External control of bacterial community structure in lakes

We investigated the importance of a regional factor for bacterial communities in lakes. External factors dominated the control of community structures in lakes with retention times up to 200 d, most likely as a result of bacterial import. Because these lakes are numerous in the boreal zone, regional processes can be of great importance for bacterial communities in general. Consequently, we propose that lakes function more like flow through systems, as opposed to the classical ‘‘lake as microcosm’’ concept. During the last decade, our knowledge about the diversity of bacteria in nature has increased enormously; however, little is understood about which factors are shaping bacterial communities (DeLong and Pace 2001; Zwart et al. 2002). Central to understanding patterns in community structure is knowledge about the relative importance of local versus regional processes (Chase 2003; Cottenie and De Meester 2004). It can be assumed that the degree of isolation, and thereby the rate of exchange of cells and genes between communities, should have consequences for which forces shape local microbial communities (Curtis and Sloan 2004; Papke and Ward 2004). In nature, a gradient in isolation of communities should exist from endosymbionts to those exhibiting a cosmopolitan distribution (Papke and Ward 2004). Early on, lakes and their ecosystems were assigned as isolated units in the ‘‘lake as microcosm’’ concept (Forbes 1887). This view of the lake was later revised, for instance, because of the discovery of the importance of allochthonous carbon for lake ecosystem function (Hessen and Tranvik 1998; Pace et al. 2004). Thus, lakes today are more often ecologically regarded as a part of a larger unit, i.e., the drainage basin (Soranno et al. 1999). Still, efforts in lake microbial ecology and diversity have largely focused on within-lake selective forces, rather than external influences on community structure and diversity, although it has been shown that lake (Crump et al. 2003; Masin et al. 2003; Lindström and Bergström 2004) and estuarine (Crump et al. 2004) community compositions can be largely influenced by inflowing bacteria. It can be assumed that the degree of isolation a lake bacterioplankton community experiences, and thereby the degree of influence by inflowing bacteria on local community structure, depends on the hydrological retention time of the lakes. Therefore, we used a range of lakes of different hydrological retention times to determine the magnitude of external influence on bacterial community structure in lakes. Our hypothesis was that lakes with short hydrological retention time should have bacterial communities that are more similar to those of the inflowing water than lakes with longer hydrological retention times because of the larger amount of imported bacteria to the former type of lake. Twelve relatively unproductive Swedish lakes with different hydrological retention times were studied (Table 1). We used denaturing gel electrophoresis (DGGE) of polymerase chain reaction (PCR)–amplified 16S ribosomal deoxyribonucleic acid (rDNA) to characterize bacterial communities in inlets and lakes. Materials and methods—Study sites and sampling: Twelve lakes were included in the study (Table 1). The lakes differ in theoretical hydrological retention time from 1 d to 10 yr. All the lakes have one inlet or one major inlet. Four lakes are situated in Lappland, northern Sweden (Lindström and Bergström 2004, 2005), and were sampled in summer 2001 (6–7 June) and 2002 (1–2 June). The eight other lakes are located in Uppland, central Sweden, and were sampled 20–28 May 2003. All lakes were sampled once, in spring, approximately 25–50 d after the start of the latest high-flow event, when flow had subsided. Temperature, pH, water color, total organic carbon, total nitrogen, total phosphorus, bacterial community composition, bacterial production, bacterial numbers, and abundance of potential bacterivores were determined in lakes and inlets (Lindström and Bergström 2004). In stratified lakes, composite samples of the epilimnion were taken, whereas in unstratified lakes, composite samples of the whole water column were taken. Assemblage composition of bacteria: The assemblage composition of bacteria in lakes and streams was determined by DGGE of PCR-amplified 16S rDNA (Lindström and Bergström 2004). Eubacterial primers GM5F (with GCclamp) and DS907-reverse (Teske et al. 1996) were used. The percent similarity between inlet and lakes in the presence and absence of DGGE bands was calculated (Lindström and Bergström 2004) as Sørensen’s index. DGGE data for Lake Örträsketsjön and Lake Stor-Sandsjön were obtained from an earlier study (Lindström and Bergström 2004). Bacterial cell budgets: Bacterial numbers were determined in lakes and streams by use of epifluorescence microscopy and acridine orange staining (Hobbie et al. 1977). Bacterial cell division rate in the lake waters was measured by use of the thymidine method (Bell 1993). Tritiated thymidine (30 or 50 nmol L21, depending on humic content of the lake water) was added to triplicates of 10-ml lake water at the time of arrival at the laboratory. The samples were incubated in darkness at in situ temperature for 60 min. The growth rate (cells L21 h21) was calculated by assuming a division rate of 2 3 1018 cells mol21 of incorporated thymidine (Bell 1990). Bacterioplankton cell budgets were constructed with the use of data on water flow, epilimnion volumes, bacterioplankton abundance, and growth rate (Lindström and Bergström