AFM imaging enables microbiologists to visualize the organization and dynamics of microbial cell walls and appendages at (near) molecular resolution, thereby answering pertinent questions that could not be addressed before. A key benefit of AFM is that the specimen need not be stained, labeled, or fixed and can be imaged under physiological conditions. By revealing the ultrastructural details of the outermost cell surface, AFM complements fluorescence microscopy, which probes the entire cell wall at lower resolution.
Cell wall architecture.
Peptidoglycan is the main constituent of bacterial cell walls. Despite the important functional roles of this polymer (mechanical strength, cell shape, and target for antibiotics), its three-dimensional organization has long been controversial (
16). In the most widely accepted model, glycan strands run parallel to the plasma membrane, arranged perhaps as hoops or helices around the short axis of the cell, resulting in a woven fabric. In the past years, AFM imaging has complemented electron cryomicroscopy and tomography techniques in providing key structural details of peptidoglycan, such as strand orientation. Much of this work has been carried out on purified sacculi by the Foster research team (for a recent review, see Turner et al. [
17]). In an initial study, they reported that the cell wall of the model rod-shaped bacterium
Bacillus subtilis has glycan strands up to 5 µm, thus longer than the cell itself (
18). The inner surface of the cell wall showed 50-nm-wide peptidoglycan cables running parallel to the short axis of the cell, together with cross striations with an average periodicity of 25 nm along each cable (
Fig. 3a). The data favored an architectural model where glycan strands are polymerized and cross-linked to form a peptidoglycan rope, which is then coiled into a helix to form the inner surface cable structures. In another study, AFM was combined with optical microscopy with fluorescent vancomycin labeling to investigate the distribution of peptidoglycan in the spherical bacterium
Staphylococcus aureus (
19). Concentric rings and knobbly surface structures were observed and attributed to nascent and mature peptidoglycan, respectively (
Fig. 3b). Peptidoglycan features were suggested to demark previous divisions and, in doing so, hold the necessary information to specify the next division plane. Peptidoglycan architecture and dynamics have also been investigated in bacteria with ovoid cell shape (ovococci), including a number of important pathogens (
20). Here, AFM images showed a preferential orientation of the peptidoglycan network parallel to the short axis of the cells, while superresolution fluorescence microscopy unravelled the dynamics of peptidoglycan assembly. The results suggested that ovococci have a unique peptidoglycan architecture not observed previously in other model organisms. Recently, the rod-shaped Gram-negative bacterium
Escherichia coli was shown to feature peptidoglycan structures running parallel to the plane of the sacculus but in many directions relative to the long axis (
21). The images also revealed bands of porosity running circumferentially around the sacculi (
Fig. 3c). Superresolution fluorescence microscopy unravelled an unexpected discontinuous, patchy synthesis pattern. A model was suggested in which only the more porous regions of the peptidoglycan network are permissive for synthesis. Accordingly, these high-resolution studies have shown that bacterial species exhibit a variety of peptidoglycan architectures, thereby contributing to new structural models of peptidoglycan arrangement.
The organization of peptidoglycan has also been visualized in living cells. In early work, changes in
S. aureus peptidoglycan architecture (nanoscale holes and concentric rings) were observed during growth (
22). High-resolution images of
Bacillus atrophaeus spores during germination revealed a porous network of peptidoglycan fibers, consistent with a honeycomb model structure for synthetic peptidoglycan oligomers (
23). Interestingly, SMFS using functionalized tips provides a means to identify and localize single peptidoglycan chains in live cells (
10,
24). Using vancomycin tips, D-Ala−
d-Ala sites of peptidoglycan were shown to locate on the equatorial rings of
Lactococcus lactis, suggesting that newly formed peptidoglycan was inserted in these regions (
24). In the same vein, Andre et al. (
10) used AFM tips modified with the lysine motif (LysM) to image peptidoglycan nanocables in
L. lactis. Using topographic imaging, they found that wild-type cells display a featureless surface morphology, while mutant cells lacking cell wall exopolysaccharides featured 25-nm-wide periodic bands running parallel to the short axis of the cell (
Fig. 2b). In addition, mapping wild-type cells with LysM tips confirmed that peptidoglycan was hidden by other cell wall constituents, while anisotropic peptidoglycan bands were detected on the mutant. Accordingly, high-resolution AFM images of sacculi and live cells have greatly contributed to refining our current perception of peptidoglycan architecture in a variety of bacterial species.
