Inhibition of Bacterial and Fungal Biofilm Formation by 675 Extracts from Microalgae and Cyanobacteria

Bacterial biofilms are complex biological systems that are difficult to eradicate at a medical, industrial, or environmental level. Biofilms confer bacteria protection against external factors and antimicrobial treatments. Taking into account that about 80% of human infections are caused by bacterial biofilms, the eradication of these structures is a great priority. Biofilms are resistant to old-generation antibiotics, which has led to the search for new antimicrobials from different sources, including deep oceans/seas. In this study, 675 extracts obtained from 225 cyanobacteria and microalgae species (11 phyla and 6 samples belonging to unknown group) were obtained from different culture collections: The Blue Biotechnology and Ecotoxicology Culture Collection (LEGE-CC), the Coimbra Collection of Algae (ACOI) from Portugal, and the Roscoff Culture Collection (RCC) from France. The largest number of samples was made up of the microalgae phylum Chlorophyta (270) followed by Cyanobacteria (261). To obtain a large range of new bioactive compounds, a method involving three consecutive extractions (hexane, ethyl acetate, and methanol) was used. The antibiofilm activity of extracts was determined against seven different bacterial species and two Candida strains in terms of minimal biofilm inhibitory concentration (MBIC). The highest biofilm inhibition rates (%) were achieved against Candida albicans and Enterobacter cloacae. Charophyta, Chlorophyta, and Cyanobacteria were the most effective against all microorganisms. In particular, extracts of Cercozoa phylum presented the lowest MBIC50 and MBIC90 values for all the strains except C. albicans.


Introduction
Bacterial biofilms are formed by aggregates of microorganisms within a complex biological system composed of assemblages of sessile cells adherent to each other or to a surface. Biofilms are The aim of the present study was to determine the antibiofilm activity of microalgae and cyanobacteria species against nine biofilm-forming human pathogens, representing the most important Gram-positive, Gram-negative, and fungal species, to search for new bioactive antibiofilm compounds using the biofilm inhibition ratio (%) and minimal biofilm inhibitory concentration (MBIC) assay.

Results
The results showed that 205 hexane extracts exhibited the best antibiofilm activity, followed by 195 extracts obtained with methanol and 189 extracts presenting inhibitory activity obtained with ethyl acetate. The rest of the extracts did not show any antibiofilm activity. Nevertheless, no significant activity was reported between methods of extraction (p > 0.05) ( Table S1). The small differences in activity among the three solvents suggested that the three solvent protocol covers a large range of compounds with different polarities, and it is more effective than extraction with only one solvent. Figure 1 provides an overview of the biofilm inhibition ratio (%) per group and solvent. The highest inhibition ratios were reported in C. albicans and E. cloacae in all solvents. Interestingly, C. albicans showed high inhibition rates above 50% of inhibition in all samples, with the exception of Glaucophyta and Miozoa methanol extracts (28.2% and 12.55%, respectively) and Rhodophyta hexane extract (34.77%). High biofilm inhibition ratios, about 35%, were found for E. cloacae. These rates were lower compared to C. albicans but still, more active in comparison with the rest of microorganisms. In the case of E. cloacae, only the methanol extract from the Miozoa phylum showed activity below 35% (9.38%). Biofilm inhibition of the nine microorganisms tested was individually analyzed in Figure 2. The inhibition rates in both cases were remarkably high and almost 50% of inhibitions were above the median value. On the other hand, in the case of S. hominis, only the methanol extract from Miozoa was able to inhibit it up to 64.22%. In addition, extract obtained from Cercozoa and Euglenophyta did not present activity against S. aureus. The aim of the present study was to determine the antibiofilm activity of microalgae and cyanobacteria species against nine biofilm-forming human pathogens, representing the most important Gram-positive, Gram-negative, and fungal species, to search for new bioactive antibiofilm compounds using the biofilm inhibition ratio (%) and minimal biofilm inhibitory concentration (MBIC) assay.

