Validation of rapid algal bioassay using delayed fluorescence in an interlaboratory ring study

Science of the Total Environment 605–606 (2017) 842–851

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Validation of rapid algal bioassay using delayed fluorescence in an interlaboratory ring study Masakazu Katsumata a,⁎, Yuko Ikushima a, Keith Bennett a, Yukiko Sato a, Ayano Takeuchi a, Norihisa Tatarazako b, Tomoyuki Hakamata a a b

Central Research Laboratory, Hamamatsu Photonics K.K., Japan National Institute for Environmental Studies, Japan

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• Rapid algal bioassay using delayed fluorescence (DF) from green alga • The DF has potential to be used as a surrogate for the measurement of algal biomass. • The method is statistically characterized in an interlaboratory ring study using reference toxicant, 3,5-dichlorophenol. • EC50–values of the test after 24 h are close enough to the conventional test to be useful for screening tests. • The intralab and interlab variabilities of EC50 at 24 h are 12% and 28%.

a r t i c l e

i n f o

Article history: Received 17 November 2016 Received in revised form 26 May 2017 Accepted 26 June 2017 Available online xxxx Editor: D. Barcelo Keywords: Chemical toxicity test Growth inhibition Alga Delayed fluorescence Surrogate of biomass

a b s t r a c t Algal growth inhibition tests are generally used to determine the toxic effects of chemical substances on algae growth. In this report, we describe a rapid and simple test procedure using delayed fluorescence (DF) to determine chemical toxicities more rapidly than the conventional 72 h or 96 h growth inhibition tests. We assess the suitability of DF to serve as an alternative endpoint for biomass production and determine the variability by an interlaboratory ring study using a typical reference toxicant 3,5-dichlorophenol (DCP). The results suggest that DF has the potential to be used as a surrogate measure of photosynthetically-active biomass in the algal growth inhibition tests. The half maximal effective concentration (EC50) values of DCP determined from the DF inhibition test in 6 h and 24 h (1.2 ± 0.3 mg/L and 2.7 ± 0.5 mg/L respectively) are in reasonable agreement with the EC50 value of DCP determined by the 72 h conventional method (1.8 mg/L). In the interlaboratory ring study, the intralaboratory and interlaboratory variabilities of the EC50 of the DF inhibition test for a 24 h exposure period are 12% and 28% respectively. DF intensity can be considered as a surrogate of living biomass with active photosynthesis, and we conclude that a 24 h exposure duration better estimates the toxic effects measured using conventional surrogate measures for dry weight such as cell counts, volume, optical density or fluorescence. © 2016 Elsevier B.V. All rights reserved.

Abbreviations: ANOVA, analysis of variance; CV, coefficient of variation; DCP, 3,5-dichlorophenol; DF, delayed fluorescence; DFI, delayed fluorescence intensity; DW, dry weight; ECX, X% effective concentration; GR, growth rate; ISO, International Organization for Standardization; NOEC, non-observed effect concentration; OECD, Organization for Economic Co-operation and Development; SETAC, Society of Environmental Toxicology and Chemistry; USEPA, United States Environmental Protection Agency. ⁎ Corresponding author at: 5000, Hirakuchi, Hamakita, Hamamatsu, Shizuoka Prefecture, Japan. E-mail address: [email protected] (M. Katsumata).

http://dx.doi.org/10.1016/j.scitotenv.2017.06.228 0048-9697/© 2016 Elsevier B.V. All rights reserved.

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

1. Introduction Algal growth inhibition tests are commonly applied for assessing toxic chemical effects on algae, a primary producer in aquatic ecosystems. Usually, growth is measured for 72 h (OECD, 2011; ISO, 2004) or 96 h (USEPA, 2012) using standard test guidelines, such as the OECD test guideline 201 (TG201) which describes the principle and criteria of the test (OECD, 2011). Growth and growth inhibition are quantified by the measurement of dry weight algal biomass as a function of time (i.e. the growth rate). Because of the difficulties measuring “dry weight” accurately, surrogate measures for the biomass such as cell counts and/ or volume, fluorescence, or optical density are described in various guidelines. TG201 requires a known conversion factor between the measured surrogate parameter and biomass. Flow cytometry was reported to be useful in the microalgae test to accurately estimate biomass dry weight (Chioccioli et al., 2014) and to analyze mixed algal populations in multispecies assays (Franklin et al., 2004). Nagai et al. (2011) proposed the utility of flow cytometry for distinguishing living from dead cells in algal growth inhibition studies. The TG201 and other standard guidelines, however, are laborintensive, time-consuming, and require tight control of algal quality. Therefore, there is a need to improve the throughput of the test, specifically by developing a rapid screening test to support the conventional 72 h growth inhibition test. The ISO test guideline (ISO, 2004) describes a shorter test protocol in its annex A; rapid screening of wastewater algal growth inhibition with a minimum test duration of 48 h. There are also several suggestions for shorter exposure tests. Effective fluorescence quantum yield ([F′m − F] / F′m), defined below, is most commonly used for toxicological investigation to evaluate photosynthesis inhibition that may result in growth inhibition (reviewed by Ralph et al. (2007)). F′m is the maximum fluorescence (with strong pulse light excitation in addition to the standard illumination for photosynthesis) and F is fluorescence from standard photosynthesis illumination. The effective quantum yield provides an indication of the photosystem activity. However, the effective quantum yield ([F′m − F] / F′m) evaluates the ratio of two chlorophyll fluorescence signals from the same sample under different light conditions. Therefore, algal cell suspensions that have a specific photosystem activity (and correspondingly, a specific ratio between F′m and F) exhibit similar effective quantum yield at different cell densities. As a result, interpretation of effective quantum yield may be complicated for the evaluation of algal suspensions with varying cell densities (i.e. growth tests) (Katsumata et al., 2012a). Delayed fluorescence (DF) has been proposed as an endpoint for evaluation of the influence of chemical substances on the growth of alga and other photosynthetic organisms (e.g. duckweed; Drinovec et al., 2004). DF can be detected for up to several minutes after light excitation in the dark (Strehler and Arnold, 1951). The source of the light

