Algal density assessed by spectrophotometry: A calibration curve for the unicellular algae Pseudokirchneriella subcapitata

1 Laboratório de Ecotecnologia e Limnologia, Instituto de Pesquisas Hidráulicas, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves 9500, 91501-970, Porto Alegre, RS, Brazil. 2 Laboratório de Ecotoxicologia, Instituto de Biociências, Universidade Federal do Rio Grande do Sul, Av. Bento Gonçalves 9500, 91501-970, Porto Alegre, RS, Brazil. 3 Laboratório de Ecologia Aquática, Faculdade de Biociências, Pontifícia Universidade Católica do Rio Grande do Sul, Av. Ipiranga 6681, 90619-900, Porto Alegre, RS, Brazil.


INTRODUCTION
Algae play a major ecological role in most aquatic ecosystems as dominant primary producers (Pfleeger et al., 1991;Lewis, 1995).Several species have been shown to be sensitive to toxicants (Geis et al., 2000;Weyers et al., 2000), making this organisms widely recommended for ecotoxicological assays to evaluate toxicity of industrial wastewater or as bioindicators for chemical compounds present in water samples (Eaton et al., 1995).In this respect, the Chlorophycea Pseudokirchneriella subcapitata (Korschikov) Hindák (previously named Raphidocelis subcapitata and Selenastrum capricornutum) is one of the most frequently used algal species for toxicity tests (Nygaard et al., 1986).
Quantifying phytoplankton is usually done by time consuming methods, as direct cell counts under *Corresponding author.E-mail: luciarrodrigues@gmail.com.microscope or measurements of cellular mass or volume.Nevertheless, indirect methods that correlate algal density to light absorbance at specific wavelengths are not only reliable, but also easy to setup for automatic monitoring systems.So, the main goal of this work is to calibrate a regression model to estimate density of P. subcapitata in water samples by using spectrophotometry absorbance values.

MATERIALS AND METHODS
Ordinary data from routine chronic toxicity tests (n=130) with P. subcapitata were used to calibrate a mathematical model to estimate the algal density as a function of light absorbance through spectrophotometry.Algal concentration was estimated by the mean number of cells obtained from direct cell count.Three subsamples of 1 mL each were screened using a counting chamber (Neubauer) and a light microscope (Zeiss Inc.) following Mcateer and Davis (1994).
Maximum absorbance was inspected by scanning a culture sample between 600 and 800 nm (Cary 1E-Varian where a and b are calibration coefficients, estimated using standard least squares procedures for linear regression after log transformation of absorbance and density data.

RESULTS AND DISCUSSION
Standard routines to estimate algal concentration include direct cell counts, chlorophyll content measurement, and absorbance or turbidity numerical correlations (EPA, 1994).When spectrophotometrical absorbance is the chosen method, a reading wavelength of 750 nm is usually recommended (EPA, 1994;Eaton et al., 1995), although values of 680 nm (Rojícková-Padrtová and Marsálek, 1999;Geis et al., 2000;Markle et al., 2000) and 687 nm (Valer and Glock, 1998) have also been used.These values are correlated to the light absorbance of chlorophyll, which could be best determined at a wavelength around 664 nm (Hersh and Crumpton, 1987;Fargasová, 1996;Rojícková-Padrtová et al., 1998).
Figure 1 presents the pattern of light absorbance for a solution with P. subcapitata screened between 600 and 800 nm.Two peaks could be observed (624 and 684 nm), with the highest absorbance obtained at 684nm, representing the wavelength of maximum sensitivity to quantify P. subcapitata samples.So on, all further analyzed samples were read in this wavelength.
Figure 2 shows the relationship between absorbance and cell density for P. subcapitata solutions.The gray line represents the adjusted absorbance equation: Absorbance (684 nm) = 7.2578E ) was used instead a simple linear coefficient as a way to minimize bias related to cell shading in increased densities.The identified b value, a little larger than one, indicates that absorbance does not increase linearly with cell density, but at increasing rates related to a cell to cell shading effect.Solving the former equation, the cell density (cells/mL) from absorbance values at 684 nm could be estimated as follow: Cell Density = e {[ln (absorbance_684) +16.439]/1.0219} .
Nevertheless, even with a high overall determination coefficient (r 2 =0.9998), a biased absorbance response was identified when densities increased from 5 million cells/mL.Figure 3 shows the percentile deviation ([observed-expected]/observed.100) according to the adjusted model.Percentile deviations are under 2.5 (%) and with a random distributional pattern for cell densities  up to 4.5 million (cells/mL).Between 4.5 and 5.0 million (cells/mL), percentile errors are all positive, but under the 2.5% threshold.Above 5.0 million (cells/mL) the adjusted model is strongly biased, with error increasing exponentially.
Valer and Glock (1998) had already presented equations to estimate algal concentrations from absorbance data for cell densities between 10 4 and 10 5 cells/mL.In the present work, by using a power function, densities of P. subcapitata of up to 5,000,000 (5x10 6 ) cells/mL were precisely estimated.Nevertheless, the proposed equation is not recommended when the measured absorbance value exceed 0.5, which may require sample dilution.

Figure 1 .
Figure 1.Pattern of light absorbance for a solution with P. subcapitata screened between 600 and 800 nm.