white background.

https://doi.org/10.7554/eLife.39813.005

We find that in A4 and A5, nuclei with the same MAT-locus (MAT-1 or MAT-2 for A4; MAT-3 or MAT-6 for A5) always harbor highly similar genotypes. This pattern results in a genetic similarity of single nuclei that is based exclusively on the MAT-locus identity (Figure 2A,B), and we find no evidence that nuclei with opposing MAT-loci cluster together in these isolates. In SL1, nuclei also segregate into two dominant clades, a finding consistent with its dikaryotic state. However, obvious links between the MAT-locus identity of its nuclei and their genotype are not always evident in this isolate (Figure 2C). For example, the nucleus 7 of SL1 harbors the MAT-5 locus, but also shows significant sequence similarity with nuclei harboring the MAT-1 locus (e.g. nucleus 8 and 12). This finding, which we confirmed using Sanger sequencing approaches (see coloured boxes in Figure 2), indicates the presence of inter-nuclear recombination in this isolate, potentially involving the MAT-locus. Other nuclei in SL1 show a similar pattern, as evidenced by the distinct mosaic genotypes of this isolate (Figure 2C). Overall, the genetic make-up of SL1 contrasts with the clear bi-allelic genetic structure of the isolates A4 and A5 (Figure 2A,B), and highlights the higher nuclear genotypic diversity harbored by this isolate.

A detailed look at individual nuclear genotypes highlights evidence of inter-nuclear recombination across the genome (Figure 3, Supplementary file 6). Examples of recombination tracts include those seen in the SL1 nuclei 14, and 15, which all share the MAT-1 locus but can sometimes carry partial genotypes found in co-existing MAT-5 nuclei, and vice versa (Figure 3, Supplementary file 6). This pattern of inter-nuclear genetic exchange occurs among many nuclei in this isolate, regardless of their mating-type, and across all of the largest scaffolds in SL1 (Figure 2, Figure 3, Supplementary file 6). In many cases, recombining genotypes encompass hundreds to thousands of base pairs, (Figure 3, Supplementary file 6). Inter-nuclear exchange between two co-existing genotypes are also seen, although more rarely, in A4 and to a much lesser extent in A5 – for example see scaffold 70 positions 100454 to 100557 (Figure 3, Supplementary file 6) for potential recombination in A4. In this example, a single recombination event between genotypes harbored by the nuclei 22 (MAT-1) and 19 (MAT-2) resulted in a genetic exchange involving at least one thousand base pairs, and similar events are found elsewhere in the genome of A4.

Selected examples of genotypes, recombination and inter-nuclear variability observed in the dikaryons SL1, A4 and A5.

In SL1, there is no obvious co-linearity between the MAT-locus of single nuclei and their genotype. White columns reflect the absence of Illumina sequencing along the nucleus at those homologous positions. The data shown are representative of the genotypes found along the first 100 scaffolds analysed in this study and are based on data available in Supplementary file 5. Variations along homologous nucleotide positions are highlighted in yellow or green, depending on the MAT-locus which the genotype is expected to associate with. The first and third columns represent, respectively the scaffold and position number where the SNP was scored using stringent filtering methodologies. The second column represents the genotype of the representative genome reference. M: Mating-type.

https://doi.org/10.7554/eLife.39813.006

We performed identical analyses using different, improved assemblies (e.g. ALLPATHS-LG; (Butler et al., 2008); Supplementary file 7) and more stringent variant callers (GATK-HaplotypeCaller (Van der Auwera, 2002), Mutect2 (Cibulskis, 2013); Supplementary file 8, Supplementary file 9), and found that these did not change the overall findings of our study. Specifically, regardless of the methods used to assemble the genome and call variants, evidence for inter-nuclear recombination is always present and more substantial in SL1 than in in A4 and A5. Accordingly, the genetic relationships among co-existing nuclei are also generally unaffected by changes in both assembly and SNP scoring methods (Supplementary files 1012).

Discussion

The Meselson effect predicts that the long-term absence of sex and recombination should lead to substantial accumulation of mutations within asexual lineages (Mark Welch and Meselson, 2000; Arkhipova and Meselson, 2005). Many asexual eukaryotes follow this prediction (Weir et al., 2016; Lovell et al., 2017; Neiman et al., 2010) but our data indicate that AMF may not. Specifically, we find that co-existing nuclei are overall very similar in sequence, showing between 0.001% to maximum of 0.38% average genome divergence, depending on the filters applied to score variation, with the vast majority of variation being in a bi-allelic state. The values obtained using basic filtering methods– that is 0.38% average nuclear genome divergence - are consistent with data obtained by others on single nuclei of the model AMF R. irregularis DAOM-197198 (Lin et al., 2014), but we also demonstrate that these are largely inflated by false positives. Future work should take this into consideration when assessing AMF genetic diversity using in-silico methodologies. Our study also shows that low nuclear polymorphism is a hallmark of many AMF species and is not only restricted to this R. irregularis DAOM-197198. It will be now interesting to see if natural AMF populations show a similarly low nuclear polymorphism.

