Light acclimation and pH perturbations affect photosynthetic performance in Chlorella mass culture

Chlorella spp. are robust chlorophyte microalgal species frequently used in mass culture. The pH optimum for growth is close to neutrality; at this pH, theoretically little energy is required to maintain homeostasis. In the present study, we grew Chlorella fusca cells in an open, outdoor, thin-layer cascade photobioreactor (TLC), under ambient photon flux at the theoretically preferred pH (7.2), and let the culture pass the exponential growth phase. Using pH drift experiments, we show that an alkalization to pH 9 supported photosynthesis in the TLC. The increased photosynthetic activity under alkaline conditions was a pH-dependent effect, and not a dissolved inorganic carbon (DIC) concentrationor light intensity-dependent effect. Re-acidification (in one step or in increments) lowered gross oxygen production and increased non-photochemical quenching in short-term experiments. Gross oxygen production and electron transport rates in PSII were uncoupled during the pH perturbation experiments. Electron transport rates were only marginally affected by pH, whereas oxygen production rates decreased with acidification. Alternative electron pathways, electron donation at the plastid terminal oxidase and state-transitions are discussed as a potential explanation. Because cell material from the TLC was not operating at maximal capacity, we propose that alkalization can support photosynthesis in challenged TLC systems.


Scenescence and mass culture
Microalgae mass culture generally aims for high biomass yields, often associated with intended nutrient starvation to increase lipid contents (Jacobsen et al. 2010, Liu et al. 2013).Nitrogen starvation lowers protein contents, RuBisCO levels, and activates bio-chemical responses that lead to an elevation of lipid bodies by a process that is not entirely understood.Algae mass-culture conditions are frequently less controlled, the exponential growth phase might have passed during the time of cell harvest and cells have experienced senescence.Under these conditions cells can exhibit acclimation properties which are comparable to nutrient starvation conditions.Lower photosynthetic performance is one result of senes-cence (Humby et al. 2013).When cells enter the stationary growth phase, chloroplast structures become disorganized, LHCII, D1, PsaA, Cyt f, and RuBisCO levels decline, and photosynthesis is down-regulated while non-photochemical quenching (NPQ) remains stable compared to the exponential growth phase (Humby et al. 2013).These substantial changes within the chloroplast might affect the cell's response to external factors such as pH and light intensity.Gardner et al. (2011) showed that a Chlorella strain demonstrated elevated lipid contents and displayed morphological changes at high pH, which indicates that the elevated pH induced stress to the cell.

Homeostasis: effect of external pH manipulation
The external pH under which the cell is capable of operating at maximal capacity is species-specific but usually range between pH 6 and 8.5 (Taraldsvik & Mykleastad 2000, Hinga 2002, Lundholm et al. 2004, Liu et al. 2007, Middelboe & Hansen 2007).Although Chlorella spp.prefer a pH between 6 and 6.5 (Myers 1953), Chlorella vulgaris can grow effectively at an alkaline pH of 10.5 (Goldman et al. 1982).Chlorella saccharophila grows optimally at pH 7.0 but can maintain homeostasis when external pH is as low as 5.0 (internal pH 7.3) (Gehl & Colman 1985).This strain can successfully grow at pH 2.5 (Beardall 1981), which shows its immense internal pH regulation capacity.Internal pH is generally maintained within narrow limits above ~pH 7 in Chlorella (Smith & Raven 1979, Beardall & Raven 1981, Sianoudis et al. 1987) and other taxa (Dixon et al. 1989, Kurkdjian & Guern 1989).Homeostasis is maintained by 2 main processes: proton binding and H + translocation between cell organelles and the external medium by active, energy-demanding, or passive ion exchange mechanisms (Smith & Raven 1979).The energy demand for active proton translocation is estimated to be 1 ATP per H + transported (Briskin & Hanson 1992).The abundance of H + ATPase complexes can be controlled by the cell via pH sensing (Weiss & Pick 1996).However, Nielsen et al. (2007) showed that light intensity (and hence energy supply) did not correlate with cell growth under extreme pH conditions, which indicates that elevated light input does not necessarily enhance the cell's capacity to regulate internal pH.In short-term pH manipulations, H + pumping is important (Bethmann & Schönknecht 2009); however, under longer-term (growth) conditions, anion exchange might be the predominant proton translocation process (Lew 2010).It is not clear how internal pH is regulated in cells that have passed their exponential growth phase.

Carbon acquisition and pH
Changes in the media's pH or supplementation of CO 2 affects the bicarbonate equilibrium.Introducing CO 2 into the medium lowers the pH and shifts the bicarbonate equilibrium towards higher concentrations of CO 2 and bicarbonate at the expense of carbonate.If the pH is increased at a constant pCO 2 , the dissolved inorganic carbon (DIC) concentration increases, while the CO 2(aq) :HCO 3 − ratio decreases.The form of DIC is relevant for the cellular DIC acquisition system, while the concentration of DIC matters when substantial rates of photosynthesis must be sustained.Most aquatic primary producers possess means of elevating the CO 2 concentration at RuBisCO, which suppresses RuBisCO's oxygenase function and provides sufficient CO 2 even when high photon flux (PF) allows high CO 2 fixation rates in the Calvin-Benson-Bassham-cycle.Most algae can regulate these CO 2 -concentrating mechanisms (CCMs) effectively (Giordano et al. 2005).Chlorella spp.possess effective CCMs and acquire both CO 2 and/or HCO 3 − actively (Shelp & Canvin 1980, Beardall 1981, Beardall & Raven 1981), and are able to raise the pH in a closed vessel to pH 11 (Myers 1953), which shows that cells are capable of carrying out effective photosynthesis.The interlinked pH and CO 2 concentration behavior makes it difficult to separate both effects.
In the present study, we tested Chlorella fusca cells grown in an open, outdoor, thin-layer cascade photobioreactor (TLC) under post-exponential growth phase conditions.We investigated if cells are susceptible to high PF, are limited by external DIC concentrations or affected by changes in the medium pH.We perturbed pH in short-term (i.e.minutes) and long-term (i.e.hours: pH drift) experiments to test the cells' capacity to regulate pH-related photosynthesis.The results show that cells performed better in alkaline conditions and were stressed by acidification of the media.

