Introduction

Culture-dependent and -independent studies have found that prokaryotic assemblages are quite diverse in aquatic habitats and contain representatives of virtually all of the roughly 40 divisions of bacteria and the major archaeal groups found so far in the biosphere [1]. Fortunately, not all of these prokaryotic groups are abundant in the plankton nor important in all biogeochemical cycles. Autotrophic and heterotrophic bacteria dominate the prokaryotic biomass in surface waters, as Archaea appear to be abundant only in the plankton of the deep oceans [2]. Among the heterotrophic bacteria, the two most abundant groups are often the Proteobacteria and the subject of this review, the Cytophaga–Flavobacteria cluster (Table 1).

1

The relative abundance of the major heterotrophic bacterial groups in aquatic ecosystems, as determined by FISH with oligonucleotide probes

System Location % of Total Prokaryotesa 
Bacteriab α β γ CF Reference 
Lakes Alpine lake 55 21 24 <2 [71
Lago di Cadagno 46 21 [72
Grosser Ostersee 46 14 [72
Lake Baikal 44 [72
Lake Constance aggregates (50 m) 50–70 10–20 40–60 <5 <5 [38
Rivers River Ter 50–90 5–7 10–33 1–4 20–65 [20
Elbe River detritus 75 5–10 50 30 10–30 [37
Spittelwasser biofilm 77 12–22 20–50 <5 <10 [73
Oceans Several oceanic sites 39–96 4–14 0–4 0–9 2–72 [72
Southern Ocean 60–95 5–10 <5–10 25–70 [74
North Sea 31–71 3–25 5–9 <1–30 [27
California coast 54–88 1–34 0–3 7–35 22–38 [24
Delaware coast 80 28 10 14 [47
Wadden Sea sediments 40–73 1.5–3.5 1.0–4.1 0.8–3.6 7.5–18 [17
System Location % of Total Prokaryotesa 
Bacteriab α β γ CF Reference 
Lakes Alpine lake 55 21 24 <2 [71
Lago di Cadagno 46 21 [72
Grosser Ostersee 46 14 [72
Lake Baikal 44 [72
Lake Constance aggregates (50 m) 50–70 10–20 40–60 <5 <5 [38
Rivers River Ter 50–90 5–7 10–33 1–4 20–65 [20
Elbe River detritus 75 5–10 50 30 10–30 [37
Spittelwasser biofilm 77 12–22 20–50 <5 <10 [73
Oceans Several oceanic sites 39–96 4–14 0–4 0–9 2–72 [72
Southern Ocean 60–95 5–10 <5–10 25–70 [74
North Sea 31–71 3–25 5–9 <1–30 [27
California coast 54–88 1–34 0–3 7–35 22–38 [24
Delaware coast 80 28 10 14 [47
Wadden Sea sediments 40–73 1.5–3.5 1.0–4.1 0.8–3.6 7.5–18 [17

The table includes only studies using probes for at least three proteobacterial subdivisions (alpha, beta and gamma) and for the Cytophaga–Flavobacteria cluster (CF). Ranges are given when the study reports several values taken over time or within a habitat. When more than depth was sampled, the surface value is reported unless noted otherwise.

a

All DAPI-positive microbes, probably nearly all bacteria with a few Archaea and fewer (to none) Eukarya.

b

Percent of DAPI-positive microbes that are recognized by the Eub338 probe.

This paper reviews recent studies that have applied molecular methods to examine uncultured Cytophaga–Flavobacteria in freshwaters and the oceans, with the ultimate goal of using this information to better understand the role of heterotrophic bacteria in carbon cycles and other biogeochemical processes. The importance of heterotrophic bacteria in biogeochemical processes is now well appreciated, but until recently geochemists and field-orientated microbial ecologists considered these microbes as if they were a single group, even though microbiologists have been accumulating for several years information about the taxonomic and phylogenetic make up (‘community structure’) of heterotrophic bacterial communities. Only recently, however, have microbial ecologists been able to link community structure with specific biogeochemical processes (‘function’) [3]. Here I summarize briefly our progress in these areas while discussing what we know about Cytophaga–Flavobacteria in aquatic habitats. This microbial group is a natural starting point because of its high abundance in many freshwater and marine systems.

Taxonomy of Cytophaga–Flavobacteria

The Cytophaga–Flavobacteria cluster belongs to a diverse bacterial division that has been labeled differently over the years. Hugenholtz et al. [1] and other authors have used ‘Cytophagales’ while Hamana and Nakagawa [4] recently stated that the division name was ‘Flavobacterium–Flexibacter–Bacteriodes–Cytophaga’. The latter name does give some indication of the diversity of this division and highlights some of the major genera (Fig. 1). The label used in the most recent edition of Bergey's Manual of Systematic Bacteriology [5] is simply ‘Bacteroidetes’, which will be used in this review. A more common label, however, is Cytophaga–Flavobacteria–Bacteriodes (or Flavobacterium–Bacteriodes–Cytophaga and other permutations), similar to the label (Cytophaga–Flexibacter–Bacteriodes) originally used by Woese and colleagues [6]. Other genera, which are prominent because of their association with humans, include Prevotella and Porphyromonas[7].

1

A phylogenetic tree of the Bacteroidetes division, illustrating the Cytophaga–Flavobacteria cluster, based on nearly complete 16S rRNA sequences. The tree was constructed using ARB and the latest data set (May 2001) available at http://www.arb-home.de/. The numbers refer to the number of strains grouped together.

1

A phylogenetic tree of the Bacteroidetes division, illustrating the Cytophaga–Flavobacteria cluster, based on nearly complete 16S rRNA sequences. The tree was constructed using ARB and the latest data set (May 2001) available at http://www.arb-home.de/. The numbers refer to the number of strains grouped together.

The Cytophaga–Flavobacteria cluster can be characterized by both phenotypic traits and 16S rRNA sequence data. As first described by Winogradsky, Cytophaga-like bacteria are unicellular, gliding, nonspore-forming Gram-negative rods, although shape can vary [8]. Colonies of many Cytophaga-like bacteria are pigmented due to flexirubin-type pigments found only in these bacteria and flavobacteria. Most notably, Cytophaga–Flavobacteria are chemoorganotrophic and are especially proficient in degrading various biopolymers such as cellulose, chitin, and pectin [8] (see also references in [9]). Another distinguishing feature is that Cytophaga–Flavobacteria are mainly aerobic, although some anaerobic strains are currently classified as Cytophaga[8]. In contrast, the bacteria within the Bacteroides subgroup are all anaerobes and are found most notably in the microflora of the human colon, making up 30% of total fecal isolates (see references in [9]).

