Chlorophyceae) and Provide Clues to the Complex Evolutionary History of Plastid Genomes in the Green Algal Tree of Life

Home , Hydrodictyaceae, Monactinus, Pseudopediastrum, Sorastrum, Sphaeropleales, Stauridium

RESEARCH ARTICLE

INVITED SPECIAL ARTICLE For the Special Issue: Using and Navigating the Plant Tree of Life

Organellar phylogenomics inform systematics in the green algal family Hydrodictyaceae (Chlorophyceae) and provide clues to the complex evolutionary history of plastid genomes in the green algal tree of life

Hilary A. McManus1,5, Karolina Fučíková2, Paul O. Lewis3, Louise A. Lewis3, and Kenneth G. Karol4

Manuscript received 5 July 2017; revision accepted 19 December PREMISE OF THE STUDY: Phylogenomic analyses across the green algae are resolving 2017. relationships at the class, order, and family levels and highlighting dynamic patterns of 1 Department of Biological and Environmental Sciences, Le evolution in organellar genomes. Here we present a within-family phylogenomic study to Moyne College, 1419 Salt Springs Road, Syracuse, New York resolve genera and species relationships in the family Hydrodictyaceae (Chlorophyceae), for 13066, USA which poor resolution in previous phylogenetic studies, along with divergent morphological 2 Department of Natural Sciences, Assumption College, 500 Salisbury Street, Worcester, Massachusetts 01609, USA traits, have precluded taxonomic revisions. 3 Department of Ecology and Evolutionary Biology, University of METHODS: Complete plastome sequences and mitochondrial protein-coding gene sequences Connecticut, Storrs, Connecticut 06269, USA were acquired from representatives of the Hydrodictyaceae using next-generation 4 Lewis B. and Dorothy Cullman Program for Molecular sequencing methods. Plastomes were characterized, and gene order and content were Systematics, The New York Botanical Garden, 2900 Southern Boulevard, Bronx, New York 10458, USA compared with plastomes spanning the Sphaeropleales. Single-gene and concatenated-gene 5 Author for correspondence (e-mail: [email protected]) phylogenetic analyses of plastid and mitochondrial genes were performed. Citation: McManus, H. A., K. Fučíková, P. O. Lewis, L. A. Lewis, and KEY RESULTS: The Hydrodictyaceae contain the largest sphaeroplealean plastomes thus K. G. Karol. 2018. Organellar phylogenomics inform systematics in far fully sequenced. Conservation of plastome gene order within Hydrodictyaceae the green algal family Hydrodictyaceae (Chlorophyceae) and provide is striking compared with more dynamic patterns revealed across Sphaeropleales. clues to the complex evolutionary history of plastid genomes in the green algal tree of life. American Journal of Botany 105(3): 315–329. Phylogenetic analyses resolve Hydrodictyon sister to a monophyletic Pediastrum, though doi:10.1002/ajb2.1066 the morphologically distinct P. angulosum and P. duplex continue to be polyphyletic. Analyses of plastid data supported the neochloridacean genus Chlorotetraëdron as sister to Hydrodictyaceae, while conflicting signal was found in the mitochondrial data.

CONCLUSIONS: A phylogenomic approach resolved within-family relationships not obtainable with previous phylogenetic analyses. Denser taxon sampling across Sphaeropleales is necessary to capture patterns in plastome evolution, and further taxa and studies are needed to fully resolve the sister lineage to Hydrodictyaceae and polyphyly of Pediastrum angulosum and P. duplex.

KEY WORDS Chlorophyta; chloroplast; Hydrodictyon; monophyly; organelle; Pediastrum; phylogeny; plastome; Sphaeropleales.

American Journal of Botany 105(3): 315–329, 2018; http://www.wileyonlinelibrary.com/journal/AJB © 2018 The Authors. American Journal of Botany is published by Wiley Periodicals, Inc. on behalf of the Botanical Society of America. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. • 315 316 • American Journal of Botany

Studies using phylogenomic data, particularly complete organellar and Pokorný, 2002; Lezine et al., 2005; Milecka and Szeroczýnska, genomes, are becoming standard, and the field of plant evolution 2005; Weckström et al., 2010). has seen an increase of such studies focusing on family-level ques- Historically, four genera have been recognized in the tions in particular (e.g., Henriquez et al., 2014; Guo et al., 2017). Hydrodictyaceae: Euastropsis Lagerheim 1895, Hydrodictyon Although organellar genome data have contributed valuable in- Roth 1797, Pediastrum Meyen 1829, and Sorastrum Kützing 1845. formation for higher-level systematics in phycology (e.g., Brouard However, several molecular phylogenetic studies, using two or three et al., 2010), until recently most studies have focused on sampling targeted DNA sequence regions, have consistently yielded Pediastrum one or a few representatives per major lineage for organellar ge- as paraphyletic. As a result, five additional genera have been recog- nome sequencing. Phylogenomic studies in green algae have re- nized: Lacunastrum H.McManus 2011, Monactinus Corda 1839, vealed numerous patterns and processes of evolution typically not Parapediastrum E.Hegewald 2005, Pseudopediastrum E.Hegewald reported in land plant plastomes. For example, extensive size vari- 2005, and Stauridium Corda 1839 (Buchheim et al., 2005; McManus ation within and among orders, as well as highly rearranged gene and Lewis, 2005, 2011; McManus et al., 2011). Though significant orders and gene gain/loss, are characteristic of the green algal lin- progress has been made in Hydrodictyaceae phylogeny and classifi- eages so far examined in this respect (Smith et al., 2013; Turmel cation, several unresolved questions remain. For example, the place- et al., 2015; Fučíková et al., 2016a, 2016b). The green algal order ment of Hydrodictyon is problematic, being resolved as either sister Sphaeropleales has recently received more attention with a broad to a monophyletic Pediastrum (Buchheim et al., 2005; McManus and sampling of representatives and increased number of molecular Lewis, 2005; Jena et al., 2014; Lenarczyk and McManus, 2016), or markers and whole organellar genomes analyzed (Farwagi et al., embedded within it rendering Pediastrum paraphyletic (Buchheim 2015; Lemieux et al., 2015; Fučíková et al., 2014b, 2014c, 2016a, et al., 2005; McManus and Lewis, 2011). Additionally, Pediastrum 2016b; McManus et al., 2017). A study of nine mitochondrial ge- contains two species (P. angulosum Ehrenberg ex Meneghini 1840 nomes within the order demonstrated conserved gene content but and P. duplex Meyen 1829), neither of which has been strongly sup- variability in intron content and genome size, and gene order analy- ported as monophyletic. A subset of P. duplex (“Group I”; McManus ses indicated potential to help resolve phylogenetic relationships at and Lewis, 2011) has been generally recovered, while the remainder the order and family levels (Fučíková et al., 2014b, 2014c; Farwagi of P. duplex has been recovered with P. angulosum as a polyphyletic et al., 2015). Characterization of the plastomes found extensive var- cluster (“Group II”; McManus and Lewis, 2011). The high genetic iation in overall size and inverted repeats (IR), and phylogenomic similarity and low nodal support, combined with considerable mor- analyses revealed support for different topologies depending on phological differences between Hydrodictyon, P. angulosum, and P. whether nucleotide or inferred amino acid sequences were used duplex, precluded taxonomic revisions. (Fučíková et al., 2016a). Clearly, we are still in the process of uncov- Within Hydrodictyaceae, mitochondrial gene order analyses ering the complexities of organellar genome evolution in green al- of four representatives indicated potential for resolving phyloge- gae. Denser taxon sampling within orders and families is the logical netic relationships (Farwagi et al., 2015), and a comparison of next step, which will allow addressing taxonomic issues associated two fully sequenced plastomes, one from Hydrodictyon reticula- with limited or convergent morphology that is a common problem tum (Linnaeus) Bory 1824 and the other from Pediastrum duplex, in microscopic green algae, and may uncover novel processes of or- showed that the two taxa shared gene order and plastome struc- ganellar genome evolution that help explain patterns seen at more ture, were more similar to each other than to other members of the inclusive taxonomic levels. Sphaeropleales and represented the largest plastomes known in the The Hydrodictyaceae (Sphaeropleales, Chlorophyta) is a remark- order (McManus et al., 2017). A phylogenetic analysis of plastome able group of green algae not only because of unique evolutionary gene order within Sphaeropleales supported the sister relationship trends discovered in their chloroplast and mitochondrial genomes, of Hydrodictyon and Pediastrum, but taxon sampling limited tests but also because of their evolving classification. They are found in of Pediastrum monophyly. freshwater and brackish systems worldwide and include ecologi- This study presents new plastomes from representatives cally important planktonic forms as well as nuisance bloomers. The of the Hydrodictyaceae, including Lacunastrum, Pediastrum, family is defined by their gametic flagellar apparatus ultrastruc- Pseudopediastrum, and Stauridium, and describes plastome archi- ture, similar asexual and sexual life cycle traits (Appendix S1, see tecture and gene order in the context of Sphaeropleales. We con- Supplemental Data with this article), and the formation of colonies ducted phylogenomic analyses on 65 protein-coding plastid (pt) with a fixed number of cells (coenobia) (Bold and Wynne, 1985; genes and 13 protein-coding mitochondrial (mt) genes and exam- Wilcox and Floyd, 1988). Morphological characters used to circum- ined the phylogenetic signal of individual genes. Together, results scribe generic and species boundaries include cell shape, presence/ from these analyses were used to address several major questions. absence and size of intercellular spaces, and patterns of the cellular First, do the Hydrodictyaceae share the trends of extensive plastome wall ultrastructure (Fig. 1A–L). The sporopollenin-like layer in the rearrangements found in other Sphaeropleales plastomes? Second, cell wall, especially of Pediastrum, has resulted in preservation in what is the sister taxon to Hydrodictyaceae? Third, is Pediastrum paleo-sediments and is used as paleolimnological indicators to re- monophyletic, with Hydrodictyon as its sister taxon (Fig. 2), and are construct lake biotopes (Komárek and Jankovská, 2001; Jankovská the two morphological species of Pediastrum resolved as distinct?

