Habitability analyses of aquatic bacteria

  • Md. Nurul Haider
  • Masahiko Nishimura
  • Minoru Ijichi
  • Ching-chia Yang
  • Wataru Iwasaki
  • Kazuhiro Kogure
Original Article


Habitability is defined as an ability of an organism to inhabit different environments. Habitability of organisms, however, cannot be inferred from analyses such as a whole genome or community structures. A recently developed database, the MetaMetaDB, gives us information from what kind of environments one particular 16S rRNA sequence data has ever been obtained, and thus enables us to infer the habitability of the bacterium in question. In order to check the applicability of this database to study the habitability of aquatic bacteria, samples collected at two Naka River stations, one estuarine station from Naka River Estuary, two coastal stations at Oarai in Ibaraki Prefecture, Japan and one station in the Kuroshio Current of the western North Pacific were examined. The phylotypes were tracked against the MetaMetaDB and it was reasonably found that the low-salinity stations were dominated by sequences with “freshwater and groundwater”, “human” and “wastewater” habitat identities, while the high-salinity stations were dominated by those with a “marine” identity. The phylotypes of low-salinity stations with a particular habitat identity were absent or rare in the high-salinity stations and vice versa. The MetaMetaDB also showed that sequences of Cyanobacteria or related phylogenetic groups may be present in the human gut, as well as the probable distribution of the relatives (ancestors/descendants/siblings) of some bacteria. These overall findings proved that the MetaMetaDB is useful as a new tool to infer microbial habitability and it gives us new information on the possible origin and ecology of microorganisms in the environments.


16S rRNA 454 Pyrosequencing MetaMetaDB Habitability Aquatic bacteria 

1 Introduction

For maintaining a population size in nature, bacteria have to multiply or inflow at a rate faster than that of the declining rate due to death (as a result of predation, bacteriolysis, etc.) or outflow from the environment. The ability of bacterial cells to adapt to an environment determines this rate and also the habitability, which means the ability to inhabit different environments. The information on habitability is important to understand how apparent community structures in the environment are formed and also how they may vary depending on how they respond to the predominant conditions. However, it is difficult to assume habitability of bacteria by widely used genetic or physiological approaches.

The development of molecular techniques has made it possible to clarify bacterial community structures without cultivation (Acinas et al. 1997; DeLong et al. 1993; Giovannoni et al. 1995; Hiorns et al. 1997). For bacterial community structures analyses, 16S rRNA gene sequences are used for the taxonomical assignment because of the adequate substitution rate to differentiate species, suitable gene size and accumulation of massive data sets in databases, such as the Greengenes (DeSantis et al. 2006), the Ribosomal Database Project (Cole et al. 2009) and SILVA (Quast et al. 2013). Most of them are readily accessible for comparative works. A single database, however, does not offer any data for comparative works among different environments, because each database usually comprises a single environment, such as terrestrial, marine or human. If those databases are somehow combined and collectively used, it is possible to search from which environments one particular sequence has ever been detected and recorded. This offers information of habitability.

The recently developed MetaMetaDB (http://mmdb.aori.u-tokyo.ac.jp/; Yang and Iwasaki 2014) contains data set of 16S rRNA sequences derived from 454 platforms in the DDBJ Sequence Read Archive (DRA). It offers environmental categories that indicate in what kind of environments each sequence was obtained and recorded. Based on this recorded information, a particular sequence can be classified into its most probable habitat, whose variety in a particular sample or in a taxonomic group can be termed as habitability. It is noteworthy that the lack of the record doesn’t mean the absence of the corresponding sequence, and that there may be some biases of habitable environments depending on the relative amount of deposited sequence data. Nevertheless, we believed that this can be a new tool to infer habitability of prokaryotic organisms in natural environments. Hiraoka et al. (Hiraoka et al. 2016) checked the habitability of two soil prokaryotic communities by the MetaMetaDB and showed that the soil affected by the tsunami in 2011 tended to contain more sequences of marine habitat compared with the unaffected soil. More investigations are, however, required to confirm the reliability or usefulness of the MetaMetaDB to prokaryotic populations in nature.

Coastal areas are under the influence of terrestrial, freshwater and offshore marine environments. Bacterial community structures are formed as a result of mixing of those different environments (Crump et al. 1999; Crump et al. 2004). We first hypothesized that salinity is a major factor controlling community structures, second, that the application of the MetaMetaDB may clarify, at least in part, the mixing of freshwater and seawater, and third, that the habitability of those attached to particles and in free-living states are different. Therefore, we collected samples from the Naka River, Naka River Estuary, Oarai coasts (both in Ibaraki Prefecture, Japan) and the Kuroshio Current in the western North Pacific Ocean. For separating particle-associated fraction (PA) and free-living (FL) fraction, seawater sample was filtrated through two filters of different pore size. The purpose of this research was to examine the applicability of the MetaMetaDB to examine the habitability of microbial communities in aquatic environments. We assumed that this database will offer new information on the origin or evolutionary processes of bacterial groups that have been overlooked.

2 Materials and methods

2.1 Sampling sites and sample collection

Surface water samples (depth 0 m) were collected at two freshwater stations, “river 1 (R1)” and “river 2 (R2)”, of the Naka River, one estuarine station, “estuarine (E)” of the Naka River Estuary, and two coastal stations, “Oarai sea side (SS)” and “Oarai port side (PS)”, in April 2015 (Table 1, Fig. 1). The samples from the off-shore surface layer (depth 5 m), “Kuroshio surface (KS)” and chlorophyll maximum depth (depth 50 m), “Kuroshio chl. (KC)”, were collected from the Kuroshio Current of the western North Pacific Ocean during the KH-14-2 cruise of th R/V Hakuho Maru (JAMSTEC) in May 2014 (Table 1, Fig. 1).
Table 1

Summary of different environmental parameters of the sampling stations



Sampling depth

Air temp. (°C)

Water temp. (°C)


Total cell count (× 105 cells mL−1)

Low-salinity stations (salinity 0.5–5.0)

 Riverine 1

36º35′59″N & 140º55′67″E






 Riverine 2

36º34′40″N & 140º57′63″E







36º33′56″N & 140º59′06″E






High-salinity stations (salinity 32–35)

 Oarai sea side

36°31′74″N & 140°59′20″E






 Oarai port side

36°30′99″N & 140°58′46″E






 Kuroshio surface

36º20′49″N & 143º 58′38″E






 Kuroshio chl.

