OUP user menu

Microbiology of aquatic surface microlayers

Michael Cunliffe, Robert C. Upstill-Goddard, J. Colin Murrell
DOI: http://dx.doi.org/10.1111/j.1574-6976.2010.00246.x 233-246 First published online: 1 March 2011

Abstract

Aquatic surface microlayers are unique microbial ecosystems found at the air–water interface of all open water bodies and are often referred to as the neuston. Unambiguous interpretation of the microbiology of aquatic surface microlayers relies on robust sampling, for which several methods are available. All have particular advantages and disadvantages that make them more or less suited to this task. A key feature of surface microlayers is their role in regulating air–water gas exchange, which affords them a central role in global biogeochemistry that is only now being fully appreciated. The microbial populations in surface microlayers can impact air–water gas exchange through specific biogeochemical processes mediated by particular microbial groups such as methanotrophs or through more general metabolic activity such as the balance of primary production vs. heterotrophy. There have been relatively few studies of surface microlayers that have utilized molecular ecology techniques. The emerging consensus view is that aquatic surface microlayers are aggregate-enriched biofilm environments containing complex microbial communities that are ecologically distinct from those present in the subsurface water immediately below. Future research should focus on unravelling the complex interactions between microbial diversity and the ecosystem function of surface microlayers in order to better understand the important but complex role of microorganisms in Earth system processes.

Keywords
  • surface microlayers
  • microbial ecology
  • biogeochemistry

Most of all, perhaps, there is assurance in the fine dust of life that remains in the surface waters. Rachel Carson (1951) in The Sea Around Us

Introduction

Aquatic surface microlayers are just a few tens of micrometers deep at the air–water interface and are physicochemically distinct from the subsurface water below. Occurring on all bodies of water, surface microlayers are unique, yet widely distributed microbial ecosystems. Naumann introduced the term neuston to distinguish microorganisms associated with the air–water interface from the subsurface plankton (Naumann, 1917).

The microbiology of aquatic surface microlayers has been studied for almost a century (Naumann, 1917) and has been reviewed periodically over recent decades (Hardy et al., 1982; Sieburth et al., 1983; Maki et al., 1993, 2002), but an appreciation of its role in global-scale processes is only now emerging. This article presents a contemporary review with a focus on molecular microbial ecology and its links to global biogeochemistry, which consolidates recent contributions to surface microlayer research.

The physicochemical nature of aquatic surface microlayers

Surface microlayers occupy a unique position in the global system that implicates them strongly in global biogeochemical fluxes such as the air–sea exchange of reactive gases and particles. Consequently, there is renewed interest in their physicochemical and microbiological compositions and how these may vary in space and time. Surface microlayers result from the accumulation of both discrete molecules and larger particles at the air–water interface to form a film that extends into the subsurface water. Historically, the depth of a surface microlayer has not been well defined, having been determined either operationally by the prevailing sampling protocol or through some combination of measured changes in physicochemical or biological parameters. Nevertheless, early descriptions of surface microlayers envisaged a distinct entity with a stratified structure comprising an upper lipid layer containing highly surface-active molecules such as fatty acids, long-chain alcohols and lipids, overlying a protein–polysaccharide layer extending into subsurface waters (Fig. 1). Lipid layer components were typically considered to be of very low solubility and to contain hydrophobic moieties extending into the air (Norkrans et al., 1980). It is this basic structure that was first collectively referred to as the surface microlayer (Norkrans, 1980; Hardy, 1982; Hermansson, 1990).

1

Schematic model of the structure of the air–water interface. The classical model shows a two layered surface microlayer, with an upper ‘dry surfactant’ lipid layer and a lower ‘wet surfactant’ protein–polysaccharide layer. Below this is the bacterioneuston, distinct from the surface microlayer (Norkrans, 1980; Hardy, 1982; Hermansson et al., 1990). This model has been revised and instead at the interface is a gelatinous film, which is a heterologous matrix (Sieburth, 1983). The gel film is enriched with TEP (Wurl & Holmes, 2008; Cunliffe et al., 2009b) and associated microbial life is intimately attached (Cunliffe & Murrell, 2009). Adapted from Sieburth (1983), Maki (1993) and Hardy (1982).

This early view of lipids as important surface microlayer components has now been revised; lipids are no longer considered to be present in sufficient concentrations (Sieburth et al., 1983; Williams et al., 1986). A new model of the sea surface microlayer advanced by Sieburth (1983) described a ‘highly hydrated loose gel of tangled macromolecules and colloids’ that is produced from dissolved organic matter (DOM) (Fig. 1). Sieburth's proposal was based on a wealth of studies and included samples collected in the mid 1960s from a Sargasso Sea ‘slick,’ during a Trichodesmium bloom, which showed elevated amylolytic activity and no proteolytic or lipolytic activity (Sieburth & Conover, 1965). Marine microlayers collected off California and Peru were enriched with DOM and particulate organic matter (POM), but their C: N: P ratios were the same as those of subsurface water. An important conclusion was that microlayer particulate material could contribute to the formation of subsurface marine snow (Williams et al., 1967). Based on these findings and those of others, Sieburth (1983) concluded that carbohydrates must be a significant component of surface microlayers.