Our analyses of single nuclei data also supports the hypothesis that essentially all fungal species have found means to recombine (Nieuwenhuis and James, 2016). In the case of AMF, our findings build on models previously proposed by Ropars et al. and Corradi and Brachmann (Ropars et al., 2016; Corradi and Brachmann, 2017), as we found evidence that nuclei with opposite MAT-loci can successfully undergo karyogamy in some dikaryotic isolates to generate new nuclear diversity through recombination (Figure 4). This discovery is unaffected by the quality of the assembly, and is supported by different methodologies used to score variation. Interestingly, in SL1, the observed inter-nuclear recombination does not appear to result from a single event – that is each nucleus shows differing recombination patterns - an indication that genetic exchanges among nuclei may have occurred independently at different locations within the same AMF multinucleated mycelium (Figure 4). In contrast, the few recombination events found in A4 involved exchange of genetic material between two co-existing and highly conserved genotypes.

Schematic representation of the three genome organizations found to date in the model AMF Rhizophagus irregularis.

Left: Most AMF analyzed using genome analysis and PCR targeted to the MAT-locus have been found to carry nuclei with the same MAT-locus. In these isolates, genetic variability is lower than dikaryotic relatives, and recombination is undetectable. Middle: The R. irregularis isolates A4 and A5 carry nuclei with two distinct MAT-loci. Evidence of recombination is very rare, and two divergent genotypes appear to co-exist in the cytoplasm. Right: In some cases, strains can harbour nuclei with two distinct MAT-loci that undergo frequent karyogamy. The frequency of karyogamy increases nuclear diversity within the mycelium.

https://doi.org/10.7554/eLife.39813.007

In fungi, diploid nuclei can undergo recombination either through meiosis, in a sexual cycle, or through aneuploidization, in a parasexual cycle involving nuclear fusion followed by random chromosomal loss. Parasexuality is rare in fungi (Paccola-Meirelles and Azevedo, 1991; Rosada et al., 2010; Seervai et al., 2013; Bennett, 2015), and this process drives recombination via mitotic events (as opposed to meiosis in a sexual cycle). It also requires the stable division of diploid nuclei for many generations to produce recombination through mitosis (Clutterbuck, 1996). To date, flow cytometry data has found that AMF nuclei are haploid (Ropars et al., 2016; Sedzielewska et al., 2011; Hosny et al., 1998) and evidence of widespread diploidy, as evidenced by the mapping of both alternate and reference reads along large portions of genome, could not be found in this study. In contrast, all AMF, including those investigated here, harbor a complete set of meiosis-related genes (Halary et al., 2011) and there is recurring evidence that conspecific isolates can exchange and recombine genetic material, presumably through sex (Riley et al., 2014; Croll and Sanders, 2009; den Bakker et al., 2010; Chen et al., 2018) . For these reasons, we argue that the inter-nuclear recombination we observed in this study is likely driven by meiotic events - that is conventional fungal sexual processes - although it is possible that parasexuality has also been at play in creating some of the observed nuclear diversity.

What causes recombination rates to vary substantially among the dikaryons SL1, A4 and A5 is presently unclear, as all cultures were established around the same time (Ropars et al., 2016; Jansa et al., 2002) and there is no available evidence that they were propagated under different in vitro conditions. Perhaps, in isolates such as A4 and A5 the co-existing genotypes are not permissive for generating recombinants (Idnurm et al., 2015) - for example the isolates have not yet met the right environmental conditions that trigger karyogamy or lack the mating compatibility necessary to undergo frequent meiosis (James et al., 2006; Heitman et al., 2013; Heitman, 2015). In this case, some of the somatic mutations and recombination found in A4 and A5 could result from exposure to mobile viral genotypes or transposable elements, as there is evidence that these are active in AMF genomes (Chen et al., 2018).

In conclusion, our findings demonstrate that arbuscular mycorrhizal fungi can generate genetic diversity via inter-nuclear recombination in the dikaryotic life-stage. This provides a novel understanding of the sexual and genetic potential of these plant symbionts and opens exciting possibilities to enhance their environmental application. In particular, we showed that recombined isolates such as SL1 contain an important source of nuclear diversity, which could lead to genetic enhancement of new AMF strains. Despite evidence of recombination, however, clonality still appears to be the prevalent mode of evolution in lab cultures of these symbionts, including for AMF dikaryons (Figure 4). From a conceptual perspective, future work should now aim to identify the mechanisms that trigger inter-nuclear recombination in AMF isolates, and determine how frequent this process is in natural populations of these widespread plant symbionts. It will also be important to establish whether recombination rates vary depending on soil conditions – for example disturbance - or specific plant hosts (Ritz et al., 2017), and how this process is reflected at the level of the transcriptome. Within this context, future investigations should also determine if specific recombined genomic regions are linked with an isolate’s phenotype (e.g. increased spore production, hyphal density, or mycorrhization rates) or with the fitness of economically important plants and crops.

Materials and methods

Single nucleus sorting, WGA, and Illumina sequencing

FACS-based sorting of single nuclei and WGA were performed according to Ropars et al., 2016 (Ropars et al., 2016). DNA concentration of the WGA samples was measured by fluorometric quantification using Qubit dsDNA HS Assay Kit (Thermo Fisher Scientific). Subsequently, sequencing libraries were constructed using Nextera XT DNA Library Preparation Kit (Illumina). Library quality and quantity were checked with a Bioanalyzer and the High Sensitivity DNA Analysis Kit (Agilent Technologies) and a Qubit with the DNA HS Assay Kit (Invitrogen). Samples were sequenced on a HiSeq1500 (Illumina) using 100 bp paired-end sequencing in rapid-run mode at LAFUGA (LMU Genecenter, Munich), resulting in 2,020,347,187 reads. Single nuclei sequencing and genome reads are available in NCBI under the following bioproject: PRJNA477348.

