Ethics statement

Live plants were collected from Cwm Idwal, UK, with permission from Natural Resources Wales, the Snowdonia National Park Authority, and the landowners (National Trust). No permit was required for this collection and no protected species were sampled. DNA was acquired from preserved modern specimens permanently deposited at Kew Gardens, UK–specimen numbers are available in S1 Table.

General approach

In this study, we selected six Isoetes for generating genome-wide DNA datasets–I. coromandelina (this specimen is referred to as I. coromandelina sensu lato to reflect the taxonomic complexity of this species–see Pantil and Rajput [60]), I. humilor, I. elatior, I. nuttallii, I. lacustris and I. andicola. These were selected to capture the deeper divergence events within this group based on previous molecular studies [18]. Analyses of nuclear ribosomal DNA available for a large number of Isoetes confirmed that the last common ancestor of the selected species likely corresponds to the last common ancestor of extant Isoetes, and the low branch length variability throughout the genus suggests the sequenced species represent a good sample of evolutionary rates within the genus (Fig 1). Sparse taxon sampling has been shown to significantly affect estimated dates in some molecular dating studies [61, 62], although not in every case [63, 64]. Rate heterogeneity likely plays an important role in the effect of sparse taxon sampling on the accuracy of molecular dating, with high levels of rate heterogeneity demanding more sampling [62]. The relatively low levels of rate heterogeneity within Isoetes (Fig 1) suggest it is a suitable group to perform molecular dating with a relatively small number of taxa. To further investigate the impact of our sampling scheme, we reanalysed the dataset of Larsén and Rydin [18], which contains 45 Isoetes species including all the major clades identified by previous studies of Isoetes [56, 65]. This dataset was reanalysed using the same constraints and BEAST settings as Larsén and Rydin [18], but with the 45 Isoetes species used reduced to the 6 closest relatives of our chosen species (I. asiatica, I. coromandelina sensu lato, I. drummondii, I. echinospora, I. kirkii and I. storkii). This resulted in an estimated crown date of Isoetes of 153.4 Ma (53.9–277.3 95% CI), only a 7% increase compared to the full species sampling. This indicates limited taxon sampling should not substantially alter the estimation of the Isoetes crown node date.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Maximum likelihood phylogram of Isoetes nuclear ribosomal internal transcribed spacer.

Branch lengths are proportional to the expected number of substitutions per site, with scale bar representing 0.05 substitutions per site. Branches in bold have bootstrap support values greater than 90. Species in bold represent data generated in this study; nuclear ribosomal internal transcribed spacers for other Isoetes are from those used in Larsén and Rydin [18].

https://doi.org/10.1371/journal.pone.0227525.g001

DNA from these species was then used to generate genome-wide datasets, and different genome partitions were analysed in isolation to get accurate estimates of divergence times. Herbarium specimens represent a useful source of DNA, particularly for globally distributed, hard-to-access groups such as Isoetes [66, 67]. Low-coverage whole-genome scans can be applied to these samples, and will yield high coverage for genomic fractions present as multiple copies, such as the organellar genomes [68]. However, highly variable evolutionary rates in chloroplast markers have been reported from seed plants [69, 70], which potentially affect the results of dating methods that differ in their assumptions of rate heterogeneity [24].

Previous studies of the chloroplast marker rbcL indicate much higher rates of sequence evolution in Selaginella than in Isoetes [18, 70, 71]. Nuclear markers can be more useful for molecular dating if they show less variation in rates among branches [24] as suggested from large scale embryophyte phylotranscriptomics [72, 73]. Genome skimming can provide nuclear sequences, but low coverage makes de novo assembly difficult. However, the sequencing reads can be mapped to a reference dataset, providing phylogenetically informative characters [74, 75]. A reference genome is available for Selaginella, but it is too distant to allow accurate mapping of reads from Isoetes. Transcriptomes provide high coverage of expressed protein-encoding genes, which represent regions of the genome allowing read mapping across distinct species [74, 75]. We consequently decided to generate and assemble a transcriptome for a single Isoetes species, which was used as a reference to map reads from low-coverage whole-genome sequencing datasets obtained from the other Isoetes species sampled from herbarium collections. The sequencing data were used to obtain chloroplast and nuclear alignments for five Isoetes species as well as a number of other land plants (mosses, ferns, lycopods, gymnosperms and angiosperms) sequenced in other studies. The phylogenetic breadth of the datasets allowed the incorporation of fossil evidence providing calibration points spread across the tree.

