Introduction

Major challenge of microbial oceanography is the elucidation of the role of microorganisms at the biogeochemical cycles of marine environments. It is known that about 30% of primary production of the ocean is performed by microorganisms. They participate in the marine carbon cycles by the consumption (degradation and absorption) or remineralization of organic matter. Marine organic matter is divided into particulate organic matter (POM) and dissolved organic matter (DOM) according to the size. Regarding the bacterial strategies in order to use organic matter, mostly two were explained; particle-attachment to POM or free-living to DOM. Interestingly, recent reports revealed that dissolved organic matters consist of various forms including transparent gels, and colloids, particles, and mucus sheets and bundles, which formed by aggregation of transparent gels. Even particles in the ocean are the smaller, the more abundant, and which are absolutely important to biogeochemical cycles in marine environments. Such small organic matters should be tasty food for bacteria and bacteria would degrade or modify them to acquire favorable organic matters for themselves. And it is necessary to define the new strategy of bacteria utilizing specific dissolved organic matter such as submicron particles, which cant be directly transported through membrane.

Through the thesis, I hypothesized that bacteria may possess submicron particles around their surface, which supposed to promote effective consumption of those particles by increasing the density of organic matter around them, and these processes were named "particle-capture", and the bacteria those have the ability of particle capture were named "Particle-capture bacteria (PCB)".

Promising objectives of the thesis are 1) the method evaluation of the separation of PCB using magnetic particles, 2) the direct observation of PCB using atomic force microscopy (AFM), and 3) the characterization of PCB community using molecular approaches (degenerating gradient gel electrophoresis, DGGE).

Materials and Method

Major methods used in this thesis are followings.

Sampling. For basic and methodological examinations, seawater samples collected from Tokyo Bay were used. Seawater samples were also obtained from Tokyo Bay, Sagami Bay, and offshore environments during the KT-05-16 cruise of R/V 'Tansei Maru' (Ocean Research Institute, The University of Tokyo, and Japan Agency for Marine-Earth Science and Technology (JAMSTEC)). Stations S1, S2, and S3 were located in the southern, middle, and northern regions, respectively, of the Kuroshio Current.

Magnetic separation. The magnetic separation method using biodegradable paramagnetic particles was examined. 1-ml seawater samples were transferred to 1.5-ml microcentrifuge tubes, and magnetic separation was initiated immediately by the addition of 20 μl of the particle suspension. The mixtures were incubated at room temperature at 10 rpm for 1 h on a sample mixer. The paramagnetic particles were then collected with a magnet and the supernatants were removed carefully. The particles were rinsed twice with artificial seawater and frozen until analysis.

Community structure analysis. Nucleic acids were extracted by using the commercialized kits. For RNA sample, cDNAs were synthesized by reverse-transcription. PCR was conducted from DNA samples and RNA-derived cDNA samples, with primers those amplify DNA fragments including V3 region of 16S rRNA with GC clamp, GC rich sequences for the DGGE analysis. Amplicons were separated by their sequences by DGGE and the banding pattern was visualized by statistical analysis. Sequencing analysis of some prominent bands excised from DGGE gels was also conducted.

AFM analysis. Bacterial cells collected from Tokyo Bay, Sagami Bay and the offshore stations were observed by AFM. Bacterial cells were differentiated from non-living particles on the basis of their size, shape and cross-section. Cells with particulate materials around them were counted, and the relative number of particle-possessing cells among the total number of bacterial cells was determined.

Results

Evaluation of Magnetic Separation. Magnetic separation method was evaluated in order to collect bacteria possessing particle-capture ability (particle-capture bacteria, PCB). Reliability was confirmed through community structure analysis. First trial to collect PCB from natural seawater was resulted that community structure of triplicate of magnetically separated samples were agreed well and were different from those of natural seawater. Close inspection into the phylotypes of magnetically separated showed α-, β-, and γ-Proteobacteria, Actinobacteria, and CFB group (Cytophaga-Flavobacterium-Bacteroides) phyla. Interestingly, several prominent bands of submicron-sized (130 nm) particle samples・were closely related to Roseobacter isolates. And a band, which belongs to the CFB group, appeared in all magnetically separated samples. It was showed that the abundances of magnetically separated bacterial cells were constant at approximately 10% of total bacteria throughout the investigations.

Secondly, incubation time with model particles were tested whether affect to the community structure of magnetically separated samples. Results showed that community structure and abundance of magnetically separated samples were mostly consistent from 1h to 8h and greatly shift around 24 h incubation suggesting 1h-incubation would be proper for efficient experimental procedures.

Spatial distribution of bacteria possessing submicron sized particles on their surface. Bacterial cells collected from samples from both coastal and open ocean environments and concentrated on Isopore filters were observed under AFM, and those surrounded by particulate materials were counted. The sizes and numbers of particles on each cell were variable, and it was difficult to distinguish any general trends. The relative numbers were high in the inner part of Tokyo Bay, but few were detected in the surface layers at other stations. Among the offshore stations, the ratios were relatively high in the middle layer (500 to 2000 m) and declined with depth.

