Introduction

Analysis of biofilm structure requires the acquisition of three-dimensional (3-D) data. Transmission electron microscopy (TEM) and scanning electron microscopy (SEM) have been used to assess the structure of biofilms (1,2). Although these techniques provide high-resolution images, they have limitations in sample penetration and require dehydration of the sample, which may result in disruptive shrinkage and loss of biofilm matrix (3,4). Fluorescence in situ hybridization (FISH) presents the latter disadvantage, but it is a good technique to identify microorganisms in biofilms due to the use of specific fluorescently labeled nucleic acid probes (5,6). To achieve this, FISH requires multicolor fluorescence microscopy, which is usually performed by confocal laser-scanning microscopy (CLSM). CLSM is the most often used technique to acquire 3-D fluorescence microscopy data, since it permits in situ nondestructive study of hydrated living biofilms (7,8,9). It also allows the study of the physical structure of biofilms, the monitoring of biofilm development, and the assessment of biofilm response to the environment (10). The images obtained from CLSM are almost free of out-of-focus sections, because only a small fraction of the out-of-focus fluorescence emission can pass through the pinhole aperture. However, this also prevents the acquisition of a large part of the fluorescence emitted by the sample, reducing detection sensitivity and requiring illumination of the sample with high light intensity that may cause bleaching of fluorescent dyes. Furthermore, confocal laser-scanning microscopes able to detect different fluorochromes can be expensive and are not available in many research centers.

The aim of the present work is to find a way of assessing the 3-D structure of a biofilm using an optical microscope with bright-field transmitted and fluorescent light. With a common optical microscope, it is almost impossible to capture stacks of images at defined distances to be used in the construction of 3-D images needed to characterize a biofilm. However, when a biofilm is observed using an optical microscope under bright-field transmitted light, it is possible to observe that the presence of cells in the x-y plane decreases proportionally the amount of light that passes the sample in the perpendicular z-axis. The same principle is used in optical density determinations in spectrophotometers to calculate the amount of biomass in samples. To be able to determine the biofilm thickness, a relation between the amount of light observed in the two-dimensional (2-D) images and the number of cells in the vertical plane was needed. To determine the number of cells stacked in the z-axis responsible for a given pixel intensity in the x-y image, the biofilms were frozen with liquid nitrogen, and the stained cells were observed under fluorescent light.

The method was validated and applied in the study of the influence of solvents on biofilm formation in the wake of our studies on solvent toxicity in organic-aqueous systems (11).

Materials and Methods

Biofilm Formation

Rhodococcus erythropolis DCL14 cells were grown on 6-well plates (BD Falcon™ polystyrene tissue-culture plate; BD Biosciences, San Jose, CA, USA). Two hundred fifty microliters solvent were added to the 5-mL cell suspension. The solvents tested were the following: 99.8% ethanol and >99.5% toluene from Merck KGaA (Darmstadt, Germany), 99% pentane from Fluka Chemie GmbH (Seelze, Germany), >99% n-hexane and >99.5% iso-octane from Honeywell Riedel-de Haën (Seelze, Germany), 95% n-heptane from Lab-Scan (Hasselt, Belgium), >99% n-octane from Merck-Schuchardt (Hohenbrunn, Germany), 99% n-nonane from Acros (Geel, Belgium), and >99% n-dodecane and 99% n-hexadecane from Sigma-Aldrich Chemie GmbH (Buchs, Switzerland). The assays were done in triplicate. The biofilms formed were observed after 48 h. Cell hydrophobicity was measured as described by de Carvalho and da Fonseca (12).

Biofilm Observation

Method calibration

Each well of the 6-well plates containing the biofilms was observed under bright-field transmitted light with a 20× lens (N.A., 0.4), and images were taken on the intersection point of a grid with lines 1 mm apart. After staining the cells with a LIVE/DEAD® BacLight™ Bacterial Viability kit (Molecular Probes®; Invitrogen, Carlsbad, CA, USA), the content of the well was frozen with liquid nitrogen and sliced into 1-mm-thick slabs. Each slice, placed on a glass slide, was observed over a glass Petri dish (containing ice to keep the samples frozen) under fluorescent light with a 100× lens (N.A., 1.3). Images showing the number of stacked cells were taken 1 mm apart from each other. The microscope used was an Olympus CX40 with an Olympus U-RFL-T burner and an U-MWB mirror cube unit (excitation filter, BP450–480; barrier filter, BA515; Olympus, Center Valley, PA, USA). Images were captured by a Cohu RGB camera (Cohu, San Diego, CA, USA). The acquisition software was Matrox Inspector 2.1 (Matrox Imaging, Dorval, QC, Canada).