843

in DF is charge recombination at a photosystem reaction center via trapping of electrons in electron acceptors during the photosynthetic chain reactions (Lavorel, 1975; Jursinic, 1986; Schmidt and Senger, 1987; Goltsev, 2009). DF is sensitive to exposure to chemical substances that disturb photosynthetic reactions (e.g. photosynthesis inhibitors, respiration inhibitors). DF represents the total amount of active photosystem (capable of trapping electrons and undergoing recombination) within the sample (e.g. algal cell suspension). Briantais et al. (1980) and Joliot and Joliot (1980) reported that the light induced proton gradient across thylakoid membranes is related to DF, and furthermore the typical inhibition of DF is caused by dissipating the proton gradient by an uncoupler. Several studies have demonstrated applicability of DF for the evaluation of toxic effects of chemicals (e.g. metals, herbicides, and photosynthetic inhibitors) on alga (Bürger and Schmidt, 1988; Scordino et al., 1996, 2008; Katsumata et al., 2008). Additional studies compare DF with measures for biomass such as cell count/volume, absorbance, and fluorescence. Drinovec et al. (2004) demonstrated DF at 24 h is more sensitive for detection of toxicity than the 72 h growth inhibition on duckweeds (Lemna minor). We reported the intensity of DF at relatively long delay times (e.g. later than 0.6 s) following short term exposure (15 min or 24 h) is a promising metric to estimate the toxicity of chemicals (herbicides and uncoupler) measured by the 72 h conventional growth test of green alga Pseudokirchneriella subcapitata (Katsumata et al., 2006, 2009). We concluded DF at 24 h is a possible endpoint to estimate the 50% effective concentration (EC50). We also demonstrated the correlation between DF and living cell density in mixtures of living and dead cells of P. subcapitata (Katsumata et al., 2010). Consequently, DF can probe photosynthetic activity for both quantity (living cells in a mixture) and quality (active photosystems per detectable cell particle), suggesting that DF may be a useful measure of algal productivity. There are numerous studies comparing DF with other surrogate measures for biomass on several algae species. These results demonstrated DF to be a suitable and sensitive method for algal toxicity testing independent of cell size, taxonomic group, and pigment composition (Zrimec et al., 2007; Leunert et al., 2013; Yamagishi et al., 2016; Breuer et al., 2016). Therefore, DF is a promising endpoint to estimate algal growth with shorter exposure durations than conventional growth tests (i.e. provides rapid evaluation). Conventionally, growth tests require preparation of test alga with a pre-incubation period of 2–3 days, which is inconvenient at best, and subject to variability leading to reduced test quality. Consequently, we developed a frozen algal suspension (cryopreservation) for preparation of test alga on the day of the test, and reported preliminary results from a ring study utilizing the frozen alga in a rapid and simple test procedure (Katsumata et al., 2012b). The analysis mainly focused on the variation of the DF and its dose-response. We have also demonstrated the use of DF to evaluate the toxicity of

Fig. 1. Preparation and exposure using reagent-algae. Reagent-algae are thawed, diluted 10-fold and pre-incubated for at 1 to 2 h. The pre-incubated algae are used to prepare control and exposure samples. The samples were incubated for up to 24 h.

844

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

industrial effluents (mixtures of unknown chemical substances), rather than a specific chemical substance (Takeuchi et al., 2014). The reported non-observed effect concentrations (NOECs) of eight industrial effluents determined by the conventional 72 h growth test using cell counting and by a rapid test using DF at 24 h provided substantially similar results (i.e. differences in the NOEC were mostly within one level of exposure concentration). Subsequently, the measurement results in the ring study (Katsumata et al., 2012b) were re-analyzed utilizing the analysis method described in detail in the effluent study (Takeuchi et al., 2014) to determine X% effective concentration (ECX) and NOEC from measurement of the DF. In this report, we characterize the properties of DF as an alternative endpoint for biomass production and determine the variability of the test method at different ECX-levels (X% effective concentrations) in an interlaboratory ring study using a typical reference toxicant 3,5dichlorophenol. 2. Materials and methods 2.1. Preparation of the test alga In the DF inhibition test, we used green algae (Pseudokirchneriella subcapitata, strain NIES-35), prepared as a reagent (hereafter, reagentalgae) to control the algal quality and quantity in the ring study. The algae for preparation of the reagent-algae were cultured in 100 mL of OECD medium in a 300 mL Erlenmeyer flask (white fluorescent light 50 μmol m−2 s−1, 24 ± 1 °C, orbital shaking). The initial cell density was 1 × 104 cells/mL. The algae were harvested after 72 h, at a cell density of 180–200 × 104 cells/mL. The average growth rate was 1.7– 1.8 day− 1, and the average cell volume was 27 fL. The reagent-algae were prepared in the following manner: an algal cell suspension was concentrated to 2000 × 104 cells/mL after harvest, then divided into 600 μL aliquots with 5% of dimethyl sulfoxide (DMSO) as cryoprotectant and then cryopreserved by deep freezing (− 80 °C), after which, the algae could be stably stored for at least 5 months at − 80 °C. The reagent-algae were transported on dry ice to the laboratories that participated in the ring study. A sample preparation procedure using the reagent-algae is shown in Fig. 1. The aliquot of algae was thawed in a warming block (37 °C, but the heater was turned off when the temperature of the cell suspension was about 0 °C) and diluted 10-fold with OECD medium to a volume of 6 mL and cell density 200 × 104 cells/mL in a glass tube (φ25 mm × 80 mm). Diluted algae were then preincubated for at least one hour in a light irradiated incubator (white fluorescent light 50 μmol m−2 s, 24 ± 1 °C) on an orbital shaking incubator specially made for the glass tube (prototype tube shaker, Hamamatsu Photonics K.K., Japan). The test system is applicable to both reagent-alga and fresh alga prepared by normal pre-incubation (e.g. 72 h). 2.2. Exposure to 3,5-dichlorophenol The pre-incubated algae were used to prepare exposure samples in a series of five concentrations, i.e. 0.1, 0.3, 1, 3, 9 mg/L, of 3,5-dichlorophenol (DCP), a major reference substance for green alga toxicity testing in the OECD TG201 standard medium (OECD, 2011) and a control sample without DCP. Both the undiluted toxicant solution DCP (100 mg/L in OECD medium) and the OECD medium (pH 8.0) used by the participating laboratories were provided from a single supply laboratory (Hamamatsu Photonics K.K.). DCP uncouples the electron transport chain reaction from the oxidative phosphorylation (photosynthetic and mitochondrial). The uncoupler migrates through the membrane, dissipating the proton gradient necessary for ATP synthesis (Escher and Schwarzenbach, 2002). The acid dissociation constant (pKa) of DCP has been reported as 8.1 (Serjeant and Dempsey, 1979). The initial concentration of algae in the test tube was 20 × 104 cells/mL. The bioavailability of DCP can be influenced by cell density. The initial cell density in

