The Bryopsis Hypnoides Plastid Genome: Multimeric Forms and Complete Nucleotide Sequence

Background: Bryopsis hypnoides Lamouroux is a siphonous green alga, and its extruded protoplasm can aggregate spontaneously in seawater and develop into mature individuals. The chloroplast of B. hypnoides is the biggest organelle in the cell and shows strong autonomy. To better understand this organelle, we sequenced and analyzed the chloroplast genome of this green alga.


Introduction
The chloroplast of plants is thought to be descended from an originally free-living cyanobacterium; most of the genes in the genome of the cyanobacterium were transferred to the nucleus of the host cell during the evolutionary transformation of the endosymbiont into the chloroplast [1,2]. Nevertheless, the chloroplast retained its own genome, which performs some essential metabolic and biosynthetic pathways, such as photosynthesis and amino acid biosynthesis [1].
Chloroplast genomes (cpDNA) were the first plant genomes to be characterized because of their small size, limited number of repeated elements, and abundance of foliar tissues. At present, complete chloroplast genome sequences have been obtained from virtually all of the major higher plant and algal lineages. Comparative analyses of those complete cpDNA sequences not only highlight considerable differences at the organizational level, but also offer clarification of the evolutionary relationships among the main groups of algae and higher plants [3][4][5][6].
Compared with higher plants, algal chloroplast genomes, especially those of the green algae, exhibit numerous extreme features [7]. The cpDNA of Helicosporidium sp. [8], a parasitic, nonphotosynthetic green alga, is only 37.5 kb in size, which is the smallest among the cpDNAs characterized, and it lacks all genes for proteins that function in photosynthesis. The cpDNA size of the siphonous alga Acetabularia sp. is more than 2000 kb, which is the largest known cpDNA of photosynthetic organisms [1,9].
In addition to cpDNA sequencing, many studies now are focused on the organization of cpDNA with other elements, such as subgenomic minicircular, plasmid-like molecules [22] and the cpDNA conformation observation [23][24][25][26][27]. Using electron microscopy, Kolodner and Tewari [23] found that instead of only monomers, some dimers of cpDNA existed in plant cells. Deng et al. [24] reported that cpDNA can exist in trimer and tetramer form. Using in-gel procedures, including pulsed-field gel electrophoresis (PFGE), restriction fragment mapping, and fluorescence microscopy, Oldenburg and Bendich [28] found that most of the cpDNA from maize seedlings was in linear or complex branched forms rather than in circles. Multimeric forms of cpDNA also have been found in brassicas using field-inversion gel electrophoresis (FIGE) [25] and in tobacco using fluorescence hybridization in situ (FISH) [26]. In contrast, algal cpDNAs, like their higher plant counterparts, have circular restriction maps, and higher order organization (such as multimeric or anomalous forms) has not been identified in algal cpDNAs to date [7].
Bryopsis sp., which is a siphonous green alga, is a unicellular coenocytic giant cell. The total contents of the multinucleate cell can be squeezed out, leaving only the cell membranes and walls [29,30]. The extruded protoplasm without a cell membrane maintains sufficient viability to regenerate into a mature individual [29]. The regenerated individual can grow up to 58 cm, which is three times longer than the wild thalli [30], suggesting that regenerated alga have an advantage over wild individuals in terms of growth. Furthermore, the Bryopsis chloroplast is the largest organelle in the cell [31]and is thought to play an important role in the protoplast regeneration process [30], thus we inferred that its cpDNA may be special regarding the genome size and coding gene compared with other algae. However, previous studies of Bryopsis almost always have been focused on morphology and mechanisms of regeneration, and the chloroplasts, especially the cpDNAs, have received little attention.
Here we describe the complete sequence of the B. hypnoides chloroplast genome and present chloroplast phylogenies based on the genomic data currently available for higher plants and algae. Additionally, we report for the first time the presence of multimeric forms of cpDNA from B. hypnoides, a phenomenon that previously was known only for higher plant cpDNAs.

