Abstract
Cell surface carbohydrates are important to various bacterial activities and functions. It is well known that different types of Bacillus display heterogeneity of surface carbohydrate compositions, but detection of their presence, quantitation and estimation of variation at the single cell level have not been previously solved. Here, using atomic force microscopy (AFM)-based recognition force mapping coupled with lectin probes, the specific carbohydrate distributions of N-acetylglucosamine and mannose/glucose were detected, mapped and quantified on single B. cereus surfaces at the nanoscale across the entire cell. Further, the changes of the surface carbohydrate compositions from the vegetative cell to spore were shown. These results demonstrate AFM-based ‘recognition force mapping’ as a versatile platform to quantitatively detect and spatially map key bacterial surface biomarkers (such as carbohydrate compositions), and monitor in situ changes in surface biochemical properties during intracellular activities at the single cell level.
1. Introduction
Carbohydrates are important components of bacterial cell surfaces, with their compositions displaying heterogeneity across strains [1]. Cell surface carbohydrates play important roles in various bacterial activities and functions, including adhesion on surfaces, molecular recognition and providing defence mechanisms to guard against unfavourable environmental conditions [2–4]. The examination of carbohydrate compositions of bacteria has been used for taxonomic differentiation, identification of different genera and types and strains of bacteria [5,6] as well as reconstruction of culturing conditions [7]. The species- and strain-specific compositions and related bacterial functions make carbohydrates excellent biomarkers to identify potential vaccine antigens, for diagnostics, as well as for elucidation of the molecular basis for bacterial virulence and pathogenicity. However, our knowledge of the organization of these cell surface polysaccharides is still limited.
Members of the Bacillus cereus group including B. cereus, B. anthracis and B. thuringiensis are genetically related species [8]. The characterization and profiling of these bacilli has great fundamental and applied importance. For instance, B. anthracis, the causative agent of anthrax, and B. cereus, a food-poisoning organism, are two major pathogenic bacilli [9,10]. Owing to their ability to form spores in response to nutrient deprivation conditions, the profiles of both vegetative cells and spores for these bacilli have been of great interest [11]. Changes in the surface carbohydrate compositions on B. anthracis and B. cereus were investigated as surface biochemical markers for tracking the spore-forming process [12]. To date, several techniques have been applied to analyse the bacterial surface carbohydrate compositions, primarily involving the use of gas chromatography–mass spectrometry (GC–MS), ion-exchange chromatography, gel chromatography and high-performance liquid chromatography [1,13,14]. However, these approaches require large populations of cells for analysis, and extensive sample preparation using biochemical methods. For example, prior to GC–MS analysis, two steps are necessary: chemical extraction and release of carbohydrates from the cell walls, followed by various derivatization methods [5,11,15,16]. While some chromatography techniques are able to separate the native carbohydrates without requiring derivatization, additional mass spectral information is required for compound identification [17,18].
Critically however, these techniques as bulk scale tools are unable to directly clarify the density and distribution of specific carbohydrates on single, living cell surfaces at the micro- or nanoscale. The high sampling requirement for conventional carbohydrate analyses has limited utility on many types of environmental samples or in situ applications with a low target cell population. In particular, a prevailing challenge is in the area of determining the spatial location of carbohydrates on the cell wall [11]. Descriptions of cell wall carbohydrates can be useful for strain classification and development of diagnostic and vaccine applications [19]. This emphasizes the real need to develop a facile, fast and versatile platform to quantify the bacterial cell surface carbohydrate compositions at the single cell level, which can in turn provide new insights into the biochemical properties of these bacteria.
Atomic force microscopy (AFM) has rapidly emerged as an important tool widely used over the past couple of decades in bacterial research [20–22]. The unique advantage of AFM is the ability not only to characterize cellular surfaces with nanoscale resolution and three-dimensional imaging, but also to measure inter- and intramolecular interaction forces with piconewton sensitivity [23–25]. Using biomolecule-modified AFM tips as probes, interaction forces between tip-bound ligands and cognate surface-bound receptors (or vice versa) can be measured [26]. In recent years, the process of collecting such force data has been further expanded with the advent of automated scanning modes that allow us to rapidly obtain spatial distributions of interaction forces [27–29]. Importantly, bacterial samples for AFM can directly investigate live cells under near physiological environments in a non-destructive fashion. Elegant work from the Dufrêne laboratory has shown the versatility of this technique for detecting specific biomarkers on different kinds of cell surfaces [30–32]. An excellent review by the same group summarized the application of AFM-based multiparametric mapping technique in cellular systems [33]. However, to date, most studies have observed localized areas on cell surfaces (typically less than 0.5 × 0.5 μm) rather than on the entire cell surfaces.