Table 1 Factors and levels in the ring study. Factors

Abbreviation

Levels

Description

Laboratory Duration of time ECX Repetition

lab time X –

7 2 3 3

L1–L7 6 h and 24 h EC50, EC25, and EC10 Individual preparation

the DF-test (20 × 104 cells/mL) was 20 to 40 times higher than the initial cell density in the conventional test (0.5–1 × 104 cells/mL). However, the typical growth rate of the reagent-alga was 1.6 day−1 in the supplying laboratory. Therefore, the expected final cell density of the control sample at 24 h was 99 × 104 cells/mL. This is similar to the final cell density in conventional tests, for example a test started with initial cell density 1 × 104 cells/mL will achieve 122 × 104 cells/mL at 72 h with growth rate of 1.6 day−1. 2.3. Measurement of algal cell density and delayed fluorescence Algal cell density was determined with a particle counter covering a detection range of 3 to 12 μm (CDA-1000, Sysmex, Japan). The DF signal from the algal suspension was measured (Takeuchi et al., 2014) by a high sensitivity luminometer (Type-7100, Hamamatsu Photonics K.K., Japan). The alga suspension was left in the dark for 60 s and was then illuminated for 30 s with white LED light (NSPW500CS, Nichia, Japan; 500 μmol m−2 s−1, Photosynthetically Active Radiation measured by quantum light meter (LI-250, LI-COR, USA)), then kept in the dark for 5 s, and finally illuminated for 1 s with the excitation light, 700 nm LED light (L700-04AU, Epitex, Japan; 50 μmol m−2 s−1). The value of the relative DF intensity (wavelength range of 640–750 nm) was recorded at 0.1 s intervals from 1 to 60 s after the excitation light was turned off. The integrated value of the DF intensity (DFI) signal from 1 to 60 s was recorded after subtracting the predetermined background dark signal. Since the DFI showed good correlation with the living algal cell content in a mixture of living and heat-treated inactivated algae, we proposed the use of DFI as a surrogate measure for algal biomass possessing growth potential (Katsumata et al., 2009; Katsumata et al., 2010). We assumed that the DFI is related to the total quantity of active photosystem in the sample. Therefore the DFI is not strictly proportional to the cell density of the algal cell if the cell volume varies during a cell cycle (i.e. the photosystem quantity in a cell varies). We convert the DFI to an equivalent cell density assuming a standard cell volume (the average cell volume is approximately 27 fL) during our preparation process. The luminometer was calibrated daily to ensure the DFI correlated with the standard cell density of the living test alga (cells/mL). 2.4. Evaluation of growth inhibition The growth inhibition tests were carried out with 3 independent replicates in each of the seven (7) laboratories. The control samples were incubated for 15 min in the incubator, at which time the initial DF signals (DFI0) from the control samples were measured. The control and exposure samples were incubated for 24 h after initial value measurements. DF signals from each sample, DFIH (H = 6 and 24), were measured at 6 h and 24 h after initial measurement. The average specific growth rate for each sample between the initial and final value measurement (i.e. day− 1) was calculated from Eq. (1), where t is time (days), i.e. H/24 (H is symbol for the incubation time in hours). Percent inhibitions (%I) of the growth rate for each exposure sample were calculated from Eq. (2), where GRC and GRE are the growth rates of control and exposure sample, respectively.
 Lnð DFI Þ−Lnð DFI Þ H −1 0 Growth rate day ¼ t

ð1Þ

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

845

individual fitting results, with predetermined inhibition percentage X (X = 50, 25, and 10). NOEC was determined with Dunnett's multiple comparison test (two-tailed test, α = 0.05). The logistic model fitting and Dunnett's test were verified using data analysis software (Origin 2016, OriginLab corp.) %IðCÞ ¼ D þ

M−D  N 1 þ KC

ð3Þ

K ECX ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi   M−D N −1 X−D

ð4Þ

2.7. Variabilities in the inter-laboratory ring study

Fig. 2. Correlation between DFI and dry weight. Correlation between DFI and dry weight of triplicates for each cell density. A line shows the fitted linear regression through the origin. The values of dry weight and DFI indicate as per liter.

%I ¼

GRC −GRE  100 GRC

ð2Þ

2.5. Analysis of DF as surrogate parameter for biomass The prepared test alga, after pre-incubation, were diluted to three cell densities, 100, 50 and 25 × 104 cells/mL and divided into 60 mL volumes with triplicates (growing alga). Another sample of 60 mL was prepared at a cell density of 50 × 104 cells/mL, heated to 50 °C for 20 min to inactivate the growth potential (inactivated alga). As we previously reported, although the inactivated alga can be detected as cell particles, it does not measurably grow for at least 24 h (Katsumata et al., 2009). 10 mL of the 60 mL algal cell suspension was used for DF measurement, and the 50 mL remainder was used for the dry weight measurement. The 50 mL algal cell suspensions were filtered with a filter membrane of 0.45 μm pore size (MF-Millipore HAWP02500, Merck Millipore) using a syringe filter housing (Swinnex SX0002500, Merck Millipore). Triplicates of 50 mL diluent (growth medium) were filtered in the same manner; these membranes are called as blank membranes. The membranes with filtered algal cells (and blanks) were dried overnight at 40 °C in a constant temperature oven. The dry weights of algal cells on the membranes were determined by subtracting the weights of the blank membranes (the average of triplicate was 26.63 mg, CV 0.33%) from the weights of the membranes including the algal cells.

The interlaboratory ring study evaluated different factors and their associated contributions to the precision of the estimation of the three parameters shown in Table 1: initial value of DFI (DFI0), growth rate of control sample (Growth rate), and X% effective concentration (ECX). The effects of specific factors were estimated using analysis of variance (ANOVA). In an effort to determine whether the variability within a lab is comparable to the variability between labs, we separately estimated the errors within laboratories, the laboratory factor (∂2lab) and total error (∂2e ) using Eq. (5) where Vlab is the variance of an laboratory factor (each laboratory's sum of squares divided by degrees of freedom of the laboratory factor); Ve is the error variance (error sum of squares (i.e. “total sum of squares” minus “explained sums of squares”) divided by the degrees of freedom of the error factor); and n is size of a laboratory's subgroup (the number of samples divided by the number of laboratories). Intralaboratory (hereinafter referred to as “intralab”) variability (repeatability), and interlaboratory (hereinafter referred to as “interlab”) variability (reproducibility), were estimated using Eqs. (6) and (7) where ∂2e is the error variance. Since the units and factors of the three analysis parameters were different, each parameter was analyzed individually. ANOVA was computed using statistical analysis software (JUSE-StatWorks V5, Institute of Japanese Union of Scientists and Engineers). 2

∂lab ¼

V lab −V e n

Interab variability ð%Þ ¼ 100 

Intralab variabilityð%Þ ¼ 100 

ð5Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2 ∂lab þ∂e Average qffiffiffiffiffi 2 ∂e Average

ð6Þ

ð7Þ

3. Results 2.6. Inhibition analysis 3.1. Analysis of DF as surrogate parameter for biomass Effective concentrations 10%, 25% and 50% (EC10, EC25 and EC50) and non-observed effect concentration (NOEC) were used to characterize the influence of DCP exposure. ECX were determined with regression using a four parameter logistic model used in pharmacology (Barlow and Blake, 1989), specifically the four parameter Hill equation (Eq. (3)). Eq. (3) can be fitted to the %I obtained from Eq. (2), i.e. as a function of the exposure concentration (C) with three parameters (M, K, N, and D). M is the maximum effect of %I. C and K respectively correspond to concentrations of a drug (i.e. DCP) and the concentration that produces half the effect, while N determines the slope of the response curve. D is the minimum value of the dose-response (which may be a negative value). Eq. (4) is obtained by solving Eq. (3) for C, which can be used to calculate X% effective concentration (ECX) using M, K, and N in

Fig. 2 shows the correlation between DFI and dry weight for each cell density. As shown in the figure, DFI and dry weight of living alga are correlated with statistical significance (R = 0.95, p b 0.01). The conversion factor between the measured DFI (counts) and dry weight (mg) was determined to be 21.1 × 10−8 counts/mg. Since both values of DFI and dry weight (DW) were obtained after subtracting the background value, the data starts from the origin (DFI = 0 and DW = 0). Therefore, the conversion factor is the coefficient of the slope of the best fit linear regression line through the origin. According to the measured cell density and dry weight in Fig. 2, the average dry weight of a single cell is 2.44 × 10−8 mg/cell, which is in the range of typical values of P. subcapitata, of 2–3 × 10−8 mg/cell (OECD, 2011). Since the reported DFI is correlated with

