Species composition shift of confined bacterioplankton studied at the level of community DNA

Using a new DNA hybridization technique that does not require culturing, we compared the species composition of natural planktonic bacterial assemblages before and after confinement in 20 1 containers for ca 2 d Although confinement is known to cause species shifts, possibly by stimulating growth of certain types of cells near the container wall, we found that such shifts were minor; 5 to 15 % of the communities had changed during confinement. The greatest shifts occurred in the samples that had the fastest bacterial growth rate measured by [Hlthymidine incorporation. Despite the minor changes in species composition, the fraction of the cultivable cells (colony forming units; CFU) increased 4to 23-fold, but amounted to < 2 % of total counts even after confinement Interestingly, CFU on unsupplemented media increased more rapidly than CFU on supplemented media. Comparisons to previous studies suggest that our use of large (20 1) containers and our efforts to minimize contamination with organic material may have decreased the 'wall effect.' We conclude that confined microbial communities did not undergo drastic changes of species composition in 2 d in 20 1 containers. It is well known that confinement of natural bacterioplankton assemblages affects the metabolic activities of confined organisms, and subsequently results in changes of species composition (ZoBell & Anderson 1936, ZoBell 1943, Ferguson et dl. 1984). Similar findings, e.g. major species composition shifts and population size changes, have been obtained with confined planktonic eukaryotic organisms (Venrick et al. 1977). One of the most distinctive changes during confinement of natural bacteria is the increase of colony forming units (CFU) on plate media, where nutrients are relatively enriched compared to natural seawater. The cells cultivable on those media are generally con* Addressee for correspondence. Present addressOceanographic and Atmospheric Sciences Division, Brookhaven National Laboratory, Upton, New York 11973, USA 0 Inter-Research/Printed in Germany sidered to require high levels of nutrients in order to proliferate. A possible mechanism of species composition change under confinement is an outgrowth of high-nutrient type cells in microhabitats near the container wall, where organic materials are adsorbed from adjacent water (ZoBell 1943). This mechanism is consistent with the observation that the increase in microbial activities under confinement becomes smaller with an increase in container volume (ZoBell & Anderson 1936, ZoBell 1943, references therein). Besides the wall effect, contained systems are considered to be different from natural systems in various ways (Ferguson et al. 1984, references therein). Previous studies examined the confinement effect on species compositions mostly by CFU, and the change of CFU has been suggested as supportive evidence of the species shift indicative of the entire community structure change. However, most naturally occurring cells are non-cultivable (CFU < 1 % of total cell counts; Ferguson et al. 1984, and this study), despite the observation that many (40 to 90 % of total count) can be shown to be metabolically active by mcroautoradiography (Fuhrman & Azam 1982). Use of cultivable cells alone may not be appropriate to investigate the species composition change of a whole community. The rich culture media used in the previous studies of confinement effects may have been too enriched to see the responses (or may have inhibited the activities) of cells which may require only slightly increased levels of nutrients (Baxter & Sieburth 1984). Lee & Fuhrman (1990) recently presented a rapid and simple method, community DNA hybridization, for species composition comparison between bacterial communities. The method uses total DNA directly extracted from natural bacterioplankton communities, and hence does not require pure cultures. Exploiting 196 Mar Ecol. Prog Ser. 79: 195-201, 1991 the observation that total DNA does not crosshybridize significantly between species (Palleroni et al. 1973, Lee & Fuhrman 1990), our technique measures the community DNA similarity (%) which is an estimate of the overlapping fraction between 2 communities due to the identical or near-identical DNA species present in both samples. The species composition similarity was presented as a summation of the minimum common fractions between 2 samples. For example, when Community A is composed of 3 species a, b and c each taking up 40, 30, 30 % of the community, respectively, and Community B with 4 species a, b, d, and e each 50, 20, 20, 10 %, then the common species are a and b, each with the minimum common fraction of 40 and 20 %, respectively. The similarity is