In order to demonstrate a whole cell profiling strategy, we investigate bacterial cell surfaces and spatially map-specific carbohydrate profiles using AFM-based recognition mapping with B. cereus (T-strain) as a model bacterium. B. cereus is an important concern in the food industry and it also shares a genetic and structural similarity with the more virulent B. anthracis [10]. Using two specific carbohydrate binding lectins, wheat germ agglutinin (WGA) and concanavalin A (Con A), as probes, we show that the carbohydrates N-acetylglucosamine (GlcNAc) and glucose/mannose (Glu/Man) can be detected, mapped and quantified on single entire bacterial cell surfaces at the nanoscale. The choice of these targets was dictated by several studies showing the heterogeneity and changes in aminosugar and simple sugars across different strains and species, making them critical targets for profiling [1,7,11,12]. We show that these carbohydrates can be measured in both vegetative and spore states in liquid environments. The cell profiles in height maps show good consistency with high adhesion force signals in force maps and vitally, show a change from the vegetative to the spore state. Directly probing the cell surfaces can result in real-time analysis and the ability to observe the effect of external microenvironments at a single cell level. The cell surface itself is useful as a diagnostic target for carbohydrate-based vaccine antigens. This work demonstrates AFM-based ‘recognition force mapping’ as a versatile platform to quantitatively detect and map the bacterial surface biomarkers (such as carbohydrate compositions) and monitor the changes of surface biochemical properties during bacterial-related activities at the single cell level.
2. Experimental
2.1. Material and instrumentation
(1-Mercaptoundec-11-yl)hexaethylene glycol (oligoethylene glycol (OEG) terminated thiol), HS-C11-(EG)6OH, and (1-mercaptohexadecanoic acid)-N-succinimidyl ester (NHS terminated thiol), HS-C15COO-NHS, were purchased from Nanocs Inc. (Boston, MA). WGA, Con A, 3-aminopropyltriethoxysilane (APTES) and glutaraldehyde were purchased from Sigma-Aldrich (St Louis, MO). Phosphate-buffered saline (PBS pH 7.4; 11.9 mM phosphates, 137 mM sodium chloride and 2.7 mM potassium chloride) and ethanol (200-proof) were purchased from Fisher Scientific. Mica was purchased from Ted Pella (Redding, CA). Ultrapure water (resistivity 18.2 MΩ cm) was obtained from a MilliQ water purification system (Millipore Scientific, MA). AC240TS cantilevers (Olympus) were used for non-contact mode imaging, whereas gold-coated TR800PB cantilevers (Olympus) were used for force mapping. Cantilevers were cleaned using a UV/ozone procleaner (BioForce Nanosciences Inc., Ames, IA) before use. Imaging and force mapping experiments were performed on an MFP-3D atomic force microscope (Asylum Research, Santa Barbara, CA).
2.2. Cell culture and sporulation conditions
Cultures of vegetative B. cereus (T-strain) were maintained at 30°C on trypticase soy agar (30 g trypticase soy broth (211768, Becton–Dickinson, Franklin Lakes, NJ) and 15 g agar (AB1185, American BioAnalytical, Natick, MA)). Starter cultures were grown by inoculating single colonies of B. cereus into 125 ml of trypticase soy broth and incubating for 24 h at 30°C and 225 r.p.m. G medium was used as the base sporulation formulation [34], and supplemented with peptone (BP9725, Fisher Scientific, Pittsburgh, PA) and tryptone (61184, Acros Organics, Waltham, MA) (both at 8 g l−1). Sporulation was performed by adding 1 ml of starter culture into 250 ml of sporulation medium and incubating at 30°C and 275 r.p.m. in an orbital shaker.
2.3. Sample preparation for atomic force microscopy
Aminopropyl-mica (AP-mica) coupled with glutaraldehyde was used to immobilize cell samples. Freshly cleaved mica was incubated in APTES vapour in a vacuum desiccator for 12 h. The AP-mica was immersed in 1 ml of 2% (v/v) glutaraldehyde water solution for 1 h and rinsed with deionized water. A 20 µl cell suspension was spotted and incubated for 30 min. The sample was immersed in 1 ml of 1% (v/v) ethanolamine PBS solution for 30 min to block the glutaraldehyde on the areas without cells. Finally, the mica was rinsed with PBS to remove the loosely immobilized cells and unreacted ethanolamine.
For making cell suspensions, 1 ml cell suspension was taken from the culture medium and placed into a sterile 2 ml Eppendorf tube. Cells were collected by low-speed centrifugation (1 min, 4°C, 4000 r.p.m.). Each pellet was re-suspended and washed two times with 2 ml sterile distilled water. Finally, the pellet was re-suspended in 2 ml PBS buffer and kept in refrigerator (4°C) before use.
2.4. Atomic force microscopy probe functionalization and recognition force mapping
AC240TS cantilever (spring constant approx. 2 N m−1, resonance frequency 70 kHz) was cleaned using high-intensity UV light to remove organic contamination and used for imaging and characterization of the surfaces in non-contact mode. Gold-coated cantilevers were cleaned in UV/ozone for 15 min. Cantilevers were functionalized as described earlier [35] by immersion in mixed thiol solution (HS-C11-(EG)6OH and HS-C15COO-NHS) in ethanol for 16 h. Cantilevers were rinsed with ethanol, and incubated in a 100 nM solution of lectin (WGA or Con A) in PBS buffer for 1 h at ambient temperature. Our group has earlier shown that this functionalization strategy limits the number of lectins on the surface to 1–3 [36]. Spring constants of the functionalized cantilevers were measured using the thermal fluctuation method.