846

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

Fig. 3. Comparison of dry weight and DFI on growing algae and inactivated algae. a) Comparison for dry weight per single cell. b) Comparison for delayed fluorescence intensity (DFI) per single cell. Columns show average of triplicates. The error bars are the standard deviations; p-value is obtained by the two-tailed t-test. Cell density of the measured algal cell suspension is 55.7 × 104 cells/mL.

living algae (Katsumata et al., 2010), Fig. 3 shows a comparison of dry weight and DFI between the growing algae and the inactivated algae at a cell density 100 × 104 cells/mL. Interestingly, the value of dry weight is not significantly different (p = 0.420) between growing alga and inactivated alga (Fig. 3a), however the DFI of the inactivated alga is significantly (p b 0.001) decreased to 7.8% of the DFI of growing algae (Fig. 3b). 3.2. Results of interlaboratory ring study 3.2.1. Average of growth curve Fig. 4a shows growth curves calculated from the average of the DFI (i.e. a biomass surrogate) obtained from 7 laboratories and 3 replications for each of several DCP exposure concentrations at 3 time points. Fig. 4b shows the growth rate of DFI (calculated from Eq. (1)) decreases with increasing exposure concentration. The effect of DCP on the growth rate during the time period 0 to 6 h (6 h) is stronger than during the time period 0 to 24 h (24 h). Especially, 3 mg/L and 9 mg/L at 6 h, and 9 mg/L at 24 h indicate negative growth rates. Since we have proposed that DFI (related to the active photosystem) is correlated with the dry weight of growing alga (Figs. 2 and 3), the observations suggest

the amount of active photosystem cells at the respective times is lower than the initial value of DFI (DFI0), possibly due to the effects of cell shrinking, bleaching, or cell lysis of the dead algae. 3.2.2. Variation of initial value of DFI Since the initial value of DFI (DFI0) is the starting point for the growth rates, variability in DFI0 is an important quality control for the test. Fig. 5 shows variability of DFI0 of control samples (i.e. exposure concentration 0 mg/L). The ANOVA table of DFI0 is given in Table 2. DFI0 is significantly different between laboratories (p b 0.001). As mentioned in the Materials and Methods section, the luminometer was calibrated daily so DFI correlated well with cell density (cells/mL). The total average (17.2 × 104 counts) is consistent with the expected initial cell density in the protocol (20 × 104 cells/mL). Intralab variability and interlab variability were 4% and 16% respectively. 3.2.3. Variation of growth rate of control sample In the ring study, each laboratory used the same model of tube shaker (specially designed for the test). However, the tube shaker was set up in light irradiated incubators previously prepared by each laboratory. To determine the effects differences between incubators, we measured the

Fig. 4. Average growth curve and growth rate by the DFI. a) Growth curve with delayed fluorescence intensity (DFI) as a surrogate of biomass. Each plot shows average of 7 laboratories with triplicates. Plots and lines indicate average of triplicates. b) Comparison of growth rate between control and exposure sample. Columns show the average of 7 laboratories with triplicates, error bar is standard error; p-value is obtained by two-tailed Dunnett's multiple comparison test, * is p b 0.05 in Dunnett's test.

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

Fig. 5. Variability of initial delayed fluorescence value (DFI0) between laboratories. Variation of initial value of DFI (DFI0) of control sample (i.e. exposure concentration 0 mg/L) from the 7 laboratories and 3 replications. Circles indicate each value of each replicate. Horizontal line indicates total average (17.2 × 104 counts).

growth rates of the control samples in each laboratory. Fig. 6 shows the growth rates at 6 h and 24 h. The variability of the growth rate for entire analysis across labs and time, and for specific times (6 h or 24 h) is presented in Table 3. As shown in Fig. 6 and Table 3, the growth rates are significantly different depending upon the laboratory (p b 0.001), as well as the interaction of laboratory factor and time period of exposure factor (time). The growth rate does not show dependence on the exposure time. The average growth rates for both time periods individually (6 h and 24 h) or together, are 1.59 ± 0.50 day−1, 1.54 ± 0.16 day−1, and 1.57 ± 0.25 day−1 respectively. This is in the range of typical values for P. subcapitata, 1.5–1.7 day−1 (OECD, 2011). Intralab and interlab variabilities in the total model (6 h and 24 h) were 15% and 31% respectively. Comparison of the variability between the 6 h and 24 h models indicates the variability of 24 h model (intralab 7%, interlab 15%) is better than the 6 h model (intralab 20%, interlab 46%). As the data is logtransformed, the variability of the 6 h model is rather large.

3.2.4. Variation of X% effective concentrations (ECX) Fig. 7 shows the variation of X% effective concentrations (ECX, X = 50, 25, and 10) at 6 h and 24 h time periods. The ANOVA table of ECX for entire analysis and for separate analyses for each time point and effective concentrations X are shown in Table 4. The average of EC50 ± 95% confidence limit (based on individual laboratory mean-values) at 6 h and 24 h is 1.2 ± 0.3 mg/L and 2.7 ± 0.5 mg/L respectively. EC50 of DCP in a conventional 72 hour alga growth inhibition test by cell counting was reported as 1.8 mg/L (Comber et al., 1995). Paixão et al. (2008) also reported EC50 of 1.79 mg/L of DCP in the 72 h test. As shown in Table 4, the differences of ECX between laboratories are significant (p b 0.001) in both the entire analysis and separate analyses. The differences between laboratories may be influenced in a complex manner by the bioavailability of DCP which is related to cell density, absorption, pH of medium, pKa of DCP, light intensity/quality, and other factors. Intralab and interlab variabilities on the entire analysis were 15% and 35%. In separate analyses, the variance in interlab variabilities decreased with larger value of X (i.e. EC10 is most inaccurate). The intralab variabilities ranged between 12% and 17%, while the interlab variabilities ranged between 28% and 43%.

847

Fig. 6. Variation of growth rate in control samples at 6 h and 24 h time periods. Variation of the growth rate in control samples at 6 h and 24 h from the 7 laboratories and 3 replications. Circles indicate growth rate in 6 h and 24 exposure time periods. Horizontal line indicates total average (1.57 day−1).

3.2.5. Variation of NOEC and EC10 Fig. 8 compares the NOEC and EC10 for 6 h and 24 h exposure time periods. For 6 h, the mode of NOEC was 0.3 mg/L (4 of 7 laboratories), 1.0 mg/L (2 of 7) or 0.1 mg/L (1 of 7). For 24 h, the mode of NOEC was 1.0 mg/L (6 of 7 laboratories) and 0.3 mg/L. This observation indicates differences in the NOEC between laboratories are within one level of exposure concentration from the mode of NOEC. The NOEC and EC10 are relatively close within each laboratory.

