Antifouling potential of enzymes applied to reverse osmosis membranes

Abstract

Many companies in the food industry apply reverse osmosis (RO) membranes to ensure high-quality reuse of water. Biofouling is however, a common, recalcitrant and recurring problem that blocks transport over membranes and decreases the water recovery. Microorganisms adhering to membranes may form biofilm and produce an extracellular matrix, which protects against external stress and ensures continuous attachment. Thus, various agents are tested for their ability to degrade and disperse biofilms. Here, we identified industrially relevant bacterial model communities that form biofilms on RO membranes used for treating process water before reuse. There was a marked difference in the biofilm forming capabilities of bacteria isolated from contaminated RO membranes. One species, Raoultella ornithinolytica, was particularly capable of forming biofilm and was included in most communities. The potential of different enzymes (Trypsin-EDTA, Proteinase K, α-Amylase, β-Mannosidase and Alginate lyase) as biofouling dispersing agents was evaluated at different concentrations (0.05 U/ml and 1.28 U/ml). Among the tested enzymes, β-Mannosidase was the only enzyme able to reduce biofilm formation significantly within 4 h of exposure at 25 °C (0.284 log reduction), and only at the high concentration. Longer exposure duration, however, resulted in significant biofilm reduction by all enzymes tested (0.459-0.717 log reduction) at both low and high concentrations. Using confocal laser scanning microscopy, we quantified the biovolume on RO membranes after treatment with two different enzyme mixtures. The application of proteinase K and β-Mannosidase significantly reduced the amount of attached biomass (43% reduction), and the combination of all five enzymes showed even stronger reducing effect (71% reduction). Overall, this study demonstrates a potential treatment strategy, using matrix-degrading enzymes for biofouled RO membranes in food processing water treatment streams. Future studies on optimization of buffer systems, temperature and other factors could facilitate cleaning operations based on enzymatic treatment extending the lifespan of membranes with a continuous flux.

Keywords: Biofilm formation; Biofouling; Confocal laser scanning microscopy; Enzymes; Reverse Osmosis Membrane; Water recovery.

Fig. 1

Overall study design. To evaluate the effect of enzyme treatments on reverse osmosis (RO) membranes, mono- or multispecies communities were cultivated in 24-well plates with the membrane present for 72 h at 25 °C. Subsequently enzymes were added, and plates were left for incubation for either 4 or 24 h, before membranes were stained with Syto9, washed with saline and imaged with an inverted confocal laser-scanning microscope (CLSM).

Fig. 2

Biofilm formation screening. Biofilm formation was quantified after 24 h and 48 h of incubation, respectively, by crystal violet staining and subsequent optical density measurements at 590 nm. More biofilm was consistently formed after 48 h (blue + turquoise bar) than after 24 h (blue bar). A) A subset of the combinations with high level of biofilm formation is presented, and so are the single species measurements. All combinations can be found in Supplementary Table 1. B) Based on the screening, twenty-three communities of up to 5 species were selected for further analysis of biofilm dynamics. Error bars represent standard error of the mean (S.E.M.) from biological triplicates. Combinations of species in bold are those selected for further temporal analysis on RO membranes. Species numbers represent the following species: 1) Escherichia coli, 2) Bacillus sp., 3) Pseudomonas proteolytica, 4) Enterobacter sp., 5) Stenotrophomonas maltophila, 6) Raoultella ornithinolytica and 7) Rothia nasimurium. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Temporal quantification of biofilm on membranes. Biofilm formation of eleven different species/compositions was imaged and quantified after 24, 48 and 72 h, respectively. Different symbols represent different species compositions and grand mean is represented by black line. Displayed images are biofilm formation of species 3467 at 24, 48 and 72h, respectively. Each image is 319.45 × 319.45 μm. Numbers represent the following species; 2) Bacillus sp., 3) Pseudomonas proteolytica, 5) Stenotrophomonas maltophila, 6) Raoultella ornithinolytica and 7) Rothia nasimurium.

Fig. 4

Estimation of anti-biofilm potential of individual enzymes. The model communities were exposed to individual enzymes for four or 24 h, respectively. Some enzymes were applied in two different concentrations (0.05 U/ml and 1.28 U/ml) to evaluate the concentration effect. Trypsin-EDTA was applied as 0.0125% solution and proteinase K at 100 μg/ml. As controls, biofilms were exposed to 0.9% saline and sodium acetate, respectively. The latter was the buffer of the α-Amylase stock. A) Biofilm reduction by enzymatic reduction after 4 h of enzyme exposure. Only β-mannosidase (1.28 U/ml) caused a reduction in biofilm formation within this short period of incubation. B) Enzymatic exposure for 24 h significantly reduced biofilm formation for all enzymes tested, independent of concentration. The higher concentration was however associated with lower P-value. Asterisks indicate P-value, *P < 0.05, **P < 0.01, ***P > 0.001 and ****P < 0.0001 (Dunnett's multiple comparison - Data points used as replicates independent of species combination). Light green circles represent biofilms of species no. 6, red squares represent species no. 7, dark green triangles represent a mix of species no. 5 + 6, blue rhombuses represent a mix of species no. 5 + 6 + 7 and cyan hexagons represent all four species together. All symbols represent one biological replicate and black lines represent the grand mean. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

The effect of enzyme mixtures on membranes. The use of enzyme Mix A (100 μg/ml proteinase K + 1.28 U/ml β-Mannosidase) significantly reduced the amount of biovolume on RO membranes after 24 h of treatment compared to the non-treated starting point and the saline control (P < 0.001, Tukey's multiple comparison). The addition of more enzymes to the cocktail (Mix B: 100 μg/ml proteinase K + 1.28 U/ml β-Mannosidase + 0.0125% Trypsin-EDTA + 1.28 U/ml α-Amylase + 1.28 U/ml alginate lyase) reduced the total biovolume even further (Mix A vs. Mix B, P < 0.0001, Tukey's multiple comparison). Letters indicate significance of P < 0.001 to others letters and unmarked controls. Images display the amount of stained biofilm without treatment (control), with Mix A and mix B, respectively with species 7 and a mix of species 5, 6 and 7 as examples. Table below images indicates the presence of enzymes in specific mixtures (green = present, red = absent). Light green circles represent biofilms of species no. 6, red squares represent species no. 7, dark green triangles represent a mix of species no. 5 + 6, blue rhombuses represent a mix of species no. 5 + 6 + 7 and cyan hexagons represent all four species together. All symbols represent a biological replicate and black lines represent the grand mean. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)