Introduction

The marine Roseobacter clade (MRC) is a monophyletic group (>87% identity in 16S rRNA genes) of bacteria within the family Rhodobacteraceae (Buchan et al., 2005). The MRC are an ecologically significant clade, representing up to 20% of bacterial cells in marine coastal waters (Buchan et al., 2005; Sowell et al., 2011). The use of both ‘omics’ and physiological experimentation has revealed that MRC bacteria harbour an extraordinary ability to metabolise a wide range of substrates to support their growth (Moran et al., 2004; Buchan et al., 2005; Newton et al., 2010). The ecological success of this clade may be in part due to their ability to utilise a variety of metabolic strategies to generate cellular energy, which allows for the more efficient utilisation of carbon (assimilation versus dissimilation; Sorokin et al., 2005; Moran and Miller, 2007; Boden et al., 2011b). For these reasons, the MRC bacteria have essential roles in both carbon and sulphur cycling, and more recently, nitrogen cycling (Buchan et al., 2005; Chen et al., 2011) within the marine environment. Ruegeria pomeroyi DSS-3 (basonym, Silicibacter pomeroyi DSS-3) is a member of the MRC, which was isolated off the coast of Georgia through enrichment with dimethylsulphoniopropionate (González et al., 2003). The genome of R. pomeroyi was sequenced in 2004 (Moran et al., 2004), and this bacterium is now a model organism enabling a better understanding of how and why marine bacteria metabolise a wide range of substrates (Moran et al., 2004; Cunliffe, 2012; Todd et al., 2012; Lidbury et al., 2014).

Trimethylamine (TMA) and trimethylamine N-oxide (TMAO) form part of the methylated amine (MA) pool found within the marine environment (King, 1984; Gibb et al., 1999; Gibb and Hatton, 2004). In the marine environment, TMAO is a compatible osmolyte for a variety of marine biota (Yancey et al., 1982; Treberg et al., 2006) and TMA is produced from the reduction of compatible osmolytes, such as glycine betaine, TMAO and choline (King, 1984; Arata et al., 1992). TMA production can also occur under aerobic conditions through oxidation of carnitine (Zhu et al., 2014), which may help explain the presence of TMA in oxygenated marine surface waters (Carpenter et al., 2012). Standing concentrations of TMA range from low nanomolar (nM) in coastal and open ocean surface waters to low micromolar (μM) in the pore water of marine sediments (Gibb et al., 1999; Fitzsimons et al., 2001; Gibb and Hatton, 2004). The ocean:atmospheric flux of MAs is important, as they can form aerosols, and are precursors for climate-active gases, such as nitrous oxide (Quinn et al., 1988; Carpenter et al., 2012). Furthermore, MAs may represent a significant proportion of the dissolved organic nitrogen pool (King, 1984; Gibb et al., 1999; Gibb and Hatton, 2004), the second largest sink of nitrogen (N) in the oceans after gaseous nitrogen (N2; Capone et al., 2008) and may help bacteria overcome severe competition for N, which is thought to be one of the limiting nutrients for ocean productivity (Zehr and Kudela, 2011).

Chen (2012) showed that representatives of the MRC can grow on TMA. While those MRC bacteria harbouring the genes necessary for TMA oxidation could all utilise TMA as a sole N source to support heterotrophic growth, only representatives from the genus Roseovarius of the MRC could grow on TMA as a sole carbon (C) source (methylotrophy). All marine bacteria that possess a functional TMA monooxygenase (Tmm; Chen et al., 2011) and a TMAO demethylase (Tdm; Lidbury et al., 2014) also have the genes necessary for the complete oxidation of methyl groups, cleaved off during catabolism of TMA (Sun et al., 2011; Chen, 2012; Halsey et al., 2012). Two different oligotrophic bacteria from the Alphaproteobacteria (Candidatus Pelagibacter ubique HTCC1062) and Betaproteobacteria (Methylophilales sp. HTCC2181), respectively, can couple TMAO oxidation to ATP production, which results in stimulation of growth (Sun et al., 2011; Halsey et al., 2012); however, these organisms fundamentally differ from members of the MRC. R. pomeroyi has the genes required for TMA catabolism (Figure 1) and can grow on TMA as a N source, but not on a sole C source, due to a lack of genes required for C assimilation via the serine cycle (Chen et al., 2011; Chen, 2012). Here we test the hypothesis that the oxidation of MAs is coupled to ATP production, providing an ecophysiological advantage to heterotrophic bacteria. We also test the hypothesis that metabolism of MAs can provide a source of remineralised N in the form of ammonia, which can be utilised by another marine bacterium.