4. Discussion In Figs. 2 and 3 we characterize the properties of DF in order to assess its suitability as a surrogate measure of biomass. The results indicate high correlation (R = 0.95) between DFI and the dry weight of alga (i.e. biomass as defined in TG201). We previously reported that the DFI is highly correlated with the cell density of growing alga (Katsumata et al., 2010). The purpose of TG201 is to determine the effects of substances on the growth of alga. Growth inhibition is quantified by measurement of the algal biomass as a function of time (i.e. the growth rate). Strictly speaking, according to this principle, only the biomass (dry weight) with growth potential should be measured. As indicated in Fig. 3, however, the conventional dry weight cannot distinguish between growing algae and inactivated algae (i.e. between living and dead cells). Since detecting cell particles at a single moment in time is insufficient to identify growth potential, it is necessary to evaluate growth rate over a long time period (e.g. 72 h) to evaluate algal growth. Critically, Fig. 3 shows a strong correlation of DFI with growing alga, and excludes dead alga. This may be an advantage for assessing chemical toxicity compared to the conventional surrogates of biomass (cell counting and dry weight measurement) which cannot distinguish between growing and inactivated algae. The results in Fig. 3 suggest the DFI has potential to be used as a simple way to measure the quantity of growing alga, similar to distinguishing living/dead cells demonstrated by flow cytometry (Nagai et al., 2011. A major difference between flow cytometry and DF is single cell identification. Flow cytometry detects individual cells, and then distinguishes between living and dead cells Therefore, the quantity of living/dead cells reported by the flow cytometry method represents the total quantity of cells discriminated

Table 2 ANOVA table for DFI0. DFI0 (counts) average ± 95% confidence limit

Factor

d.f.

Variance

F

p-Value

σ2

Variability (CV%) Within-lab

Between-lab

172,300 ± 20,187

lab error

6 14

2.2E+09 4.8E+07

46.6 –

b0.001 –

7.3E+08 4.8E+07

4

16

d.f., degrees of freedom; F, variance ratio; σ2, estimating component of variance. CV%, coefficient of variation (percentage of the standard deviation to the mean). 95% confidence limits are calculated based on the individual laboratory mean-values.

848

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

Table 3 ANOVA table for growth rate. Growth rate (day−1) average ± 95% confidence limit

Factor

d.f.

Variance

F

p-Value

σ2

Variability (CV%) Within-lab

Between-lab

Entire analysis

lab time lab ∗ time error lab error lab error

6 1 6 28 6 14 6 14

1.105 0.025 0.407 0.059 1.367 0.105 0.145 0.012

18.8 0.4 6.9 – 13.0

b0.001 0.522 b0.001 – b0.001 – b0.001 –

0.174 – – 0.059 0.421 0.105 0.044 0.012

15

31

20

46

7

15

Separate in time

1.57 ± 0.25

6h

1.59 ± 0.50

24 h

1.54 ± 0.16

11.8 –

d.f., degrees of freedom; F, variance ratio; σ2, estimating component of variance. CV%, coefficient of variation (percentage of the standard deviation to the mean). 95% confidence limits are calculated based on the individual laboratory mean-values.

using fluorescence thresholds for chlorophyll and SYTOX green. DFI is the total signal of measured cells in the test tube, and represents the total amount of active photosystem in the measured cells. Although the number of active photosystem within the individual cells (or individual chloroplasts in the cells) may not be uniform, it is reasonable to expect DF represents the total number of active photosystems within each test tube. Fig. 4 shows the average growth curve and growth rate measured by DFI. The growth rates of alga with DCP concentrations of 3 mg/L and 9 mg/L for a 6 h exposure, and 9 mg/L for a 24 h exposure are negative; the DFI from the measured algal suspension at 6 h for these concentrations is lower than the DFI at 0 h (DFI0). As Briantais et al. (1980), and Joliot and Joliot (1980) suggested, the observed DF inhibition may reflect dissipation of the proton gradient across the thylakoid membrane or ATPase inhibition by DCP (an uncoupler). Fig. 4a also suggests that the effect of DCP in the first 6 h exposure period is higher than in the later 6 h to 24 h time period. The dynamics of growth inhibition should be interpreted in the context of bioavailability of DCP to alga (e.g. adsorption, uptake, mode of toxic action and kinetics). The observation of growth curves by DFI with exposure to DCP is similar to the growth curves obtained using living cell density under exposure to pentoxazone reported using flow cytometry (Nagai et al., 2011). The uptake of a chemical substance by algal cells and transport to the target site is related to the lipophilicity indicated by the log n-octanol/water partition coefficient (Kow). Furthermore, hydrolytic degradation of the exposure chemical is one of the reasons for decreasing effects in longer exposure periods. DCP is a lipophilic compound with Kow 3.62–3.68

(pentoxazone Kow 4.66), and is an uncoupler that transports protons through membranes. These chemical properties may be associated with rapid uptake of DCP into the thylakoid membrane, and rapid dissipation of the proton gradient (uncoupling), thereby inhibiting emission of DF. The decreased effect of DCP exposure in later times, 6 h to 24 h, may be related to the consumption of DCP in the cell, or adsorption to cell surface and the test tube. Since the concentration of the toxicant was not subjected to chemical analysis, change in concentration due to metabolism, adsorption, abiotic decomposition processes etc. cannot be excluded. On the other hand, DCP is not easily hydrolyzed in the growth medium. Bioavailability of DCP can also be influenced by cell density. Based upon the average growth rates in the ring study (e.g. 1.57 day− 1), the alga in the control samples at the initial density grows 4.8-fold to 96 × 104 cells/mL at 24 h. Similarly, TG201 cultures with an initial cell density of 1 × 104 cells/mL will attain 111 × 104 cells/mL at 72 h at the same growth rate. Thus, in view of the final cell density, we assume the difference in initial cell density does not significantly affect the results at 24 h. Certainly, lower initial cell density is better. The DF-test using the reagent-alga can be conducted with a lower initial cell density (e.g. 5 × 104 cells/mL) (Takeuchi et al., 2014). On the other hand, generation is one of the most important test conditions of the chronic toxicity test. For example, both the 72 h growth test in OECD (2011) and ISO (2004) require the growth of test alga to be 16 times or more at 72 h (also applicable in the 48 h short-term test in the ISO, 2004). Referring to the typical growth rate of 1.57 day−1 of the ring study, growth in our test was 1.5 times at 6 h, and 4.8 at 24 h. This growth may be

Fig. 7. Variation of ECX, in 6 h and 24 h time periods. Variation of X% effective concentrations (ECX, X = 50, 25, and 10) in 6 h and 24 h time periods at the 7 laboratories and 3 replications. Symbols indicate ECX in 6 h and 24 h exposure time periods.

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

849

Table 4 ANOVA table of X% effective concentrations in 6 h and 24 h model. ECX (mg/L) average ± 95% confidence limit

–

Entire analysis

Separate in time and X

6h

24 h

EC50

1.2 ± 0.3

EC25

0.7 ± 0.2

EC10

0.4 ± 0.1

EC50

2.7 ± 0.5

EC25

1.8 ± 0.4

EC10

1.1 ± 0.3

Factor

d.f.