Effect of various size and substances of particles. Community structure analysis to see the changes on different types of particles to associated were performed using 8 types paramagnetic particles; 130nm, dextran; 150nm, silica; 250nm, dextran; 300nm, silica; 500nm, dextran; 500nm, silica; 6um, silica; 6u, polylactic acid. As consistent with above results, community structures of magnetically separated samples were different from those of natural seawater. And also among magnetically separated samples of different particles, it was clearly shown that community structures were changed size- and substance-dependently. Submicron sized samples (130nm to 300nm) were shown consisting similar community structure, while they were different from large particle (6um) samples. Regarding on substance of particles, same substance samples were shown consisting similar community structures although size of particles they associated was slight different, 130nm and 250nm of dextran, 150nm and 300nm of silica, respectively. These results were consistent among Yokohama Port and Shinagawa Port samples.

Interesting thing was 500nm particle samples. Patterns of community structure shifts were different between Yokohama Port and Shinagawa Port; at Yokohama Port, 500nm samples were presented middle of other submicron samples and large particle samples, while at Shinagawa Port, they were close to large particle samples.

Effect of metabolic activity. DNA- and RNA-derived DGGE banding patterns were compared between submicron sized (130nm) particle samples and large (1um) particle samples from the'total'seawater and the 3um filtrate, at Sinagawa port, Yokohama port, and Aburatsubo port, respectively. DNA-derived profiles showed consistently clear differences between submicron and large particle samples through 3 different stations. While, RNA-derived profiles varied among stations, showing inconsistent different patterns between submicron and large particle samples; at Yokohama station, 1um samples and the 130nm sample from total seawater and 1um samples from 3um filtrate were plotted closely together, while the 130nm sample from 3um filtrate seawater were plotted apart. At Sinagawa station, all samples were plotted apart regardless the size of particles and sampling fractions. At Aburatsubo station, except the 1um sample from total, all samples were plotted close.

Phylotypes assumed as key members who drove above results were suggested from sequencing analysis. Members that only appeared among RNA level were closely related to members of α-proteobacteria; Metholybacter and Xanthobacter and uncultured γ-proteobacteria.

Effect of ionic interaction between PC bacteria and particles. Effect of ionic interaction on PC was investigated performing experiments adding chelating agent, EDTA (ethylenediaminetetraacetic acid). EDTA volume was adjusted from preliminary experiments using famous model marine bacteria, Vibrio parahaemolyticus, and added to natural seawater. Increasing the concentration of EDTA, the relative number of PC bacteria decreased subsequently. Additional experiments about the effect of flagella on PC were performed. 3 types of mutants were used; YM4, no flagella; YM18, polar flagella, YM19, lateral flagella. Relative number of PC was high from YM19. And other 3 strains including wild type was shown similar % of PC to the total abundance.

Discussion

In aquatic environments large numbers of submicron particles are generally present, and their turnover and fate are of considerable ecological importance in biogeochemical cycles in marine environments. From the results of this thesis, it was suggested that some bacterial community would capture submicron sized particulate organic matter such as submicron particles (SMPs). Their abundance was about 10% of total bacterial abundance of surface water and relatively high in the inner part of Tokyo Bay and at depths of 500 to 2000 m in the open ocean. Mesopelagic to bathypelagic is known as the last frontier in the earth and recently several international big projects are conducting in order to investigate biogeochemical process to solve a riddle on the marine carbon cycles. And there are known as sinking particles are abundant and bacterial respiration rate and exoenzymatic activity are high. With high relative abundance of particle-capture bacteria, possibilities were that particle-capture bacterial community might play important role in the biogeochemical cycles around mesopelagic to bathypelagic.

And it was also suggested that particle-capture ability might be the fundamental way for survival of marine bacteria. Particle-capture bacterial community were distributed phylogenetically broadly among α-, β-, γ-proteobacteria, CFB group, Actinomyces, Firmicutes, etc. Close inspection into phylotypes of PCB suggested that particle-capture as a fundamental way for survives. For example, Roseobacter sp., which is the typical free-living bacteria, may be necessary to uptake nutrients from DOM. And also in the case of couple of members of CFB, they live close to the phytoplankton may capture submicron particulate matter leaked from upper trophic. Comparison between DNA and RNA level community structure analysis revealed that metabolic activity should be hardly related with particle-capture by bacteria suggesting other physiochemial factors might affect particle-capture. Then I tested whether ionic interaction between bacteria and SMPs would consist of particle-capture process using chelating agent. Results showed that relative particle-capture bacterial number decreased as chelating agent added, which might suggest ionic interaction between bacteria and SMPs would consist of particle-capture. That is, surface charge of bacteria that might be by nature or affected by surroundings, possibly would be one of important factor for particle-capture. Additional experiments suggested that surface features which bacteria possess by nature, might affect PC such as lateral flagella and ionic status. In summary this thesis proposed that:

1. The method of collecting particle-capture bacteria was developed in this thesis, and suggested the novel concept for bacteria to utilize submicron particles.