Twenty-five years on from Sieburth's pioneering work, Wurl & Holmes (2008) reported an enrichment of transparent exopolymer particles (TEP) in sea surface microlayers collected from several locations around Singapore (Table 1). TEP are sticky gel particles produced in the water column from the coagulation of DOM precursors and are operationally defined as particles formed from acidic polysaccharides that are stainable with alcian blue (Alldredge et al., 1993; Passow, 2002; Engel et al., 2004). TEP play a pivotal role in the marine carbon cycle by acting as substrates for the transfer of dissolved carbon to the particulate pool (Verdugo et al., 2004), readily forming aggregates in the water column with other particles such as detritus (Alldredge et al., 1993). TEP-induced aggregates consequently become incorporated into marine snow, ultimately sinking and transporting nutrients to the deep ocean (Alldredge & Gotschalk, 1988; Azam & Long, 2001).

View this table:
1

Comparative summary of the average enrichment factors of common parameters determined in several recent microlayer studies

Sampling detailsInorganicOrganicBiological
Marine/freshwaterSML sampling methodSML sampling depth (μm)SS sampling depth (m)AmmoniumNitrateNitritePhosphateDissolved organic carbonParticulate organic carbonDissolved organic nitrogenTransparent exopolymer particlesChlorophyll aBacterial abundanceArchaea abundanceVLPs
Santa Barbara Channel (Wurl et al., 2009)MGP501NDNDNDNDNDNDND1.7NDNDNDND
Johor Strait (Estuarine) (Wurl & Holmes, 2008)MGP501NDNDNDND1.6NDND1.8NDNDNDND
Singapore Strait (Wurl & Holmes, 2008)MGP501NDNDNDND1.3NDND1.3NDNDNDND
Subtropical Eastern Atlantic (Reinthaler et al., 2008)MGP≤8009.3422.21.5ND3.3NDND1NDND
Western Mediterranean Sea (Reinthaler et al., 2008)MGP≤800139533.11.8ND4.1NDND1NDND
Mediterranean Sea – Barcelona (Joux et al., 2006)MMS≤4400.5ND1.22.3NDNDNDND31ND3
Mediterranean Sea – Banyuls (Joux et al., 2006)MMS≤4410.5ND1.32.5NDNDNDND11ND1
South Pacific (Obernosterer et al., 2007)MMSND5NDNDNDND14NDND11NDND
Fjord mesocosm (Cunliffe et al., 2009b)MMS4000.8NDNDNDND1.3NDND2.58ND1NDND
Lowland lakes (two lakes) (Denmark) (Baastrup-Spohr & Staehr, 2009)FMS292ND1.6NDNDND1.11NDND1NDNDND
Montane lakes (10 lakes) (Pyrenees) (Auguet & Casamayor, 2008)FMS4001NDNDNDNDNDNDNDNDND13ND
Montane lakes (six lakes) (Alps) (Hörtnagl et al., 2010)FMS9000.2NDNDNDND1.2ND1.1NDND1NDND
  • ND, not determined in the study; M, marine; F, freshwater; MS, mesh screen; GP, glass plate; SML, surface microlayer; SS, subsurface.

Not all TEP sink; they can also ascend the water column (Azetsu-Scott & Passow, 2004). Seawater containing TEP precursors collected from the Santa Barbara Channel were used to generate TEP de novo, which was subsequently shown to be positively buoyant. The ascending TEP were able to aggregate with latex spheres that mimicked solid particles, such as microbial cells, thus demonstrating a mechanism for surface microlayer formation.

Microlayer TEP have so far been quantified in situ in a small number of locations (Table 1). TEP were quantified in experimental mesocosms using Norwegian fjord waters augmented with nitrate and phosphate to stimulate phytoplankton blooms (Cunliffe et al., 2009b). The results showed significant TEP enrichment (P≤0.04) in surface microlayer samples compared with subsurface water (Fig. 2). TEP are also produced by freshwater biota (Passow, 2002). As yet undescribed homologous processes therefore also seem likely in freshwater systems.

2

Surface microlayer enrichment of TEP at the end of an artificially induced phytoplankton bloom mesocosm experiment. The values presented are means (n=3). Subsurface water (SS) and surface microlayer (SML). Figure modified from Cunliffe et al. (2009b).