In his early review of bacterial systematics, Woese [10] did not use the term ‘Cytophaga–Flavobacteria cluster’, but he did divide 19 16S rRNA sequences from Bacteroidetes bacteria into two clusters, the Bacteroides and the Flavobacterium subdivisions. The number of Bacteroidetes sequences has increased by over 10-fold since Woese's review, so there are now several other groups in the Bacteroidetes division, but most of the Cytophaga and Flavobacteria species still cluster together (Fig. 1). However, one sign of problems in Bacteroidetes systematics is that the genus name Cytophaga is scattered throughout the entire Bacteroidetes phylogenetic tree, as illustrated by the tree presented recently by Weller et al. [7]. It seems that the call by Nakagaw and Yamasato [11] to divide Cytophaga into several genera and higher taxa has not been fully heeded.

The large diversity in rRNA sequences complicates several molecular methods for examining members of Bacteroidetes and other bacterial groups, but its impact on one method, fluorescence in situ hybridization (FISH) with oligonucleotide probes, is critical for this review. By far the most widely used FISH probe for Cytophaga–Flavobacteria is CF319a [9]. Although often termed a probe for Cytophaga–Flavobacteria, CF319a also hybridizes to some members of the Bacteroides–Prevotella cluster [7]. What is also noteworthy is that CF319a does not recognize several Cytophaga–Flavobacteria[7], implying that studies using this probe have underestimated Cytophaga–Flavobacteria abundance. One strain not recognized by the CF319a probe is Cytophaga hutchinsonii; the genome of this bacterium is now being sequenced by the Joint Genomic Institute (http://www.jgi.doe.gov/). Even though CF319a misses several Bacteroidetes genera, it remains the most general probe for the Bacteroidetes division [7]. Rather than replacements for CF319a, new probes are needed to examine clusters within the Bacteroidetes division, including those recognized by the CF319a probe. Weller et al. [7] suggested some new probes, but these need to be applied to environmental samples. Eilers et al. [12] used new probes to count culturable Cytophaga–Flavobacteria in the North Sea; the FISH data indicated that these strains were about 6% of total prokaryote abundance (DAPI counts) in April through September, which is a large fraction for a single bacterial lineage. Weller et al. [7] mentioned that they were unable to design habitat-specific probes that recognize bacteria found uniquely in freshwaters, the oceans, or soils.

The abundance of Cytophaga–Flavobacteria in nature, clone libraries and data bases

Bacteria in the Bacteroidetes division can be found in just about every habitat in the biosphere, including kusaya [13] (a Japanese delicacy consisting of putrid fish), rumens [14], hydrothermal vents [15], rocks and sea-ice in Antarctica [16], and sediments of lakes and the oceans [17]. Cytophaga–Flavobacteria seem particularly common in the oceans, according to FISH. In fact, in many oceanic habitats, the Cytophaga–Flavobacteria cluster is the most abundant of all bacterial groups (Fig. 2), accounting for as much as half of all bacteria potentially identified by FISH. Generally, β-proteobacteria are the most abundant group in freshwater systems (Table 1), although Cytophaga–Flavobacteria can dominate mature freshwater biofilms [18], and they were abundant in two Japanese rivers [19] and a river-reservoir system in Spain [20]. Cytophaga–Flavobacteria can be abundant in sediments [21], and in one study of Wadden Sea sediments, CF319a-positive bacteria were the most abundant group even in anoxic layers [17].

2

Summary of FISH results for the Baltic [72], North Sea [27], California [24], and the Antarctic circumpolar current in the Southern Ocean [74]. The Greek letters indicate the subdivisions of proteobacteria and ‘CF’ is Cytophaga–Flavobacteria. See Table 1 for other environments. The error on FISH data is on the order of 20%.

2

Summary of FISH results for the Baltic [72], North Sea [27], California [24], and the Antarctic circumpolar current in the Southern Ocean [74]. The Greek letters indicate the subdivisions of proteobacteria and ‘CF’ is Cytophaga–Flavobacteria. See Table 1 for other environments. The error on FISH data is on the order of 20%.

In contrast to the FISH results, few Cytophaga–Flavobacteria sequences have been found by most studies of clone libraries of 16S rRNA genes from free-living bacterial assemblages. In one review of these studies, Giovonnoni and Rappé[22] noted that clone libraries from marine systems were dominated by the following heterotrophs, in descending order: three groups of α-proteobacteria (SAR11, Roseobacter, and SAR116); a γ-proteobacterial group (SAR86); and Actinobacteria. The Cytophaga–Flavobacteria cluster was not in this list. Several Cytophaga–Flavobacteria sequences have been retrieved from particulate detritus communities [23] but few from free-living microbial assemblages, which usually make up >90% of total microbial abundance in the oceans.

There are two possible explanations for the difference between the FISH and clone library results. One is simply that all clone libraries to date have been constructed from sites or at times when Cytophaga–Flavobacteria were relatively sparse. After all, bacterial community structure varies greatly among habitats and over time. Another explanation is that clone-library construction is biased against the Cytophaga–Flavobacteria cluster. This latter explanation appears to the case.

Two studies have compared FISH and clone libraries of 16S rRNA genes from the same water sample. FISH results indicated that the Cytophaga–Flavobacteria cluster was the most abundant bacterial group along a transect off the coast of California [24], although a few stations had large numbers of α-proteobacteria. In contrast to the FISH results, libraries from two sites of this transect were dominated by 16S rRNA genes from α-proteobacteria, and only five of 169 clones were from Cytophaga–Flavobacteria. Similar to other studies [22], the α-proteobacterial group SAR 11 was abundant in one of the libraries (12% of all clones), but SAR 11-like bacteria were <1% of total bacterial abundance according to FISH. Clones from the γ-proteobacterial group SAR 86, which has attracted much attention recently [25,26], were also abundant in the libraries (about 10%), but these bacteria averaged only 2% of total abundance according to FISH. Eilers et al. [27] also found discrepancies between FISH and clone-library results. Specifically, these authors did not find any 16S rRNA Cytophaga–Flavobacteria clones in a library of 54 clones from a North Sea sample, yet this bacterial group was the most abundant one as identified by FISH. In short, the Cytophaga–Flavobacteria cluster is under-represented and the α-proteobacteria over-represented in clone libraries, compared with the FISH data (Fig. 3).

