Complete chloroplast genome for Caulerpa racemosa (Bryopsidales, Chlorophyta) and comparative analyses of siphonous green seaweed plastomes

Genoma completo del cloroplasto de Caulerpa racemosa (Bryopsidales, Chlorophyta) y un análisis comparativo de los plastomas de algas marinas sifonáceas.

Daryl W. Lam¹, Juan M. Lopez-Bautista^1*

¹Department of Biological Sciences, The University of Alabama, Tuscaloosa, Alabama, United States of America

e-mail:jlopez@ua.edu

Introduction

The division Chlorophyta is a species rich group strictly comprised of green algal organisms that contain chlorophylls a and b (Graham et al. 2009).  The Chlorophyta lineage is sister to the Streptophyta lineage, a plant division that includes some green algae, mosses, liverworts, hornworts, and vascular plants (Lewis & McCourt 2004; Leliaert et al. 2012).  The split between the two lineages has been estimated to have occurred ~936 million years ago (Becker 2013; Parfrey et al. 2011). Although not as well studied as the Streptophyta, chlorophytan species exhibit a vast array of morphological and ecological diversity. They exhibit a wide array of morphologies, a few examples include: single celled organisms that form palmelloid colonies (Zechman et al. 2010), flagellated free-living organisms (Pröschold et al. 2001), multicellular macrophytic blades (Melton et al. 2015), filamentous thalli with a single nucleus per cell (Rindi & Lopez-Bautista 2009), filamentous thalli with multiple nuclei per cell (Leliaert et al. 2007), calcified uninuclear siphons (Olsen et al. 1994), and multinuclear siphons (Lam & Zechman 2006).  The Chlorophyta are ubiquitous and can be found in marine, freshwater, and terrestrial habitats (Graham et al. 2009). Recent systematic assessments of this division based on phylogenomic analyses inferred a large well-supported monophyletic group, which included taxa in the classes Chlorodendrophyceae, Chlorophyceae, Pedinophyceae, Trebouxiophyceae, and Ulvophyceae and was branded as the “core Chlorophyta” clade (Fučíková et al. 2014).
Within the “core Chlorophyta”, the order Byopsidales (class Ulvophyceae) is comprised of 564 marine species and a single freshwater species all with siphonous thalli (Guiry & Guiry 2015).  In other words, the vegetative thallus is a single undivided cell, which contains multiple nuclei that arose from mitosis without cytokinesis (Leliaert et al. 2015). A systematic revision of the Bryopsidales erected two suborders, the Bryopsidineae (including the genera Bryopsis, Codium, and Derbesia) and the Halimedineae (Caulerpa, Halimeda, and Udotea), and was based on morphological attributes within the grouping (Hillis-Colinvaux 1984).  Based on Hillis-Colinvaux’s assessment (Hillis-Colinvaux 1984) there are several important morphological features that distinguish the two suborders. The Halimedineae have two types of plastids (heteroplasty), amyloplasts and chloroplast, while Bryopsidineae species have only chloroplasts (homoplasty).  Sexual reproduction characteristics also differ in the two suborders. The Bryopsidineae reproduce with septa separating the reproductive cells from the rest of the thallus, while the Halimedineae lack septa and reproduce in a holocarpic manner where most of the vegetative thallus is filled with gametes (Hillis-Colinvaux 1984).  After the release of gametes, the greater majority of the protoplasm is lost with an empty “ghost thallus” remaining (Kooistra 2002). It should be noted that Caulerpella is an exception to this rule as the genus has been reported to be nonholocarpic (Fama et al. 2002).  Molecular phylogenetic analyses based on the large subunit of the plastid-encoded gene RuBisCO supported the monophyly of both suborders and placed the cryptic genus Pseudocodium, which superficially looks like Codium but exhibits heteroplasty with Halimedineae taxa (Lam & Zechman 2006).  Subsequent phylogenies using the same molecular marker inferred that the marine genera Avrainvillea and Cladocephalus formed a robustly supported clade along with the freshwater species Dichotomosiphon tuberosus (Curtis et al. 2008).  A time-calibrated multilocus (five plastid markers and one nuclear locus) phylogeny inferred the separation of the two suborders to have occurred in the Early Paleozoic, approximately 465 million years ago (Verbruggen et al. 2009).
Clearly, molecular based systematic analyses have increased our understanding of bryopsidalean evolutionary processes.  Recent technological advances in high throughput DNA sequencing currently offer a wealth of gene data at a fraction of the cost of Sanger-based sequencing methods. Because chloroplasts are abundant in most macrophytes and each individual chloroplast contains multiple copies of genomic DNA, plastid genomes (plastomes) are a common target for botanical based studies (Lutz et al. 2011). The Bryopsidales are not an exception to this trend, with three completed chloroplast genomes currently published (Lü et al. 2011; Leliaert & Lopez-Bautista 2015).  Within the suborder Halimedineae, the complete chloroplast genome (105,200 base pairs (bp)) of Tydemania expeditionis was recently assembled and annotated (Leliaert & Lopez-Bautista 2015).  For the suborder Bryopsidineae two species of Bryopsis have been assembled and annotated.  B. hypnoides (153,429 bp) and B. plumosa (106,859 bp) have published plastomes (Lü et al. 2011; Leliaert & Lopez-Bautista 2015).  All three plastomes were published as circular contigs (Lü et al. 2011; Leliaert & Lopez-Bautista 2015).
In order to infer the evolutionary history of this order of green algae, we have sequenced, de novo assembled, and performed annotations for the chloroplast genome of one bryopsidalean species, Caulerpa racemosa.  Although there have been chloroplast genome analyses for C. sertularoides (Leh-man & Manhart 1997) through Southern hybridization/analysis and restriction fragments analysis and partial plastid genome data (~ 30,000 bp, 23 genes) for C. filiformis, C. racemosa represent the first completely sequenced genome for the family Caulerpaceae. Here we present, a phylogenomic inference based on protein-coding plastid genes.  In addition, we infer genomic rearrangement events for all taxa in the order with currently completed plastomes.