36º20′53″N & 143º 58′29″E

50 m




No data

The samples of the freshwater and brackish water (from river, estuary and coastal) were collected on 10 April, 2015, and Kuroshio Current samples of the western North Pacific Ocean during the KH-14-2 cruise of the R/V Hakuho Maru (JAMSTEC) on 23 May, 2014

Fig. 1

Sampling locations (satellite image: Imagery ©2014 Google, TerraMetrics, Data SIO, NOAA, U.S. Navy, NGA, GEBCO, Map data ©2014 Google, ZENRIN). R1, river 1; R2, river 2; E, estuary; SS, Oarai sea side; PS, Oarai port side; KS, Kuroshio surface; KC, Kuroshio chlrophyll maximum

2.2 Filtration and sample preparation

For the river, estuary and coastal samples, 10 L of water were collected in a sterilized screw-capped plastic bag and carried back to the laboratory in an ice box within 2–3 h after sampling. Two litters of the water samples were filtered through a 0.22-μm pore-sized Sterivex-GP pressure filter unit (Merck, Billerica, MA, USA) for the total (T) bacterial fraction. Another 2 L of water samples were pre-filtered through a 47-mm, 3.0-μm pore-sized Nuclepor polycarbonate track-etched membrane filter (GE Healthcare Life Science, Chiyoda-ku, Tokyo, Japan) for particle-associated (PA) fractions and then through a 0.22-μm pore-sized sterivex filter unit for free-living (FL) bacterial fractions. The samples from the R/V Hakuho Maru were collected by using a Niskin bottle and filtered within 1–2 h after collection onboard. The seawater samples were treated as is stated above. The membrane filters and the sterivex cartridge filters were stored at −20 or −80 °C until further processing.

2.3 Environmental parameters and total cell count

The air and water temperature of the samples from river, estuary and the coast were measured by a mercurial thermometer. Their salinity was determined by a refractometer (IS/Mill-E, As One, ATAGO, Japan). For cruise samples, salinity, water temperature and depth were obtained by using an SBE 911 plus CTD system (Sea-Bird Electronics, Inc., Washington DC, USA). To enumerate the bacterial abundance, the collected water samples were fixed with formalin (2% final concentration) and stored at 4°C in dark until enumeration. One mL sample was filtered onto a 25-mm2, 0.2-µm pore-sized Isopore™ membrane filter (Merck Millipore Ltd., Tullagreen, Carrigtwohill Co. Cork, Ireland), stained with DAPI (4′, 6-diamidino-2-phenylindole) mix solution [(5.5 parts Citiflour (Citiflour), 1 part Vectasheild (Vector Laboratories) and 0.5 parts phosphate-buffered saline (PBS), with DAPI (final concentration 2 μg ml−1)] and examined under an Olympus BX-51 epifluorescence microscope (Olympus Opticals, Tokyo, Japan; Porter and Feig 1980). Approximately 40–60 images were taken from each filter, and total prokaryotic cells in 10 randomly selected images were counted. The total count was calculated as the average of triplicate samples.

2.4 DNA extraction

DNA was extracted by ChargeSwitch Forensic DNA Purification Kits (Invitrogen™, Carlsbad, CA, USA) with slight modifications, using ZircoPrep Mini (FastGene™, Nippon Genetics Co. Ltd., Bunkyo-ku, Tokyo, Japan) for bead beating prior to the lysis process. Bead beating was done using a Micro Smash (MS-100R, Tomy Medico., Ltd., Tokyo, Japan) at 5000 rpm and 4°C for 30 s for each filter with great care to avoid contamination. The extracted DNA was then cleaned with NucleoSpin gDNA Clean-up kit (MACHEREY–NAGEL GmbH & Co. KG, Neumann-Neander-Str., Düren, Germany) according to the manufacturer's protocol and stored at −30 °C until the following treatments.

2.5 Bacterial 16S rRNA gene amplification and pyrosequencing

The hypervariable V1–V3 regions of bacterial 16S rRNA gene was amplified by polymerase chain reaction (PCR) using the forward primer 27F with multiple identifiers (MIDs): 5′-CCATCTCATCCCTGCGTGTCTCCGACTCAGXXXXXXXXXXAGAGTTTGATCMTGGCTCAG-3′ and the reverse primer 519R with adaptor: 5′-CCTATCCCCTGTGTGCCTTGGCAGTCTCAG(GWATTACCGCGGCKGCTG)-3′; where Xs represents the sample-specific multiplex identifier-MID (Kim et al. 2011). PCR reactions were carried out in 20 μL of the mixture consisted of 2 μL of DNA template, 13.1 μL of molecular biological grade double-distilled water, 0.6 μL of each primer at 5 μM, 2 μL of 10 × TaKaRa Ex Taq Buffer, 1.6 μL of TaKaRa dNTP mixture (2.5 mM each), and 0.1 μL of XUnits TaKaRa Ex Taq HS Polymerase (TaKaRa, Kusatsu, Shiga, Japan) in triplicate. Thermal cycling was carried out with the following conditions: initial denaturation at 94 °C for 4 min, 25 cycles of the denaturation at 98 °C for 10 s, annealing at 55 °C for 30 s and elongation at 72 °C for 1 min, and final elongation at 72 °C for 10 min. After amplification, the presence of the desired length of the partial 16S rRNA gene was confirmed by agarose gel electrophoresis and any sort of contamination was carefully verified by observing the bands of the triplicates of the same samples. After confirming the length, PCR products were purified and normalized using Agencourt AMPure XP (Beckman Coulter Inc., Beverly, MA, USA) according to the guidance of the 454 Sequencing Amplicon Library Preparation Method Manual (GS Junior Titanium Series 2012). The purified PCR products were then quantified using Quant-iT™ PicoGreen® dsDNA Assay Kit (Thermo Fisher Scientific, Eugene, OR, USA). After quantification, the PCR products were sequenced using 454 GS Junior sequencer (Roche Diagnostics, 454 Life Sciences Corp., Branford, CT, USA) at the Atmosphere and Ocean Research Institute, the University of Tokyo (Kashiwa, Chiba, Japan), according to the manufacturer’s protocol for the 454 GS Junior Titanium Series.

