Recent development in molecular approaches allows access to genetic structure and diversity of indigenous microbial communities. In contrast, the functional analysis of microorganisms in their environment is still hampered by methodological limitations. Analysis of total proteins expressed at the whole community level (metaproteome) has been proposed to characterize the functional structure of microbial communities in their environment. However, developments are still required to perform such analysis. Our aim was to optimize methods to extract and characterize metaproteome of indigenous microbial communities. Experiments were first conducted in monoxenic bacterial cultures, and various methods were examined to define a procedure of protein extraction ensuring an efficient recovery regardless of the taxonomic affiliation of the cells. These developments were next applied to characterize the metaproteome from indigenous bacterial communities in freshwater samples. Bacterial cells were recovered from water using a high-speed density gradient centrifugation method before protein extraction and fingerprinting. The reactivity and sensitivity of this metaproteomic approach were tested by analyzing the variations of protein fingerprints according to perturbations (cadmium or mercury contamination). The genetic structure of the corresponding communities was also characterized by automated ribosomal spacer analysis (ARISA) DNA fingerprinting. Both protein and DNA fingerprints were statistically analyzed. Results obtained showed that the method developed for protein recovery and fingerprinting was efficient, sensitive, and reproducible. Both the functional and genetic structures of the freshwater bacterial community were complex and varied with perturbations. These variations occurred at both population and protein expression levels and were specific to the perturbation applied.