Completing the nuclear genome sequence, from telomere to telomere, without gaps or ambiguities, required two distinct subprojects: (i) filling in all 46 previously existing gaps between contigs, and (ii) sequencing the 34 (of 40) chromosome ends that had not been sequenced previously by the shotgun method 9. Using PCR with sequence-specific primers to amplify portions of the C. merolae bacterial artificial chromosomes (BACs) that contained gap regions of interest, and then sequencing the resulting subclones, we have reduced the number of gaps between contigs from 46 to zero (Figure 1). A variety of methods (Additional files 1, 2) were then used to sequence all of the chromosomal ends that had not been sequenced previously (Figure 1). The global size of the 46 gaps that we have sequenced in this study is 46 469 base pairs. These sequences are G+C-rich (61%) compared with that of the whole nuclear genome (55%), and six of the 46 have extremely high G+C contents (70–76%). Furthermore, sequences corresponding to 13 previous gaps have extremely high nucleotide identity (99.4–100%) to other genomic regions. It turns out that the nuclear genomic sequence previously reported 9 contains misassembled and/or incorrect sequences of >20 kbp in total. The resulting complete genome sequence contains 16546747 base pairs and covers 100% of the chromosomal genome, from telomere to telomere. The numbers of base pairs in each of the 20 nuclear chromosomes are given in Table 1. These 20 linear nuclear chromosomes plus the two circular organellar DNA molecules 1718 comprise the entire genome of the organism, and contain 16728945 base pairs (Table 1).

The total number of histone genes varies greatly by species. Animal genomes typically have several hundred to several thousand histone genes organized as a set of tandemly repeating quintets of the five major histone gene types 1920. In contrast, histone genes are located on more than one chromosome in some organisms, such as the ultra-small green alga Ostreococcus 12 and the flowering plant Arabidopsis 4. In our previous study, the complete organization of the histone cluster area in the C. merolae genome was left unresolved, because of several gaps within this region of sequence 9. However, our present complete sequence establishes that all C. merolae histone genes are localized on chromosome 14 and form a small cluster (around 48 kbp long) that includes two copies of the three core histone genes (H2A, H2B and H4), three copies of the H3 gene, and a single copy of the linker-histone H1 gene (Figure 2). In Saccharomyces cerevisiae 2, Schizosaccharomyces pombe 6, Dictyostelium discoideum 11, Encephalitozoon cuniculi 5 and Ostreococcus tauri 12 the histone genes also exhibit small copy numbers (up to three), as in C. merolae, but they are dispersed across more than two chromosomes. Thus, the C. merolae histone genes are present in the most compact cluster yet described.

Whole-genome shotgun sequencing suggested that the repeat unit in C. merolae telomeres is AATGGGGGG 9, but in that study only six of the 40 chromosome ends were examined 9 (Figure 1). Here the sequencing of all remaining termini confirmed that AATGGGGGG is the telomere repeat sequence in all C. merolae chromosomal ends. In most plants, the telomeres are composed of many copies of the sequence TTTAGGG 21, and the C. merolae telomere sequence, AATGGGGGG, has never yet been found elsewhere.

Telomere length varies among plants from approximately 0.5 kbp in the green alga Chlorella vulgaris to 150 kbp in tobacco 2223. Telomere restriction fragment analyses using the (CCCCCCATT)3 probe revealed AATGGGGGG repeats varying from 400 to 700 bp in the total chromosomal ends in the C. merolae genome (data not shown). Our Southern blot analysis using a specific genomic probe suggested that C. merolae chromosome 15 has around a 400-bp telomere repeat sequence at the left end (see Additional file 3). However, the longest telomeric repeats that could be sequenced in this study were 2.5 repeats (AATGGGGGGAATGGGGGGAATGGG) in the right end of chromosome 1, possibly because long stretches of AATGGGGGG repeats are difficult to clone or sequence using conventional techniques used in this and previous studies 9.

The putative telomerase catalytic subunit, telomerase reverse transcriptase (TERT) is transcribed (CMD110C) in C. merolae 9. In the C. merolae genome, we found two possible telomerase RNA subunit genes in chromosomes 13 and 16, based on two transcripts, CMM123T and CMP131T, which included UUCCCCCCAUU and CCAUUCCCCCCAUU sequences, respectively. The telomerase RNA template sequence (CCCCCCAUU) has been detected in only these two hypothetical non-coding RNA genes among all the predicted genes in the present 100% complete nuclear genome. These are the first convincing candidate telomerase-RNA subunit genes in plant and algal lineages.