3

Frequency of clones from various bacterial groups in libraries of 16S rRNA genes compared with their relative abundance as estimated by FISH in California coastal waters. The Greek letters indicate the subdivisions of proteobacteria and ‘CF’ is Cytophaga–Flavobacteria. Data from Cottrell and Kirchman [24].

3

Frequency of clones from various bacterial groups in libraries of 16S rRNA genes compared with their relative abundance as estimated by FISH in California coastal waters. The Greek letters indicate the subdivisions of proteobacteria and ‘CF’ is Cytophaga–Flavobacteria. Data from Cottrell and Kirchman [24].

Of course, FISH is not perfect [28]. One problem is that a relatively large fraction (ca. 50%) of the prokaryotic cells sometimes cannot be probed with any oligonucleotide probe, including the Eub338 probe, which should recognize all bacteria (but see reference [29]). However, Cottrell and Kirchman [24] argued that, for their samples, the composition of clone libraries would still overestimate the α-proteobacterial abundance as determined by FISH, even if all bacteria not detected by FISH were α-proteobacteria. Nonspecific hybridization with CF319a to proteobacteria, which would lead to overestimating Cytophaga–Flavobacteria abundance, has not been observed to date (R. Amann, pers. comm.). On the contrary, it seems likely that CF319a misses members of the Cytophaga–Flavobacteria cluster (see above), thus leading to underestimates of their numbers.

The PCR step is an obvious place to look for biases against the Cytophaga–Flavobacteria cluster. Suzuki et al. [30] pointed to the mismatches between EubA (the reverse primer used by Cottrell and Kirchman [24,47]) and the 16S rRNA sequences from their Bacteroidetes clones. However, these mismatches seem insufficient to explain the discrepancy, because clone libraries made with other primer pairs also have few Cytophaga–Flavobacteria 16S rRNA genes [24]. For example, a library of 190 clones made with primers 27F and LSU1933R had only three Bacteroidetes clones compared with 67 from α-proteobacteria [30]. The library of Eilers et al. [27] made with primers 8f and 1542r had four α-proteobacterial clones (and 49 from the γ-proteobacterial subdivision) but none from the Bacteroidetes division. It seems unlikely that primer mismatches explain the paucity of Bacteroidetes clones in all of these libraries. Biases other than primer mismatches need to be considered.

Regardless of the explanation, relatively few 16S rRNA sequences from the Cytophaga–Flavobacteria cluster are in the databases. According to the RDP database (http://www.cme.msu.edu/rdp; June 2000 release), the entire Bacteroidetes division is represented by 781 sequences or about 5% of all 16S rRNA sequences, almost an order of magnitude less than the nearly 7000 sequences from the Proteobacteria division. The discrepancy is even larger when sequences of only uncultured bacteria are considered. A search of GenBank (May 2001) turned up 93 16S rRNA gene sequences when ‘Cytophaga’ and ‘uncultured’ were used as keywords. An analogous search for uncultured proteobacterial sequences yielded over 1700 entries. Glöckner et al. [31] reported new Cytophaga–Flavobacteria 16S rRNA sequences from several freshwater habitats, but more sequence data are still needed.

Degradation of organic material by Cytophaga–Flavobacteria

Heterotrophic bacteria are the dominant organisms in aquatic habitats using dissolved organic material (DOM), which is one of the largest pools of organic carbon in the biosphere, and thus is critical in carbon budgets and cycles. Uptake of DOM is the first step in the microbial loop that ultimately mineralizes about 50% of primary production [32]. Microbial ecologists interested in carbon cycling would like to know the role of various bacterial groups in taking up DOM. The Cytophaga–Flavobacteria cluster is one obvious starting point because of its high abundance, but also because these bacteria may have a specialized role in DOM uptake and degradation. Cultured isolates of Cytophaga–Flavobacteria are well known to be proficient in degrading biopolymers such as cellulose and chitin, part of the high molecular mass (HMW) fraction of DOM. This DOM fraction is being intensively examined by organic geochemists because concentrations and fluxes of HMW DOM are high, especially in the oceans [33].

Two lines of evidence indicate that Cytophaga–Flavobacteria have a special role in using HMW DOM in aquatic environments, not just in laboratory cultures. First, several studies have demonstrated that Cytophaga–Flavobacteria are enriched on particulate organic detritus in marine habitats [23,34–36]. In contrast, detritus in the Elbe River [37] and in Lake Constance [38] was always dominated by β-proteobacteria, although Cytophaga–Flavobacteria became more abundant over time on artificial organic aggregates made from riverine as well as marine DOM [39,40]. One major phenotypic property of Cytophaga–Flavobacteria, use of HMW organic material, would help explain the success of these bacteria on both particulate detritus and in free-living assemblages.

It could be argued that the high abundance of Cytophaga–Flavobacteria in the free-living assemblage is not due to use of HMW organic material (or any DOM component) but rather the release of these bacteria from detritus-associated communities. However, concentrations and fluxes of particulate detritus are usually too small, much less than those of DOM in most aquatic habitats, to support the large, detritus-free Cytophaga–Flavobacteria biomass observed at least in the oceans (Table 1). Detritus-associated microbial communities certainly release Cytophaga–Flavobacteria, but these cells must be able to survive in the free-living community in order to result in their high abundance. Similarly, the observation that similar ribotypes of Cytophaga–Flavobacteria are found free-living and detritus-associated communities [41] can be explained by use of organic compounds available to both communities, that is, HMW organic material, rather than exchange between these two communities.

The detritus examined by previous studies was either macroscopic material collected by scuba divers or suspended material caught on filters with large pore sizes. Little is known about the microbes associated specifically with other forms of detritus, such as gels [42], transparent polymers [43] and submicron particles [44]. The boundary between the free-living community and microbes associated with these detrital forms is probably not sharp.