2.5.1 Sequence data accession number

The raw sequence data were deposited in the DDBJ Sequence Read Archive databases under the accession number DRA004565.

2.6 Sequence analyses

The subsequent analysis, quality checking and arrangement were done using the open-sourced MOTHUR program (Schloss et al. 2009) following the guidelines available in the operation manual for the 454 (http://www.mothur.org/wiki/454_SOP). The separately run and obtained data files were fused together after removing the tags and primer sequences, and after trimming (qwindowaverage = 35, qwindowsize = 50, minlength = 200). For the next step, the data set was made more workable by selection of only the unique sequences. Then, similar sequences were aligned using the “silva.nr_v119.align” file as reference (silva.nr_v119; Pruesse et al. 2007). Sequencing errors were further reduced by screening, filtering and de-noising through the pre-cluster method (Huse et al. 2010). The chimeras were checked and removed using chimera.uchime (Edgar et al. 2011). In order to improve the data quality, sequences were subsequently classified using the Ribosomal Database Project (Maidak et al. 1996) reference files, and the inactive components, such as the chloroplast, mitochondria and organelles affiliated with “former” bacterial sequences, were removed using the “remove.lineage” command from our dataset. The qualified high-quality sequences were used to generate the distance matrix and for assigning of clusters to operational taxonomic units (OTUs) at a 97% identity level (Schloss and Westcott 2011). A representative sequence from every OTU was used for classification by running the MOTHUR program based on the “silva.nr_v119.tax” file (silva.nr_v119). To standardize the number of reads sequenced between samples, they were randomly re-sampled according to the sample with the fewest reads (2838 reads) using the MOTHUR program; this was done based on OTU files clustered at a 0.03 cut-off level.

2.7 Tracking the habitability of the identified representative sequences

To test the habitability, the obtained 16S rRNA gene sequences were tracked against the MetaMetaDB and gathered information on their habitability above 97% (species level), 95% (genus level), 90% (family level), 85% (order level) and 80% (class level) of identity as default output of the database according to Kirchman (Kirchman 2012). In order to verify the robustness and any sort of biases, the same sequences were verified against both the latest (data by November 6, 2014) and the previous version (data by March 19, 2014). The latest version of MetaMetaDB contains 2,949,852 representative 16S rRNA sequences from 61 diverse environments while the previous one contained 2,737,833 representative 16S rRNA sequences (http://mmdb.aori.u-tokyo.ac.jp/download.html; Yang and Iwasaki 2014).

The MetaMetaDB generates a list of hits in each level of identity and the significant hits are considered to determine the microbial habitability indices (MHIs; Yang and Iwasaki 2014), as a set of ratios of environmental categories. According to Yang and Iwasaki (Yang and Iwasaki 2014), MHIs can be calculated by the following equation
$${\text{MHIc}}^{\left( e \right)} = \frac{{Nc^{\left( e \right)} { \log }\frac{{R_{total} }}{R\left( e \right)}}}{{\sum e\left( {Nc^{\left( e \right)} { \log }\frac{{R_{total} }}{R\left( e \right)}} \right)}},$$
where e is environmental category and c is identity threshold; Nc(e) is the number of hits that are marked by e and above identity c; Rtotal is the total number of sequences in the database and R(e) is the total number of sequences marked by e in the database. The denominator of this equation is a summation of the hit number weighted by the logarithmic weighting factor, so it is used to make MHI a value between 0 and 1. For a particular phylotype, the larger the MHI value for an environmental category, the higher the possibility to be found in that environment. The most probable (adapted) habitat of a particular OTU means the environmental category which had the largest MHI value in case of multiple significant hits in this study. Hereafter it is referred to as habitat identity. For convenience, the information of several smaller habitats/environmental categories was combined together and grouped in a “major habitat”. For example, the habitats “sea beach”, “coral reef”, “marine waters”, “marine sediments”, “hydrothermal vents” etc. were grouped into a single major habitat, “marine”. Similarly, the habitats like “human skin”, “human oral”, “human gut”, “human lunch” etc. were grouped into a single major habitat, “human” (please see Supplementary Table 1 for more details). The habitats that did not contribute to 0.5% or more in at least one sample were grouped into “others” at a 97% level of identity. For better understanding and comparison, similar major groups were kept also at an 85% level of identity. As we expected that the 97% (species level) would give us more accurate information, we had initially analyzed the phylotypes at a 97% level of identity. However, at this level, most of the phylotypes were not assigned to any environmental categories. After inputting a query of 16S rRNA sequences, the MetaMetaDB runs a BLAST search against the representative 16S rRNA sequences and results in a hit list; the higher the level of identity, the lower will be the number of hits. Thus, we also analyzed the phylotypes at an 85% (order level) level which covered more than 80% of obtained phylotypes. The phylotypes assigned to a particular habitability were identified by aligning them against their taxonomy.