Sequence acquisition

Live Isoetes lacustris were sampled from Cwm Idwal, Wales and maintained at the University of Sheffield in 40 x 30 x 25 cm transparent plastic containers, with a substrate of sand to a depth of 5 cm, and the containers filled to the top with deionised water. These were placed in a Conviron growth chamber with a 12-h day/night cycle, 495 μmol m²s^-1 light, temperature at 20°C during the day and 18°C at night, and CO₂ at 400 ppm for six days. To maximise the number of transcripts retrieved, leaves from three individuals were sampled 3 hours after dark and 3 hours after light and stored immediately in liquid nitrogen. We also generated a transcriptome for Littorella uniflora (Plantaginaceae), another species of aquatic plant that shares aquatic CAM photosynthesis [76]. Individuals from this species were also sampled from Cwm Idwal and were grown under a variety of conditions before sampling their leaves as described above. Dried specimens were deposited in the Sheffield University Herbarium (I. lacustris–DW1, L. uniflora–DW2).

RNA was extracted from the sampled leaves using the RNeasy® Plant Mini Kit (Qiagen), following the manufacturer protocol, with the addition of on-column DNase I digestion (Qiagen RNase-Free DNase Set). We then added 2.5 μl SUPERase-In^TM RNase inhibitor (Invitrogen) to 50 μl of extracted RNA to stabilise it. RNA was quantified using a gel electrophoresis, RNA 6000 Nano chips (Aligent) in an Aligent 2100 Bioanalyser, and a Nanodrop 8000. Samples were then prepared for Illumina sequencing using the TruSeq® RNA Sample Prep Kit v2 (Illumina). Paired-end sequencing was performed on an Illumina HiSeq 2500 platform available at Sheffield Diagnostic Genetics Service in rapid mode for 100 cycles, with 24 libraries pooled per lane of flow cell (other samples were from the same or different projects).

DNA from herbarium specimens of five Isoetes species were acquired from the DNA Bank from the Royal Botanical Gardens, Kew (S1 Table). This was supplemented with one silica gel dried leaf each of I. lacustris and L. uniflora collected from the field as described previously. Whole genome sequencing of these seven samples was performed at the Genotoul from the University of Toulouse, using previously described protocols [74, 77]. Each sample was sequenced on a 24^th of a lane of a flow cell, with other samples from various projects. Raw sequencing reads were cleaned using NGS QC toolkit v2.3.3 [78] by removing adapter sequences, reads with ambiguous bases and reads with less than 80% of positions with a quality score above 20. Low quality bases (q<20) were removed from the 3' end of remaining reads. Species identity and branch length variability within the genus were assessed by assembling the nuclear ribosomal internal transcribed spacer (nrITS) using NOVOPlasty 2.5.9 [79]. The assembled sequences were aligned to nrITS sequences used in Larsén and Rydin (2015) [18] using MAFFT v7.164 [80]. A phylogeny for this marker was then produced using RAxML v8.2.11 [81], with a GTR + G + I substitution model, identified as the best-fit substitution model through hierarchical likelihood ratio tests (Fig 1).

Chloroplast data matrix

Cleaned reads from Isoetes and Littorella corresponding to the chloroplast genomes were assembled using NOVOPlasty, with a 39-bp kmer and a seed sequence of the I. flaccida chloroplast genome [71]. In cases where a circular chloroplast genome was not produced, contigs were aligned to the I. flaccida chloroplast genome using blastalign [82], and reads corresponding to regions of the reference chloroplast genome not covered by the contigs were used as seed sequences to assemble new contigs. All contigs were subsequently realigned to the reference genome, and overlapping contigs were merged. Chloroplast genome assemblies from 24 additional species representing the major embryophyte taxa, including two Selaginella species, were downloaded from NCBI database [83–101] (S2 Table).

Chloroplast protein-coding genes were identified from all chloroplast genomes using DOGMA [102] and coding sequences were extracted using TransDecoder v2.1.0 [103]. A total of 64 genes were identified and aligned by predicted amino acids using t-coffee [104] and MAFFT. Gene alignments were manually inspected and trimmed using AliView [105]. Twelve of them (clpP, cysaA, psi_psbT, rpl16, rpl21, rps15, ycf1, ycf2, ycf3, ycf10, ycf66, ycf68) were discarded either due to poor homology or alignment difficulties, or because sequences were obtained from less than 10 of the 33 chloroplast genomes analysed. The remaining 52 chloroplast genes were concatenated, producing a 55,542 bp matrix, with 33,582 polymorphic and 25,501 parsimony informative sites. A maximum likelihood phylogeny was generated using RAxML, with a GTR + G + I model of sequence evolution, determined to be the best-fit model using hierarchical likelihood ratio tests. The same matrix was later used for molecular dating.

Nuclear data matrices

Cleaned RNAseq reads of I. lacustris were assembled using Trinity v2.3.2 [103], resulting in 285,613 contigs with an average length of 689 bp. A similar procedure yielded 159,920 contigs for L. uniflora, with an average length of 769bp. For each species, the longest open reading frames (ORFs) were extracted using TransDecoder, and for each unigene the contig with the longest ORF was used to build a reference dataset. Cleaned reads from the whole-genome sequencing for each of the Isoetes species were then separately mapped to this reference dataset using bowtie2 v2.3.2 [106] in local mode to avoid excluding reads overlapping exon/intron boundaries. Alignments with MAPQ quality below 20 were excluded using SAMtools v1.5 [107]. The SAMtools mpileup utility was then used to generate for each species consensus sequences from the reads mapping to each I. lacustris transcript.