The second line of evidence for HMW DOM use by uncultured Cytophaga–Flavobacteria comes from a new approach that combines microautoradiography and FISH [45,46]. Using a version of this approach (‘Micro-FISH’), Cottrell and Kirchman [47] examined uptake of protein and chitin and monomers of these two biopolymers by bacterial assemblages in coastal waters of Delaware (USA). Bacteria belonging to the Cytophaga–Flavobacteria cluster were the most abundant microbes using protein and chitin (Fig. 4 illustrates chitin use), but surprisingly few of these bacteria were among the microbes using amino acids, although a large fraction using N-acetylglucosamine were also Cytophaga–Flavobacteria. The Micro-FISH data suggest that amino acid uptake was dominated by α-proteobacteria in these coastal waters. Consistent with these results, Schweitzer et al. [38] found net amino acid uptake when lake snow aggregates were dominated by α-proteobacteria, but net release of amino acids when α-proteobacteria declined and the abundance of Cytophaga–Flavobacteria and β-proteobacteria increased.

4

Uptake of chitin and N-acetylglucosamine by bacterial groups in Delaware coastal waters, as determined by Micro-FISH. The Greek letters indicate the subdivisions of proteobacteria. The bar labeled with ‘?’ represents those bacteria not identified by any of the group-specific probes. Data taken from Cottrell and Kirchman [47]. The error on FISH data is on the order of 20%.

4

Uptake of chitin and N-acetylglucosamine by bacterial groups in Delaware coastal waters, as determined by Micro-FISH. The Greek letters indicate the subdivisions of proteobacteria. The bar labeled with ‘?’ represents those bacteria not identified by any of the group-specific probes. Data taken from Cottrell and Kirchman [47]. The error on FISH data is on the order of 20%.

These data do not indicate that Cytophaga–Flavobacteria are incapable of using amino acids and other monomers. Since compounds larger than ca. 500 Da cannot be directly transported across cell membranes, even biopolymer-degrading bacteria need to be able to take up monomers or small oligomers resulting from the hydrolysis of HMW DOM. However, biopolymer-degrading bacteria may be less proficient than other bacteria in using low molecular mass compounds. Conceivably, local concentrations of low molecular mass byproducts of HMW DOM hydrolysis are sufficiently high to alleviate the need for biopolymer-hydrolyzing bacteria to maintain high-affinity transport systems for monomers. Uptake by bacteria with low-affinity transport systems is probably not detected by the Micro-FISH method, which is designed to assay the most active cells.

Another approach for exploring the role of Cytophaga–Flavobacteria in using HMW DOM would be to follow the abundance of specific bacterial groups during times or locations when the input of HMW DOM is high, such as during phytoplankton blooms. A few studies have tried to follow Cytophaga–Flavobacteria abundance during blooms in marine [48–51] and freshwater habitats [52], but it is difficult to draw any general conclusions with these data. Eilers et al. [12] did observe large maxima in Cytophaga–Flavobacteria during short-lived phytoplankton blooms in the North Sea; this bacterial group was the most abundant through most the year but other bacterial groups also increased when Cytophaga–Flavobacteria increased. Nonetheless, it is intriguing to note that Cytophaga–Flavobacteria responded more than other bacteria when cyanobacterial biomass was added to anoxic sediments [53], and Cytophaga–Flavobacteria increased six- to eight-fold in a glacial stream following a storm that washed in allochthonous organic material [54]. Of course, factors other than DOM uptake will affect the success of Cytophaga–Flavobacteria and other bacteria in natural environments.

The great interest in uncultured Cytophaga–Flavobacteria, evident in the papers reviewed above, has not been matched by relevant work on cultured representatives. Biopolymer degradation by cultured Cytophaga–Flavobacteria was examined by studies published over 20 years ago (see review by [8]), and other studies of interest to ecologists are nearly as old [55]. More current research seems lacking. One specific gap is the paucity of sequence data on enzymes catalyzing biopolymer hydrolysis. Even though biopolymer hydrolysis by Cytophaga–Flavobacteria is well known, GenBank has only three cellulase and no chitinase sequences from the entire Bacteroidetes division; over 70% of the cellulase sequences are from Gram-positive bacteria while the chitinases sequences are mainly from γ-proteobacteria. The whole genome sequencing of C. hutchinsonii already has found several genes for presumptive endoglucanases, but more sequence data are needed to elucidate the ecology and physiology of these bacteria.

Growth rates of Cytophaga–Flavobacteria and other bacterial groups

Data on population growth rates are essential in examining the autecology and biogeochemical role of bacterial groups in natural environments. More specifically, growth rate data may help explain the abundance of Cytophaga–Flavobacteria in some environments. One of the rare studies with relevant data is that of Jürgens et al. [56]. These authors used FISH to follow over time the abundance of several bacterial groups after filtration of lake water through a 250-μm mesh net. Even though small protist grazers and viruses were not removed, bacterial abundance increased in these experiments because the screen removed the crustacean zooplankton Daphnia, a major bacterivore in lakes. Jürgens et al. found that the net growth rate of Cytophaga–Flavobacteria was nearly two-fold faster than that of the other bacterial groups they examined (Fig. 5). These rates are some evidence that the high abundance of Cytophaga–Flavobacteria is due to high growth rates.

5

Growth rates of bacterial groups in Daphnia-free water. ‘All cells’ refers to the growth rate of all DAPI-positive microbes. ‘Eub338+’ is the growth rate of the bacteria recognized by the Eub338 probe. The Greek letters indicate the subdivisions of proteobacteria. Data taken from Jürgens et al. [56].

5

Growth rates of bacterial groups in Daphnia-free water. ‘All cells’ refers to the growth rate of all DAPI-positive microbes. ‘Eub338+’ is the growth rate of the bacteria recognized by the Eub338 probe. The Greek letters indicate the subdivisions of proteobacteria. Data taken from Jürgens et al. [56].

However, other experiments indicate that Cytophaga–Flavobacteria may not always grow faster than other bacteria. In experiments similar to those conducted by Jürgens et al., but with 0.8-μm filtration, Šimek et al. [57] found that Cytophaga–Flavobacteria (those recognized by the CF319a probe) grew the slowest in a reservoir in South Bohemia, although only marginally more slowly than β-proteobacteria recognized by the BET42a probe. A subgroup of β-proteobacteria grew the fastest. In experiments with samples from the Plymouth Sound and English Channel, Fuchs et al. [58] found that both γ- and α-proteobacteria grew faster than Cytophaga–Flavobacteria in dilution cultures in which microbes were diluted 10-fold with filter-sterilized water.