Results
Growth of the germinated aggregation of protoplasts of B. hypnoides The protoplasm extruded from wild B. hypnoides immediately aggregated into numerous balls and fine strands when they were mixed with natural seawater ( Figure 1A). The aggregations were covered with gelatinous envelopes within 3 min ( Figure 1B). Some of the aggregations germinated after 24 hours ( Figure 1C-E), and the germinated aggregations then developed into mature individuals ( Figure 1F). As in wild B. hypnoides, the mature regenerated alga could also develop into the rhizoid, which was used as an anchor, and the thallus.

Dimensions of chloroplasts in the B. hypnoides thallus
The size of chloroplasts in the B. hypnoides thallus varied from 3 to 16 mm, but most were 8-12 mm in length (Figure 2A), suggesting that the size diversity of the chloroplasts existed in the B. hypnoides thallus. This result was confirmed by the results of sucrose density gradient centrifugation ( Figure 2B), in which five clear discrete green bands appeared in the centrifugation tubes. Then all the bands were recovered separately and observed under the light microscope. We found that all the five bands were intact chloroplasts except for a few cell debris; the chloroplasts in the five bands were of different sizes, demonstrating the presence of five kinds of chloroplasts. When Structure and gene organization of the B. hypnoides chloroplast genome The B. hypnoides cpDNA sequence assembles into a circle of 153,429 bp; Figure 3 illustrates its gene map. Overall, the GC content of the cpDNA is 33.1%, which is comparable with that of Chlamydomonas (34.5%), Chlorella (31.6%), Pseudendoclonium (31.5%), and Pyramimonas (34.7%). Like its C. sertularoides and C. fragile homologs, the chloroplast genome of B. hypnoides does not contain the inverted repeat that is commonly found in many chloroplast genomes. A total of 111 functional genes, including 69 potential protein-coding genes, 5 ribosomal RNA genes, and 37 tRNA genes, were identified (Table 1). In addition, 29 open reading frames (ORFs) were identified with a threshold of 300 bp. All genes are present in a single copy, and this gene content is typical for chlorophyte cpDNAs. The sequence of B. hypnoides was most similar to that of C. vulgaris when compared with other completely sequenced chlorophyte cpDNAs. Table 2 compares the gene content of B. hypnoides cpDNA with that of other Ulvophyceae, Trebouxiophyceae, and Chlorophyceae (UTC) algal cpDNAs that have been completely sequenced to date. A common set of 84 genes is shared by these genomes. Relative to ulvophytes (O. viridis and P. akinetum), seven protein genes (chlI, minD, psaI, psaM, rpl19, ycf1, and ycf20) are absent from B. hypnoides cpDNA. Two genes, cysA and cysT, that encode sulfate transport proteins are absent in the ulvophytes O. viridis and P. akinetum but are present in the trebouxiophytes C. vulgaris and L. terrestris and in our B. hypnoides cpDNA.
Eleven introns in the B. hypnoides chloroplast are distributed among ten genes, among which rrs exhibits two introns and atpA, psaA, psbB, rbcL, rpl2, rpl5, rpl23, trnL-UAA and ycf3 each contain one intron. These introns vary from 348 to 2466 bp in size. According to their secondary structures, five introns belong to the group I family [32], and three of these carry an internal ORF encoding a putative LAGLIDADG homing endonuclease. The intron of rpl2 is commonly present in the chloroplast genomes of land plants, however it not found in the completely sequenced chlorophyte cpDNAs. The introns of rpl5 and rpl23 in Bryopsis are the first found in the known chloroplast genomes of Viridiplantae, and blast searches of these two intron sequences against the GenBank database failed to detect any homologous introns in other organisms. Figure 4A shows the results of PFGE analysis of cpDNAs of the wild B. hypnoides. In the cpDNAs from the different types of chloroplasts purified by sucrose density gradient centrifugation, at least four clear bands can be seen in every lane, each corresponding to a type of B. hypnoides chloroplast. The southern hybridization with labeled probes of the rbcL gene showed all the four bands were positive ( Figure 5), suggesting that the four bands represented all of the cpDNAs from B. hypnoides. The four bands were located at 150 kb, 300 kb, 450 kb, and 600 kb, all of which were in multiple relations. The relative amounts of the four bands in PFGE, which corresponded to monomers, dimers, trimers, and tetramers of the B. hypnoides cpDNAs, were estimated to be about 1, 1/2, 1/4, and 1/8 based on molecular hybridization. Figure 4B shows the results of PFGE analysis of cpDNAs from the regenerated B. hypnoides. These results are almost the same as those shown in Figure 4A, which confirms that the cpDNAs from the regenerated individual are similar to those of the wild alga.

