Departamento de Botânica, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil
Correspondence: Alessandra Giani, Departamento de Botânica, Universidade Federal de Minas Gerais, Av. Antonio Carlos 6627, 31270-010 Belo Horizonte, MG, Brazil. Tel.: 55 31 34092681; fax: 55 31 34092671; e-mail: firstname.lastname@example.org
Cyanobacteria can form blooms and in these situations they dominate the phytoplanktonic community, reaching extremely high densities. In the domain Bacteria, high population densities can stimulate a phenomenon known as quorum sensing, which may produce several modifications in the cell physiology. Very little is known about quorum sensing in Cyanobacteria. Because of their planktonic way of life, quorum sensing should be more evident during a bloom event. In this work, we tested whether cell density could shape the production of bioactive compounds produced by Cyanobacteria. The experiments consisted of two treatments, where cultures of Cyanobacteria were maintained at low and high cellular densities through a semi-continuous set-up. Analyses were performed by HPLC-PDA and MALDI-TOF MS. Seventeen peptides were detected and 14 identified, including microcystins, aeruginosins, cyanopeptolins and microviridins. The results showed that cellular density seems to have a significant effect on the peptides production. Most of the compounds had significantly higher cellular quotas in the higher-density treatment, although microviridins and an unknown peptide were produced only at low density. These results may hint at a possible role for quorum sensing in triggering the production of several cyanobacterial peptides.
Cyanobacteria can form blooms in aquatic environments such as lakes, reservoirs and oceans; these situations consist of dominance of the phytoplankton community by one or a few Cyanobacteria species and an exponential increase in the biomass of the dominant group (Oliver & Ganf, 2002). Blooms have received considerable worldwide attention, because they are natural events in which populations of Cyanobacteria can reach extremely high densities. As most blooms are toxic, they represent a severe problem. It is acknowledged that nutrient enrichment of aquatic ecosystems promotes their appearance and persistence (Paerl et al., 2001). Phosphorous has been considered the primary nutrient source responsible for algal biomass accumulation in freshwater systems (Schindler et al., 2008) and together with nitrogen it may support the development of blooms when present in excess (Downing et al., 2001). Besides nutrient-rich conditions, other factors favour intensive cyanobacterial growth, such as temperatures above 20 °C, persistent vertical stratification, calm surface waters and low flushing rates (Reynolds, 1987; Paerl, 1988; Shapiro, 1990). After its establishment, a bloom can persist for several months, and in tropical regions, they can in some extreme situations become permanent (McGregor & Fabbro, 2000; Figueredo & Giani, 2009).
Cyanobacteria can produce several kinds of oligopeptides whose synthesis is usually achieved by nonribosomal peptide synthetase (NRPS) systems (for a full review see Welker & von Döhren, 2006). The major classes of these peptides are aeruginosins (Murakami et al., 1994), cyanopeptolins (Martin et al., 1993), microginins (Okino et al., 1993), microviridins (Ishitsuka et al., 1990), anabaenopeptins (Harada et al., 1995) and microcystins (Carmichael, 1992). The functions of these compounds are still unclear; however, it is known that some peptides are toxic to zooplankton (Agrawal et al., 2001; Czarnecki et al., 2006) and that microcystins might be related to physiological processes such as photosynthesis (Long et al., 2001; Young et al., 2005) and iron metabolism (Martin-Luna et al., 2006). Microcystins are also toxic to humans and other mammals (Sivonen & Jones, 1999). The increase of cyanobacterial blooms and the toxic nature of some of their peptides create major public health and water treatment problems (Watson et al., 2000; Watson, 2004).