The growth rates in the dilution culture experiments conducted by Fuchs et al. [58] were consistent with flow cytometric data. Fuchs et al. found that the proteobacterial subdivisions dominated the high-protein cells sorted by flow cytometry, whereas cells in the low-protein cells mostly belonged to the Cytophaga–Flavobacteria cluster. Zubkov et al. [59] used leucine incorporation to estimate growth rates of three groups of bacteria sorted by DNA content. These authors found that the group dominated by Cytophaga–Flavobacteria had the lowest growth rate in the mixed layer of the Celtic Sea. Growth rates of the three sorted groups were similar in the pycnocline and deep layer.

Impact of grazing on Cytophaga–Flavobacteria

The abundance of the total bacterial community and perhaps of specific bacterial groups is relatively constant over short time periods (day time scale) in most aquatic habitats, because growth of these microbes is matched by mortality due to both viral lysis and grazing. Generally, grazing accounts for most of bacterial mortality [60], but the impact of grazing and viral lysis on individual bacterial groups is largely unknown. Phages of Cytophaga–Flavobacteria have been examined in laboratory experiments [61,62], but nothing is known about the impact of viral lysis on natural communities of Cytophaga–Flavobacteria. Turning the tables, Cytophaga-like bacteria are known to lyse cyanobacteria [63], but the significance of this process in nature is unclear. More is known about the impact of grazing, a topic that was reviewed recently by Hahn and Höfle [64].

The impact of grazing can be deduced by following the abundance of bacterial groups in incubations with and without bacterivores. The experiments conducted so far have been with freshwater assemblages. In these experiments, the abundance of either α- or β-proteobacteria increased after removal of grazers, but in all experiments the absolute and relative abundance of Cytophaga–Flavobacteria increased substantially, more so than observed in the controls [56,65,66]. These data imply that Cytophaga–Flavobacteria and, less consistently, one of the proteobacterial subdivisions are preferentially grazed by bacterivores.

This ‘preference’, however, may simply result from the high growth rates of these bacterial groups. Since grazing rates increase with prey density, rapidly growing Cytophaga–Flavobacteria will in effect be grazed on more heavily than slow-growing bacteria [67,68]. However, growth rate could affect cell properties that do attract heavy grazing activity independent of prey density.

One such cell property is cell size. Since grazing pressure increases with cell size, large, fast growing cells may be grazed on more heavily than small, slow-growing cells. There is little information about growth rates (except for the data reviewed before) and cell size of Cytophaga–Flavobacteria and other uncultured bacteria in nature. Some experiments have examined an extreme case where the bacterial prey becomes too large to be eaten by protist grazers. In laboratory experiments of simple predator–prey systems, bacteria can form long filaments when a bacterivore is present [64], and thus escape being eaten. The capacity to form these filaments seems to be common among the bacterial groups examined to date, and the morphology of both proteobacteria and Cytophaga–Flavobacteria can shift to filaments under grazing pressure in simple laboratory experiments [64].

Cell properties other than size also may affect grazing. In cultures, heterotrophic protists often grow on some bacterial species better than on others, and grazing rates are often higher on live than on heat-killed bacteria [60]. Cell size does not explain these observations. We know little about how grazer-sensitive cell properties vary among the major bacterial groups.

How much do we need to know about bacterial diversity?

This review has discussed the ecology of the entire Cytophaga–Flavobacteria cluster and of three subdivisions of proteobacteria, as if these bacterial groups were simple collections of a few microbes, which of course they are not. In part, we have little choice but to use these broad phylogenetic labels because of limitations in our understanding of bacterial diversity in aquatic environments. These limitations in turn affect our methods (e.g. designing new FISH probes) for examining aquatic bacteria. In particular, we still know little about the diversity of Cytophaga–Flavobacteria in aquatic habitats and about which clades within this cluster we should focus on. In spite of these limitations, however, I suggest that the studies reviewed here were able to make useful generalizations about the ecology of these bacteria. In spite of great differences among the aquatic habitats discussed here, we can see some consistent patterns in how Cytophaga–Flavobacteria and other bacterial groups respond to experimental manipulations and vary with locations and the seasons. But, in other cases, we may need more data about subgroups within the Cytophaga–Flavobacteria cluster. The question then becomes, what phylogenetic level is most appropriate for examining the biogeochemical roles of Cytophaga–Flavobacteria and other uncultured bacterial groups?

The appropriate level will depend on the specific question. At one extreme, we already know that a strain-specific approach is needed to examine pathogenicity or antibiotic production by bacteria (but see [69]). At the other extreme, it may be adequate to consider all heterotrophic bacteria together as a single functional unit when examining DOM uptake and mineralization in aquatic habitats. An argument against lumping all heterotrophic bacteria together is that we now know that uncultured Cytophaga–Flavobacteria and other bacterial groups differ in their use of DOM. Still, before diving into studies of DOM use by individual species (or finer phylogenetic levels), which may number in the hundreds to thousands in lakes and the oceans, it may be useful to keep in mind that current models of the carbon cycle have a single compartment for all heterotrophic bacteria (and Archaea) using DOM [70]. So, even adding four groups (the Cytophaga–Flavobacteria and three proteobacterial subdivisions) would drastically change (and complicate) current models.

Although microbial ecologists should not be limited by models, keeping them in mind may help ensure that information about microbial community structure will be useful in examining larger issues, such as the role of microbes in carbon cycles and global climate change. Elucidating the links between community structure and function is likely to enrich our understanding of not only microbial processes but also of the diversity of microbial communities in natural environments. We are still in the beginning stages of examining structure–function links, but even now it seems clear that Cytophaga-Flavobacteria are important members of the bacterial community in many aquatic environments.

Acknowledgements

I thank Matt Cottrell, Jakob Pernthaler and two anonymous reviewers for their comments on this review and Rudi Amann and Frank-Oliver Glöckner for their technical input. This review was prepared with support by grants from the US Department of Energy (BIOMP program) and the National Science Foundation.

References

[1]
Hugenholtz
P
Goebel
B.M
Pace
N.R
(
1998
)
Impact of culture-independent studies on the emerging phylogenetic view of bacterial diversity
.
J. Bacteriol.
 
180
,
4765
4774
.
[2]
Karner
M.B
DeLong
E.F
Karl
D.M
(
2001
)
Archaeal dominance in the mesopelagic zone of the Pacific Ocean
.
Nature
 
409
,
507
510
.
[3]
Gray
N.D
Head
I.M
(
2001
)
Linking genetic identity and function in communities of uncultured bacteria
.
Environ. Microbiol.
 
