Fluctuations in microcystin concentrations, potentially toxic Microcystis and genotype diversity in a cyanobacterial community from a tropical reservoir

Highlights

• We studied a cyanobacterial bloom in a tropical reservoir with molecular tools.

• Major Microcystis genotypes were detected along with other rare genotypes.

• Potentially toxic Microcystis cells and microcystin were present in all samples.

• Microcystin concentrations were not correlated with the number of toxic cells or the dominance of certain genotypes.

• We question whether qPCR is appropriate for use in MC risk assessment.

Abstract

In the study of cyanobacterial communities, molecular methods can be used to detect and quantify toxin-producing strains and assess genotype diversity. These are promising methods for estimating the risk of exposure to the cyanobacterial toxin microcystin (MC), assuming that the abundance of toxic genotypes is related to the MC concentration. This assumption has been tested in many studies, most of which were conducted in temperate and subtropical aquatic environments, and has not always been met. In this study, we collected cyanobacterial samples from a tropical reservoir and tested whether the MC concentration could be explained by variations in the number or ratio of toxin-producing genotypes or by the presence of certain Microcystis genotypes. The genetic diversity of cyanobacteria was determined by cpcBA sequencing, and the proportion of potential MC-producing genotypes was quantified by qPCR using mcyB as the target gene. Cyanobacteria dominated during the entire sampling period (88.4–99.8% of the total phytoplankton biomass), with Microcystis, Anabaena and Cylindrospermopsis occurring as the main genera. Three major Microcystis genotypes were recognized along with other rare or unique genotypes. MCs were detected in all samples (MC-LR, RR and YR), with most concentrations being less than 200 ng L⁻¹. Microcystis was the most abundant group during all the study period and was the main MC-producing genus. Potentially toxic Microcystis cells were present in all samples, varying from 10% to 100% of the total Microcystis. MC concentrations were not correlated with the number or ratio of toxic cells of Microcystis determined by qPCR or with the predominance of certain genotypes. These results indicate the limitations of using molecular methods such as qPCR in MC risk assessment.

Introduction

Cyanobacterial blooms can alter the aquatic food web and can be toxic to wildlife, domestic animals and humans (Codd et al., 2005). Microcystis is the most frequently reported genus in freshwater blooms worldwide (O’Neil et al., 2012, Paerl and Otten, 2013, Srivastava et al., 2013). Some Microcystis strains produce microcystins (MCs), a group of toxic cyclic heptapeptides encompassing more than 80 variants (Dittmann et al., 2013), which are the most common toxins produced by cyanobacterial blooms in freshwater (WHO, 2003, Figueiredo et al., 2004, O’Neil et al., 2012).

In natural populations of cyanobacteria, MC concentrations are determined based on the number of toxin-producing cells and the rate of MC synthesis, which is, in turn, affected by a variety of environmental factors (Figueiredo et al., 2004, Meissner et al., 2013). MC production may be constitutive in cells containing functional mcy genes and may vary 2–3-fold in response to environmental conditions (Meissner et al., 2013, Paerl and Otten, 2013). However, the seasonal variability in MC concentrations during Microcystis blooms exceeds the physiological variability in intracellular MC content reported for isolated Microcystis strains (Kardinaal et al., 2007). Thus, it appears that changes in MC concentrations during the development of Microcystis blooms mainly result from the seasonal succession of toxic and non-toxic genotypes.

The dynamics of Microcystis blooms have recently been studied using molecular techniques, exploring both the diversity and succession of genotypes as well as the variation in the proportion of toxic cells. Studies describing the genotypic diversity of Microcystis populations have reported contradictory findings, showing either selection for specific genotypes during bloom development (Kardinaal et al., 2007, Briand et al., 2009, Kim et al., 2010, Miller et al., 2013) or the maintenance of high genetic diversity within the population (Ye et al., 2009, Pobel et al., 2012, Wang et al., 2012, Zhu et al., 2012). Regarding the detection of mcy⁺ genotypes during bloom events, in some instances, the proportion of toxic genotypes remained relatively constant and the mcy copy numbers paralleled the Microcystis cell densities, thus prevailing at the peak of the bloom (Yoshida et al., 2007, Ha et al., 2009, Ye et al., 2009, Kim et al., 2010, Te and Gin, 2011, Srivastava et al., 2012). In other cases, the proportion of mcy⁺ cells varied, and some studies indicated a predominance of toxic genotypes in the emergence and/or senescence phases of a bloom (Kardinaal et al., 2007, Briand et al., 2009, Sabart et al., 2010, Martins et al., 2011). However, several additional cases showed no evidence for a relationship between the proportion of toxic cells and the bloom stage (Hotto et al., 2008, Davis et al., 2009, Conradie and Barnard, 2012, Pobel et al., 2012).

As bloom development is related to environmental factors, many studies have focused on the relationship between the relative abundance of potential MC-producing genotypes and physicochemical variables. Regarding nutrient concentrations, several studies have shown that nitrogen and/or phosphorus levels were positively correlated with the abundance of potentially toxic Microcystis cells (Yoshida et al., 2007, Davis et al., 2009, Davis et al., 2010, Ha et al., 2009, Rinta-Kanto et al., 2009, Te and Gin, 2011, O’Neil et al., 2012, Srivastava et al., 2012). Others indicated that MC-producing genotypes were positively correlated with pH (Ye et al., 2009) or water–surface temperature (Conradie and Barnard, 2012). However, no consensus has emerged to date, as evidenced by the aforementioned studies showing that when a particular abiotic factor is considered, contradictory results are reported, and, in some cases, no strong relationship is observed between any tested physiochemical variable and the abundance of mcy⁺ cells (Hotto et al., 2008, Martins et al., 2011).

In the majority of frequently cited studies, the correlation between the abundance of toxic genotypes and MC concentrations has been tested. Although positive relationships have been observed in certain cases (Hotto et al., 2008, Davis et al., 2009, Davis et al., 2010, Ha et al., 2009, Rinta-Kanto et al., 2009, Martins et al., 2011, Te and Gin, 2011, Conradie and Barnard, 2012, Srivastava et al., 2012), in other reports, a direct relationship between these two variables is not clear (Ye et al., 2009, Sabart et al., 2010). Furthermore, although cyanobacterial blooms are commonly recorded in the tropics, very few studies have employed molecular tools to investigate the diversity or potential toxicity of these populations compared to temperate and subtropical aquatic environments (Te and Gin, 2011). In temperate aquatic environments, Microcystis experiences a benthic phase during winter followed by recruitment in spring that results in planktonic proliferation and blooms in summer (Brunberg and Blomqvist, 2002, Kim et al., 2010). In contrast, in tropical environments, Microcystis can be found in the water column throughout the entire year, occasionally occurring as the predominant phytoplankton species (Soares et al., 2013). Therefore, an understanding of the dynamics of genotypes and production of MCs in natural populations of cyanobacteria in tropical areas can help reveal factors responsible for the predominance of certain species and prevalence of toxic strains.

In this study, we collected cyanobacterial samples from a eutrophic tropical reservoir where blooms have been recorded annually over the last 15 years (Soares et al., 2009). The three main genera in this system are Microcystis, Cylindrospermopsis and Anabaena, which are also the main bloom-forming genera of cyanobacteria in Brazil (Soares et al., 2013). We tested the hypothesis that MC concentrations can be explained by variations in the number or ratio of toxic genotypes or by the presence of certain genotypes of Microcystis. The results highlight the need to assess whether molecular methods such as qPCR are appropriate for use in MC risk assessment.