Results
The results showed that 205 hexane extracts exhibited the best antibiofilm activity, followed by 195 extracts obtained with methanol and 189 extracts presenting inhibitory activity obtained with ethyl acetate. The rest of the extracts did not show any antibiofilm activity. Nevertheless, no significant activity was reported between methods of extraction (p > 0.05) ( Table S1). The small differences in activity among the three solvents suggested that the three solvent protocol covers a large range of compounds with different polarities, and it is more effective than extraction with only one solvent. Figure 1 provides an overview of the biofilm inhibition ratio (%) per group and solvent. The highest inhibition ratios were reported in C. albicans and E. cloacae in all solvents. Interestingly, C. albicans showed high inhibition rates above 50% of inhibition in all samples, with the exception of Glaucophyta and Miozoa methanol extracts (28.2% and 12.55%, respectively) and Rhodophyta hexane extract (34.77%). High biofilm inhibition ratios, about 35%, were found for E. cloacae. These rates were lower compared to C. albicans but still, more active in comparison with the rest of microorganisms. In the case of E. cloacae, only the methanol extract from the Miozoa phylum showed activity below 35% (9.38%). Biofilm inhibition of the nine microorganisms tested was individually analyzed in Figure 2. The inhibition rates in both cases were remarkably high and almost 50% of inhibitions were above the median value. On the other hand, in the case of S. hominis, only the methanol extract from Miozoa was able to inhibit it up to 64.22%. In addition, extract obtained from Cercozoa and Euglenophyta did not present activity against S. aureus.   Only the results from phyla with a high number of extracts tested (Charophyta, Chlorophyta and Cyanobacteria) were statistically analyzed. The two-way ANOVA showed that both solvents and phylum significantly influenced the biofilm formation rates (p < 0.001) compared to growth control.
To further investigate which phylum and method of extraction (solvent) was more effective in biofilm inhibition, Tukey's multiple comparison test was also performed. Charophyta presented the best inhibition ratios against E. coli, P. aeruginosa, S. aureus, S. epidermidis, and C. parapsilopsis (Tables  S2-S4). Cyanobacteria induced biofilm inhibition versus K. pneumoniae and S. hominis. On the other hand, the three phyla presented the same effectiveness against both C. albicans and E. cloacae (Tables  S2-S4).
C. albicans and C. parapsilopsis were inhibited by 308 extracts. Biofilm formation among Gramnegative strains was inhibited by 202 extracts (30%) with equal distribution among species. Among Gram-positive strains, biofilm formation was inhibited by 69 of the 675 samples (10.2%). Only the results from phyla with a high number of extracts tested (Charophyta, Chlorophyta and Cyanobacteria) were statistically analyzed. The two-way ANOVA showed that both solvents and phylum significantly influenced the biofilm formation rates (p < 0.001) compared to growth control.
To further investigate which phylum and method of extraction (solvent) was more effective in biofilm inhibition, Tukey's multiple comparison test was also performed. Charophyta presented the best inhibition ratios against E. coli, P. aeruginosa, S. aureus, S. epidermidis, and C. parapsilopsis (Tables S2-S4). Cyanobacteria induced biofilm inhibition versus K. pneumoniae and S. hominis. On the other hand, the three phyla presented the same effectiveness against both C. albicans and E. cloacae (Tables S2-S4).
C. albicans and C. parapsilopsis were inhibited by 308 extracts. Biofilm formation among Gram-negative strains was inhibited by 202 extracts (30%) with equal distribution among species. Among Gram-positive strains, biofilm formation was inhibited by 69 of the 675 samples (10.2%).
Cercozoa and the unknown group showed activity at the lowest concentration in all the groups of microorganisms, except for C. albicans. Among all microorganism, the Cryptophyta, Euglenophyta, and Glaucophyta groups showed the best antibiofilm activity.