3
,
481
492
.
[4]
Hamana
K
Nakagawa
Y
(
2001
)
Polyamine distribution profiles in the eighteen genera phylogenetically located within the Flavobacterium–Flexibacter–Cytophaga complex
.
Microbios
 
106
,
7
17
.
[5]
Ludwig
W.
Klenk
H.-P.
(
2001
)
Overview: A phylogenetic backbone and taxonomic framework for procaryotic systematics
. In:
Bergey's Manual of Systematic Bacteriology
  (
Boone
D.R.
Castenholz
R.W.
, Eds.).
Springer-Verlag
,
New York
, pp.
49
65
.
[6]
Olsen
G.J
Woese
C.R
Overbeek
R
(
1994
)
The winds of (evolutionary) change: breathing new life into microbiology
.
J. Bacteriol.
 
176
,
1
6
.
[7]
Weller
R
Glöckner
F.O
Amann
R
(
2000
)
16S rRNA-targeted oligonucleotide probes for the in situ detection of members of the phylum Cytophaga-Flavobacterium-Bacteroides
.
Syst. Appl. Microbiol.
 
23
,
107
114
.
[8]
Reichenbach
H.
Dworkin
M.
(
1991
)
The order Cytophagales
. In:
The Prokaryotes
  (
Balows
A.
Truper
H.G.
Dworkin
M.
Harder
W.
Schleifer
K.H.
, Eds.).
Springer-Verlag
,
New York
,
356
379
.
[9]
Manz
W
Amann
R
Ludwig
W
Vancanneyt
M
Schleifer
K.H
(
1996
)
Application of a suite of 16S rRNA-specific oligonucleotide probes designed to investigate bacteria of the phylum Cytophaga–Flavobacter–Bacteroides in the natural environment
.
Microbiology (UK)
 
142
,
1097
1106
.
[10]
Woese
C.R
(
1987
)
Bacterial evolution
.
Microbiol. Rev.
 
51
,
221
271
.
[11]
Nakagawa
Y
Yamasato
K
(
1993
)
Phylogenetic diversity of the genus Cytophaga revealed by 16S ribosomal RNA sequencing and menaquinone analysis
.
J. Gen. Microbiol.
 
139
,
1155
1161
.
[12]
Eilers
H
Pernthaler
J
Peplies
J
Glöckner
F.O
Gerdts
G
Amann
R
(
2001
)
Isolation of novel pelagic bacteria from the German Bight and their seasonal contributions to surface picoplankton
.
Appl. Environ. Microbiol.
 
67
,
5134
5142
.
[13]
Satomi
M
Kimura
B
Takahashi
G
Fujii
T
(
1997
)
Microbial diversity in Kusaya gravy
.
Fish. Sci.
 
63
,
1019
1023
.
[14]
Ramsak
A
Peterka
M
Tajima
K
Martin
J.C
Wood
J
Johnston
M.E.A
Aminov
R.I
Flint
H.J
Avgustin
G
(
2000
)
Unravelling the genetic diversity of ruminal bacteria belonging to the CFB phylum
.
FEMS Microbiol. Ecol.
 
33
,
69
79
.
[15]
Sievert
S.M
Kuever
J
Muyzer
G
(
2000
)
Identification of 16S ribosomal DNA-defined bacterial populations at a shallow submarine hydrothermal vent near Milos Island (Greece)
.
Appl. Environ. Microbiol.
 
66
,
3102
3109
.
[16]
Smith
M.C
Bowman
J.P
Scott
F.J
Line
M.A
(
2000
)
Sublithic bacteria associated with Antarctic quartz stones
.
Antarct. Sci.
 
12
,
177
184
.
[17]
Llobet-Brossa
E
Rossello-Mora
R
Amann
R
(
1998
)
Microbial community composition of Wadden Sea sediments as revealed by fluorescence in situ hybridization
.
Appl. Environ. Microbiol.
 
64
,
2691
2696
.
[18]
Manz
W
Wendt-Potthoff
K
Neu
T.R
Szewzyk
U
Lawrence
J.R
(
1999
)
Phylogenetic composition, spatial structure, and dynamics of lotic bacterial biofilms investigated by fluorescent in situ hybridization and confocal laser scanning microscopy
.
Microb. Ecol.
 
37
,
225
237
.
[19]
Kenzaka
T
Yamaguchi
N
Tani
K
Nasu
M
(
1998
)
rRNA-targeted fluorescent in situ hybridization analysis of bacterial community structure in river water
.
Microbiology (UK)
 
144
,
2085
2093
.
[20]
Šimek
K
Armengol
J
Comerma
M
Garcia
J.C
Kojecka
P
Nedoma
J
Hejzlar
J
(
2001
)
Changes in the epilimnetic bacterial community composition production, and protist-induced mortality along the longitudinal axis of a highly eutrophic reservoir
.
Microb. Ecol.
 
42
,
359
371
.
[21]
Ravenschlag
K
Sahm
K
Amann
R
(
2001
)
Quantitative molecular analysis of the microbial community in marine Arctic sediments (Svalbard)
.
Appl. Environ. Microbiol.
 
67
,
387
395
.
[22]
Giovannoni
S.J.
Rappé
M.S.
(
2000
)
Evolution, diversity and molecular ecology of marine prokaryotes
. In:
Microbial Ecology of the Ocean
  (
Kirchman
D.L.
, Ed.), pp.
47
84
.
Wiley Interscience
,
New York
.
[23]
DeLong
E.F
Franks
D.G
Alldredge
A.L
(
1993
)
Phylogenetic diversity of aggregate-attached versus free-living marine bacterial assemblages
.
Limnol. Oceanogr.
 
38
,
924
934
.
[24]
Cottrell
M.T
Kirchman
D.L
(
2000
)
Community composition of marine bacterioplankton determined by 16S rDNA clone libraries and fluorescence in situ hybridization
.
Appl. Environ. Microbiol.
 
66
,
5116
5122
.
[25]
Béjà
O
Aravind
L
Koonin
E.V
Suzuki
M.T
Hadd
A
Nguyen
L.P
Jovanovich
S
Gates
C.M
Feldman
R.A
Spudich
J.L
Spudich
E.N
DeLong
E.F
(
2000
)
Bacterial rhodopsin: Evidence for a new type of phototrophy in the sea
.
Science
 
289
,
1902
1906
.
[26]
Béjà
O
Spudich
E.N
Spudich
J.L
Leclerc
M
DeLong
E.F
(
2001
)
Proteorhodopsin phototrophy in the ocean
.
Nature
 
411
,
786
789
.
[27]
Eilers
H
Pernthaler
J
Glöckner
F.O
Amann
R
(
2000
)
Culturability and in situ abundance of pelagic bacteria from the North Sea
.
Appl. Environ. Microbiol.
 