FIGURE 1. Photographs of representative taxa within Hydrodictyaceae taken with a Nikon E800 equipped with differential interference contrast optics. (A) Pediastrum duplex AL0403MN; (B) Pediastrum duplex SL0404MN; (C) Pediastrum duplex EL0201CT; (D) Pediastrum angulosum ACOI 1354; (E) Pediastrum duplex PL0501b; (F) Pediastrum angulosum KP0301SC; (G) Hydrodictyon reticulatum; (H) Pseudopediastrum boryanum IL0402MN; (I) Pseudopediastrum boryanum ML0410MN; (J) Pseudopediastrum integrum ACOI 577; (K) Pseudopediastrum sp. CL0201VA; (L) Stauridium tetras ACOI 84. Scale bar = 20 μM except in B (10 μM) and G (50 μM). March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 317

A B C

D E F

G H I

J K L 318 • American Journal of Botany

the personal collection of HAM. Herbarium barcodes (NY) are listed in Table 1.

Genomic data Total genomic DNA was extracted from living cells using the PowerPlant Pro DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA, USA) or following a CTAB extraction protocol (Doyle and Doyle, 1987). TruSeq library preparation was performed by Woodbury Genome Center A B at Cold Spring Harbor Laboratories, followed FIGURE 2. Illustration of two phylogenetic scenarios being tested. Scenario (A) depicts a by sequencing on Illumina HiSeq2500 to pro- monophyletic Pediastrum with Hydrodictyon recovered as sister lineage; (B) depicts a nonmono- duce 2×101-bp paired-end reads. Data were phyletic Pediastrum, with Hydrodictyon sister to one of the two groups, Group I (P. duplex) or trimmed, paired, and assembled de novo us- Group II (P. duplex + P. angulosum). ing the program Geneious v.9.1.5 (Biomatters, Auckland, New Zealand). Contigs containing pt and mt genome fragments were iden- Fourth, do all genes support the same phylogenetic conclusion or tified for each strain using BLASTN (National Institutes of Health, is there conflicting among-gene phylogenetic signal that is sympto- Bethesda, MD, USA; http://blast.ncbi.nlm.nih.gov/) after an initial de matic of Sphaeropleales? Despite the common assumption that orga- novo assembly. Geneious was then used to map reads to the selected nellar genomes function as a single locus, recombination and lateral fragments in a series of reference assemblies until longer fragments gene transfer have been documented, especially in plant mt genomes were obtained that could be joined into a single sequence; due to (e.g., Archibald and Richards, 2010 and references within) and, to assembly difficulties, no new complete mitochondrial genomes are a much lesser extent, in plastomes (e.g., Rice and Palmer, 2006). presented in this study. Annotations of the complete plastomes and Moreover, recent genomewide studies have shown strong phyloge- fragmentary mitochondriomes were done using DOGMA (Wyman netic conflict among genes from the same cellular compartment in et al., 2004; dogma.ccbb.utexas.edu/), BLAST, tRNAscan-SE 2.0 green algae, even though recombination or lateral gene transfer were (Lowe and Chan, 2016), RNAweasel (Lang et al., 2007; http://me- not directly implied (e.g. Fučíková et al., 2014c). Together, these gasun.bch.umontreal.ca/RNAweasel/), and Geneious. Genome maps analyses shed light on evolution in the Hydrodictyaceae and the were drawn using OrganellarGenomeDRAW (Lohse et al., 2013, unique mode of plastome evolution in this group. Although fewer http://ogdraw.mpimp-golm.mpg.de/), and the script illustrating the plastome rearrangements exist within the family compared to across number of genome rearrangements was written in JavaScript using the Sphaeropleales, the genome size varies considerably across the the D3 library (https://d3js.org). BLASTX similarity searches were family, predominantly due to infiltration of novel genes (ORFs) and used to characterize introns and freestanding open reading frames expansion of noncoding regions. We show that key phylogenetic (ORFs) ≥300 bp. The plastome ORFs were translated into protein relationships are supported with phylogenomic data, including the sequences and subsequently subjected to a 40% pairwise identity replacement of Hydrodictyon sister to a monophyletic Pediastrum. ciprocal comparison using BLASTP to find putatively homologous ORFs across the family (Addou et al., 2009). Repeats were identified using the method of Benson (1999) at http://tandem.bu.edu/ MATERIALS AND METHODS trf/trf.html. GenBank accession numbers are listed in Table 1 and Appendices S2 and S3. Cultures Phylogenomic analyses Specimens selected for this study include five strains of Pediastrum duplex (Fig. 1A–C, E), two strains of P. angulosum (Fig. 1D, F), one Initial single-gene translation-aided alignments generated by of Hydrodictyon reticulatum (Fig. 1G), two of Pseudopediastrum Geneious v.9.1.5 were inspected visually, and ambiguously aligned boryanum (Turpin) E.Hegewald 2005 (Fig. 1H, I), one of codons were masked for subsequent removal by a custom set of Pseudopediastrum integrum (Nägeli) M.Jena & C.Bock 2014 Python scripts similar to those used in Fučíková et al. (2016a). Each (Fig. 1J), an unidentified species of Pseudopediastrum (Fig. 1K), one data set was analyzed under a nucleotide GTR+I+G model parti- Lacunastrum gracillimum, and one Stauridium tetras (Ehrenberg) tioned by codon position, under a codon model, and under an amino E.Hegewald 2005 (Fig. 1L). Specimens were obtained from H. A. acid model. Because of the noncanonical genetic code used in chlo- McManus’s culture collection, from the Culture Collection of Algae rophycean mitochondria, a script (codon_fix.py) was developed to at the University of Texas, Austin (UTEX), and the Cooimbra recode all TAG codons as ambiguous (NNN) for the codon analyses Culture Collection of Algae, Cooimbra, Portugal (ACOI). All cul- of mitochondrial data. Each analysis was carried out in the maximum tures were maintained at 20°C under a 16 h light/8 h (L:D) cycle on likelihood (ML) framework using the program Garli (Zwickl, 2006) agar slants. The agar slants were prepared following McManus and and Bayesian framework (BI) using the program MrBayes (Ronquist Lewis (2011). Voucher material for each strain is deposited in The et al., 2012). Five searches for the best tree and 100 ML bootstraps New York Botanical Garden William and Lynda Steere Herbarium were conducted in Garli, and 5 million generations in each of two (NY), with duplicate specimens deposited in the George Safford runs were carried out in MrBayes. Configuration files and MrBayes Torrey Herbarium at the University of Connecticut (CONN), and in blocks are available in Appendix S4. March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 319

