I. Introduction

I-i. Apicomplexan parasites and their mitochondria as drug targets

The spread of infectious diseases has long been a big worldwide problem. Despite the generations of efforts in preventing the spread, millions of people die in a year. Malaria is one of the most devastating infectious diseases, which is majorly prevalent in tropical and subtropical areas. The malaria parasites Plasmodium spp. are members of the phylum Apicomplexa, a large eukaryotic assemblage comprised of unicellular parasites. This phylum encompasses so many causative agents of human infectious diseases such as Toxoplasma and Cryptosporidium, which occasionally cause fatal illnesses. It is an urgent task to develop effective chemotherapy for these parasitic diseases, so that finding potential drug targets is highly needed.

The mitochondrion is a eukaryotic organelle functioning as the center of energy metabolism in cells. Especially, the mitochondrial electron-transport chain (mtETC) directly contributes to the energy production by forming the proton gradient which drives the ATP synthase. Intensive studies have revealed that the mtETC of apicomplexans are divergent and therefore can be promising drug targets. For instance, the cyanide-resistant alternative oxidase (AOX) in Cryptosporidium is sensitive to ascofuranone, and atovaquone specifically binds to the complex III of Plasmodium mtETC, interfering with the electron-flow. Another notable feature of apicomplexan mitochondria is observed in their genomes. Generally, apicomplexan mt genomes are 6-8kb long linear molecules, the known smallest mt genomes. This mt genome element contains only three protein-coding genes and unusually fragmented rRNA genes, and the protein-coding genes seem to lack canonical start and stop codons. Given the importance of the gene products and the probable difference in the expression mechanism between parasites and mammalian hosts, they can also be drug targets. As there remain several problems against the establishment of chemotherapy, such as the emergence of drug-resistant strains, it is necessary to further characterize the parasite's mitochondria and examine their potential as drug targets.

I-ii. Perkinsus, a shellfish pathogen closely related to apicomplexans

Some of the characteristic features of the apicomplexan mitochondria are shared in the related lineages. In particular, the sister lineage dinoflagellates share some features seen in mt genomes with apicomplexans. Researchers have recently paid attentions to the ancestral species of apicomplexans and dinoflagellates, as it can be anticipated that a comprehensive understanding on physiology of many apicomplexan and dinoflaegllate species would be obtained. One closely related organism to apicomplexans is a shellfish pathogen Perkinsus, which is branched between the two groups. This close relationship, along with the similarities in the ultrastructure and life cycle, implies that Perkinsus shares several biological features with apicomplexans and is expected as a model organism that would characterize mitochondria of apicomplexans and dinoflagellates. It is also notable that Perkinsus can be maintained in host-free in vitro culture and sufficient amount of proteins for biochemical assays are easily obtained. Accordingly, it is of great significance to characterize the mitochondrial energy conversion system including the mtETC functions and mt gene expression system of Perkinsus, from an evolutionary viewpoint.

In such backgrounds, I started to investigate mitochondria of Perkinsus. As almost all current studies on Perkinsus do not focus on mitochondria, the main purpose of this study is to create a platform to characterize Perkinsus mitochondria. I attempted to i) establish a preliminary preparation protocol for Perkinsus mitochondria, ii) outline the Perkinsus mtETC by in silico analyses and biochemical experiments, and iii) determine and analyze the mt gene sequence.