Subsequently, lectin-functionalized cantilevers (spring constant approx. 0.15 N m−1, resonance frequency 15 kHz) were used to obtain force data on cells. Each experiment was repeated on at least five different B. cereus cell surfaces to demonstrate reproducibility. Force–distance curves were obtained by collecting a series of sequential force curves in an m × n grid. Each force curve was obtained at the same loading rate (135 nN s−1, at a ramp velocity of 900 nm s−1) by pressing the cantilever to a low trigger point (200 pN), allowing binding to occur (contact time 0.1 s), and then retracting. All force maps were obtained by collecting approximately 80 × 80 force curves over a defined area (approx. 3 × 3 μm2), estimating the unbinding force values, and displaying these values by scale of colour. The height maps of the same area were generated simultaneously as the force mapping occurred. Blocking experiments were performed by a buffer exchange of PBS containing 5 μg ml−1 lectin in the fluid cell followed by force mapping. As a control experiment, a tip without lectin attached (only HS-C11-(EG)6OH) was used to map the cell surface. All images including height and force maps were performed using custom routines in Igor Pro v. 6.32 A (Wavemetrics Inc., OR) for fast processing of data.
3. Results and discussion
The intersection of AFM, with its capabilities of high-resolution imaging and force spectroscopy, with microbiology has provided an exciting strategy to study organisms both biophysically and biochemically [21,37]. The application of force spectroscopy for detecting specific recognition sites on cell surfaces and spatially mapping them has been an important step in this regard. For instance, Francius et al. [38] used two lectin-modified probes, Con A (glucose/mannose) and Pseudomonas aeruginosa lectin (galactose), to detect carbohydrates on Lactobacillus rhamnosus strain GG (LGG) surfaces. The force maps showed the Glu/Man and Gal distributions of the LGG wild-type are markedly different from those of the mutant strains. Andre et al. [31] performed recognition force mapping on Lactococcus lactis (LL) cells to detect specific peptidoglycans. The force maps combined with topographic images demonstrated that the peptidoglycans were arranged linearly and parallel to the short cell axis. However, these studies focused on small areas of the cells (400 × 400 nm2 on LGG [38] and 500 × 500 nm2 on LL cells [31]) with height and force channels obtained separately. These approaches zooming in smaller area are accurate for specific target detection and distribution mapping. However, for quantitative analysis of surface composition on single cells, mapping larger areas including entire cells can provide an efficient strategy for spatial co-localization and real-time analysis. For instance, consider that bacterial vegetative cells and spores are approximately two orders of magnitude larger in terms of surface area (3.4 ± 0.3 µm2 and 2.2 ± 0.3 µm2 for the B. cereus cells in this work). As we show in the following, observation of the entire cell morphology provides a unique opportunity to profile single, live cells for a range of targets.
3.1. Nanoscale imaging and probing of Bacillus
To observe the morphology and spatially locate carbohydrate markers using AFM, the first step involves proper sample preparation. Two requirements are needed: (i) cells should be firmly fixed on substrate to avoid dragging by an AFM tip, especially in liquid, and (ii) cells should be maintained in the native state to preserve lectin-reactive sites on the surface. Here, a simple chemical fixation method was used to immobilize the cell samples on substrate: AP-mica was modified with glutaraldehyde to form an aldehyde-terminated surface. Following aldehyde functionalization, the substrate was gently rinsed with sterile water to remove unreacted glutaraldehyde, which might damage the bioactivity of carbohydrates on cell surfaces. For morphology observation, both B. cereus vegetative cells and spores were initially imaged in air to show their high-resolution surface topography, including surface nanostructures (figure 1) [39]. Imaging was also performed in liquid environments (electronic supplementary material, figure S1). Figure 1a,b shows B. cereus vegetative cells are rod-shaped and form vegetative cell chains. A quantitative analysis was performed on several cell samples from different areas and images (n = 20). The dimensions of these vegetative cells were 2.8 ± 0.5 µm in length and 1.2 ± 0.2 µm in width, consistent with previously reported dimensions [40,41]. Compared with the rod-shaped vegetative cells, the B. cereus spore shows a characteristic ovoid shape (figure 1c,d). In addition to prior verification using phase contrast microscopy, several morphological characteristics typical of spores are observed by AFM imaging: (i) surface ridges extending along the entire spore length, a characteristic surface feature owing to the folding of the spore coat in the dehydrated state (marked by top and middle arrows in figure 1d) [42,43], (ii) an ultrathin capsule (approx. 20 nm) surrounding the spores, presumably the exosporium, another characteristic feature of B. cereus spores (marked by bottom arrow in figure 1d) and (iii) a significant length decrease following spore formation.