Figure 1
figure 1

Proposed model for methylated amine catabolism in the marine bacterium Ruegeria pomeroyi DSS-3. Text in brackets denotes the locus tag of the corresponding gene in R. pomeroyi. CH2=H4F, 5,10-methylene tetrahydrofolate; CO2, carbon dioxide; DMA, dimethylamine; Dmm, dimethylamine monooxygenase; GMA, gamma-glutamylmethylamide; GmaS, gamma-glutamylmethylamide synthetase; MgdABCD, N-methylglutamate dehydrogenase; MgsABC, N-methylglutamate synthase; MMA, monomethylamine; NMG, N-methylglutamate; TMA, trimethylamine; TMAO, trimethylamine N-oxide; TmoXWV, ATP-dependent TMAO transporter (Lidbury et al., 2014).

Full size image

Materials and methods

Growth conditions

R. pomeroyi DSS-3 was maintained in the laboratory on marine agar 2216 (Difco, Sparks, MD, USA). Gentamicin (10 μg ml−1) was added to maintain mutant strains Δtmm::Gm and Δtdm::Gm (Lidbury et al., 2014). For all experiments R. pomeroyi (wild type and mutants) was grown in marine ammonium mineral salts (MAMS) medium (Schäfer, 2007) using glucose as the sole carbon source. MAMS medium was modified from Schäfer (2007) and contained (per litre): NaCl, 20 g; (NH4)2SO4, 1 g; MgSO4·7H2O, 1 g; CaCl2·2H2O, 0.2 g; FeSO4·7H2O, 2 mg; Na2MoO4·2H2O, 20 mg; KH2PO4, 0.36 g; K2HPO4, 2.34 g; plus 1 ml of SL-10 trace metals solution (Schäfer, 2007). Vitamins were prepared as described previously (Chen, 2012). Continuous culture work was performed using a glucose-limited (5 mM) chemostat using the methods previously described by Boden et al. (2011b). To avoid precipitants forming in the medium during autoclaving, NH4Cl was substituted for (NH4)2SO4. Steady state was achieved after five dilutions and the dilution rate was set at 0.05 h−1.

Citreicella sp. SE45 (a gift from Dr Alison Buchan) was also maintained and grown using the same methods. Both strains were incubated at 30 °C on a rotary shaker (150 r.p.m). Methylomonas methanica MC09 (Boden et al., 2011a) was maintained on MAMS plates using methane (5%) as the sole carbon source. For growth experiments, M. methanica was grown in MAMS medium using methanol (2 mM) as the sole carbon source and incubated at 25 °C.

Determination of biomass (mg dry weight l−1)

R. pomeroyi cultures (500 ml) were grown on glucose and ammonium with or without TMA (3 mM) to an optical density at 540 nm(OD540) ∼1.4. Cells were diluted to 0%, 25%, 50%, 75% (n=3) in MAMS and the OD540 was recorded prior to filtration onto 0.22-μm nitrocellulose filter pads (Millipore, Darmstadt, Germany). Cells trapped on the filter pads were washed twice with 15-ml sterile deionised water to remove salts and other debris before being placed in a drying oven at 60 °C. Filter pads were repeatedly weighed until a constant weight was achieved. A standard curve was plotted for OD540 against dry weight (Supplementary Figure S1). For all conversions of OD540 to dry weight, a constant of 1 OD unit at OD540=254 mg dry weight (wt) l−1 was applied.

Variable cell counts of R. pomeroyi during carbon/energy starvation

R. pomeroyi was grown in MAMS with TMA (3 mM) or TMAO (3 mM) as the sole N source to a final OD540 ∼0.5. Cells were re-suspended in MAMS with no exogenous C and then aliquoted (20 ml) into 125-ml serum vials (n=3) with either no exogenous C (control), or TMA (1 mM) or TMAO (1 mM). For cell counts, serial dilutions were generated (n=3) and 10 μl were spotted (n=3) on ½ YPSS (per litre; 2 g yeast extract, 1.25 g peptone, 20 g sea salts (Sigma-Aldrich, Gillingham, UK) plates and incubated at 30 °C. TMA and TMAO were quantified by ion-exchange chromatography as described previously (Lidbury et al., 2014).