A characteristic of surface microlayers is that many of their various components occur at higher concentrations than in subsurface water. This enrichment effect is numerically expressed as the enrichment factor (EF), which is defined simply as the concentration of a specific constituent in the microlayer divided by its concentration in subsurface water directly below the microlayer. Many studies have quantified the EFs of a diverse range of microlayer components in different aquatic systems (Table 1) (Williams et al., 1967; Carlson et al., 1983; Williams et al., 1986; Munster et al., 1998; Agogue et al., 2004; Joux et al., 2006; Obernosterer et al., 2007; Reinthaler et al., 2008; Wurl & Holmes, 2008; Baastrup-Spohr & Staehr, 2009; Cunliffe et al., 2009b; Wurl et al., 2009; Hörtnagl et al., 2010). Direct comparison between studies is problematic because there is currently no consensus as to the most appropriate strategy for sampling surface microlayers; different microlayer samplers yield varying operationally defined depths. Moreover, subsurface reference depths used to define EF also vary considerably, typically between 0.03 and 5 m (Table 1). However, it is important to note that for most biological oceanographic parameters considered to date, a subsurface reference depth of≤5 m has no significant effect on the resulting EF of marine surface microlayers (Agogue et al., 2004).

In comparison with the subsurface water, microlayers are physically more stable because of surface tension at the air–water interface (Hardy, 1982). Samples collected from the Santa Barbara Channel showed the sea surface microlayer to be stable under typical oceanic wind conditions (Wurl et al., 2009). Surface-active substances, total dissolved carbohydrates, chromophoric DOM and TEP were significantly enriched at the air–sea interface at wind speeds >6 m s−1, and up to a maximum observed wind speed of 9.6 m s−1 (Wurl et al., 2009). As the average wind speed over the oceans is 6.6 m s−1 (Archer & Jacobson, 2005), it must be concluded that the sea surface microlayer is ubiquitous under most prevailing weather regimes.

Surface tension and wind action can also significantly contribute to the spreading dynamics of surface microlayers through entraining and further concentrating bacterial cells within the microlayer film-spreading front (Hale & Mitchell, 1997). The ability of surface films to influence the nature of water bodies has been known for some time (Pockels et al., 1891). For example, surface films can significantly modify surface tension and dampen capillary waves (Wei & Wu, 1992).

The location of surface microlayers makes them a highly dynamic system (Fig. 3). Atmospheric inputs to surface films include wet/dry deposition, air–sea gas and aerosol transfer. In some ocean areas, the deposition of terrestrial dust can significantly impact surface productivity by contributing growth-limiting nutrients such as phosphorus, nitrogen and iron (Behrenfeld et al., 2006). Exchange with the atmosphere is strongly influenced by the physicochemical and microbiological nature of surface microlayers. For example, surface films reduce gas transfer velocities at the air–sea interface by dampening capillary waves (Frew et al., 1990; Frew et al., 2005). Also, the composition of marine aerosols formed from bursting bubbles at the sea surface changes in response to the occurrence of dense microlayers containing organic compounds originating from phytoplankton blooms (O'Dowd et al., 2004; Russell et al., 2010). Marine aerosols also contain microorganisms, including viruses, proteinaceous material and gel particles (Kuznetsova et al., 2005). Bubble bursting and aerosol formation is an important transport mechanism for microlayer components (Russell et al., 2010).

3

Inputs, outputs and processes within aquatic surface microlayers. Adapted from Hardy (1982) and Maki (2002).

Water column processes regulate the accumulation of material in surface microlayers (Fig. 3). Stratification and TEP formation drives the upward flux of material via positive buoyancy and promotes aggregation (Azetsu-Scott & Passow, 2004; Wurl et al., 2009). Microbial communities are highly active within microlayers. Surface microlayer samples collected from the Atlantic Ocean and Mediterranean Sea had significantly higher rates of bacterial respiration than subsurface water (Reinthaler et al., 2008). Extracellular enzyme activity, which converts POM to DOM, is enhanced in microlayers (Kuznetsova & Lee, 2001) and DOM formed within the microlayer can be subsequently exported to subsurface water below.

Unlike those in the underlying waters, organisms within surface microlayers receive maximal UV radiation, which has the potential to cause direct DNA damage or indirect damage via the formation of destructive intermediates such as reactive oxygen species. There are conflicting reports of the effects of UV radiation on the neuston. Analysis of amino acid utilization in surface microlayers off California showed neuston communities not to be measurably affected by either visible or UV radiation (Carlucci et al., 1985), while a similar study in Chesapeake Bay showed deleterious effects from both (Bailey et al., 1983). A culture-based study found that resistance to UV was not significantly different for bacterioneuston and bacterioplankton isolates from the Mediterranean Sea (Agogue et al., 2005b). Nevskia ramosa is a bacterium that has been isolated from freshwater surface microlayers and identified as being specifically adapted for life in the neuston (Glockner et al., 1998; Sturmeyer et al., 1998; Pladdies et al., 2004). A strain of the species that was isolated from a freshwater water lake was shown to be more sensitive to UV radiation than Escherichia coli; however, it had a very effective photorepair mechanism (Sturmeyer et al., 1998).