Figure 1. Bacillus cereus vegetative cells and spores in air. Panels (a) and (c) are height maps of vegetative cells and spore, respectively. Panels (b) and (d) are the corresponding amplitude images. (Online version in colour.)
3.2. Spatial recognition of surface carbohydrates on vegetative cells
To spatially recognize and map carbohydrates on B. cereus, force spectroscopy experiments were conducted in a liquid microenvironment using cantilevers functionalized with different lectins—Con A (which displays specificity to Glu/Man) and WGA (specific to the amino sugar GlcNAc). Figure 2 shows the height and force (interaction) maps of the cell surfaces for each lectin. For clarity, these images have been overlaid. Force curves were analysed automatically using a custom algorithm in IgorPro that can rapidly differentiate between specific binding and non-specific interactions and also determine binding forces. Curves classified as specific binding events are manifested in a cantilever deflection observed as a nonlinear delayed retraction curve with a different slope from that of the contact region [36]. Examples of the raw data of individual layers and representative force traces are shown in the electronic supplementary material, figure S2. An excellent correspondence can be observed between the cell profiles in the height map and a simultaneous increase of the recognition events at the same areas. In comparison, there was very low adhesion on the background. For WGA as the probe, figure 2b shows the cell profiles revealing correspondingly higher recognition events compared with cells probed with Con A (figure 2a). The non-overlaid raw images are presented in the electronic supplementary material, figure S3. Histogram analysis of force curves shows that the forces in the Con A–Glu system are higher than for the WGA–GlcNAc system, consistent with previous reports [44] (electronic supplementary material, figure S4). It can be seen that in Con A–Glu system, the main adhesion forces are around 120 pN and some are around 60 pN. This may be because at a pH > 7, ConA exists as a tetramer, in which four binding sites can act independently. In addition, the elasticity on single cell surface is not homogeneous, which may result in different contact area between the tip and surface at different locations. Thus, there are possibilities of multivalent interactions as observed [36]. The overall distribution is observed to be uniform across the surface and not concentrated at edges. The three-dimensional data obtained allow a rendering of the entire cell as shown in figure 3a. A further advantage of mapping the entire cell surface rather than zooming in on a smaller area on single cell surface is that the carbohydrate compositions of the cell surface can be quantitatively estimated (figure 3b). As a quantitative measure, the binding was calculated as a percentage of interaction traces to the total number of traces collected on the cell surface. This binding percentage is analogous to a molar concentration, because each interaction event represents the binding of a single (or double) molecule pair. The density of Glu/Man (9.8 ± 3.9%) is calculated to be much lower than that of GlcNAc (40.2 ± 7.2%) on B. cereus cell surfaces, where the data are from five independent experiments on different cells. To verify our results, we referenced earlier studies which analysed vegetative B. cereus cell walls. Comparison of surface carbohydrate compositions among several strains is summarized in table 1. Although different strains show the variation of carbohydrate compositions, it can be easily seen that the amount of GlcNAc is higher than that of Glu/Man in each strain with a variation from approximately 2 : 1 to approximately 10 : 1. Our calculated value of 4 : 1 on B. cereus strain T vegetative cells falls within this expected range. Among these different strains, the strains 03BB102 and G9241 show similar molar ratio (approx. 4 : 1) with B. cereus strain T in this study. Interestingly, our results are in excellent consistency with the reported values on B. cereus AHU 1356 strain, wherein the lysozyme digests of the cell wall showed that the polysaccharide component was composed of GlcNAc, ManNAc, GalNAc and Glc in ratios of 4 : 1:1 : 1 [45].
| strain | carbohydrate composition (% ± s.d.) | molar ratio Glu/Man : GlcNAc | ||
|---|---|---|---|---|
| Glc | Man | GlcNAc | ||
| ATCC 14579 | 27.7 ± 2.2 | n.d. | 45.2 ± 2.5 | 0.75 |
| F666 | 24.5 ± 7.9 | n.d. | 32.3 ± 6.5 | 0.93 |
| ATCC 10987 | 2.6 ± 1.3 | n.d. | 25.7 ± 3.0 | 0.12 |
| 03BB102 | 5.1 ± 0.8 | 0.9 ± 0.5 | 22.9 ± 2.7 | 0.27 |
| G9241 | 5.2 ± 0.7 | n.d. | 21.7 ± 1.0 | 0.29 |
| 03BB87 | 2.5 ± 0.9 | n.d. | 28.0 ± 3.2 | 0.11 |

Figure 2. Overlay of force and height profiles for the carbohydrates on B. cereus vegetative cell surfaces studied with their corresponding lectins. Two sets of experiments are presented. (a,c) Glu/Man probed with Con A. (b,d) GlcNAc probed with WGA. Typical force curves are also shown at the bottom of the figure (scale bars, 1 µm). (Online version in colour.)