The unknown group (represented by two species) presented the highest activity in all microorganisms except for E. cloacae. Thus, the MBIC 50 was 64 µg/mL, and the MBIC 90 was 256 µg/mL for E. coli, P. aeruginosa, and C. parapsilopsis. K. pneumoniae and S. hominis showed MBIC 50 and MBIC 90 values of 64 and 128 µg/mL, respectively. Finally, S. aureus and S. epidermidis presented an MBIC 50 of 128 µg/mL and an MBIC 90 of 256 µg/mL.
Among the Gram-negative bacteria (Table 1), E. cloacae was inhibited by the lowest MBIC values of all the extracts tested. In addition to the results described above, the Cryptophyta and Rhodophyta phyla demonstrated activity against E. cloacae with an MBIC 50 and an MBIC 90 ranging from 16 to 256 µg/mL and from 16 to 128 µg/mL, respectively.
The Cryptophyta showed activity against P. aeruginosa with MBIC 50 and MBIC 90 values of 64 and 512 µg/mL, respectively.
Among Gram-positive bacteria (Table 2) Among Candida spp., the greatest activity was reported in C. albicans ( Table 3). Three of the 11 groups were able to inhibit biofilm formation with MBIC 50 values of 8 µg/mL. The lowest activity was reported by the Euglenophyta with an MBIC 50 of 8 µg/mL and an MBIC 90 of 16 µg/mL. Similar results were found with the Cryptophyta showing also an MBIC 50 of 8 µg/mL and an MBIC 90 of 128 µg/mL. Finally, the Glaucophyta presented an MBIC 50 and an MBIC 90 of 8 and 256 µg/mL, respectively. Interestingly, the Euglenophyta and Glaucophyta were only active against biofilm formation in C. albicans.
On the other hand, Rhodophyta species demonstrated activity against C. parapsilopsis with an MBIC 50 of 64 µg/mL and an MBIC 90 of 512 µg/mL.

Discussion
The present study was designed to determine the effect of new bioactive compounds from microalgae and cyanobacteria species on biofilm formation. We focused on these microorganisms because they are widely distributed in marine and freshwaters and are an important population in all strata, contributing as primary producers in open water systems [25][26][27][28]. Crude extracts are a heterogeneous mixture of polar and non-polar compounds. The selection of an efficient method of extraction is important for performing successive assays. For that reason, we used a method combining three different solvents (hexane [non-polar], ethyl acetate [polar], and methanol [polar]) for extracting a wide range of biological samples in order to collect the greatest range of polar and non-polar compounds.
The purpose of hexane is to extract polar compounds from the mixture, such as triacylglycerides (TAG), while methanol and ethyl acetate, polar solvents, can extract many biological compounds (polar and non-polar metabolites), such as fatty acids (FA). In fact, previous studies have reported that FA have the ability to inhibit biofilm formation. For example, oleic acids block bacterial adhesion in S. aureus [29]. The cis-2-decenoic acid synthetized by P. aeruginosa is able to induce the dispersion of established biofilms and inhibit biofilm formation [30].
Indeed, recent studies have demonstrated that small FA messengers inhibit cell-cell communication, achieving biofilm dispersion and are considered to be a quorum sensing inhibitor [31]. Hence, with three consecutive extraction solvents we were able to collect a wide range of lipids including fatty acids (FA), waxes, sterols, hydrocarbons, ketones, and pigments (carotenoids, chlorophylls, and phycobilins) [32], which have been described as possessing high antibacterial and antibiofilm activity [23,33]. The findings of the one-way ANOVA data support the claim that three solvents are necessary to obtain a wide range of molecules.
Only three phyla (Charophyta, Chlorophyta, and Cyanobacteria) were included in the two-way ANOVA analysis due to the number of samples extracted. Nevertheless, the fact that the other phyla showed interesting levels of biofilm inhibition rates (Figure 1) suggests that they are also good candidates for further studies.
