Photoinactivation of Photosystem II in Prochlorococcus and Synechococcus

The marine picocyanobacteria Synechococcus and Prochlorococcus numerically dominate open ocean phytoplankton. Although evolutionarily related they are ecologically distinct, with different strategies to harvest, manage and exploit light. We grew representative strains of Synechococcus and Prochlorococcus and tracked their susceptibility to photoinactivation of Photosystem II under a range of light levels. As expected blue light provoked more rapid photoinactivation than did an equivalent level of red light. The previous growth light level altered the susceptibility of Synechococcus, but not Prochlorococcus, to this photoinactivation. We resolved a simple linear pattern when we expressed the yield of photoinactivation on the basis of photons delivered to Photosystem II photochemistry, plotted versus excitation pressure upon Photosystem II, the balance between excitation and downstream metabolism. A high excitation pressure increases the generation of reactive oxygen species, and thus increases the yield of photoinactivation of Photosystem II. Blue photons, however, retained a higher baseline photoinactivation across a wide range of excitation pressures. Our experiments thus uncovered the relative influences of the direct photoinactivation of Photosystem II by blue photons which dominates under low to moderate blue light, and photoinactivation as a side effect of reactive oxygen species which dominates under higher excitation pressure. Synechococcus enjoyed a positive metabolic return upon the repair or the synthesis of a Photosystem II, across the range of light levels we tested. In contrast Prochlorococcus only enjoyed a positive return upon synthesis of a Photosystem II up to 400 μmol photons m-2 s-1. These differential cost-benefits probably underlie the distinct photoacclimation strategies of the species.


Introduction
Synechococcus and Prochlorococcus picocyanobacteria numerically dominate the open ocean phytoplankton community [1,2]. Synechococcus is most abundant in mesotrophic open ocean surface waters and coastal regions, and at lower abundances in nutrient-depleted areas [1,3]. Strains of Synechococcus tolerate a range of temperatures [4,5] and can thus extend to high a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 oxidative damage to the PsbA protein through uncontrolled oxidation of amino acids [42,43]. In samples dominated by this photoinactivation mechanism the rate constant of photoinhibition increases linearly with increasing irradiance [44,45]. Furthermore, Mn absorbance in the visible range is sufficient to account for the quantum, "per-photon", yield of photoinhibition [18,41,[46][47][48].
A second photoinactivation mechanism involves production of the reactive oxygen species singlet oxygen ( 1 O 2 ) [48-52] by a a normal triplet-state oxygen molecule reacting with a triplet excited state sensitizer molecule, the 3 P680 state of chlorophyll, yielding 1 O 2 and the singlet ground state of the sensitizer molecule, P680 [48,53]. As electrons are then passed from one molecule to the next of the photosynthetic electron transport chain, short-lived radical intermediate states are generated. In PSII, if electron acceptors are available, electron transfer will happen before the radical intermediate can interact with anything else [18]. However, as the PQ pool becomes fully reduced, forward electron transport may be blocked at the level of Q A or Q B [18,50,54,55]. After primary charge separation a short lived 3 [P680 + Pheo -] charge pair is formed in its triplet ground state [56,57]. 3 P680 thus occurs even under normal conditions and increases when forward electron flow slows through limitations downstream of Q A or Q B [58,59]. Triplet chlorophyll, 3 Chl, a functional equivalent of 3 P680, also occurs in the PSII antenna light harvesting complex in reactions independent of electron transport events [18,52,54,60,61] but proportional to the effective size of the photosynthetic antenna, which determines the average duration of excitation migration through the antenna.
Both the direct absorbance and inactivation of the Mn 4 Ca complex and reactive oxygen dependent photoinactivations occur [41,55,62,63] but show differing responses to incident light. The extinction coefficient (or target size) for a photon hitting the Mn 4 Ca complex is constant and so the probability of a single photon causing a photoinactivation should be independent of incident light level, while the rate of photoinactivation should show a simple linear dependence upon treatment light level. Furthermore, because Mn has higher absorbance in the blue and UV wavelength range and low absorbance in the red wavelength range, the probability of a photoinactivation is higher for a blue photon than for a red photon, no matter the spectral properties of the light harvesting antenna [18,41,64,65]. The probability of direct photoinactivation will, however, vary depending upon the optical screening and package effects around PSII which alter the achieved light incident upon the Mn 4 Ca complex. Since package effect varies with cell size larger cells show a lower probability of direct photoinactivation [65,66].