There is a substantial amount of information and research on the factors controlling and affecting the appearance and persistence of cyanobacterial blooms. In addition, the knowledge about the production of microcystins and other cyanobacterial peptides is increasing. However, there is not much information about possible changes in the physiology of the cyanobacterial cells during a bloom event and its potential connection with the production of secondary metabolites. Some authors have previously observed changes in the relative amount of toxic genotypes in a bloom (Briand et al., 2008; Okello et al., 2010; Sabart et al., 2010; Pimentel & Giani, 2013). Furthermore, Wood et al. (2011) found that the expression of the mcyE gene during a bloom event was not constitutive, but was influenced by cell concentration and it could produce increases of up to 28-fold in microcystin levels. In mesocosm experiments, Wood et al. (2012) also observed that microcystin cell quota increased significantly with cell density.
Several bacterial species have the ability to release signalling compounds in their environment (Waters & Bassler, 2005), and when these molecules reach a threshold, they trigger a coordinated response able to change the gene expression pattern of the affected cells and, as a consequence, the metabolism and physiology of the bacterial population. The term quorum sensing is used to describe this density-dependent phenomenon (Fuqua et al., 1994). It is known that quorum sensing can control different biological functions as motility, aggregation, swarming, conjugation, luminescence, virulence, symbiosis, biofilm differentiation, antibiotics biosynthesis and others (Swift et al., 2001; Waters & Bassler, 2005; Williams et al., 2007). Up to date, a wide variety of molecules responsible for quorum sensing in bacteria have been isolated (for a review, see Williams et al., 2007). The phenomenon of quorum sensing is believed to be widespread in the bacteria domain (Miller & Bassler, 2001), but concerning Cyanobacteria, there is still a lack of information.
In this work, we examined under semi-continuous experimental conditions whether low and high cellular densities could affect peptide concentrations. The idea was that experimentally maintained high-density cultures would mimic a situation similar to a bloom event. The experiments were run using three different species of Cyanobacteria, isolated from Brazilian freshwater systems.
Materials and methods
The strains used in these experiments were Microcystis panniformis (strain Mp9), M. aeruginosa (strain Ma26) and a Radiocystis fernandoii (strain R28). Microcystis is one of the most studied genera and is known for its ability to form blooms (Oliver & Ganf, 2002) and produce secondary metabolites such as microcystins and others peptides (Welker et al., 2006). Radiocystis (Komárek & Komárková-Legenerová, 1993) is less common and less studied than Microcystis, but it is known to form blooms in tropical regions (Sant'Anna et al., 2008) and produce microcystins and other peptides (Vieira et al., 2003; Lombardo et al., 2006; Pereira et al., 2012). All strains are maintained in the culture collection of the Phycology Laboratory of the Botany Department in the Federal University of Minas Gerais and were isolated from Furnas reservoir (20°40′S; 46°19′W), located in the south-eastern region of Brazil.
Experiments were performed in two treatments, low cell density and high cell density. Each treatment was prepared in triplicate and received a different amount of inoculum, varying from 20 to 100 mL, to create two different cell densities, and the final volume was completed to 250 mL with WC medium (Guillard & Lorenzen, 1972). Growth conditions were 12 h light/12 h dark photoperiod, temperature 20 ± 1 °C and 65 μmol m−2 s−1 of irradiance. To avoid differences between experiments due to faster nutrient consumption in the higher-density treatment, experiments were conducted in semi-continuous cultures. Semi-continuous cultures allow the maintenance of a constant level of biomass and nutrient-replete growth under controlled conditions. Every other day, a constant volume of the experimental culture was removed and the same amount of fresh medium was added, to maintain the low or high cell density conditions within fixed limits. Under these settings, the experiments ran for 6 days, and cultures were kept at low and high cell density and exhibited similar growth rates. Samples were taken every 2 days to follow growth, and a minimum of 400 cells were counted in a Fuchs-Rosenthal hemocytometer. A 20 mL sample was filtered on the last day and used to measure chlorophyll, according to the methodology described by Nusch (1980). The rest of the culture was freeze-dried for further biochemical analyses.
