Structural study of an exocellular polysaccharide of Bacillus circulans

Sugar composition and FT-IR analysis of exopolysaccharides produced by microbial isolates from paper mill slime deposits

Thirty exopolysaccharides (EPS) produced by bacteria isolated from biofilms or slimelayers from different paper and board mills in Finland, France and Spain were subjected to size exclusion chromatography and sugar compositional analysis. High performance size exclusion chromatography (HPSEC) analysis revealed that some samples were composed of several molecular weight populations. These samples were fractionated by size exclusion chromatography and pooled accordingly. Principal components analysis (PCA) of the sugar compositions of the different pools indicated the presence of glucans and mannans caused by insufficient removal of the carbon or nitrogen source (yeast extract) from the bacteria growth medium leading to an overestimation of the glucose and mannose level in the sample, respectively. From the point of view of slime problems the EPS populations are the most important for multivariate analysis. Four groups of EPSs have been recognized by PCA analysis: a group of EPSs produced by Enterobacter and related genera similar to the regularly reported colanic acid; a group of Methylobacterium EPSs having high galactose and pyruvate levels and two groups that showed less dense clusters produced by Bacillus and related genera, showing high mannose and/or glucose levels and Klebsiella EPSs that showed galactose with rhamnose as major characteristic sugar moieties. Fourier transform infrared spectroscopy (FT-IR) of the same samples followed by discriminant partial least squares regression (DPLS) and linear discriminant analysis (LDA) showed that, when used with a well-defined training set, FT-IR could be used clustering instead of time-consuming sugar composition analysis. The Enterobacter and Methylobacetrium EPS groups could be recognized clearly. However the fact that this could hardly be done for the other two groups in the dataset indicates the importance of a larger and well-defined training or calibration set. The potential to use FT-IR, as a tool for pattern recognition and clustering with respect to EPS structures produced by micro organisms isolated from a paper mill environment is discussed.