Materials and Methods

Field collection and DNA extraction/sequencing
C. racemosa (voucher UNA00072801) was collected from La Parguera, Isla Magueyes, Enrique, Puerto Rico, USA on November 11, 2014. DWL identified these samples based on the two taxonomic keys (Littler & Littler 2000; Littler et al. 2008). DNA was extracted from silica-gel-dried C. racemosa using the E.Z.N.A. Plant DNA Extraction Kit (Omega Bio-tek Norcross, GA USA) following the protcol therein. For this species, a paired-end 101 bp library was constructed using the standard Illumina Truseq adapter. MiSeq DNA sequencing was performed at Cold Spring Harbor Laboratory (Cold Spring Harbor, NY, USA) on a multiplexed sequencing run that contained two other algal species. This run generated 2,862,510 paired-end reads (2 x 101 bp).

Genome assembly
The dataset was downloaded to the Alabama Supercomputer Center’s Dense Memory Cluster (DMC).  Fastq formatted files were quality checked using the software packages fastQValidator (https://github.com/statgen/fastQValidator) and fastqc (http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc/).  The A5 Assembly pipeline (Tritt et al. 2012) was used to remove ambiguous and low quality portions of the raw read data, correct for sequencing errors, assembly, scaffolding, and reassembly using 120 gigabytes of RAM and a single processing core. In order to identify the chloroplast genome from the rest of the data, all contigs from the de novo assembly were compared against a custom local dataset comprised of protein coding sequences (CDS) and ribosomal RNA (rRNA) regions from previously published bryopsidalean chloroplast genomes: Bryopsis hypnoides (GenBank accession number NC_013359.1), Bryopsis plumosa (NC_026795.1), and Tydemania expeditionis (NC_026796.1) via MegaBLAST (Morgulis et al. 2008) with an E-value threshold of  < 1x10-10 through Geneious v 7.9 (Biomatters, http://www.geneious.com) using ten processing threads.  The linear contigs had identical sequence data (~100 bp) on both ends.  Subsequently, one of the identical sequence ends was removed and the contig was circularized in Geneious.  In order to edit the nucleotide data for the contigs, the original raw read data was trimmed of low-quality bases and adapters sequences by Trim Galore! (http://www.bioinformatics.babraham.ac.uk/projects/trim_galore/).  Trimmed reads were mapped on to the chloroplast contig in Geneious and a majority rule consensus sequences was created. 

Gene annotation
Annotations were made in Geneious by mapping the aforementioned previously published gene regions to the de novo plastome.  For protein-coding genes open reading frames (ORFs) were identified in Geneious using a bacterial/plant plastid genetic code.  Matching annotations were double-checked by BLASTx (Gish & States 1993) against the non-redundant NCBI database.  Relatively large ORFs (> 700 bps) that did not match any of the aforementioned gene annotations were are also queried against the NCBI non-redundant database for potential gene homology to other organisms in GenBank.  Intron-exon boundaries for protein coding genes were identified by translational alignment in Geneious. For rRNA genes these boundaries were inferred by MAFFT (Katoh & Standley 2013) alignments as implemented in Geneious.  Transfer RNA (tRNA) gene predictions were made using tRNAscan-SE (Lowe & Eddy 1997) with the following parameters: search mode set to “Organellar”, searching with Cove only (cutoff score = 15), Covariance model tRNA2.cm, max intron + var. length = 40, and pseudogene checking disabled.  The C. racemosa plastome is available for download as NCBI/GenBank accession KT946602.   
    