The MBIC assay was used to determine the effectiveness of the sample extractions against biofilm producer microorganisms. Among bacteria, lower extract concentrations are needed to inhibit biofilm formation in Gram-negative in comparison with Gram-positive bacteria. Among the yeasts, lower extract concentrations are needed to inhibit biofilm formation in C. albicans compared to C. parapsilopsis. This finding was unexpected, and these extracts are being further investigated in an attempt to elucidate the mechanism underlying the antifungal action of these extracts.
The microalgae and cyanobacteria from which extracts generated were very diverse, and differences were observed in the level of bioactivity of different groups.
It is interesting to note that extracts from the Cercozoa phylum were very active against the eight target microorganisms in this study, except C. albicans, which was relatively unaffected. The Cercozoa phylum is a very diverse lineage of unicellular amoeboid organisms that are mostly heterotrophic and that can have a very complex cellular ultrastructure and behavioral patterns. Common examples in marine environments include the biomineralizing radiolarians and Foraminifera. A single relatively minor group within this lineage, the chlorarachniophytes, has acquired green chloroplasts by secondary endosymbiosis at some point in its evolutionary history and therefore qualifies as microalgae. Very little is known about the ecology or metabolism of the 14 described species in this group, but members appear to be good candidates for further investigation in the context of the search for new bioactive molecules. However, these data should be interpreted with caution because the Cercozoa phylum was represented by only three marine culture strains in our study, all of which are relatively difficult to grow (low growth rates and maximum cell abundances, adherence to the surface of culture vessels), meaning scale-up of cultures is likely to be challenging.
Several studies have described antimicrobial activities in microalgae and cyanobacteria species [34], but no results about antibiofilm against clinical pathogens, except those from Lauritano et al. [24], have been reported to date. Lauritano's group found two species belonging to the Leptocylindrus genus showing strong antibiofilm activity against S. epidermidis when they are grown in N-starved medium.
This issue may be important since resistance against antimicrobial agents changes depending on the expression of the phenotype. Planktonic microorganisms show greater susceptibility against the current therapies available than biofilm-forming microorganisms [35,36]. On the other end of the bioactivity spectrum, the Chlorophyta and Charophyta, which are both lineages of green microalgae, consistently had high average MBIC values. The Chlorophyta are known for their ability to synthesize a variety of bioactive compounds such as lipids and derivative polyunsaturated fatty acids (PUFAs). Two PUFAs, docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), have been demonstrated to exhibit antibacterial and antibiofilm properties [36,37]. Extracts from some chlorophyte strains had very low MBIC values and the high average values may be a reflection of the fact that the Chlorophyta was the most represented phylum in our study. Some microalgal groups appear to be specifically active against certain pathogens. For instance, extracts from Rhodophyta species exhibited relatively high antibiofilm activity against E. cloacae, while biofilm formation in C. albicans was particularly sensitive to extracts from Cryptophyta, Euglenophyta, and Glaucophyta (three completely unrelated lineages). Phyla such as Cyanobacteria, Haptophyta and Ochrophyta genera, all of which were fairly well represented in our study, consistently exhibited intermediate average MBIC values. These results can be explained due to several secondary metabolites such as circular or linear lipopeptides, amino acids, FA, macrolides, and amides with antibacterial activity [21].

Microalgae and Cyanobacteria Extracts
A total of 225 species of microalgae and cyanobacteria (belonging to the phyla Cercozoa, Charophyta, Chlorophyta, Cryptophyta, Cyanobacteria, Euglenophyta, Glaucophyta, Haptophyta, Miozoa, Ochrophyta, Rhodophyta, and 2 unknown species) were tested. They were supplied by Blue Biotechnology and Ecotoxicology Culture Collection (LEGE-CC) at CIIMAR, Centro Interdisciplinar de Investigação Marinha e Ambiental (CIIMAR) [38], the Coimbra Collection of Algae, University of Coimbra (ACOI), and the Université Pierre et Marie Curie-Paris 6 (UPMC). ACOI microalgae strains were collected mainly from freshwater habitats in Portugal. Cyanobacteria were collected mainly from beaches along the Portuguese coast (Atlantic Ocean) but also from brackish and freshwater systems (rivers, lakes, and estuaries). Only one of the cyanobacteria strains tested in this study was isolated from Chile (lake).
