Most oxyphototrophs use visible light (VIS, 400–700 nm) for chlorophyll (Chl) a-driven photosynthesis, while cyanobacteria in the genus Acaryochloris largely employ Chl d, thereby enabling them to use near-infrared radiation (NIR, >700 nm) for oxygenic photosynthesis ^{1, 2} . Two of the strains are well described in their growth and photopigment composition: the type strain Acaryochloris marina MBIC11017 was isolated from the didemnid ascidian Lissoclinum patella from coral reef habitats in Palau ^{2, 3} and was later genome sequenced ⁴ . Since its first discovery, other Acaryochloris strains have been obtained from Japanese macroalgae (Strain Awaji ⁵ ), from surfaces in a hypertrophic lake in the US (strain CCMEE5410) ⁶ , from swipes of coral substrate collected on Heron Island, Australia (strain HICR111A) ⁷ , and most recently from an Australian mangrove (strain MPGRS1) ⁸ and stromatolites in Shark Bay, Western Australia (strain ssball1) ⁹ .

The light microenvironment in natural habitats occupied by Acaryochloris spp. has a high contribution of NIR relative to visible light ^{10– 12} and such habitats appear to create similar niche differentiation with bacteria carrying specialized photopigments such as Chl d/ f or bacteriochlorophylls ^{10, 13} . The notion of a global distribution of Chl d and cyanobacteria in the genus Acaryochloris ^{11, 14} further reinforces the need to obtain information on the photobiology of Chl d-containing oxyphototrophs. Understanding the adaptive mechanisms in oxyphototrophs capable of using wavelengths beyond VIS is of interest as it provides information concerning the usability, stress levels and limitations associated with NIR-driven oxygenic photosynthesis.

Of all biological pathways, photosynthetic electron transport is particularly prone to produce reactive oxygen species (ROS) due to the very high (positive) redox potential of the primary donor of photosystem II needed to oxidize water, and the low redox potential of the primary electron acceptor of photosystem I needed to reduce ferredoxin; here, singlet oxygen ( ¹O ₂) is produced by PSII and superoxide anions ( ^-O ₂)/hydrogen peroxide (H ₂O ₂) in the Mehler-ascorbate peroxidase (MAP) pathway of PSI ^{15– 17} . ROS encompasses the production of singlet oxygen, superoxide anions, hydrogen peroxide and hydroxyl radicals (OH), all of which are derived through the local energization of O ₂. If not properly quenched by protective mechanisms, ROS can damage proteins, DNA and other cellular macromolecules, and this damage can ultimately lead to cell death. Known quenching mechanisms encompass enzymes such as superoxide dismutase and catalase or non-enzymatic antioxidants like glutathione, carotenoids and α-tocopherol (vitamin E) ¹⁸ . In plants, ROS and in particular ¹O ₂ production has been shown to occur at photosystem II upon illumination with visible light ^{19, 20} . In cyanobacteria, shorter wavelengths such as ultraviolet radiation (UVR, <400 nm) are known to induce ROS, causing DNA damage, lipid-peroxidation and overall decreased photosynthetic efficiency ^{21, 22} . To our knowledge, no study has investigated the effect of NIR on ROS production in cyanobacteria.

Relative levels of ROS can be estimated through measurements of for example gene expression ²³ , ROS-sensitive fluorescence probes ²⁴ and enzyme activity ²⁵ . These methods provide integrated values of ROS production over incubation time intervals ranging from minutes to hours. In this study we used a fast and sensitive laser photo-acoustic gas detector ²⁶ that can measure the ethylene produced from the reaction of certain types of ROS with the substrate 2-keto-4-thiomethylbutyric acid (KMBA), previously added to the samples. Such near real-time ROS detection is valuable in determining the immediate effect of treatments on the physiological state and stress level within living organisms. KMBA is thought to diffuse into intact cells ²⁷ and, when supplied at saturating concentrations, outcompetes other radical scavenging mechanisms. In the KMBA assay, the butyric acid moiety reacts with ROS like peroxynitrite, hydroxyl radicals and peroxyl radicals ²⁸ , resulting in the formation of ethylene, which can then be quantified. In other studies, KMBA has been used to test the antioxidant capacity of radical scavengers via their ability to inhibit ethylene formation relative to a control reaction (total oxyradical scavenging capacity, TOSC) ^{28, 29} .