Gene duplication and losses are common in nuclear genomes–polyploidy is common in Isoetes, including in I. lacustris which has previously been identified as a decaploid [53]. Therefore, a combined reciprocal best blast and phylogenetic approach was adopted to identify groups of co-orthologs covering I. lacustris and the other land plants. Families of homologous ORFs generated by the method of Vilella et al. [108] were downloaded from EnsemblPlants. In total, 4,516 homolog families highly conserved among land plants (containing at least one sequence from Physcomitrella patens, Selaginella moellendorffii, Amborella trichopoda, Oryza sativa, Arabidopsis thaliana and Theobroma cacao) were used for subsequent ortholog identification.

Transcriptome and coding sequence data from seven additional species representing different embryophyte groups were retrieved from the literature [109–115] (S3 Table) and ORFs were extracted. Reciprocal best protein BLAST searches assigned ORFs of I. lacustris, L. uniflora and the additional embryophyte species to homolog families, with a minimum match length of 50 amino acids and e-value of 10⁻⁷. The expanded homolog families were then aligned according to their protein sequences using MAFFT, and phylogenies were constructed using RAxML and the GTR + G + I model, which fits most genes and is therefore appropriate for constructing large numbers of gene trees [116–118]. The longest sequence of each monospecific clade of sequences belonging to the same species was identified using custom scripts to remove transcripts representing the same gene or genes that duplicated after the divergence from all other species. These sequences were then realigned and a new phylogeny was inferred. Sets of 1:1 orthologs were then identified as clades containing exactly one gene per species, resulting in 30,258 groups of co-orthologs. Of these, 2,165 contained more than nine species, including I. lacustris, S. moellendorffii and either P. patens or Ceratodon purpurea, which were needed to use some of the fossil calibration points. By restricting our analysis to these 1:1 orthologs, we eliminate the possibility of non-orthologous genes resulting from gene or genome duplications being considered as orthologs. These 2,165 orthogroups were realigned, and consensus sequences of the genome skimming data were added to the alignments. Only the 782 orthogroups containing sequences for I. coromandelina sensu lato, which is necessary to capture the earliest split among extant species of Isoetes [18] (Fig 1), were considered further. New phylogenetic trees were inferred from these datasets, and genes failing to recover the monophyly of the vascular plants, Isoetopsida (Isoetes plus Selaginella, [119]) or Isoetes were considered phylogenetically uninformative and excluded. The remaining 292 datasets were deemed suitable for the phylogenetic problem addressed here, and were used for molecular dating. A phylogenetic tree was inferred separately for each of these markers, and a maximum likelihood phylogeny was also inferred using the 694,437 bp concatenated alignment, which was 41.14% complete with 443,864 polymorphic and 316,350 parsimony informative sites.

Calibration points

Time-calibrated trees were inferred from the different markers using the same set of calibration points. To date the crown node of extant Isoetes, a fossil constraining the crown node of extant Isoetes would be ideal–such a constraint would require a fossil containing a synapomorphy from one of the two descendant branches of this node. Previous studies have identified a geographically diverse group of Isoetes including I. coromandelina sensu lato as the outgroup to the rest of the Isoetes [18, 56, 65]. Whilst the I. coromandelina complex itself contains a number of features initially thought to identify this as diverging earliest from other extant members of the group, its presence within “Clade A” identifies these features as derived [18, 55]. No morphological features reliably distinguishing Clade A and the rest of the Isoetes appear to exist [18, 55]. Therefore, no fossil will contain features distinguishing these two groups, so fossils cannot provide a hard minimum age for this node. Within the Isoetes crown group, no morphological features clearly divide different clades [18, 42, 58]. A number of features vary between taxa and clades, such as corm lobation and glossopodium structure, but these are either not widely characterised across the genus or show multiple transitions within clades [52, 120]. The fossil record therefore does not allow implementing hard minimum ages for nodes within Isoetes.

The nearest node to the Isoetes crown node for which reliable synapomorphies are available is the crown node of the Isoetopsida (Isoetes plus Selaginella). The Isoeptopsida are a well supported clade appearing in the Devonian, containing synapomorphies such as a heterospory and a ligule [119, 121]. Isoetalean lycopsid trees are considered to form a clade within the Isoetales, being more closely related to Isoetes than Selaginella. This assessment is based on shared synapomorphies including bipolar growth from a shoot like “rhizomorph” structure and secondary woody tissue [48]. Arborescent lycopsids are known from the Frasnian [122, 123] (382.7–372.2 Ma), although the rhizomorph root structure could not be identified in these early fossils. However, discovery of a putatively homosporous arborescent lycopsid (the Isoetales are heterosporous) suggests that arborescence could be a convergent phenotype within the lycopods [124]. As multiple examples of isoetalean arborescent lycopsids, including rhizomorphs, are known from Famennian strata [125, 126] (358.9 to 372.2 Ma), a minimal age of 358 Ma was implemented using a uniform distribution between 358 and 485 Ma.