Organism, growth conditions and growth history
The chlorophyte microalga Chlorella fusca (Culture Collection of Marine Microalgae, ICMAN-CSIC, Cádiz, Spain) was grown in an outdoor thin-layer cas cade photobioreactor (TLC3; TLC number is given to enable comparison with to other studies using the same experi mental setup) in September 2012 in southern Spain (Málaga), using Bold's Basal Medium modified with 3-fold nitrate content plus the addition of vitamins (3N-BBM-V) (Andersen et al. 2005).The cell suspension was held in an open tank (~120 l), pumped to the top of a flat panel slide (4 m 2 , 2% inclination), and allowed to flow back to the tank by gravity (surface to volume ratio 27 m −3 , 145 l working volume).The cells received full sunlight on the cascade (for a duration of ~11 s) and were exposed to low light conditions in the tank.Completion of a single cycle took approximately 70 s.Macronutrients (NaNO 3 , MgSO 4 and KH 2 PO 4 ) were added once a day using common farming fertilisers (Welgro Hydroponic, Comercial Química Massó); nu trient concentrations were ≥540 mg l −1 (PO 4 − and SO 4 ), and ≥650 mg l −1 (NO 3 − ; R. Abdala pers.obs.).The TLC was aerated; pure CO 2 gas was injected into the air stream (final 1%) and the pH was maintained at between 7.2 and 7.6.pH drift experiments showed very limited capacity of the cells to increase the pH by means of photosynthesis (see Fig. 1A), even when DIC was added in form of NaHCO 3 (see Fig. A1A in the Appendix), or when diluted with fresh medium (Fig. A1C).We therefore tested the photophysiological fitness in a light acclimation ex periment (see Figs. 2 & 3), which showed that photosynthetic capacity was not severely im paired.However, increasing the pH by substantial addition of NaHCO 3 (20 mM) in the TLC increased pH drift (see Fig. 1).pH perturbance experiments were carried out after pH adjustments had been performed in the TLC.Cell numbers de creased for 2 d prior to the experiments (see Figs. 4-7) (from ~1 × 10 6 to ~4 × 10 5 cells ml −1 ), but remained stable thereafter (counted using a Neubauer chamber; R. Abdala pers.obs.).At that point, the TLC had been maintained for 12 d, at a temperature of ~25°C.The TLC used in the present study was part of a nutrient-concentration effect study.A TLC with replete nutrients concentrations (TLC presented in the present study), was maintained parallel to a TLC with low nitrogen (TLC-N) and another TLC with lowered sulphate concentrations (TLC-S).We expected the TLC with replete nutrients to show highest cell numbers.However, maximal cell concentrations were lower than expected (TLC of the present study, i.e. replete nutrients = 2 × 10 7 cells ml -1 , TLC-S = 2.3 × 10 7 cells ml -1 , and TLC-N = 1 × 10 6 cells m -1 l; R. Abdala pers.comm.).The experiments in the present study were carried out after the maximal TLC capacity had passed, and cell suspen-sions experienced grazing, grazer treatment and were contaminated with Nannochloris sp. and Chlamy domonas sp. cells in variable concentrations.
More information about the hydrodynamics of the TLC used and light exposure effects in the TLC on photo synthetic performance is given in Jerez (2014, this Theme Section).Further information on photoacclimation in the TLC will be published elsewhere (J.C. Kromkamp et al. unpubl.).

pH drift experiments
Samples were withdrawn from the TLC in the morning, transferred to 20 ml glass scintillation vials, and either continuously exposed to ambient sunlight or shaded by neutral density filters.pH was measured several times a day (Basic 20, Crison Instruments), calibrated with standard laboratory solutions.Measurement duration was standardized to 1 min vial −1 to avoid erratic gas exchange amongst replications.Temperatures were in the range of 25 to 28°C and maintained by placing vials in a non-controlled tub water bath and regularly exchanging the cooling fluid.

Light acclimation experiment
Cell suspensions for continuous light exposure experiments were withdrawn from the TLC in the evening.250 ml were filled into 500 ml conical glass flasks, which were kept in the dark overnight and then exposed to ambient light conditions (HL, high light) the following day while aerated with ambient air (~1 l min −1 ).Shading was achieved by neutral density filters to 70% (ML, medium light) or 30% (LL, low light) of the ambient PF.Temperature conditions were maintained at approximately 25 ± 3°C by placing flasks in a styrofoam tub filled to one-third with water (which was replaced when needed).At given times, 2 ml samples were withdrawn, immediately placed in an AquaPen fluorometer (AquaPen, P.S.I.) and fast fluorescence induction curves measured.Maximum and effective quantum yield of PSII (F v /F m and F v '/F m ') and Vj (accumulation of the 'primary' acceptor of PSII, Q A − ) were computed by the software provided by the instrument.Initial samples were dark-incubated overnight.Other samples were measured in the absence of actinic light, which allows Q A − to oxidize.However, because samples were not dark-acclimated prior to measurements, nor exposed to ambient actinic light, residual NPQ can lower the fluorescence signal and decrease variable fluorescence.This results in F 0 ' dur-ing the O phase of the fast fluorescence induction curve, rather than F' or F 0 (i.e.quantum yields are F v '/F m ' rather than ΔF/F m ') (van Kooten & Snel 1990, Kromkamp & Forster 2003).