Photodegradation of DOM is another important UV effect in surface microlayers and it may be a primary transformation mechanism (Fig. 3). Photochemically produced low-molecular-weight carbonyl compounds such as acetone and acetaldehyde are readily available for microbial uptake within the microlayer, which can therefore reduce mixing with subsurface water or escape into the atmosphere via air–sea exchange (Zhou & Mopper, 1997).

To summarize, what are aquatic surface microlayers and why is this microbial habitat different from subsurface water? Physical forces and the presence of a biogenic gelatinous film layer constitute a microbial habitat at the air–water interface, where compounds and particles can concentrate. Surface microlayer habitats are unique as they interact with both the atmosphere and the hydrosphere simultaneously.

Sampling of surface microlayers

The most commonly used devices for surface microlayer collection are mesh screens, often referred to as Garret screens (Garrett, 1965), glass plates (Harvey, 1965; Harvey & Burzell, 1972) and membranes (Crow et al., 1975; Kjelleberg et al., 1979), all of which have inherent advantages and disadvantages (Fig. 4) (Table 2). Mesh screens rely on the collection of water in the interstitial spaces of a mesh. Metal or plastic mesh is stretched over and secured to a hand-held frame. The mesh screen is oriented horizontally and lowered through the microlayer into the subsurface water before being slowly withdrawn in the same way (Fig. 4). Water retained within the mesh is then poured into a sampling vessel for processing as required. Glass plates (typically 20–30 cm2) exploit the adhesion of a surface microlayer sample to a clean glass surface. Unlike the mesh, the glass plate is oriented vertically, lowered through the microlayer into the subsurface water and slowly raised back out. The adhered sample is then removed using a wiper blade (e.g. neoprene). Surface microlayers can also be sampled by adhesion to sterile membrane filters placed directly onto the water surface (Crow et al., 1975; Kjelleberg et al., 1979). There are distinct advantages over mesh screens and glass plates in that the membranes do not come into contact with the subsurface water (Fig. 4). The potential for contamination is virtually eliminated using forceps; individual membranes are lifted from the water surface with the surface microlayer attached and placed into storage vessels for processing. Various types of membrane material are available; to date, both polycarbonate and PTFE have been used (Table 2) (Crow et al., 1975; Kjelleberg et al., 1979; Cunliffe et al., 2009a).

4

Mesh screen and membrane surface microlayer samplers in use.

View this table:
2

Summary of the advantages and disadvantages of the three most commonly used surface microlayer sampling methods

MethodSampling depthAdvantagesDisadvantages
Mesh screen150–400 μm (Carlson et al., 1982; Cunliffe et al., 2009a)Collects a relatively large sample volume, which facilitates more extensive downstream analysisAs the sampling depth is deep, the sample collected can be a mixture of both surface microlayer and subsurface water (Cunliffe et al., 2009a)
Potential for contamination between sampling events
Difficult to standardize operation (Hatcher & Parker, 1974)
Glass plate20–150 μm (Harvey, 1965; Harvey & Burzell, 1972; Cunliffe et al., 2009a)Same as mesh screensSame as with mesh screens
Cheap and easily available
MembranesPolycarbonate: 35–42 μm (Franklin et al., 2005; Cunliffe et al., 2009a) PTFE: 6 μm (Cunliffe et al., 2009a)Cheap and easily available Collects exclusively microlayer sample Relatively easier to standardize operation No bias associated with molecular analysis of bacterial communities (Cunliffe et al., 2009a)Collects a relatively small sample volume, which limits downstream analysis Difficult to use in high winds Potential bias associated with bacterial cell counts (Agogue et al., 2004)
  • * Sampling depth is determined from the volume of water sampled and the surface area of the sampler.

  • PTFE, polytetrafluoroethylene (Teflon®).

Agogue et al. (2004) compared methods for sampling surface microlayers of the Mediterranean Sea for subsequent estimates of chlorophyll concentrations, bacterial production, bacterial cell numbers (total and cultivable), viruses, Synechococccus cells, picoeukaryotes, nanoeukaryotes, flagellates and ciliates. They concluded that there was no sampling bias associated with either the mesh screen or the glass plate. However, PTFE and polycarbonate membranes showed a bias in the numbers of bacterial cells (total and cultivable) removed from the microlayer.