Quantification of intracellular ATP concentrations

R. pomeroyi wild type and mutant strains were grown using either TMA or TMAO as the sole nitrogen source, and cells were harvested by centrifugation (10 min; 8000 g) at late exponential phase (1 × 109 cells) and washed twice to remove exogenous C. Cells were re-suspended in MAMS medium minus glucose, given TMA (1 mM), TMAO (1 mM) or no exogenous energy source and then aliquoted (500 μl) into 2 ml microcentrifuge tubes (n=3). Cells were left for 16 h before adding a further 500 μl of each test compound. After 1 h, 100 μl of cell suspension was mixed with 100 μl of BacTiter Glo cell viability kit (Promega, Fitchburg, WI, USA) and incubated for 5 min before recording luciferase activity on a Luminoskan Ascent microplate luminometer (Thermo Scientific, Waltham, MA, USA). A standard curve was generated using ATP standards according to the manufacturer’s guidelines.

Co-culture of R. pomeroyi and Methylomonas methanica MC09

R. pomeroyi wild type and the mutant, Δtmm::Gm (Lidbury et al., 2014), were grown using either TMA or TMAO as the sole N source (OD540∼0.3). Cells were re-suspended in fresh medium containing 1 mM methanol. For each strain, triplicate cultures were set up using either TMA or ammonium chloride as the sole N source (1 mM). M. methanica was grown using methanol as the C source (2 mM) and ammonia (0.5 mM) as the limiting nutrient until the onset of stationary phase. A 5% (v/v) inoculum of M. methanica (∼107 cells) was added to each R. pomeroyi culture. Co-cultures were incubated at 25 °C on a rotary shaker (150 r.p.m.). For M. methanica cell counts, serial dilutions were generated (n=3) and 10 μl were spotted (n=3) on MAMS plates with methane as the sole C source and incubated at 25 °C.

Results

TMA and TMAO oxidation increases R. pomeroyi growth yields when grown on glucose

R. pomeroyi oxidised TMA and TMAO in the presence of both glucose and ammonia in the culture medium (Figures 2a and b). The rate of TMA and TMAO oxidation was greatest through exponential growth, but did continue throughout the stationary phase when glucose was exhausted from the medium (data not shown). TMA oxidation by wild-type cells resulted in a greater final growth yield (OD540=2.91±0.05; Figure 2a) compared to the mutant, Δtmm::Gm (OD540=2.063±0.06), which was unable to catabolise TMA (Lidbury et al., 2014). TMAO oxidation in wild-type cells (Figure 2b) also led to an increase in final growth yield (OD540=2.46±0.02) compared to the mutant, Δtdm::Gm (OD540=1.94±0.07), which cannot oxidise TMAO (Lidbury et al., 2014). TMA oxidation to TMAO could still function in the Δtdm::Gm mutant, resulting in the accumulation of extracellular TMAO in the medium (Supplementary Figure S2).

Figure 2
figure 2

(a) Catabolism of TMA during growth of R. pomeroyi wild type (grey circles) and the Δtmm::Gm mutant (white circles) on glucose and ammonium. TMA in the culture medium was quantified throughout growth for both wild type (grey diamonds) and the mutant (white diamonds). Note that the y axis is not presented as a logarithmic scale. (b) Catabolism of TMAO during growth of R. pomeroyi wild type (grey circles) and the Δtdm::Gm mutant (white circles) on glucose and ammonium. TMAO in the culture medium was quantified throughout growth for both wild type (grey diamonds) and the mutant (white diamonds). Note that the y axis is not presented as a logarithmic scale. (c) Final growth yields of R. pomeroyi wild-type and mutant strains, Δtmm::Gm and Δtdm::Gm, grown on glucose and ammonium (black bars) and supplemented with either 5 mM TMA (white bars) or 5 mM TMAO (grey bars). Error bars denote s.d. Results presented are the mean of triplicate cultures.