In contrast, the probability of generation of 1 O 2 shows a positive dependence upon increasing incident light level, which drives saturation of electron carriers out of PSII, leading to an increased probability of reactive oxygen species production [18]. Furthermore the antenna size of the PSII light harvesting complex will positively affect the per-photon probability of photoinactivation of PSII [67] since a larger antenna size supports greater photon capture, resulting in increased excitation pressure at a given light level [16,18,48]. Fig 1 presents a schematic overview of these fluxes and fates of excitonic energy within PSII, drawing upon concepts reviewed elsewhere [18,44,52,62,68]. Incident photons are absorbed by antenna pigment proteins, whose spectral properties vary depending upon the taxon and acclimation state of the cell. Within the antenna complex both regulated and nonregulated processes [69] lead to dissipation of a variable fraction of the excitons as heat. A quantitatively minor fraction of excitons are also dissipated from the antenna as fluorescence, contributing to F O level emission. The remaining, variable, fraction of excitons are transferred into the inner antenna of the PSII reaction centre, which is highly conserved across taxa. Of these excitons a small fraction are again dissipated as fluorescence from the pigments of the reaction centre, contributing to F O level emission but also to the variable fluorescence F V emitted by closed PSII centres. The remaining fraction of the excitons provoke PSII photochemistry, parameterized as σ PSII (A 2 quanta -1 ) [70][71][72], the effective absorption cross section for PSII photochemistry. PSII photochemistry can lead to charge separation and electron transport, parameterized as the yield F PSII , or to fast recombinations resulting in heat or delayed light emission [52]. If excitation delivery is high relative to downstream electron transport excited intermediates can accumulate [18]. This provokes production of singlet oxygen Incident photons (downward waves) are differentially absorbed (ā*) depending upon the particular absorbance spectra of the antenna pigment protein complexes serving PSII, indicated as a green trapezoid to reflect the diversity of antenna pigments and protein structures across taxa. A variable fraction of the excitons is dissipated as heat (yelloworange arrow, NPQ); a further small fraction of excitons are dissipated as fluorescence emitted from the antenna (red arrow). The remaining excitons are transferred into the inner antenna of the PSII reaction centre (downward red arrows). Of these excitons a small but variable fraction are again dissipated as fluorescence (red arrow) from the pigments of the reaction centre. The remaining excitons provoke PSII photochemistry (σ PSII ; A 2 quanta -1 ) leading to charge separation and electron transport with a variable yield (Φ PSII ), or to fast recombinations resulting in heat emission (yellow-orange arrow, NPQ). If excitation is high relative to downstream electron transport excited intermediates can accumulate. This provokes production of singlet oxygen ( 1 O 2 ), which leads to an increase in the probability of PSII inactivation, herein parameterized Φ i PSII , a dimensionless yield for inactivation of a PSII by a photon delivered through the PSII antenna. Short wavelength photons can also be directly absorbed by the manganese cluster of the oxygen evolving complex (thin violet wave), leading to direct photoinactivation. This figure is based upon concepts reviewed in [18,44,52,62,68].
( 1 O 2 ), which leads to an increase in the probability of PSII inactivation, herein parameterized as the yield for inactivation of PSII by a photon delivered through the PSII antenna, F i PSII . F i PSII is thus predicted to increase with rising excitation pressure [62,68]. A few short wavelength photons can also be directly absorbed by the manganese cluster of the oxygen evolving complex, leading to photoinactivation whose probability per photon does not change with excitation pressure [41,48,62,73]. Under low to moderate blue light this direct path can dominate the effective absorption cross for inactivation of PSII on the basis of incident photons, parameterized as σ i (A 2 quanta -1 ) [44, 66,74,75]. The relative magnitudes of the two categories of photoinactivation thus depends upon the spectral quality of the light relative to the antenna spectra of the organism, and upon the physiological state of the organism [76], which influences excitation pressure and capacities for reactive oxygen species detoxification.