The dry material was extracted three times with methanol 75% (v/v). The procedure was undertaken with sonication on ice followed by centrifugation (15 min, 8900 g). The extract was purified by reverse-phase chromato-graphy using SPE-C18 cartridges (Waters, Sep-Pak Vac 3cc – 500 mg) as described by Lawton and Edwards (2001). The purified extract was dried in Speed-Vac system and diluted in a known volume of methanol 75% (v/v). The analyses were performed using HPLC (Waters Alliance 2695) with a photodiode array detector (PDA – Waters 2996) at 225 and 238 nm wavelength. The column used was a Waters Symmetry C18 (4.6 × 250 mm I.D., 5 μm ODS). Mobile phase A was acetonitrile, containing 0.1% (v/v) trifluoroacetic acid (TFA), and mobile phase B was water, containing 0.1% (v/v) TFA. The chromatographic run consisted of a linear gradient from 30% A to 34% in 33.5 min then 40% for 6.5 min with a flow rate of 1 mL min−1. The peptide quantification was performed dividing the peak area by the dry weight of the lyophilized material, obtaining a measurement of the relative change in the peptide concentration; this method was chosen because of the lack of standards for most peptides. Dry weight was used to standardize the measurements, as in previous experiments it showed high correlation with cellular biovolume (Pereira et al., 2012). This relationship biovolume/dry weight was tested statistically by regression analysis, and results showed highly significant correlation for all strains (R2 ≥ 0.9).
Peptides were identified by collecting the HPLC fractions and submitting them to analysis in a MALDI-TOF-TOF Autoflex III mass spectrometer (Bruker Daltonics, Billerica, MA). The products were mixed with α-cyano-4-hydroxycinnamic acid matrix solution (1 : 1, v/v) and left to dry at room temperature in a MALDI target plate Anchorchip 600 (Bruker Daltonics). The peptide masses were obtained using a reflector mode and compared with known cyanobacterial metabolites. Known and unknown peptides were then fragmented using the LIFT fragmentation mode (MS/MS), and the fragment patterns were analysed according to Erhard et al. (1999) and Welker et al. (2006).
anova tests were used to compare the means of the two treatments (high and low cell densities). The parameters analysed were peptide concentration (area mg−1), cell density (cell mL−1), chlorophyll (μg mg−1) and growth rate (μ day−1). The statistical analyses were performed with jmp version 7 software.
Fourteen peptides were identified according to their fragmentation pattern and UV absorbance: 4 aeruginosins, 2 cyanopeptolins, 7 microcystins and 1 microviridin. Three metabolites could not yet be identified. Some peptides were produced at both high and low cell density, while others appeared in only one of the treatments (Table 1). Figure 1a and b shows the chromatographic pattern found in low and high cell density treatments for strain R28, Fig. 1c and d presents results for strain Mp9, and Fig. 1e and f for strain Ma26.
Table 1. Peptides produced by the studied strains
M + H
Characteristic Ions – M + H
Masses (M + H) are given in daltons.
Homo variant of the amino acid tyrosine in position 2.
Strain R28 produced four microcystins. The variants RR and YR were the major constituents, while the variants FR and WR were produced in lower concentration. Strains Ma26 and Mp9 produced the well-studied microcystin-LR and a nonidentified variant with a molecular weight of 976 daltons. There is no evidence in the literature of a microcystin molecule with 976 daltons, but the peptide we found presented the same fragmentation pattern and UV absorbance spectra that is characteristic of microcystins (see Supporting Information, Fig. S1; Fig. 1). Strain Ma26 also produced a different microcystin-YR, which seems to be a homovariant, showing the amino acid tyrosine in position 2. This compound had a molecular mass of 1058 daltons, which is the same mass of a nonindentified peptide (peptide 1057) produced by strains Ma26 and Mp9. However, the microcystin presented the fragmentation pattern and UV absorbance spectra that are typically found in microcystins, while the peptide 1057 had different fragmentation pattern and UV absorbance spectra. Furthermore, retention time was different for each one of the two compounds, characterizing them as different peptides.