3 Results

3.1 Environmental parameters and total cell count

The sampling locations, depth, air temperature, water temperature, salinity and total cell count are shown in Table 1. Salinity showed a typical gradient from freshwater (about 0.5 in the river stations), brackish water (about 5.0 in the estuary station to about 32 in the Oarai sea side station) to off-shore marine water in Kuroshio waters (about 35). The highest total cell count was at the Oarai port side station and the lowest in Kuroshio surface waters. The counts for two riverine stations, one estuarine and one sea side station were similar (Table 1).

3.2 Bacterial community structures

A total of 114,010 reliable quality sequences consisting of 11,178 different phylotypes (OTUs) were obtained (Supplementary Fig. 1, the rarefaction curve showed the number of phylotypes obtained against the total number of sequences for each sample). Bacterial community structure at a phylum level is presented in Fig. 2. As for Proteobacteria, however, five classes such as Alpha-, Beta-, Gamma-, Delta- and Epsilonproteobacteria are shown separately; for Bacteroidetes, the class Flavobacteriia is shown. The phyla that were less than 1% in all the samples were combined together and referred to as “others” and those with no known information/no classification as “unclassified”. The community structures of the estuarine station and the river stations (low-salinity stations) in all fractions were very similar (Fig. 2). The phylum Bacteroidetes was the most dominant one followed by the phylum Proteobacteria (Fig. 2). At stations with high salinity, i.e., in the Oarai port side and sea side stations, the relative contributions of the phylum Proteobacteria were higher than that of the phylum Bacteroidetes (Fig. 2). Among classes in Proteobacteria, Betaproteobacteria were dominant in freshwater (river) and brackish (estuary) environments, whereas most of them were shared by Alpha- and Gammaproteobacteria at high-salinity environments (Oarai and Kuroshio stations; Fig. 2). In the Kuroshio seawater, the phyla Cyanobacteria and Verrucomicrobia also appeared as dominant groups depending on the sample (Fig. 2). As for the PA and FL fractions, there were no clear differences except for samples from Kuroshio, especially the samples from the chlorophyll maximum layer (Fig. 2).
Fig. 2

Bacterial community structures at a phylum level. PA particle-associated, FL free-living and T total bacterial composition. As for Proteobacteria, five classes such as Alpha-, Beta-, Gamma-, Delta- and Epsilonproteobacteria are shown separately; for Bacteroidetes, the class Flavobacteriia is shown separately. The phyla less than 1% in all the samples were combined together and referred to as “others” and those with no known information/no classification as “unclassified”

3.3 Habitability assessment

Habitability of the obtained sequences was assessed contrasting against both the old (March 19, 2014) and the latest version (November 6, 2014). The results were very similar in both versions (Fig. 3, Supplementary Fig. 2) with no remarkable biases. So, the results of the latest version were considered, and the habitability was obtained at each phylogenetic level (Table 2). Among the 11,178 phylotypes, about 4.63, 14.3, 45.4, 79.5 and 94.4% of them were identified at the 97% (species), 95% (genus), 90% (family), 85% (order) and 80% (class) levels of identity, respectively (Table 2). No habitat identity (no matching with the database sequences) was available for 5.61% of the phylotypes (Table 2). The habitability was assessed initially at a 97% (species) level of identity. The river and estuarine stations (the low-salinity stations) were mostly occupied by the groups assigned as “freshwater and groundwater” and “human” and these two covered 42–59% of the relative abundance at this level (Table 2 and Fig. 3a), while all the assigned habitats collectively covered 44–60% of the relative abundance. In contrast, high-salinity stations were mostly occupied by the groups assigned as “marine” and “plants and roots”, and all the assigned habitats collectively covered about 5.7–26% of the relative abundance. No phylotype with habitat identity of “freshwater and groundwater” or “human” was found in the Kuroshio waters at this level (Fig. 3a).
Fig. 3

Habitability deduced from the MetaMetaDB (version November 6, 2014) at a 97% (species) level of identity and b 85% (order) level of identity. The percent contribution to the relative abundance was calculated considering only the OTUs of those assigned to different habitat identities at a 97% (a) or 85% (b) level of identity. Abbreviations are the same as in Fig. 2

Table 2

Percentage of phylotypes assigned to habitability at the different level of identity with their contribution to the relative abundance

Level of identify

Number of phylotypes assigned out of 11,178

% of phylotypes assigned

% contribution to the relative abundance by the assigned phylotypes

At low-salinity stations

At high-salinity stations

97% (species)





95% (genus)





90% (family)





85% (order)





80% (class)





No identity





The contribution was obtained by adding the relative abundance data of the individual phylotypes assigned to different habitat identities at a particular level of identity. The phylogenetic levels were marked as default output of the database

While at 97% (species) level, only 4.63% of the phylotypes were covered, at 85% (order) level, about 80% of sequences matched those in the MetaMetaDB (Table 2). At this level, a clear contrast was also found between the low and high-salinity stations (Table 3 and Fig. 3b). At the former stations, about a half was occupied by those assigned as “freshwater and groundwater”, followed by “wastewaters” and “plants and roots” (Table 3 and Fig. 3b). In contrast, “marine” type shared the considerable part of the sequences derived from high-salinity stations (Table 3 and Fig. 3b). Especially, nearly 90% were “marine” type sequences at the Kuroshio chlorophyll maximum depth (Fig. 3b). Among high-salinity stations, there were some differences between Oarai stations and Kuroshio stations. Sequences assigned as “sediments and soil” were seen at Oarai stations more frequently than Kuroshio ones, especially at port side. In Kuroshio surface water, some sequences of bacteria with “human” habitat identity were noticed (Fig. 3b). The habitability of the PA and FL fractions did not show marked differences (Fig. 3b).
Table 3

