Phycocyanin-specific absorption coefficient: Eliminating the effect of chlorophylls absorption

The applicability of algorithms for estimation of phycocyanin (PC) concentration based on light spectral reflectance heavily depends on the specific absorption of the pigment. But the determination of PC-specific absorption coefficient is not a straightforward task, as PC optical activity is overlapped by absorption of chlorophylls. The aim of our study was to determine a PC(625)—the specific absorption coefficient of PC at 625 nm, in samples with PC concentrations ranging from 0.5 mg m to 126.4 mg m and varying proportions of chlorophylls a, b, and c in the samples. The effect of chlorophylls was subtracted from total absorption at 625 nm using Chl a absorption at 675 nm as a reference; the contribution of the chlorophylls was computed on the basis of their relative absorption at 625 nm and concentrations. The major effect on the precision of a PC(625) determination was imposed by Chl a absorption, but the effect of the accessory chlorophylls ought to be accounted for in order to arrive at reliable PC-specific absorption values. a PC(625) varied widely, from 0.002 mg m to 0.027 mg m, and decreased with increase of PC concentration. As PC concentration exceeded 10 mg m, a PC(625) was almost invariable, slightly oscillating around 0.007 mg m 22 and similar to PC specific absorption coefficient estimated by alternative methods. It is clear that a universal value is unfeasible, but a PC(625) of 0.007–0.008 mg m 22 seems an appropriate value for use in algorithms destined for estimation of phycocyanin in typical mesotrophic and eutrophic inland waters dominated by cyanobacteria. A useful way to determine the potential of vegetation to capture light is by measuring mass-specific light absorption. Practically, the value of specific light absorption is vital for the development of analytical models for determination of pigment concentration (Gons 1999), and for algorithms destined to estimate phytoplankton primary productivity (Marra et al. 2007). In the case of phytoplankton, it usually defined as chlorophyll a (Chl a)-specific absorption (Kiefer and SooHoo 1982). Synthetic datasets of reflectance spectra showed that the relationship between optical information at the redNIR domain, essential for development of algorithms for Chl a estimation in productive waters, heavily depends on Chl aspecific absorption value at 665 nm (Gilerson et al. 2010). The measurement of phytoplankton absorption is accomplished using cell suspension or particles captured on a glass fiber filter. The specific absorption of a pigment in vivo is calculated by division of the absorption value at a given wavelength by the concentration of the pigment in question, and in the case of Chl a is a straight forward task, as the absorption peak is located beyond the optical activity of other pigments and is only slightly influenced by detritus and colored dissolved organic matter. The variation of the specific absorption coefficient of Chl a was extensively studied and the inverse relationship between pigment concentration and its specific absorption is well documented (Bricaud et al. 1995; Allali et al. 1997; Dall’Olmo and Gitelson 2005; Gurlin et al. 2011). Determination of specific absorption coefficient of other pigment than Chl a is complicated due to overlapping absorption of the pigments themselves, as well as by the effects of the noncellular water constituents. Of particulate importance is to establish real values for the water soluble phycobilins, signature pigments for cyanobacteria (Rowan 1989); some representatives of that phylum are toxinproducers and thus confer a potential environmental risk (Carmichael 1997). Cyanophytes often dominate phytoplankton in marine and freshwater ecosystems, and current knowledge leads to the conclusion that they will increase in *Correspondence to: [email protected] Present address: Department of Civil and Environmental Engineering, Technion, Haifa, Israel 157 LIMNOLOGY and OCEANOGRAPHY: METHODS Limnol. Oceanogr.: Methods 13, 2015, 157–168 VC 2015 Association for the Sciences of Limnology and Oceanography doi: 10.1002/lom3.10015 importance in the foreseen future (Paerl and Huisman 2008). With the increasing importance of these organisms attempts are made to improve methods for monitoring, among others by means of remotely sensed information, mostly by acquisition of spectral information in the visible and near-infrared domains (Simis et al. 2005, 2007; Hunter et al. 2009; Mishra et al. 2013, 2014; Ogashawara et al. 2013). In cyanobacteria dominating phytoplankton in inland waters, the major pigment is mostly phycocyanin (PC). PC absorption peaks around 610–620 nm and impairs a blue-green color in combination with Chl a. Phycobilins are located in protein-based structures, phycobilisomes, and the procedures applied for extraction of phycobilins are those used for most proteins, i.e., in aqueous solutions (Colyer et al. 2005). Those procedures are laborious and require specialized equipment, needed for cell disruption for the liberation of protein cellular components. The