Microalgae cultures were performed by CIIMAR, ACOI, and UPMC. Briefly, samples were cultured until stationary phase for 15 or 30 days, depending on the strain, with air bubbling, and temperature and pH ranged between 20 and 30 • C and between 6 and 8, respectively.
Cyanobacteria strains were cultured for 30-45 days, depending on the strain, without bubbling, with daily manual shaking between 30 s and 1 min every day and temperature ranging between 25 and 30 • C. To culture cyanobacteria Z8 [39] and modified Z8 medium supplemented with 25 g/L of synthetic sea salts (Tropic Marin, Juliao do Tojal, Portugal) and B12 vitamin were used. The Z8 medium was supplemented because all marine cyanobacterial strains require artificial salts and some of them require the vitamin to grow. Cultures were performed under 14 h light /10 h dark cycles with a light intensity of 10-30 mol photons/m2/s).
Microalgae biomass was collected by centrifugation at 4000 rpm for 15 min. Pellets were frozen at −80 • C and lyophilized. Freeze-dried biomass was disrupted using a ceramic mortar, previously exposed to liquid nitrogen or with ultrasounds at 240 W, 35 kHz for 5 min.
Sequential extractions from hexane (non-polar), ethyl acetate to methanol (polar) were carried out. Each pellet was extracted by adding 3 × 20 mL of each solvent to centrifuge tubes, occasional vortexing and centrifuged at 4500 rpm for 15 min. Supernatant was recovered, transferred to glass vials, and fully dried in the rotary evaporator. Extracts were stored under dark dry conditions until use to avoid hydrolysis of bioactive molecules. The extracts were resuspended in 1 mL of 6% dimethyl sulfoxide (DMSO) (Panreac Applichem, Barcelona, Spain) immediately before the bioassay.

Microorganisms and Culture Conditions
Bacterial strains were stored in skim milk (BD) at −80 • C. In the present study, the strains were characterized in terms of biofilm formation, using the crystal violet protocol, as is shown in Table 4. E. coli, K. pneumoniae, E. cloacae, and P. aeruginosa were cultured for 24 h at 37 • C in aerobic conditions in Luria Broth agar (Condalab, Barcelona, Spain).

Minimal Biofilm Inhibitory Concentration (MBIC)
The MBIC assay was performed by the broth microdilution assay described in the CLSI document M7-A7 with some modifications [40]. The culture media used for biofilm formation experiments was M63 supplemented with 0.25% glucose for E. coli, Luria Bertani broth supplemented with 0.25% glucose for K. pneumoniae and P. aeruginosa, Tryptic Soy Broth (TSB) supplemented with 0.25% glucose for Gram-positive bacteria, and Yeast Nitrogen Base (YNB) for Candida spp. These culture media improve the biofilm formation in the corresponding bacterial specie.
For all strains, with an exception of C. albicans, two-fold dilution series of each microalgae extract in culture media was inoculated with 50 µL of a 0.5 McFarland Standard (corresponding to an inoculum of 5 × 10 6 cells/well) and incubated for 48 h at 37 • C (or 30 • C in the case of E. coli) in aerobic conditions without shaking. A negative control (culture medium without inoculum) and a positive control (culture medium with inoculum) were included in each plate. All the plates were covered with adhesive foil lids to avoid evaporation.
The biofilm susceptibility assay for P. aeruginosa was performed using the Calgary protocol as described previously [41] with one modification. The bacterial biofilm was formed by immersing the pegs of a modified polystyrene microtiter lid (catalog no. 445497; Nunc TSP system, Nunc, Roskilde, Denmark).