Among aeruginosins, we found a total of four variants, aeruginosin 98B, aeruginosin 101 and aeruginosin 608, were produced by strain Ma26, and aeruginosin 684 and aeruginosin 101 were produced by strain Mp9. Aeruginosins were the main peptide class produced by strain Ma26, with aeruginosin 101 being the major compound, followed by aeruginosin 608. Only two cyanopeptolins were found in the strains used in these experiments, cyanopeptolin 1071 produced exclusively by R28 strain and cyanopeptolin 1044 produced by strains R28 and Mp9. For strain R28, both cyanopeptolins were produced at low and high cell densities, but the production was significantly higher in the high-density (HD) treatment for cyanopeptolin 1071. In this strain, the production of cyanopeptolin 1043 was higher in the HD treatment, but not significant. However, for the strain Mp9, the cyanopeptolin 1043 was only found in the HD treatment. It is possible that the production of this peptide in strain Mp9, growing at low cell density, is kept at basal levels that could not be measured by the HPLC technique for lack of sensitivity of the method. Only one microviridin was found in the experiments, the microviridin 1709, produced by strain 28: no other strains produced microviridins.
Three nonidentified peptides were detected in the experiments. Their fragmentation pattern could not be associated with any class of cyanobacterial peptides and additional biochemical studies are needed to establish whether they belong to an existing class of peptides or not. Among these peptides there are two compounds (P 1102 and P 1057) that had increased production at high cellular densities and one compound that was seen only at low cell density (P 562).
Figure 2 shows the peptide concentrations in all strains and treatments. The amount of all substances produced by strain 28 was significantly higher in the HD treatment, with the exception of cyanopeptolin 1043 which showed no significant difference between treatments and the microviridin 1709 that was detected only in the low-density (LD) treatment (Fig. 2a). The amounts of microcystin-LR and microcystin 976 produced by strain Mp9 were significantly higher in the HD treatment, the other compounds were only detected in the HD treatment, and peptide 562 was detected in two different peaks only in the LD treatment (Fig. 2b). For all compounds observed in strain Ma26, a significantly greater amount was produced in the HD treatment, with the exception of the peptide 562 which was detected only in the LD treatment (Fig. 2c and d).
The difference between cell density (cell mL−1) and chlorophyll concentration (μg mg−1) in the low and high cell density treatments was statistically significant (Table 2). Growth rates showed no significant difference between treatments, as cultures were maintained stable in a semi-continuous set-up. For strain Ma 26, growth rate was 0.14 (LD) and 0.16 day−1 (HD), for Mp9, 0.16 (LD) and 0.15 day−1 (HD), and for R28, 0.06 (LD) and 0.05 day−1 (HD).
Table 2. Cell density, chlorophyll a concentration and growth rate in the three strains studied at low and high cell density treatments
Cyanobacteria are known to produce several types of bioactive oligopeptides (Welker & von Döhren, 2006) and they also form blooms where cell density is ten to hundred times higher than in normal phytoplanktonic populations (Oliver & Ganf, 2002). In this research, we investigated whether cell density could affect the production of some of these peptides. Our results showed that the production of peptides was significantly different at low cell and high cell densities, suggesting the existence of a quorum sensing phenomenon in planktonic Cyanobacteria: the higher cell density may have modified the communication patterns among cells, and as a result, peptide concentrations changed.