The same sampling methods were later re-evaluated specifically for molecular microbial analysis of the surface microlayer. DNA was extracted from surface microlayer and cognate subsurface (0.4 m below the surface) samples from the Tyne Estuary, UK (salinity 7) (Cunliffe et al., 2009a). Bacteria 16S rRNA gene amplicons from the DNA samples were profiled using denaturing gradient gel electrophoresis (DGGE) and compared objectively using computer analysis software. This revealed the bacterial community profiles from surface microlayer samples collected using both membrane types to be 78% similar to those collected from the subsurface water and surface microlayer samples collected using the mesh screen and glass plate. In order to establish whether this difference was due to selective bias of the membranes, as reported by Agogue et al. (2004), subsurface water was also sampled using the membranes and processed in the same way. Bacteria 16S rRNA gene profiles obtained from subsurface membrane samples clustered with those from the subsurface, mesh screen and glass plate sample. From this, it was concluded that membranes do not show a selective sampling bias in terms of the microbial community structure.

The studies of Agogue et al. (2004) and Cunliffe et al. (2009a) are the only ones to date to have compared surface microlayer sampling strategies with a specific focus on microbiological parameters. As these were performed in different aquatic ecosystems and using somewhat different approaches, it is not possible to definitively comment on state-of-the-art surface microlayer sampling. Clearly, this is an area that deserves further work, the ultimate aim of which should be to unequivocally establish which sampling protocols are the most suitable for specific applications.

Air–water gas exchange

Aquatic environments are significant sources and/or sinks for a wide range of globally significant trace gases, including carbon dioxide (CO2), carbon monoxide (CO), methane (CH4), nitrous oxide (N2O), halocarbons and a range of sulfur-containing gases including dimethylsulfide (Fig. 3). The air–water exchange of these gases is critical to global element cycles and climate change. An important focus of recent microlayer research has therefore been to establish the specific role of the sea surface microlayer in air–sea gas exchange.

The sea surface microlayer has long been implicated indirectly in both the consumption and the production of atmospheric trace gases (Conrad & Seiler, 1988), and there is now supporting evidence that general microbial activity and metabolism at the air–sea interface can regulate air–sea gas exchange. The ocean is a major sink for anthropogenic CO2 (Sabine et al., 2004). Air–sea CO2 exchange is controlled by the disequilibrium between the partial pressure of CO2 at the ocean surface and in the atmosphere (ΔpCO2). A study in the subtropical Atlantic showed that ΔpCO2 was dependent on net microbial community metabolism in the surface layer (≤2 cm) and was independent of the metabolism of the subsurface community (5 m below) (Calleja et al., 2005). Respiration and gross primary production of the microbial community within the top 2 cm of the ocean were, respectively, more than sevenfold and tenfold greater than those at 5 m depth. Net autotrophic metabolism within the uppermost 2 cm created a CO2 demand and drove CO2 uptake at some locations, while at others, net heterotrophy led to surface CO2 supersaturation and CO2 emission to the atmosphere (Calleja et al., 2005). In the Baltic Sea, microelectrode measurements of aggregates at the air–water interface during a cyanobacterial bloom showed high rates of photosynthesis. Oxygen fluxes (also determined using microelectrodes) were higher at the air–sea interface than between the aggregates and underlying water (Ploug et al., 2008).

Despite increasing interest in the role of the bacterioneuston in air–sea gas exchange, there have only been three surface microlayer studies to date in which gas transfer rates were formally measured under controlled conditions. Using a free-floating gas exchange box in the tropical Atlantic Ocean, Conrad & Seiler (1988) found a significant mismatch between the invasive and the evasive exchange of CH4, CO, N2O and hydrogen, which they ascribed to microbial gas consumption by the bacterioneuston, the only plausible gas sink during the experiments. This early work provided the impetus for further study in which Frost (1999) similarly found an invasion–evasion mismatch of up to 8% for CH4 in the coastal North Sea. Later work in a laboratory gas exchange tank with added methanotrophs confirmed the potential for active bacterioneuston control of trace gas exchange. Air–water CH4 exchange differed by up to 10% relative to both N2O and the inert tracer sulfur hexafluoride (all normalized to constant diffusivity), consistent with active metabolic control of CH4 exchange by the bacterioneuston (Upstill-Goddard et al., 2003).

The common conclusion from these studies is that the bacterioneuston is closely involved in the cycling of at least some climatically active trace gases and that it is potentially both a small gas source and a small gas sink, dependent on the prevailing microbial and biogeochemical regimes.

