It is generally accepted that food structure has an influence on bacterial growth (1). In the past decades, a myriad of research has been conducted on growth of bacteria in solid-food (model) systems (2-7). These studies used only one gelling agent, resulting in a homogeneous microstructure. In reality, food products are most often heterogeneous, as different phases (e.g., fat, water, proteins, etc.) are present. In addition, most studies are macroscopic, focusing on the overall population, whereas bacterial growth occurs at the level of the individual bacteria. In this study, the effect of a heterogeneous microstructure on the growth of Escherichia coli JM-109 DE3 was investigated by means of a nondestructive technique, i.e., confocal laser scanning microscopy (CLSM). This work is a first step in the process of unraveling the mechanism of bacterial growth in heterogeneous microstructures, i.e., microstructures consisting of different phases (8).For the study, a loop of an E. coli JM-109 DE3 pRSETb-Venus stock culture, kindly provided by the Department of Chemistry, KU Leuven, was transferred to an Erlenmeyer flask containing 20 ml of lysogeny broth medium enriched with 20 l of ampicillin and placed for 7 h at 37°C to obtain the preculture. Media were prepared by mixing 0.185 g brain heart infusion (BHI) (Oxoid, United Kingdom) and 0.146 g NaCl (AnalaR Normapur; VWR, Belgium) with different ratios of gelatin (from bovine skin, type B; Sigma, USA) and dextran (from Leuconostoc spp.; Sigma, Denmark) (M r , ϳ500,000) in glass tubes with a screw cap (see Table 1 for the composition of the different mixtures). After addition of 5 ml of distilled water, samples were placed in a water bath (GR 150 S12; Grant, United Kingdom) at 70°C for 12 min. In the next step, 10 l of 0.01% rhodamine B (R953; Aldrich, Germany) and 5 l of ampicillin were added and the mixture was filter sterilized by pushing it through a 0.2-m-pore-size filter (Filtropur S 0.2; Sarstedt, Germany) with the aid of a syringe (10-ml Norm-Ject; Henke Sass Wolf, Germany). Samples were inoculated with 10 l of preculture. Well chambers (Nunc Lab-Tek [USA] chambered borosilicate coverglass system) were filled with 300 l of inoculated medium and allowed to solidify at room temperature. After approximately 40 h, images were taken with a commercial laser scanning microscope (FV 1000; Olympus). The fluorescent probe rhodamine B and Venus fluorescent protein were used to visualize, respectively, the gelatin phase (see, e.g., reference 9) and the bacterial cells. The associated excitation wavelengths were 561 and 488 nm, respectively, and emission maxima were at 625 and 528 nm. The emission ranges recorded were 570 to 670 nm and 505 to 555 nm. A 60ϫ oil immersion objective was used with a numerical aperture of 1.35. Digital image files were acquired in .tif format at a resolution of 512 by 512 pixels. This results in an image resolution of 0.414 m per pixel. Experiments are performed twice, whereas multiple images were taken from each mixture. The images in Fig. 1 a...