A maximum age constraint of the crown node of all land plants was set based on the appearance of cryptospores in the fossil record. These abundant spores are considered a likely synapomorphy of early land plants [127]. Their appearance in the fossil record is therefore likely to occur soon after the origins of land plants, making them appropriate for setting a maximum age for land plants [18, 128]. The earliest unequivocal cryptospores are found in the early Middle Ordovician [129] (473–471 Ma). However, pre-Middle Ordovician terrestrial sediments are rare [130], and as no unequivocal cryptospores are found in pre-Ordovician rocks [128, 131], the beginning of the Ordovician (485 Ma) was used as a conservative upper limit for the age of land plants. Whilst other molecular dating studies have estimated the age for this node to be comparable or older than this date (e.g.[132–134]), these are based on comparable fossil evidence but with soft maxima assigned to this node, allowing older age estimates than the hard maximum approach used in the present study. This maximum age was used to constrain the crown node of the liverworts plus the rest of vascular plants in the chloroplast dataset, and the crown node of the bryophytes plus vascular plants in the nuclear dataset. The minimum age of the same node in both cases was constrained by the early vascular plant macrofossil, Baragwanathia longifolia from the Ludlow epoch in the Silurian at 421 Ma [135–137]. This is of a similar age to other putative vascular plant fossils such as Cooksonia in the Wenlock epoch [138–140].

For the chloroplast dataset, trees were rooted by constraining each of the liverworts and the rest of the land plants to be monophyletic [136]. For the nuclear dataset, which only contained bryophytes and vascular plants, the tree was rooted by enforcing the monophyly of each of these two groups. Details of fossil calibrations as outlined in Parham et al. [141] are outlined in S4 Table.

Molecular dating software

Molecular dating was performed using r8s [142] and BEAST [143], two commonly used relaxed-clock methods that differ in their general approach and the strategy used to assign rates to internal branches of the phylogeny. r8s implements a semiparametric method that uses a penalised likelihood approach to assign rates among branches [144]. The smoothing parameter, which determines the extent to which rates vary among branches, is determined for each dataset using an empirical approach [142]. The method takes a phylogram as input, assumes no uncertainty in topology, and uses a simplified model of nucleotide substitution. BEAST implements a highly parametrised Bayesian method that samples trees generated from nucleotide data using an explicit model of sequence evolution [145]. When using the relaxed molecular clocks implemented in BEAST, rates are uncorrelated across the tree, but an overall distribution of rates is assumed, with the mean and standard deviation inferred from the data.

For r8s, version 1.81 was used, with the “TN” algorithm and additive penalty function. Cross validation was performed for a range of smoothing parameters from 10⁻² to 10⁶, increasing by a power of 10^0.5 each time, and the best smoothing parameter was used for molecular dating. Confidence intervals were obtained by generating 100 bootstrap pseudo-replicates using seqboot [146] and obtaining branch lengths for each of these using RaxML (GTR+G+I) while constraining the trees to the topology generated by the original dataset. These trees were then individually dated using r8s, providing a distribution of ages across the pseudo-replicates. This approach was used to date the chloroplast dataset, the concatenated nuclear dataset, as well as individual nuclear markers.

For BEAST, version v1.8.4 was used. A lognormal relaxed clock was adopted with a GTR + G + I model of nucleotide substitution with four rate categories and a birth-death speciation prior. For the concatenated chloroplast markers, four independent analyses were run for at least 20,000,000 generations and appropriate burn-in periods (at least 10%) were assigned by inspection of the traces using Tracer v1.6 [143]. For individual nuclear genes, BEAST was run for 3,000,000 generations (based on observing convergence times with a subset of genes) with a burn-in period of 50%. Dating the concatenated nuclear dataset was computationally too intensive with this approach. We therefore randomly subsampled 55,743bp (approximately the length of the chloroplast alignment) from the 694,437bp nuclear alignment eight times, and analysed these subsamples with BEAST. The same parameters as the individual nuclear genes were used, with the exception that BEAST was run twice for 10,000,000 generations for each alignment, with a burn-in period of 10%, with convergence verified using Tracer. These were combined using logcombiner and treeannotator to produce a maximum clade credibility tree. For comparison we performed r8s on these alignments as described previously.