Fast fluorescence induction curves
Fast fluorescence induction curves represent the consecutive reduction of electron carriers in PSII and the plastoquinone (PQ) pool within about 1 s of illumination with saturating light.After the O phase (first data point at t = 50 µs), Q A gets reduced until the J phase is reached (100% Q A -, 0% Q A , 2 ms), while the consecutive reduction of Q B and formation of plastoquinol (PQ pool reduction) is associated with the inflection (I phase) and the maximal fluorescence signal (P phase at t = 1 s) by a fully reduced electron transport chain from the PSII reaction centre to the PQ pool (Tomek et al. 2001, Zhu et al. 2005).This concept is commonly accepted, however, alternative interpretation of fast fluorescence induction curves have been frequently presented due to the high number of fluorescence quenchers in PSII and the photosynthetic unit (Strasser et al. 1995, Bukhov et al. 2004, Stirbet & Govindjee 2012).The reduction state of the PQ pool, for instance, was not a determinant factor for maximal fluorescence signals during P, but the reduced PQ pool has been shown to affect the J phase (Tóth et al. 2005(Tóth et al. , 2007)).Nevertheless, we will interpret data of this study following Zhu et al. (2005), where signals show the accumulation of the following electron carrier: , PQ is fully reduced.Vj was calculated as (J − O) / (P − O), with J phase = F 2ms (where F = fluorescence), O = F 50 µs , P = F 1s . 1 − Vj is indicative for the electron transport capacity past Q A .Low values represent a lower probability that electrons are efficiently transported.Note that P = F m (or F m ') in multiple-turnover saturation pulse fluorometers (e.g.pulse-amplitude modulation, PAM), while J is equivalent to maximal fluorescence in a single-turnover fluorometer (e.g.fast repetition rate fluorometer, FRRf).

pH perturbations: simultaneous PAM, O 2 and pH measurements
Cell suspensions were withdrawn from the TLC and placed in a 20 ml glass scintillation vial.The lid was modified to allow submersion of a standard pH electrode (Basic 20, Crison Instruments), and an oxygen optode (Microx TX3 PreSens) into the cell sus-pension, while minimizing the contact surface of the cell suspension with ambient air.The optical fibre of a Mini-PAM (Walz) was placed against the top onethird of the vial and orientated perpendicular to the actinic (white LED) light source, which provided a constant PF of 140 µmol photons m −2 s −1 .Higher actinic PF was achieved by usage of the internal light source of the fluorometer and exposed cells to 400 µmol photons m −2 s −1 when switched on.However, cells moved into and out of the additional light field since only a part of the scintillation vial was exposed to the additional PF.This can limit comparison with the FRRf experiment, where the entire volume of the cell suspension was ex posed to PF conditions.Cell suspensions were stirred con tinuously.After a ~4 min acclimation phase, additions of known amounts of HCl perturbed the pH, which was recorded manually every 30 s.To test for DIC limitation, 300 µl of NaHCO 3 (1 mM final concentration) were injected at the times indicated.Experiments were conducted from morning until noon, before high ambient light could cause inhibition.

pH perturbations: fast repetition rate fluorescence measurements
FRRf fluorometers use a different excitation and measurement protocol compared to PAM instruments.While PAM measures F m or F m ', during a multiple turnover (of PSII) saturation pulse (approx.800 ms duration), FRRf uses a sequence of subsaturating flashlets to consecutively reduce PSII to a single turnover.The FRR fluorometer used (Fast-Tracka Mark II, Chelsea Technology Group) was equipped with a bench-top illumination extension (FastAct, Chelsea Technology Group) and was set to apply a sequence of 100 flashlets, each 1 µs long and spaced 50 µs apart.Data from 12 such sequences were averaged by the standard FRRf software (Fast-Pro) and produced quantum yields (F v /F m , ΔF/F m ') of a single turnover in PSII, i.e. all Q A reduced, PQ-pool not affected.An iterative curve fit of a single excitation flashlet sequence provides measurements of the functional absorption cross-section of PSII (in nm 2 ).A sequence of single measuring flashlets after the single turnover excitation protocol shows Q A − re-oxidation kinetics (τ PSII ), i.e. the electron transport capacity from a fully reduced Q A .Further information about FRRf fluorometry can be found in Kolber & Falkowski (1992, 1993) and Kolber et al. (1998).
A total of 3 ml of microalgae culture was withdrawn from the TLC on the same day as PAM pH perturbation experiments were carried out, and immediately placed in the FRR fluorometer.An internal LED provided actinic PF of 3, 140, 400, 140 and 0 µmol photons m −2 s −1 in this sequence for the times indicated.Acidification 5 min after the start of the protocol was achieved by addition of 8.6 µl of 0.5 M HCl.After 7 min, the pH was brought back up by injection of 51 µl 0.1 M NaOH.Samples were homogenized by aeration between each saturation flashlet application, which enables gas exchange between the cell suspension and air, and complicates data interpretation regarding the bicarbonate system and pH.

RESULTS pH drift experiments
pH drift was monitored daily.Until the pH was increased in the TLC, incubations could elevate the pH by a maximum of 0.5 pH units (examples are given in Fig. A1 in the Appendix).To test if cells were DIC limited, we added small amounts of DIC (400 and 800 mM), which increased the initial value (pH 7.11) marginally (pH 0.06 and 0.23 units respectively).Incubations with increased DIC concentrations showed similar pH rise kinetics compared to the control (Fig. A1A,B).pH drift experiments where fresh media was added to cell suspensions from the TLC (1:1 by volume) did not stimulate pH drift compared to the control, neither did CO 2 bubbling performed before the experiment (Fig. A1C).This indicates that cells were not nutrient starved or depleted.When the pH in the TLC was adjusted (from pH 7 to 9), pH drift increased (initial pH 8.47, final pH 9.57; Fig. 1).No pH increase was recorded in suspensions collected before the DIC addition.A stronger pH drift was observed on the following day (initial pH 8.65, final pH 11.5 ± 0.06; Fig. 1B).A similar effect was pre viously observed in an other TLC (TLC2) (S.Ihnken pers.obs., J. C. Kromkamp et al. unpubl.),where the pH was increased from pH 7 to 8.5 and drift experiments showed increased pH rise (initial pH 8.5, final pH 10.5).
Fig. 1B shows that shielding the vials at 70% of ambient PF (i.e.ML) resulted in the highest final pH.Higher and lower PF levels showed a similar pH drift capacity, with some variation due to photoinhibition (HL), or light limitation (LL).Control flasks, which experienced the same treatment but remained closed for the entire period, showed similar final pH values compared to vials that were used for pH measurements, indicating that the measurements and the accompanying brief exposure to air did not influence the results.In samples which experienced full ambient PF (HL), however, a lower final pH was found in the control vials (pH 11.09 ± 0.25 and pH 10.02 ± 0.18 for opened and continuously closed vials respectively).