Habitability of the phylotypes at 85% (order) level of identity

Mostly adapted habitats (MHIs)

Contributions at low-salinity stations (%)

Contributions at high-salinity stations (%)

Fish (17–100%)



Freshwater and groundwater (14–100%)



Human (14–100%)



Marine (16–100%)



Oil production facilities (22–100%)



Plants and roots (17–100%)



Sediments and soil (18–100%)



Wastewaters (17–100%)



Others (ant, bioreactors, compost, epibiont, food, gut, ice, pig; 23–100%)



The percentage of contribution by a particular habitat identity was calculated after accumulating the contribution to the total relative abundance by the phylotypes assigned to a particular habitat

MHIs microbial habitability indices

3.4 Major groups assigned to each habitat identity

The phylotypes assigned to a particular habitat identity at a 97% (species) level were identified by aligning them against their phylogenetic group (Fig. 4 and Supplementary Tables 2–5). In general, there were clear differences between the low- and high-salinity stations. About 14–20% of those in the “freshwater and groundwater” (Supplementary Table 2) and 25–41% of those in “human” (supplementary Table 3) were shared by the genus Flavobacterium (Flavobacteriales) of phylum Bacteroidetes in the low-salinity stations. This genus was also present in the Oarai sea side station with up to 1.4% of “freshwater and groundwater” and 3.3% of “human” in the total. The contribution of this genus was insignificant at the Oarai port side station and nonexistent at the Kuroshio Current samples (Supplementary Tables 2 and 3). In contrast, 3.0–18.3% was occupied by “marine” at high-salinity stations (Supplementary Table 4). Among different members of the phylum Bacteroidetes, the genera NS5 marine group (Flavobacteriales), NS9 marine group (Flavobacteriales) and Polaribacter (Flavobacteriales) contributed about 0.2–4.1%, 0.1–4.1% and 0.2–1.9% to the total, respectively, at high-salinity stations. The genus Ulvibacter (Flavobacteriales) of the same phylum (Bacteroidetes) shared 2.4–3.5% of the total at the Oarai sea side station. The SAR86 clade (Oceanospirillales; Gammaproteobacteria) contributed 0.46–9.0% to the total at high-salinity stations. On the other hand, in the “marine” habitat identity, only 0.4–2.6% relative abundance was covered at the low-salinity stations and these above-mentioned genera were scarce or absent there (Supplementary Table 4). As for the “plants and roots” (supplementary Table 5), 0.3–6.8% was covered at the high-salinity stations and 0.02–0.3% at the low-salinity stations. The group NS4 (Flavobacteriales) of phylum Bacteroidetes was dominant in this “plant and roots” habitat identity, and among different high-salinity stations, the highest dominancy was found in Kuroshio chlorophyll maximum water, with about 6.5% assigned to the FL bacterial fraction. The genus Marinoscillum (Cytophagales) of the same phylum (Bacteroidetes) was found only in the Kuroshio waters (Supplementary Table 5).
Fig. 4

Major groups of bacteria of the MetaMetaDB appraisal samples summarized from different projected habitats at a 97% level of identity. The cells highlighted with blue color indicate these groups contributed 0.5% or more to the relative abundance in at least one sample type (i.e., PA, FL or T). The contribution was calculated considering the relative abundance of each of the OTU assigned to these habitat identities. The cells highlighted with red color indicate the same groups were totally absent (0% contribution to the relative abundance) in those projected habitat identities. R1, R2, E, SS, PS, KS and KC mean riverine 1, riverine 2, estuarine, Oarai sea side, Oarai port side, Kuroshio surface and Kuroshio chlorophyll maximum stations, respectively

At an 85% level of identity, more than 20% of the sequences in FL and PA fractions from Kuroshio surface water was unexpectedly assigned to the “human”, more specifically, “human gut” (Supplementary Table 6). Those sequences were mostly shared by the order “Cyanobacteria subsection I” of the phylum Cyanobacteria. The same group was also found to contribute about 4% of the total in the FL fraction of Kuroshio chlorophyll maximum water. In other sampling points, this group was absent or scarce (Supplementary Table 6).

4 Discussion

The study was conducted to examine the applicability of the MetaMetaDB to examine the habitability of the bacterial communities in river, estuarine, coastal and offshore environments. In order to confirm the applicability, we assumed that the following conditions should be met. First, as for samples obtained from different salinity, the MetaMetaDB gives reasonable explanations based on the salinity differences. Second, the MetaMetaDB shows the influence of terrestrial or anthropogenic influences on some coastal samples, but not on offshore samples. Third, as for some phylogenetic groups, the MetaMetaDB gives information which is overlooked by ordinary community structure analyses. For the first, at a 97% level of identity, habitability showed clear and expected differences between the low- and high-salinity stations. The bacterial communities at two rivers and one estuarine water stations were mostly occupied by the phylotypes from “freshwater and groundwater” and “human”, while the high-salinity coastal stations were mostly occupied by the phylotypes assigned as “marine” and “plants and roots”. The similar tendency was also found at the 85% (order level) level of identity. For the second, at a 97% level, sequences assigned to “freshwater and groundwater” or “human” were present, especially at the sea side of Oarai stations, but were absent at the Kuroshio area. At an 85% level, sequences assigned to “sediment and soil” were seen at Oarai stations, whereas those shared only a minor portion in the Kuroshio area. It is noteworthy that the Oarai sea side sometimes receives freshwater from the Naka River (Matsu-ura et al. 2010; Uda 2010), which might bring some phylotypes assigned to the “freshwater and groundwater”. For the third, we found two examples, Bacteroidetes and sequences assigned to the “human” in Kuroshio water (see below).

