Application of Real-Time PCR To Estimate Toxin Production by the Cyanobacterium Planktothrix sp.

Quantitative real-time PCR methods are increasingly being applied for the enumeration of toxic cyanobacteria in the environment. However, to justify the use of real-time PCR quantification as a monitoring tool, significant correlations between genotype abundance and actual toxin concentrations are required. In the present study, we aimed to explain the concentrations of three structural variants of the hepatotoxin microcystin (MC) produced by the filamentous cyanobacterium Planktothrix sp., [Asp, butyric acid (Dhb)]-microcystin-RR (where RR means two arginines), [Asp, methyl-dehydro-alanine (Mdha)]-microcystin-RR, and [Asp, Dhb]-microcystin-homotyrosine-arginine (HtyR), by the abundance of the microcystin genotypes encoding their synthesis. Three genotypes of microcystin-producing cyanobacteria (denoted the Dhb, Mdha, and Hty genotypes) in 12 lakes of the Alps in Austria, Germany, and Switzerland from 2005 to 2007 were quantified by means of real-time PCR. Their absolute and relative abundances were related to the concentration of the microcystin structural variants in aliquots determined by high-performance liquid chromatography (HPLC). The total microcystin concentrations varied from 0 to 6.2 μg liter⁻¹ (mean ± standard error [SE] of 0.6 ± 0.1 μg liter⁻¹) among the samples, in turn resulting in an average microcystin content in Planktothrix of 3.1 ± 0.7 μg mm⁻³ biovolume. Over a wide range of the population density (0.001 to 3.6 mm³ liter⁻¹ Planktothrix biovolume), the Dhb genotype and [Asp, Dhb]-MC-RR were most abundant, while the Hty genotype and MC-HtyR were found to be in the lowest proportion only. In general, there was a significant linear relationship between the abundance/proportion of specific microcystin genotypes and the concentration/proportion of the respective microcystin structural variants on a logarithmic scale. We conclude that estimating the abundance of specific microcystin genotypes by quantitative real-time PCR is useful for predicting the concentration of microcystin variants in water.During the last decade, genetic methods have significantly increased our understanding of the distribution of genes that are involved in the production of toxins within cyanobacteria that occur in fresh and brackish water (45). Although genetic methods can indicate only the potential risk of toxin synthesis and do not provide information about the actual toxin concentrations, quantitative real-time PCR has been increasingly applied for monitoring the toxin-producing genotypes of cyanobacteria in water (26, 33, 44). The development of real-time PCR methods was driven primarily by its potential (i) as an early-warning tool as well as to monitor toxin-producing cyanobacteria and (ii) to identify those factors that lead to a dominance/repression of toxin-producing genotypes versus nontoxic genotypes. For the first aim, it is essential that the abundance of toxin-producing cyanobacteria can be related to the concentration of the respective toxic substance in water. A few studies showed that the concentration of certain toxic genotypes was linearly related to the respective toxin concentrations, e.g., for the most common group of hepatotoxins, the microcystins (MCs) (7, 12, 14), and for the related nodularin (19). Both microcystins and nodularins are known to be potent inhibitors of eukaryotic protein phosphatases 1 and 2A, resulting in a health hazard to humans and the environment (9). In contrast, no correlation was found (37, 50), or even the opposite was reported, by other studies, i.e., that the measurement of microcystin-producing genotypes is not a satisfactory method for use in monitoring programs in order to predict the toxic risk associated with cyanobacterial proliferation (3). For microcystins, these contrasting results may be due to several reasons: (i) several genera producing microcystins frequently coexist in water bodies, and therefore, not all microcystin producers may have been identified; (ii) the semilogarithmic calibration curves limit the accuracy in estimations of genotype numbers and proportions (for example, the only laboratory comparison carried out so far revealed that among the three laboratories tested, the proportions of toxic genotypes were overestimated or underestimated by 0 to 72% and 0 to 50%, respectively [42]); and (iii) inactive mutants that contain the respective genes, however, which have been inactivated in toxin production through the insertion of transposable elements, may co-occur and decrease toxin production in a given population (6). Nevertheless, the real-time PCR technique is the only quantitative technique available for estimating the proportion of potential toxin-producing genotypes in water. The development of automated and field-applicable real-time PCR methods (e.g., see reference 35), in particular, may contribute to a more widespread integration of real-time PCR into routine monitoring programs in the future.In the present study, we attempted to quantify microcystin-producing genotypes in total as well as quantify the specific genotypes that were shown to encode different microcystin structural variants characterized for strains isolated from lakes in the Alps (23): (i) the methyl-dehydro-alanine residue (Mdha) genotype, which was found to synthesize structural variants containing only Mdha in position 7; (ii) the butyric acid (Dhb) genotype, which was found to contain Dhb instead of Mdha in the same position; and (iii) the homotyrosine (Hty) genotype, which was found to contain Hty and Leu in position 2 but never Arg. The Hty variant has always been found to co-occur with Dhb in position 7 of the molecule (24). Consequently, the Hty genotype forms a subgroup of the microcystin-producing population composed of the Mdha and Dhb genotypes. The following hypotheses were tested: (i) as only one microcystin-producing organism (Planktothrix sp.) is of quantitative importance in those lakes (32), the total microcystin concentration should be predictable from the sum of Mdha and Dhb genotypes; (ii) given that all Planktothrix genotypes are amenable to cultivation, all the structural microcystin variants found in the field samples should have been described for the strains isolated previously (23); and (iii) as, on average, the proportion of the inactive microcystin genotypes was found to be low and rather stable (<6.5% [32]), their occurrence should not reduce the ability to predict microcystin concentrations from genotype abundance. For this purpose, the phytoplankton in 12 lakes of the Alps in Austria, Germany, and Switzerland was monitored both with an inverted microscope as well as by means of real-time PCR over the course of 2 years (2005 to 2007). In parallel, microcystin concentrations in aliquots were determined by means of high-performance liquid chromatography (HPLC). We show that the abundance of specific microcystin genotypes can be related to the corresponding microcystin concentrations in water on a logarithmic scale over a range of trophic conditions. The proportion of certain genotypes encoding the synthesis of a specific microcystin variant significantly correlates with the concentration of the respective microcystin variant. We argue that these genotype-toxin concentration relationships are of great importance for the justification of real-time PCR use in monitoring programs.