Towards a bioleaching biofilm image analysis platform: Semi-automatic quantification of biofilm development on pyrite by differential cell labeling

Abstract

Leaching microorganisms catalyze the dissolution of metal sulfides (MS) by attaching to mineral surfaces and producing extracellular polymeric substances (EPS), which among other characteristics, allows an efficient leaching of the mineral. Attachment to MS is affected by diverse factors, such as growth conditions, the previous/present presence of other colonizers or their EPS, cell-to-cell communication mechanisms, among others. In natural environments biofilms (BF) form as complex communities with multispecies interactions. Further insights in the study of BF development could help to improve the comprehension of interactions between different colonizing species in the mineral surface, however BFs are generally composed by a mixture of microorganisms, therefore, non- invasive BF analysis tools have to be further developed to allow species-specific recognition of cells in BF images. Bacteria are heterogeneously distributed in pyrite grains, therefore strategies such as, the analysis of large sets of BF images, are implemented to try to overcome this difficulty. A first step towards this objective is to recognize cells in axenic cultures. In this work BFs of acidophilic bacterial species were studied using epifluorescence microscopy (EFM) and three image analysis software packages. EFM images were acquired from pyrite samples with BFs formed on the mineral surface under different conditions, BFs of Acidiferrobacter sp. SP3/III Leptospirillum ferrooxidansT and Leptospirillum ferriphilumT were studied. The signal spots from 480 images were quantified by using semi-automatic software.

Lipophilic membrane dyes are fluorophores composed by a polar head and two hydrophobic alkyl chains with affinity for hydrophobic solvents, when the membrane dye is incorporated into a lipid membrane it produces enhanced fluorescence. The use of membrane cell tracing dyes may be used to follow up BF development, since it has the advantage to distinguish some species in mixed cultures, even when sizes and shapes of these are similar, by differential staining and counter-staining with DNA-binding dyes, however a relation between the quantified signal from lipophilic membrane dyes and mineral attached cells visualized by DNA-binding dyes, has not been defined. A comparison of the image analysis software packages: Zen 3.0, Fiji and a custom-made script was performed, obtaining similar relative quantification. It was found that the custom script obtained lowest quantification results, while Zen 3.0 identifies a higher number of colonies than the other 2 software for the same images. This is explained by the manual selection of the colonies step, that Zen 3.0 allows. Finally, the custom script was used to quantify the effect of the addition of synthetic QS signal molecules acyl-homoserine lactones (HSL) and diffusible signaling factors (DSF) on BF formation and cell dispersion. Interestingly, in L. ferrooxidansT HSLs external addition of 1 mM mixtures of C12-HSL & O-C12-HSL or C14-HSL & O-C14-HSL reduced the number of attached cells on pyrite grains at 70 h, while for L. ferriphilumT the attachment at 48 h diminished upon the addition of 1 mM mixtures of C12-HSL & O-C12-HSL or 1 mM O-C14-HSL. Changes in the pattern of dispersion or attachment were not possible to be measured upon DSF addition, given the current sample size.