Full size image

We conducted an initial screen using a plate assay method, whereby R. pomeroyi was grown on glucose-limited MAMS plates with or without TMA (3 mM). Colonies grew larger in the presence of TMA, suggesting a greater proportion of the glucose was assimilated into the biomass (Supplementary Figure S3). We then carried out further experiments to quantify the enhanced growth yield due to the addition of either TMA or TMAO by quantifying dry weight of R. pomeroyi wild type and the mutants. R. pomeroyi was grown in batch culture under glucose-deplete conditions and either supplemented with or without TMA (5 mM) or TMAO (5 mM). Wild-type cells grown on glucose alone reached a final biomass of 504±14.3 mg dry wt l−1 (Figure 2c) and when supplemented with either TMA or TMAO, a final biomass of 616±8.9 mg dry wt l−1 (+22%) and 626±12.6 mg dry wt l−1 (+24%) was achieved, respectively. The Δtmm::Gm mutant, which cannot catabolise TMA, had no increase in final biomass (519±21.4 mg dry wt l−1) compared with the glucose-only cultures (534±14.3 mg dry wt l−1); however, when supplemented with TMAO, the final biomass was 664±13.3 mg dry wt l−1 (+24%; Figure 2c). Supplementing the Δtdm::Gm mutant with either TMA or TMAO did not result in any increase in final biomass (glucose=489±14.5; +TMA=453±20.6; +TMAO=487±31.7 mg dry wt l−1). When wild-type R. pomeroyi cells were grown in a glucose-limited chemostat (dilution rate=0.05 h−1), we also observed a 30.4% increase in growth yield when supplemented with TMA (5 mM), whereas the growth yield of the mutant, Δtmm::Gm, did not change (Table 1).

Table 1 Growth yields for R. pomeroyi strains grown in a glucose-limited chemostat at a growth rate of 0.05 h−1 with or without TMA (5 mM)
Full size table

Citreicella sp. SE45, which was isolated from a salt marsh (USA), is another member of the MRC and can also grow on TMA as a sole N source, but not as a sole C source (Chen, 2012). Salt marshes are typified by having high concentrations of MAs, including TMA, derived from the anaerobic degradation of compatible osmolytes such as glycine betaine (King, 1984). When Citreicella sp. SE45 was grown using glucose-deplete MAMS medium, the addition of TMA led to an increase in final growth yield (Supplementary Figure S4), thus demonstrating that catabolism of TMA can also enhance chemoorganoheterotrophic growth of another closely related bacterium.

TMA increases the growth rate of R. pomeroyi when grown on glucose

We also observed a direct correlation between specific growth rates and varying concentrations of TMA in the medium (Figure 3a). The specific growth rate increased from 0.061±0.002 (h−1) for cells incubated with no TMA to 0.087±0.003 (h−1) for cells incubated with 3 mM TMA. Likewise, the final growth yield increased from 484±10.39 (no TMA) up to 600±8.79 (3 mM TMA; Figure 3b). Using intermediate concentrations of TMA (0.5–1 mM) resulted in an intermediate increase in growth rates and growth yields compared with glucose-only cultures. Together, these data confirm that oxidation of MAs can enhance chemoorganoheterotrophic growth on glucose in R. pomeroyi.

Figure 3
figure 3

A comparison of the specific growth rates (a) and final growth yields (b) of the wild-type R. pomeroyi grown on glucose and ammonium when supplemented with increasing concentrations of TMA, using a starting inoculum that was pre-incubated with TMA (24 h). (c) The final growth yield of R. pomeroyi after 7 days during which four additions of glucose (100 μM) were added every 24–48 h. Cultures were incubated with TMA (2 mM) or thiosulphate (2 mM) or both and the same concentrations were added every 48 h. Error bars denote s.d. Results presented are the mean of triplicate cultures.

Full size image

We also observed a synergistic effect of the enhancement of heterotrophic growth when R. pomeroyi was incubated with two exogenous energy sources (TMA+thiosulphate) during incubations where low concentrations of glucose (100 μM) were stochastically added (every 24–48 h) four times (400 μM total C). Cells incubated without a supplementary energy source (TMA or thiosulphate) reached a final growth yield of 31.7±1.5 mg dry wt l−1 (Figure 3c). Cells incubated with either TMA or thiosulphate alone reached a final growth yield of 42.2±4.7 and 44.3±5.4 mg dry wt ml−1, respectively. Cells incubated with both TMA and thiosulphate reached a final growth yield of 70.8±4.9 mg dry wt ml−1, which equates to over a twofold increase in biomass.