Based on the end sequencing of all 40 chromosomal termini completed here, all of the subtelomeric regions in the C. merolae nuclear genome have been unveiled (Figure 3). Eight types of homologous DNA elements (up to 20 kbp), with unique gene contents, were found adjacent to the telomere repeats in the 40 chromosomal ends (Figure 3). Two of the eight types exist in only two ends; element H for 3L (left end of chromosome 3) and 16R, and element L for 3R and 11R. PH elements are the most frequent and are recognized in 6L, 7R, 8L, 8R, 11L, 14R, 18R and 19R, whereas nine termini (namely, 2L, 6R, 12R, 13L, 13R, 15L, 15R, 16L and 20R) showed no homology to other terminal regions. This situation is different from that of S. cerevisiae, in which core X elements exist in all chromosomal ends 24. In the ultrasmall, obligate intracellular parasite E. cuniculi (Microsporidia) 5 and the nucleomorphs (the reduced nuclei of the eukaryotic endosymbionts) of the cryptophyte Guillardia theta and chlorarachniophyte Bigelowiella natans 252627, both of the subtelomeric regions of all chromosomes harbor rRNA genes. However, such features are not recognized in Cyanidioschyzon (Figures 1, 3) and the ultra-small green Ostreococcus 12, and thus seem to represent some evolutionary convergence resulting from the intracellular parasitism or endosymbiosis in E. cuniculi and the nucleomorphs.

The number of transposable elements in various eukaryotes differs widely 282930, and of all the non-symbiotic and non-pathogenic eukaryotes previously studied, yeast and pufferfish have the fewest: about 3% of the genome 2930. However, transposons constitute only about 0.7% of the C. merolae genome (by assuming that the average gene size of RT- and transposase-related elements are 4 kbp and 3 kbp, respectively). The completed C. merolae genome sequence contains only 26 class I elements (retrotransposons) and eight class II elements (transposons) 31 (Figure 1). The transcripts of three of the 26 retrotransposons contained an intact reverse transcriptase open-reading frame, and a BLASTX search suggested that all 26 of the retrotransposons were most closely related to non-long terminal repeat (non-LTR) retrotransposons. So, although LTR retrotransposons are widely distributed in plants and animals 31, we could not find any of them in the C. merolae genome. All eight of the transposase-related elements of C. merolae were most closely related (E-value = 0.071–6 × 10^-7) to transposase sequences of bacterial transposons. In short, C. merolae is one of the two non-symbiotic eukaryotes with an extremely low abundance of transposable elements (Table 2). In the endosymbiotic reduced eukaryotes, Encephalitozoon cuniculi and nucleomorphs 5252627, however, transposable elements seem lacking in chromosomes.

In contrast to this, 253 copies of a novel interspersed repetitive element were found in the C. merolae genome. All copies have a truncated ORF that is weakly related (BLASTX E-value = 10^-5-4 × 10^-2) to a putative protein, WSV486, that is encoded in the genome of the shrimp spot syndrome virus 32. These repetitive elements have an average size of 3.2 kbp, are distributed randomly on all chromosomes, and altogether comprise about 5% of the genome. Because these elements exhibit transcriptional activity 9, they may contribute to genomic or cellular functions in C. merolae in the same manner as repetitive DNA does in other eukaryotes 14.

Previously constructed C. merolae BAC clones 9 were used for filling in previously existing gaps between contigs. PCRs of BAC clones containing unsequenced regions were carried out using Taq polymerase with GC buffer (Takara LA Taq; Takara Bio Inc., Osaka, Japan) and specific primers complementary to sequences flanking the gaps. DNA walking annealing control primer technology (DNA Walking SpeedUp™ Kit; Seegene, Seoul, Korea) was also used to directly amplify unknown sequences adjacent to known sequences within a contig. PCR products were sequenced by cycle sequencing (Big-Dye Terminator Cycle Sequencing Kit v 3.1; Applied Biosystems, Foster City, CA, USA), except for a single gap in chromosome 10, which was filled by a sequencing reaction performed using in vitro transcription (CUGA Sequencing Kit; Nippon Genetech Co., Ltd., Tokyo, Japan).

PolyC-tailing and the anchor primer method, the inverse PCR method and the asymmetric PCR method were used for sequencing the ends of chromosomes. For details, see Additional files 1 and 2.

Assembling of sequence data and two strategies for gene prediction have been described previously 9.

To determine the complete sequences of the histone cluster area in the C. merolae genome, we carried out NotI and ApaI subcloning of the BAC clone GESZ2-b20, which included possible histone clusters in chromosome 14 9. Restriction-enzyme analysis, Southern blot analysis with histone-related probes and end sequencing of the subclones revealed relative positions of the subclones on chromosome 14 (Figure 2). Six gaps between contigs/fragments in the histone cluster area 9 were filled using primer walking of the subclones. For details, see Additional files 1 and 4.

The 100% chromosome sequences are accessible under the DDBJ with accession numbers AP006483–AP006502 (chromosome 1–20). Sequences and annotation are available at http://merolae.biol.s.u-tokyo.ac.jp/.