66
,
3044
3051
.
[28]
Amann
R.I
Ludwig
W
Schleifer
K.H
(
1995
)
Phylogenetic identification and in situ detection of individual microbial cells without cultivation
.
Microbiol. Rev.
 
59
,
143
169
.
[29]
Daims
H
Bruhl
A
Amann
R
Schleifer
K.H
Wagner
M
(
1999
)
The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: Development and evaluation of a more comprehensive probe set
.
Syst. Appl. Microbiol.
 
22
,
434
444
.
[30]
Suzuki
M.T
Beja
O
Taylor
L.T
DeLong
E
(
2001
)
Phylogenetic analysis of ribosomal RNA operons from uncultivated coastal marine bacterioplankton
.
Environ. Microbiol.
 
3
,
323
342
.
[31]
Glöckner
F.O
Zaichikov
E
Belkova
N
Denissova
L
Pernthaler
J
Pernthaler
A
Amann
R
(
2000
)
Comparative 16S rRNA analysis of lake bacterioplankton reveals globally distributed phylogenetic clusters including an abundant group of actinobacteria
.
Appl. Environ. Microbiol.
 
66
,
5053
5065
.
[32]
Ducklow
H.
(
2000
)
Bacterial production and biomass in the oceans
. In:
Microbial Ecology of The Oceans
  (
Kirchman
D.L.
, Ed.), pp.
85
120
.
John Wiley and Sons
,
New York
.
[33]
Amon
R
Benner
R
(
1996
)
Bacterial utilization of different size classes of dissolved organic matter
.
Limnol. Oceanogr.
 
41
,
41
51
.
[34]
Crump
B.C
Armbrust
E.V
Baross
J.A
(
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
3204
.
[35]
Rath
J
Wu
K.Y
Herndl
G.J
DeLong
E.F
(
1998
)
High phylogenetic diversity in a marine-snow-associated bacterial assemblage
.
Aquat. Microb. Ecol.
 
14
,
261
269
.
[36]
Fandino
L.B
Riemann
L
Steward
G.F
Long
R.A
Azam
F
(
2001
)
Variations in bacterial community structure during a dinoflagellate bloom analyzed by DGGE and 16S rDNA sequencing
.
Aquat. Microb. Ecol.
 
23
,
119
130
.
[37]
Böckelmann
U
Manz
W
Neu
T.R
Szewzyk
U
(
2000
)
Characterization of the microbial community of lotic organic aggregates (‘river snow’) in the Elbe River of Germany by cultivation and molecular methods
.
FEMS Microbiol. Ecol.
 
33
,
157
170
.
[38]
Schweitzer
B
Huber
I
Amann
R
Ludwig
W
Simon
M
(
2001
)
Alpha- and beta-Proteobacteria control the consumption and release of amino acids on lake snow aggregates
.
Appl. Environ. Microbiol.
 
67
,
632
645
.
[39]
Grossart
H.P
Ploug
H
(
2000
)
Bacterial production and growth efficiencies: Direct measurements on riverine aggregates
.
Limnol. Oceanogr.
 
45
,
436
445
.
[40]
Grossart
H.P
Ploug
H
(
2001
)
Microbial degradation of organic carbon and nitrogen on diatom aggregates
.
Limnol. Oceanogr.
 
46
,
267
277
.
[41]
Riemann
L.
Winding
A.
(
2001
)
Community dynamics of free-living and particle-associated bacterial assemblages during a freshwater phytoplankton bloom
.
Microb. Ecol.
 
43
.
[42]
Chin
W.C
Orellana
M.V
Verdugo
P
(
1998
)
Spontaneous assembly of marine dissolved organic matter into polymer gels
.
Nature
 
391
,
568
572
.
[43]
Alldredge
A.L
Passow
U
Logan
B.E
(
1993
)
The abundance and significance of a class of large, transparent organic particles in the ocean
.
Deep-Sea Res. I
 
40
,
1131
1140
.
[44]
Koike
I
Hara
S
Terauchi
K
Kogure
K
(
1990
)
Role of sub-micrometre particles in the ocean
.
Nature
 
345
,
242
244
.
[45]
Lee
N
Nielsen
P.H
Andreasen
K.H
Juretschko
S
Nielsen
J.L
Schleifer
K.H
Wagner
M
(
1999
)
Combination of fluorescent in situ hybridization and microautoradiography – a new tool for structure–function analyses in microbial ecology
.
Appl. Environ. Microbiol.
 
65
,
1289
1297
.
[46]
Ouverney
C.C
Fuhrman
J.A
(
1999
)
Combined microautoradiography–16S rRNA probe technique for determination of radioisotope uptake by specific microbial cell types in situ
.
Appl. Environ. Microbiol.
 
65
,
1746
1752
.
[47]
Cottrell
M.T
Kirchman
D.L
(
2000
)
Natural assemblages of marine proteobacteria and members of the CytophagaFlavobacter cluster consuming low- and high- molecular-weight dissolved organic matter
.
Appl. Environ. Microbiol.
 
66
,
1692
1697
.
[48]
Pinhassi
J
Azam
F
Hemphala
J
Long
R.A
Martinez
J
Zweifel
U.L
Hagström
A
(
1999
)
Coupling between bacterioplankton species composition, population dynamics, and organic matter degradation
.
Aquat. Microb. Ecol.
 
17
,
13
26
.
[49]
Riemann
L
Steward
G.F
Azam
F
(
2000
)
Dynamics of bacterial community composition and activity during a mesocosm diatom bloom
.
Appl. Environ. Microbiol.
 
66
,
578
587
.
[50]
Schafer
H
Servais
P
Muyzer
G
(
2000
)
Successional changes in the genetic diversity of a marine bacterial assemblage during confinement
.
Arch. Microbiol.
 
173
,
138
145
.
[51]
Lebaron
P
Servais
P
Troussellier
M
Courties
C
Muyzer
G
Bernard
L
Schafer
H
Pukall
R
Stackebrandt
E
Guindulain
T
Vives-Rego
J
(
2001
)
Microbial community dynamics in Mediterranean nutrient-enriched seawater mesocosms: changes in abundances, activity and composition
.
FEMS Microbiol. Ecol.
 