TABLE 1. Summary of Hydrodictyaceae plastomes. %Coding includes all CDS (including ORFs), tRNAs and rRNAs (both IRs); %GC content includes both IRs; Genes includes CDS (including ORFs), tRNAs and rRNAs (both IRs). NYBG Mean herbarium Size assembly % Coding Noncoding No. Species Strain GenBank barcode (bp) coverage GC (%) (bp) Repeats Genes IR (bp) Acutodesmus obliquus UTEX 393 DQ396875 n/a 161,452 — 26.9 56.0 55,454 125 106 12,022 Chlorotetraedron incus SAG 43.81 KT199252 n/a 193,197 — 27.1 46.7 94,081 120 106 13,490 Hariotina reticulata UTEX 1365 KY792638 - KY792700 n/a n/a n/a n/a n/a n/a n/a n/a n/a Hydrodictyon HAM0289 KY114065 02334980 225,641 — 32.1 59.6 91,268 389 111 18,296 reticulatum Lacunastrum ACOI 392 MF276976 02335035 187,836 70X 29.8 55.2 84,136 94 120 10,307 gracillimum Neochloris aquatica UTEX 138 KT199248 n/a 166,767 — 30.3 50.9 54,026 28 102 18,217 Pediastrum angulosum ACOI 1354 MF276977 02335031 219,917 29X 32.6 56.4 95,855 120 144 18,118 Pediastrum angulosum KP0301SC MF276978 02335032 195,439 147X 32.2 57.4 83,299 81 154 11,960 Pediastrum duplex AL0403MN MF276979 02335033 214,352 545X 33.0 53.1 100,621 158 141 14,208 Pediastrum duplex PL0501b MF276980 02335034 198,980 164X 31.5 54.1 91,369 91 124 12,889 Pediastrum duplex EL0201CT KY114064 02334981 232,554 ˆ 32.6 53.3 108,632 233 125 19,830 Pediastrum duplex SL0404MN MF276981 02335036 220,663 60X 33.6 54.9 99,402 185 149 13,557 Pediastrum duplex UTEX 1364 MF536514 - MF536522 n/a n/a n/a n/a n/a n/a n/a n/a n/a Pseudopediastrum IL0402MN MF276982 02335037 206,668 99X 31.2 53.8 95,472 129 127 17,775 boryanum Pseudopediastrum ML0410MN MF276983 02335038 200,630 129X 29.5 57.1 86,067 92 119 15,614 boryanum Pseudopediastrum ACOI 577 MF276984 02335039 187,268 104X 29.2 57.4 79,836 123 118 14,365 integrum Pseudopediastrum sp. CL0201VA MF276985 02335040 211,386 30X 29.3 54.8 95,648 213 133 18,343 Stauridium tetras ACOI 84 MF276986 02335041 143,108 37X 27.3 70.0 42,894 32 111 10,391

Although organellar genomes commonly are assumed to evolve frequencies, and GTR exchangeabilities, a flat Relative Rate as a single unit and without recombination, this assumption has prior (Fan et al., 2011) on partition subset relative rates, an been challenged by a number of studies (e.g., Andre et al., 1992; Exponential(1) prior on the discrete gamma shape parameter, Boynton et al., 1987). Therefore, the same set of analyses was per- and a hierarchical Exponential(μ) prior on edge lengths with an formed for each single-gene data set, for the concatenated data InverseGamma(2.1, 1/1.1) hyperprior on μ. The prior on tree set of all plastid protein-coding genes (65 genes after exclusion of topologies was such unless a tree topology explains the data better the hypervariable ftsH), and for the concatenation of the 13-mt (by an amount equal to 1 log likelihood unit) than a topology with protein-coding genes. Given the short internal branches in the phy- one edge collapsed, the less-resolved tree topology will be favored. logeny and the possibility of heteroplastidy or biparental plastid This gently favors polytomous trees while still allowing resolved inheritance, SVDquartets (singular value decomposition scores) trees to predominate if substitutions provide sufficient support for (Chifman and Kubatko, 2014, 2015), implemented in PAUP* v.4.0a all edges. The Python script used to run Phycas for each single- (build152, 2017) (Swofford, 2002, http://phylosolutions.com/paup- gene analysis is available in the Appendix S4 and may be consulted test/) was used to infer a species phylogeny of the concatenated data for other details. set under the coalescent model to test for incomplete lineage sort- To examine the relationship of evolutionary divergence and ing. All stripped alignments and resulting trees (Garli best ML trees plastome rearrangements, the average pairwise HYK85+gamma and Garli bootstrap tree files, Bayesian consensus trees) are availa- nucleotide distances for the concatenated pt gene data set, deter- ble in Appendix S4. mined with PAUP* v.4.0a152 (Swofford, 2002) was compared with The support for select topologies was summarized from single- the pairwise DCJ distances determined using the Mauve plugin gene analyses in three ways, separately for ML and BI. First, we (Darling et al., 2010) as implemented in Geneious. compared the proportions of the numbers of genes supporting each To test for phylogenetic signal in pt intergenic spacer regions, topology with ≥0.9 Bayesian posterior probability (BPP), while sup- these regions were annotated in five closely related taxa making up port values <0.9 BPP were considered unresolved. Second, the num- the Hydrodictyon-Pediastrum group and examined for similarity ber of genes favoring a particular topology (≥0.9 BPP) was broken and phylogenetic signal. down by analysis type (nt, codon, aa), and third, statistical support (BS and BPP) for each topology in each analysis was assessed with the aid of SumTrees (part of the DendroPy package, Sukumaran and RESULTS Holder, 2010) and a custom Python script. Phycas v. 2.2.1 (Lewis et al., 2015) was used to verify that Plastomes support for hypotheses of interest were not due to the polytomy artifact (Lewis et al., 2005) in selected single-gene analyses. The The new fully assembled plastomes of Hydrodictyaceae ranged in model used was GTR+I+G partitioned by codon position, flat size from 143,108 bp to 220,663 bp (Table 1). Each plastome con- Dirichlet priors on the proportion of invariable sites, nucleotide tained two single-copy (SC) regions and two copies of an inverted 320 • American Journal of Botany

Pseudopediastrum boryanum MF276983 200,630 bp

FIGURE 3. Gene map of Pseudopediastrum boryanum plastome (ML0410MN, MF276983). The inverted repeats (IRA and IRB) that separate the genome into two single-copy regions are indicated on the inner circle, along with the nucleotide content (G/C = dark grey, A/T = light grey). Genes labeled on the outside of the outer circle are transcribed clockwise, genes on the inside counterclockwise. Gene boxes are color-coded by functional group as shown in the key. repeat (IR) between 10,307 bp and 18,343 bp in size (Fig. 3; members of Chlorophyceae. The number of genic introns ranged Appendix S5). The mean assembly coverage ranged from 29× to from two to 10 in various positions and some shared common in- 545× (Table 1). All plastomes shared the same set of genes, except sertion sites (Figs. 4, 5). Introns were of Group I or Group II type that the trnY(gua) gene is absent in Stauridium tetras (Table 2). The and some contained ORFs with conserved domains with GIY-YIG, IRs share the same set of genes and zero to four total introns in HNH, or LAGLIDADG homing endonucleases (Fig. 5; Appendix the rrl and rrs rRNA genes. The psaA gene was trans-spliced as in S6). Between 32 and 389 small repeated sequences were found. other members of the Sphaeropleales, with exon 1 located in the One to 36 freestanding open reading frames (ORFs) ≥300 bp small SC and exons 2 and 3 in the large SC. The family contains a of unknown function were annotated in the plastomes, some con- trnL(uaa) intron that is seen in many Sphaeropleales and in other taining conserved domains (Fig. 4; Appendix S6). The reciprocal March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 321

TABLE 2. List of plastid-encoded genes annotated for Hydrodictyaceae. ts = trans-spliced; * = intron-containing gene in some taxa (see Fig. 6); X2 = duplicated gene not in inverted repeat (IR); X2IR = duplicated gene in IR; ✜ = absent in Stauridium tetras. Gene class Genes Ribosomal RNAs rrf x2IR, rrl * x2IR, rrs * x2IR Transfer RNAs trnA-UGC x2IR, trnC-GCA, trnD-GUC, trnE-UUC x2, trnF-GAA, trnG-UCC, trnH-GUG, trnI-CAU, trnI-GAU x2IR, trnK-UUU, trnL-UAA*, trnL-UAG, trnMe-CAU, trnMf-CAU, trnN-GUU, trnP-UGG, trnQ-UUG, trnR-ACG, trnR-UCU, trnS-GCU x2IR, trnS-UGA, trnT-UGU, trnV-UAC, trnW-CCA, trnY-GUA ✜ ATP synthase atpA, atpB *, atpE, atpF, atpH x2IR, atpI Chlorophyll biosynthesis chlB, chlL, chlN Cytochrome petA, petB, petD *, petG, petL Photosystem I psaA ts, psaB *, psaC, psaJ Photosystem II psbA *, psbB *, psbC, psbD *, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbM, psbN, psbT, psbZ * Ribosomal proteins rpl2, rpl5, rpl14, rpl16, rpl20, rpl23, rpl32, rpl36, rps2, rps3, rps4, rps7, rps8, rps9, rps11, rps12, rps14, rps18, rps19 RNA polymerase rpoA, rpoBa, rpoBb *, rpoC1, rpoC2 Hypothetical proteins ftsH, ycf1, ycf3, ycf4, ycf12 Miscellaneous proteins ccsA, cemA, clpP, infA, rbcL *, tufA

40% protein similarity comparisons of all ORFs revealed all plas- Gene order analyses tomes shared a large homologous ORF that contained a conserved domain with a putative reverse transcriptase and intron Evaluation of pt gene order revealed extreme conservation across maturase (Appendix S6). Other ORFs shared similarity between the family, with only one or two rearrangements between most taxa two to four taxa, with most (13) putatively homologous ORFs be- and Pseudopediastrum integrum and Stauridium tetras, and three ing shared by the sister taxa P. duplex AL0403MN and P. duplex rearrangements separating the two taxa (Fig. 6). Stauridium tetras SL0404MN. and P. integrum differ from the rest of the family within a 53,500-nt Of the 116–149 intergenic spacer regions annotated in five closely region in the smaller SC with regard to the position of several gene related taxa making up the Hydrodictyon-Pediastrum group, only five sets: tufA-trnV-trnQ-rpoC1, rps4-trnL-clpP-rps18-petB-trnI, trnG- were alignable and ranged from 85 to 539 nucleotides in length. Too atpA-psbI-cemA-rpl12-trnF-trnE, and psaB-ccsA-psbZ-psbM-trnR- little information was present to be used for inferring phylogeny. trnP-atpB (Fig. 7).