Among the Bacteroidetes phylotypes assigned to “freshwater and groundwater” and “human” at a 97% level, genus Flavobacterium (Flavobacteriales) contributes predominantly. The members of genus Flavobacterium (Bergey et al. 1923) are widely distributed in soil and freshwater habitats, and some of them are pathogenic for fish (Bernardet et al. 1996; Kim et al. 2009; Wang et al. 2006). The phylotypes assigned to “marine” contain the NS5 marine group (Flavobacteriales), NS9 marine group (Flavobacteriales), genus Polaribacter (Flavobacteriales) and genus Ulvibacter (Flavobacteriales). The genus Polaribacter was first isolated from a polar marine environment (Gosink et al. 1998) and since then, it has been found in various regions, including coastal areas of Japan (Fukui et al. 2013; Teeling et al. 2012). The genus Ulvibacter was first isolated from green alga (Nedashkovskaya et al. 2004) as well as from coastal sea water (Baek et al. 2014). The genus Marinoscillum was first isolated from a marine sponge (Seo et al. 2009). The phylotypes assigned as “plants and roots” were mostly occupied by the NS4 and genus Marinoscillum (Cytophagales) of the Bacteroidetes phylum that was only present in the Kuroshio waters. Thus, the MetaMetaDB showed that there are multiple members with different habitat identities in the Bacteroidetes phylum, which was overlooked by ordinary community structure analyses.

The second case of our unexpected result was the presence of sequences with habitat identity of “human” in Kuroshio surface water. Some bacterial sequences with habitat identity of “human” were noticed. Di Rienzi et al. (Di Rienzi et al. 2013) tried the whole genome reconstruction of human feces and proposed a new candidate phylum, Melainabacteria, a sibling to Cyanobacteria. The Melainabacteria are non-photosynthetic, anaerobic and obligately fermentative, and are also present in soil and aquatic environments. Although further detailed analyses are required, there is a possibility that members belonging to this group were present in Kuroshio samples.

Previous studies reported that the phylum Bacteroidetes is widely distributed in both freshwater and marine habitats (Amaral-Zettler et al. 2010; Glockner et al. 1999; Kirchman 2002; Kirchman 2012). The classes Alphaproteobacteria, (DeLong et al. 2006; Fuhrman and Davis 1997; Lopez-Garcia et al. 2001; Pham et al. 2008) and Gammaproteobacteria are generally the two most dominant groups in high-salinity or marine environments, but are relatively rare in freshwater or low-salinity waters. On the contrary, the class Betaproteobacteria and the phylum Actinobacteria are more abundant in freshwater habitats, but less so in marine habitats (Cottrell and Kirchman 2003; Crump et al. 1999; Herlemann et al. 2011; Kan et al. 2008; Kirchman et al. 2005; Kirchman 2012; Murray et al. 1996). We have also found that the Kuroshio water was dominated by the phyla Cyanobacteria in both PA and FL fractions, and Verrucomicrobia, especially in PA fractions of the chlorophyll maximum layer. The abundance of the phylum Cyanobacteria in the Kuroshio area were reported in previous studies (Juan et al. 2011; Kataoka et al. 2009). There were also indications that the phylum Verrucomicrobia is significantly abundant in PA fractions to compare to the FL in some areas (Crespo et al. 2013; Freitas et al. 2012) and more abundant in the chlorophyll maximum layer (Crespo et al. 2013). Thus, the present results on bacterial community structures are consistent with typical distribution patterns of the phyla Bacteroidetes, Proteobacteria, Actinobacteria, Cyanobacteria and Verrucomicrobia.

The results did not show clear differences in both community structure and habitability among PA and FL fractions of bacteria, with the only exception in the Kuroshio chlorophyll maximum layer, suggesting the exchanges of populations between FL and PA states occur at a reasonably fast rate. However, this doesn’t exclude the possibility that there may be some differences at a finer spatiotemporal scale.

In conclusion, in order to examine the applicability of the MetaMetaDB to examine habitability, the bacterial community structures in environments with different salinity were examined. The MetaMetaDB showed that the low-salinity stations were dominated by the sequences with “freshwater and groundwater”, “human” and “wastewater” habitabilities, while the high-salinity stations were dominated by those with “marine” habitat identity. These overall findings proved that the MetaMetaDB is useful as a new tool to infer microbial habitability.

Supplementary material

10872_2017_449_MOESM1_ESM.docx (391 kb)
Supplementary material 1 (DOCX 391 kb)
10872_2017_449_MOESM2_ESM.docx (71 kb)
Supplementary material 2 (DOCX 70 kb)