Genome assembly, read mapping and SNP prediction and filtering

SPAdes v. 3.10 was used to assemble the SL1, R. diaphanus (MUCL 43196), and R. cerebriforme (DAOM227022) genomes using default parameters (Bankevich et al., 2012), and the resulting contigs were then scaffolded using SSPACE (Boetzer et al., 2011). Reads from single nucleus sequencing were cleaned using trim_galore with default parameters and were then mapped to the respective genome assembly using BWA-Mem with –M parameter (bwa mem –M) (Li, 2013). SPAdes assemblies are publicly available on NCBI (R. diaphanus: QZLH00000000, R. cerebrifrome: QZLG00000000, SL1 SPAdes: QZCD00000000, and SL1 ALLPATH-LG: QZCC00000000) along with the core meiosis SL1 genes (RAD21/Rec8: MH974797, MND1: MH974798, DMC1: MH974799, Spo11: MH974800, HOP2: MH974801, MSH4: MH974802, MSH5: MH974803).

The whole genome reads are available on NCBI SRA: R. cerebriforme (SRR7418134 and SRR7418135), R. diaphanus (SRR7418169 and SRR7418170), and R. irregularis SL1 (SRR7418171 and SRR7418172). The single nucleus reads are also deposited (R. cerebriform: SRR7411799 to SRR7411813, R. diaphanus: SRR7410308 to SRR7410319, R. irregularis A1: SRR7416439 to SRR7416450, R. irregularis A4: SRR7416451 to SRR7416464, R. irregularis A5: SRR7416431 to SRR7416438, R. irregularis C2: SRR7416465 to SRR7416473, and R. irregularis SL1: SRR7411814 to SRR7411829).

Two filtering methods were used to detect variants, which are referred to here as ‘basic’ or ‘stringent’. In the basic filtering procedure, initial variant calling was done using freebayes with the following parameters: -p 1 m 30 K -q 20 C 2; namely a ploidy of one, a minimum quality of mapped reads of 30, a minimum base quality of 20, and a minimum set of reads supporting alternative allele of two (Garrison and Marth, 2012). Predicted variant positions were then filtered by keeping only those where reads supporting alternate alleles outnumbered reads supporting the reference allele by a ratio of 10:1.

The strict filtering method was built on top of the basic filtering procedures and includes three additional criteria. The first criterion was the coverage of the reference assembly from the original reads. Specifically, reads were mapped to the reference assembly and only candidate SNP positions with coverage that ranged between 69% and 131% of average coverage along the genome reference (close to 50/50 ratio) were kept. The second criterion was that the remaining candidate SNP positions also needed to keep a proportion of reference allele to alternate allele between 26% to 74% in the SNP calling via original assembly reads. Finally, the third criterion was confirming the single copy nature of the SNP using BLAST procedures. In this case, the 100 bp upstream and downstream regions overlapping the scored SNPs were BLAST against the reference genome. If BLAST results returned more than two good hits (e-value of better than 0.001) this region was considered to be a multi-copy region and thus discarded from downstream analyses.

Distance matrix comparisons

To detect inter-nuclear sequence divergence, pairwise comparison of SNP data was performed for co-existing nuclei of all isolates. In this analysis, two datasets were created. The first dataset includes only all variable regions, as defined by the total number of filtered SNP positions shared between the nuclei (Supplementary file 6). The other dataset includes the number of covered positions shared between the nuclei (Supplementary file 2).

SNP dispersion analysis

Raw SNP predictions from single nucleus reads were cleaned by removing SNP that were longer than length one. Each SNP position was converted into percent of total length of scaffold one and then plotted with an alpha value of 0.2.

Validation of SNPs and MAT-loci via PCR and Sanger sequencing using DNA from single nuclei

PCR-reactions were run with 0.13 ng/µL to 0.33 ng/µL of single nucleus DNA as a template, 0.5 µM forward and reverse primers each (Supplementary file 13 and 0.2 mM dNTPs. For reactions performed with Phusion HF Polymerase, 1x Phusion HF Buffer and 0.02 U/µL Phusion HF Polymerase and for reactions with GoTaq Hot-Start Polymerase, 1x Green Buffer, 2.5 mM MgCl2 and 0.025 U/µL GoTaq Hot-Start Polymerase were used. The reactions were performed in a total volume of 15 µL or 30 µL. The amplification was run with a touchdown PCR program: 95°C/2 min - [95°C/30 s; T1/30 s (T1 declining by 0.5°C every cycle) -; 72°C/30 s]x10 - [95°C/30 s; T2/30 s; 72°C/30 s]x35–72°C/5 min. T1 and T2 are specific annealing temperatures for different primer combinations (see Supplementary file 9). PCR cleanup was performed with 2.9 U/µL Exonuclease I (New England Biolabs) and 0.14 U/µL Shrimp Alkaline Phosphatase (rSAP, New England Biolabs) at 37°C for 5 min and heat inactivated at 85°C for 10 min. Sanger sequencing reactions were performed at the Genomics Service Unit (LMU).