Phylogenetic position of B. hypnoides chloroplasts
To elucidate the overall position of B. hypnoides in the plastid phylogeny of algae/land plants, a global analysis was performed using a subset of 31 taxa (see supplemental data) and 14,160 aligned characters. Thirty-one organisms were included as representatives of algae and higher plants and included the green lineage (streptophyte and chlorophyte lineage) and the non-green lineage (red and chromist lineage). The best tree identified two distinct lineages: the green lineage and the non-green lineage. Moreover, the chlorophytes and streptophytes formed two distinct green lineages ( Figure 5). Unexpectedly, the trees inferred with ML and MP methods both identified a clade uniting B. hypnoides and four complete sequenced members of the Chlorophyceae, this clade received 90% and 62% bootstrap value in MP and ML analyses, respectively. For other chlorophytes, N. olivacea represents the first branch of the Chlorophyta, which is also reported in recent phylogenetic analyses [6,12], but the branching order of the other four Prasinophyceae (O. tauri, P. parkeae, P. provasolii and Monomastix sp.) showed the different topology in the ML and MP analyses. In ML tree, P. provasolii is sister to all other UTC algae (bootstrap 51%); in the MP topology, it forms a moderately supported clade with O. tauri, Monomastix sp. and P. parkeae. For the UTC algae, the topological differences are also seen in the C. vulgaris clade, which was clustered with two Ulvophyceae (O. viridis and P. akinetum) in the MP tree, whereas in the ML tree, it located at the basal position within the UTC algae.
The relationships observed for the non-green algae and streptophytes taxa in the phylogeny are congruent with recently published phylogenies based on whole chloroplast genome sequences [6,33]. Within the Streptophyta, the clade uniting Chlorokybus and Mesostigma was placed basally and received strong bootstrap support. The other four Charophyceae were clustered with two higher plants (N. tabacum and M. polymorpha). In terms of the non-green lineage examined in this study, P. purpurea and G. tenuistipitate var. liui formed a strongly supported lineage that is sister to the clade uniting the G. theta, R. salina, O. sinensis and E. huxleyi. Two other red algae, C. merolae and C. caldarium, robustly clustered in a separate clade.

Discussion
The autonomy of the B. hypnoides chloroplast Our previous studies showed that the chloroplasts of B. hypnoides had strong autonomy for the following reasons: (1) The chloroplasts in B. hypnoides had greater vitality than other organelles under unfavorable conditions [34]; (2) chloroplasts purified by sucrose density gradient centrifugation can aggregate into spheres, although they cannot develop into mature individuals [35]; and (3) The chloroplasts from siphonous algae possess great vitality and can maintain activity in other cells or other body cavities. For example, a symbiotic association exists between siphonaceous green algae (Acetabularia, Bryopsis, Caulerpa, and Codium) plastids and some marine sea slug species that are highly specialized herbivores that feed on siphonalean algae by puncturing the cells and sucking out the contents. The chloroplasts of siphonous algae are not always digested and can lodge in the body of the animal and conduct photosynthesis for at least 3 months [36][37][38][39][40][41]. Moreover, genes supporting photosynthesis have been acquired by the host animal via horizontal gene transfer, and the encoded proteins are retargeted to the chloroplast [42]. In summary, compared with chloroplasts of other algae and higher plants, B. hypnoides chloroplasts have great vitality and independence. Because of these extraordinary traits, we conducted research both on the genome sequence and the conformation of the cpDNA.