Molecular microbial ecology of aquatic surface microlayers

Early studies that utilized molecular methodologies to study microbial ecology in the sea surface microlayer offered conflicting conclusions (Agogue et al., 2005a; Franklin et al., 2005). Agogue et al. (2005a) compared, using single-strand conformation polymorphism (SSCP) of Bacteria 16S rRNA genes, the bacterioneuston community structure with the bacterioplankton community structure at two sample sites: the oligotrophic Bay of Banyuls-sur-Mer, France, and a moderately eutrophic area off Olympic Harbour in Barcelona, Spain (Table 3). In their SSCP profiles, the authors could detect no consistent difference between the bacterioneuston and the bacterioplankton at either site (Agogue et al., 2005a). By contrast, surface microlayer samples collected off the UK North Sea coast contained a distinct bacterioneuston community (Franklin et al., 2005). Analysis of Bacteria 16S rRNA gene libraries (≈500 clones) constructed from DNA isolated from the surface microlayer and subsurface water showed the bacterioneuston to be distinct from the bacterioplankton and dominated by only two genera: Vibrio spp. (68% of clones) and Pseudoalteromonas spp. (21% of clones) (Franklin et al., 2005).

View this table:
3

Summary of molecular microbial ecology studies of aquatic surface microlayers

Study locationSML sampling method and depth (μm)Domain studiedAnalysis detailsResults summaryReferences
North SeaPC40Bacteria16S rRNA gene librariesBacterioneuston was dominated by two genera (Vibrio and Pseudoalteromonas)Franklin et al. (2005)
Mediterranean SeaMSNDBacteria16S rRNA gene SSCPNo consistent difference between bacterioneuston and bacterioplankton communitiesAgogue et al. (2005a)
Pacific Ocean (Hawaiian Islands)PC40Bacteria and Archaea16S rRNA gene DGGEBacterioneuston communities on different sides of Oahu Island (>10 km apart) were more similar than cognate subsurface bacterioplankton communities (0.4 m below)Cunliffe et al. (2009a)
Fjord mesocosm (Norway)PC40Bacteria16S rRNA gene DGGE and T-RFLPA distinct and reproducible bacterioneuston formed. Dominant taxa included the families Flavobacteriaceae and AlteromonadaceaeCunliffe et al. (2009c)
MS400
Fjord mesocosm (Norway)MS400Eukarya18S rRNA gene DGGEDistinct and reproducible protist communities formed in the microlayer, dominated by SML-specific Cercozoa populationsCunliffe & Murrell (2010)
Blyth Estuary (UK)PC40Bacteria and Archaea16S rRNA gene clone libraries and DGGESML DGGE profiles for both Bacteria and Archaea were different to subsurface DGGE profilesCunliffe et al. (2008)
Tyne Estuary (UK)PC42Bacteria16S rRNA gene DGGEDeeper surface samplers (i.e. GP and MS) can collect microlayer samples that are diluted with underlying subsurface water. As a result, analysis of samples can underestimate the differences between microlayer and subsurface water communitiesCunliffe et al. (2009a)
PTFE6
GP150
MS400
Montane Lakes (Spain)MS400ArchaeaCARD-FISH and 16S rRNA gene librariesArchaea in the neuston were mainly Crenarchaeota, subsurface Archaea were mainly EuryachaeotaAuguet & Casamayor (2008)
Montane Lakes (Alps)MS900BacteriaCARD-FISHBoth the surface microlayer and the subsurface water were dominated by BetaproteobacteriaHörtnagl et al. (2010)
  • PC, polycarbonate membrane; PTFE, polytetrafluoroethylene (Teflon®) membrane; MS, mesh screen; GP, glass plate; T-RFLP, terminal restriction fragment length polymorphism; CARD-FISH, catalyzed reporter deposition FISH.

More recently, surface microlayer samples were collected from three sites in the Pacific Ocean, close to the Hawaiian Island of Oahu (Table 3) (Cunliffe et al., 2009a). DNA was extracted from the surface microlayer and subsurface water samples and 16S rRNA genes of Bacteria and Archaea were amplified and profiled using DGGE. The similarity of the DGGE profiles was determined using unweighted pair with arithmetic mean dendrograms that were calculated from similarity coefficients. Bacterioneuston DGGE profiles from two sampling sites on opposite sides of the island were more similar to each other than the cognate bacterioplankton DGGE profiles. Corresponding Archaea 16S rRNA gene DGGE profiles from the same samples did not show this result (Cunliffe et al., 2009a). The conclusion to draw from these data is that Bacteria evidently respond more readily to the sea surface microlayer environment than do Archaea.

One other recent study involved a mesocosm comparison of microbial diversity in the bacterioneuston and the cognate bacterioplankton (Table 3) (Cunliffe et al., 2009c). Although mesocosm data should be interpreted with caution because they may not completely emulate the processes occurring in open water in situ, they can nevertheless facilitate carefully controlled and replicated experiments designed to resemble, as closely as possible, natural plankton successions (Davis et al., 1982). The study of Cunliffe et al. (2009c) compared the dynamics of the bacterioneuston and the bacterioplankton during an artificially induced phytoplankton bloom. Surface microlayer samples and cognate subsurface water samples (sampling depth 0.75 m) were collected at the start, middle and end of the bloom (Fig. 5). Two surface microlayer sampling methods were used: mesh screen (sampling depth ∼400 μm) and polycarbonate membranes (sampling depth ∼40 μm). Bacterioneuston and bacterioplankton communities in the samples were profiled using both DGGE and terminal restriction fragment length polymorphism of Bacteria 16S rRNA genes. Throughout the experiment, a distinct bacterioneuston community formed that was conserved between the replicated mesocosms (Cunliffe et al., 2009c).