34
,
255
266
.
[52]
van Hannen
E.J
Zwart
G
van Agterveld
M.P
Gons
H.J
Ebert
J
Laanbroek
H.J
(
1999
)
Changes in bacterial and eukaryotic community structure after mass lysis of filamentous cyanobacteria associated with viruses
.
Appl. Environ. Microbiol.
 
65
,
795
801
.
[53]
Rossello-Mora
R
Thamdrup
B
Schafer
H
Weller
R
Amann
R
(
1999
)
The response of the microbial community of marine sediments to organic carbon input under anaerobic conditions
.
Syst. Appl. Microbiol.
 
22
,
237
248
.
[54]
Battin
T.J
Wille
A
Sattler
B
Psenner
R
(
2001
)
Phylogenetic and functional heterogeneity of sediment biofilms along environmental gradients in a glacial stream
.
Appl. Environ. Microbiol.
 
67
,
799
807
.
[55]
Höfle
M.G
(
1984
)
Transient responses of glucose-limited cultures of Cytophaga johnsonae to nutrient excess and starvation
.
Appl. Environ. Microbiol.
 
47
,
356
362
.
[56]
Jürgens
K
Pernthaler
J
Schalla
S
Amann
R
(
1999
)
Morphological and compositional changes in a planktonic bacterial community in response to enhanced protozoan grazing
.
Appl. Environ. Microbiol.
 
65
,
1241
1250
.
[57]
Šimek
K
Pernthaler
J
Weinbauer
M.G
Hornak
K
Dolan
J.R
Nedoma
J
Masin
M
Amann
R
(
2001
)
Changes in bacterial community composition and dynamics and viral mortality rates associated with enhanced flagellate grazing in a mesoeutrophic reservoir
.
Appl. Environ. Microbiol.
 
67
,
2723
2733
.
[58]
Fuchs
B.M
Zubkov
M.V
Sahm
K
Burkill
P.H
Amann
R
(
2000
)
Changes in community composition during dilution cultures of marine bacterioplankton as assessed by flow cytometric and molecular biological techniques
.
Environ. Microbiol.
 
2
,
191
201
.
[59]
Zubkov
M.V
Fuchs
B.M
Burkill
P.H
Amann
R
(
2001
)
Comparison of cellular and biomass specific activities of dominant bacterioplankton groups in stratified waters of the Celtic Sea
.
Appl. Environ. Microbiol.
 
67
,
5210
5218
.
[60]
Strom
S.L.
(
2000
)
Bacterivory: interactions between bacteria and their grazers
. In:
Microbial Ecology of the Oceans
  (
Kirchman
D.L.
, Ed.), pp.
351
386
.
Wiley-Liss
,
New York
.
[61]
Richter
C.A
Pate
J.L
(
1988
)
Temperate phages and bacteriocins of the gliding bacterium Cytophaga johnsonae
.
J. Gen. Microbiol.
 
134
,
253
262
.
[62]
Jiang
S.C
Kellogg
C.A
Paul
J.H
(
1998
)
Characterization of marine temperate phage-host systems isolated from Mamala Bay, Oahu, Hawaii
.
Appl. Environ. Microbiol.
 
64
,
535
542
.
[63]
Rashidan
K.K
Bird
D.F
(
2001
)
Role of predatory bacteria in the termination of a cyanobacterial bloom
.
Microb. Ecol.
 
41
,
97
105
.
[64]
Hahn
M.W
Höfle
M.G
(
2001
)
Grazing of protozoa and its effect on populations of aquatic bacteria
.
FEMS Microbiol. Ecol.
 
35
,
113
121
.
[65]
Šimek
K
Kojecká
P
Nedoma
J
Hartman
P
Vrba
J
Dolan
J.R
(
1999
)
Shifts in bacterial community composition associated with different microzooplankton size fractions in a eutrophic reservoir
.
Limnol. Oceanogr.
 
44
,
1634
1644
.
[66]
Langenheder
S
Jurgens
K
(
2001
)
Regulation of bacterial biomass and community structure by metazoan and protozoan predation
.
Limnol. Oceanogr.
 
46
,
121
134
.
[67]
Pernthaler
J
Posch
T
Šimek
K
Vrba
J
Amann
R
Psenner
R
(
1997
)
Contrasting bacterial strategies to coexist with a flagellate predator in an experimental microbial assemblage
.
Appl. Environ. Microbiol.
 
63
,
596
601
.
[68]
Suzuki
M.T
(
1999
)
Effect of protistan bacterivory on coastal bacterioplankton diversity
.
Aquat. Microb. Ecol.
 
20
,
261
272
.
[69]
Long
R.A
Azam
F
(
2001
)
Antagonistic interactions among marine pelagic bacteria
.
Appl. Environ. Microbiol.
 
67
,
4975
4983
.
[70]
Fasham
M.J.R
Boyd
P.W
Savidge
G
(
1999
)
Modeling the relative contributions of autotrophs and heterotrophs to carbon flow at a Lagrangian JGOFS station in the northeast Atlantic: The importance of DOC
.
Limnol. Oceanogr.
 
44
,
80
94
.
[71]
Pernthaler
J
Glöckner
F.O
Unterholzner
S
Alfreider
A
Psenner
R
Amann
R
(
1998
)
Seasonal community and population dynamics of pelagic bacteria and archaea in a high mountain lake
.
Appl. Environ. Microbiol.
 
64
,
4299
4306
.
[72]
Glöckner
F.O
Fuchs
B.M
Amann
R
(
1999
)
Bacterioplankton compositions of lakes and oceans: a first comparison based on fluorescence in situ hybridization
.
Appl. Environ. Microbiol.
 
65
,
3721
3726
.
[73]
Brümmer
I.H.M
Fehr
W
Wagner-Döbler
I
(
2000
)
Biofilm community structure in polluted rivers: Abundance of dominant phylogenetic groups over a complete annual cycle
.
Appl. Environ. Microbiol.
 
66
,
3078
3082
.
[74]
Simon
M
Glöckner
F.O
Amann
R
(
1999
)
Different community structure and temperature optima of heterotrophic picoplankton in various regions of the southern Ocean
.
Aquat. Microb. Ecol.
 
18
,
275
284
.