Figure 3. (a) Three-dimensional rendering of a B. cereus vegetative cell showing Con A binding sites to Glu/Man. (b) Analysis of the rupture forces showing binding of lectins and target carbohydrates on B. cereus vegetative cells: 1, WGA on AFM tip/GlcNAc on cell surface; 2, Con A on AFM tip/Glu/Man on cell surface; 3, WGA on AFM tip/GlcNAc on cell surface blocked by WGA in buffer; 4, Con A on AFM tip/Glu/Man on cell surface blocked with Con A in buffer; 5, unmodified AFM tip (without lectin attached) on cell surface. Blocking and bare tip experiments show a decrease in binding (calculated as the ratio of the number of points showing recognition events to the total number of points collected on the cell surface). The data are from five independent experiments for each system.
Several control experiments were carried out to confirm the specificity of the pair-wise lectin–carbohydrate interactions. First, an unmodified (‘bare’) Au cantilever (functionalized in a similar manner with only OH–PEG–thiol modified but without attached lectin) was used to probe the cell surface. Minimal specific recognition events can be observed on the cell surfaces (figure 4a), implying that the observed recognition between the lectin-functionalized cantilever and the cell surfaces are specific in nature. Second, cell surfaces were mapped using lectin probes following the addition of relevant (free) lectins in buffer. This blocks the surface carbohydrates making them unlikely to show up when probed by the corresponding lectin. Figure 4b,c shows the force mapping of B. cereus cell surface by Con A and WGA-modified AFM probes, respectively. A significant decrease of high adhesion force areas is observed (quantitatively shown in figure 3), with a few non-specific high adhesion spots, confirming that the interactions observed are specific in nature and not the result of the lectin-modified tips binding non-specifically to the cell surface. These experiments further demonstrate the potential of this strategy to work in a complex medium in real time, by observing changes in the cell surface on adding exogenous substances to the microenvironment.
Figure 4. Control experiments on B. cereus cell surfaces. (a) Bare tip: overlaid height and force map of cell surface probed with a bare tip (without lectin attached). (b) Blocking experiment: height and force map of cell surface probed with Con A lectin in the presence of free Con A that blocks Glu/Man binding sites. (c) Blocking experiment: height and force map of cell surface probed with WGA lectin in the presence of free WGA that blocks GlcNAc binding sites. (Scale bars, 1 µm.) (Online version in colour.)
3.3. Spatial recognition of surface carbohydrates on spores
Following our study of vegetative cells, we investigated B. cereus spore surfaces. As a typical spore-forming cell, B. cereus is able to metabolically transform into oval, dormant cells in response to unfavourable environmental conditions [46]. It has been shown that vegetative B. cereus cells have different carbohydrate compositions in comparison with spores. These compositional biomarkers switches are important to potentially discriminate between vegetative or spore stages of Bacillus as well as to detect single cells [47]. Using recognition force mapping, GlcNAc and Glu/Man were spatially mapped on B. cereus spore surfaces. Figure 5 shows the spatial distribution of the carbohydrates on B. cereus spore surfaces. Rare projection spots on the height images may result from biopolymers that could have been pulled out by the AFM tip. The carbohydrate compositions of the spore surface can again be quantitatively estimated (figure 6). Similar control experiments including unmodified probes and blocking experiments were conducted. As expected, the binding probability decreased significantly, which further confirms the specificity of lectin–carbohydrate interactions on the spore surface as well (figure 6).
Figure 5. Overlay of force and height profiles for the carbohydrates on B. cereus spore surfaces studied with their corresponding lectins. Two sets of experiments are presented. (a,c) Glu/Man probed with Con A. (b,d) GlcNAc probed with WGA. Typical force curves are also shown at the bottom of the figure. (Scale bars, 1 µm.) (Online version in colour.) Figure 6. Analysis of the rupture forces showing binding of lectins and target carbohydrates on B. cereus spores: (a) WGA on AFM tip/GlcNAc on spore surface; (b) Con A on AFM tip/Glu/Man on spore surface; (c) WGA on AFM tip/GlcNAc on spore surface blocked by WGA in buffer; (d) Con A on AFM tip/Glu/Man on spore surface blocked with Con A in buffer. (e) Unmodified AFM tip (without lectin attached) on spore surface. Blocking and bare tip experiments show a significant decrease in binding.