  1. Acinas SG, Rodrı ́guez-Valera F, Pedro ́s-Alio ́ C (1997) Spatial and temporal variation in marine bacterioplankton diversity as shown by RFLP finger-printing of PCR amplified 16S rDNA. FEMS Microbiol Ecol 24:27–40CrossRefGoogle Scholar
  2. Amaral-Zettler L, Artigas LF, Baross J et al (2010) A global census of marine life. In: McIntyre AD (ed) Life in the World’s Oceans: diversity, distribution, and abundance. Wiley-Blackwell, Oxford, pp 221–245CrossRefGoogle Scholar
  3. Baek K, Jo H, Choi A, Kang I, Cho J-C (2014) Ulvibacter marinus sp. nov., isolated from coastal seawater. Int J Syst Evol Microbiol 64:2041–2046CrossRefGoogle Scholar
  4. Bergey DH, Harrison FC, Breed RS, Hammer BW, Huntoon FM (eds) (1923) Bergey’s manual of determinative bacteriology. Williams and Wilkins, BaltimoreGoogle Scholar
  5. Bernardet JF, Segers P, Vancanneyt M, Berthe F, Kersters K, Vandamme P (1996) Cutting a Gordian knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl and Tait 1978). Int J Syst Bacteriol 46:128–148CrossRefGoogle Scholar
  6. Cole JR, Wang Q, Cardenas E, Fish J, Chai B et al (2009) The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res 37:D141–D145CrossRefGoogle Scholar
  7. Cottrell MT, Kirchman DL (2003) Contribution of major bacterial groups to bacterial biomass production (thymidine and leucine incorporation) in the Delaware estuary. Limnol Oceanogr 48:168–178CrossRefGoogle Scholar
  8. Crespo BG, Pommier T, Fernández Gómez B, Pedrós Alió C (2013) Taxonomic composition of the particle-attached and free-living bacterial assemblages in the Northwest Mediterranean Sea analyzed by pyrosequencing of the 16S rRNA. Microbiologyopen 2(4):541–552CrossRefGoogle Scholar
  9. Crump BC, Armbrust EV, Baross JA (1999) Phylogenetic analysis of particle-attached and free-living bacterial communities in the Columbia River, its estuary, and the adjacent coastal ocean. Appl Environ Microbiol 65:3192–3204Google Scholar
  10. Crump BC, Hopkinson CS, Sogin ML, Hobbie JE (2004) Microbial biogeography along an estuarine salinity gradient: combined influences of bacterial growth and residence time. Appl Environ Microbiol 70:1494–1505CrossRefGoogle Scholar
  11. DeLong EF, Franks DG, Alldredge AL (1993) Phylogenetic diversity of aggregate-attached vs. free-living marine bacterial assemblages. Limnol Oceanogr 38:924–934CrossRefGoogle Scholar
  12. DeLong EF, Preston CM, Mincer T et al (2006) Community genomics among stratified microbial assemblages in the ocean’s interior. Science 311:496–503CrossRefGoogle Scholar
  13. DeSantis TZ, Hugenholtz P, Larsen N, Rojas M, Brodie EL et al (2006) Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Appl Environ Microbiol 72:5069–5072CrossRefGoogle Scholar
  14. Di Rienzi SC, Sharon I, Wrighton KC, Koren O, Hug LA et al (2013) The human gut and groundwater harbor non-photosynthetic bacteria belonging to a new candidate phylum sibling to Cyanobacteria. eLife 2:e01102CrossRefGoogle Scholar
  15. Edgar RC, Haas BJ, Clemente JC, Quince C, Knight R (2011) UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27:2194–2200CrossRefGoogle Scholar
  16. Freitas S, Hatosy S, Fuhrman JA, Huse SM, Welch DBM, Sogin ML, Martiny AC (2012) Global distribution and diversity of marine Verrucomicrobia. ISME J 6:1499–1505CrossRefGoogle Scholar
  17. Fuhrman JA, Davis AA (1997) Wide spread archaea and novel bacteria from the deep sea as shown by 16S rRNA gene sequences. Mar Ecol Prog Ser 150:275–285CrossRefGoogle Scholar
  18. Fukui Y, Abe M, Kobayashi M, Saito H, Oikawa H, Yano Y, Satomi M (2013) Polaribacter porphyrae sp. nov., isolated from the red alga Porphyra yezoensis, and emended descriptions of the genus Polaribacter and two Polaribacter species. Int J Syst Evol Microbiol 63:1665–1672CrossRefGoogle Scholar
  19. Giovannoni SJ, Mullins TD, Field KG (1995) Microbial diversity in oceanic systems: rRNA approaches to the study of unculturable microbes. NATO ASI Ser Ser G Ecol Sci 38:217–248Google Scholar
  20. Glockner FO, Fuchs BM, Amann R (1999) Bacterioplankton compositions of lakes and oceans: a first comparison based on fluorescence in situ hybridization. Appl Environ Micro-biol 65:3721–3726Google Scholar
  21. Gosink JJ, Woese CR, Staley JT (1998) Polaribacter gen. nov., with three new species, P. irgensii sp. nov., P. franzmannii sp. nov. and P. filamentus sp. nov., gas vacuolate polar marine bacteria of the Cytophaga–Flavobacterium–Bacteroides group and reclassification of ‘Flectobacillus glomeratus’ as Polaribacter glomeratus comb. nov. Int J Syst Bacteriol 48:223–235CrossRefGoogle Scholar
  22. Herlemann DPR, Labrenz M, Jürgens K, Bertilsson S, Waniek JJ, Andersson AF (2011) Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J 5:1571–1579CrossRefGoogle Scholar
  23. Hiorns WD, Methe BA, Nierzwicki-Bauer SA, Zehr JP (1997) Bacterial diversity in Adirondack Mountain lakes as revealed by 16S rRNA gene sequences. Appl Environ Microbiol 63:2957–2960Google Scholar
  24. Hiraoka H, Machiyama A, Ijichi M, Inoue K, Oshima K, Hattori M, Yoshizawa S, Kogure K, Iwasaki W (2016) Genomic and metagenomic analysis of microbes in a soil environment affected by the 2011 Great East Japan earthquake tsunami. BMC Genom. doi:10.1186/s12864-016-2380-4 Google Scholar
  25. Huse SM, Welch DM, Morrison HG, Sogin ML (2010) Ironing out the wrinkles in the rare biosphere through improved OTU clustering. Environ Microbiol 12(7):1889–1898CrossRefGoogle Scholar