Distinctive features of the B. hypnoides chloroplast genome
The genome size, arrangement, and IR-lacking structure of the chloroplast genome of B. hypnoides more closely resemble that of C.
vulgaris cpDNA [15] than its O. viridis [17] and P. akinetum [5] homologs. In addition to the 84 conserved genes that exist in all of the completely sequenced UTC algal cpDNAs, B. hypnoides shares an additional 14 genes with C. vulgaris cpDNA, compared to 12 and 10 genes with O. viridis and P. akinetum, respectively (Table 2).
The most notable feature of the B. hypnoides chloroplast genome is that the rRNA locus consists of five genes: rrn23, rrn16, rrn7, rrn5, and rrn3. The same situation can be found only in C. reinhardtii cpDNA [18], as genes rrn7 and rrn3 are absent from all other completely sequenced chlorophyte cpDNAs. Similar to C. reinhardti, the rRNA gene cluster of B. hypnoides is arranged in the order rrn16-rrn5-rrn23-rrn3-rrn7; however, the typical rRNA operon in B. hypnoides has been broken in half and the SSU and LSU genes are distributed at opposite ends of the gene map circle, as is found in the ulvophytes C. sertularoides [43] and C. fragile [44]; this might be an outcome of the loss of the inverted repeat.
Another surprising feature of the B. hypnoides chloroplast gene repertoire is the presence of 10 unusual tRNA genes that have not been found in other completely sequenced chlorophyte cpDNAs. Five of them (trnA-AGC, trnE-CUC, trnI-AAU, trnV-CAC, and trnV-AAC) correspond to those identified in some embryophyte cpDNAs, whereas the other five (trnA-CGC, trnK-CUU, trnP-AGG, trnQ-CUG, and trnT-AGU) have been found previously only in some bacterial genomes. These unusual tRNA genes may not be essential for plastid function in green algae or may not be functional genes; they also might be involved in the special physiological functions of B. hypnoides.
Wolfe et al. [45] reported that the cp genome of Epifagus virginiana, a plant with non-photosynthetic chloroplasts, encodes  only 25 proteins because the photosynthetic machinery and the corresponding genes are not needed. Glöckner et al. [46] identified several genes unique to the cp genome of C. caldarium that correlate to its special environmental conditions. Thus, it seems that loss or gain of function is accompanied by changes of genes in the cp genome. Herein we originally presumed that the size and gene  The multimeric forms conformation of B. hypnoides cpDNA Figure 4 shows that the cpDNA of B. hypnoides has multimeric forms, including at least monomers, dimers, trimers, and tetramers, and these are similar to the multimeric cpDNA forms of higher plants, such as Arabidopsis, tobacco, and peas [24,26,47]. The five different sizes of B. hypnoides chloroplasts showed the same cpDNA characteristics in multimeric forms, suggesting that the relative amounts of the different forms of B. hypnoides cpDNA had no relationship to chloroplast dimension.
Green alga is thought to have been the progenitor of higher plants [10][11][12]; B. hypnoides is a green alga thus, its cpDNA conformation may be similar to that of higher plants. However, a PFGE study of the C. reinhardtii chloroplast genome did not reveal the higher order organization that was observed previously for higher plant cpDNAs [18]. Our phylogenetic tree revealed that B. hypnoides is located in the same clade as C. reinhardtii (Figure 5), whereas their cpDNA conformations were quite different. This finding indicates that the cpDNA of green algae is structurally plastic.
Lilly et al. [26] proposed that the formation of higher order multimeric molecules may result from recombinational events or the random cleavage and fusion of replication intermediates. Recombinational events correlate to the presence of IRs, as revealed in both Arabidopsis and tobacco [26,48]. Recombination between IRs that occurs on two separate monomers results in a dimer [49]; subsequent recombination events between molecules presumably produce multimers. However, our study demonstrated that multimeric forms of cpDNA also exist in the IR-lacking chloroplast genome, which suggests that the higher order organization of the chloroplast genome maybe less related to the presence of IRs than was previously thought. Our results indirectly support the alternative explanation that multimeric molecules were produced by the random cleavage and fusion of replication intermediates. However, this explanation requires confirmation further.

Similar conformation of cpDNA between wild and regenerated B. hypnoides
Our previous studies showed that the regenerated alga had an advantage over the wild individual in terms of growth and that biochemical compositions differed between the wild and the regenerated alga [30,34]. The regenerated alga can grow up to 58 cm in length, which is three times longer than the wild form [30]. Wang and Tseng [34] reported that when the regenerated alga was on the decline, the organelles aggregated in the thallus and then moved to the outside; next, one organelle aggregation located outside of the thallus germinated and developed into a mature alga. From this, we inferred that the DNA in the organelles of the regenerated alga, especially the cpDNAs, underwent changes during organelle aggregation and later development into a mature individual. The cpDNAs from the regenerated B. hypnoides were similar in size and conformation to the cpDNA of the wild alga (Figure 4), and both exhibited monomeric, dimeric, trimeric, and tetrameric forms of cpDNAs. Thus, the differences between the wild alga and the regenerated individual probably result from gene expression rather than alteration of the genome, especially the cpDNAs.