2. Direct observation by AFM revealed the presence of particle-capture bacteria, and particle-capture bacteria is dominant around mesopelagic and bathypelagic. In the context of biogeochemical cycles of marine environments, it is suggested that particle-capture bacteria may play an important role in the marine carbon cycles around deep sea.

3. Community structure analysis of particle-capture bacteria from DNA and RNA derived samples suggested physiochemical properties might be the important factors for possessing particle-capture ability. In the context of microbial oceanography, particle-capture may be the ability by nature of bacteria

本論文は6章からなり、第1章は序論、第2章は微粒子捕獲細菌の観察及びその空間的分布の観察、第3章は、磁性分離法による微粒子捕獲細菌の分離、第4章は微粒子捕獲細菌と代謝活性の関係、第5章は微粒子捕獲における物理及び化学的要因、第6章は総合考察について述べられている。それぞれの章の概要は以下の通りである。

第1章 序論

海洋には、大きさが10 nm~1μmの微粒子(サブミクロン粒子)が細菌数の10~100倍の数存在し、細菌によるその代謝は海洋での物質循環を理解する上で重要である。しかし、海洋細菌とサブミクロン粒子との相互作用は明らかにされていない。本論文では、細菌と海洋物質間の相互作用が行われる細菌表面周辺 (Bactsphere)を新しい微生物生態系の概念として提案し、海洋中に多く存在する微粒子を効果的に利用するために海洋細菌がそれらを細胞表面に捕獲するという仮説を立てた。この仮説を証明するため、微粒子捕獲細菌の観察および分離を行い、また、微粒子捕獲のメカニズムを解明することを目的とした。

第2章 微粒子捕獲細菌の観察及びその天然海洋環境における空間的分布の観察

天然海水中の細胞表面に微粒子を保持している細菌を観察し、水平および鉛直分布を調べるため、原子間力顕微鏡を用いて細胞表面に微粒子を保持している細菌を観察・計数した。その結果、微粒子を細胞表面に持っている細菌が海洋で広く分布していることが分かった。特に、沿岸表層及び外洋中深層において多いことが明らかになった。

第3章 磁性分離法による微粒子捕獲細菌の分離

天然海水から微粒子捕獲細菌を分離するため、磁性分離法を確立した。天然海水に磁性粒子をモデル粒子として添加し、磁石を用いて磁性粒子を捕獲している細菌群を分離した。それらの群集組成を変性剤勾配ゲル電気泳動法(DGGE)で解析した結果、特定の細菌群集が再現性よく分離されることが分かった。また、微粒子捕獲細菌群集は天然海水全体の細菌群集または付着細菌群集と異なることが分かった。

第4章 微粒子捕獲細菌と代謝活性の関係

微粒子捕獲細菌の代謝活性を測定するため、磁性分離法により分離された微粒子捕獲細菌の16S rRNAの定量を行い、浮遊細菌の16S rRNAと比較した。その結果、微粒子捕獲細菌は浮遊細菌より相対的に高い代謝活性を持っていることが分かった。また、DNA及びRNAレベルでの群集組成をDGGEで解析した。その結果、DNAおよびRNAレベルで見られる群集組成は異なり、微粒子捕獲細菌群集には様々な代謝活性を持つ細菌群により構成されていることが示唆された。

第5章 微粒子捕獲における要因の解明

微粒子捕獲のメカニズムを解明するため、各種要因(微粒子のサイズ、材質、細胞表面の構造(ベン毛)、2価イオンの存在)と細菌の微粒子捕獲の関係を調べた。異なる微粒子のサイズ及び材質により異なる細菌群集が分離された。また、ベン毛の変異株と野生株における微粒子捕獲能の比較により、ベン毛が微粒子捕獲に影響を与えることが分かった。キレート剤であるEDTAを用い、微粒子捕獲細菌数と群集組成を調べた結果、EDTAの添加により微粒子捕獲細菌数が減少し、群集組成も変化した。以上より、微粒子と細菌間にイオン結合が存在することが明らかになった。

第6章 総合考察

本研究では、海洋環境中での微粒子捕獲細菌の存在を明らかにすることができた。また、微粒子捕獲は、微粒子の大きさや材質、ベン毛が影響し、微粒子と細菌の細胞表面ではイオン結合が存在することが明らかになった。本研究は、細菌と微粒子の相互作用を明らかにした最初の報告である。微生物生態学において、細菌の微粒子捕獲は海洋炭素循環における微生物の役割に対する新たな知見をもたらすと期待できる。

なお、本論文2章、3章は池本栄子、吉田弘明、木暮一啓との共同研究であるが、論文提出者が主体となって分析および検証を行ったもので、論文提出者の寄与が十分であると判断する。

以上の結果を総合し、博士(環境学)の学位を授与できると認める。