Whatever the photoinactivation mechanism(s), as light level increases so will the rate of inactivation of PSII centers which must be countered through removal of inactive PSII proteins and synthesis of new ones to maintain photosynthetic activity. A damaged PsbA protein is degraded by the FtsH and/or Deg proteases [77][78][79], allowing the reconstitution and reactivation of the PSII center with a newly synthesized PsbA. Cells respond to increasing light by activating repair mechanisms including the up-regulation of psbA transcript levels, the gene encoding the PsbA protein [48, [80][81][82][83]. Most cyanobacteria, although not Prochlorococcus, have small psbA gene families encoding two or more isoforms of the PsbA proteins with different photochemical properties [84][85][86] which are differentially induced depending upon irradiance and physiological state. The very repair cycle needed to regenerate PSII is itself strongly sensitive to inhibition by reactive oxygen [87][88][89][90] so net photoinhibition can result from both direct photoinactivation and from concomitant inhibition of the counteracting repair processes.
Photoinactivation and the regeneration of PSII through the PSII repair cycle imposes a significant metabolic burden upon the organism [91][92][93][94], which in turn places ecological limits upon their achieved niches and productivity [95,96]. Given the ecological importance of Synechococcus and Prochlorococcus, and their differential strategies to manage and exploit excitation [8,[95][96][97] we sought to characterize the photoinactivation of PSII in these species. We used controlled culturing to grow representative strains of Synechococcus and Prochlorococcus under low and higher light. We then tracked their susceptibility to photoinactivation of PSII under a range of light levels, when counteracting PSII repair was blocked. We used a blue light treatment as an approximation of the blue-enriched light field in marine waters, and a red light treatment because the comparison with blue light was likely to be mechanistically informative [41,48,68]. We then used the composition of PSII and standard pathways to estimate the metabolic return on investment for repair or synthesis of a PSII, for each species under a range of light levels.

Culturing Conditions and Growth Rates
Synechococcus sp. WH8102 was grown in PCR-S11 media [98] while Prochlorococcus marinus MED4 was grown in Pro99 media [99,100]. Semi-continuous cultures for each species were maintained at low light (30 or 75 μmol photons m -2 s -1 ) or high light (260 μmol photons m -2 s -1 ) at 23-24˚C under fluorescent lights with a 12:12 light:dark cycle with transitions from dark to light at 08:00 hours. A first round of semi-continuous cultures was grown in 50 mL tubes with plastic caps allowing airflow. A second round of semi-continuous cultures was grown in clear 6-well plates with each well containing 6.5 mL of culture. Data from both rounds of semi-continuous culturing was pooled for this manuscript after analyses of variance showed no significant effect of culture round upon measured variables (data not presented).
The growth of the cultures was tracked daily by fluorescence emission (Molecular Devices SpectraMax Gemini EM, Sunnyvale, California). Prochlorococcus was excited at 440 nm and the fluorescence emission was read at 680 nm. Synechococcus was excited at 550 nm and the fluorescence emission was read at 650 nm. For both genera the Relative Fluorescence Units (RFU) of their emission was tracked over elapsed time to calculate exponential growth rates for each sample using the equation: Where RFU t is the fluorescence reading at elapsed time "t" and RFU 0 is the fluorescence reading at time 0. The cultures were kept in exponential growth phase by 1 in 5 dilutions with fresh media every 5-7 days. Cultures were diluted and/ or harvested near the end of exponential phase to harvest the most biomass possible while the culture was still in exponential phase.