Phylogenetic reconstruction using Isoetes nrITS sequences

The maximum likelihood phylogeny based on nrITS broadly agrees with phylogenies previously inferred for the group (Fig 1) [18], and confirms that the species sampled for genomic analyses encapsulate the first divergence within the Isoetes species for which molecular data is available (Fig 1). These six samples are moreover representative of the group in terms of branch lengths, and therefore evolutionary rates (Fig 1). The samples sequenced in this study generally cluster with other individuals from the same species sequenced previously, with the exception of Isoetes nuttallii. The newly sequenced sample of I. nuttallii groups with I. asiatica and I. echinospora, disagreeing with the topologies found in other studies [18, 55]. The sample in this study was collected on Knight Island in the Prince William Sound, Alaska, USA, outside the known range of I. nuttallii which extends up to British Columbia at its northernmost [147]. I. asiatica, I. echinospora and I. occidentalis and their interspecific hybrids occur throughout Alaska [148]. The underwater growth habit and large megaspores in this specimen are inconsistent with previously identified I. nuttallii specimens that are typically emergent [148, 149]. These features are consistent with the hexaploid I. occidentalis, a close relative of I. asiatica that has large megaspores, an underwater growth habit and is native to the area [43, 56, 149]. Furthermore, two duplicates of this collection have been subsequently redesignated as I. occidentalis. We therefore refer to this specimen as I. occidentalis throughout the remainder of this manuscript and associated data files.

Phylogenetic reconstruction and dating based on the chloroplast genome

The maximum likelihood phylogeny based on chloroplast markers recapitulated major land plant relationships and expected relationships within the Isoetes clade, with I. coromandelina sensu lato being sister to the rest of samples (Fig 2A). The tree was well resolved, with only the Ceratophyllum/eudicot split receiving less than 95% bootstrap support. Branch lengths were highly variable, particularly between Isoetes and Selaginella, with the latter having accumulated approximately 4.5 times more substitutions than Isoetes since their most recent common ancestor (Fig 2A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Maximum likelihood phylograms of concatenated chloroplast and nuclear markers.

Phylograms are shown for a) concatenated chloroplast markers and b) concatenated nuclear markers. Branch lengths are proportional to the number of expected substitutions per site, with scale bar representing a) 0.07 and b) 0.2 substitutions per site. All bootstraps support values are 100 with the exception of the branch separating Ananas comosus from the clade containing Ceratophylum demersum in b), which has a support value of 85.

https://doi.org/10.1371/journal.pone.0227525.g002

Based on chloroplast markers, r8s estimated the age of the crown group of Isoetes at 24.2 Ma with an optimum smoothing parameter of 1000 identified by cross validation, and a 95% bootstrap confidence interval of 22.8–25.9 Ma (near the Oligocene-Miocene boundary; Table 1). Decreasing the value of the smoothing parameter resulted in an increased age of the Isoetes crown group, with a smoothing value of 0.01 giving a crown age of Isoetes of 219 Ma (Late Triassic; Fig 3). Whilst low smoothing values result in over-fitted models that perform poorly in cross validation, high levels of smoothing may produce rates that are nevertheless poor predictors of branch lengths in particular parts of the tree. For high smoothing values, the ratio of the effective rate (the branch length divided by the estimated time elapsed) to the rate assigned by the model was 0.33 for the stem branch of Isoetes (Fig 4), showing that the branch is significantly shorter than would be expected for the assigned rate and divergence time. On the other hand, the average ratio for the crown branch lengths remains near 1 for all smoothing values, indicating that the crown branches are close to the expected values for the assigned rates and divergence times (Fig 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Effect of different smoothing factors on Isoetes crown date estimation in r8s.

Estimated crown dates for Isoetes produced by r8s for concatenated chloroplast (green) and nuclear (blue) datasets are shown for a range of smoothing factors. The best fitting smoothing factor, as identified by cross validation, is highlighted for each dataset by a filled circle.

https://doi.org/10.1371/journal.pone.0227525.g003

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Rate assignment on crown and stem branches of Isoetes in r8s.

The ratio of effective vs. assigned rates is shown for different smoothing factors in r8s for the stem branch of Isoetes (solid lines) and for the Isoetes crown branches (average; dashed lines), for the concatenated chloroplast (green) and nuclear (blue) datasets. Solid black line represents effective/assigned rate ratio of 1, for reference.

https://doi.org/10.1371/journal.pone.0227525.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Estimates of Isoetes crown date.

https://doi.org/10.1371/journal.pone.0227525.t001

For the same chloroplast markers, BEAST estimated the crown age of Isoetes at 23.2 Ma (95% HPD = 6.4–46.8 –; middle Oligocene; Table 1), similar to the value obtained with the optimum level of smoothing in r8s. Unlike in r8s, rates in BEAST can vary throughout the tree, but their distribution is assigned a priori–in this case a lognormal distribution. Rates in the maximum clade credibility tree accordingly follow a lognormal distribution—the log-transformed rates following a straight line on a quantile-quantile plot, indicating the rates are distributed lognormally (Fig 5). Notably, the Isoetes stem branch is assigned the lowest rate in the tree and the crown branches assigned rates closer to the average rates in the rest of the tree (Fig 5). While these rate assignments lead to a lognormal distribution that satisfies the priors, they result in a lower rate in the Isoetes stem branch compared to the crown branches.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Quantile-quantile plot of BEAST rates for concatenated chloroplast markers.

