Figure 2. Light dependence of oxidation of dis- solved Fe(II) by anoxygenic photoautotrophic bacteria. A: Relationship between Fe(II) oxida- tion rate and light intensity is linear for Rho- dobacter ferrooxidans strain SW2 (squares) and Thiodictyon sp. strain F4 (triangles). Data represent average of three independent exper- iments with error bars showing standard devi- ation. B: Oxidation of Fe(II) by strain SW2 in presence of full 34W incandescent light bulb spectrum (squares) and in presence of same light bulb in combination with a filter that cuts off wavelengths >650 nm (triangles). In both cases, light intensity was adjusted to 600 lux by positioning culture bottles at appropriate distances from light source. Graphs are repre- sentatives of at least two independent experiments.
types of these bacteria at different light intensities (Rhodobacter fer- rooxidans strain SW2 and Thiodictyon sp. strain F4).
All bacteria strains were cultivated in a freshwater mineral me- dium with either dissolved Fe(II), hydrogen, or acetate at pH 6.8–6.9 at 24 C (F4, KoFox) or 16–18 C (SW2), as described previously (Kappler and Newman, 2004). Cultures were incubated in front of 34W tungsten incandescent light bulbs (Watt-Miser, General Electric, 380 lumens). Experiments with different light intensities were performed by incubating the cultures at different distances to the light bulb. The light intensity was measured with a Traceable dual-range light meter (Control Company, Friendswood, Texas). Wavelength-dependent Fe(II) oxidation rates were determined by incubating the cultures behind a Borofloat light filter (Edmund Optics, Barrington, New Jersey), which
dation rates observed at
were analyzed by linear different light intensities
regression of the oxidation during exponential growth.
To analyze for Fe(II) in cultures of Fe(II)-oxidizing bacteria, 200
L culture suspension was withdrawn with glove box, filtered with 0.5 mL nylon (0.22 Corning, New York), and analyzed for Fe(II) (Stookey, 1970).
a syringe in an anoxic m) filter tubes (Costar, with the ferrozine assay
We calculated the light attenuation by a layer of phototrophic
Fe(II)-oxidizing bacteria present in a water column (Fig. 1) as follows. Absorbance spectra of cultures with known cell density were recorded, background absorption by medium components was subtracted, and attenuation coefficients (K) for different wavelengths were calculated using the dependencies (1) of absorbance (A) on light intensity [A log (I0/I)] (I is the light intensity after it passes through the sample; I0 is the initial light intensity) and (2) of attenuation coefficients on light
intensity and thickness of sample (z) [K
(LN(I/I0)/z)] for a layer
z containing 106 cells/mL. The attenuation coefficient for the bacterial layer was convoluted with the attenuation coefficients given for pure water (Smith and Baker, 1981) to create the dashed lines in Figure 1.
RESULTS AND DISCUSSION
The phototrophic oxidation of dissolved Fe(II) by both the purple sulfur bacterium Thiodictyon sp. strain F4 and the purple nonsulfur bacterium Rhodobacter ferrooxidans strain SW2 showed a linear de- pendence on the light intensity, up to 400 and 600 lux, respectively,
corresponding to 8 and 12
mol quanta m
(Fig. 2A). At higher
light intensities the oxidation rates did not increase linearly owing to saturation effects, as has been observed for other anoxygenic photo- synthetic bacteria, where saturation at 1500 lux (Chromatium vinos- um) (van Gemerden, 1980) or 1000 lux (Rhodopseudomonas palustris)
(Uemura et al., 1961) has been documented.
In addition to light intensity, the metabolic rate of phototrophic organisms is dependent on light quality, owing to the selective absorp- tion of particular wavelengths of light by (bacterio)chlorophyll (Bchl) and carotenoids. For example, Bchla, Bchlb, and Bchlc, present in pur- ple sulfur, purple nonsulfur, and green sulfur Fe(II)-oxidizing bacteria, have absorption maxima between 800 and 880 nm (Bchla), at 1020 nm (Bchlb), and at 750 nm (Bchlc). In contrast, the carotenoids of the Fe(II)-oxidizing organisms (SW2: spheroidene, spheroidenone, OH- spheroidene; F4: rhodopinal; KoFox: chlorobactene) maximally absorb at much lower wavelengths (360–517 nm; see Data Repository1). Be- cause light of wavelengths lower than 300 and higher than 600 nm will readily be absorbed or scattered by water within the top few meters of the ocean (Fig. 1), this can be expected to limit the rate of Fe(II) oxidation by anoxygenic phototrophs. To determine to what extent pho- totrophic Fe(II) oxidation could be sustained in deeper water, Fe(II) oxidation rates were measured with strain SW2 using only wavelengths
650 nm (Fig. 2B). The oxidation rate decreased (to 0.08 mM/day)
but was still
20% of the rate obtained in the presence of the full light
Fe(II) oxidation can creased efficiency.
In the water column of modern aqueous environments, such as the Black Sea and stratified lakes, anoxygenic phototrophs are typically found at depths below cyanobacteria because oxygen inhibits anoxy- genic phototrophic metabolism (Repeta et al., 1989; Falkowski and Raven, 1997). Note that in these ecosystems cyanobacteria colonize the surface mixed layer down to the depth of a strong pycnocline that prevents (or at least reduces to very low values) the mixing between
¨¨ the surface oxic water and the deeper anoxic water (Ozsoy and Unlu¨ata,
1997). Light is still available below the mixed layer, with an intensity that allows anoxygenic photosynthesis. If organisms like strain SW2 were present in the Precambrian ocean, a similar strong density gra- dient would have been required to separate the surface mixed layer from the deeper anoxic water in which they were living. We therefore assume that the pycnocline was shallower than the photic depth. Given
1GSA Data Repository item 2005170, isolation and identification of pig- ments, is available online at www.geosociety.org/pubs/ft2005.htm, or on request from firstname.lastname@example.org or Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301, USA.
GEOLOGY, November 2005