During a bloom event, the elevated cellular density of cyanobacterial populations creates a favourable environment for quorum sensing. In this kind of situation, the high cellular density would intensify the accumulation of signalling molecules in the environment, thus optimizing the conditions for the occurrence of quorum sensing. Gobler et al. (2007) investigated the dynamics and toxicity of Cyanobacteria in a eutrophic lake in New York and found that the expression of the mcyE gene, which is part of the microcystin operon, was higher during periods of high cellular density and declined during months of lower cell density. Even though the authors did not link their results with quorum sensing, their findings may hint to a potential correlation between the existence of this phenomenon and the mcyE gene expression. Similar results were also found by Wood et al. (2011) who observed that higher levels of microcystin per cell (up to 28-fold) occurred when cell concentration increased from 70 000 to 4 000 000 cells mL−1 and corresponded to changes in the mcyE gene expression. In mesocosm experiments, Wood et al. (2012) also detected a significant up-regulation of microcystin cell quota upon increasing the concentration of Microcystis cells. Of course, caution has to be used when correlating laboratory experiments with natural environments, as experiments represent conditions that are simpler than natural scenarios; nevertheless, they allow the isolation and testing of a specific factor and the understanding of single phenomena. In our experiments, the high cell density cultures represent a bloom situation, and the results obtained suggest significant differences in the production of secondary metabolites between bloom and nonbloom events.
The experiments were made in semi-continuous cultures, which is an important procedure that guarantees that a low cell density culture does not attain high cellular density and that a high cell density culture does not reach stationary phase. This method assured that the strains maintained their cell density almost stable during the experimental period, which is evidenced by the nonsignificantly different growth rates measured in both situations for all strains tested.
How the quorum sensing modulates metabolic responses in Cyanobacteria is still poorly understood, but it is known that acylhomoserine lactones (AHLs), which is a class of autoinducers produced by several gram-negative bacteria, are produced and can influence the physiology of the organisms. Sharif et al. (2008), for example, showed that Cyanobacteria from the genus Gloeothece produce an N-octanoyl-homoserine lactone, the accumulation of which is able to change the expression of 43 different proteins. Romero et al. (2011) also showed that AHLs with different side chains can inhibit nitrogen fixation in Anabaena, while Van Mooy et al. (2012) found that phosphorous acquisition is increased in marine Trichodesmium consortia in the presence of AHLs.
Our results are an indication of a potential link between quorum sensing and production of secondary metabolites in Cyanobacteria as shown by the different peptide pattern at low and high cell densities. For most of the compounds measured, the production was stimulated in the high cell density treatment, with a significant increase in concentration usually from 1.5 up to 4 times at the HD, and in some cases even reaching 8.3 (Fig. 2) and 89.6 times higher values, respectively, for microcystin-LR in Mp9 (Fig. 2b) and microcystin-YR* in Ma26 strain (Fig. 2d). Interestingly, some peptides were present only in the LD treatment, having their production suppressed when the cellular density of the culture was higher, as observed for microviridin 1709 in R28 strain and peptide 562 in Mp9 and Ma26 strains. In a mesocosm experiment, Wood et al. (2012) found an increase in microcystin quota in situations where the cell number per mL reached 2.9 million and 7 million. In our experiments, the number of cells in the HD treatment varied from 1.9 to 3.1 million per mL, and even though the experiments had different designs and conditions, the results were similar, which corroborates the idea that microcystin production is stimulated in situations of high cellular density.
Although there is a lot of new information on cyanobacterial peptides, it is still difficult to understand their functions in the physiology and ecology of Cyanobacteria; some of them can inhibit proteases from zooplanktonic organisms, characteristic of protection against grazing (Agrawal et al., 2001; Rohrlack et al., 2003; Von Elert et al., 2004; Czarnecki et al., 2006). There are also indications that microcystins may be involved in light-related photosynthetic processes (Long et al., 2001; Young et al., 2005). It is known that no specific group of oligopeptide is essential for the growth of Cyanobacteria, as natural populations are composed of producing and nonproducing strains, for each peptide and peptide class (Fastner et al., 2001; Rohrlack et al., 2001; Welker et al., 2004), and no clear advantage has yet been seen for producing over nonproducing strains (Kaebernick et al., 2001). In our experiments, we saw that the peptide pattern changes not only in concentration, but also in peptide composition according to the cellular density of the culture. All strains had some peptides that appeared only in the low cell density treatment, and strain Mp9 had several peptides that exclusively appeared in high cellular density, which is in accordance with the idea that these compounds are not essential for survival, at least individually, although they may offer some advantage to the individual strains in particular situations.