From figure 5, the density of Glu/Man (45.8 ± 6.5%) is observed to be much higher than that of GlcNAc (14.8 ± 4.2%) on the spore surfaces. This trend which is opposite to the vegetative surface studied above is intriguing. Previous studies of spore carbohydrate compositions on different strains have shown that GlcNAc < Glu/Man for most strains although some variations can be observed from the vegetative cell to spore form [12,47]. Compared with the vegetative cells, the change of carbohydrate compositions in spores may be due to the lectin probe primarily interacting with the surface of the exosporium, known to envelop spores of B. cereus. Surface interactions on vegetative cells would also likely involve peptidoglycan components of the cell wall. We also do not rule out the interaction of the lectin Con A with teichoic acid, a significant component of Gram-positive bacilli cell walls. While this interaction is cross-reactive to Glu/Man, earlier works have shown that the ratio of teichoic acid and other components remains roughly the same in both vegetative cells and spores [48]. The exosporium is a complex mixture of carbohydrates, lipids and proteins with an overall composition that is distinct from the cell wall of vegetative cells [13]. In previous studies, the carbohydrates were released from whole spores using hydrolysis, such that carbohydrates both on the surface and inside the spore would be analysed. In contrast, AFM force mapping primarily probes the surface of spores. Compared with the vegetative cells, the new synthetic protective layers on spore surfaces during sporulation may lead to the change of carbohydrate compositions when detecting surface carbohydrates using AFM force mapping. Among these newly formed surface structures of spores, the outermost exosporium, a transparent loose-fitting membrane which is mainly composed of highly glycosylated glycoprotein BclA, known to envelop spores of B. cereus [49,50]. Our estimated carbohydrate composition value (GlcNAc : Glu/Man) is in excellent agreement with a previously reported value of 1 : 3 (molar ratio) measured from the B. cereus T-strain spore (same strain as this study) exosporium [13], in which only the outer exosporium was extracted and analysed.
Carbohydrate compositions obtained by recognition force mapping can therefore be used as significant parameters for tracking bacterial-related activities. An important consideration in whole cell analysis is the curvature of the cell which can result in artefacts at the cell–substrate interface. In these experiments, we calculated the aspect ratio (height/width) for a number of cells and spores (n = 10). For vegetative cells, aspect ratio is 0.30 ± 0.05, and for spores 0.66 ± 0.08. On areas such as the top of the cell, this problem is minimized. In the case of vegetative cells, there were also few issues at the cell substrate interface owing to the flatness of these cells. Here, we have investigated several cells in different orientations to consider these concerns. Probing the whole cell provides quantitative spatial distributions beyond yes/no answers. It should be noted that both GlcNAc and Glu/Man can be found on vegetative cells and spores, which implies that they are not necessarily standalone carbohydrate markers for all cell types. An interesting extension would be to detect spore-specific biomarkers and figure out presence/absence. An example would be rhamnose, the main sugar component of BclA which is spore-specific [51]. However, the rhamnose-binding lectin and antibody of BclA are currently not commercially available. The ability of this technique is therefore limited to targets that can be directly probed by specific binding ligands. However, several advantages including the uncovering of spatial, high-resolution and potential real-time information on whole, live cells under physiological environments is a powerful tool in the development of bacterial profiling strategies.
3.4. Resolution of recognition force mapping
In recent years, there has been rapid progress in developing new modes in AFM-based imaging starting with topography and recognition imaging. Among these, quantitative imaging and peak force tapping methods are able to image cell surface morphology and biophysical properties (e.g. elasticity and adhesion) simultaneously at high speed and high resolution [52,53]. The higher lateral resolution provided by peak force tapping has been found to be attractive when imaging smaller areas, including single proteins [53]. While these are extremely promising strategies, we note that there are advantages of the technique presented here: for instance, there is no requirement of any specialized equipment. The experiments presented were conducted on a standard, widely used AFM capable of performing force spectroscopy using regular, functionalized cantilevers.
In addition to accuracy and specificity, spatial resolution of this technique is a key parameter of recognition force mapping. For complex surfaces over whole cell areas, it is important to balance demands of spatial resolution and accuracy [54]. For example, increasing the spatial resolution can provide a better look at cell surfaces. However, force spectroscopy at two points too close to each other may not be optimal as surface targets such as carbohydrates are flexible polymers that may display multiple interactions. Moreover, an increase in data collection time causes further challenges including possible damage to the integrity and functionality of a biofunctionalized tip owing to repeated force curves, or thermal drift over long periods of time. The high level of control of scan size, scan speed and loading as reported in this paper results in an accurate estimation of surface composition. In these experiments, a ramp velocity of 0.9 μm s−1 was used, which reduces time of experiment while minimizing the effect of hydrodynamic drag force and hysteresis [55,56]. One complete force map on an area of approximately 4 × 4 µm with a resolution of 80 × 80 (6400 force curves) can be easily obtained in less than 1 h without drift, tip damage or multiple interactions. We envision that in conjunction with the biophysical tools and strategies discussed here, new instrumental tools can further add to the repertoire of ways to probe bacterial cells at the single cell level and in real time.