Glycopolymers represent another class of cell wall constituents which fulfill important functions, such as protecting the cell against unfavorable environmental conditions, mediating cellular recognition, and promoting biofilm formation. Stukalov et al. used AFM and transmission electron microscopy to study capsular polysaccharides of four different Gram-negative bacterial strains (
25). While electron microscopy analysis revealed capsules for some but not all of the strains, AFM allowed the unambiguous identification of the presence of capsules on all strains. Moreover, AFM visualized bacterial cells within the capsules, indicating that the technique is capable of probing subsurface features. Francius et al. (
9) probed the cell surface polysaccharides of the probiotic bacterium
Lactobacillus rhamnosus GG (
Fig. 2a). AFM images of the cells in buffer revealed a rough morphology decorated with nanoscale waves. These features reflected extracellular polysaccharides, as they were hardly seen in a mutant impaired in exopolysaccharide production. In addition, SMFS with tips functionalized with lectins was used to identify single polysaccharide chains, demonstrating the coexistence of polysaccharides of different nature on the cell surface. Although teichoic acids are known to play important roles during cell elongation and cell division (
26), we know little about the relationships between the spatial localization of these components and their functional roles. To address this issue, AFM was combined with fluorescence microscopy to map the distribution of wall teichoic acids (WTAs) in
Lactobacillus plantarum (
11). Phenotype analysis of wild-type and mutant strains revealed that WTAs are required for proper cell elongation and cell division. Nanoscale imaging by AFM showed that strains expressing WTAs have a highly polarized surface morphology, the poles being much smoother than the side walls (
Fig. 2c). SMFS and fluorescence imaging with specific lectin probes demonstrated that the polarized surface structure correlates with a heterogeneous distribution of WTAs, the latter being absent from the surface of the poles. These findings show that the polarized distribution of WTAs in
L. plantarum plays a key role in controlling cell morphogenesis.
How about cell wall proteins? AFM has been intensively used to image proteins in purified membranes, at subnanometer resolution directly in aqueous solutions. These high-resolution studies are not covered here, as there are several reviews available on the subject (
3,
27). An important challenge in membrane protein research is to increase the temporal resolution of AFM in order to monitor dynamic processes (
28). In recent years, the Ando research group has made remarkable progress in developing new high-speed AFM instruments (
28). While the time required to record a high-resolution image with conventional AFMs is about 60 s, high-speed technology makes it possible to obtain 10 images per second. This enabled them to observe dynamic molecular processes in photoactivated bacteriorhodopsin, showing that illumination of this light-driven proton pump induces major structural changes within 1 s (
29). Also, high-speed AFM enabled the Scheuring team to track the motion of the outer membrane protein F (OmpF) from
E. coli (
30). High-resolution movies revealed that the proteins were widely distributed in the membrane as a result of diffusion-limited aggregation. Although the overall protein motion scaled with the local density of proteins in the membrane, individual protein molecules could also diffuse freely or become trapped by protein-protein interactions. From these data, they determined an interaction potential map and an interaction pathway for a membrane protein. Of note, the high-speed technology has also been applied to living bacteria, revealing the molecular dynamics of the cell surface (
31). The bacterial outer membrane was covered with a net-like structure with slowly diffusing holes, presumably reflecting porin trimers. Collectively, the above studies have contributed to better understanding of the structural organization of microbial constituents, including peptidoglycan, glycopolymers, and membrane proteins.
Cell wall remodeling.
Understanding how cell walls remodel in response to growth or to drugs and how such structural dynamics correlate with changes in biophysical properties are important topics in cellular microbiology. AFM imaging allows researchers to track dynamic structural changes, while force spectroscopy provides a means to correlate these changes with differences in cell wall rigidity. In their pioneering work, Plomp et al. used AFM to probe the high-resolution structural dynamics of single
Bacillus atrophaeus spores germinating under native conditions (
23). AFM images revealed previously unrecognized germination-induced alterations in spore coat architecture and topology as well as the disassembly of outer spore coat rodlet structures. Combined AFM-fluorescence imaging enabled us to visualize how the fungal pathogen
Candida albicans takes advantage of the yeast- to hyphal-phase transition to facilitate piercing and escape from phagocytes (
15). Besides growth, environmental stresses can also greatly alter the microbial cell wall. By way of example, the Dague team explored the effects of heat stress on the structural and mechanical properties of
Saccharomyces cerevisiae (
36). Heat stress induced the formation of circular rings on the cell surface and increased the cell wall stiffness with a concurrent increase in chitin content. Analysis of mutants suggested that the circular features reflect defective bud scars or bud emergence sites during temperature stress.