The quantile-quantile plot of log₁₀ transformed branch rates is shown for the concatenated chloroplast dataset in BEAST. The values for the Isoetes stem branch (blue) and crown branches (red) are highlighted.

https://doi.org/10.1371/journal.pone.0227525.g005

For both r8s and BEAST, a date of 23–29 Ma (Oligocene) is obtained via the implicit or explicit inference of a decrease in the rate of evolution along the stem branch, with rates in the crown branches being more similar to those in the rest of the tree. This assumption results from the model, and is not necessarily correct, urging for independent evidence.

Phylogenetic reconstruction and dating based on nuclear markers

The concatenated nuclear phylogram also recapitulated major land plant relationships (Fig 2B). The topology of the Isoetes clade was consistent with that of the chloroplast phylogeny, with I. coromandelina again being sister to all other species. Despite overall longer branch lengths in the concatenated nuclear phylogeny, variation among groups was reduced. Particularly, the total branch lengths from the common ancestor of Isoetes and Selaginella were much more similar than in the chloroplast phylogeny, with Selaginella having accumulated approximately 1.25 times more mutations than Isoetes since their last common ancestor (Fig 2B). However, the ratio of the average crown branch length to stem length in the Isoetes lineage was very similar between the nuclear and chloroplast markers; approximately 5.8 for the chloroplast dataset and 5.6 for the nuclear dataset (Fig 2).

Dating of the concatenated matrix of nuclear markers in r8s gave an estimated crown node age of Isoetes at 58.9 Ma (Table 1), with an estimated stem node age of 358 Ma, at an optimum smoothing value of 0.1. Unlike with the chloroplast markers, the date of the Isoetes crown node was similar across all smoothing values tested (Fig 3). Increased smoothing values led to increases in the disparity between effective and assigned stem rates (Fig 4), although this was low compared with the concatenated chloroplast alignment (0.82 vs 0.33 for the stem branch for a smoothing value of 10⁶). As with the chloroplast markers, the disparity between effective and assigned rates in crown branches was low across the range of smoothing values (Fig 4). The conservation of the effective rates in the stem and crown branches of Isoetes across a range of smoothing parameters indicates that the average rates predicted across the entire nuclear tree are a relatively good fit to both stem and crown branches of Isoetes (Fig 4). This suggests that stem and crown branches of Isoetes have similar rates, which is consistent with their highly similar length ratios between the chloroplast and nuclear trees (Fig 2).

Dating individual nuclear genes in r8s resulted in a wide range of optimum smoothing values (S1 Fig). Low smoothing values frequently resulted in gradient check failures, indicating a single optimum solution is not reached (S1 Fig). For genes reaching a single optimum, the median estimated crown date for Isoetes was 46.4 Ma with 95% of estimates between 16.1 and 85.8 Ma and 50% of results between 31.9 and 58.3 Ma (Table 1). Overall, the estimated dates form a unimodal distribution (S2 Fig). While low values of the smoothing parameter increased the age estimates, all values above 10 yielded estimates centred around 50 Ma, similar to those based on the optimum smoothing values (S2 Fig). As with the chloroplast datasets, increasing smoothing values resulted in a decreased effective/assigned stem rate (S3 Fig). The disparities for the optimum smoothing values were again reduced compared to the chloroplast data (S3 Fig), indicating the globally optimum smoothing values for the individual nuclear markers fit the stem and crown branches of the Isoetes better than in the chloroplast dataset.

Dating individual genes using BEAST gave a median estimate of 47.6 Ma for the crown of Isoetes, with 95% of estimates between 24.1 and 90.1 Ma, and 50% between 39.2 and 58.6 Ma. The ages obtained for individual genes were highly correlated between r8s and BEAST (linear model, slope = 0.94, p-value < 0.001; R² = 0.64; Fig 6). Linear modelling suggested a significant but small effect of the percent completeness of the alignments on the estimate for the crown age of Isoetes, with a larger effect from the average completeness of Isoetes sequences (S5 Table). However, the adjusted R² for this latter effect was 0.059 for values obtained with r8s and 0.042 for those obtained with BEAST, indicating that the completeness of the alignment has relatively little impact on the estimated dates. BEAST dating of 55,743bp subsamples of the concatenated dataset gave a mean crown date for Isoetes as 54.5 Ma (mean 95% HPD 27.9–85.2Ma; Table 1; Fig 7). The eight individual subsamples gave very similar estimates of the mean crown date, with a standard deviation of 2.8Ma between the different subsamples. r8s gave a slightly older estimate, 62.9 Ma (mean 95% HPD 54.0–62.9; Table 1) with a standard deviation of 2.9Ma between the mean estimates of the different subsamples.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. r8s versus BEAST Isoetes crown estimates for individual nuclear genes.