existing laboratory methods for phycobilins extraction do not suit large-scale work, as mostly required for field samples, and therefore, may introduce large margin of experimental error (Millie et al. 1992). Moreover, the bond between the phycobilins and the protein is not broken following cell disruption, and the spectrum showed by the released phycobilin is species specific, and dependent on the interaction between the protein and the chromophore. So far, a major effort was directed to develop an optical technique for remote estimation of PC, a water-soluble pigment common in freshwater cyanobacteria. The absorption signal of PC is detectable from reflectance spectra in eutrophic waters near 625 nm (Dekker 1993; Jupp et al. 1994; Gitelson et al. 1999; Schalles and Yacobi 2000). Following the same logic as for Chl a retrieval (Gitelson et al. 1985), a semianalytical algorithm for PC estimation has been proposed (Simis et al. 2005). The algorithm attributed the absorption signal in the 625-nm band to both PC and Chl a and to the absorption of Chl a around 670 nm. The initial tests have shown that the algorithm generally provides significant overestimations of the PC, except when extremely high PC concentrations associated with massive blue-green algae blooms were observed in the lakes (Simis et al. 2007). These authors conclude that correction for the red absorption by chlorophyll c (Chl c) and chlorophyll b (Chl b) is needed to yield more realistic PC assessments in inland waters. Thus, the state-of-the-art for PC retrieval can be summarized as follows: (1) PC prediction best matches observed values only during periods of high relative abundance of cyanobacteria in the plankton community, and the algorithm in its current form (Simis et al. 2005) is considered suitable for detection of the PC concentration above 50 mg m in cyanobacteria-dominated waters. (2) Applying a fixed value of the specific absorption coefficient for PC, a PC(625), caused an overestimation of the PC concentration that increases drastically at lower concentrations of that pigment. (3) The results of Simis et al. (2005) algorithm calibration suggest strong (up to 15-fold) seasonal variation in a PC(625). This difference is probably the result of package effect as well as various physiological states of algae (nutrient/light effects). (4) The presence of pigments other than PC and Chl a, and the variable influence of Chl a on retrieved absorption at 625 nm, are potential causes of error in PC retrieval. (5) Due to significant and systematic overestimation of Chl a by the algorithm developed by Gons et al. (2000) for Chl a < 20 mg m, there are significant errors in subtraction of Chl a contribution at 625 nm in the Simis et al. (2005) algorithm. It is likely that non-phytoplankton absorption as well as absorption by Chl c and Chl b may explain this tendency to overestimate Chl a (Gitelson et al. 2008). Absorption of all chlorophylls partly overlaps and also overlaps the range of the optical activity of PC. It is therefore necessary to take into account the effect of Chl a and the accessory Chl b and Chl c absorption in the spectral range of PC absorption in order to achieve a reliable estimation of PC-specific absorption. The goals of this study were: (1) to determine the net effect of PC absorption at the wavelength around 625 nm (where a peak is observed in absorption spectra and a small but discernible trough in reflectance spectra); (2) to establish the quantitative relationship between PC concentration and specific absorption coefficient of PC; (3) to define the minimum concentration of PC or relative proportion of it to Chl a where the optical activity of PC is detectable in natural samples. Materials and Procedures Sampling sites Water samples were collected in Lake Kinneret (Israel) in eight different campaigns (June, July, August, September, and November 2012 and June and July 2013) and in three campaigns in Berlin lakes (Germany) in August 2013. Experimental work Collected water samples were kept in a dark box, shipped to the laboratory, and processed within four hours from the first collection. Qualitative microscopic examination was undertaken to determine the dominant components of phytoplankton, identification limited to the level of genus. Subsamples of 200–250 mL in Lake Kinneret, and 40–200 mL in Berlin lakes were filtered onto glass fiber filter (Whatman GF/F) using vacuum of < 15 kPa and the wavelength-dependent absorption of the collected particles was determined spectrophotometrically in the range from 400 nm to 750 nm. To elucidate the optical impact of phytoplankton pigments, the sample (origin) was bleached for 20 min by 1% sodium hypochlorite solution, following Ferrari and Tassan (1999). The depigmented tripton (bleached) displayed an exponential decline of absorption from 400 nm toward longer wavelengths. The absorption of the depigmented material was subtracted from the absorption of the origin, resulting in a wavelengthdependent description of phytoplankton pigments absorption. The pigments absorption wavelength-dependent Yacobi et al. Phycocyanin-specific absorption coefficient