Light acclimation status
Fig. 2 shows the light acclimation capacities in a separate experiment where cells were withdrawn from the TLC and continuously exposed to ambient PF.F v '/F m ' decreased with increasing ambient PF in all treatments.The lowest quantum yields were found in full light exposed samples after the highest ambient PF, while ML cells had already started to recover at this point (t = 15:00 h).HL cells recovered from low midday yields in the evening.LL cells lowered values only marginally around noon.F v '/F m ' depressions in HL and ML samples were partly caused by impairment of electron transport capacities past Q A . 1 − Vj decreased by 1 unit at the highest PF, but recovered later during the day (Fig. 2B).Low re-oxidation turnover for electron carriers past Q A , as shown by 1 − Vj, were also visible in low P phase values during the fast fluorescence induction in HL samples (Fig. 3).The lowest P phase values were found in the afternoon (15:00 h, HL), but samples re covered thereafter.J phase signals were only re pressed in HL samples but remained low in the evening.In LL samples, the J phase was hardly affected; however, P phase values increased over the course of the day, with higher values in the evening than in the morning.
To test if cells suffered from DIC limitation, samples were withdrawn from the experiment (Fig. 3) at 3 times (morning, noon, afternoon), exposed to actinic PF of approximately 140 µmol photons m −2 s −1 for 2 min, spiked with DIC to a final concentration of 300 µM, and ΔF/F m ' followed for 4 min by a Mini-PAM.ΔF/F m ' was not enhanced by DIC additions (see Fig. A2 in the Appendix).

pH perturbation: decreasing pH
Perturbation experiments were carried out with subsamples of the TLC, 2 d after the pH was in creased from ~7.2 to ~9.Acidification of the medium lowered the fluorescence signal rapidly (Fig. 4A).Both F ' (i.e.steady-state fluorescence under actinic light conditions) and F m ' (maximal fluorescence mea sured during a saturation pulse by the fluorometer) decreased within approximately 1 s to a lower state when the pH was dropped.Continuously recorded fluorescence signals indicate a recovery from acidification, as can be seen by a slow increase in F ' and F m ' until the pH was raised.The decrease in F m ' was caused by a rapid increase in NPQ.However, lower pH did not affect relative electron transport rates (rETR) effectively; values remained stable after HCl additions.
Acidification decreased gross oxygen evolution from 11.2 ± 1.74 to 3.7 ± 2.61 µM l −1 min −1 (data averaged from Figs. 4 & 5).This is surprising as fluorescence measurements suggest a steady continuation of rETR.Acidification therefore led to an uncoupling of oxygen and variable fluorescence based measurements of photosynthesis.
Supplementing suspensions with DIC (1 mM) did not increase rETR.Photosynthetic O 2 production, how ever, appeared to be stimulated in 1 sample when the DIC amendment was performed under acidic conditions, but not in another (Fig. 4B but not C).Cells of the different TLCs were tested regularly for DIC limitation where the photosynthetic performance remained unaffected by DIC additions at low or high growth pH (data not shown).Efficient quantum yields are indicative of the efficiency of photon usage for photosynthetic electron transport.1 − Vj represents the probability that an electron in Q A − moves further to electron carriers including the plastoquinone (PQ) pool.The higher the value (efficient F v '/F m ', or 1 − Vj), the higher the photosynthetic competence of the sample.Data show mean ± SD (n = 3; n = 2 for fluorescence induction curves shown in HL for t > 15:00 h; where n = 2 due to failure of aeration in 1 replicate)

pH perturbation: increasing pH
Increasing the pH rapidly increased F ' and F m ' to initial values or higher (Fig. 4).As a result, NPQ decreased to initial levels or lower.Negative NPQ values occurred when initial F m ' was lower than F m ' after the pH treatment.Because NPQ was calculated using the first F m ' value instead of F m , negative NPQ values can occur, and indicate a lower NPQ state compared to values at the start of the experimental protocol.Photosynthesis measured by rETR was not affected by the single-step pH increase.Photosynthetic oxygen production, however, was stimulated by addition of NaOH.Initial values were not quite restored, which could have been be due to high oxygen concentrations in the vials at the end of the measurements due to a buildup of O 2 by photosynthesis.

Photon flux effect at low and high pH
When the actinic PF was in creased 4-fold at low pH, rETR slowly increased during the entire period of elevated PF (Fig. 5).Decreasing F m ' values caused NPQ to increase until the pH was raised after 5 min.pH elevation caused a similar response under high PF conditions compared to low PF conditions (Fig. 4).PF elevation after cells were exposed to pH shifts did not induce a different response compared to PF elevation at low pH (Figs. 4 & 5).No pattern was found for oxygen production rates regarding PF.

Gradual pH amendments
When the pH was incrementally lowered, continuously recorded fluorescence and NPQ acclimation was visible (Fig. 6) in contrast to single-step pH perturbations, where very little acclimation was visible.Accli ma tion to pH increments was visible at pH > 7.0, where alkalization increased the fluorescence signal, decreased NPQ and elevated oxygen production with highest values between pH 7 and 8. Gradual increase of pH mirrored responses to decreasing pH.Clearly, F ' decreased with acidification, and increased with alkalization.