FIGURE 4. Number of freestanding ORFs and introns in each taxon sampled for this study. Blue bars represent ORFs; green bars represent introns occurring within genes. Hydrodictyon reticulatum contains the most ORFs in the family, Stauridium tetras the fewest. 322 • American Journal of Botany

trnL atpB petD psaA psaB psbA psbB psbC psbD psbZ rbcL rrlrrs (uaa) TR 89, 417 Pediastrum duplex AL0403MN 2795 269 504* 2651 34 TR 89, Pediastrum duplex SL0404MN 1173 645 269 34 2123 Pediastrum duplex KY114064 522 TR 89, 1772 1172 1173 269 2637 2646 34 Pediastrum duplex UTEX 1364 522 TR 89, 269, 34 C1797 Pediastrum angulosum ACOI 1354 TR 89, 1260 2749 269 1772 276 582* 2893 34

TR 89, 555* 2217 2740 Pediastrum duplex PL0501b 900 882 582* 34 269 744 2884 TR 89, 269, Pediastrum angulosum KP0301SC 522 34 C1797 TR 89, 414 Hydrodictyon reticulatum KY114065 813 1098* 34 269 553* TR 89, 2019 34 Pseudopediastrum boryanum IL0402MN 269 2860 TR 89, 269 Pseudopediastrum boryanum ML0410MN 1173 2030 485 34 C1797 Pseudopediastrum integrum ACOI 577 TR 89, 93* 534 269 34 TR 89, 744 65* 2607 489 34 Pseudopediastrum sp. CL0201VA 105* 269 TR 89, 549 Lacunastrum gracillimum ACOI 392 1772 900 699 34 269 744 TR 89, 462 Stauridium tetras ACOI 84 1173 522* 2712 506 34 269 699

G I G II HNH GIY-YIGLAGLIDADG FIGURE 5. Distribution of introns in the plastid genes of Hydrodictyaceae examined in this study. Numbers indicate insertion site within genes, and colors indicate type of intron and homing endonuclease, if present (color code located at bottom of the figure). GIY-YIG = GIY-YIG homing endonuclease, LAGLIDADG = LAGLIDAD G homing endonuclease, HNH = H-N-H homing endonuclease. Introns with an asterisk (*) are unique to the family Hydrodictyaceae.

Relationship between genetic divergence and genome Plastid concatenated—The topology was identical for the concat- rearrangements enated nucleotide analyses performed by Garli and MrBayes. All five Garli ML trees had identical topology and nearly identical Pairwise evolutionary distance, a proxy for time since divergence branch lengths, and this was the case for other types of analyses of among taxa, is correlated with the number of rearrangements across the concatenated data sets. Support was absolute for all nodes in the order (Fig. 8). Comparisons of taxa within Hydrodictyaceae, the Bayesian analysis and for all but one node (Chlorotetraëdron + representing more shallowly diverging taxa, are distinct from com- Hydrodictyaceae, BS = 92) in the ML bootstrap analysis (Appendix parisons across families in Sphaeropleales, representing taxa sep- S7). A near-identical result was obtained from the amino acid arated by larger divergences. The only within-genus comparison analyses. previously reported in Sphaeropleales (that of two Bracteacoccus Of the relationships of special interest, Pediastrum duplex and P. species) is characterized by few rearrangements (Fučíková et al., angulosum strains grouped together, with Hydrodictyon as their sis- 2016a), similar to what is seen within Hydrodictyaceae. ter taxon. The genus Pseudopediastrum was also resolved as strongly monophyletic, with Pseudopediastrum sp. being sister to the rest of Mitochondrial genes this clade. Lacunastrum gracillimum was sister to the Pediastrum, Hydrodictyon, and Pseudopediastrum clade, and Stauridium tet- Thirteen mt genes were extracted from assembled fragments of the ras was sister to the rest of the Hydrodictyaceae. Chlorotetraëdron mt genomes: atp6, atp9, cob, cox1, cox2, cox3, nad1, nad2, nad3, was recovered sister to Hydrodictyaceae in the plastome data set nad4, nad4L, nad5, nad6. Zero to three introns were identified in (Appendix S7). the cox1 gene. All but one of the cox1 introns was Group IB, two of which contained a LAGLIDADG homing endonuclease. The cox1 intron in Pediastrum angulosum KP0301SC was Group II. Introns Mitochondrial concatenated—Phylogenetic results of the mt were not identified in the remaining mt genes. analyses under Garli and MrBayes predominantly agreed with those from the pt data analyses. One notable exception was the Phylogenetic analyses placement of Neochloris sister to Hydrodictyaceae rather than Chlorotetraëdron. Most nodes received high ML bootstrap values, Figure 8 presents a summary phylogeny that illustrates the overall and all nodes received strong BI support values (Appendix S8). relationships supported by the pt and mt concatenated gene analy- Unlike in the pt analyses, the amino acid and nucleotide analysis ses, where there are conflicting results, and which lineages comprise results differed for the mt data, both within the Pediastrum duplex Pediastrum Group I and Group II. clade and in the relative placement of Pseudopediastrum sp. In the aa March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 323

S. t

ACOI 84

ex L. gracillimum etras

ACOI 392 dupl

P. KY114064

ex dupl Ps. boryanum P. SL0404MN ML0410MN

x P. duple N AL0403M Ps. boryanum IL0402MN

P. angulosum KP0301SC

tegrum Ps. in 7 ACOI57

P. angulosum ACOI 1354

A . sp. Ps

CL0201V P. duple

PL0501 1

5 x 2 3

KY11406

H. reculatum FIGURE 6. Summary of gene rearrangements between pairs of taxa. Each colored block corresponds to a plastome. Lines between blocks represent the total number of gene rearrangements between plastomes; the thicker lines represent a greater number of rearrangements. analysis, Pseudopediastrum sp. was sister to the rest of Pediastrum, of Neochloris, Chlorotetraëdron, and Hydrodictyaceae (CN, CH, Hydrodictyon, and Pseudopediastrum strains, whereas it grouped and NH) received support from single-gene analyses in the pt and with other Pseudopediastrum taxa in the nucleotide analysis (and the mt genes (Appendix S9). The results of pt gene Bayesian analy- in the pt data analyses). ses show 24 of 65 (~37%) pt genes supported CH, with 12 of these (~18%) with ≥0.90 BPP. The results of mt gene Bayesian analyses Single-gene analyses—Support for three main topologies, showed three of 13 (~23%) genes supporting CH and two (~15%) Chlorotetraëdron + Neochloris, or monophyletic Neochloridaceae supporting NH with ≥0.90 BPP, with four (~30%) genes showing (CN), Chlorotetraëdron + Hydrodictyaceae (CH), Neochloris + conflict among the three types (nt, aa, codon) of analyses. Garli re- Hydrodictyaceae (NH) was tested with three single-gene analysis sults (not shown) generally agreed with Bayesian results; best and types: nt, aa, and codon. Overall, results varied in nodal support bootstrap tree files are included in the supplements, as well as the and often in topology. In particular, all three possible arrangements support summaries in Appendix S10. 324 • American Journal of Botany

LSC IR SSC

region compared is 53500 nt rpl2 3 rpl2 rpl1 9 rps1 4 tufA trnV trnQ rpoC 1

Stauridium tetras (ACOI 84) trnI trnL rps4 trnE trnF rpl1 2 psbI atpA trnG pet B rps1 8 clpP cemA atpB trnP trnR psbM psbZ ccsA psaB trnE trnW rpoA trnM e rpl2 3 rpl2 rpl1 9 rps1 4 trnG atpA psbI rpl1 2 trnF trnE psbZ cemA psaB tufA trnV ccsA trnR trnP trnQ psbM atpB rpoC 1

Pseudopediastrum boryanum (ML0410MN) trnI pet B trnL rps1 8 clpP rps4 trnE trnW rpoA trnM e Most Hydrodictyaceae clpP rps1 8 pet B rpl2 3 rpl2 rpl1 9 rps1 4 trnG atpA psbI rpl1 2 trnF trnE rps4 trnL trnI cemA e Pseudopediastrum integrum (ACOI577) rpoC1 trnQ trnV tufA atpB trnP trnR psbM psbZ ccsA psaB trnE trnW rpoA trnM

FIGURE 7. Summary of inferred plastid genome rearrangements between Stauridium tetras, Pediastrum integrum, and the remaining Hydrodictyaceae taxa included in this study (represented by Pseudopediastrum boryanum). Rearrangements involve the small single copy region (SSC) and two blocks of genes.