Microcystins are characterized by the amino acid (2S, 3S, 8S, 9S)-3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid (Adda, position 5) (Carmichael, 1992). In this study, we found seven different variants of these compounds. Our experiments showed that all microcystin variants were produced at both cell densities; nevertheless, their production was stimulated when the cell density was higher. Because of their toxicity to humans and other mammals, microcystins are cause of concern for human health and water treatment plants (Watson, 2004), and as a consequence, a wide range of information is already available. Our findings add new knowledge to the ecology and physiology of this compound, because the fact that quorum sensing may stimulate the production of microcystins is of importance to the management of lakes, reservoirs and water treatment plants.
Aeruginosins stand as a class of compounds formed by linear peptides characterized by a hydroxy-phenyl lactic acid (Hpla) at the N terminus, the amino acid 2-carboxy-6-hydroxyoctahydroindole (Choi) and an arginine derivative at the C-terminus (Murakami et al.,1995). In strain Ma26, aeruginosins were the main peptide class. The production pattern for aeruginosins was the same as for microcystins, with significantly higher concentrations observed at the HD treatment. Only two cyanopeptolins were found in the strains used in these experiments. This class of peptides is characterized by the amino acid 3-amino-6-hydroxy-2-piperidone (Ahp), and the cyclization of the peptide ring by an ester bond of the b-hydroxy group of threonine with the carboxy group of the terminal amino acid (Martin et al., 1993). Cyanopeptolins also presented a production response similar to the observed for microcystins and aeruginosins.
The resembling patterns found for microcystin, aeruginosin and cyanopeptolin may reflect their biosynthesis route, which is based on NRPS systems. We observed similar responses for the three classes of peptides, and each strain had a major peptide belonging to a different class (microcystin for R28, an unknown peptide for Mp9 and aeruginosin for Ma26). Such results point to the fact that these compounds may have interchangeable functions. These observations, added to the knowledge that cyanobacterial natural populations consist of a mix of producing and nonproducing strains for each class of peptide (Fastner et al., 2001; Welker et al., 2004), are good indicators of the multiplicity of profiles found in nature with potential similar cellular function.
Microviridin is the largest known cyanobacterial oligopeptide (Ishitsuka et al., 1990), and this class is characterized by a multicyclic structure established by secondary peptide and ester bonds and a side chain of variable length; its amino acids are all in l-configuration and the only nonproteinogenic unit is the N-terminal acetyl group. In our experiments, only one variant was found in strain R28 and this peptide was produced only at low cell density, opposite to the observed for the other classes of peptides. As previously noted by some authors (Welker & von Döhren, 2006; Philmus et al., 2008), microviridins are synthesized ribosomally and their structure is finalized by post-translational modifications. The different form of synthesis might be one explanation for the opposite behaviour of this microviridin, and also an indication that its function might be different when compared to the peptides synthesized by NRPS.
There are still many open questions concerning the complex connections between Cyanobacteria, blooms, quorum sensing and production of secondary metabolites. Genetic studies and field experiments are needed to expand the knowledge on these interactions. Our results suggest that quorum sensing may be an important mechanism in regulating the physiology of bloom forming Cyanobacteria and that this phenomenon can play a significant role in the production of toxic and nontoxic peptides in these organisms.
This study was supported by a scholarship to D.A.P. from CAPES (Coordenação de Aperfeiçoamento do Pessoal Docente, Brazil), and by funds provided by Furnas Centrais Hidroeletricas, CNPq (Conselho Nacional de Pesquisa), FAPEMIG (Fundaçao de Apoio a Pesquisa de Minas Gerais) and CEMIG (Companhia Eletrica de Minas Gerais) to A.G.