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
    Mycorrhizal Symbiosis
    1. JD James
    2. SE Smith
    3. DJ Read
    (1998)
    71–72, Plant Growth Regulation, Mycorrhizal Symbiosis, Springer.
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
    The frequency of sex in fungi
    1. BPS Nieuwenhuis
    2. TY James
    (2016)
    Philosophical Transactions of the Royal Society B: Biological Sciences 371:20150540.
    https://doi.org/10.1098/rstb.2015.0540
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
    GATK best practices
    1. GA Van der Auwera
    (2002)
    Current Protocols in Bioinformatics 11:11.10.1–11.1011.
  55. 55
  56. 56
  57. 57

Decision letter

  1. Patricia J Wittkopp
    Senior Editor; University of Michigan, United States
  2. Francis Martin
    Reviewing Editor; Université de Lorraine, France
  3. Rene Geurts
    Reviewer; Wageningen University, Netherlands
  4. Peter Young
    Reviewer; University of York, United Kingdom
  5. Michael Seidl
    Reviewer

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Single nucleus sequencing reveals evidence of inter-nucleus recombination in arbuscular mycorrhizal fungi" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen a Reviewing Editor, and Patricia Wittkopp as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Rene Geurts (Reviewer #1); Peter Young (Reviewer #2); Michael Seidl (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

Arbuscular mycorrhizal fungi (AMF) in the Glomeromycotina are ecologically important symbionts which promote plant growth and diversity. The AMF mycelium is unique among eukaryotes as it harbors hundreds to thousands of nuclei within one continuous cytoplasm and mechanisms that could trigger sexual processes in these organisms have long been elusive. This manuscript describes an elegant study to provide an answer to a long-standing question in AMF biology. Can AM fungi experience genetic recombination, despite the lack of obvious phases of sexual reproduction? Chen et al. tackled this question by assessing the hypothesis that karyogamy and genetic recombination are taking place among nuclei co-existing in dikaryotic isolates by comparing sequence data obtained from single AMF nuclei. They quantified SNP segregation in dikaryotic and homokaryotic Rhizophagus strains. Whole genome sequences, as well as tens of single nuclei sequence data, were generated of a total of seven strains: five from Rhizophagus irregularis (3 dikaryotic and 2 homokaryotic strains), and two from the related species R. diaphanus and R. cerebriforme (both homokaryotic). SNP segregation analysis provides solid evidence that genetic recombination between nuclei occurred, although at a very low rate. These findings support the contention that AMF can generate genetic diversity via inter-nuclear recombination in the dikaryotic life-stage.

The manuscript is generally well written, the illustrations and tables are clear and informative, and the applied methodology is appropriate to support the results and the conclusion, i.e. inter-nuclear recombination in AMF.

However, we have significant concerns to which extent this data and the derived conclusions present a significant step forward. Our main concerns are listed below:

Essential revisions:

1) First, evidence for recombination in different AMF have been reported previously (e.g. Croll et al., 2009, and references therein). Second, this manuscript builds on a paper from your group that was published two years ago (Ropars et al., 2016). The main new claim is evidence for recombination between nuclei in a dikaryon. In fact, Ropars et al. already claimed to have "Sporadic evidence for recombination" in their Figure 3B, though we don't see any evidence for recombination in the dikaryotic strain A5. The only putative 'recombination event' was in the homokaryotic A1, which is stated to have 'no recombination' in the current manuscript. Data as presented in Figure 1B for isolates A1, A4 and A5 have been reported previously (see Figure 1 in Ropars et al., 2016). It appears that the present paper is an extension of your seminal observations reported in Ropars et al., 2016, and provides new datasets: (i) by sequencing isolate DAOM-240409, which is considerably more divergent than A4 and A5, and (ii) by analysing a 'genome-wide' set of SNPs rather than focus on few SNP positions.

Recommendations:

– Please, carefully check for inconsistencies between the present manuscript and Ropars et al., 2016.

– Please emphasize more strongly in your Introduction/Discussion the novelty of the present findings. What's new?

2) The present manuscript presents sequence data on three strains not previously studied: Rhizophagus irregularis DAOM-240409, R. diaphanus MUCL-43196 and R. cerebriforme DAOM-227022. There is also analysis of sequence data for strains A1, C2, A4 and A5, but it is not clear whether this includes new data or is based on the sequencing already carried out by Ropars et al., 2016.

Recommendations: Please clarify.

3) The focus is on the dikaryotic strain DAOM-240409, and you claimed that there has been extensive recombination between co-existing nuclei in this strain, in contrast to A4 and A5, which are also dikaryotic but show only limited signs of recombination. This is certainly backed up by the similarity matrices in Figure 2. The sequencing of PCR products (Supplementary Figure 1) is less convincing, since template switching during PCR can often generate artefactual 'recombinant' sequences. The single-nucleus sequences provide much better evidence (Figure 3).

Recommendations: Please clarify and comment.

4) One important caveat is the quality of the assembly of DAOM-240409. Despite higher sequencing coverage, this strain has twice as many scaffolds as the other dikaryons, and a much larger assembly size. Although you mention this (Introduction), you offer no explanation. There may be a connection between the assembly and the high frequency of apparent recombination, but it is not clear whether the high scaffold number is somehow a consequence of recombination, or the apparent recombination is an artefact of an assembly that is poor for technical reasons. We need a more careful consideration of this question, based on some analysis of the reasons for the high scaffold number and assembly size. This could reveal new information of real biological interest. For example, is this related to any difference in transposable element density/activity/categories? There is no mention of the gene annotations in the present manuscript. Can you assess the quality of the DAOM-240409 assembly by looking at this gene annotation?

