A TEST OF THE OXYGEN PARADOX USING ANTIOXIDANT-DEFICIENT CYANOBACTERIA

Oxygenic photosynthesis evolved in an anaerobic environment that contained several orders of magnitude more CO2 than present-day air. As photosynthesis oxygenated the environment, the formation of reactive oxygen species (ROS) would have become more common. When oxygen is present, photosystem I can accidentally transfer a single electron to molecular oxygen, resulting in the formation of superoxide anions. Subsequently, superoxides can be converted to hydrogen peroxide and/or hydroxyl radicals. Additionally, excited singlet-state oxygen can form within both photosystems I and II (Asada and Takahashi, 1987). All ROS indiscriminately react with biological molecules, resulting in stress that can lead to cell death. Similar reactions occur with electron transport proteins (e.g., cytochromes) of non-photosynthetic organisms. The antioxidant system detoxifies ROS by adding additional electrons to radicals and quenching excited-state molecules (Fridovich, 1989). Therefore, the co-evolution of the antioxidant system and oxygenic photosynthesis presents a paradox. Without antioxidants, oxygenic photosynthesis is self-destructive. However, without photosynthesis, there may not have been any selective pressure for the evolution of antioxidants. We considered two possible hypotheses regarding the sequence of photosynthesis-antioxidant co-evolution. If photosynthesis evolved first, the large diffusion gradient produced by the anoxic environment would have allowed oxygen to diffuse out of the cells before being converted to ROS. An antioxidant system would not be needed until the environment became oxygenated. Alternatively, antioxidant systems may have evolved first in response to UV-induced and other non-biogenic oxidative stresses. The presence of an antioxidant system would have protected earlier anoxygenic photosynthesizers, and provided a pre-adaptation for the later evolution of oxygenic photosynthesis. To test these hypotheses, we grew wild-type and antioxidant-deficient strains of Synechococcus sp. PCC7942 in air and in simulated primordial atmospheres (2.5% CO2 in N2). The katG strain lacks catalase; the tplA strain lacks a thioredoxin-peroxidase-like enzyme (Perelman et al., 2003); and the sodB strain (Herbert et al. 1992) lacks iron superoxide dismutase (FeSOD). If the presence of intact antioxidant enzymes was required under primordial conditions, we would expect the growth rates of the wild-type strains to be higher than those of the antioxidant mutant strains. Conversely, if antioxidants were not needed under primordial conditions, we would expect the growth rates of the wild-type and mutant strains to be identical. Cyanobacterial cultures were grown in liquid BG-11 medium at 27°C, and illuminated with cool-white fluorescent tubes at 50 μmol photons m s. Chloramphenicol or spectinomycin was added as a selective agent to stock cultures of antioxidant deletion strains. Antibiotics were not added to experimental cultures. Control cultures were bubbled with filtered air; experimental cultures were bubbled with 2.5% CO2 in N2 (Thomas and Herbert, 2005). As a control for the effects of CO2, additional cultures were bubbled with 2.5% CO2 in air. Growth was measured as optical density at 720 nm. Damage to photosystem II was quantified by measuring chlorophyll fluorescence (FV/FM) (Campbell et al., 1998). Damage to photosystem I was measured by changes in P700 redox kinetics (Sacksteder and Kramer, 2000). Both photosystem measurements were made with a Walz PAM-101 chlorophyll fluorometer. Data were analyzed with Microsoft Excel 2003, α = 0.05.