Variance

F

p-Value

σ2

Variability (CV%) Within-lab

Between-lab

lab time X lab ∗ time lab ∗ X time ∗ X lab ∗ time ∗ X error lab error lab error lab error lab error lab error lab error

6 1 2 6 12 2 12 84 6 14 6 14 6 14 6 14 6 14 6 14

3.11 40.50 13.93 0.66 0.08 1.63 0.01 0.04 0.47 0.03 0.21 0.01 0.08 0.00 1.44 0.10 1.08 0.06 0.66 0.04

75.8 987.2 339.6 16.0 1.9 39.7 0.3 – 17.6 – 23.3 – 23.3 – 13.9 – 16.8 – 16.6 –

b0.001 b0.001 b0.001 b0.001 0.058 b0.001 0.994 – b0.001 – b0.001 – b0.001 – b0.001 – b0.001 – b0.001 –

0.170 – – – – – – 0.041 0.148 0.027 0.068 0.009 0.024 0.003 0.445 0.103 0.339 0.064 0.206 0.040

15

35

14

36

13

39

14

41

12

28

14

34

17

43

d.f., degrees of freedom; F, variance ratio; σ2, estimating component of variance. CV%, coefficient of variation (percentage of the standard deviation to the mean). 95% confidence limits are calculated based on the individual laboratory mean-values.

disparaged in comparison to the criterion of the conventional test; however, we suggest that a 24 h short-term test with 4.8 times growth would be useful. Of course, the DF-test duration can be extended. As shown in Fig. 5 and Table 2, intralab and interlab variability of the initial value (DFI0) are 4% and 16% respectively. DFI0 is significantly different between the laboratories. Since the laboratories used the reagent-algae prepared by the same method, the differences in DFI0 were probably caused by dilution of the cell suspension, or by variation of DFI0 yield per cell density. In this ring study, the laboratories are volunteers with expertise in biochemical experiments and toxicity measurements. Therefore we expect variation in skills of the participants is not the major reason for the variation in DFI0. The first 1 h preincubation after thawing is important to restore the photosystem activity of the reagent-alga. Differences in growth condition (e.g. actual temperature and light intensity on algal cell suspension) during the preincubation period may cause variations in photosystem activity. Although the growth medium was supplied from the same laboratory (pH 8.0), there is the possibility medium conditions varied between laboratories during the first 1 h pre-incubation. There is furthermore some chance that the yield of DFI0 for identical cell densities is due to variation in the ratio of growing alga to inactivated alga, as shown in Fig. 3. Inactivation of reagent-algae may occur in transportation and storage if the temperature is not maintained at deep freezing (−80 °C) temperatures.

Fig. 6 indicates the average growth rate of all control is 1.57 ± 0.25 day− 1, within in the range of established typical values, 1.5– 1.7 day−1 (OECD, 2011). The results suggest DFI is usable for calculation of growth rate, similar to other surrogates of biomass. On the other hand, as shown in Table 3, the growth rate of control samples varies between laboratories, and with interaction of laboratories and time periods. The intralab variability of the 24 h model is 7%. This satisfies a criterion of the TG201, variation of average growth rates during the test period in control cultures must not exceed 7%. On the other hand, the intralab variability for 6 h is relatively poor (20%), suggesting the 6 h exposure period does not provide stable growth rate, or uncontrolled experimental factors may be too large; 6 h may simply be too short to evaluate alga growth. The interlab variability (46% at 6 h, 15% at 24 h) is much higher than the intralab variabilities. The variation between laboratories is likely caused by differences in the growth conditions and incubators, including light condition and pH. For example, the light condition is not only defined by the light intensity (μmol m−2s−1) but also position or spectra of the light source, and variation (increasing) of pH during the test period may affect the observed growth rate. Since we did not request pH measurements from each institute (to simplify the workload on volunteer institutes), we are unable to analyze the effects of pH. Moreover as shown in Fig. 3b, variation in the ratio of growing alga and inactivated alga from sample to sample may

Fig. 8. Variation of NOEC and EC10 in 6 h and 24 h exposure time periods. Comparison of NOEC and EC10 in 6 h and 24 h exposure time periods. a) NOEC and EC10 in 6 h exposure time periods. b) NOEC and EC10 in 24 h exposure time periods. Circles indicate the individual EC10 of each replicate, and NOEC determined by Dunnett's multiple comparison test (twotailed test, α = 0.05).

850

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

EC10 of the alga DF inhibition test can be considered as nearly equivalent to NOEC.

Table 5 Intralab and interlab variabilities in similar studies. Reference

Duration

Endpoint

Variability

USEPA (2002)

72 h

EC25

Shieh et al. (2001)

96 h

EC10

interlab intralab intralab

Blaise et al. (1986)

96 h

EC50

intralab

Toxicant 24% 11% 19% 23% 20% to 25%

KCl Cd Ni Effluent

The variabilities are summarized important test result that can be related our study.

contribute to variability of DFI yield. The intralab and interlab variabilities of the 24 h growth rate may be acceptable; however, the variabilities in the 6 h growth rate are relatively poor. As shown in Fig. 7 and Table 4, difference of ECX (X = 50, 25 and 10) between laboratories is significant (p b 0.001). On the other hand, the ratios between maximum and minimum value of ECX (average of triplicates in Fig. 7) in each time period (6 h and 24 h) are within a factor of 2.0–3.4. The lowest ratio of 2.0 is in 24 h EC50 (maximum 3.4 mg/L and minimum 1.7 mg/L); the highest ratio 3.4 is in 24 h EC10 (1.7 mg/L and 0.5 mg/L). This difference may be acceptable for a toxicity screening test, as the purpose is to identify those compounds that require further quantitative evaluation. A deficiency of the ring study is not requesting pH measurement from each institute. Since the pKa of DCP is 8.1, even if the medium is provided with pH 8.0, drift of the pH during the 24 h growth test may considerably vary the toxicity of DCP. The supply laboratory confirmed the drift of pH is typically b0.5 under the same conditions as in the ring study; however, variable pH drift could have contributed to the difference in ECX among laboratories. The values of EC50 are highest among the ECX, and the values ECX is higher at 24 h than at 6 h exposure. Intralab and interlab variability of EC50 at 24 h are observed to be 12% and 27%. Variabilities are highest for EC10, at 17% (intralab EC10, 24 h) and 43% (interlab EC10, 6 h). As shown in Table 5, other studies have shown interlab and interlab variabilities in the range of 11%, and 20–25% respectively. The 24 h variabilities in our ring study are reasonable in the context of these previous works. As supplementary data, the supplying laboratory evaluated the EC50 of DCP using the reagent-alga under the same test conditions by both DFI and cell density (ten individual preparations). In this test, the EC50 in 24 h by DFI and cell density measurement are determined to be 2.6 mg/L (CV 18%) and 2.5 mg/L (CV 18%) respectively. The EC50 (or CV) at 24 h determined by DFI or cell density are similar to the values obtained in the ring study. The EC50 at 6 h by DFI and cell density are determined as 1.2 mg/L (CV 17%) and 3.9 mg/L (CV 28%) respectively. The EC50 (or variability) at 6 h by DFI is also similar to the values determined by the ring study, whereas EC50 (or CV) at 6 h by the cell density is higher than at 6 h by DFI in the results of the ring study or the supplementary data. Further analysis is necessary to understand the causes of the difference between DFI and cell density at 6 h exposure. However the difference may potentially relate to differences in detection of growing (living) and inactivated (dead) cells by DFI and cell density, suggesting DFI would not be expected to perfectly correlate with conventional cell count/volume at 6 h. On the other hands, this difference may provide additional information about the algal photosynthesis inhibition dynamics, specifically whether algae cells possesses active photosynthesis. Therefore DFI might be better considered as surrogate of biomass specifically with active photosynthesis, while excluding biomass without active photosynthesis. The result of toxicity estimation by DFI may be this therefore expected to differ from conventional cell count/volume. The differences between DFI and conventional surrogate metrics may furthermore be dependent upon the exposure duration after the toxicant inhibits photosynthesis activity, i.e. estimated effect of DCP on DFI at 6 h is stronger than the effect on the conventional surrogate measures. Longer exposure duration (i.e. 24 h) may correlate better with the conventional surrogate measures. As shown in Fig. 8, the NOEC and EC10 are relatively close, and are similar to results of a previous validation study (Shieh et al., 2001).