5) Reviewer #3 emphasized that the manuscript would significantly benefit from analyses aimed to provide a better understanding into the processes underlying and modulating the observed recombination (as suggested in the Discussion). For instance, while you suggest the operation of a 'meiotic-like' process, mainly due to the presence of a complete set of meiosis-related genes (Halary et al., 2011) and the absence of aneuploid nuclei, no further experimental evidence is provided. They suggested that alternative genome sequencing and assemblies (e.g. based on Oxford Nanopore or PacBio data) approaches should yield contiguous genome assemblies which should enable the analyses of recombination hotspots (Supplementary file 3) in high resolution. I assume that these additional genome sequencing cannot be carried out in time for the resubmission owing to the difficulty in obtaining high quality genome DNA from AMF for PacBio or Nanopore sequencing. Please comment.

https://doi.org/10.7554/eLife.39813.057

Author response

Essential revisions:

1) First, evidence for recombination in different AMF have been reported previously (e.g. Croll et al., 2009, and references therein).

It is essential to distinguish between two distinct patterns of recombination:

A) Past studies = PCR/Cloning based evidence for recombination among conspecific AMF isolates – e.g. Croll and Sanders, 2009, Den Bakker et al., 2010, Riley et al., 2014 and others.

B) Our study = Direct, single nucleus based evidence recombination within the AMF mycelium – i.e.how inter-nucleus recombination generates new genetic diversity in the absence of observable mating.

The origin/nature of the recombination reported by references mentioned in point (A) is very different from the one we identified in this paper. In particular, as opposed to the studies cited above, our study provide direct, single nucleus based quantifiable evidence of inter-nuclear recombination, while other studies citing evidence of recombination had to infer its existence based on patterns found in distant organisms.

More importantly, the above mentioned studies are all based on data obtained through PCR and bacterial cloningof single gene sequences, assuming a priori that the genes analyzed and compared among isolates were all single-copy and orthologues. Unfortunately, AMF genome needed data to back such claims was unavailable when these studies were first published.

Consequently, such claims of recombination can all be explained by a myriad of mechanisms including for example, gene conversion through paralogy (which is extensive in AMF), balancing selection, or accelerated mutation rate at the loci in some of the loci/strains investigated and, perhaps most importantly, PCR/Cloning artefacts. Indeed, to our knowledge, the only other studies suggesting that recombination may derive from inter-nuclear recombination within the AMF mycelium are those of Kuhn et al., 2001, and Galdolfi et al., 2003. Both studies are now known to have likely analyzed artificial recombinants resulting from PCR cloning procedures. Specifically, both studies identified recombination in a locus named the “Bip gene” (an Hsp90 protein), which was later found to be multi-copy and monomorphic using AMF genome data (Tisserant et al., 2013; Chen et al, 2018) as well as single nuclear genome sequences (Lin et al., 2014).

Some issues with previous claims of recombination and high heterokaryosis in AMF have been recently discussed in Corradi and Brachmann, 2017. We believe it would be confusing to resurrect these settled debates, which is the reason why we avoided discussing this particular topic in the original version of our paper.

However, we understand the editor’s and reviewer’s concerns, and we thus now briefly discuss some of the past work on AMF inter-isolate recombination in our revised paper, stressing how our study finally provides an answer to a long-lasting debate regarding the presence of inter-nuclear recombination within the AMF mycelium, particularly in the dikaryotic life-stage.

Second, this manuscript builds on a paper from your group that was published two years ago (Ropars et al., 2016). The main new claim is evidence for recombination between nuclei in a dikaryon. In fact, Ropars et al. already claimed to have "Sporadic evidence for recombination" in their Figure 3B, though we don't see any evidence for recombination in the dikaryotic strain A5. The only putative 'recombination event' was in the homokaryotic A1, which is stated to have 'no recombination' in the current manuscript.

We are sorry for the confusion the term has created. In our view, the term “sporadic” in Figure 3B legend of Ropars et al. 2016 was synonym of “isolated”. It specifically referred to the single mutation found in the homokaryotic isolate A1, which you mentioned.

In this case, the mutation could either represent a nucleus-specific somatic mutation, or a past inter-isolate recombination with the isolate C2. This situation is reminiscent of the studies mentioned earlier, where one observation (i.e. the A1 singleton) can be explained in very different ways.

In stark contrast to this, our analyses of 80+ single nuclei from seven strains provide for the first time direct, genome-wide evidence that AMF can generate genetic diversity within their mycelium by recombining nuclei with opposing MAT-loci. This is a completely new discovery, with outstanding implication for understanding AMF biology and genetics, and their application as biofertilizers.

Data as presented in Figure 1B for isolates A1, A4 and A5 have been reported previously (see Figure 1 in Ropars et al., 2016). It appears that the present paper is an extension of your seminal observations reported in Ropars et al., 2016, and provides new datasets: (i) by sequencing isolate DAOM-240409, which is considerably more divergent than A4 and A5, and (ii) by analysing a 'genome-wide' set of SNPs rather than focus on few SNP positions.

In our original version of the paper, we specifically state that our findings build on models proposed by Ropars et al., 2016 and Corradi and Brachmann, 2017. Specifically, we finally provide direct evidence that nuclei recombine in dikaryotic AMF isolates – i.e. AMF have found ways to generate genetic diversity in the absence of sexual reproduction. Previously, this was only a testable hypothesis.