For Candida spp., a loopful of yeasts, from overnight culture, were washed twice with 3 mL of phosphate-buffered saline (PBS; pH 7.2; Ca 2+ and Mg 2+ -free), and the optical density of the suspension was adjusted to 0.38 at 520 nm. Two-fold dilution series of each microalgae extract in culture media was inoculated with 50 µL of a 0.38 OD (corresponding to an inoculum of 5 × 10 6 cells/well) and incubated for 48 h at 37 • C in aerobic conditions without shaking.
For all microorganisms, after incubation, liquid culture was carefully removed and washed once with PBS and dried at 65 • C for at least 20 min. Biofilms were stained with 100 µL of 1% (v/v) solution of crystal violet (CV) and incubated for 10 min at room temperature. Afterwards, the CV was completely removed, washing once with PBS and heat-fixed at 65 • C for 60 min.
The CV was eluted by the addition of 200 µL of 33% acetic acid. Biofilm formation was measured at 580 nm using a Microplate reader (EPOCH). The MBIC was defined as the lowest concentration of drug that resulted in a three-fold decrease of the optical density of 580 nm (OD 580 ) in comparison with the positive growth-control value. The MBIC 50 and MBIC 90 values were calculated for all the strains.

Statical Analysis and Data Processing
The biofilm inhibition rates were calculated using the equation: 100 × (1 − OD 580 of the test/OD 580 of non-treated control). The MBIC 50 and MBIC 90 were defined as the lowest concentration that caused 50% and 90% inhibition on the formation of biofilm. Statistical analyses were performed in Prism 8 software (GraphPad Software, Inc., La Jolla, CA, USA). One-way ANOVA was used to compare the effects of solvents against bacterial biofilms. Two-way ANOVA with Tukey's test was used to compare the phylum and solvents using OD values. Differences were considered statistically significant when p < 0.05. A circular dot plot was created with Tableau Software.

Conclusions
The present study provides initial large-scale evidence that microalgae and cyanobacteria are rich sources of substances with antibiofilm activity. Results demonstrated the importance of (1) employing a comprehensive extraction protocol in order to increase the chances of detecting bioactive substances; (2) testing extracts against a range of target microorganisms (because sensitivity to individual extracts differed among the organisms tested here); and (3) testing extracts from a broad range of source organisms (because significant differences were observed in the level of activity of extracts from different microalgal groups). Large-scale screening programs like this are extremely useful for identifying organisms that warrant further study as producers of bioactive substances of interest. Overall, the findings of this study provide insights for new opportunities provided by oceans and freshwater systems in the fight against biofilm infections. Further studies will be made in order to determine the active compounds responsible of the antibiofilm activity as well as their toxicity to mammal cells because at extract level toxicity could be due to other components different from the active one.
Supplementary Materials: The following are available online at http://www.mdpi.com/2079-6382/8/2/77/s1. Table S1: Three methods of extraction analyzed by one-way ANOVA. DF: degrees of freedom; P value: p < 0.05 were considered statistically significant. Table S2: Results of Gram-negatives strains comparing with treatment (phylum) and method of extraction (solvent). Two-way ANOVA followed by post hoc Tukey's test were used. MD: mean difference; SD: standard error of the difference; CI of diff: confidence interval of difference 95%; P value: p < 0.05 were considered statistically significant. Table S3: Results of Gram-positives strains comparing with treatment (phylum) and method of extraction (solvent). Two-way ANOVA followed by post hoc Tukey's test were used. MD: mean difference; SD: standard error of the difference; CI of diff: confidence interval of difference 95%; P value: p < 0.05 were considered statistically significant. Table S4: Results of Candida spp. comparing with treatment (phylum) and method of extraction (solvent). Two-way ANOVA followed by post hoc Tukey's test were used. MD: mean difference; SD: standard error of the difference; CI of diff: confidence interval of difference 95%; P value: p < 0.05 were considered statistically significant.