Photosystem II Physiological Characterization
The photoinactivation treatments and measurements were carried out in Photon Systems Instruments FL3500 super-heads (Drasov, Czech Republic) with a lab-built aluminum plug for temperature control. For each treatment, a 2 mL sample was loaded into a cuvette with a micro stir-bar and plugged to hold the temperature at 23-24˚C through circulation of cooling fluid through the aluminum plug. The super-head provided three capacities: firstly, application of sub-saturating flashlets of adjustable light level of 1.2 μs duration in red (655 nm LED) or blue (455 nm LED); secondly, detection of the fluorescence emission resulting from the flashlets; and thirdly, delivery of actinic light of adjustable intensity in red or blue (Table 1). We used the super-head flashlet capacity to perform fast repetition rate (FRR) chlorophyll fluorescence induction profiles, which were fit to a four parameter model ( Table 2) to quantify PSII physiology [70] using PSIWORX-R data handling and fitting software implemented in the R environment (A. Barnett, http://sourceforge.net/projects/psiworx). The FRR induction curves were driven by a train of 40 flashlets of 1.2 μs each separated by 2.0 μs dark, for a cumulative flashlet train duration of 128 μs, shorter than the 200-500 μs for transfer of an electron from Qa to Qb, and almost 1 order of magnitude less than the subsequent~1000 μs that it takes for the bound Qb pool of PSII to equilibrate with the plastoquinone pool [16,70]. FRR inductions were measured in the presence of background actinic light, and then again 2 s after the end of actinic light exposures.

Whole-Cell Spectroscopy
An absorbance spectrum was taken for each sample using an Olis CLARiTY DSPC spectrophotometer (Bogart, Georgia). The integrating, internally reflective DSPC cavity records accurate absorbance spectra from dilute suspensions of phytoplankton cells with high light scattering relative to low absorbance. Spectra were taken in the visible range, λ = 390-750 nm, at 1 nm intervals. Spectra of media blanks were subtracted from their respective genera and a conversion was performed to correct for chamber reflectivity using Fry's Method [101]. This data transform takes the total absorbance spectra from the integrating, internally reflective cavity, and corrects it to the equivalent absorbance expected from a hypothetical 1 cm path length spectrophotometer cuvette measurement without scattering. Fig 2A compares typical absorbance spectra for Prochlorococcus and Synehococcus with the emission wavelength profiles of the LED lamps used to provide the light treatments.

Photoinactivation Treatments
To measure photoinactivation the overlapping effects of the PSII repair cycle and non-photochemical quenching must both be considered [44,68]. Lincomycin at a final concentration of 500 μg mL -1 , was used to inhibit prokaryotic ribosomes and thereby block the PSII repair cycle during the photoinactivation treatments [48, 73,97,102]. Samples were subjected to light shift treatments of 10 consecutive periods of 327 s each, totaling 3270 s at a range of levels of blue or red treatment light ( Table 1). The choice of 327 s was imposed by a limitation on the maximum time between sequential measurements in the software control of the fluorometer. Fig 2B presents a schematic overview of the photoinactivation treatment and measurement protocol. A FRR measurement was taken 2s after each illumination period and analyzed to extract photophysiological parameters ( Table 2) including σ PSII 0 2s and F V 0 2s/F M 0 2s estimates corresponding to their respective preceding period.
