Quantification of Biofilm Morphology

J. B. Xavier¹, A. M. Reis², A. Schnell³, S. Wuertz³, E. S. Gilbert⁴, S. E. Cowan⁴, J. D. Keasling⁴, D. C. White⁵, J. S. Almeida⁶

¹ITQB/Univ. of Nova Lisboa, Oeiras, PORTUGAL; ²FCT/Univ. Nova Lisboa, Mt Caparica, PORTUGAL; ³Inst. Water Quality Control and Waste Management, Technical Univ. of Munich, Munich, GERMANY; ⁴Dept. of Chemical Engineering, Univ. of California, Berkeley, CA; ⁵Ctr. Environ. Biotechnology, Univ. of Tennessee, Knoxville, TN; ⁶ITQB and FCT/ Univ. of Nova Lisboa, Oeiras, PORTUGAL

ABSTRACT -The morphogenesis of biofilms has been shown to include self-organized structures, which establishes these microbial consortia as quasi-organisms. The resulting microbial aggregate often exhibits specialized structures such as channels and axis of development that are conserved throughout biofilm formation. From an engineering point of view, the effect of biofilm structure on process performance due to internal mass-transfer limitations as been a major concern since the early studies leading to the definition of Thiele modulus. Biofilm bioreactors, in particular those using membrane supported growth, have been observed to lead to higher volumetric productivities than alternative designs. However, the biological constraints defining the boundaries for volumetric interface area, porosity and turtuosity are still poorly understood. Confocal Laser Scanning Microscopy (CLSM) is the method of choice to investigate structure of live biofilms. Processing of CLSM images in order to correct for contrast, alignment and tilting is hereby described. The final image was segmented and further processed to characterize biofilm morphology. The quantification of morphological traits was pursued by identifying the developmental axis (DA) defined by the lowest diffusion resistance map technique (LDM). The geometric arrangement of DAs was further analyzed by determining its fractal dimension. The analysis outlined above leads to two main conclusions 1) the development of effective thickness measures that relate to mass transfer limitations; 2) the identification of regions within the biofilm whose geometry is, sequentially, geared towards attachment to the substratum, cell growth and cell detachment.

click here to enlarge

METHODS – Biofilms can be cultivated in vitro by inoculating laboratory flow cells. Staining methods such as commercially available probes or green fluorescence protein (GFP) expression are used for fluorescence image acquisition using CSLM. The choice of staining method is system specific and done according to the individual characteristics and the purpose of the analysis. Thin optical cross sections of the biofilm are imaged using CSLM. The entire three-dimensional biofilm structure is recorded by scanning along the biofilm depth, and the stacks of cross sections are stored as digital images. The software tools for quantitative analysis use a stack of optical cross sections by converting it into a matrix of data points, or three-dimensional image.
Prior to quantitative image processing, the raw images from CSLM undergo a series of preprocessing steps which include
Correction of flow cell tilting
Solid substratum surface detection
Contrast calibration (specially relevant for binary image processing)

VOLUMETRIC FEATURES – BIOFILME MEASUREMENTS FROM BINARY IMAGES - Image segmentation is the process of assigning the picture voxels to the recognizable image elements (e.g. distinguishing cell material from liquid media)The final result is a binary image: a 3D matrix with two possible values for each entry – e.g. 1 for biofilm, 0 for the background. The quantification of volumetric features is achieved by the analysis of the segmented (binary) images. They include direct measurements such as

And also heuristic measurements based on mass transfer properties of the biofilm structure. These are features obtained from the numeric integration of mass transfer differential equations

click to enlarge

QUANTIFICATION OF BIOFILM MORPHOLOGY THROUGH DETERMINATION OF DEVELOPMENTAL AXIS - Developmental axis (DA) are defined as elements within biofilm were accessibility to the bulk medium is the lowest. The method for DA identification uses numeric integration of internal mass transfer differential equations to compute a 3D least diffusion-resistance map (LDM). This method does not require segmentation and yields results that are independent of the original image contrast settings. Local biofilm morphology is accessed through the fractal dimension of developmental axis, D_f, a structural parameter that results from scale independence characteristics of the morphology. Df allows identification of regions within the biofilm with geometry geared towards attachment (support), growth and detachment according to its value and the proximity to the Euclidean dimensionality (D_e) of surface (D_e=2), lines (D_e=1) or points (D_e=0) respectively.

TWO SPECIES BIOFILM - Validation is a key step in the development of mathematical tools. The case study of this work has been mixed species biofilm. Validation of software tools is underway through diversification of the objects of study. CSLM images of a dual species biofilms comprised of two pseudomonads, Pseudomonas sp. strain GJ1 and Pseudomonas putida strain DMP1, were also analyzed. The biofilm was visualized through GFP expression of strain GJ1 and addition of a universal red stain (Syto 59, Molecular probesTM). Effective diffusivity ratio, porosity and turtuosity profiles of the structure were computed.

FIG6(click to enlarge) – Dual species pseudomonas biofilm stained with the GFP (strain GJ1) and a universal stain (Syto59, Molecular Probes TM).

INTERFACIAL AREA OF PHENANTHRENE CRYSTALS - In order to quantify the effect of the presence of phenathrene crystals in bacterial growth a method was developed to measure Interfacial area of phenathrene crystals using CSLM images. Segmentation of the images was effectuated using RATS (robust automatic threshold selection). Statistical methods were used to determine the error of the segmentation procedure on the interfacial area measurements.

FIG7(click to enlarge) – Bacterial communities (green) in the presence of phenanthrene crystals (red)

DENITRIFICATION - Formation of a mixed species denitrifying biofilm can be studied in detail by scanning the same site over time, thanks to the non-destructive characteristics of CSLM. A flow cell was inoculated with a mixed species consortium, obtained from a waste water treatment plant, and growth was followed for 28h. 3D images were collected using CSLM at 16h, 21h, 24h and 28h. The biofilm was stained with Syto 9 stain (Molecular Probes TM) prior to each image acquisition. Images were analyzed for effective thickness and morphology through determination of developmental axis. Time course analysis revealed exponential growth of effective thickness for the monitored period and a growth rate of 0.084 h-1 (thickness doubling time of 8.3 h). Developmental axis allowed classification of regions within the biofilm accordingly to the morphology.

FIG 8(click to enlarge) - denitrifying biofilm stained with the live/dead kit (Molecular Probes TM). Green and red represent live and dead cells respectively.

Acknowledgments:
This work is financially supported by the Foundation for Science and Technology/ M.C.T., PORTUGAL, through the grant PRAXISXXI/BD/18285/98 .