12 Sphaeropleales Hydrodictyaceae 10

4 wise number of rearrangements air P Bracteacoccus minor v. Bracteacoccus aerius 2

0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 Pairwise sequence divergence of coding genes FIGURE 8. Comparison of number of gene rearrangements vs. sequence divergence across the order Sphaeropleales (purple circles) and within Hydrodictyaceae (orange squares). Note Hydrodictyaceae represents similar conservation as presented between two species of Bracteacoccus, with greater variation between taxa across Sphareopleales. March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 325

The relationships among Hydrodictyon and Pediastrum strains synteny and a low number of plastome rearrangements compared to were investigated using a similar approach testing the following the exceptionally high numbers of rearrangements seen across other three topologies, monophyletic Pediastrum duplex, Hydrodictyon members of Sphaeropleales. Instead of extensive gene order changes, + P. duplex Group I, and Hydrodictyon + P. duplex Group II. plastome expansion from introns, ORFs, and other nonstandard While the concatenated analyses decisively recovered Pediastrum genic regions is a dominant form of evolution in Hydrodictyaceae. as monophyletic with Hydrodictyon sister, there was some dis- Extra, repetitive DNA is thought to facilitate rearrangements, serv- cordance among single-gene analyses (Bayesian results shown ing as hotspots of recombination (e.g., Bondreau and Turmel, 1996; in Appendices S10 and S11). Fifty pt genes (~77%) and seven mt Weng et al., 2013). If the expansion of repetitive DNA in the plas- genes (~54%) did not yield strong support for any of the three tome within Hydrodictyaceae ultimately leads to the accumulation tested topologies across the analysis types (nt, aa, codon). Ten pt of unique rearrangements over evolutionary time, and this process genes (~15%) and five mt genes (~38%) supported monophyletic is ongoing and independently giving rise to unique rearrangements Pediastrum with Hydrodictyon as its sister taxon. The two other within other families in the order, then the number of rearrangements topologies, Hydrodictyon + P. duplex Group I and Hydrodictyon + observed when comparing two members of Sphaeropleales from dif- P. duplex Group II, received phylogenetic support from three pt and ferent families is expected to be higher than within families. In this one pt genes, respectively (Appendix S11A), and varying but rela- study, pairwise evolutionary distance was generally correlated with tively low gene (Appendix S11B) and sum BPP (Appendix S11C) the number of rearrangements across the order (Fig. 8), but within- support in the nt and aa analyses. No mt genes supported either of family comparisons were primarily limited to Hydrodictyaceae. these two topologies across analysis types. To determine whether this pattern is supported more generally in Thirteen pt genes were initially found to offer signal to resolve rela- Sphaeropleales, we clearly need to analyze plastome data of many tionships between Hydrodictyon and Pediastrum (Appendix S11A), more species within other sphaeroplealean families. but by allowing the sampling of polytomous trees, additional Phycas analyses narrowed the number of “decisive” pt genes that strongly Introns—The trans-spliced psaA gene within the family shares the supported (using 0.9 BPP as the cutoff) a monophyletic Pediastrum same splice sites 89 and 269 with each other and other Sphaeropleales, with Hydrodictyon as its sister taxon to five (rpoBa, rpoC1, rps2, with the exception of Acutodesmus obliquus (Turpin) Hegewald & ycf1, tufA). Of the pt genes that supported the Hydrodictyon + P. Hanagata 2000, which only has two trans-spliced exons, separated duplex Group I and Hydrodictyon + P. duplex Group II topologies, at site 269. Two of the taxa, Pediastrum angulosum KP0301SC and only one (psaA, Hydrodictyon + P. duplex Group I) was strongly Pseudopediastrum boryanum ML0410MN, have an additional cis- supported. The five mt genes showing higher than 0.9 BPP support spliced intron at site 1797, similar to Ourococcus multisporus and for monophyletic Pediastrum when sampling of polytomy trees was P. duplex UTEX 1364 (Fučíková et al., 2016b). Interestingly, the allowed were cox3, nad1, nad2, nad5, and nad6, while cox1 resolved P. duplex strains included in the present study that are closely re- Hydrodictyon sister to Group II. SVDquartet analyses recovered a lated to UTEX 1364 do not contain the cis-spliced exon. Most of phylogeny that placed Chlorotetraëdron as sister to Hydrodictyaceae the introns in Hydrodictyaceae share insertion sites with other taxa and resolved Hydrodictyon as sister to monophyletic Pediastrum. in Sphaeropleales or are not unique introns in the particular gene. Nine introns appear to be completely unique to the family, some of which are restricted to a single strain (Fig. 5), indicating recent DISCUSSION independent invasions into the plastomes. In our study, rpoBb sequence expansion was found in all three How do plastomes evolve in Hydrodictyaceae compared members of the P. angulosum clade. This expansion roughly begins to the rest of Sphaeropleales? around site 1000 (variable due to the variable nature of the gene itself) and does not interrupt the reading frame. This expansion appears Plastome size, structure, and content—The current sampling of relatively conserved in sequence across the three Pediastrum strains, plastomes in the Hydrodictyaceae has permitted a detailed investi- and BLAST reveals similar sequences in the rpoBb gene of the volvo- gation of how plastomes vary among closely related taxa. The family caleans Lobochlamys segnis, Pleodorina starrii, and Oogamochlamys contains the largest plastomes of the Sphaeropleales sequenced thus gigantea, all >39% coverage and >70% sequence similarity and no far, represented by Pediastrum duplex at 232,554 nt (McManus et al., interruption of the rpoBb reading frame. Additional chlorophycean 2017). The second smallest plastome of the order is represented in this hits exhibit lower BLAST scores. So far only one chlorophycean taxon study by Stauridium tetras at 102,718 nt. Within the family, from 30 is known to have a true intron in rpoBb—the recently characterized to 46.9% of the plastome is noncoding sequence (Table 1). The plas- Koshicola spirodelophila (Watanabe et al., 2016). The intron in this tome size variation across the family is largely attributed to a trend in species is inserted at site 1905, interrupts the rpoBb reading frame, expansion of intergenic spacer regions and the acquisition of novel and has a short but recognizable Group II intron-derived element. ORFs (Table 1, Fig. 4), is similar to what is seen across Sphaeropleales The evolution of the rpoB gene was discussed in Novis et al. (2013), (Lemieux et al., 2015; Fučíková et al., 2016a, 2016b). The IRs in largely with respect to the expansions of coding sequence thought to Hydrodictyaceae range in size from 10,391 bp to 19,830 bp and har- have led to the fragmentation of rpoB into rpoBa and rpoBb. bor the atpH gene, similar to two related genera, Chlorotetraëdron The relative paucity of introns represented in the rbcL gene in both and Neochloris. The conclusions of Fučíková et al. (2016a) suggesting this study and Fučíková et al. (2016b) is not due to the lack of introns a dynamic evolution of IRs across the Sphaeropleales (6,375–35,503 in that gene, but instead to taxon sampling. McManus et al. (2012) bp), also due to the expansion of spacer regions and introduction of discovered two Group I introns of distinct origins, each inserted at introns, are supported in this study. a specificrbcL site, 462 or 699, in multiple hydrodictyaceaen taxa Despite the dramatic range in plastome size seen among the mem- and within the genus Bracteacoccus (only the 699 intron). The hyd- bers of Hydrodictyaceae compared in this study, there is surprising rodictyaceaen taxa typically contained only one of the introns, with 326 • American Journal of Botany