II. Attempts for preparation of enriched Perkinsus mitochondria

Initially, I attempted to establish an enrichment protocol for Perkinsus mitochondria. To mildly disrupt the cells, I tried chemical disruption methods using digitonin in the presence or absence of cellulase, as the Perkinsus cell is surrounded by the cell wall. But these methods did not lyse the cell membrane and cell wall, unfortunately. Among several physical disruption methods tested, the N2 cavitation method was utilized, which is also employed in cell homogenization of Plasmodium to obtain mitochondria. In preliminary experiments, Perkinsus cells were disrupted by incubation with N2 gas dissolved at 300 psi at 4°C, and the crude organellar fraction was subsequently ultracentrifuged at 23,000 rpm for 1 h. The resultant sample was separated into 24 fractions by peristaltic pump. Several fractions clearly showed the activity of the mtETC complex II, succinate dehydrogenase (SDH). This is the first biochemical evidence of mitochondria with an active mtETC in Perkinsus. Among them, fractions Nos. 12-14 showed higher total and specific SDH activities, indicating that these fractions contained more mitochondria. For a purification method, I need to modify and improve the protocol. The specific distribution of mitochondria will be endorsed by co-distribution of other mtETC enzyme activities and by immunoblotting assay. It should be ensured that other organelles like plastids and endoplasmic reticulum are not co-distributed with mitochondria. Also, the intactness of mitochondria has to be ensured for density-based fractionation. For an easier method to see if the cells are broken, I am now establishing transformed cell lines which have green fluorescent protein signal in mitochondria. After these refinements, it will surely contribute to the specific studies of Perkinsus mitochondria.

III. Genomic and biochemical characterization of Perkinsus mtETC proteins

To outline the Perkinsus mtETC, genomic and biochemical analyses were conducted. Biochemical assays detected activities of the mtETC complexes II, III, IV and AOX, and their combined activities as well. Then BLASTp-based sequence searches were performed in public database using mtETC protein sequences from related species as queries. As a result, I revealed that Perkinsus possessed homologs for main subunits of the mtETC complexes II-IV and the FOF1-ATP synthase with high sequence similarity to those of related organisms. Curiously, I did not find homologs for CybL, CybS (complex II), cytochrome b (complex III), cytochrome c oxidase subunit 1 (COX1), and COX3 (complex IV). In spite of their absence, the activities of SDH and succinate-ubiquinone reductase (SQR) which were attributed to the complex II were clearly observed by biochemical assays. This indicated the presence of the CybL and CybS, which are possibly divergent as suggested for Plasmodium. Similarly, the activities of the complexes III and IV were demonstrated although the three important subunits were missing. Importantly, these three genes are encoded exclusively in mt genome in apicomplexans and dinoflagellates. It was proposed that these genes are highly divergent, or that Perkinsus mt genome sequences were just not added to the public database. Additionally, I found homologs for unconventional enzymes: the type II nicotiamide dinucleotide dehydrogenase (NDH2) observed in many apicomplexans, and the AOX. The apparent ratio of AOX contribution to the quinol oxidation was around 40%. As a whole, I confirmed here that the Perkinsus mtETC contained conventional complexes II-IV, the alternative type of NDH, and the additional terminal oxidase AOX, with detectable activities. The presence of alternative enzymes supports the similarity in the mtETC organization between apicomplexans and Perkinsus.

IV. Analysis of Perkinsus mitochondrial gene and its specific expression system

IV-i. Determinatoin of the full-length mRNA sequence of P. marinus cox1 and its genomic localization

It is intriguing that the size and the organization of the mt genome vary greatly among the eukaryotic taxa, and the mt gene expression system is more subjected to code change like codon reassignment and recoding, than nuclear system. As mentioned before, apicomplexans possess highly unusual mt genomes and the gene expression system remains to be characterized. Interestingly, dinoflagellate mt genomes encode the same set of genes as apicomplexans, despite that they are composed of heterogeneous DNA molecules. Since Perkinsus mt gene sequences were not found at all by BLAST-based sequence searches in public database, I tried to obtain the P. marinus cryptic mt genes by molecular biological approaches. I discovered two tiny AT-rich fragments in P. marinus contigs in public database showing high predicted-amino-acid identity with the partial COX1 sequence of a dinoflagellate. Based on the features of flanking sequences, these cox1-like fragments were found to be the remnant DNA that had been transferred from mt genomes into the nuclear genome. As it was anticipated that these fragments were similar to the mitochondrial counterparts they had originated from, I prepared PCR primers based on these sequences and amplified the inner part of the authentic mitochondrial cox1 from P. marinus. After sequencing, I amplified and determined the full-length cox1 mRNA sequence by performing RACE. This mRNA was 1434 nt long, and the overall AT content was 80.9%. As a whole, this gene was similar to cox1 of apicomplexans and dinoflagellates with an E-value less than 10-70; hereafter, I refer to this sequence as Pmcox1.