The scatterplot shows the estimates of the Isoetes crown date in r8s and BEAST for each individual nuclear gene. Line represents output of linear model using lm() function in R v3.5.2.

https://doi.org/10.1371/journal.pone.0227525.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Maximum clade credibility tree for combined nuclear subsamples in BEAST.

Node labels represent ages (Ma), blue bars represent 95% mean HPD intervals.

https://doi.org/10.1371/journal.pone.0227525.g007

Nuclear analysis supports a recent origin of extant Isoetes

In this study, we used phylogenomics to estimate the age of Isoetes, a group of lycopods often interpreted as “living fossils”. Using molecular dating with calibration points on deep branches of the land plant phylogeny, we found very different dates for the crown of Isoetes using the chloroplast and nuclear datasets, at 23–24 Ma (Oligocene) and 45–60 Ma (Paleocene and Eocene), respectively (Table 1). These differences are unlikely to be caused by the dating methods employed, since BEAST and r8s produced almost identical dates (Table 1; Fig 6), despite the very distinct ways in which these two programs deal with rate variation among branches. Subsets of 55,743bp (approximately the same size as the chloroplast alignment) of the concatenated nuclear alignment gave dates consistent with the other nuclear datasets, indicating alignment size was not the cause of this disparity either (Table 1). Instead, the incompatibilities between estimates based on nuclear and chloroplast datasets probably arise from differences in rate variation among branches. Branch lengths varied widely between Selaginella and Isoetes chloroplast markers, and throughout the chloroplast tree (Fig 2A), as previously reported [18, 69–71]. These high levels of variability make low levels of smoothing in r8s relatively poor fits to the data, as rates are poorly correlated between nearby branches on the tree, resulting in poor cross-validation scores. This results in a high smoothing value being identified as optimum for the chloroplast tree, effectively forcing a uniform rate on the tree that is determined by the average root-to-tip branch length. That, in turn, results in a high rate fitted to the Isoetes branch that is a poor match to its relatively short branch lengths overall (Fig 2A, Fig 4). The overall model likelihood in r8s is calculated as the sum of the log likelihood of each branch [142], meaning a date for the crown node will be assigned primarily to optimise the fit for the numerous crown branches as opposed to the single stem branch. For higher rates as smoothing values increase, younger Isoetes crown dates are therefore assigned. This results in a good fit for the short crown branches, but an increasingly worse fit for the stem branch (Fig 4), where the high rate and long temporal duration predict a long branch, which is not observed in the data. Similarly, in BEAST, the lognormal prior distribution results in a relatively low rate assignment on the stem branch compared to the crown branches, which leads to a better fit to the lognormal distribution across branches than if all crown branches had a low rate (Fig 5). We conclude that the high rate variability hampers accurate molecular dating using the chloroplast data. By contrast, the individual and concatenated nuclear datasets have a small disparity between estimated and effective rates, and the crown age estimate for Isoetes is consistent across genes (Figs 4 and S2), both in the concatenated versus individual datasets (Table 1) and between BEAST and r8s (Fig 6). The more consistent rates make the nuclear dataset more appropriate for estimating divergence times.

We conclude, based on our nuclear genome-wide analyses, that the diversity of extant Isoetes most likely originated during the Paleogene, between 45 and 60 Ma (Paleocene-Eocene; Table 1), although an origin from the Late Cretaceous to the Micoene is within the 95% confidence interval (16–86Ma; Table 1). This conclusion sharply contrasts with previous estimates of the crown group Isoetes of 147–251 Ma (Triassic to Jurassic) [18, 57, 58]. Kim and Choi [57] used Triassic I. beestonii to provide a narrow lognormal prior with an offset of 245.5, a mean of 1.5 and standard deviation of 0.5Ma for the age of crown Isoetes, resulting in an estimate of the crown age at 251Ma. Pereira et al. [58] used Jurassic I. rolandii to provide a minimum age for the crown node of 145 Ma, resulting in an estimate of the crown age at 147Ma (145–154 95% CI). In both these studies, fossils are used to a priori strongly constrain the crown node of Isoetes to old ages. However, these fossils do not provide evidence that the split between “clade A” and the rest of Isoetes had occurred, as no synapomorphies are known from the extant members of these clades that could be preserved in fossils. The differences between these previous studies and our own emphasise the impact of fossil calibrations on date estimates. The study of Lars00E9n and Rydin (2015) [18] used fossil calibrations consistent with those used in the present study, but nevertheless estimated a crown age of 147Ma [96–215 95% CI]. Their study was based on three markers, with only rbcL aligning with sequences outside of the genus. The markers that do not align outside Isoetes should not affect the crown node age estimate as they do not inform the ratio of crown to stem substitutions. However, including noncoding markers such as nrITS with much higher levels of substitution than rbcL may result in a young Isoetes crown node giving a poor model fit for a single partition. Indeed, maximum likelihood phylogenies of the full dataset have Isoetes crown branch lengths that are five times longer when compared to the same taxa only with rbcL (S6 Table). To investigate the effects of this imbalance, we reanalysed the dataset using BEAST with the same parameters as Larsén and Rydin (2015) using only Isoetes species for which all three markers were available, finding a similar crown age of 145.8 Ma [88–208.7 95% CI]. Removal of the non-coding markers available solely for Isoetes species results in a crown estimate for Isoetes of 40.5 [22.6–61.6 95% CI] Ma, comparable with the results of the present study (S7 Table; BEAST files used for reanalysis available as File S6). We conclude that calibration points and molecular data both strongly impact age estimates in the case of Isoetes.