Stauridium tetras the only taxon known to harbor both. McManus of discordant phylogenetic signal among genes, possibly due to hori- et al. (2012) found the rbcL 462 introns in Hydrodictyaceae form a zontal gene transfer, incomplete lineage sorting, ancient paralogy, or close relationship with rbcL 462 introns in the Volvocales, and dis- recombination (e.g., Birky, 2001; Rice and Palmer, 2006; Rice et al., tribution of the intron indicates horizontal transfer between the lin- 2013; Sullivan et al., 2017). Further discrepancy in signal from in- eages. The 699 introns are closely related to rRNA L2449 introns in dependent organelles can arise because each organelle is inherited the ulvophytes, and distribution supports a hypothesis of horizontal independently. Such is the case when mating experiments between transfer between the lineages (McManus et al., 2012). Chlamydomonas smithii R.W.Howshaw & H.Ettl 1966 and C. reinhardtii P.A.Dangeard 1888 (Chlamydomonadales) were performed, ORFs—The comparison of freestanding ORFs within Hydrodictyaceae and the chloroplast was inherited maternally while the mitochondrion highlights the extensive invasion, spread, and subsequent degrada- was inherited paternally, and in some cases, a biparental chloroplast tion/loss of ORFs. Closely related taxa share the highest number of was formed, resulting in opportunities for recombination and segre- similar ORFs, and relatively few that were ancestrally inherited appear gation (Kuroiwa et al., 1982; Boynton et al., 1987; Greiner et al., 2014). to have persisted. McManus et al. (2017) reported an ORF that con- Additionally, transmission of a segment of C. smithii mt genome was tained a conserved domain typical of a virus replication-associated spread to the mt genome of C. reinhardtii (Boynton et al., 1987). protein (circular DNA virus-8) isolated from a freshwater pond in Studies of the reproductive biology and organelle inheritance patterns Antarctica. The absence of additional occurrences in the family of are necessary in the Neochloridaceae to test the hypothesis of relaxed similar ORFs supports a hypothesis that it originated from a relatively organelle inheritance and/or formation of biparental organelles. recent invasion event. The distribution of ORFs in the hydrodictyaceaen plastomes supports two hypotheses: (1) ancestral inheritance Phylogeny of Hydrodictyaceae followed by degradation and multiple losses across the family, and (2) recent, independent acquisitions in each lineage, highlighting the Is Pediastrum monophyletic?—Concatenated analyses of protein- susceptibility of these plastomes to invasion by novel elements. coding genes from both organellar genomes strongly resolve Hydrodictyon as sister to P. duplex + P. angulosum, thereby contra- What is the sister lineage to Hydrodictyaceae? dicting previous one- and two-gene analyses of McManus and Lewis (2011). Most of the individual genes analyzed under Bayesian infer- A phylogenomic study of plastomes sampled across the ence (nt, aa, and codon models) do not contain substantial signal Sphaeropleales (Fučíková et al., 2016a) found conflicting phyloge- to resolve the relationships between Hydrodictyon and Pediastrum, netic signal between the nucleotide and amino acid analyses when but some pt genes (rpoBa, rpoC1, rps2, ycf1, tufA, psaA) and mt examining the relationship of Hydrodictyaceae to related families. genes (cox3, nad1, nad2, nad5, nad6, cox1) that do could be con- The nucleotide analysis in that study recovered Chlorotetraëdron as sidered for future studies within that group, which should include sister to Hydrodictyaceae (represented by one strain of Pediastrum additional strains covering more of hydrodictyacean phylo- and duplex), while the amino acid analysis recovered Chlorotetraëdron morphodiversity. Even with conflict found in one pt gene (psaA) + Neochloris in that position. Given only one representative of and one mt gene (cox1), among-gene conflict does not appear to be Hydrodictyaceae was included, it was proposed that increas- due to deep coalescence, as SVD quartets recovered the same tree ing the taxon sampling may resolve the conflict. The increased found with noncoalescent methods. sampling of the Hydrodictyaceae in this study did not unequiv- Among-gene conflict in phylogenetic signal could have a num- ocally resolve the CN/CH/NH conflict between types of analy- ber of causes. Lateral gene transfer is an often-dismissed one, par- ses as anticipated by Fučíková et al. (2016a). A slight majority of ticularly in plastids, in which it is extremely rare, or at least rarely plastid genes (24/65), analyzed three ways (nt, codon, aa), prefers documented (Rice and Palmer, 2006). In land plants, single-locus Chlorotetraëdron as sister to Hydrodictyaceae (Appendix S9), with inheritance of plastomes is assumed based on large amounts of ge- 12 of those genes showing high support (≥0.90 BPP). In contrast, nome data, but applying the same assumption of plastome stability the nt and aa mitochondrial analyses favored Neochloris as the sis- to green algae may not be justified. For instance, the plastomes of ter lineage to Hydrodictyaceae, while the codon analyses favored Hydrodictyaceae and other Sphaeropleales are laden with tandem Chlorotetraëdron as sister to Hydrodictyaceae (Appendix S9). The repeats, which are thought to be associated with the potential for conflict between the pt and mt data could be due to challenges in within-genome recombination (e.g., Odahara et al., 2015). Intron analyzing data with complex molecular evolution. It is noteworthy movement is also rampant, and therefore lateral gene movement that organellar genomes of Neochloris exhibit some unusual fea- may be possible, though not demonstrable with the current data. tures, including a loss of the chlB, chlL, and chlN from the plastome, Our study took the approach of probing single genes for their phy- and an apparently convergent juxtaposition of rrs fragments in the logenetic signal, because single genes are still commonly used to mitochondrial genome (Fučíková et al., 2014b). Such features do draw taxonomic conclusions in algae. No matter the source of con- not necessarily imply horizontal gene transfer, but may point to un- flicting signal, if some genes tell a different phylogenetic story than usual events in the evolution of the organellar genomes, some of others, gene choice will have an effect on taxonomic conclusions— which might also affect phylogenetic signal. Additionally, the diver- in our case, selecting psaA vs. tufA for phylogenetic inference sity of Neochloridaceae is represented by only two taxa, and a fo- would decide whether Pediastrum remains a monophyletic genus cus on the phylogenetically critically positioned genus Tetraëdron or whether erecting a second Pediastrum-like taxon is warranted. Kützing 1845, which is also closely related to Hydrodictyaceae and Our phylogeny (Fig. 9) supports taxonomic revisions made historically placed within the family, will be important to further by Buchheim et al. (2005) that retained the genera Hydrodictyon test the placement of Chlorotetraëdron and Neochloris. and Pediastrum. However, the sister relationship of the net-like There is a broad assumption that genes of organellar genomes be- Hydrodictyon with the two-dimensional Pediastrum depicts an inter- have as one unit, yet many investigations have demonstrated evidence esting scenario of morphological evolution. Hydrodictyon colonies March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 327

Hydrodictyaceae Pediastrum duplex AL0403MN I Pediastrum duplex SL0404MN P ediastrum Pediastrum duplex KY114064

Pediastrum duplex UTEX 1364

Pediastrum angulosum ACOI 1354 II

Pediastrum duplex PL0501b

Pediastrum angulosum KP0301SC

Hydrodictyon reticulatum KY114065

Pseudopediastrum boryanum IL0402MN pediastrum Pseudo- Pseudopediastrum boryanum ML0410MN

Pseudopediastrum integrum ACOI 577

Pseudopediastrum sp. CL0201VA

Lacunastrum gracillimum ACOI 392

Stauridium tetras ACOI 84

Chlorotetraedron incus SAG 43.81 Neochloridaceae Neochloris aquatica UTEX 138

Acutodesmus obliquus UTEX 393 Scenedesmaceae Hariotina reticulata UTEX 1365

FIGURE 9. Summary phylogeny of concatenated plastid genes. Stars indicate nodes that differ in the resulting mitochondria phylogeny. Scendesmaceae included as outgroup taxa, I = Group I Pediastrum, II = Group II Pediastrum, as described in McManus and Lewis (2011) and McManus et al. (2011).

could result from an increase in size of the individual cells and inter- among P. angulosum isolates, as found in previous analyses cellular spaces of a two-dimensional Pediastrum-like colony, likely (McManus and Lewis, 2011). The morphological variability is not beginning at the zoospore stage. Each zoospore would need to con- reflected at the molecular level, at least for genes thus far sequenced, nect with two to three other zoospores at each end, then, after loss and could be indicative of incomplete lineage sorting due to a re- of flagella, extreme expansion of the cells would create large intercel- cent and rapid speciation event (Wendel and Doyle, 1998; Funk lular spaces and lengthened individual cells, resulting in the mesh and Omland, 2003; Alverson, 2008). Another possibility is hybrid- form. Marchant and Pickett-Heaps (1972) include micrographs of ization and biparental inheritance of organelles, as demonstrated a young Hydrodictyon coenobium with marginal cells similar to in strains of Chlamydomonas (Kuroiwa et al., 1982; Boynton et al., lengthened Pediastrum marginal cells, with two small processes 1987; Greiner et al., 2014). A recent study of Late Glacial (~10–13 on either end of the cell. The processes are lost during the growth ka) fossil Pediastrum indicated difficulty assigning species name to of the Hydrodictyon colony, so are typically not observed in adult some samples that exhibited intermediate characteristics between stages. Environmental factors such as water depth, temperature, P. angulosum and P. duplex (Turner et al., 2014), indicating poten- pH, conductivity, or nutrient levels could be driving factors in the tial for discovery of extant intermediate forms. morphological shift, and worthwhile exploring further. Additionally, characterizing which gene or suite of genes is responsible for such a morphological shift would shed light on the evolution of the genus CONCLUSIONS and the genetic underpinnings of such distinct colony forms. Genomic studies across the green algae are uncovering new lineages, Are Pediastrum species phylogenetically distinct?—Despite mor- highlighting a wide diversity of genomes in size and architecture, phological differences between Pediastrum duplex and P. angu- and phylogenomic analyses are aiding in the resolution of relation- losum, all analyses recover Pediastrum duplex (PL0501b) nested ships not supported by previous molecular phylogenetic studies (e.g., 328 • American Journal of Botany