5

Bacteria 16S rRNA gene DGGE profiles from the start, middle and end of a mesocosm experiment. DGGE profiles show each replicate (n=3) from the subsurface water (SS) and from the surface microlayer sampled using a mesh screen (MS) and polycarbonate membranes (PC). Beside each DGGE profile is the associated unweighted pair with arithmetic mean dendrogram showing the similarity of the lanes in the DGGE profiles. The arrows show which DGGE bands were excised and sequenced. Copyright © American Society for Microbiology, Applied and Environmental Microbiology, 75, 2009, 7173–7181 (doi:10.1128/AEM.01374-09).

The most dominant bands in the DGGE profiles of mesocosm bacterioneuston were sequenced and found to be similar to two families: the Flavobacteriaceae (Dokdonia sp. and Krokinobacter sp.) and the Alteromonadaceae (Alteromonas spp. and Glaciecola sp.) (Fig. 5) (Cunliffe et al., 2009c). Alteromonas and Glaciecola have previously been shown to dominate surface microlayer samples collected from a UK North Sea estuary (salinity 31) (Cunliffe et al., 2008). Pseudoalteromonas spp., closely related to the genera Alteromonas and Glaciecola (i.e. they share the same order: Alteromonadales), was a dominant component of sea surface microlayer samples collected from the coastal North Sea (Franklin et al., 2005).

The proposed role for the bacterioneuston in the regulation of air–water gas exchange (Conrad & Seiler, 1988; Frost et al., 1999; Upstill-Goddard et al., 2003), as outlined earlier, requires the presence of specific microlayer bacterial populations with the necessary metabolism to measurably affect trace gas concentrations and subsequent exchange fluxes. Estuaries are significant sources of CH4 (Upstill-Goddard et al., 2000) and CO (Stubbins, 2001). The diversity of functional genes that encode subunits of methane monoxygenase (mmoX) and carbon monoxide dehydrogenase (coxL) in estuarine surface microlayer samples (sampling depth ∼40 μm) was different from the diversity of the same genes in subsurface water (sampling depth 0.4 m) (Table 3) (Cunliffe et al., 2008). This indicates that the required necessary potential for the consumption of these trace gases is at least present in the estuarine surface microlayer and that the bacterial populations there may be specifically adapted for this environment.

Even fewer studies of molecular microbial ecology have been conducted on freshwater systems. Auguet & Casamayor (2008) conducted a detailed study of 10 oligotrophic lakes in the Pyrenean Mountain range. Archaea communities in the microlayer were compared with subsurface communities using catalyzed reporter deposition FISH and 16S rRNA gene sequence analysis (Table 3). Archaea populations in the microlayers were mainly Crenarchaeota, whereas subsurface populations mainly comprised of Euryarchaeota. Within the lake microlayers, Archaea cells determined using 4′,6-diamidino-2-phenylindole counts constituted up to 37% of the total cell numbers, making them a numerically significant component of the neuston (Auguet & Casamayor, 2008). We propose that the microbiological community should now adopt a new term to describe Archaea associated with aquatic surface microlayers: the archaeoneuston.

Joux et al. (2006) have shown that virus-like particles (VLP) are also enriched in the sea surface microlayer (Table 1). At the two coastal Mediterranean sites, VLP enrichment was generally low (mean EF=1.1); however, some ‘hotspots’ had larger enrichments (mean EF=5.1). Compared with the bacterioneuston and the archaeoneuston, even less is known about the virioneuston. Because of the intense exposure of the sea surface microlayer to UV radiation discussed earlier, the virioneuston could be inactive or at least UV disrupted. Virioplankton are often attached to subsurface aggregates. For example, River Danube subsurface aggregates (sampling depth 30 cm) had attached virus populations of up to 5.4 × 109 individuals cm−3 (Luef et al., 2009). Viruses no doubt are attached to microlayer aggregates also and are subsequently transported via sprays and aerosols (discussed above) (Kuznetsova et al., 2005).

Food webs in surface microlayers

Trophic interactions or food webs in aquatic environments are exemplified by the microbial loop, in which protists such as flagellates and ciliates prey on bacterial and archaeal cells and are in turn preyed upon by relatively larger aquatic organisms, such as copepods (Pomeroy, 1974; Azam et al., 1983). Trophic interactions influence the structure of microbial loop communities. For example, selective protist grazing can have profound structuring effects on their bacterial prey communities as protists can target particular bacterial groups such as those in a specific size range (Pernthaler, 2005).