The number of taxa sampled in this study is however lower than in these previous studies. Reduced taxon sampling has been shown to have an impact in some [61, 62], but not all [63, 64], cases, with high levels of rate heterogeneity likely requiring increased taxon sampling [62]. The relatively low levels of rate heterogeneity in Isoetes (Fig 1) indicate this is unlikely to affect our age estimates, and reduction of the taxon sampling in the comprehensive Larsén and Rydin [18] dataset by 87% only resulted in a 7% change in the estimated age of the Isoetes crown date (see Materials and Methods). Reanalysis of the entire Larsén and Rydin [18] dataset only with markers alignable outside Isoetes resulted in a similar age estimate to the present study, despite the significant differences in taxon sampling (S7 Table). These considerations suggest that rather than taxon sampling, the distribution of nucleotide data among groups explain the differences between our study and that of Larsén and Rydin [18]. Therefore, while using appropriate fossil calibrations is always critical, the choice of molecular data can also have a large impact on estimated dates.

Despite our improved molecular dataset and careful assignment of fossils, the long gap between the Isoetes crown node and the nearest available calibration points presents a challenge in appropriate rate assignment for any node-based dating approach. Total-evidence based approaches [150] may be able to leverage the rich Isoetalean lycopsid fossil record [42, 48] to inform estimates of the rate along this branch. Nevertheless, groups such as Isoetes represent a particular challenge for molecular dating, necessitating careful treatment of fossil and molecular data, and the modelling approaches that use these datasets to produce age estimates. Our approach generated nucleotide data that are homogeneously distributed among taxonomic groups, and the fossil evidence is used cautiously, even though this results in a great distance between the calibrated nodes and our node of interest, the crown node of Isoetes. These considerations allow disentangling the significant effects of methodological variation, rates of molecular evolution, and treatment of fossils on the molecular dating of a group of “living fossils”.

Despite morphological stasis, Isoetes recently expanded

The relatively young age of the Isoetes crown node indicates that despite morphologically similar forms appearing in the Triassic [46–48], all modern Isoetes are descended from a single lineage in the early Cenozoic. This indicates that the fossil Isoetites from the Jurassic, and morphologically similar plants from earlier epochs, are likely stem relatives of extant Isoetes. The results contrast with the expectation for “living fossil” taxa, that extant species members are the last members of once diverse lineages, diverging long in the past [19, 20]. This is consistent with a number of studies in some “living fossil” plant groups such as cycads [13], bryophytes [31] and Ginkgo [26] showing relatively recent origins of extant species of these groups, despite long periods of morphological stasis. It is important to note, however, that other groups fit with the traditional expectations for”living fossils” [17, 27, 28]. When compared to other spore-producing plants, the pattern identified in this study is similar to the high levels of diversification seen in ferns and Lycopodiaceae since the Cretaceous [132, 151, 152], but contrasts with steady patterns of diversification over time in Selaginella [153]. Within both “living fossil” taxa and spore-producing plants, there appears to be a variety of patterns of lineage-diversification through time.

The global distribution of extant Isoetes indicates that this lineage was able to successfully colonise the globe in a relatively short amount of time. This contradicts the conclusions of previous studies that, based on older estimates for the age of the Isoetes crown node, explain current distributions by vicariance due to continental drift [18, 57, 58]. Based on our age estimates, geographic disparities within several subclades of Isoetes–such as Larsen and Rydin’s [18] Clade B containing Mediterranean, North American and Indian species–indicate that long distance dispersal events have been relatively common Cenozoic Isoetes. It is important to note that numerous geographic disparities would remain with older age estimates. For example, the closely related Indian and Australian clades in Larsen and Rydin’s Clade E diverged less than 15 Ma despite the separation of these continents during the Jurassic [154]. Further studies of Isoetes dispersal rates and mechanisms, which are poorly understood, are required [18, 155–157]. It should however be noted that many relationships within the Isoetes are poorly supported, and rely on a small number of genetic markers and taxon samples–and indeed many additional cryptic species within current taxa may exist [158]. Further studies will be required to fully explain extant distributions of Isoetes species.

The rapid global spread of extant Isoetes strongly contrasts with the expectation that the distribution of “living fossil” taxa are the remnants of potentially larger ancestral ranges [17, 18]. Our results show that despite having undergone little morphological change for hundreds of millions of years, rather than being the declining remnants of a bygone era, the modern Isoetes species instead represent recent arrivals onto the world stage.