To determine the genomic localization of Pmcox1 and to infer the Perkinsus mt genome structure, I conducted Southern hybridization using total P. marinus DNA. Probes for the nuclear large subunit ribosomal DNA hybridized to the stacked, high molecular-weight, chromosomal DNA in the uncut DNA sample. In sharp contrast, Pmcox1 signals constituted a smear in the low molecular-weight region (<10kb) of uncut genomic DNA, which is far lower than the expected position for chromosomal DNA. The smear signals suggested that Pmcox1 resides on relatively small, heterogeneous non-chromosomal DNA. These hybridization data are congruent with previous reports on other dinoflagellates, suggesting that Pmcox1 is encoded on multiple heterogeneous DNA molecules, which is similar to the structure inferred for dinoflagellate mt genomes.

IV-ii. Frameshifts at all AGG and CCC codons suggested for translation of Pmcox1

Pmcox1 lacked canonical start and stop codons in terminal regions, similar to apicomplexan and dinoflagellate mt genes. This indicates that Perkinsus also utilizes alternative mechanisms for initiation and termination of translation in mitochondria, as inferred also in the apicomplexans and dinoflagellates. Unexpectedly, this mRNA was not translated in a single reading frame with standard codon usage; several stop codons appeared in all three frames. By BLASTx-based searches, I identified eleven partial COX1-like amino acid sequences that appeared separately in all three reading frames. Most intriguingly, the conserved AGGY (8 sites) or CCCCU (2 sites) motifs appeared to coincide with the transitions between the coding-blocks, and AGG and CCC always appeared in-frame preceded by the predicted COX1-coding blocks (Figure 1). Based on these observations, I postulated a modified decoding which could shift the reading frame during translation: specific frameshifts at every in-frame AGG and CCC. Accordingly, I prepared a putative PmCOX1 amino acid sequence in the following manner. I eliminated the A residues of the AGGY motifs and made a +1 frameshift, making GGY instead of AGG in-frame. I also deleted the first two C residues of CCCCU motifs and made a +2 frameshift, making CCU instead of CCC in-frame. Along this manner, all the 11 "blocks" were connected into one consecutive coding sequence.

This predicted PmCOX1 sequence based on my frameshift model retains the conserved residues essential for COX functions, thereby it reinforces the validity of the frameshift model. The PmCOX1 sequence also conserves the glycine and proline residues, which are most common in the proximity of the AGGY and CCCCU motifs, respectively. Moreover, the highly conserved tryptophans coded by UGA were deduced only with the frameshift model for PmCOX1, which is often observed in mt genes of other organisms. The frameshift model is further supported by the conservation of all the ten frameshift motifs in the cox1 ortholog from another Perkinsus species, P. olseni. Besides, the cob-like fragment found in the P. marinus whole-genome shotgun assembly contained five in-frame AGG which would induce +1 frameshift to connect discontinuous COB-like fragments. Considering the absence of such frameshift motifs in Perkinsus nuclear genes, it was assumed that an unconventional event occurred during translation exclusively in mitochondria of Perkinsus species. Notably, this is the first evidence of a frameshift-involving translation system in protist mitochondria.