High time resolution measurements by FRRf
Fig. 7 shows variable fluorescence responses to pH perturbations in different PF levels.After a 3 min relaxation phase in very low light, the low actinic light (140 µmol photons m −2 s −1 ) was switched on and led to a temporary increase in F ' and F m ' followed by a drop and a quasi-steady state just before HCl was added (Fig. 7, dashed line), which decreased F ' and F m ' further.NPQ was activated when the actinic light was switched on and reached maximum values briefly after acidification.rETR, however, was only marginally disturbed, which corroborates PAM measure ments.The cells started to recover from acid addition within approximately 40 s, as shown by NPQ down-regulation.Increasing the actinic light intensity to 400 µmol photons m −2 s −1 (Fig. 7, light grey bars) perturbed the NPQ down-regulation only briefly (~1 min).rETR clearly increased under additional light, but was down-regulated upon addition of NaOH and consequent rise of the pH.This is in contrast to PAM measurements, where rETR remained stable upon an alkalization at high PF (Fig. 5).Acidification and alkalization appears to induce 2 phases, one imme diately after the pH adjustment for a duration of approximately 1 min.In the secondary phase, the parameters F', F m ' and NPQ developed in the opposite direction until the conditions were altered following the measurement protocol.When cells were transferred to lower actinic PF, the effective quantum yields increased, mainly due to a relaxation of NPQ as F ' increased only slightly.In a brief dark period, F ' decreased to initial values and F m ' responded marginally, suggesting that photoprotective mechanisms were not active at the last light phase of the experimental protocol.Photoinhibition was absent, as shown by F v /F m values which were not repressed by the experimental treatment (0.37 ± 0.08 and 0.37 ± 0.07 for the first and last quantum yield respectively).While acidification lowered Q A − re-oxidation kinetics (τ PSII ) only slightly, addition of NaOH increased τ PSII almost 2-fold (Fig. 7 C), which suggests a Q A − re-oxidation acceleration.The PSII functional cross section (σ PSII ) decreased when the actinic PF of 140 µmol photons m −2 s −1 was switched on due to a NPQ decrease.Acidification caused a slow rise in σ' PSII until initial values were restored.The increase of the PF from 140 to 400 µmol photons Gross oxygen evolution 7.7 µM l -1 min -1 (0.90) Gross oxygen evolution 12.3 µM l -1 min -1 (0.93)  (C) shows a repeat measurement from (B) for error estimation purpose.pH perturbations were achieved by addition of small volumes (≤300 µl) of acid (HCl) or base (NaOH).DIC (NaHCO 3 − solution) was added as indicated by arrows to yield final 1 mM.Note that continuous fluorescence lines, pH and oxygen concentrations do not read on any y-axis.pH was measured every 30 s m −2 s −1 only caused a minor dip in this 'recovery process'.Interestingly, addition of NaOH caused a slow decrease in σ' PSII .Note that the changes in σ' PSII are generally in the opposite direction to NPQ, as might be expected, but that the kinetics differ.

Light intensity stress
Chlorella spp.are known to grow well, even under challenging conditions.In the TLC used in the present study, the cells were exposed to full sunlight for some tens of seconds, then remained for some time in low light (Jerez et al. 2014).This light treatment challenges the photosynthetic apparatus of the cells and requires effective regulation of photosynthesis and photoprotection.Although diatoms perform better than green algae under fluctuating light (Wagner et al. 2006), chlorophyta convert photon energy efficiently to biomass (Wilhelm & Jakob 2011) and are successfully used for mass culture.To test the strain used for high light resistance, we exposed cells continuously to full strength sunlight for an entire day (≤1800 µmol photons m −2 s −1 , daily irradiance dose ~150 kW m −2 ).Cells were able to cope with continuous full strength ambient light conditions as shown by fast fluorescence induction curves.Quantum efficiencies and Q B reduction kinetics were lower in high light, but could recover in the afternoon.Frequently, P phase values were lower than J, which could be caused by a selective operation of the active fluorescence quenchers, or indicate photodamage in PSII.Photodamage impairs Q A − reduction kinetics and might lower the degree of Q B and PQ pool reduction capacities, potentially due to the proximity of Q B to the susceptible D1 protein (Aro et al. 1993, Jansen et al. 1999).However, the full photosynthetic poten-  where F m '* = the first F m ' value of the experimental protocol.F m was not measured as cells were not transferred to the dark prior to measurements.For legends and protocol explanations refer to Fig. 4 tial could be restored in lower PF in the late afternoon and early evening.Samples were dark-acclimated only very briefly (≤2 s), which will oxidize primary electron acceptors in PSII (Q A − ), but not secondary electron acceptors or the PQ pool (Strasser et al. 1995), or relax photoprotective NPQ.Although the rapid phase of energy dependent quenching (q E ) might relax due to a dissipation of the ΔpH-gradient within seconds, xantho-phyll cycle-mediated q E requires minutes, as do statetransitions (Müller et al. 2001, Ihnken et al. 2011a).Residual NPQ in illuminated samples can explain the obvious difference between initial samples, which were dark-acclimated overnight, and samples taken at low PF.Unexpected, however, was that even LL samples showed similar F v '/F m ' values compared to ML and HL samples in the morning, where the PF at the vessel surface in LL treatments was ≥200 µmol photons m −2 s −1 .In situ rapid light curves showed light saturation at approximately 500 µmol photons m −2 s −1 (J.C. Kromkamp pers.comm.), which indicates that LL cells were light-limited except for the highest PF at noon.Interesting, even LL samples appear to be photoinhibited around noon and needed to repair in the afternoon, as shown by a lag of F v '/F m ' up-regulation after the highest light values in the diurnal cycle.Theoretically, chlorophyta perform better under continuous PF (as provided in the PF experiment) compared to fluctuating light (in the TLC) (Wagner et al. 2006), but Chlorella cells can acclimate to various, and fluctuating, PF regimes without compromising growth (Kroon et al. 1992a,b), or be stimulated by fluctuating light compared to continuous light exposure (Wijanarko et al. 2007).However, it is possible that a change in the light exposure treatment af fected the photosynthetic performance in the present study.Cells experienced fluctuating light under growth conditions in the TLC and were then exposed to continuous PF for 1 d in the PF experiment.Photoacclimation may take hours to days, and is an energy-dependent process (Post et al. 1984, Wilhelm & Wild 1984, Havel kova-Dousova et al. 2004).It is possible that LL samples needed longer to adjust to the changed PF regime due to lower energy capture compared to ML and HL.However, LL conditions in pH drift experiments did not restrain pH drift, which shows that cells were still able to perform well, at least when the initial pH was high (Fig. 1B).
Generally, cells were resistant to severe photo-damage due to effective repair, or efficient and progressive photoprotection.These results show that cells were able to cope with in situ, and partly very high, PF without major damage.The very restricted pH drift at pH ~7 was therefore not due to impairment of the photosynthetic machinery.