4. Conclusion
In summary, an AFM recognition force mapping strategy was demonstrated to quantitatively detect and spatially map surface carbohydrates on vegetative and spore cells of B. cereus at the single cell level. By calculating the binding using different lectin probes, WGA and ConA, we show that surface molar ratios of Glu/Man : GlcNAc range approximately 1 : 4 on a vegetative cell but switch to approximately 3 : 1 on a spore. These values are in excellent quantitative agreement with previously reported bulk analyses using GC–MS, while providing clear spatial distributions on the cell surface. Control experiments were used to confirm the specificity of the lectin–carbohydrate interactions. Taken together, the correlations between the force and height maps clearly show that recognition force mapping can spatially locate specific targets on entire cell surfaces. One of the significant advantages of the recognition force mapping strategy described here is the simultaneous collection of both forces and height data which renders the analysis of the single spore quickly (each image can be obtained and analysed in less than 1 h). AFM-based ‘recognition force mapping’ is shown to be a versatile platform to quantitatively detect and map bacterial surface biomarkers, specifically carbohydrate compositions, but easily extendable to a host of cell-surface targets, while monitoring changes in surface biochemical properties during intracellular processes at the single cell level.
Acknowledgement
The authors thank Cristina Stanciu for help with cell and spore culture.
Funding statement
This research was partially supported by a grant from the VCU PeRQ fund.
Footnotes
References
- 1
Fox A, Stewart GC, Waller LN, Fox KF, Harley WM& Price RL . 2003Carbohydrates and glycoproteins of Bacillus anthracis and related bacilli: targets for biodetection. J. Microbiol. Methods 54, 143–152. (doi:10.1016/S0167-7012(03)00095-2). Crossref, PubMed, ISI, Google Scholar - 2
Kolter R& Greenberg EP . 2006Microbial sciences: the superficial life of microbes. Nature 441, 300–302. (doi:10.1038/441300a). Crossref, PubMed, ISI, Google Scholar - 3
Weintraub A . 2003Immunology of bacterial polysaccharide antigens. Carbohydr. Res. 338, 2539–2547. (doi:10.1016/j.carres.2003.07.008). Crossref, PubMed, ISI, Google Scholar - 4
Ofek I& Sharon N . 1988Lectinophagocytosis: a molecular mechanism of recognition between cell-surface sugars and lectins in the phagocytosis of bacteria. Infect. Immun. 56, 539–547. Crossref, PubMed, ISI, Google Scholar - 5
Choudhury B, Leoff C, Saile E, Wilkins P, Quinn CP, Kannenberg EL& Carlson RW . 2006The structure of the major cell wall polysaccharide of Bacillus anthracis is species-specific. J. Biol. Chem. 281, 27 932–27 941. (doi:10.1074/jbc.M605768200). Crossref, ISI, Google Scholar - 6
Rao SS, Mohan KVK& Atreya CD . 2010Detection technologies for Bacillus anthracis: prospects and challenges. J. Microbiol. Methods 82, 1–10. (doi:10.1016/j.mimet.2010.04.005). Crossref, PubMed, ISI, Google Scholar - 7
Colburn HA, Wunschel DS, Antolick KC, Melville AM& Valentine NB . 2011The effect of growth medium on B. anthracis Sterne spore carbohydrate content. J. Microbiol. Methods 85, 183–189. (doi:10.1016/j.mimet.2011.02.015). Crossref, PubMed, ISI, Google Scholar - 8
Helgason E, Okstad OA, Caugant DA, Johansen HA, Fouet A, Mock M, Hegna I& Kolsto A-B . 2000Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis: one species on the basis of genetic evidence. Appl. Environ. Microbiol. 66, 2627–2630. (doi:10.1128/aem.66.6.2627-2630.2000). Crossref, PubMed, ISI, Google Scholar - 9
Liu YL, Elsholz B, Enfors SO& Gabig-Ciminska M . 2007Confirmative electric DNA array-based test for food poisoning Bacillus cereus. J. Microbiol. Methods 70, 55–64. (doi:10.1016/j.mimet.2007.03.011). Crossref, PubMed, ISI, Google Scholar - 10
Ivanova N, 2003Genome sequence of Bacillus cereus and comparative analysis with Bacillus anthracis. Nature 423, 87–91. (doi:10.1038/nature01582). Crossref, PubMed, ISI, Google Scholar - 11
Leoff C, Saile E, Sue D, Wilkins P, Quinn CP& Carlson RW . 2008Cell wall carbohydrate compositions of strains from the Bacillus cereus group of species correlate with phylogenetic relatedness. J. Bacteriol. 190, 112–121. (doi:10.1128/JB.01292-07). Crossref, PubMed, ISI, Google Scholar - 12