The initial dark period allowed for the measurement of σ PSII after relaxation of any nonphotochemical quenching induced by growth conditions. F V 0 2s/F M 0 2s taken after 327 s under the sample growth light gave the maximal photochemical yield of PSII in the light acclimated state. F V 0 2s/F M 0 2s measured after each of 8 successive 327 s periods under the treatment light tracked the maximum photochemical yield in response to the light shift treatment. Finally, the terminal tenth low light period from 2943 to 3270 s then allowed the sample to reverse any minimal fluorescence with PSII open, estimated for cells under excitation, excluding cumulative influence of photoinactivation. [108,111] σ PSII effective absorption cross section, Å 2 photon -1 , for PSII photochemistry [70] σ PSII 0 effective absorption cross section, Å 2 photon -1 , for PSII photochemistry under excitation [70] σ PSII 0 2s effective absorption cross section, Å 2 photon -1 , for PSII photochemistry 2 s after excitation [110] Yield of PSII photoinactivations per photon delivered to PSII through the effective absorption cross section for PSII photochemistry σ PSII 0 this work   Table 1. The tenth period was a low light recovery phase of 20 μmol photons m -2 s -1 . The black triangle indicates a Fast Repetition chlorophyll fluorescence induction measurement taken after the initial dark period, using 40 flashlets of 1.2 μs duration, spaced by 2 μs darkness, which cumulatively delivered a single turnover saturating flash over 128 μs. From this induction changes caused by sustained NPQ which persisted through the 2s dark periods before measurements, but which could relax over the final 327 s under low light in the absence of protein synthesis, blocked by lincomycin. The influence of sustained NPQ on photochemical yield was then estimated as: where Influence of Sustained NPQ was forced to be !0. The Influence of Sustained NPQ on photoinhibition were then corrected by adding the value to F V 0 2s/F M 0 2s measured after each treatment light interval. Therefore, we attributed any remaining decline in the maximum photochemical yield of (+) lincomycin samples to photoinactivation, which did not reverse in the absence of protein synthesis under the final 327 s of low light (Fig 2B). For the cultures in this study this correction had little influence on the subsequent curve fits, although in other taxa sustained NPQ is a significant factor confounding estimations of the photoinactivation of PSII [103].

Quantitation of Photoinactivation
The corrected F V 0 2s/F M 0 2s values were then plotted against cumulative incident photons for the treatment light period.
Cumulative Incident Photons photons m À 2 ¼ ð6:022 Â 10 17 photons mmol À 1 Þ Â S n¼2À 9 ðLight Level of Period n mmol photons m À 2 s À 1 Â Duration of Period n sÞ ð3Þ Where cumulative incident photons had units of quanta m -2 , treatment light level had units of μmol photons m -2 s -1 , and the duration of each period was a constant 327 s.
A single-phase exponential decay curve was fit to the data using the following equation: where σ i is a target size parameterization [44, 66,74] of the probability of a photoinactivation per incident photon with units of m 2 quanta -1 . Higher σ i values indicate a higher probability of photoinactivation per incident photon, and a lower σ i indicates a lower susceptibility to photoinactivation per incident photon. Fig 3A presents representative photoinactivation curves measured in the presence of lincomycin, under red or blue light treatments.

Correcting σ i for Photons Delivered to PSII
To correct for effects on σ i caused by spectral or regulatory differences in delivery of excitation to PSII through the light harvesting antennae, susceptibility to photoinactivation was recalculated using cumulative photons delivered to PSII, as opposed to the previous σ i calculation using cumulative photons incident upon the cells. A single-phase exponential decay was used curve we used PSIWORX-R script to extract estimates for F 0 , F M and σ PSII . Each green triangle thereafter indicates the timing of a chlorophyll fluorescence induction measurement taken after an illuminated period, represented in the inset above the treatment trace. At each measurement point we captured an induction applied in the presence of continuing actinic irradiance to extract estimates for F S , F M 0 and σ PSII 0 . We then interrupted the actinic irradiance for 2 s of darkness to allow PSII centres to re-open, followed by another fluorescence induction to extract estimates for F 0 0 , F M 0 (2s) and σ PSII 0 (2s). The post-induction fluorescence relaxation phase of the full Fast Repetition and Relaxation fluorescence profile is omitted from this schematic diagram for clarity. to calculate the yield of photoinactivation of PSII on the basis of photons delivered to PSII (F i PSII ): The parameter F i PSII is a formally dimension-less yield (PSII inactivated photon -1 = quanta quanta -1 ) that expresses the number of PSII photoinactivation events per photon delivered to PSII through the effective absorption cross section for PSII photochemistry. We earlier expressed a conceptually similar concept [44] as the ratio of the effective absorption cross section for PSII photochemistry, σ PSII to the effective absorption cross section for photoinactivation, σ i . Herein we used time-specific σ PSII 0 2s values measured every 327 s throughout the light-shift photoinactivation treatment to give the effective absorption cross section of PSII over that 327 s, in order to calculate the photons delivered to PSII over that period. Fig 3B presents representative photoinactivation curves measured in the presence of lincomycin, under red or blue light treatments plotted versus cumulative photons delivered to PSII.