pH or DIC effects?
pH drift experiments showed very low activity at approximately pH 7.2.The low photosynthetic activity at this pH is surprising, as Chlorella sp.showed highest growth rates at pH 7 or lower (Myers 1953, Goldman et al. 1982), and can raise the pH to values well over pH 10 (Myers 1953) due to the presence of efficient CCMs and bicarbonate uptake capacity (Shelp & Canvin 1980, Beardall 1981, Beardall & Raven 1981).We were not able to measure the DIC concentration, but suspected a DIC limitation when the pH was low.At an acidic pH, the total DIC concentration is much lower than at an alkaline pH.At pH 7.0, for instance, [DIC] = 74 µM, while at pH 9.0 [DIC] = 6300 µM when solutions are in air equilibrium (temperature = 25°C, salinity = 0), calculated using R; (R Development Core Team 2013) package seacarb 3.0 (Lavigne & Gattuso 2011).The in situ DIC concentrations would have been much higher because cell suspensions were maintained at pH ~7.2 by aeration with CO 2 gas.DIC addition experiments did not stimulate pH drift, nor rETR.The DIC additions were sufficiently substantial to enable photosynthesis for several hours if DIC would have been limiting (gross O 2 production ~10 µM l −1 min −1 ≈ CO 2 consumption; DIC additions 400, 800 mM).We therefore concluded that the DIC concentration was not restricting photosynthesis in the TLC, or in pH drift experiments.After the pH was raised by bicarbonate addition in the TLC, pH drift was stimulated.Final pH values were high and showed substantial DIC acquisition and photosynthetic capacity compared to other studies and species (Merrett et al. 1996, Choo et al. 2002, Ihnken et al. 2011b).An in crease in pH also elevated growth in a Chlorella mass-culture (Castrillo et al. 2013).Similar conclusions were made using marine algae in situ.The H + concentration, and not the DIC concentration, was the primary factor for cell regulation, growth and photosynthesis in moderate pH manipulation experiments (Lundholm et al. 2005, Hansen et al. 2007, Middelboe & Hansen 2007).Unfortunately, it is not clear what drives these effects due to the complexity of pH and inter-relationships with different cell functions (Smith & Raven 1979, Felle 2001).However, it can be concluded that impaired pH drift was neither due to DIC limitation, nor to PF effects in the present study.

pH-dependent acclimation patterns
O 2 evolution was repressed when the pH was decreased from alkaline conditions to pH 6.5.Cells grown at various pH levels show speciesdependent pH effects (Schneider et al. 2013).Cyanobacteria growth was depressed by acidification (Wang et al. 2011), while seagrass photosynthesis decreased with increasing pH (Invers et al. 1997).
Unfavorable pH conditions increased dark respiration and repressed photosynthetic activity in Euglena (Danilov & Eke lund 2001) and cyanobacteria, while F v /F m was not affected by pH (Liao et al. 2006).
It was expected that a pH increase would be less favored by Chlorella.Alkaline conditions increase the likelihood of cell damage due to increased concentrations of reactive oxygen species at high pH (Liu et al. 2007).At pH 7, internal pH values are similar to the external medium (Smith & Raven 1979, Beardall & Raven 1981, Gehl & Colman 1985, Sianoudis et al. 1987), which suggests an energetic advantage at this pH due to low costs for homeostasis.Proton regulation can be energetically expensive (Smith & Raven 1979, Briskin & Hanson 1992) and involve regulative ATPase activity (Weiss & Pick 1996).Active proton pumping might only be em ployed by the cell during short-term regulation (Beth mann & Schönknecht 2009), while steady-state homeostasis might be facilitated mainly by anion exchange (Lew 2010) in Chlorella fusca by a K + /H + antiport system (Tromballa 1987).
However, the substantial pH perturbations in the present study were carried out in a shock-mode, which is likely to require active H + pumping and entail ATP consumption (Bethmann & Schönknecht 2009).Because the intra-cellular pH could not be determined, it is not clear if (and how fast) homeostasis was reached.Photoautotrophic cells can regulate internal pH in time scales of minutes (Bethmann & Schönknecht 2009), but no information of rapid internal pH regulations (in seconds) was found in the literature for Chlorella.Limitations of internal pH maintenance requires extreme external pH (Gehl & Colman 1985), or osmotic stress (Goyal & Gimmler 1989) and induced internal pH changes of approximately 0.9 and 0.15 units respectively.F ', and F m ' values related positively with pH in the present study: the higher the pH, the higher the fluorescence signal.NPQ was activated by acidification.Similar results have been found for thylakoid fragments (Rees et al. 1992, Heinze & Dau 1996).In these studies, fluorescence responded in seconds, and a steady-state was reached within 1 min.These results can be explained by the quenching of the NPQ component q E , which was directly triggered by pH, irrespective of photosynthetic ETR.A lumen pH of 6 clearly activates q E (Kramer et al. 1999).Similarly, Jajoo et al. (2012) showed that pH changes can induce a thylakoid membrane re-organisation in pea thylakoid membranes.However, in the present study intact cells were used, and cell compartment membranes as well as pH regulatory mechanisms are expected to regulate internal pH effectively.If external pH perturbation could affect q E directly, the cells' pH regulation would be severely impaired.The exponential growth phase was passed in the TLC, which is very likely to have lowered the cell fitness.Cells in the stationary phase show distinctively different genetic activity, photophysiological characteristics (Humby et al. 2013) and are more susceptible to external stress factors (Randhawa et al. 2013).However, senescence effects are under-investigated (Humby et al. 2013).It appears likely that cells exploit alternative electron transport routes to increase energy dissipation to compensate for decreased photosynthetic energy quenching capacity.During ionic stress in higher plants, electron acceptance at the plastid terminal oxidase site (PTOX) accounted for up to 30% of the electron transport in PSII (Stepien & Johnson 2009).Because PTOX activity consumes oxygen while electron transport is facilitated, this mechanism could explain the discrepancy between rETR and gross oxygen evolution in the present study.Electrons are being transported and donated to molecular oxygen at the PTOX, with water as an end product (Kuntz 2004, Cardol et al. 2011).However, increasing, not decreasing, pH supported PTOX-mediated O 2 consumption (Josse et al. 2003), i.e. reaction kinetics operate in the opposite direction and are therefore not likely to explain low oxygen evolution rates in acidified samples in the present study.It is still un known if PTOX activity can generally act as a safety valve under stress conditions in higher plants, partly due to open questions regarding reduction kinetics in this pathway (Trouillard et al. 2012, Laureau et al. 2013).Nevertheless, it is possible that this mechanism is employed under pH shock conditions in Chlo rella as electron redox systems can participate in homeostasis (Houyoux et al. 2011).
A segregation between ETR and oxygenetic photosynthesis can be explained by state-transitions (Ihnken et al. 2014).Further studies must evaluate if state-transitions occur in Chlorella under conditions of pH stress and test the hypothesis that q T is the reason for changes in variable fluorescence readings and the seeming segregation of rETR and O 2 evolution.A deviation from ETR/O 2 photosynthesis linearity can also be explained by other mechanisms, such as cyclic electron flow around PSII. Cyclic electron transport in PSII suggests photosynthetic electron transport, but electrons are cycled within PSII and do not contribute to photosynthesis.Al ternatively, nitrate reduction, elevated oxygen consuming processes such as the Mehler reaction (Asada 2000, Heber 2002), photorespiraton/ chloro respiration (Beardall et al. 2003, Peterhansel & Maurino 2011) or mitochon drial respiration can cause deviation between fluorescence and oxygen based measurements.
A pH effect on the inter-PS-system redox chain was clearly shown in the present study.Q A − re-oxidation kinetics (τ PSII ; Fig. 7C) increased with alkalization.Faster Q A − re-oxidation, most likely due to a higher oxidation state of the PQ pool, was not facilitated by elevated electron transport, as rETR decreased when the base was added.This clearly shows that alternative electron routes are employed by the cells when the pH of the external medium is perturbed.However, rapid alkalization causes cells to aggregate temporarily (Castrillo et al. 2013), which can cause self-shading similar to the package-effect (Berner et al. 1989).It is therefore possible that high τ PSII values are artifacts caused by cell aggregation after addition of the alkaline solution.

