© 2014 Nature America, Inc. All rights reserved. PROTOCOL NATURE PROTOCOLS | VOL.10 NO.1 | 2015 | 199 INTRODUCTION Since its first development in 1986 (ref. 1), AFM has developed into a powerful tool that has opened the doors to the nanoworld 2 . AFM is particularly well suited for biology, as it allows multiple characterizations (topography, mechanical and adhesive proper- ties) of living cells in their physiological environments. However, a prerequisite for such AFM experiments is the immobilization of the biological samples probed. However, this crucial step is often a challenge, as samples have to be immobilized individually and firmly enough to withstand the lateral forces exerted by the AFM tip, but without altering their cellular integrity. Immobilization of cells for AFM experiments Several techniques have been used to immobilize living cells. Living cells such as microorganisms can be chemically fixed on a solid substrate using glutaraldehyde or (3-aminopropyl)triethox ysilane (APTES) 3 , or they can be immobilized on gelatin-coated surfaces 4 . However, these techniques can, respectively, modify the interface of the biological sample, or they can pollute the AFM tips, leading to artifacts. Another strategy is to trap round- shaped cells such as bacterial cocci and yeast cells in the pores of polycarbonate membranes by filtration 5 or in lithographically patterned substrates by gentle drying 6 . The filtration technique has been widely used over recent years 7–10 , although it is time- consuming and cells can be exposed to mechanical forces when trapped in the pores. We therefore developed a new and versatile strategy in 2011 (refs. 11–13), which consists of immobilizing round single living cells in microstructured PDMS stamps by convective/capillary deposition. We have demonstrated that this generic protocol can be used to immobilize different types of round cells, ranging from small cocci bacteria to yeasts cells and even algae, by tuning the geometry of the PDMS stamp patterns (Supplementary Data). This approach was also exploited to immobilize yeast cells of dif- ferent species, Saccharomyces cerevisiae and Candida albicans 14–16 , as well as spores of Aspergillus fumigatus 17 . Statistical significance of results obtained from AFM: the new challenge Statistical analysis of results is desirable. As AFM is a tool adapted for single-cell analysis, this technology requires the analysis of multiple cells in order to achieve statistical confidence. This can- not be performed using techniques such as immobilization in a pore filter, as the deposition of the cells on the surface is random and the rate of filled pores is low and not controlled. By using our method of immobilization, it is possible to generate arrays of cells; therefore, AFM results on an array of 100 cells can be performed using different AFM modes, such as multiparametric imaging, chemical force microscopy, single-molecule force spec- troscopy and single-cell force spectroscopy 14,18–21 . Such setups are effectively ‘lab chips’ for AFM analysis. With the progress made in AFM data processing software 22 , it is thus possible to take all the data acquired to analyze in a reasonable time period, so as to generate relevant and statistically significant results for biologi- cal studies. Overview of the procedure Immobilization of living cells in PDMS stamps involves three sequential stages. The first stage is the generation of a glass and chromium mask that harbors microstructured patterns, and the transfer of these patterns onto a silicon wafer. The second stage consists of the preparation of a corresponding PDMS stamp. Finally, the third stage is the assembly of the living cells into the PDMS microstructured stamps. The generation of a silicon master is achieved by photolithography, followed by pattern transfer Generation of living cell arrays for atomic force microscopy studies Cécile Formosa 1–4 , Flavien Pillet 2,5 , Marion Schiavone 1,2,6 , Raphaël E Duval 3,4,7 , Laurence Ressier 2,8 & Etienne Dague 1,2 1 Centre Nationale de la Recherche Scientifique (CNRS), Laboratoire d’Analyse et d’Architecture des Systèmes (LAAS), Toulouse, France. 2 Université de Toulouse; LAAS, Institut des Technologies Avancées en Sciences du Vivant (ITAV), Institute de Pharmacologie et de Biologie Structurale (IPBS), Toulouse, France. 3 CNRS, Unité Mixte de Recherche (UMR) 7565, Laboratoire Structure et Réactivité des Systèmes Moléculaires Complexes (SRSMC), Vandœuvre-lès-Nancy, France. 4 Université de Lorraine, UMR 7565, Faculté de Pharmacie, Nancy, France. 5 CNRS, IPBS, UMR 5089, Toulouse, France. 6 L’Institut National de la Recherche Agronomie (INRA), UMR 972 Laboratoire d’Ingénierie des Systèmes Biologiques et des Procédés (LISBP), Toulouse, France. 7 ABC Platform, Nancy, France. 8 Université de Toulouse, Laboratoire de Physique et Chimie de Nano-Objets (LPCNO), L’Institut National des Sciences Appliquées (INSA)-CNRS-Université Toulouse III-Paul Sabatier, Toulouse, France. Correspondence should be addressed to E.D. (edague@laas.fr). Published online 31 December 2014; doi:10.1038/nprot.2015.004 Atomic force microscopy (AFM) is a useful tool for studying the morphology or the nanomechanical and adhesive properties of live microorganisms under physiological conditions. However, to perform AFM imaging, living cells must be immobilized firmly enough to withstand the lateral forces exerted by the scanning tip, but without denaturing them. This protocol describes how to immobilize living cells, ranging from spores of bacteria to yeast cells, into polydimethylsiloxane (PDMS) stamps, with no chemical or physical denaturation. This protocol generates arrays of living cells, allowing statistically relevant measurements to be obtained from AFM measurements, which can increase the relevance of results. The first step of the protocol is to generate a microstructured silicon master, from which many microstructured PDMS stamps can be replicated. Living cells are finally assembled into the microstructures of these PDMS stamps using a convective and capillary assembly. The complete procedure can be performed in 1 week, although the first step is done only once, and thus repeats can be completed within 1 d.