Many important antibiotics, including β-lactams (penicillin) and glycopeptides (vancomycin), target microbial cell walls. Owing to its ability to monitor drug-induced surface alterations in microbial pathogens, AFM has opened up new possibilities for understanding the mode of action of antibiotics and for screening new antimicrobial molecules capable to fight resistant strains (
37–42). Using real-time imaging, Francius et al. (
40) captured the structural dynamics of
S. aureus cells exposed to lysostaphin, an enzyme that specifically cleaves the peptidoglycan cross-linking pentaglycine bridges and that represents an interesting potential alternative to antibiotics. The enzyme induced major changes in cell surface morphology (swelling, splitting of the septum, and nanoscale perforations) and cell wall mechanics, which were attributed to the digestion of peptidoglycan, leading eventually to the formation of osmotically fragile cells. Similarly, the
P. aeruginosa cell wall was demonstrated to be structurally and biophysically affected at the nanoscale by two reference antibiotics, ticarcillin and tobramycin, with the cell wall stiffness decreasing dramatically after treatment (
41). In a related study, the effect of a polycationic calixarene-based guanidinium compound, CX1, on an
P. aeruginosa multidrug-resistant strain was investigated (
42). CX1 caused substantial alteration of the cell wall morphology (increased roughness and perforations) and a major drop in the cell wall stiffness. Further analysis of artificial membranes suggested that CX1 destroys the outer membrane of the bacteria. Treatment of
C. albicans with the antifungal agents flucytosine and amphotericin B led to perforation and deformation of the cell wall (
43). Greater cell wall damages were observed when the drugs were combined with allicin, an organic compound from garlic (
44). Caspofungin, a novel antifungal drug that targets the synthesis of cell wall β-1,3-
d-glucans, caused major morphological and structural alterations of the
C. albicans cell wall, which correlated with a change in the cell wall mechanical strength (
45,
46). Moreover, the drug induced the massive exposure of the cell adhesion protein Als1 on the cell surface and led to increased cell surface hydrophobicity, two features that triggered cell aggregation (
45).
The mode of action of antimicrobial peptides has also been examined (
47–51). Among these peptides, colistin is being used in combination with other antibiotics to treat and control chronic lung infections in cystic fibrosis patients. The mechanism of action seems to involve electrostatic interactions between cationic peptides and the outer membranes of Gram-negative bacteria. Supporting this view, AFM showed that various bacterial species treated with colistin have disrupted cell surfaces, increased stiffness, and decreased adhesive properties (
47–49). In contrast, treatment of
B. subtilis with the peptide trichokonin VI induced collapse of the cell wall, increased roughness, and caused a progressive decrease in cell stiffness (
50), suggesting that the leakage of intracellular materials is a possible mechanism of action. AFM was used to probe the interaction of chrysophsin-3 with
Bacillus anthracis in sporulated, germinated, and vegetative states (
51). Unlike sporulated and germinated cells, vegetative cells became stiffer after treatment, an effect attributed to loss of water content and cellular material from the cell due to disruption of the cell membrane.
These investigations indicate that AFM imaging has the potential to become an important tool in antimicrobial therapy and pharmacology. A key direction for future research is to improve the temporal resolution of the technique so that fast cell wall remodeling can be monitored (
28). Using the high-speed technology, Fantner et al. (
52) observed, in real time, the effect of the antimicrobial peptide CM15 on individual
E. coli cells. The results suggested that bacterial killing is a two-stage process consisting of an incubation phase, followed by an execution phase in which most of the damage is completed in less than a minute. In the future, it is anticipated that AFM, and more specifically high-speed imaging, will allow us to better understand the action mode of antimicrobial agents, including antibiotics, antimicrobial peptides, and innovative compounds like nanoparticles.