DIC acquisition
Costs for DIC acquisition are expected to be lower at pH 6.5 due to the higher CO 2 :HCO 3 − ratio compared to pH 9.5.Chlorella possesses effective CCMs and can actively acquire both CO 2 and HCO 3 − (Shelp & Canvin 1980, Beardall 1981, Beardall & Raven 1981, Matsuda & Colman 1995a,b).Predominantly, bicarbonate CCMs are anion transporters which can operate as symport or antiport systems using Na + (Badger  et al. 2002) or Cl − (Young et al. 2001).Because anion regulation is also involved in homeostasis, HCO 3 − acquisition and consumption might be related to internal pH regulation.Cells acclimated to alkaline pH might combine bicarbonate acquisition and pH regulation, although uptake of HCO 3 − does not appear preferable as bicarbonate must dissociate to CO 2 (which is fixed in the Calvin-Benson-Bassham-cycle) and OH − (which hydrolyses to water under 'consumption' of H + ).Nevertheless, cells ex hibited higher photosynthetic rates at alkaline pH compared to measurements taken at pH 6.5.Compared to optimal growth conditions, RuBisCO contents are down-regulated when cells enter the stationary growth phase (Humby et al. 2013).A CCM up-regulation can account for low RuBisCO levels (Beardall et al. 1991).If CCM induction is controlled by pH in Chlorella, an alkalization might allow higher photosynthetic rates due to CCM up-regulation, or activation.

CONCLUSIONS
The present study shows 2 surprising pheno mena.Firstly, cells showed higher performance at elevated pH, a DIC concentration-and light in tensityindependent response.As this is in conflict with the majority of the literature values (where Chlorella spp.preferred pH ≤ 7), the precondi tioning of the cells (having passed the maximal growth phase) is likely to have influenced the outcome of the experiments.The second surprising finding is related to the deviation of oxygen and fluorescence based measurements of photosynthesis.Acidification lowered gross oxygen production, but not electron transport rates in PSII.The reason for this might be related to state-transitions which could not be detected due technical limitations in the present study.Nevertheless, alkalinisation clearly increased the physiological performance of cultures that are not operating at maximum efficiency.
Acknowledgements.The authors express their special gratitude to the local coordinators of the bio-technology workgroup, Irene Malpartida and Roberto Abdala.Special thanks to the Spanish GAP coordinators Félix L. Figueroa and Jesús Mercado, and the Spanish GAP committee María Segovia, Nathalie Korbee, Roberto Abdala, Rafael Conde, Francisca de la Coba, Andreas Reul and Irene Malpartida.