Fučíková et al., 2014a; Leliaert et al., 2016). Most organellar genome Alverson, A. J. 2008. Molecular systematics and the diatom species. Protist 159: studies of the Chlorophyceae have included relatively few represent- 339–353. atives of each major lineage, and none have presented dense within- Andre, C., A. Levy, and V. Walbot. 1992. Small repeated sequences and the struc- family sampling. The order Sphaeropleales has recently received ture of plant mitochondrial genomes. Trends in Genetics 8: 128–132. Archibald, J. M., and T. A. Richards. 2010. Gene transfer: anything goes in plant more attention, with an increased number of molecular markers and mitochondria. BMC Biology 8: 147. organellar genomes analyzed (Fučíková et al., 2014b, 2014c, 2016a, Benson, G. 1999. Tandem repeats finder: a program to analyze DNA sequences. 2016b; Farwagi et al., 2015; McManus et al., 2017). Nucleic Acids Research 27: 573–580. This study presents a broad sampling of plastomes and frag- Birky, C. W. 2001. The inheritance of genes in mitochondria and chloroplasts: mentary mitochondriomes within a family of Sphaeropleales. The laws, mechanisms, and models. Annual Review of Genetics 35: 125–148. results highlight the variation found between a number of closely Bold, H. C., and M. Wynne. 1985. Introduction to the algae. Prentice Hall, related taxa. The overall architecture of the plastomes is conserved, Englewood Cliffs, NJ, USA. with size differences caused by nonstandard genic regions, introns, Bondreau, E., and M. Turmel. 1996. Gene rearrangements in Chlamydomonas and freestanding ORFs, and some diversity in mitochondriome gene chloroplast DNAs are accounted for by inversions and by the expansion/con- order was demonstrated by Farwagi et al. (2015). Conservation of traction of the inverted repeat. Plant Molecular Biology 27: 351. Boynton, J. E., E. H. Harris, B. D. Burkhart, P. M. Lamerson, and N. W. Gillham. plastomes in Hydrodictyaceae might indicate that with increased 1987. Transmission of mitochondrial and chloroplast genomes in crosses of within-family sampling across the Sphaeropleales, we may see trends Chlamydomonas. Proceedings of the National Academy of Sciences, USA 84: toward more gradual shifts in plastome architecture, and an ability to 2391–2395. study patterns of genome evolution of the order. The additional taxon Brouard, J. S., C. Otis, C. Lemieux, and M. Turmel. 2010. The exceptionally large sampling will also allow us to capture a more complete evolutionary chloroplast genome of the green alga Floydiella terrestris illuminates the evo- history of intron gain and loss, and test hypotheses of the observed lutionary history of the Chlorophyceae. Genome Biology and Evolution 12: freestanding ORF distribution and evolution in the plastomes. 240–256. The phylogenomic approach has offered resolution in the place- Buchheim, M., J. Buchheim, T. Carlson, A. Braband, D. Hepperle, L. Krienitz, ment of Hydrodictyon as sister to a monophyletic Pediastrum. M. Wolf, and E. Hegwald. 2005. Phylogeny of the Hydrodictyaceae However, despite their morphological distinctness, P. angulosum (Chlorophyceae): inferences from rDNA data. Journal of Phycology 41: 1039–1054. remains paraphyletic to P. duplex. Is there a switch between the Chifman, J., and L. Kubatko. 2014. Quartet inference from SNP data under the two phenotypes that is easily triggered by environmental cues? Do coalescent. Bioinformatics 30: 3317–3324. they readily hybridize? Further analyses with additional taxa as well Chifman, J., and L. Kubatko. 2015. Identifiability of the unrooted species tree as studies that focus on hybridization potential and the stability topology under the coalescent model with time-reversible substitution pro- of phenotypic traits under varying environmental conditions are cesses, site-specific rate variation, and invariable sites. Journal of Theoretical needed to gain a better understanding of the relationship between Biology 374: 35–47. P. angulosum and P. duplex. Darling, A. E., B. Mau, and N. T. Perna. 2010. progressiveMauve: Multiple genome alignment with gene gain, loss and rearrangement. PLoS One 5: e11147. https://doi.org/10.1371/journal.pone.0011147. ACKNOWLEDGEMENTS Doyle, J. J., and J. L. Doyle. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. Fan, Y., R. Wu, M.-H. Chen, L. Kuo, and P. O. Lewis. 2011. Choosing among par- The authors thank the Special Issue Editor, Pamela Diggle, and two tition models in Bayesian phylogenetics. Molecular Biology and Evolution reviewers for their valuable comments and edits on this manuscript. 28: 523–532. We thank Stéphane L. Marty for assistance with the figures and il- Farwagi, A., K. Fučíková, and H. A. McManus. 2015. Phylogenetic patterns of lustrations and Alexander Weir at SUNY-ESF for use of his Nikon gene rearrangements in four mitochondrial genomes from the green algal E800 microscope. Analyses were carried out at the Computational family Hydrodictyaceae (Sphaeropleales, Chlorophyceae). BMC Genomics Biology Core Facility of the University of Connecticut. This ma- 16: 826. terial is based upon work supported by the National Science Fučíková, K., F. Leliaert, E. D. Cooper, P. Skaloud, S. D’Hondt, O. De Clerck, Foundation under Grant No. DEB-1036466, DEB-1036448, DEB- C. F. D. Gurgel, et al. 2014a. New phylogenetic hypotheses for the core Chlorophyta based on chloroplast sequence data. Frontiers in Ecology and 1354146, DEB-1213675 and DEB-1519316, Le Moyne College Evolution 2: 63. Faculty Research & Development grant, and the Le Moyne College Fučíková, K., P. O. Lewis, D. González-Halphen, and L. A. Lewis. 2014b. Gene Office of the Provost. arrangement convergence, diverse intron content, and genetic code modifi- cations in mitochondrial genomes of Sphaeropleales (Chlorophyta). Genome Biology and Evolution 6: 2170–2180. SUPPORTING INFORMATION Fučíková, K., P. O. Lewis, and L. A. Lewis. 2014c. Putting incertae sedis taxa in their place: a proposal for ten new families and three new genera in Sphaeropleales Additional Supporting Information may be found online in the (Chlorophyceae, Chlorophyta). Journal of Phycology 50: 14–25. supporting information tab for this article. Fučíková, K., P. O. Lewis, and L. A. Lewis. 2016a. Chloroplast phylogenomic data from the green algal order Sphaeropleales (Chlorophyceae, Chlorophyta) re- veal complex patterns of sequence evolution. Molecular Phylogenetics and Evolution 98: 176–183. Fučíková, K., P. O. Lewis, and L. A. Lewis. 2016b. Comparative analyses of LITERATURE CITED chloroplast genome data representing nine green algae in Sphaeropleales (Chlorophyceae, Chlorophyta). Data in Brief 7: 558–570. Addou, S., R. Rentzsch, D. Lee, and C. A. Orengo. 2009. Domain-based and Funk, D. J., and K. E. Omland. 2003. Species-level paraphyly and polyphyly: fre- family-specific sequence identity thresholds increase the levels of reliable quency, causes, and consequences, with insights from animal mitochondrial protein function transfer. Journal of Molecular Biology 387: 416–430. DNA. Annual Review of Ecology, Evolution, and Systematics 34: 397–423. March 2018, Volume 105 • McManus et al.—Hydrodictyaceae phylogenomics • 329