Phylogenomic analyses
Maximum likelihood tree based on atpA, atpB, atpE, atpF, atpH, atpI, clpP, infA, petA, petB, petG, psaA, psaB, psaC, psaJ, psbA, psbB, psbC, psbD, psbE, psbF, psbH, psbI, psbJ, psbK, psbL, psbN, psbT, rbcL, rpl2, rpl5, rpl14, rpl16, rpl20, rpl23, rpl36, rps3, rps4, rps7, rps8, rps9, rps11, rps12, rps14, rps18, rps19, tufA, ycf3, ycf4, and ycf12 genes with a GTR + G model implemented in RAxML (Stamatakis 2014).  Poorly aligned regions were masked using Gblocks (Talavera & Castresana 2007).  Nodal support values were based on 1000 bootstrap replicates.  Prasinophyte taxa were used to root this phylogeny. 
From Treebase (http://treebase.org/), the nucleotide alignment from Fučíková et al. (2014) was downloaded (study 16203 id M24024).  The concatenated dataset was separated into individual alignments by gene.  Transcriptome sequence data for the taxon Codium decorticatum was removed from this dataset because its nucleotide similarity was nearly identical to published sequences for Bryopsis plumosa (Leliaert & Lopez-Bautista 2015). We added corresponding gene data from the following ulvophycean species:  Bryopsis plumosa (GenBank NC_026795.1), Caulerpa racemosa (this study), Tydemania expeditionis (NC_026796.1), and Ulva sp. (GenBank KP720616.1). As in Fučíková et al. (2014), the psaM, rpl32, and the highly variable ycf1 gene were not included in the concatenated dataset. The single gene datasets were realigned with MAFFT and concatenated into a single alignment using Sequence Matrix (Vaidya et al. 2011). Hypervariable regions of the alignment were removed from the dataset using the Gblocks (Talavera & Castresana 2007) web server http://molevol.cmima.csic.es/castresana/Gblocks_server.html) allowing for smaller final blocks, gap positions within the final blocks, less strict flanking positions and many contiguous non-conserved positions. Phylogenetic analyses were performed in RAxML version 8.0.24 (Stamatakis 2014) with a GTR+G model with rapid bootstrapping (1000 replicates) and subsequent maximum likelihood search on the Alabama supercomputer DMC with 10 cores and 25 gigabytes of RAM.

Analysis of genomic rearrangements
The C. racemosa plastome was aligned using the progressive Mauve algorithm (Darling et al. 2004) using the full alignment option and automated calculation of locally co-linear block scores. 

Results and Discussion

For C. racemosa, a total of 2,322,843 reads were mapped to the 176,522 bp circular contig with an average coverage per site at 1701.3x and had an overall nucleotide similarity of 99.3% (Fig. 1).  There were 76 protein coding genes, 27 tRNA genes, 3 rRNA genes, 18 introns, and 15 ORFs (≥ 700 bps in length) that were annotated to the C. racemosa plastid genome. This is currently the largest known plastid genome for the order Bryopsidales. C. racemosa was missing the ycf47 gene that was found in the bryopsidalean species Bryopsis plumosa and Tydemania expeditionis. However, it should be noted that the overall gene content for C. racemosa is quite similar to that of the previously published bryopsidalean plastomes (Lü et al. 2011; Leliaert & Lopez-Bautista 2015). C. racemosa contained the tilS pseudogene at positions 62,978 to 62,908 (reverse orientation).  Likewise, the tilS pseudogene was found in both the Bryopsis and Tydemania plastomes (Leliaert & Lopez-Bautista 2015).
Since the overall gene content for C. racemosa among Bryopsidales is quite similar, the differences in genome size are partially due to variance of intergenetic space and introns.  C. racemosa has longer intergenetic regions as compared to Bryopsis plumosa and Tydemania expeditionis. One example of intronic differences is exemplified in the alignment of rrl rRNA gene.  The C. racemosa rrl gene contained six introns that increased the gene length from ~3 kbp to over 9 kbp when compared to T. expeditionis. 
In the C. racemosa plastome, ORFs (≥ 700 bps in length) are numerous (15 ORFs total). ORFs 4, 5, and 8 when queried against the NCBI nr database were found to be significantly similar (E-values of 6.00 x 10-53, 2.00 x 10-155, and 9.00 x 10-53, respectively) to bacterial DNA methylase proteins.  Also in the C. racemosa ORF 6 was significantly similar (E-value < 0.00) to a type I bacterial restriction endonuclease.  Restriction endonucleases are used in bacteria as a means of defense against foreign DNA (Arber & Linn 1969).  These restriction endonucleases digest DNA at specific restriction sites (Arber & Linn 1969).  Meanwhile, the bacteria’s host genome is protected biochemically by DNA methylation.  It is possible that C. racemosa’s plastids seem to exhibit a similar defense mechanism against foreign DNA invasion as many prokaryotic organisms.  However, this hypothesis still needs further investigation as expression-based evidence (transcriptomics) is necessary to support this theory. That being said, the most likely means of acquirement is via horizontal gene transfer from bacteria to the host’s plastid genome. Multiple bacterial communities have been found inside of the Caulerpa thallus (Delbridge et al. 2004).  Furthermore, ORFs of bacterial origin have been found in the plastomes of Bryopsis and Tydemania from previously published studies (Leliaert & Lopez-Bautista 2015), suggesting that transfer of bacterial genes into bryopsidalean plastome is fairly common in the order Bryopsidales. However, plastomes for more Bryopsidales species are necessary to elucidate the evolution of bacterial ORFs in these genomes.
Phylogenomic analyses