Fig. 1 .
Fig. 1. pH drift experiments with Chlorella fusca cells collected (A) before and after the pH was increased by addition of NaHCO 3− in the thin-layer cascade photobioreactor (TLC) (refer to 'Results' section for DIC-effects versus pH-effects).(B) Light intensity effect on pH-adjusted suspensions to full-strength ambient photon flux (HL), medium light (ML: 70% of ambient photon flux), and low light (LL: 30%).Cells were collected from the TLC in the morning, and exposed to natural sunlight throughout the day.Glass scintillation vials shielded cells from a high fraction of UV light.Vials in (A) were shaded to ML levels.Temperature was kept below 28°C.Data show means ± SD when plot symbol is extended (n = 3).Times are given in 24 h format

Fig. 3 .
Fig. 3. Fast fluorescence induction curves of Chlorella fusca cell suspensions continuously exposed to HL, ML and LL (see Fig.1) at different times of the day.A 250 ml cell suspension was withdrawn from the TLC, filled in 500 ml conical glass flasks, bubbled over night and exposed to measurement conditions on the following day.Samples were withdrawn from experimental conditions and dark-acclimated for ≤2 s before the measurement was performed.Data were normalized to O (initial fluorescence at t = 50 μs); vertical bars range over 0.5 units.Data show mean ± SD (n = 3; n = 2 for fluorescence induction curves shown in HL for t > 15:00 h where n = 2 due to failure of aeration in one replicate)

Fig. 4 .
Fig. 4. Chlorella fusca.Photosynthesis and photoprotection in response to rapid pH perturbations.From to top to bottom in each panel, data show gross O 2 evolution and correlation coefficient (in brackets), pH, continuous fluorescence recording, non-photochemical quenching (NPQ), relative electron transport rates (rETR) and actinic photon flux (grey shaded area; 140 µmol photons m −2 s −1 in the low photon flux, LL, phase; 450 µmol photons m −2 s −1 in the high PF, HL, phase) measured simultaneously.rETR and photoprotection NPQ were determined using variable fluorescence with rETR = ΔF/F m ' × photon flux, and NPQ = (F m '* − F m ') / F m ', where F m '* = the first F m ' value of the experimental protocol.F m was not measured as cells were not transferred to the dark prior to measurements.Data show responses to pH amendment at (A) constant photon flux, while in (B) and (C) the light intensity was increased at the end of the protocol.(C) shows a repeat measurement from (B) for error estimation purpose.pH perturbations were achieved by addition of small volumes (≤300 µl) of acid (HCl) or base (NaOH).DIC (NaHCO 3 − solution) was added as indicated by arrows to yield final 1 mM.Note that continuous fluorescence lines, pH and oxygen concentrations do not read on any y-axis.pH was measured every 30 s

Fig. 5 .
Fig. 5. Combined effects of pH perturbance and photon flux on photosynthesis and photoprotection in Chlorella fusca.Photon flux was (A) elevated when cells were exposed to acidic conditions and maintained while pH was brought back up again in a single mode, or (B) by 2 consecutive base additions to initial values.The low light relaxation phase was intermitted in (C).Relative electron transport rates (rETR) and photoprotection non-photochemical quenching (NPQ) were determined using variable fluorescence with rETR = ΔF/F m '* photon flux, and NPQ = (F m '* − F m ') / F m ', where F m '* = the first F m ' value of the experimental protocol.F m was not measured as cells were not transferred to the dark prior to measurements.For legends and protocol explanations refer to Fig.4

Fig. 6 .
Fig. 6.Chlorella fusca.Effects on photosynthesis and photoprotection by consecutive pH adjustments at constant photon flux.(A) and (B) show separate experiments performed on the same day.A DIC addition was performed at the end of (B) (arrow) with a final 0.7 µM

Fig. 7 .
Fig. 7. Chlorella fusca.Fluorescence parameters measured by FRRf in response to manipulated pH and actinic light conditions.High initial pH (9.3) was lowered by addition of HCL (dashed line) to a theoretical pH 6.5.After 5 min, the cell solution was alkalized by injection of NaOH (dotted line) to a theoretical pH 8.5.Grey scale bars indicate actinic photon flux: 3, 140, 400 and 0 µmol photons m −2 s −1 .Data in (A) show minimal fluorescence in darkness or actinic light (F'), maximal fluorescence during a single turnover flashlet application (F m ') and effective (or maximal) quantum yields; (B) shows relative electron transport rates (rETR) and non-photochemical quenching (NPQ).The iterative curve fit during the single turnover fluorescence emission allows determination of the functional absorption cross-section of PSII (σ PSII ); (C) Re-oxidation kinetics of Q A − after the single turnover are shown by τ PSII .Data show mean ± SD for n = 3 Experiments were carried out under support from the Department of Ecology of the Universidad de Málaga.The Netherlands Institute for Sea Research (NIOZ) provided acknowledged travel funds for J.C.K. and S.I.Three anonymous reviewers contributed through constructive criticism.LITERATURE CITED Andersen RA, Berges JA, Harrison PF, Watanabe MM (2005) Recipes for freshwater and seawater media.In: Andersen RA (ed) Algal culturing techniques.Elsevier Academic Press, London, p 429−538 Aro EM, McCaffery S, Anderson JM (1993) Photoinhibition and D1 protein degradation in peas acclimated to different growth irradiances.Plant Physiol 103: 835−843 Asada K (2000) The water-water cycle as alternative photon and electron sinks.Philos Trans R Soc Lond B Biol Sci 355: 1419−1431 Badger MR, Hanson D, Price GD (2002) Evolution and diversity of CO 2 concentrating mechanisms in cyanobacteria.Funct Plant Biol 29: 161−173 Beardall J (1981) CO 2 accumulation by Chlorella saccharophila (Chlorophyceae) at low external pH: evidence for active transport of inorganic carbon at the chloroplast envelope.J Phycol 17: 371−373 Beardall J, Raven JA (1981) Transport of inorganic carbon and the 'CO 2 concentrating mechanism' in Chlorella emer sonii (Chlorophyceae).J Phycol 17: 134−141 Beardall J, Roberts S, Millhouse J (1991) Effects of nitrogen limitation on uptake of inorganic carbon and specific activity of Ribulose-1, 5-bisphosphate carboxylase oxygenase in green microalgae.Can J Bot 69: 1146−1150 Fig. A1.pH drift experiments.(A) Cells were supplemented with NaHCO 3− ; (B) shows a scatter plot of (A) for samples without DIC amendment and samples where an 800 mM bicarbonate supplement was performed.(C) Shows additional pH drift experiments from TLC2 and TLC3, where cell suspensions were bubbled with CO 2 prior to pH drift experiments, and samples that were diluted with fresh, nutrient replete medium (50:50 by volume).Data are mean ± SD (n = 3)