Greiner, S., J. Sobanski, and R. Bock. 2014. Why are most organelle genomes Milecka, K., and K. Szeroczyńska. 2005. Changes in macrophytic flora and transmitted maternally? Bioessays 37: 80–94. planktonic organisms in lake Ostrowite, Poland, as a response to climatic Guo, X., J. Liu, G. Hao, L. Zhang, K. Mao, X. Wang, D. Zhang, T. Ma, Q. Hu, I. A. and trophic fluctuations. Holocene 15: 74–84. Al-Shehbaz, and M. A. Koch. 2017. Plastome phylogeny and early diversifi- Novis, P. M., R. Smissen, T. R. Buckley, K. Gopalakrishnan, and G. Visnovsky. cation of Brassicaceae. BMC Genomics 18: 176. 2013. Inclusion of chloroplast genes that have undergone expansion mis- Henriquez, C. L., T. Arias, J. C. Pires, T. B. Croat, and B. A. Schaal. 2014. leads phylogenetic reconstruction in the Chlorophyta. American Journal of Phylogenomics of the plant family Araceae. Molecular Phylogenetics and Botany 100: 2194–2209. Evolution 75: 91–102. Odahara, M., Y. Masuda, M. Sato, M. Wakazaki, C. Harada, K. Toyooka, and Y. Jankovská, V., and J. Pokorný. 2002. Palaeoecology of a medieval fishpond sys- Sekine. 2015. RECG maintains plastid and mitochondrial genome stability tem (Vajgar, Czech Republic). Folia Geobotanica 37: 253–273. by suppressing extensive recombination between short dispersed repeats. Jena, M., C. Bock, C. Behera, S. P. Adhikary, and L. Krienitz. 2014. Stain sur- PLoS Genetics 11: e1005080. vey on three continents confirms the polyphyly of the genus Pediastrum Rice, D. W., and J. D. Palmer. 2006. An exceptional horizontal gene transfer in (Hydrodictyaceae, Chlorophyceae). Fottea, Olomouc 14: 63–76. plastids: gene replacement by a distant bacterial paralog and evidence that Komárek, J., and V. Jankovská. 2001. Review of the green algal genus Pediastrum; haptophyte and cryptophyte plastids are sisters. BMC Biology 4: 31. implication for pollen-analytical research. Bibliotheca Phycologica 108: Rice, D. W., A. J. Alverson, A. O. Richardson, G. J. Young, M. V. Sanchez-Puerta, 1–125. J. Munzinger, K. Barry, et al. 2013. Horizontal transfer of entire genomes via Kuroiwa, T., S. Kawano, S. Nishibayashi, and C. Sato. 1982. Epifluorescent mi- mitochondrial fusion in the angiosperm Amborella. Science 324: 1468–1473. croscopic evidence for maternal inheritance of chloroplast DNA. Nature Ronquist, F., M. Teslenko, P. van der Mark, D. L. Ayres, A. Darling, S. Höhna, 298: 481–483. B. Larget, et al. 2012. MrBayes 3.2: efficient Bayesian phylogenetic infer- Lang, B. F., M.-J. Laforest, and G. Burger. 2007. Mitochondrial introns: a critical ence and model choice across a large model space. Systematic Biology 61: view. Trends in Genetics 23: 119–125. 539–542. Leliaert, F., A. Tronholm, C. Lemieux, M. Turmel, M. S. DePriest, D. Smith, D. R., T. Hamaji, B. J. Olson, P. M. Durand, P. Ferris, R. E. Michod, J. Bhattacharya, K. G. Karol, et al. 2016. Chloroplast phylogenomic analyses re- Featherston, H. Nozaki, and P. J. Keeling. 2013. Organelle genome complex- veal the deepest-branching lineage of the Chlorophyta, Palmophyllophyceae ity scales positively with organism size in volvocine green algae. Molecular class. nov. Scientific Report 6: 253–267. Biology and Evolution 30: 793–797. Lemieux, C., A. T. Vincent, A. Labarre, C. Otis, and M. Turmel. 2015. Chloroplast Sukumaran, J., and M. T. Holder. 2010. DendroPy: a Python library for phyloge- phylogenomic analysis of chlorophyte green algae identifies a novel lineage sis- netic computing. Bioinformatics 26: 1569–1571. ter to the Sphaeropleales (Chlorophyceae). BMC Evolutionary Biology 15: 264. Sullivan, A. R., B. Schiffthaler, S. L. Thompson, N. R. Street, and X. R. Wang. Lenarczyk, J., and H. A. McManus. 2016. Testing the boundaries of the green 2017. Interspecific plastome recombination reflects ancient reticulate evo- algal species Pediastrum alternans (Chlorophyceae) using conventional, ge- lution in Picea (Pinaceae). Molecular Biology and Evolution 34: 1689–1701. ometric morphometric and phylogenetic methods. Phycologia 55: 515–530. Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other Lewis, P. O., M. T. Holder, and K. E. Holsinger. 2005. Polytomies and Bayesian methods), version 4.0 build 152 (2017). Sinauer, Sunderland, MA, USA. phylogenetic inference. Systematic Biology 54: 241–253. Turmel, M., C. Otis, and C. Lemieux. 2015. Dynamic evolution of the chloroplast Lewis, P. O., M. T. Holder, and D. L. Swofford. 2015. Phycas: software for Bayesian genome in the green algal classes Pedinophyceae and Trebouxiophyceae. phylogenetic analysis. Systematic Biology 64: 525–531. Genome Biology and Evolution 7: 2062–2082. Lezine, A. M., J. C. Duplessy, and J. P. Cazet. 2005. West African monsoon var- Turner, F., R. Pott, A. Schwarz, and A. Schwalb. 2014. Response of Pediastrum iability during the last deglaciation and the Holocene: evidence from fresh in German floodplain lakes to Late Glacial climate changes. Journal of water algae, pollen and isotope data from core KW31, Gulf of Guinea. Paleolimnology 52: 293–310. Palaeogeography, Palaeoclimatology, Palaeoecology 219: 225–237. Watanabe, S., K. Fučíková, P. O. Lewis, and L. A. Lewis. 2016. Koshicola Lohse, M., O. Drechsel, S. Kahlau, and R. Bock. 2013. Organellar spirodelophila gen. et sp. nov. (Chaetopeltidales, Chlorophyceae), a novel GenomeDRAW—a suite of tools for generating physical maps of plastid and green alga associated with the aquatic angiosperm Spirodela polyrhiza. mitochondrial genomes and visualizing expression data sets. Nucleic Acids American Journal of Botany 103: 1–11. Research 41(Web Server issue): W575–W581. Weckström, K. J. Weckström M. Yliniemi, and A. Korhola. 2010. The ecology of Lowe, T. M., and P. P. Chan. 2016. tRNAscan-SE On-line: Search and contextual Pediastrum (Chlorophyceae) in subarctic lakes and their potential as paleo- analysis of transfer RNA genes. Nucleic Acids Research 44(W1): W54–W57. bioindicators. Journal of Paleolimnology 43: 61–73. Marchant, H. J., and J. D. Pickett-Heaps. 1972. Ultrastructure and differentia- Wendel, F. W., and J. J. Doyle. 1998. Phylogenetic incongruence: window intro tion of Hydrodictyon reticulatum. VI. Formation of the germ net. Australian genome history and molecular evolution. In D. E. Soltis, P. S. Soltis, and J. J. Journal of Biological Science 25: 1199–1213. Doyle [eds.], Molecular systematics of plants. II. DNA sequencing, 265–296. McManus, H. A., and L. A. Lewis. 2005. Molecular phylogenetics, morpholog- Kluwer, Boston, MA, USA. ical variation and colony-form evolution in the family Hydrodictyaceae Weng, M.-L., J. Blazier, M. Govindu, and R. Jansen. 2013. Reconstruction of the (Sphaeropleales, Chlorophyta). Phycologia 44: 582–595. ancestral plastid genome in Geraniaceae reveals a correlation between ge- McManus, H. A., and L. A. Lewis. 2011. Molecular phylogenetic relationships nome rearrangements, repeats, and nucleotide substitution rates. Molecular in the freshwater family Hydrodictyaceae (Sphaeopleales, Chlorophyceae), Biology and Evolution 31: 645–659. with an emphasis on Pediastrum duplex. Journal of Phycology 47: 152–163. Wilcox, L. W., and G. L. Floyd. 1988. Ultrastructure of the gamete of Pediastrum McManus, H. A., L. A. Lewis, K. Fučíková, and P. Haugen. 2012. Invasion of duplex (Chlorophyceae). Journal of Phycology 24: 140–146. protein coding genes by green algal ribosomoal group I introns. Molecular Wyman, S. K., R. K. Jansen, and J. L. Boore. 2004. Automatic annotation of orga- Phylogenetics and Evolution 62: 109–116. nellar genomes with DOGMA. Bioinformatics 20: 3252–3255. McManus, H. A., D. J. Sanchez, and K. G. Karol. 2017. Plastomes of the green Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of algae Hydrodictyon reticulatum and Pediastrum duplex (Sphaeropleales, large biological sequence datasets under the maximum likelihood criterion. Chlorophyceae). PeerJ 5: e3325. Ph.D. dissertation, University of Texas at Austin, Austin, TX, USA.