Contribution of Aerobic Photoheterotrophic Bacteria to the Carbon Cycle in the Ocean

See allHide authors and affiliations

Science  29 Jun 2001:
Vol. 292, Issue 5526, pp. 2492-2495
DOI: 10.1126/science.1059707


The vertical distribution of bacteriochlorophylla, the numbers of infrared fluorescent cells, and the variable fluorescence signal at 880 nanometers wavelength, all indicate that photosynthetically competent anoxygenic phototrophic bacteria are abundant in the upper open ocean and comprise at least 11% of the total microbial community. These organisms are facultative photoheterotrophs, metabolizing organic carbon when available, but are capable of photosynthetic light utilization when organic carbon is scarce. They are globally distributed in the euphotic zone and represent a hitherto unrecognized component of the marine microbial community that appears to be critical to the cycling of both organic and inorganic carbon in the ocean.

Although closely related to purple photosynthetic bacteria, aerobic anoxygenic photoheterotrophs (AAPs) are obligate aerobes, with unusually high concentrations of carotenoids (1–3), low cellular contents of bacteriochlorophyll a (BChla) (4), and while containing photosynthetic reaction centers (RC) and light harvesting complex I (LHI), they often lack LHII (3). Photosynthetic energy conversion has been confirmed in several species (5–8), but most known AAPs have been isolated from organic-rich environments (9–11), where they appear to be heterotrophic. Recently, AAPs were found throughout the surface waters of the oligotrophic ocean (12), however, their abundance, distribution, and potential ecological importance were unknown. Here, we report quantitative measurements of the vertical distribution of AAPs and BChla in the open ocean, determine the photosynthetic competence of these organisms, and evaluate their contribution to the marine carbon cycle.

To characterize the vertical distributions of AAPs and their photosynthetic properties we used an Infrared Fast Repetition Rate (IRFRR) fluorescence transient technique (12, 13). Samples obtained from discrete depths were analyzed within 60 min of sampling (14). We assessed the distribution of AAPs using the BChla fluorescence signal at 880 nm (Fig. 1A), and measured bulk BChlaby high-performance liquid chromatography (HPLC) (15) (Figs. 1A and 2A). The BChlaconcentrations reached a maximum of 3 to 5 ng/liter at about 30 m and decreased to levels <0.01 ng/liter below 150 m (Fig. 1A). Chlorophyll a (Chla), which in the open ocean is only found in oxygenic phytoplankton, was about 150-fold more abundant. The vertical distribution of Chla, however, was closely correlated with BChla (Fig. 1B). The HPLC-based estimates of BChla corresponded to the in vivo fluorescence at 880 nm, allowing us to use the IRFRR signal (which reflects the radiative losses from LH antennae), to derive the concentration of BChla in situ and vice versa.

Figure 1

Vertical distribution of fluorescence, Chla, BChla, and microbial cells on July 2000 at station C354-004. (A) Amplitude of the 880-nm fluorescence signal characteristic of AAPs phototrophic bacteria (○), the corresponding HPLC-based BChla distribution (□), and temperature (broken line). (B) Amplitude of the 685-nm signal characteristic of oxygenic phytoplankton (○) and HPLC-based Chla (□). (C) Infrared cell counts (○) and DAPI stained cell counts (□).

To quantify the representation of AAPs, we counted the BChla-containing and total, 4′6-diamidino-2-phenylindole (DAPI) stained cell numbers by epifluorescence microscopy (16) (Fig. 1C). Model II regression analysis of the relationship between the fluorescence signals, pigment concentrations, and cell counts reveals a significant correlation between the 880-nm fluorescence emission and BChla (r2 = 0.68,F = 62.6, Fig. 3A), and the IR fluorescent cell counts (r2 = 0.49,F = 21.6, Fig. 3B). From these data, we calculated the average cellular BChla content at 1.2 × 10−19 mol per cell. The morphology of a representative isolate from the surface waters (cylindrical motile cells, approximately 1.2 μm long, 0.7 μm in diameter) allows us to estimate the cell volume at about 0.5 μm3, a cell wet weight of 0.5 pg, and a cell dry weight of 0.05 pg, yielding a BChla/dry weight ratio of about 2.4 μmol/g. This ratio is similar to that of Erythrobacter longus(17), but much higher than that ofCitromicrobium bathyomarinum (18). Assuming 36 BChla molecules/RC+LHI (19), we estimated about 2000 RCs per cell. The cellular BChla content calculated here is about an order of magnitude lower than that ofRhodobacter sphaeroides (20), and the BChla/RC ratio is also about fivefold less than that of typical purple nonsulfur bacteria (19). However, the effective photosynthetic absorption cross section measured at 470 nm (12) (about 62 Å2) was comparable to that measured in a laboratory culture of R. sphaeroides(about 70 Å2). By comparison, the effective absorption cross section in photosystem II reaction centers in phytoplankton averaged 420 Å2, consistent with 200 to 300 Chla/RCII. Although the rate of photon absorption/RC was sevenfold less than in their oxygenic planktonic counterparts, AAPs display a similar light utilization efficiency per unit of chromophore.