thus 40 + 20 = 60 %. Therefore, the similarity is an index based on the 2 parameters of the community structure, ire. species and their frequencies. This approach was tested with mixtures of DNA from pure lab cultures, by hybridizing the DNA mixtures in a reciprocal fashion (probe and target switched) and comparing the observed similarities to the theoretical ones calculated from the known compositions of the mixtures (Lee & Fuhrman 1990). The measured similarities approximated the expected ones under most circumstances. Although the discriminating power of DNA hybridization led to a suggestion that bacterial strains of > 70 % DNA homology belong to one species (Wayne et al. 1987), strict interpretation of community DNA similarity to within a few percent is not possible, partly because there is a slight cross-hybridization between species (generally < 10 %). Details of the concept, tests and evaluations, and limitations of the method are presented in Lee & Fuhrman (1990). In this study, confinement effects on bacterial species compositions were examined by the community DNA hybridization method. An important difference from previous studies was the use of larger (20 1) volume containers, needed to obtain sufficient DNA for the hybridizations. We also investigated changes of CFU, cell concentrations, and growth rates under confinement. The community DNA hybridization showed ca 5 to 15 % of species composition shift while the CFU were always < 2 % of the total cell count. Materials and methods. Sampling: Naturally occurring marine planktonic bacteria were sampled (Table 1) from an oligotrophic Pacific station (550 km west of San Diego, USA) at 25 m (OpenPac25) and 100 m (OpenPaclOO) water depths, and from a jetty at Playa Del Rey, Santa Monica Bay, Los Angeles, CA (CoastalPac). Other samples included in the community DNA comparisons were from a beach at Crane Neck, Long Island, New York, USA, (LIS) and the Caribbean Sea (Carib; 100 km west of Dominica). To collect the seawater, we used Niskin bottles (OpenPac and Carib samples), or clean plastic buckets thoroughly rinsed with seawater (CoastalPac and LIS samples). The water collected with the buckets was transferred to clean (see below) 20 1 polyethylene containers for transport and storage (1.e. confinement). Confinement experiment: The seawater samples were transferred to and stored in 20 1 polyethylene containers (Cubitainer; Consolidated Plastics Co., Inc., Twinsburg, OH, USA) which were soaked overnight with 10 % HC1 and rinsed thoroughly with the sample seawater. No treatment was given except aging in 20 1 containers. The OpenPac25 sample was stored on the ship deck in the shade for 48 h, and the OpenPaclOO sample was stored in the ship's lab for 32 h. The CoastalPac sample was stored indoors in the lab at University of Southern California for 48 h. Storage temperatures were 14 to 18 'C (OpenPac) and 21 'C (CoastalPac), whereas in situ temperatures were 17 OC (OpenPac25), 13 OC (OpenPaclOO), and 20 'C (CoastalPac). Cell collection for DNA: Bacterial cells were collected by pressure filtration as described previously (Lee & Fuhrman 1990). In brief, the water was first pre-filtered through type AE glass fiber filters (142 mm diameter; Gelman Sciences Inc., Ann Arbor, MI, USA) to remove larger particles and eukaryotic cells, and bacterial cells were collected on 0.22 pm pore size Durapore filters (142 mm diameter; Millipore Corp., Bedford, MA, USA). The filtrations were done in a relatively short period of time (total of 1.5 to 3 h after sample collection, depending on the volume filtered). Subsamples were taken from unfiltered seawater, intermediate filtrate (between AE filter and Durapore filter) and final filtrate. Filtering efficiencies were monitored from subsamples by the acridine orange direct count (AODC) method (Hobbie et al. 1977). Filtration was done in the same way for stored seawater samples at the end of confinement. DNA extraction: We followed the protocol of Fuhrman et al. (1988) to extract DNA from the cells collected on Durapore filters, with one more step of purification with phenol/chloroform/isoamyl alcohol (24: 6: 1 by vol, pH 8.0). After final purification and redissolution in TE (10 mM Tris, 1 mM EDTA, pH 8.0) at the concentrations of 100 to 500 ng p l l , DNA was quantified by Hoechst 33258 dye (bisbenzimde; Sigma Chemical Co., St. Louis, MO, USA) fluorometry (Paul & Myers 1982). Probe and target DNA preparation: Probe DNA was labeled with [alpha-^S]dATP (DuPont, NEN Research Products, Boston, MA, USA) by nick translation and purified as described previously (Lee & Fuhrman 1990). Probes were dried by vacuum centrifugation Lee & Fuhrman: Species composition shift of confined microorganisms 197 Table 1. Dates and locations of natural bacterioplankton samples