Again, this is something that was previously unknown to exist in AMF and represents a landmark finding in the field.

More generally, the new data we provide is as follows:

– 3 new genomes assemblies: Rhizophagus irregularis DAOM-240409, R. diaphanus and R. cerebriforme. We would like to point out that DAOM-240409 is not phylogenetically “divergent”, but simply shows evidence of higher recombination and heterokaryosis compared to relatives. All assemblies used in this study have now been deposited in Genbank. Accession numbers are available in the revised version of the manuscript.

– Genome data from 80+ single nuclei isolated from seven Rhizophagus isolates, including nuclei from the R. irregularis isolates A1, A4, A5 and C2. A prior study provided data and analyses from 7 single nuclei (Lin et al., 2014) isolated from the model AMF R. irregularis DAOM-297198. Most of these analyses centered on genome content and on challenging the presence of heterokaryosis in AMF, but found no evidence of recombination in this homokaryotic isolate.

– In contrast, our analyses of single nuclei from many homo/dikryotic strains provide the first direct evidence that co-existing nuclei with opposing MAT-loci recombine within the AMF mycelium – i.e. we demonstrate how these organisms can create genetic diversity in the absence of observable sex.

– Demonstrate that false SNP positives are very abundant in analyses of genomes and single nuclei, and identify ways to drastically reduce their identification. This is particularly important for our area of research, which is interested in detecting the degree of genetic variability among/within isolates.

– We provide first direct (single nucleus sequencing) evidence that nuclear polymorphisms is always low in AMF, regardless of the species investigated. Previous knowledge in this area has always been inferred from single-gene or whole-genome datasets, which is always error-prone and gives a very approximate measure of genetic diversity – e.g. SNP/Kb counts.

Overall, the only publicly available data we used is represented by the genome assemblies of A4, A5, A1 and C2, which were used to map their respective single nuclei reads. All single nuclei from these isolates were newly obtained in our present study.

In the revised version of the paper, the Table 1 has been modified to highlight that the assemblies from A4, A5, A1 and C2 were previously obtained by Ropars et al., 2016.

[…]

2) The present manuscript presents sequence data on three strains not previously studied: Rhizophagus irregularis DAOM-240409, R. diaphanus MUCL-43196 and R. cerebriforme DAOM-227022. There is also analysis of sequence data for strains A1, C2, A4 and A5, but it is not clear whether this includes new data or is based on the sequencing already carried out by Ropars et al., 2016.

Recommendations: Please clarify.

The genome analyses presented in Figure 1 – i.e. allele plots – are based on new analyses of publicly available genome sequence data from Ropars et al., 2016 (assembly and genome reads), and new references we provide for the isolates DAOM-240409, R. diaphanus MUCL-43196 and R. cerebriforme DAOM-227022.

For clarity, this information has now been included in the revised figure legend.

3) The focus is on the dikaryotic strain DAOM-240409, and you claimed that there has been extensive recombination between co-existing nuclei in this strain, in contrast to A4 and A5, which are also dikaryotic but show only limited signs of recombination. This is certainly backed up by the similarity matrices in Figure 2. The sequencing of PCR products (Supplementary Figure 1) is less convincing, since template switching during PCR can often generate artefactual 'recombinant' sequences. The single-nucleus sequences provide much better evidence (Figure 3).

Recommendations: Please clarify and comment.

We were originally surprised to see how well the recombination patterns correlated with our findings – i.e. DAOM-240409 and A4 showed evidence of recombination and found none in A5 – and thus decided to keep those results in the paper.

However, we agree that PCR//Cloning can artifactually produce these patterns (as we noted in the original manuscript) and that these could detract the reader from the solid evidence of inter-nuclear recombination we identified using single nuclei sequencing.

Thus, we decided to remove this figure from the revised version of the manuscript.

We have now modified the text accordingly. The modified section now reads as follows: “Here, we tested the hypothesis that karyogamy and inter-nuclear recombination occur in the AMF mycelium using a single-nucleus sequencing approach. […] Genome data from R. irregularis SL1 (also known as DAOM-240409), R. diaphanus and R. cerebriforme represent new additions to public databases from representatives of the Glomeromycotina, and all sequencing data were assembled using SPAdes (Bankevich et al., 2012) to facilitate comparisons with published genome data.”

4) One important caveat is the quality of the assembly of DAOM-240409. Despite higher sequencing coverage, this strain has twice as many scaffolds as the other dikaryons, and a much larger assembly size. Although you mention this (Introduction), you offer no explanation. There may be a connection between the assembly and the high frequency of apparent recombination, but it is not clear whether the high scaffold number is somehow a consequence of recombination, or the apparent recombination is an artefact of an assembly that is poor for technical reasons. We need a more careful consideration of this question, based on some analysis of the reasons for the high scaffold number and assembly size. This could reveal new information of real biological interest. For example, is this related to any difference in transposable element density/activity/categories? There is no mention of the gene annotations in the present manuscript. Can you assess the quality of the DAOM-240409 assembly by looking at this gene annotation?

We agree that assessing the assembly quality of DAOM-240409 is essential for this study, especially given the higher recombination rates we found in this study.

We aimed to address these concerns by performing additional analyses that follow-up on suggestions from reviewer #3. These analyses are aimed to test the completeness of the assembly and if technical/analyses errors could have generated some of the higher genetic diversity identified in the DAOM-240409 assembly.