Oxidation of TMA and TMAO enhances cell survival and viability during energy starvation

R. pomeroyi was grown on TMA as a sole N source to induce the enzymes (Figure 1) involved in MA catabolism, for example, Tmm, Tdm and GmaS, prior to re-suspension in a fresh minimal medium with no C or energy source. Cells were either supplemented with TMA or TMAO or had no exogenous energy source (control). Both TMA and TMAO were rapidly catabolised over 8 days, although the rate of TMAO catabolism slowed during the final 2 days (Figure 4a). At the start of energy starvation, the number of viable cells in all cultures was 4.0 × 109 cells ml−1 (Figure 4b). After 4 days, the number of viable cells incubated in the control cultures decreased to 7.4 × 108, whereas the cell numbers were 2.2 × 109 ml−1 in the presence of TMAO and 1.1 × 109 ml−1 in the presence of TMA, respectively. After 8 days, the number of viable cells from cultures with no exogenous C decreased to 2.9 × 107 ml−1, whereas +TMAO and +TMA cultures had 9.0 × 108 and 7.5 × 108 ml−1 cells, respectively. In summary, the number of viable cells surviving periods of energy starvation was an order of magnitude greater when cells were incubated with either TMA or TMAO.

Figure 4
figure 4

(a) Quantification of TMA (white squares) and TMAO (grey squares) during incubations with energy-starved R. pomeroyi cells. (b) Quantification of viable cells in carbon and energy-starved R. pomeroyi cultures incubated with either no exogenous carbon (black circles), TMA (white circles) or TMAO (grey circles). Error bars denote s.d. Results presented are the mean of triplicate cultures.

Full size image

To confirm that cells do indeed generate ATP from the oxidation of MAs, cells were energy-starved overnight prior to the addition of either TMA (1 mM) or TMAO (1 mM) and incubated for a further 2 h. Wild-type cells incubated with either TMA or TMAO had 93.6±4.2 and 92.1±7.8 zeptomoles ATP per cell, respectively (Figure 5), whereas the intracellular concentration of ATP was lower for cells in the no substrate control (58.3±9.7 zeptomoles ATP per cell). Incubating the mutant, Δtmm::Gm, with TMA resulted in no increase in intracellular ATP (54±5.3 zeptomoles ATP per cell) compared with the no substrate control (52.2±8.1 zeptomoles ATP per cell), whereas incubation with TMAO did result in an increase in intracellular ATP (80.7±4.9 zeptomoles ATP per cell). As expected, incubation with TMA or TMAO did not result in an increase of intracellular ATP concentrations for the Δtdm::Gm mutant (control=56.4±3.4; TMA=55.7±2.1; TMAO=56.1±2.2 zeptomoles ATP per cell).

Figure 5
figure 5

Quantification of intracellular ATP concentrations from R. pomeroyi cultures energy-starved for 18 h before incubation for a further 2 h with either 1 mM TMA (white bars), 1 mM TMAO (grey bars) or no exogenous carbon source (black bars). Error bars denote s.d. Results presented are the mean of triplicate cultures.

Full size image

Metabolism of TMA remineralises nitrogen (ammonification)

As R. pomeroyi can metabolise MAs in order to generate energy, we hypothesised that the amine group would undergo remineralisation to ammonia and subsequent cellular release from cells could provide a source of N for other marine microorganisms (Figure 6a). To test this hypothesis, we designed a co-culture experiment with R. pomeroyi and the methylotrophic bacterium, Methylomonas methanica MC09 (Boden et al., 2011a). We inoculated a C-starved and N-starved R. pomeroyi culture (∼108 ml per cells) with M. methanica (∼107 ml per cells) and supplied methanol (1 mM) as the only C source in the system, as methanol is only utilised by M. methanica. Cultures were either supplemented with ammonium chloride (1 mM) or TMA (1 mM) prior to incubation. Incubation of wild-type R. pomeroyi with methanol and TMA resulted in no growth, while TMA was depleted from the medium (data not shown). Addition of ammonium chloride resulted in growth of M. methanica when incubated with either wild type (3.9 × 108 ml−1) or the Δtmm::Gm mutant (3.3 × 108 ml−1), confirming that R. pomeroyi does not inhibit growth of M. methanica (Figure 6b). Wild-type cells of R. pomeroyi depleted TMA from the medium, resulting in growth of M. methanica (2.6 × 108 ml−1); however, no growth of M. methanica occurred (2.8 × 107 ml−1) during incubation with the Δtmm::Gm mutant of R. pomeroyi, as a consequence of no TMA degradation during the 9-day incubation period (Figure 6c).

