Analysis of Biofilm Structure and Gene Expression Using Fluorescence Dual Labeling Eric S. Gilbert, †,§ Artem Khlebnikov, †,| Stacie E. Cowan, ‡,⊥ and Jay D. Keasling* ,†,‡ Department of Chemical Engineering, University of California, Berkeley, California 94720, and University of California, Joint Bioengineering Graduate Program, Berkeley and San Francisco, California 94720 The use of biofilms for the degradation of recalcitrant environmental contaminants or for the production of secondary metabolites necessitates understanding and controlling gene expression. In this work, dual labeling with green fluorescent protein (GFP) variants was used to investigate inducible gene expression in a biofilm. Colocalization of GFP emissions was used to determine regions of attached cells and to correlate structure and activity within the biofilm. The labeling strategy reported here is unique in that the two GFP signals were distinguished by differential excitation rather than differential emission. Introduction Biologically catalyzed processes may be enhanced when the requisite microorganisms are cultured in biofilms. Potential benefits include increased reactivity in com- parison to suspended culture due to the comparatively high cell density (1) and more efficient separation of reaction products from microorganisms in the effluent. Processes that could be improved through the use of biofilm reactors include the biosynthesis of fine chemicals and bioremediation of mixed waste, particularly those chemicals that require multiple enzymatic modifications by multiple organisms. To study the distribution and the activity of biofilm members in response to changing conditions, markers are necessary to indicate a cell’s location within the biofilm and also its level of activity with respect to the function of interest. Moreover, three-dimensional imaging of bio- films by confocal scanning laser microscopy (CSLM), a powerful technique for elucidating the morphology and intercellular processes of microbial biofilms (2), neces- sitates spectrally distinct fluorescent signals. Dual labeling in combination with green fluorescent protein (GFP) for analysis of biofilms has been ac- complished using fluorescent stains, such as propidium iodide (3) and soluble nucleic acid dyes (4), or with fluorescently labeled oligonucleotide probes (5). In the present work, dual labeling with GFP and a soluble nucleic acid dye was used to determine the extent of gene expression after induction of an engineered E. coli strain grown in a monoculture biofilm. Subsequently, dual labeling with two GFP variants in a second E. coli strain was used to locate the metabolic activity and also to identify the population of cells attached to a cover glass. Materials and Methods Two E. coli strains, DH10B (F - -mcrA ∆(mrr-hsdRMS- mcrBC) φ80dlacZ∆M15 ∆lacX74 deoR recA1 endA1 araD139 ∆(ara, leu)7697 galU galK1 rpsL nupG; Life Technologies, Inc., Gaithersburg, MD) and 33456 (wild- type, (6)), were grown at 37 °C in minimal salts medium (MSM) with 0.1% glycerol, 0.1% Casamino acids, and 100 μg mL -1 ampicillin. Both strains harbored pCSAK50, a tightly regulated, arabinose-inducible, pMB1-derived plasmid containing a modified gfpuv gene (7). Strain DH10B also contained pSMC2, a pUCP-based plasmid with a gfpmut2 variant (8). pCSAK50 was induced with 0.02% arabinose. Biofilms were cultured in bench-scale parallel plate flow cells (working volume of 0.35 mL) at ambient temperature. Reactors were operated in batch (start-up) and continuous modes at a constant flow rate (0.862 mL min -1 ) using a peristaltic pump. Microscopic examinations were done with a Nikon Diaphot inverted microscope (Nikon Inc., Tokyo, Japan) equipped with a Bio-Rad MRC1024 scanning laser con- focal imaging system (Bio-Rad Inc., Hercules, CA). In the case of dual GFP imaging, the 363 nm line from a UV laser (Innova Technology/Coherent Enterprises, Santa Clara, CA) and the 488 nm line from a Kr/Ar laser (Bio- Rad Inc., Hercules, CA) were directed into the UBHS excitation and T1 emission filters. Samples were visual- ized using a Nikon 60× Plano-Apo oil immersion objec- tive, and the signals were sequentially collected (with a delay of approximately 4 s between scans) using 522/35 band-pass filters and directed in different channels. In the case of dual labeling with GFP and the nucleic acid stain, 488 nm light from the Kr/Ar laser was directed into the E2 excitation and T1 emission filters. Images were captured as described above. Prior to imaging, biofilms were stained with 20-μM red fluorescent nucleic acid stain SYTO 59 (Molecular Probes Inc., Eugene, OR). Green and red signals were collected separately using 522/35 and 605/32 band-pass filters, respectively. The obtained images were processed using Confocal Assistant 4.02 (Todd Clark Brelje) and Adobe Photoshop 5.0 (Adobe Systems, Inc., San Jose, CA) software. * Ph: 510-642-4862. Fax: 510-643-1228. E-mail: keasling@ socrates.berkeley.edu. † Department of Chemical Engineering. ‡ Joint Bioengineering Graduate Program. § Present address: Georgia State University, Department of Biology, 24 Peachtree Center Avenue, Atlanta, GA 30303. | Present address: Aviron Pharmaceuticals, Mountain View, CA 94043. ⊥ Present address: Procter & Gamble, Cincinnati, OH. 1180 Biotechnol. Prog. 2001, 17, 1180-1182 10.1021/bp0101031 CCC: $20.00 © 2001 American Chemical Society and American Institute of Chemical Engineers Published on Web 10/26/2001