Fluorescence excitation spectra, recorded at the emission maximum of 875 nm (21) (Fig. 2B), indicate that the oceanic AAPs utilize carotenoids as an efficient, auxiliary photosynthetic pigment. These carotenoids harvest light between 460 and 550 nm (Fig. 3B), which penetrates relatively deeply into the water column, whereas infrared light in the 800- to 900-nm range is about 1000-fold attenuated owing to absorption by water (22). The efficiency of excitation transfer from carotenoids to RC (defined as the ratio of quanta transferred to the RC to the quanta absorbed) measured on our isolates of AAPs (Fig. 2B) was relatively low, increasing from approximately 5 to 20% as the organic content in the growth medium decreased. From the IRFRR data acquired in natural AAP populations, we estimated the maximal cellular photosynthetic fluxes at about 1 × 1010 photons, or 2.2 × 10−9 J per cell per day. Assuming a cellular carbon content of about 2.5 fmol (60% of cell dry weight), the maximal cellular photosynthetic energy fluxes could reach 1 × 106 J/mol of carbon under high irradiance levels, and average 1 ×105 J/mol of carbon in the euphotic zone. Assuming a specific growth rate of one cell per day, and a respiratory efficiency of 5 × 105 J/mol carbon, about 20% of the cellular energy requirement would be satisfied by photosynthetic electron transport.

Figure 2

(A) HPLC chromatogram of fluorescent pigments from a surface sample (2 m depth) collected at station C354-004. Excitation was at 365 nm, emission at 780 nm, with 20-nm slits. These wavelengths were chosen to maximize the signal from BChla, while minimizing the signal from the more abundant pigments, Chla and Chlb. (Inset) Fluorescence emission spectrum of the peak eluting at 16.7 min in (A). Excitation was at 365 nm and slits were 20 nm. Virtually identical retention times and fluorescence emission spectra were recorded for BChladerived from R. sphaeroides, and isolate NAP1 obtained from the coastal North Atlantic Ocean. Lettuce extract served as standards for Chla and Chlb. (B) Fluorescence excitation spectra recorded at 875 nm (thick gray line) and absorption spectra (black line), both measured on whole cells of NPP1 isolate from Northeastern Pacific Ocean. Quantum efficiency of excitation transfer from carotenoids to the reaction center (broken line) was about 20%. (C) P versus I curves of carbon fixation (•) (23) and IRFRR based estimates of the photosynthetic electron transport rates (□) (12) in NAP1 isolate obtained from the coastal North Atlantic Ocean. The continuous line represents a numerical fit of carbon fixation data to a hyperbolical model, y =a 0 +a 1 x/(a 2 +x), where y is the CO2 fixation, andx is the irradiance.

Figure 3

Relationships between fluorescence, pigment, and cell counts at four stations in the vicinity of the Juan de Fuca Ridge. (A) Model II regression analysis of HPLC-based BChla and fluorescence signal at 880 nm. (B) Model II regression analysis infrared epifluorescence cell count and fluorescence signal at 880 nm. (C) Model II regression analysis of infrared and DAPI-stained cell counts.

Oceanic AAPs are capable of light-dependent CO2 fixation (23) (Fig. 2C). Under saturating light, the rates of light-dependent CO2 fixation in laboratory-grown isolates are relatively low, averaging at about 0.43 mol carbon/mol RC/s, with a maximum quantum yield of about 1% (mol carbon/mol quanta absorbed). These yields are about an order of magnitude lower than in phytoplankton, and comparable to that reported previously inErythrobacter sp. (24). We estimate the daily cellular rates of CO2 fixation at about 0.08 fmol of carbon, or 3% of the cellular carbon content. As the cells were grown in 2 to 20 mM organic medium, maintaining specific growth rates of four per day, CO2 fixation contributed to about 1% of the total carbon anabolism. In the open ocean, where dissolved organic matter (DOM) is three orders of magnitude less abundant, the relative contribution of the CO2 fixation may be significantly higher. Nevertheless, the potential contribution of AAPs to the oceanic carbon cycle is determined by their ability to supplement, or substitute respiration with the light-driven generation of ATP and reductants for carbon anabolism, preserving the existing organic carbon. Isolates of oceanic AAPs grown in 20 mM organic medium displayed about 20 to 40% higher specific growth rate under day-night cycle relative to dark-incubated cultures, resulting in a fivefold higher biomass accumulation within 6 days. Such a mechanism of light-enhanced preservation of the organic carbon will affect the extent of the new production in the upper ocean, given that AAPs are abundant within the microbial community.