Figure 6
figure 6

(a) Schematic diagram of the flow of nitrogen in a co-culture system involving R. pomeroyi and Methylomonas methanica MC09. Ammonia liberated from the catabolism of TMA can be used by another bacterium to support its growth. (b) The cell count of Methylomonas methanica MC09 (B) after incubation for 9 days with either R. pomeroyi wild type (A) or Δtmm::Gm mutant (A') and supplemented with either ammonium chloride (1 mM) or TMA (1 mM). (c) Quantification of TMA during incubation with wild type (white triangles) or Δtmm::Gm mutant (grey triangles). Error bars denote s.d. Results presented are the mean of triplicate cultures.

Full size image

Discussion

Methylated one-carbon compounds were originally thought to be substrates primarily for a specialised guild of bacteria, the methylotrophs (Chistoserdova et al., 2009; Chistoserdova, 2011); however, recent evidence has implicated marine heterotrophic bacteria in the catabolism of these compounds (Chen et al., 2011; Sun et al., 2011; Lidbury et al., 2014). Although a small percentage of isolates of the MRC can grow on TMA and TMAO as a sole C source, the majority appear to be able to only utilise these compounds as a sole N source, while maintaining the genes predicted to be involved in oxidation of the methyl groups (Chen, 2012). We show that R. pomeroyi and also Citreicella sp. SE45 can oxidise TMA and TMAO to help stimulate growth on an organic substrate. The implications for this are (1) catabolism of MAs results in the more efficient conversion of organic substrates into biomass, which provides an ecological advantage to these bacteria (Moran and Miller, 2007); (2) the turnover of MAs in the marine environment is likely to be rapid during times of high primary productivity due to an influx of organic substrates from phytoplankton exudation and cell death; (3) marine heterotrophic bacteria are likely to be an efficient biological sink for these compounds, retarding their flux into the atmosphere; (4) the metabolism of MAs as an energy source results in the remineralisation of MAs to ammonium, which can in turn support the growth of other microbial communities in the environment.

The ecological success of the MRC may be in part due to the utilisation of a wide range of both organic and inorganic compounds for the generation of cellular energy. Although TMAO oxidation has been shown to provide ATP for Candidatus Pelagibacter ubique HTCC1062 (SAR11 clade), no effect on the ecophysiology of the bacterium was identified (Sun et al., 2011). Our study revealed that TMA and TMAO oxidation could enhance both the growth rate and growth yield of R. pomeroyi. This is in agreement with previous work demonstrating that a methylotroph had a higher-specific growth rate and higher growth yield as a result of co-oxidation of TMAO alongside its growth on methanol (Halsey et al., 2012). Cells with higher intracellular concentrations of ATP can respond faster to fluxes of organic matter associated with phytoplankton through ATP-mediated transport (Steindler et al., 2011). Both SAR11 and Roseobacter cells devote a large amount of resources into the production of ABC-transporter systems to help facilitate the rapid uptake of essential nutrients (Sowell et al., 2008; 2011; Williams et al., 2012; Gifford et al., 2013). Therefore, bacteria of the MRC and SAR11 clades capable of generating ATP from the catabolism of TMA and TMAO may have an ecological advantage through the efficient scavenging of nutrients in the surface waters. The production of ATP through the oxidation of thiosulphate to sulphate helps Citreicella thiooxidans grow more efficiently on organic substrates (Sorokin et al., 2005). This trait is widespread within the MRC (Newton et al., 2010) and R. pomeroyi has enhanced growth when incubated with thiosulphate (Moran et al., 2004). In our study, the growth of R. pomeroyi during additions of glucose was enhanced through the co-catabolism of both TMA and thiosulphate, thus demonstrating how utilisation of multiple exogenous energy sources can enhance growth. Both TMA and thiosulphate are ‘energy rich’, in the sense that they can generate between seven and eight ATP molecules from the oxidation of one TMA or thiosulphate molecule. In contrast, carbon monoxide is a relatively ‘energy poor’ compound, only liberating two electrons, which does not appear to result in an enhancement of growth for R. pomeroyi (Cunliffe, 2012). The utilisation of MAs as a supplementary energy source is consistent with a growing body of data that points towards the success of certain heterotrophic bacterial groups that can generate energy from a wide range of sources, including reduced organic carbon compounds (Eiler, 2006; Moran and Miller, 2007; Boden et al., 2011b; Green et al., 2011; Steindler et al., 2011; Sun et al., 2011).