IV-iii. Possible mechanisms for the unusually frequent frameshift in translation of Perkinsus mt genes

I propose two possible mechanisms for this unusual frameshift in Perkinsus mitochondrial translation system. The first one is a ribosomal frameshift observed in a wide range of organisms during translation. In the case of +1 ribosomal frameshift, observed also in some mt genes, a rarely used codon or a stop codon in the ribosome A site is suggested to induce the ribosome to stall and allow the reading frame to be subsequently shifted forward by skipping 1 base. Based on previous studies, I hypothesized that ribosomes in Perkinsus mitochondria skip the A residue in the first position of the in-frame AGG in the AGGY motif and the first two C residues in the CCCCU motif by shifting forward by one base at in-frame CCC (Figure 2A). Alternatively, specialized tRNAs that recognize non-triplet codons may be utilized at frameshift sites. Naturally occurring deviant tRNAs recognize four-base codons and act as suppressors of nonsense mutations, and artificial tRNAs bearing modified loops can recognize quadruplet and even quintuplet codons. In the case of Pmcox1, specialized tRNAs may recognize AGGY (for glycine) and CCCCU (for proline) to enable the proposed frameshifts (Figure 2B).

Regardless of the mechanism, it should be noted that the 100% frameshift efficiency has never been observed so far. Lower frameshift efficiencies are not lethal to organisms with frameshift-dependent genes, because there is only one or at most two frameshifts per one gene. In contrast, frameshift must occur at as many as ten sites to produce a complete COX1 protein in Perkinsus, which is by far the highest frequency among the relevant studies. With only one failure in any of ten frameshifts due to low efficiency, a truncated PmCOX1 protein will be synthesized which will affect the respiration. The frameshift-promoting elements like upstream Shine-Dalgarno-like sequences or downstream pseudoknot structures were not found around the frameshift sites in Pmcox1. Based on these observations, I suggest that the complete translation of Pmcox1 requires a quite accurate mechanism for highly frequent and efficient frameshifts. I will analyze the actual PmCOX1 sequence, and also investigate the translation machinery in Perkinsus mitochondria to understand the mechanisms allowing these "extensive" frameshifts.

V. Conclusions and perspectives

This study is a pioneering work on Perkinsus mitochondria, and contains two major important points with great scientific significance. The first point is that I created the foundation for studying the Perkinsus mitochondria. I outlined the Perkinsus mtETC by genomic and biochemical approaches, and determined and characterized the first Perkinsus mt gene sequence. I demonstrated the similarities in the mtETC organization and mt gene expression machinery between Perkinsus and apicomplexan parasites. These observations, along with the ease in handling, will prompt us to consider that Perkinsus may be a highly suitable material to investigate the properties of mtETC enzymes and the mt gene expression machinery of apicomplexans, both of which are the candidates of antiparasitic drug targets.

The second is that I discovered an unusual phenomenon: extremely frequent frameshifts in translation. Identifying the components employed in translation like tRNAs and ribosomes in Perkinsus mitochondria would delineate a specific molecular mechanism which extends our general view on decoding and might give rise to invaluable insights into the essence of translation. Furthermore, considering its phylogenetic position and the threat to fisheries, this work and future related studies will be informative also to the evolutionary protistology and fisheries sciences. As a conclusion, this study encompasses basic and essential information in studying mitochondria of Perkinsus as a relative of apicomplexans, and presents a discovery of an unusual and tantalizing phenomenon, making contributions to a wide range of scientific fields.

Figure 1. Schematic model of Pmcox1 mRNA and the COX1-like blocks. COX1-like amino acid sequences distributed across three reading frames are represented by gray boxes labeled with Roman numerals. Red and blue bars at the end of COX1-like blocks indicate AGGY motifs containing in-frame AGG, and CCCCU motifs including in-frame CCC, respectively. The nucleotide sequence and its three frame translation around the junction of the coding blocks I-III are described, with AGG and CCC codons colored in red and blue, respectively.

Figure 2. Possible mechanisms of frameshift-dependent translation at AGG and CCC codons of Perkinsus mt genes.