Photoinactivation of Picocyanobacteria
Testing the Response of Φ i PSII to PSII Closure If photoinactivation is driven by reactive oxygen by-products of excitation and electron fluxes through PSII [18,52,105] the susceptibility to photoinactivation should increase as excitation of PSII increases relative to electron transport away from PSII. To determine whether excitation pressure, the balance between excitation delivery to PSII, and electron flow away from PSII [106], indeed affected susceptibility to photoinactivation [18,52] we used the parameter 1 -q P [107] where a value of 0 indicates all PSII are open and ready to accept excitation and a value of 1 indicates all PSII are closed and thus more susceptible to generation of reactive oxygen species. We calculated 1 -q P for a given treatment after the initial 327 s exposure to the treatment light, as the value remained relatively unchanged thereafter with further treatment light exposure (data not shown). In these calculations we used a calculated F 0 0 (

PSII Electron Transport versus Electron Equivalents to Recycle PSII
In order to determine the number of photochemical events performed per s by PSII under the treatment light, we used the effective absorption cross section of PSII under treatment light after 2 s of darkness (σ PSII 0 2s), the proportion of open PSII when exposed to treatment light (q P ) and the intensity of the treatment light (I) [113,114]: In order to determine the number of photoinactivation events per PSII per s, we multiplied the effective cross section for photoinactivation (σ i ) (Å 2 quanta -1 ) by the intensity of the treatment light (I) (1 μmol photons m -2 s -1 = 6.022 × 10 −3 quanta Å -2 s -1 ). We then estimated the investment in metabolic electron equivalents of ATP and reductant needed in order to either synthesize a complete de novo PSII (1.96 × 10 5 e-/PSII), or to recycle most PSII protein subunits while proteolytically degrading and re-synthesizing the equivalent of new PsbA and PsbD proteins (5.77 × 10 3 e-/(PsbA + PsbD)).
We arrived at these metabolic electron equivalent costs by summing the amino acid composition of the proteins of PSII [92,115] and accounting for the ATP costs of ribosomal protein synthesis [116][117][118], the electron and ATP costs for assimilation of CO 2 and NO 3 to amino acids [119][120][121], and the ATP costs for FtsH protease turnover of PsbA and PsbD subunits [122,123]. We converted ATP costs to electron equivalents with the textbook ratio of 1.33 e -/ ATP for respiratory electron transport. These approximations serve to place the cost of PSII inactivation on a common denominator for comparison with electron generation by PSII.

Results & Discussion
σ i varies with Treatment Light Intensity and with Excitation Pressure on PSII σ i estimates were derived from individual photoinactivation time courses in the presence of lincomycin (Fig 3A) for combinations of species, growth light level, treatment light colour and treatment light intensity (Table 1). Fig 4 presents these estimates of σ i plotted versus the treatment light intensity. For fitting of regressions each σ i estimate was weighted by the inverse of its 95% confidence interval, plotted as error bars on the points, to account for variability in the precision of individual estimates of σ i (Fig 3A).
Under red light σ i was low for both Synechococcus and Prochlorococcus. For Synechococcus the Y intercept of the regression of σ i versus treatment light intensity was not significantly different from 0, but there was a significant increase in photoinactivation with increasing intensity of red light treatment. In Prochlorococcus the Y intercept was small and there was no significant increase in σ i with increasing intensity of red light treatment. Under blue light treatment σ i was significantly higher than under red light, consistent with expectations [41,48,65,68,124,125]. Synechococcus cultures grown under high light fell on regression with intercept and slope distinct from Synechococcus cultures grown under lower light. When treated under blue light the Prochlorococcus low light and higher light grown cultures all fell on a common regression but there was an increase in σ i with increasing intensity of blue light treatment.