Numerically, AAPs constitute approximately 11.3 ± 1.7% of the total microbial community in the euphotic zone in the Northeastern Pacific Ocean at 48°N, 128°W (Figs. 1C and 3C). The Bchla/Chla ratio at this location, about 0.8%, increased to as much as 10% in the oligotrophic waters of the Eastern Pacific Ocean at 14°N, 104°W (12). Such a dramatic change in the BChla/Chla ratio almost certainly reflects an increase in the relative abundance of AAPs within the microbial community. Thus, the calculated 11.3% of the total cell count at 14°N, 104°W may be greatly exceeded in the oligotrophic ocean, and the globally averaged BChla/Chla ratio may be as high as 5 to 10%.

We isolated 10 AAP strains from the Coastal North Atlantic Ocean, Northeastern Pacific Ocean, Equatorial Pacific Ocean, Southern Ocean, and Mediterranean Sea (25). After 8 to 10 days incubation on organic-poor agar plates in the dark, they all formed small pink colonies, which on replating with the same medium frequently segregated into purple and yellow isolates.

When grown in an organic-poor, autotrophic liquid medium, all the isolates displayed IRFRR fluorescence transients characteristic of photosynthetic electron transport, similar to that measured in the natural water samples. When transferred to an organic-rich medium, the purple isolates displayed a 40% decline in the amplitude of the IRFRR fluorescence transient within 24 hours, whereas the yellow isolates lost this signal completely, showing no accumulation of BChla and RC despite maintaining high specific growth rates of about six per day. The growth rate decreased during the next 7 days as nutrients became depleted, the fluorescence transients were restored in both isolates, and BChla accumulation resumed. Such a response to nutrient conditions indicates that oceanic AAPs are capable of controlling the expression of their photosynthetic apparatus; i.e., they are facultative phototrophs, switching to a mostly heterotrophic metabolism under organic-rich conditions, where photosynthesis presumably offers less of an advantage. We observed this behavior within a dense particle layer in the upper portion of the euphotic zone, where we detected a strong fluorescence signal at 880 nm, but with unusually low amplitude of the variable fluorescence.

Phylogenetic analysis of the 16S ribosomal RNA gene (16S rDNA) (26) from a North Atlantic isolate (NAP1) placed it in the Erythrobacter-Citromicrobium-Porphyrobac- tercluster within the α-4 subclass of the Proteobacteria (3), which forms a relatively isolated group with respect to other AAPs (27). The 16S rDNA sequence of NAP1 showed high sequence similarity with those of Erythrobacter longus and Erythrobacter litoralis (98.2 and 97.9%, respectively), suggesting that this isolate belongs to theErythrobacter genus. A preliminary characterization of a purple isolate from the Northeastern Pacific Ocean (NPP1), based on both the restriction fragment length polymorphism (RFLP) and sequence of the 16S rDNA, revealed that this isolate is closely related to NAP1.

Our data indicate that, like cyanobacteria, AAPs are ubiquitous in the euphotic zone of the open ocean. We have observed IRFRR fluorescence transients characteristic of bacterial photosynthesis in every ocean surface water sample analyzed to date. All of these water samples yielded isolates displaying similar fluorescence characteristics. The isolates from the Northeastern Pacific and coastal North Atlantic Oceans, (the most extensively characterized) are morphologically, biophysically and phylogenetically similar. Oceanic AAPs occupy a uniform environment characterized by relatively low concentrations of DOM, and are exposed to high irradiance levels. They are photosynthetically competent in situ and utilize carotenoids as a major LH pigment. All are readily cultivated on organic-poor media. These common features suggest that oceanic AAPs are probably represented by a relatively uniform, widespread class.

We speculate that the phylogenetically related, yet phenotypically diverse aerobic anoxygenic photosynthetic bacteria, discovered in a variety of ecological niches over the last 20 years, may have speciated from a common ancestral oceanic AAP. Possible evolutionary adaptations may have ranged from a permanent loss of photosynthetic activity to the development of a regulatory mechanism that controls the level of expression of the photosynthetic apparatus in response to nutrient concentrations.

The close correlation between AAPs and oxygenic phototrophs in the euphotic zone indicates that they coexist in a tightly linked nutrient cycle. Because AAPs are unable to utilize water as an electron donor, they most likely rely on exudants produced by oxygenic photoautotrophs to supply reductants. AAPs will use the available DOM if present at sufficient concentration, but are also capable of photosynthetic CO2 fixation under DOM-deficient conditions. Their photosynthetic efficiency and spectral light utilization is similar to that of the oxygenic phototrophs, further explaining their co-occurrence with phytoplankton in the water column. These facultative photoheterotrophs coexist with oxygenic photoautotrophs, contributing to a light-controlled component of microbial carbon and redox cycle, the details of which are presently unknown.

  • * To whom correspondence should be addressed. E-mail: zkolber{at}


View Abstract

Navigate This Article