本研究はマラリア原虫をはじめとする寄生性生物群アピコンプレックス類とその近縁系統群に着目しており、細胞内のエネルギー転換の中心であるミトコンドリアについて生化学的および分子生物学的解析を行った。本研究では系統学的にアピコンプレックス類とその姉妹系統群である渦鞭毛藻類との中間に位置する貝類寄生虫Perkinsusを用いて、そのミトコンドリア電子伝達鎖(mtETC)とミトコンドリア遺伝子(mt遺伝子)発現系の解析を行い、下記に示す結果を得ている。

1. Perkinsusミトコンドリア濃縮画分を得るため、P. marinus培養細胞を様々な手法を用いて破砕し、Perkinsusミトコンドリアの調製法の条件検討を行った。その結果、ガス平衡化圧を300 psiに調整したN2 cavitation法とパーコール密度勾配超遠心法を組み合わせることにより、mtETC複合体であるコハク酸脱水素酵素(SDH)の高い活性を示す画分が得られた。これは酵素活性を有するmtETCがPerkinsus細胞に存在する初めての生化学的証拠である。さらに、Perkinsusミトコンドリア濃縮画分を得る予備的手法を確立したといえる。

2. PerkinsusのmtETCに存在すると考えられる様々な酵素の活性を生化学的手法により測定したところ、mtETC複合体II、III、IVの明らかな活性が確認され、アピコンプレックス類Cryptosporidiumに含まれるAlternative oxidase(AOX)の活性も検出された。また、暫定公開されているP. marinusゲノムデータベースにおいて近縁生物のホモログ配列を用いたBLAST検索を行った結果、AOXや、多くのアピコンプレックス類が有するタイプII NADH脱水素酵素のような特殊な酵素を含む、mtETC複合体の主要サブユニットのホモログ配列が見つかった。このことから、PerkinsusのmtETCはアピコンプレックス類のそれと類似していることが示された。

3. P. marinusのゲノムデータベースにおけるBLAST検索ではPerkinsusのmt遺伝子と考えられる配列は見つからなかった。そこで、PCRおよびRACEを行うことで、Perkinsusのmt遺伝子としては初の報告となる、cox1のmRNAの完全長配列(1434 nt)を決定した。近縁生物のmt遺伝子と同様に、本遺伝子の末端領域には標準的な開始コドン・終止コドンが見出されなかった。またサザンハイブリダイゼーションの結果から、本遺伝子は数kb程度の比較的小さなDNA分子の集合体として形成されている、渦鞭毛藻類のミトコンドリアゲノムに類似した構造のDNA分子にコードされていることが示された。

4. cox1の塩基配列と推定アミノ酸配列を注意深く比較することによって、本遺伝子の翻訳時にAGGコドン(8箇所)とCCCコドン(2箇所)において+1フレームシフトが起こっていることが示唆された。これらのコドンを含むフレームシフトモチーフは近縁種P. olseniのcox1でも完全に保存されていた。また、P. marinusゲノムデータベースから見出されたもう1つのmt遺伝子cob様配列についてもAGGコドン(5箇所)でのフレームシフトを適用することで、近縁生物ホモログに極めて類似したアミノ酸配列が推定された。以上の結果から、Perkinsusのmt遺伝子発現特異的にこのフレームシフトが起こっていると考えられた。このようなフレームシフトは既に幾つかの生物から報告されているが、1遺伝子に10回という頻度は異常に高頻度であり、例えばコドン認識能が特殊化したtRNAが関わるような、極めて効率の高いフレームシフト機構が存在すると考えられた。

本研究ではまず、PerkinsusのmtETCが近縁生物群アピコンプレックス類のそれと類似した特徴を持つことが明らかとなった。Perkinsusは純粋培養系が確立しており扱いが容易であり細胞収量も高いことから、培養に宿主を必要とするアピコンプレックス類よりも生化学的解析に適した生物である。今後、本研究を基盤としPerkinsusを用いた研究が進められることによって、アピコンプレックス類と近縁生物の持つミトコンドリアの機能についてさらなる知見が得られると期待できる。さらに本研究では、貝類寄生虫Perkinsusのmt遺伝子発現系において高頻度なフレームシフトという特徴的な現象が起こっていることが示された。この分子機構を解明することは、遺伝子発現における読み枠の維持という、全生物に共通する機構の理解に大きく貢献すると考えられる。以上のように本研究は様々な側面において学術的貢献が期待でき、学位の授与に値するものと考えられる。