The absorbance spectra of the cultures varied with species (Fig 2A) and with growth light (data not shown). The photosynthetic physiologies of the species are also distinct with evidence for down-stream limitations on the rate of electron transport away from PSII in

Photoinactivation of Picocyanobacteria
Prochlorococcus when compared to Synechococcus [11,126,127]. Therefore a given treatment light level or colour could have differing functional implications for the cell depending upon species and prior light regime [76]. We therefore decided to replot the σ i estimates versus the excitation pressure upon PSII imposed by each given combination of treatment light intensity and colour (Fig 5A and 5B). These replots show patterns very similar to the plots of σ i versus treatment light intensity (Fig 4A and 4B), except that under red light treatment Prochlorococcus shows a small positive response of σ i to increasing excitation pressure.
The Y intercepts on these plots represent the inherent photoinactivation potential of an incident photon in the absence of any excitation pressure, and therefore in the absence of any electron transport or risk of reactive oxygen generation by PSII or related electron flows. Under red light treatment both Synechococcus and Prochlorococcus show Y intercepts not significantly different from 0 for these plots of σ i versus increasing excitation pressure. Thus at Fig 5. σ i versus excitation pressure on Photosystem II. σ i plotted against excitation pressure, measured as 1-q P after 300 s of treatment light exposure. Under red or blue treatment light, q P showed remained nearly steady from 327 s onwards (data not shown) so the plotted values for 1-q P were taken after the first period of 327 exposure to treatment light. For regressions each σ i derived from an individual treatment timecourse (Fig 3A) was weighted by the inverse of its 95% confidence interval (error bars on the points) to account for variability in the precision of individual estimates of σ i . . σ i estimates derived from red light treatment (low light grown cultures, red inverted triangles; high light grown cultures, red diamonds) fell on a single regression (solid red line, slope = 3.484 × 10 −5 ± 1.222 × 10 −5 , intercept not significantly different from zero, p>0.05;, R 2 = 0.5755, dotted red lines denote 95% confidence intervals). σ i estimates derived from blue light treatment (low light grown cultures, blue triangles; high light grown cultures, blue squares) fell on a single regression (solid blue line, slope = 1.262 × 10 −4 ± 5.667 × 10 −5 , intercept not significantly different from zero, p>0.05; R 2 = 0.2259, dotted blue lines denote 95% confidence intervals) (S4 Statistics). doi:10.1371/journal.pone.0168991.g005 Photoinactivation of Picocyanobacteria the Y intercept, in the absence of electron transport, red photons show no detectable intrinsic toxicity to PSII. The low level of photoinactivation under red light is thus quantitatively attributable to by-products of electron transport through PSII.
Under blue light treatments Synechococcus grown under lower light, and Prochlorococcus, σ i shows Y intercepts significantly higher than 0, demonstrating an intrinsic toxicity of blue light to PSII [48, 62,68,124], even in the absence of electron flow and its concomitant reactive oxygen species byproducts. Synechococcus grown under higher light showed low (Fig 4A) or insignificant (Fig 5A) Y intercepts for σ i , and thus showed a lower intrinsic susceptibility to blue light photoinactivation, which could relate to expression of alternate isoforms of the PsbA protein [84][85][86] in Synechococcus.
The rising slopes of σ i with increasing excitation pressure upon PSII represent the outcome of a complex balance among multiple paths [18,52,105] of potential reactive oxygen species production related to electron fluxes (Fig 1) and the counteracting detoxification mechanisms whose presence and induction can vary both with species and with prior growth conditions [8,76].