In this study, we report the effect of light intensity and spectral composition on ROS generation, as measured in real-time using a laser-photoacoustic gas detector in three different strains of NIR utilizing cyanobacteria belonging to the genus Acaryochloris, including a new strain (named Acaryochloris CRS), isolated from phototrophic biofilms growing on dead coral branches collected on Heron Island, Australia.

We aimed to determine the stress levels associated with Chl d-driven oxygenic photosynthesis and tested NIR, VIS and narrower wavebands for their capacity to induce ROS in three strains of Acaryochloris: i) The Acaryochloris marina type strain MBIC11017 ^{2, 3} , ii) strain HICR111A ⁷ , and iii) a novel strain, named CRS, that we isolated from dead coral branches which, based on 16S rRNA gene sequencing, grouped within the genus Acaryochloris ( Figure 1). Acaryochloris strains MBIC11017 and HICR111A are both well described in terms of their photopigmentation, genomic content, ultrastructure and their capability to perform photoacclimation ^{4, 7, 30} . To test whether photoacclimation, i.e., light-dependent change in pigment levels, was associated with increased resistance or sensitivity towards ROS, we acclimated the strains to NIR or VIS prior to experiments. By taking advantage of the real-time ROS detection method, we could for the first time demonstrate the immediate effects of NIR, VIS and other wavelengths on the ROS pools within living cyanobacteria Specifically, we found that:

(i) Depending on strain and previous adaptation, ROS levels were lower in cells exposed to NIR than in those exposed to VIS ( Figure 2A);

(ii) Exposure to shorter wavelengths such as blue and cyan, generated the most ROS in strain MBIC11017 and HICR111A, while less ROS was produced upon exposure to longer wavelengths (green, amber and red) ( Figure 2C).

(iii) Strain CRS generated less ROS upon exposure to VIS when previously acclimated to NIR, while strains MBIC11017 and HICR111A appear more sensitive to VIS when adapted to NIR ( Figure 2A).

In Acaryochloris, VIS irradiance is primarily absorbed by the photopigments Chl d (with maximum absorption occurring at 440–470/710 nm), Chl a (440–470/675 nm), carotenoids (440–520 nm) and if present, phycobiliproteins (560–650 nm). NIR provides a more targeted stimulation of photosynthesis and is almost exclusively absorbed by Chl d. At comparable photon irradiances (VIS = 340–480 µmol m ^-2 s ^-1 versus NIR = 400 µmol m ^-2 s ^-1), we found that, depending on strain and previous adaptation, ROS levels were lower in cells exposed to NIR than in those exposed to VIS ( Figure 2A). Based on pulse-amplitude modulated (PAM) fluorometry measurements, the light intensities used in our experiments are known to saturate relative-electron transport rates in the type strain MBIC11017 ¹² . Photoinhibition was not observed even at higher photon irradiance, but we hypothesize that prolonged exposure to relatively high irradiances (10–20 fold more irradiance than during culturing for 15–20 min) could result in the over-reduction of the primary acceptors on the PSI and PSII side, resulting in the production of ROS ^{15, 23} . Alternatively, ROS could be the result of photosensitized light-harvesting pigments ³¹ ; however, in intact light-harvesting complexes the efficiency of electron transfer towards the reaction centers is usually outcompeting the formation of long-lived (ROS-forming) chlorophyll triplet states ³² . This appears particularly true for the unique phycobiliprotein antenna rods in A. marina MBIC11017, in which excitation electron transfers to PSII were found to be significantly faster than in Chl a-containing cyanobacteria ³³ . Additionally, it is known that within light-harvesting complexes carotenoids are outcompeting O ₂ in the de-excitation of triplet chlorophyll states ³⁴ . All three strains used in the current study were found to contain the carotenoid zeaxanthin ( Table 1), which can play a crucial role in the quenching of singlet oxygen and general non-photochemical quenching of excited Chl states ^{32, 35} .

Exposure to shorter wavelengths, such as blue (470 nm) and cyan (495 nm) light generated the most ROS in strains MBIC11017 and HICR111A, while less ROS was produced upon exposure to longer wavelengths (green, amber and red) ( Figure 2C). Blue and cyan light-induced ROS production in MBIC11017 and HICR111A is probably due to spectral overlap with the Soret-band absorption of Chl a/ d (440–470 nm) ( Figure 2B) and the above-mentioned mechanisms in ROS generation. Red (645 nm), amber (595 nm) and green (535 nm) light overlaps with the absorption spectra of phycobiliproteins which, if present, aid in light harvesting and excitation energy transfer towards the photosystems ³³ . Strain MBIC11017 is known to express the phycobiliproteins phycocyanin and allophycocyanin ³⁶ , while strain HICR111A reportedly lacks phycobiliproteins ⁷ .