– A BUSCO analysis was carried out on all assemblies to obtain a general overview of their completeness/coding capacity. We find that all R. irregularis dikaryotic assemblies almost identical in term of completeness – i.e. 90% for A4, 91% for DAOM-240409 and 92% for A5. Please note that those values are slightly lower than those reported in Chen et al., 2018, because they were obtained on genome assemblies, as opposed to annotated protein sets.

– As proposed by the reviewer, we compared TE abundance in DAOM-240409 with other isolates, and found that all show similar TE content. Thus, the higher fragmentation of the DAOM-240409 does not seem to result from higher TE counts.

– We tested whether the higher assembly fragmentation in DAOM-240409 was due to much higher genome size using K-mer counts, an approach that gives an approximate estimation of genome size based on sequencing reads. We find that all Rhizophagus isolates similar genome sizes – ranging between 130-150Mb.

– Higher assembly sizes and fragmentation are also found in the other dikaryons, particularly A4 but also A5, so fragmentation is clearly correlated with nuclear complexity.

For all these reasons, we believe that the larger size and genome fragmentation of the DAOM-240409 assembly is better explained by higher recombination/heterokaryosis rates, although we cannot conclusively rule out other scenarios.

These additional analyses have now been included in the revised version of the paper to better highlight our hypothesis that the higher DAOM-240409 fragmentation is the result of higher nuclear genetic diversity – i.e. recombination/heterokaryosis.

The text now highlights this and reads as follows: “Higher fragmentation is also seen in other dikaryons, particularly A4, and thus appears to result from the presence of higher intra-mycelial genetic diversity in this part of the AMF cycle – i.e. presence of two distinct nucleotypes challenges the assembly procedures. Interestingly, BUSCO, K-mer graphs and transposable elements (TE) analyses all show that SL1 harbors a gene repertoire, a genome size and TE counts that are very similar to those of other isolates analysed in this study (Table 2, Supplementary file 1)”.

Regardless of what has caused of the assembly fragmentation, it is noteworthy that we took drastic steps to reduce the detection of false positives by applying very stringent methodologies on the most reliable scaffolds – i.e. top 100 only. In summary, we have been hyper-conservative in our analyses.

5) Reviewer #3 emphasized that the manuscript would significantly benefit from analyses aimed to provide a better understanding into the processes underlying and modulating the observed recombination (as suggested in the Discussion). For instance, while you suggest the operation of a 'meiotic-like' process, mainly due to the presence of a complete set of meiosis-related genes (Halary et al., 2011) and the absence of aneuploid nuclei, no further experimental evidence is provided. He/she suggested that alternative genome sequencing and assemblies (e.g. based on Oxford Nanopore or PacBio data) approaches should yield contiguous genome assemblies which should enable the analyses of recombination hotspots (Supplementary file 3) in high resolution. I assume that these additional genome sequencing cannot be carried out in time for the resubmission owing to the difficulty in obtaining high quality genome DNA from AMF for PacBio or Nanopore sequencing. Please comment.

Our response to these comments are separated into two categories.

A) Regarding the absence of further experimental evidence for meiosis vs. parasexuality:

As discussed in the original manuscript, the observed inter-nuclear recombination can result from two processes – i.e. meiosis or parasexuality. Our goal is to always discuss all plausible scenarios for the patterns we observed.

In the original paper, we argue that meiosis better explained our findings for three main reasons. We still believe that these arguments are sound.

In particular:

– We find that a complete set of meiosis genes is found in all isolates investigated.

– Past evidence of inter-isolates recombination in AMF support the existence of conventional mating as seen in sexual fungi. Specifically, heterokaryon incompatibility prevents nuclear mixing in natural fungal populations, making it unlikely that inter-isolate recombination in AMF is not driven by mating processes and subsequent meiotic-like events.

– To date, only cytometry evidence for haploid nuclei was found.

– Parasexual cycles can drive recombination via mitotic events. However, these require the stable division of diploid cells for a number of generation to trigger mitotic recombination, but we found no sequence-based evidence for diploidy in our single nucleus analyses.

– The high frequency of recombination found in DAOM-240409 is not compatible with a parasexual cycle, as recombination generally happens at low frequency during this process.

The paper was revised to discuss this additional information. The modified text reads are follows: "In fungi, diploid nuclei can undergo recombination either through meiosis, in a sexual cycle, or through aneuploidization, in parasexual cycle involving nuclear fusion followed by random chromosomal loss. […] For these reasons, we argue that the inter-nuclear recombination we observed in this study is likely driven by meiotic events – i.e. conventional fungal sexual cryptic process – although it is possible that parasexuality has also been at play in creating some of the observed nuclear diversity…”

As original stated, we still fully acknowledge the potential contribution of parasexuality in creating some of the observed recombination.

B) Regarding long-read sequencing:

We agree with the referee that having access to “long-reads” sequencing data would be very helpful in this study. Our long-term goal is to acquire optimal, ideally chromosome-level, assemblies using such technologies. To this end, we are presently trying to acquire large quantities of High Molecular Weight DNA from all AMF dikaryotic isolates, but this is has been more challenging than we had anticipated.