Φ i PSII versus Excitation Pressure on PSII to Reconcile Species and Growth Light Effects
We next asked whether the photoinactivation patterns could be reconciled by considering the yield of photoinactivation on the basis of photons delivered to PSII via the light harvesting antenna. We therefore estimated a new parameterization, F i PSII to express the yield of photoinactivation of PSII relative to photons driving PSII photochemistry, as measured by the effective absorption cross section for photochemistry, σ PSII . For plots of F i PSII versus excitation pressure (Fig 6) Synechococcus and Prochlorococcus from both low and higher prior growth lights all fell on a common regression under red light treatment, and on a common regression for blue light treatment. The two treatment lights gave equivalent slopes with increasing excitation pressure, but were distinguised by a higher Y intercept for blue light treatments. Thus the intrinsic toxicity of blue light, likely through direct photoinactivation of the manganese cluster of PSII [41,48] manifests as a strong potential for photoinactivation by blue photons even when electron flow through PSII is negligible. With rising excitation pressure both blue and red light drive comparable increases in photoinactivation, when expressed on the basis of excitation actually delivered to PSII photochemistry, rather than on the basis of incident photons. The relative importance of these distinct photoinactivation paths [62,68] will depend upon cellular absorbance spectra and the light level, but more particularly upon the physiological excitation pressure upon PSII, a proxy for the probability of generation of reactive oxygen species. Under physiologically low to moderate levels of light direct photoinactivation by blue light can dominate, and can be parameterized as a simple target size [66,[73][74][75]. This conditions would often prevail in marine systems. As excitation pressure rises either through increasing light in near-surface environments or through factors that restrict the down-stream removal of electrons from PSII [39,106] the reactive oxygen species -dependent photoinactivation paths related to electron transport will increase and even predominate. These patterns become apparent across spectrally diverse species with different growth histories by expressing photoinactivation on the basis of photons delivered to PSII, plotted versus excitation pressure, rather than the treatment light level per se.
For regressions each F i PSII derived from an individual treatment time course (Fig 3B) was weighted by the inverse of its 95% confidence interval (error bars on the points) to account for variability in the precision of individual estimates of F i PSII .

The Return on Investment for Photosystem II in the Face of Photoinactivation
The inactivation of PSII and the counteracting repair processes impose a significant, and variable, metabolic burden upon photoautotrophs [91][92][93][94]119] that interacts with nutrient supply [129,130] and with cell size [66]. We thus sought to use the optically simple Synechococcus and Prochlorococcus to compare the photochemical return on investment from a PSII, to the costs for net biosynthesis of a PSII from inorganic precursors, or to the costs to degrade and replace the PsbA and PsbD subunits of the complex (Fig 7). Our present analyses is restricted to nutrient replete cultures, but nutrient limitation likely interacts with these processes [131,132]. Photochemical events through PSII were estimated as (σ PSII 0 2s) × q P × E. Photoinactivation events  Photoinactivation of Picocyanobacteria  Fig 7 plots a narrow black segment along the X axes to show the low metabolic cost needed to repair a photoinactivated PSII through degradation and re-synthesis of new PsbA/PsbD proteins. Across the tested range of blue or red treatment light levels Synechococcus PSII electron transport stayed well above this cost to repair a photoinactivated PSII, so PSII repair was always a metabolically viable strategy for our Synechococcus cultures, at least when growing under nutrient repletion. In contrast in Prochlorococcus much lower PSII electron transport, possibly limited by downstream metabolic capacity [127] means that the photochemical return from a PSII drops to levels similar to the metabolic cost to repair a PSII at light levels of 1200 μmol photons m -2 s -1 for red or for blue treatment lights. The wider grey segment shows the metabolic cost for net biosynthesis of a new PSII from inorganic precursors at a rate sufficient to counter photoinactivation. Again, Synechococcus enjoys a positive net return on newly synthesized PSII up to and beyond 1200 μmol photons m -2 s -1 , under both blue and red treatment lights, at least when growing under nutrient repletion and near-optimal temperatures. In marked contrast for Prochlorococcus under blue light the net cost to biosynthesize a new PSII exceeds the anticipated photochemical return from the new PSII at light levels abovẽ 300 μmol photons m -2 s -1 . Under the informative but less ecophysiologically relevant red light the lower rates of photoinactivation mean Prochlorococcus enjoys a positive return upon a new PSII up to light levels of~1200 μmol photons m -2 s -1 . These contrasting patterns in Synechococcus and Prochlorococcus support findings [8] that Synechococcus tends towards active mechanisms to cope with increasing excitation. In contrast Prochlorococcus tends to merely endure transient exposure to excess excitation and might trend towards mixotrophy under some conditions [133] when maintaining photosynthesis becomes untenable. These contrasting patterns also underlie findings of different excitation tolerance thresholds for the different picocyanobacteria [95,96], and their niche partitioning [97].