  26. Juan L, Zhang Y, Junde D, Youshao W, Lei C, Jingbin F, Hongyan S, Dongxiao W, Si Z (2011) Spatial variation of bacterial community composition near the Luzon strait assessed by polymerase chain reaction-denaturing gradient gel electrophoresis (PCR-DGGE) and multivariate analyses. Afr J Biotechnol 10(74):16897–16908Google Scholar
  27. Kan J, Evans S, Chen F, Suzuki M (2008) Novel estuarine bacterioplankton in rRNA operon libraries from the Chesapeake Bay. Aquat Microb Ecol 51:55–66CrossRefGoogle Scholar
  28. Kataoka T, Hodoki Y, Suzuki K, Saito H, Higashi S (2009) Tempo-spatial patterns of bacterial community composition in the western North Pacific Ocean. J Marine Sys 77(1–2):197–207CrossRefGoogle Scholar
  29. Kim JH, Kim KY, Cha CJ (2009) Flavobacterium chungangense sp. nov., isolated from a freshwater lake. Int J Syst Evol Microbiol 59:1754–1758CrossRefGoogle Scholar
  30. Kim M, Morrison M, Yu Z (2011) Evaluation of different partial 16S rRNA gene sequence regions for phylogenetic analysis of microbiomes. J Microbiol Meth 84(1):81–87CrossRefGoogle Scholar
  31. Kirchman DL (2002) The ecology of Cytophaga-Flavobacteria in aquatic environments FEMS Microbiol. Ecol. 39:91–100Google Scholar
  32. Kirchman DL (2012) Processes in microbial ecology. Oxford University Press, New YorkGoogle Scholar
  33. Kirchman DL, Dittel AI, Malmstrom RR, Cottrell MT (2005) Biogeography of major bacterial groups in the Delaware Estuary. Limnol Oceanogr 50:1697–1706CrossRefGoogle Scholar
  34. Lopez-Garcia P, Lopez-Lopez A, Moreira D, Rodriguez-Valera F (2001) Diversity of free-living prokaryotes from a deep-sea site at the Antarctic Polar Front. FEMS Microbiol Ecol 36:193–202CrossRefGoogle Scholar
  35. Maidak BL, Olsen GJ, Larsen N, Overbeek R, McCaughey MJ, Woese CR (1996) The Ribosomal Database Project (RDP). Nucleic Acids Res 24(1):82–85CrossRefGoogle Scholar
  36. Matsu-ura T, Uda T, Kumada T, Sumiya M (2010) Sand accumulation in wave-shelter zone of Oharai Port and change in grain size of seabed materials on nearby coast. In: Proc 32nd ICCE, sediment 63:1–11. http://journals.tdl.org/ICCE/article/view/1077/pdf_179
  37. Murray AE, Hollibaugh JT, Orrego C (1996) Phylogenetic compositions of bacterioplankton from two California estuaries compared by denaturing gradient gel electrophoresis of 16S rDNA fragments. Appl Environ Microbiol 62:2676–2680Google Scholar
  38. Nedashkovskaya OI, Kim SB, Han SK, Rhee MS, Lysenko AM, Falsen E, Frolova GM, Mikhailov VV, Bae KS (2004) Ulvibacter litoralisgen. nov., sp. nov., a novel member of the family Flavobacteriaceae isolated from the green alga Ulva fenestrata. Int J Syst Evol Microbiol 54:119–123CrossRefGoogle Scholar
  39. Pham VD, Konstantinidis KT, Palden T, DeLong EF (2008) Phylogenetic analyses of ribosomal DNA-containing bacterioplankton genome fragments from a 4000 m vertical profile in the North Pacific Subtropical Gyre. Environ Microbiol 10:2313–2330CrossRefGoogle Scholar
  40. Porter KG, Feig YS (1980) The use of DAPI for identifying and counting aquatic microflora. Limnol Oceanogr 25(5):943–948CrossRefGoogle Scholar
  41. Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig W, Peplies J, Glöckner FO (2007) SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res 35:7188–7196CrossRefGoogle Scholar
  42. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T et al (2013) The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res 41:D590–D596CrossRefGoogle Scholar
  43. Schloss PD, Westcott SL (2011) Assessing and improving methods used in operational taxonomic unit-based approaches for 16S rRNA gene sequence analysis. Appl Environ Microbiol 77(10):32193226CrossRefGoogle Scholar
  44. Schloss PD, Westcott SL, Ryabin T et al (2009) Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75(23):7537–7541CrossRefGoogle Scholar
  45. Seo HS, Kwon KK, Yang SH, Lee HS, Bae SS, Lee JH, Kim SJ (2009) Marinoscillum gen. nov., a member of the family ‘Flexibacteraceae’, with Marinoscillum pacificum sp. nov. from a marine sponge and Marinoscillum furvescens nom. rev., comb. nov. Int J Syst Evol Microbiol 59:1204–1208CrossRefGoogle Scholar
  46. Teeling H, Fuchs BM, Becher D et al (2012) Substrate-controlled succession of marine bacterioplankton populations induced by a phytoplankton bloom. Science 336:608–611CrossRefGoogle Scholar
  47. Uda T (2010) Impacts on sandy beach and habitat of Japanese hard clams due to construction of port breakwater. In: Inter Symposium on Integrated Coastal Management for Marine Biodiversity in Asia. Collective abstracts B-1, pp 28–34Google Scholar
  48. Wang Z-W, Liu Y-H, Dai X, Wang B-J, Jiang C-Y, Liu S-J (2006) Flavobacterium saliperosum sp. nov., isolated from freshwater lake sediment. Int J Syst Evol Microbiol 56:439–442CrossRefGoogle Scholar
  49. Yang C-C, Iwasaki W (2014) MetaMetaDB: a database and analytic system for investigating microbial habitability. PLoS One 9(1):e87126. doi:10.1371/journal.pone.0087126 CrossRefGoogle Scholar

Copyright information

© The Oceanographic Society of Japan and Springer Japan KK 2017

Authors and Affiliations

  1. 1.Atmosphere and Ocean Research InstituteThe University of TokyoKashiwaJapan
  2. 2.Department of Biological Sciences, Graduate School of ScienceThe University of TokyoTokyoJapan
  3. 3.Faculty of FisheriesBangladesh Agricultural UniversityMymensinghBangladesh

Personalised recommendations