5. Conclusion DF of alga prepared as a reagent is examined in this study to assess its suitability as a rapid screening test for chemical toxicity, and as a surrogate measure of biomass. The DF intensity (DFI) significantly correlates with the dry weight of alga in control solutions; the calculated growth rate in the ring study (1.57 day− 1) is in the range of typical values (1.5–1.7 day−1). Furthermore, in a test containing living and dead alga, DF is shown to be specific for living biomass in a mixture of live and dead alga. Algal generation and final density are both important controls for the OECD TG201 test. The average growth in the 24 h growth test shows a 4.8–fold increase from the initial cell density to the final cell density (96 × 104 cells/mL), demonstrating both the generation capability of the test system and achieving a final density comparable to that specified in the OECD TG201 guidelines. Therefore, DF thus has potential to be used as a surrogate for the measurement of algal biomass in algal growth inhibition tests. A ring study using DFI with seven laboratories evaluates the concurrence and repeatability of results this protocol with for effective concentrations (ECx, X = 10, 25, 50) for DCP obtained with the conventional method. As development of a rapid protocol is one of the goals of this study, both 6 h and 24 h exposure times are evaluated. The EC50 values for 6 h and 24 h exposures (1.2 ± 0.3 mg/L and 2.7 ± 0.5 mg/L), differ from the EC50 determined using the 72 h conventional method (1.8 mg/L). DFI measures the total quantity of active photosystem in the sample and would not be expected to precisely equivalent to either the total dry weight of cells (which may contain dead cells) or the number of living cells (which may have variable internal photosystem activity). Measurement of DFI at any point in time requires only 1 min, and therefore measures the condition of the alga at that point in time. Our measurements show DCP toxicity estimated by DFI at 6 h is greater than at 24 h, and the EC50 at 6 h measured by DFI is lower than the EC50 measured by conventional methods at 72 h. Toxicity as determined by a one minute measurement of photosynthetic activity using DFI is likely to differ from methods such as conventional cell count/volume measurement that estimate growth over long times. The ratio between maximum and minimum of the ECX determined by averaging the values obtained from sample triplicates between laboratories is roughly 3 (range 2.0–3.4 times), while the differences of the NOEC determined by DFI are within one level of exposure concentration. Factors contributing to intralaboratory and interlaboratory variability are discussed in detail in the text. Importantly, the 24 h exposure test is required for stable growth in the control culture, and variability is within the OECD specification of b 7%. When using DF inhibition with reagent alga for rapid screening estimation of ECX, we therefore recommend using a 24 h exposure time to better estimate the toxic effects measured using conventional surrogate measures for dry weight such as cell counts, volume, optical density or fluorescence. The intralab and interlab variabilities of EC50 (24 h) for DCP (12% and 28%, respectively) are comparable to variabilities reported in other studies using conventional methods. These observations support the suitability of the reagent-algae DF inhibition test for rapid estimation of chemical toxicity. Acknowledgements The international ring study was carried out in 2011 and 2012 with support from Taku Tanaka, Kumiai Chemical Industry Co., LTD.; Mikio Kikuchi, Kanagawa Institute of Technology; Koji Arizono, Prefectural University of Kumamoto; Norio Oishi, Bio Safety Research Center; Masanobu Kawanishi, Osaka Prefecture University, Naoyuki Yokobori; Sumika Chemical Analysis Service, Ltd.; Hiroshi Yamamoto, University of Tokushima; Yoshie Tsuboi, Chemicals Evaluation and Research

M. Katsumata et al. / Science of the Total Environment 605–606 (2017) 842–851

Institute; Takafumi Mizuno, Unitika Environmental Technical Center, LTD.; Yohei Shimasaki, Kyushu University; Tatsuhiro Niino, Mitsubishi Chemical Medience Corporation. The ring study was managed with support by Yoshio Sugaya, National Institute for Environmental Studies; and Kimiko Kazumura, Takashi Koike, Yoshiki Maeda, Hiroshige Takada, Ken Nozaki, and Katsuhiro Kobayashi, Hamamatsu Photonics K.K. The first research report of the ring study was published at the SETAC Asia Pacific Annual Meeting 2012 (Katsumata et al., 2012b). This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References Barlow, R., Blake, J.F., 1989. Hill coefficients and the logistic equation. Trends Pharmacol. Sci. 10 (11), 440–441. Blaise, C., Legault, R., Bermingham, N., Van Coilloe, R., Vasseur, P., 1986. A simple microplate algal assay technique for aquatic toxicity assessment. Toxicity Assess. Int. Q. 1, 261–281. Breuer, F., Dören, L., Ebke, K.P., 2016. Comparison of four measuring techniques to assess growth inhibition in standardized tests with seven freshwater algae and cyanobacteria. Toxicol. Environ. Chem. 98 (8), 848–859. Briantais, J.-M., Vernotte, C., Picaud, M., Krause, G.H., 1980. Chlorophyll fluorescence as a probe for the determination of the photo-induced proton gradient in isolated chloroplasts. Biochim. Biophys. Acta 591 (1), 198–202. Bürger, J., Schmidt, W., 1988. Long-term delayed luminescence: a possible fast and convenient assay for nutrition deficiencies and environmental pollution damages in plants. Plant Soil 109, 79–83. Chioccioli, M., Hankamer, B., Ross, I.L., 2014. Flow cytometry pulse width data enables rapid and sensitive estimation of biomass dry weight in the microalgae Chlamydomonas reinhardtii and Chlorella vulgaris. PLoS ONE 9 (5), e97269. Comber, M.H.I., Smyth, D.V., Thompson, R.S., 1995. Assesment of the toxicity to algae of colored substances. Bull. Environ. Contam. Toxicol. 55, 922–928. Drinovec, L., Drobne, D., Jerman, I., Zrimec, A., 2004. Delayed fluorescence of Lemna minor: a biomarker of the effects of copper, cadmium, and zinc. Bull. Environ. Contam. Toxicol. 72 (5), 896–902. Escher, B.I., Schwarzenbach, R.P., 2002. Mechanistic studies on baseline toxicity and uncoupling of organic compounds as a basis for modeling effective membrane concentrations in aquatic organisms. Aquat. Sci. 64, 20–35. Franklin, N.M., Stauber, J.L., Lim, R.P., 2004. Development of multispecies algal bioassays using flow cytometry. Environ. Toxicol. Chem. 23 (6), 1452–1462. Goltsev, V., 2009. Delayed fluorescence in photosynthesis. Photosynth. Res. 101 (2–3), 217–232. International Organization for Standardization (ISO), 2004. 8692:2004 Water Quality Fresh Water Algal Growth Inhibition Test With Unicellular Green Algae. Joliot, P., Joliot, A., 1980. Dependence of delayed luminescence upon adenosine triphosphatase activity in Chlorella. Plant Physiol. 65 (4), 691–696. Jursinic, P.A., 1986. Delayed fluorescence: current concepts and status. In: Govindjee, Amesz, J., Fork, D.C. (Eds.), Light Emission by Plants and Bacteria. Academic Press, pp. 291–328. Katsumata, M., Koike, T., Nishikawa, M., Kazumura, K., Tsuchiya, H., 2006. Rapid ecotoxicological bioassay using delayed fluorescence in the green alga Pseudokirchneriella subcapitata. Water Res. 40 (18), 3393–3400. Katsumata, M., Takeuchi, A., Kazumura, K., Koike, T., 2008. New feature of delayed luminescence: preillumination-induced concavity and convexity in delayed luminescence decay curve in the green alga Pseudokirchneriella subcapitata. J. Photochem. Photobiol. B Biol. 90 (3), 152–162. Katsumata, M., Koike, T., Kazumura, K., Takeuchi, A., Sugaya, Y., 2009. Utility of delayed fluorescence as endpoint for rapid estimation of effect concentration on the green alga Pseudokirchneriella subcapitata. Bull. Environ. Contam. Toxicol. 83, 484–487.