The amoeba Acanthamoeba actively feed on bacterial cells at the water–air interface of freshwater ponds (Preston et al., 2001). Microscopical analysis of Acanthamoeba at the water–air interface shows that amoebae move in a similar way to amoebae on solid surfaces. Oxytricha bifaria, a freshwater hypotrich ciliate, is also able to move across bacterial films at the air–water interface as if it were solid substratum and graze on the resident bacterioneuston (Ricci et al., 1991).

Microscopical surveys of sea surface microlayers have revealed the presence of both motile and sessile forms of flagellate and ciliate protists, indicating that complex protist communities are present (Sieburth, 1983; Joux et al., 2006). To our knowledge, only one published study has used molecular biological methods to determine the diversity of protists in surface microlayers (Table 3). Replicate samples collected from the surface microlayer (sampling depth ∼400 μm) and subsurface water (sampling depth 0.75 m) during a phytoplankton bloom experiment in a fjord mesocosm were analyzed using DGGE of Eukarya 18S rRNA gene amplicons. Protist community structure in the surface microlayer samples was different from that in subsurface water, with two taxa appearing to dominate the microlayer samples: Cercozoa and, to a lesser extent, Ciliophora (Cunliffe & Murrell, 2010). This study implies that the protist components of the microbial loop in the fjord surface microlayer are significantly different from those in subsurface water.

Protist grazing in surface microlayers may also contribute to the physicochemical structure of the film, including altering air–water gas exchange properties (discussed above). Ciliates in particular are able to produce significant amounts of surface-active organic compounds that are part of the surface microlayer film (Kujawinski et al., 2002).

Microbial niches in the surface microlayer – the importance of attached cells

FISH analysis of bacterioneuston communities in samples collected from a freshwater pond showed that bacterial cells are more aggregated in the microlayer than they are in subsurface water (Fig. 6) (Cunliffe & Murrell, 2009).

6

Aggregation of bacterial cells in the surface microlayer of a freshwater pond. Bacterial cells in the subsurface water from 0.4 m below the surface are more dispersed. The bacterial communities were challenged with FISH probes targeting almost all known Bacteria (EUBI-III-Cy5), Betaproteobacteria (Bet42a-Cy3) and Gammaproteobacteria (Gam42a-FLOUS). Figure modified from Cunliffe & Murrell (2009).

Aggregates in the water column are readily colonized by microorganisms (Smith et al., 1992; DeLong et al., 1993; Passow & Alldredge, 1994; Crump et al., 1999; Verdugo et al., 2004), and both attachment and detachment can be rapid. Bacterial attachment to artificial aggregates (agar spheres) had typical residence times of around 3 h (Kiørboe et al., 2002). Attached bacterioplankton communities in the water column have a higher temporal variability and diversity compared with the free-living bacterial community (Rooney-Varga et al., 2005). Specific attached bacterial communities can develop (DeLong et al., 1993; Crump et al., 1999), and following bacterial colonization, bacterivorous flagellates and ciliates can also attach (Kiørboe et al., 2003).

The extracellular enzymatic conversion of POM into DOM in the water column by attached bacterial cells is uncoupled and results in substantial quantities of DOM being released (Azam & Long, 2001). Microlayers are enriched with aggregates, DOC and show elevated levels of extracellular enzyme activity (Kuznetsova & Lee, 2001). Aggregates in the surface microlayer probably undergo comparable bacterial attachment/detachment processes and are dissolved in a manner similar to subsurface aggregates. The significance of attached microbial cells to buoyant aggregates in the microlayer is proposed to be a distinguishing ecological characteristic of the neuston (Fig. 1).

Future perspectives

The marine environment is vast, covering >70% of the surface of the Earth, but there have only been a few microbiological studies of the sea surface microlayer using molecular biology techniques and these, in total, have covered <1 km2 (Table 3). Efforts must therefore focus on a wider range of environments and involve sustained monitoring programs in order to gain a more comprehensive understanding of this vast ecosystem. A continuing challenge is to better understand the links between microbial diversity and ecosystem function. For surface microlayer research to make progress in the future, multidisciplinary studies that acknowledge the close links between microbiological and biogeochemical processes should be high on the research agenda.

Acknowledgements

We acknowledge the generous support of the UK Natural Environment Research Council through its Surface Ocean Lower Atmosphere Study (SOLAS) directed program. We also extend our heartfelt thanks to all those colleagues involved in our various surface microlayer projects and without whom our research would not have been possible. Special thanks are due to Hendrik Schäfer for kindly providing the photographs used in Fig. 4.

Footnotes

  • Editor: Ferran Garcia-Pichel

References

View Abstract