The greater number of viable cells in R. pomeroyi cultures incubated with TMA and TMAO is consistent with the notion that exogenous energy sources will be preferentially used instead of endogenous C stores in order to maintain cellular integrity. This also resulted in R. pomeroyi maintaining higher intracellular ATP concentrations during periods of energy starvation. Representatives of the SAR11 clade and Vibrio spp. start to break down and respire endogenous carbon when energy-starved and this process is significantly reduced when incubated in the light, through proteorhodopsin-mediated energy production (Gómez-Consarnau et al., 2010; Steindler et al., 2011). This results in a greater number of viable cells and also larger, more active cells during periods of energy starvation (Gómez-Consarnau et al., 2010; Steindler et al., 2011).

In marine surface waters, primary production is often limited by N availability, and this has a direct effect on the amount of organic matter exported to the deep ocean (Eppley and Peterson, 1979; Falkowski et al., 1998; Zehr and Kudela, 2011). The microbially mediated remineralisation of N (ammonification) following phytoplankton decomposition has previously been demonstrated in a laboratory study, which suggested that this process may occur in seawater (Garber, 1984). Here we demonstrate a ‘proof of concept’, whereby the turnover of TMA resulted in the release of remineralised N in the form of ammonia, which was subsequently taken up by another bacterium and used to support growth. As a number of MRC species are frequently associated with phytoplankton blooms (González et al., 2000; Buchan et al., 2005; Wagner-Dobler et al., 2009; Hahnke et al., 2013; Nelson et al., 2014), we predict that this N remineralisation process may take place with several different ‘nitrogen-rich’ compounds, for example, glycine betaine, choline and carnintine. This process has strong implications for the ‘microbial loop’, which ultimately controls the level of both primary and secondary production in the world’s oceans (Azam et al., 1983). N-rich compounds may represent a source of ammonia in the oceans, as the C in these compounds is catabolised to generate energy (Sun et al., 2011; Halsey et al., 2012). This process may reduce the amount of N lost to the sub-photic zone through the sinking of cell debris and particles, and may provide a feedback between the phytoplankton and heterotrophic bacteria (Azam et al., 1983; Garber, 1984). Interestingly, in bacteria from the SAR11 clade, N limitation does not induce any of the genes involved in the catabolism of MAs, while energy starvation (in the dark) does induce some (Steindler et al., 2011; Smith et al., 2013). Moreover, an ammonium transporter (SAR_1310) located adjacent to the genes involved in MA catabolism is only induced under nitrogen-replete conditions and it has been proposed that this transporter is involved in ammonia export (Smith et al., 2013). All bacteria of the MRC and SAR11 clades capable of utilising MAs have a homologue of the transporter adjacent to genes involved in MA catabolism. Homologues of the putative ammonium exporter related to both SAR11 and MRC clades are highly expressed in surface waters of the coast of Georgia (Gifford et al., 2013). At this site, genes involved in the catabolism of TMAO are also highly expressed in bacteria related to the SAR11 and MRC clades (Gifford et al., 2013). The function of this proposed ammonium transporter warrants further investigation, as it may have a pivotal role in the release of ammonium through remineralisation of organic nitrogen in marine surface waters. Together these data strengthen the hypothesis that MAs are primarily catabolised to generate cellular energy, which in turn remineralises ammonium through methylamine oxidation.

In summary, catabolism of MAs by a heterotrophic bacterium enhances chemoorganoheterotrophic growth as well as enhancing the survival of energy-starved cells. In turn, this liberates inorganic N (ammonification) that can be subsequently used by other microbes. As there are no data regarding in situ residence times and turnover rates of MAs in the surface waters of the oceans, our recent findings may help to predict the likely fate of these compounds in which rapid microbial consumption of MAs may present an oceanic sink and retard their flux from the oceans to the atmosphere.