851

Katsumata, M., Koike, T., Kazumura, K., Takeuchi, A., Sugaya, Y., 2010. Characteristics of delayed fluorescence for endpoint of algal growth inhibition test to estimate chemical toxicity. Jpn. J. Environ. Toxicol. 13 (1), 59–67. Katsumata, M., Bennett, K., Takeuchi, A., Kobayash, Y., Kazumura, K., Sato, Y., Koike, T., Tatarazako, N., Sugaya, Y., 2012a. Application of delayed fluorescence to estimate influence of chemical mixtures on alga. 6th SETAC World Congress/SETAC Europe 22nd Annual Meeting 2012a, TU 210 (poster presentation). Katsumata, M., Tatarazako, N., Tanaka, T., Kikuchi, K., Arizono, K., Oishi, N., Kawanishi, M., Yokobori, N., Yamamoto, H., Tsuboi, Y., Mizuno, T., Niino, T., Shimasaki, Y., Sugaya, Y., 2012b. First results from a ring study of a new rapid alga test estimating chemical influence on algal growth. SETAC Asia Pacific Annual Meeting 2012, 3D-2-5 (oral presentation). Lavorel, J., 1975. Luminescence. In: Govindjee (Ed.), Bioenergetics of Photosynthesis. Academic Press, pp. 223–317. Leunert, F., Grossart, H.P., Gerhardt, V., Eckert, W., 2013. Toxicant induced changes on delayed fluorescence decay kinetics of cyanobacteria and green algae: a rapid and sensitive biotest. PLoS ONE 8 (4), e63127. Nagai, T., Ishihara, S., Yokoyama, A., Iwafune, T., 2011. Effects of four rice paddy herbicides on algal cell viability and the relationship with population recovery. Environ. Toxicol. Chem. 30 (8), 1898–1905. Organization for Economic Cooperation and Development (OECD), 2011. OECD Guideline for Testing of Chemicals No. 201. Freshwater Alga and Cyanobacteria, Growth Inhibition Test (Paris, France). Paixão, S.M., Silva, L., Fernandes, A., O'Rourke, K., Mendonça, E., Picado, A., 2008. Performance of a miniaturized algal bioassay in phytotoxicity screening. Ecotoxicology 17 (3), 165–171. Ralph, P.J., Smith, R.A., Macinnis-Ng, C.M.O., Seery, C.R., 2007. Use of fluorescence-based ecotoxicological bioassays in monitoring toxicants and pollution in aquatic systems: review. Toxicol. Environ. Chem. 89 (1–4), 589–607. Schmidt, W., Senger, H., 1987. Long-term delayed luminescence in Scenedesmus obliquus I. Spectral and kinetics properties. Biochim. Biophys. Acta 890, 15–22. Scordino, A., Triglia, A., Musumeci, F., Grasso, F., Rajfur, Z., 1996. Influence of the presence of atrazine in water on the in-vivo delayed luminescence of Acetabularia acetabulum. J. Photochem. Photobiol. B Biol. 32, 11–17. Scordino, A., Musumeci, F., Gulino, M., Lanzanò, L., Tudisco, S., Sui, L., Grasso, R., Triglia, A., 2008. Delayed luminescence of microalgae as an indicator of metal toxicity. J. Phys. D. Appl. Phys. 41 (15), 155507 (7 pp). Serjeant, E.P., Dempsey, B., 1979. International Union of Pure and Applied Chemistry. Commission on Equilibrium Data; International Union of Pure and Applied Chemistry. Commission on Electrochemical Data. Pergamon Press, Oxford; New York. Shieh, J.N., Chao, M.R., Chen, C.Y., 2001. Statistical comparisons of the no-observed-effect concentration and the effective concentration at 10% inhibition (EC10) in algal toxicity tests. Water Sci. Technol. 43 (2), 141–146. Strehler, B.L., Arnold, W., 1951. Light production by green plants. J. Gen. Physiol. 34, 809–820. Takeuchi, A., Katsumata, M., Koike, T., Takata, Y., Itasu, Y., Kusui, T., 2014. Comparison of the conventional algal growth inhibition tests using cell counting and algal bioassay using delayed fluorescence: application to industrial effluents. J. Water Environ. Technol. 12 (4), 367–377. United States Environmental Protection Agency (USEPA), 2002. Test method green alga, Selenastrum capricornutum, growth test, method 1003.0. Short-term Methods for Estimating the Chronic Toxicity of Effluents and Receiving Waters to Freshwater Organisms, fourth editionEPA 821-R-02-013. United States Environmental Protection Agency (USEPA), 2012. Ecological Effects Test Guidelines OCSPP 850.4500: Algal Toxicity. Yamagishi, T., Katsumata, M., Yamaguchi, H., Shimura, Y., Kawachi, M., Koshikawa, H., Horie, Y., Tatarazako, N., 2016. Rapid ecotoxicological bioassay using delayed fluorescence in the marine cyanobacterium Cyanobium sp. (NIES-981). Ecotoxicology 25 (10), 1751–1758. Zrimec, M.B., Drinovec, L., Zrimec, A., Tišler, T., 2007. Delayed fluorescence in algal inhibition tests. Cent. Eur. J. Biol. 2 (2), 169–181.