Evolution of the chlorophycean chloroplast genome
The basal position of the Prasinophyceae in the Chlorophyta is well established, but the branching order of the Ulvophyceae, Trebouxiophyceae, and Chlorophyceae remains unresolved [5,51]. There are two hypotheses concerning the divergence order of the UTC lineages: 1) phylogenetic inferences from cpDNAencoded proteins and genes favor the hypothesis that the Ulvophyceae is a sister to the Trebouxiophyceae [5]; and 2) chloroplast phylogenies inferred from gene order [5] and mitochondrial phylogenies inferred from proteins or genes [51] suggest that the Ulvophyceae share a sister relationship with the Chlorophyceae. Our phylogenetic analyses of 42 chloroplast proteins revealed that the Ulvophyceae clade I (O. viridis and P. akinetum) is a sister to the Trebouxiophyceae I (C. vulgaris). However, the B. hypnoides (Ulvophyceae II) chloroplast genome is closely related to the Chlorophyceae. Although the B. hypnoides chloroplast genome shares similarities with its O. viridis and P. akinetum counterparts in terms of gene order, its IR-lacking structure and gene content are quite different; we therefore inferred that Ulvophyceae was non-monophyletic group, B. hypnoides is located in the different phylogenetic lineage with O. viridis and P. akinetum. Moreover, our phylogenetic analyses favor the premise that the Trebouxiophyceae and Prasinophyceae were also non-monophyletic groups, thus it was difficult to state a precise taxonomic relationship among the UTC lineages in this study. These different phylogenetic results may be caused by the insufficient taxon sampling. So studies on additional chloroplast genome data, especially from the UTC algae, will be very useful for determining the phylogenetic relationships among the major lineages of Chlorophyta.

Materials
Bryopsis hypnoides Lamouroux was collected from the intertidal zone of Zhanqiao Pier, Qingdao, China (36u39N, 120u199E). The fresh algae, which were rinsed with plenty of autoclaved seawater and brushed with a soft brush to remove the surface microbial and epiphytic organisms, were cultured in autoclaved sea water under irradiance of 25 mmol m 22 ? s 21 with a 16 h light: 8 h dark regime at room temperature [30].

The aggregation of organelles in the protoplasm
Thalli of B. hypnoides were cut into small pieces and then placed in eight layers of sterilized gauze to squeeze out the protoplasm.
The extruded protoplasm was mixed with an equal volume of sterilized seawater (pH 8.3) and gently rocked. The organelles in the protoplasm aggregated into spheres of different sizes, and the aggregated spheres were cultured into mature B. hypnoides (the regenerated B. hypnoides) under the culture conditions described above.

Chloroplast and cpDNA purification
The protoplasts squeezed from the wild B. hypnoides were added to double volume ice-cold extraction buffer (400 mM sucrose, 50 mM Tris, 20 mM EDTA, 0.2% BSA, 0.2% b-mercaptoethanol, pH 7.8) and then filtered through four layers of cheese cloth. The filtrate was centrifuged at 800 g at 4uC for 10 min, and the pellet was suspended in the extracted buffer. Most of the pellets were found to be chloroplasts under microscopic examination, so they were layered onto the sucrose density gradient from 10% to 60%. The gradient was centrifuged at 150,000g for 90 min, and five clear green bands appeared in the tube after centrifugation. The bands were removed from the tube separately, and dialyzed against the rinse buffer (400 mM sucrose, 50 mM Tris, 0.5%, BSA pH 7.8) for 5 hours to remove sucrose, then were observed under the microscope.
The purified chloroplasts were washed twice with the buffer (50 mM Tris-HCl, pH 8.0, 25 mM EDTA), suspended in the lysis buffer (50 mM Tris, pH 8.0, 25 mM EDTA, 2% SDS, 50 mg/ml proteinase K), and incubated at 40uC for 3 hours with gently shaking, following by being centrifuged at 10 000 g at 4uC for 15 min. The supernatant was extracted several times with phenol/ chloroform and then precipitated with cold ethanol. CsCl density gradient ultracentrifugation was used for further purification of the cpDNA. CsCl and Hoechst dye No. 33258 were added to the crude cpDNA, and the mixture was centrifuged at 240,000g at 20uC for 38 hours with a Beckman Ti 80 rotor. cpDNA bands were visualized under UV illumination and then recovered. The cpDNA was precipitated and dissolved in TE buffer after both the Hoechst dye 33258 and CsCl were removed. The cpDNA from the regenerated B. hypnoides then was purified as described above.