Comparable ROS levels were observed in strains HICR111A and MBIC11017 under yellow and green light, suggesting the presence of pigments absorbing these wavelengths or the possibility of other light-induced ROS production mechanisms. Spectrophotometric analysis of the strains showed weak absorption in the phycobiliprotein-specific region within all three strains ( Figure 3). This would corroborate the excitation energy transfer to PSI and II in strain HICR111A and could explain the observed ROS production under yellow and green light. However, this would also refute previous reports on the absence of phycobiliproteins in this strain ⁷ . Given that phycobiliproteins were not purposely extracted and further analyzed in the present study, we can at this point only speculate about their presence and relative expression under different growth conditions.

There is a long history of associating pigment compositions within phototrophs with the spectral composition of ambient light and exposure history: for recent work see Stomp et al. 2007 ³⁷ and Apel and Hirt 2003 ²³ , respectively. These two factors are likely to determine the sensitivity of phototrophs to irradiance and their capability to cope with ROS levels generated upon irradiation. Interestingly, we found that strain CRS generated very little ROS upon exposure to VIS when previously acclimated to NIR, while strains MBIC11017 and HICR111A appear to be more sensitive to VIS when previously adapted to NIR ( Figure 2A). HPLC analysis revealed a much higher ratio of zeaxanthin/α-carotene in NIR-acclimated CRS cells than in the other two strains ( Table 1). In A. marina MBIC11017, α-carotene was found to be an integral part of both photosystem reaction center cores ^{38, 39} and zeaxanthin is predominantly found in the periphery of light-harvesting complexes ³⁵ .

Based on the higher zeaxanthin/α-carotene ratios in NIR-adapted cells of all three strains, we hypothesize that there are slightly more antenna complexes (zeaxanthin) per reaction center cores (α-carotene) in NIR-adapted cells than in those pre-adapted to VIS. Given that the antenna complexes are predominantly composed of Chl d and thus absorb in the NIR part of the light spectrum, this chromatic photoacclimation is expected and further corroborated by higher Chl d/zeaxanthin ratios in NIR-adapted strains in this study. Besides their light harvesting capability, the carotenoids, zeaxanthin and α-carotene, are also known for their antioxidative capabilities ^{15, 34} . Particularly zeaxanthin was shown to play a crucial role in non-photochemical quenching and energy dissipation from sensitized chlorophyll molecules or singlet oxygen ¹⁵ . We hypothesize that the higher contribution of zeaxanthin in the NIR-adapted strain CRS could aid in effectively capturing NIR but also in quenching almost all ROS produced during illumination with either VIS or NIR.

In addition, it is possible that strain-specific differences in ROS production are associated with dissimilarities in growth forms: Strain HICR111A forms cell aggregates ⁷ , and so does strain CRS ( Figure 2B), whereas strain MBIC11017 usually grows as dispersed cells ⁷ but can be immobilized into biofilms ⁴⁰ . The formation of aggregates in strain HICR111A and CRS might provide photoprotection through self-shading, a behavior reportedly less pronounced in strain MBIC11017 ⁷ . Both HICR111A and CRS originate from concretionary substrata in shallow reef flats, a high irradiance habitat. In contrast, strain MBIC11017 was isolated from a didemnid ascidian ² , a light environment depleted of VIS but with sufficient NIR ^{10– 12} .

Based on these first measurements, we suggest that through utilization of NIR, Acaryochloris can avoid a considerable amount of light stress, while harvesting a portion of the electromagnetic radiation spectrum not used by other oxyphototrophs. Additionally, aggregation of certain strains could protect against excess amounts of ROS generated during high irradiance exposure. Overall, this would add a further argument as to why Acaryochloris is a successful and apparently globally widespread oxyphototroph ¹¹ .

To the best of our knowledge, we are the first to provide data on ROS levels associated with NIR-driven oxygenic photosynthesis. Unfortunately, due to measuring time constraints at the trace gas facility, we were unable to provide replicated measurements of ROS levels under various light regimes. This emphasizes the need to perform additional measurements of NIR induced ROS production in A. marina and other far-red utilizing oxyphototrophs.