Consequently, as recognized by the editor, it will take us well over two months to implement optimal conditions for AMF DNA extractions, obtain top assemblies, and analyze our single nuclear data using these technologies. Nevertheless, we understand the concern of the reviewer, and thus we aimed to address it using two independent approaches. Both of these methods confirm our previous findings:

– We first tested if a better assembly for DAOM-240409 would remove evidence of recombination in this isolate. To this end, we used ALLPATHS-LG and obtained a significantly better assembly for SL1 – i.e. 6642 scaffolds, as opposed to 29249 for the SPAdes assembly. We find that using identical methods on this new assembly still shows substantial evidence of recombination in this isolate.

Furthermore, single nuclei pairwise genotype similarities based on the ALLPATH-LG assembly are very similar to those found based on the SPAdes assembly.

– Secondly, we tested if recombination could also be found using known stringent “variant callers” such as GATK HaplotypeCaller and Mutect2. Both callers again supported the existence of significant inter-nucleus recombination in DAOM-240409.

These additional analyses have now been included in the revised version of the manuscript and are discussed, accordingly in the paper.

The revised text reads as follows: “Performing identical analyses on improved assemblies (ALLPATHS-LG; Butler et al., 2008; Supplementary file 7) and with more stringent variant callers (GATK-HaplotypeCaller (Van der Auwera et al., 2002), Mutect2 (Cibulskis et al., 2013); Supplementary files 8 and 9) did not change the overall findings. […] Accordingly, the genetic relationships among co-existing nuclei are also generally unaffected by changes in both assembly and SNP scoring methods (Supplementary files 10-12)”.

Accordingly, several new supplementary files were added to the revised manuscript.

https://doi.org/10.7554/eLife.39813.058

Article and author information

Author details

  1. Eric CH Chen

    Department of Biology, University of Ottawa, Ottawa, Canada
    Contribution
    Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  2. Stephanie Mathieu

    Department of Biology, University of Ottawa, Ottawa, Canada
    Contribution
    Formal analysis, Investigation, Visualization, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  3. Anne Hoffrichter

    Institute of Genetics, Faculty of Biology, Ludwig Maximilian University of Munich, Munich, Germany
    Contribution
    Resources, Validation, Investigation, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  4. Kinga Sedzielewska-Toro

    Institute of Genetics, Faculty of Biology, Ludwig Maximilian University of Munich, Munich, Germany
    Contribution
    Resources, Methodology
    Competing interests
    No competing interests declared
  5. Max Peart

    Department of Biology, University of Ottawa, Ottawa, Canada
    Contribution
    Resources, Formal analysis, Validation, Methodology
    Competing interests
    No competing interests declared
  6. Adrian Pelin

    Department of Biology, University of Ottawa, Ottawa, Canada
    Contribution
    Data curation, Formal analysis, Validation
    Competing interests
    No competing interests declared
  7. Steve Ndikumana

    Department of Biology, University of Ottawa, Ottawa, Canada
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared
  8. Jeanne Ropars

    Department of Biology, University of Ottawa, Ottawa, Canada
    Present address
    Ecologie Systematique et Evolution, CNRS, Université Paris Sud, AgroParisTech, Université Paris Saclay, Orsay, France
    Contribution
    Conceptualization, Writing—review and editing
    Competing interests
    No competing interests declared
  9. Steven Dreissig

    Leibniz Institute of Plant Genetics and Crop Plant Research, Gatersleben, Germany
    Contribution
    Resources, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-4766-9698
  10. Jorg Fuchs

    Leibniz Institute of Plant Genetics and Crop Plant Research, Gatersleben, Germany
    Contribution
    Resources, Methodology
    Competing interests
    No competing interests declared
  11. Andreas Brachmann

    Institute of Genetics, Faculty of Biology, Ludwig Maximilian University of Munich, Munich, Germany
    Contribution
    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Investigation, Methodology, Writing—review and editing
    Competing interests
    No competing interests declared
  12. Nicolas Corradi

    Department of Biology, University of Ottawa, Ottawa, Canada
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing
    For correspondence
    ncorradi@uottawa.ca
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7932-7932

Funding

Natural Sciences and Engineering Research Council of Canada

  • Nicolas Corradi

Ontario Ministry of Research, Innovation and Science (ER13-09-190)

  • Nicolas Corradi

Deutsche Forschungsgemeinschaft (BR 3527/1-1)

  • Andreas Brachmann

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Timothy Y James and Linda Bonen for their comments on an earlier version of the manuscript. NC is supported by the Discovery program from the Natural Sciences and Engineering Research Council of Canada (NSERC-Discovery), an Early Researcher Award from the Ontario Ministry of Research and Innovation (ER13-09-190), and the ZygoLife project funded by the National Science Foundation (DEB 1441677). AH, KS-T, and AB are supported by the German Research Foundation (DFG grants BR 3527/1–1 and PA 493/11–1).

Senior Editor

  1. Patricia J Wittkopp, University of Michigan, United States

Reviewing Editor

  1. Francis Martin, Université de Lorraine, France

Reviewers

  1. Rene Geurts, Wageningen University, Netherlands
  2. Peter Young, University of York, United Kingdom
  3. Michael Seidl

Publication history

  1. Received: July 3, 2018
  2. Accepted: November 16, 2018
  3. Version of Record published: December 5, 2018 (version 1)

Copyright

© 2018, Chen et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 992
    Page views
  • 124
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Evolutionary Biology
    2. Plant Biology
    David Winship Taylor, Hongqi Li
    Insight
    1. Evolutionary Biology
    2. Plant Biology
    Qiang Fu et al.
    Research Article