cpDNA sequencing
The purified cpDNA was sheared by nebulization, and 1,50063,000 bp fragments were recovered by electroelution after agarose gel electrophoresis. These fragments were treated with T4 DNA polymerase and cloned into the SmaI site of PUC18. After transformation of electrocompetent E. coli TOP10 cells (Invitrogen, Carlsbad, CA), recombinant plasmids were isolated and nucleotide sequences were determined with the PRISM dye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA) on a DNA sequencer (model 373; Applied Biosystems) using T3 and T7 primers. Sequencing data were accumulated to 106 coverage for all PCR fragments; remaining gaps were cloned by PCR. The determined sequences were accumulated, trimmed, aligned, and assembled using the Phred-Phrap (Phil Green, University of Washington, Seattle, WA, USA) and Consed programs [52]. The fully annotated B. hypnoides chloroplast genome sequence has been deposited in GenBank with accession number GQ892829.

PFGE and southern blot analysis
The PFGE assay was performed by the method described by Sambrook and Russell [53]. Brifely, the purified chloroplasts were mixed with an equal volume of low melting point agarose, which was dissolved in rinse buffer containing 20 mM EDTA, and 90 ml of the mixture was added into every pole of the mould (Bio-Rad, Richmond, Calif), then solidified at 4uC. The solid gels were put into the lysis buffer (0.01 M Tris, 0.45 M EDTA, pH 7.8, 2% SLS, 10 mg/ml proteinase K) and incubated at 50uC for 40 h, during which the buffer was changed twice. The gels then were rinsed six times with TE (10 mM Tris, 1 mM EDTA, pH 8.0) and then stored at 4uC. Pulsed-field gel electrophoresis (PFGE) was performed on a Bio-Rad CHEF Mapper TM according to the manufacturer's instructions (CHEF Mapper TM and CHEF Mapper XA Pulsed Field Electrophoresis Systems). After PFGE, the gel was dyed with 0.5 mg/ml EB for 30 min, and the result was observed by Pharmacia Biotech ImagemasterH VDS.
A DNA fragment encoding the B. hypnoides rbcL gene (1145 bp; GenBank accession no. AY566304) was prepared from genome DNA as described previously [54]. The rbcL gene then was labeled according to the directions provided in the Dig High Prime DNA Labeling and Detection Starter Kit I (Roche, Germany) and used as a probe for southern blot analysis. DNA in the gel was denatured and transferred to the positively charged nylon membrane (Osmonics, Westborough, MA, USA) [53] and crosslinked under UV for 90 seconds, then hybridized with the probe following the instructions for the Roche Kit.

Sequence and phylogenetic analyses of cpDNA
Genes were identified using Blast homology searches provided by the National Center for Biotechnology Information (NCBI) server (http://www.ncbi.nlm.nih.gov/BLAST/). Protein-coding genes and positions of ORFs were determined using ORFFIN-DER at NCBI. tRNA genes were annotated using the tRNAscan-SE program [57]. Intron boundaries were located by modeling intron secondary structures [32,56] and by comparing the sequences of intron-containing genes with those of intron-less homologues using FRAMEALIGN of the Wisconsin package. Circle graphs were generated using the CGView program [57].

Supporting Information
Data S1 Algal and land plant chloroplast genomes examined in the phylogenetic analyses Found at: doi:10.1371/journal.pone.0014663.s001 (0.02 MB DOC)