The 50-gene nucleotide alignment of chlorophytan algae was analyzed via maximum likelihood and the resulting phylogeny is depicted as Fig. 2. This tree was rooted with Prasinophytes (sensu lato) taxa. The green algal class Chlorophyceae formed a separate robustly supported (100% bs) clade.  However, the monophyly of the classes Trebouxiophyceae and Ulvophyceae were unresolved. In contrast, the order Bryopsidales formed a robustly supported claded (100% bs).  Within the order Bryopsidales, the suborders Bryopsidineae and Halmedineae formed separate robustly supported clades (100% bs and 100% bs, respectively), thus supporting the systematic/taxonomic assessment of Hillis-Colinvaux (1984) based on purely morphological observations.  In general, the presented phylogenomic analysis is quite similar to previously published trees (Fučíková et al. 2014 ; Lam & Zechman 2006; Lieliaert & Lopez-Bautista 2015; Melton et al. 2015).  This being said, more plastid gene data from a more diverse set of chlorophytan species is required before a robustly supported tree for this division of green algae can be inferred.

Chloroplast genome rearrangements

Based on the inferred monophyly of both suborders in the 50-gene phylogenomic analysis, separate Mauve alignment analyses were conducted for the Bryopsidineae and Halimedineae. These analyses included all currently completed chloroplast genome data for the order Bryopsidales (Lü et al. 2011; Leliaert & Lopez-Bautista 2015; and this study). Fig. 3 represents the plastid genomes of the two members of the suborder Bryopsidineae where colored locally collinear blocks (LCB) represent homologous sequence regions that do not contain major rearrangements.  These results were quite similar to the results of Leliaert and Lopez-Bautista (2015). The number of genomic rearrangement in the suborder Bryopsidineae is fewer than the rearrangement events inferred for the suborder Halimedineae (Fig. 4). However, this might be an artifact of taxon sampling as Fig. 3 has two species from the same family/genus (Bryopsidaceae/Bryopsis), while Fig. 4 has two species from two different families (Caulerpaceae and Udoteaceae).

Conclusions.

The circular plastid genome of C. racemosa (176,522 bp) is currently the largest completely sequenced plastome for the Bryopsidales.  For comparison, the complete chloroplast genome Tydemania expeditionis is 105,200 bp in length. Completed plastomes of Bryopsis hypnoides (153,429 bp) and B. plumosa (106,859 bp) have also been published (Lü et al. 2011; Leliaert & Lopez-Bautista 2015).  Overall the genome contained similar gene complements to previously published bryopsidalean plastomes. The larger genome size of C. racemosa was due to differences in number and size of introns, intergenetic spacers, and ORFs. This species had ORFs that were most likely transferred horizontally from bacterial communities. In addition, C. racemosa had ORFs that were significantly similar to bacterial methyl-transferases and a restriction endonuclease genes. Phylogenomic inferences based on a 50-gene dataset supported the current taxonomy and previously published phylogenetic studies. Mauve based synteny analyses suggested several genomic rearrangement events for taxa in both suborders, however there were more extensive rearrangement events in the Halimedineae. Although these genomes are a large step toward the elucidation of the chlorophytan evolution, more plastid genomes from more green algal species are necessary in order to fully unravel the history of this green algal division. 

Acknowledgements.

We thank Trey Melton, Ana Tronholm, and David Ward for sample collection. Jeremy McDoniell, Hendrick Adams, and Dustin Steffan assisted DWL with DNA extractions and laboratory protocols.  We acknowledge the Alabama Supercomputer Cluster for analysis time on its high performance supercomputer that was used for both short-read assembly and phylogenomic inference. To the College of Arts & Sciences and the Department of Biological Sciences of The University of Alabama for their support. The National Science Foundation through Assembling The Tree of Life for Green Algae GRAToL (DEB 1036495).

REFERENCES.

Recibido: 21 septiembre
Revisado: 24 octubre
Corregido: 10 noviembre
Aceptado: 11 de noviembre
Revisores: 2 revisores anónimos