Fox A, Black GE, Fox K& Rostovtseva S . 1993Determination of carbohydrate profiles of Bacillus anthracis and Bacillus cereus including identification of O-methyl methylpentoses by using gas chromatography-mass spectrometry. J. Clin. Microbiol. 31, 887–894. Crossref, PubMed, ISI, Google Scholar - 13
Matz LL, Beaman TC& Gerhardt P . 1970Chemical composition of exosporium from spores of Bacillus cereus. J. Bacteriol. 101, 196–201. Crossref, PubMed, ISI, Google Scholar - 14
Amano K, Hazama S, Araki Y& Ito E . 1977Isolation and characterization of structural components of Bacillus cereus AHU 1356 cell walls. Eur. J. Biochem. 75, 513–522. (doi:10.1111/j.1432-1033.1977.tb11552.x). Crossref, PubMed, Google Scholar - 15
Brown WC . 1973Rapid methods for extracting autolysins from Bacillus subtilis. Appl. Microbiol. 25, 295–300. Crossref, PubMed, Google Scholar - 16
Fox A, 1990Chemotaxonomic differentiation of legionellae by detection and characterization of aminodideoxyhexoses and other unique sugars using gas-chromatography mass-spectrometry. J. Clin. Microbiol. 28, 546–552. Crossref, PubMed, ISI, Google Scholar - 17
Bruggink C, Maurer R, Herrmann H, Cavalli S& Hoefler F . 2005Analysis of carbohydrates by anion exchange chromatography and mass spectrometry. J. Chromatogr. A 1085, 104–109. (doi:10.1016/j.chroma.2005.03.108). Crossref, PubMed, ISI, Google Scholar - 18
Guignard C, Jouve L, Bogeat-Triboulot MB, Dreyer E, Hausman JF& Hoffmann L . 2005Analysis of carbohydrates in plants by high-performance anion-exchange chromatography coupled with electrospray mass spectrometry. J. Chromatogr. A 1085, 137–142. (doi:10.1016/j.chroma.2005.05.068). Crossref, PubMed, ISI, Google Scholar - 19
Grandbois M, Dettmann W, Benoit M& Gaub HE . 2000Affinity imaging of red blood cells using an atomic force microscope. J. Histochem. Cytochem. 48, 719–724. (doi:10.1177/002215540004800516). Crossref, PubMed, ISI, Google Scholar - 20
Muller DJ& Dufrene YF . 2011Atomic force microscopy: a nanoscopic window on the cell surface. Trends Cell Biol. 21, 461–469. (doi:10.1016/j.tcb.2011.04.008). Crossref, PubMed, ISI, Google Scholar - 21
Dupres V, Alsteens D, Andre G& Dufrene YF . 2010Microbial nanoscopy: a closer look at microbial cell surfaces. Trends Microbiol. 18, 397–405. (doi:10.1016/j.tim.2010.06.004). Crossref, PubMed, ISI, Google Scholar - 22
Dufrene YF . 2008Towards nanomicrobiology using atomic force microscopy. Nat. Rev. Microbiol. 6, 674–680. (doi:10.1038/nrmicro1948). Crossref, PubMed, ISI, Google Scholar - 23
Florin EL, Moy VT& Gaub HE . 1994Adhesion forces between individual ligand-receptor pairs. Science 264, 415–417. (doi:10.1126/science.8153628). Crossref, PubMed, ISI, Google Scholar - 24
Jalili N& Laxminarayana K . 2004A review of atomic force microscopy imaging systems: application to molecular metrology and biological sciences. Mechatronics 14, 907–945. (doi:10.1016/j.mechatronics.2004.04.005). Crossref, ISI, Google Scholar - 25
Wang CZ& Yadavalli VK . 2014Investigating biomolecular recognition at the cell surface using atomic force microscopy. Micron 60, 5–17. (doi:10.1016/j.micron.2014.01.002). Crossref, PubMed, ISI, Google Scholar - 26
Clausen-Schaumann H, Seitz M, Krautbauer R& Gaub HE . 2000Force spectroscopy with single bio-molecules. Curr. Opin. Chem. Biol. 4, 524–530. (doi:10.1016/s1367-5931(00)00126-5). Crossref, PubMed, ISI, Google Scholar - 27
De Pablo PJ, Colchero J, Gomez-Herrero J, Baro AM, Schaefer DM, Howell S, Walsh B& Reifenberger R . 1999Adhesion maps using scanning force microscopy techniques. J. Adhes. 71, 339–356. (doi:10.1080/00218469908014547). Crossref, ISI, Google Scholar - 28
Kienberger F, Ebner A, Gruber HJ& Hinterdorfer P . 2006Molecular recognition imaging and force spectroscopy of single biomolecules. Acc. Chem. Res. 39, 29–36. (doi:10.1021/ar050084m). Crossref, PubMed, ISI, Google Scholar - 29
Almqvist N, Bhatia R, Primbs G, Desai N, Banerjee S& Lal R . 2004Elasticity and adhesion force mapping reveals real-time clustering of growth factor receptors and associated changes in local cellular rheological properties. Biophys. J. 86, 1753–1762. (doi:10.1016/s0006-3495(04)74243-5). Crossref, PubMed, ISI, Google Scholar - 30
Alsteens D, 2013Unraveling the nanoscale surface properties of chitin synthase mutants of Aspergillus fumigatus and their biological implications. Biophys. J. 105, 320–327. (doi:10.1016/j.bpj.2013.05.040). Crossref, PubMed, ISI, Google Scholar - 31
Andre G, Kulakauskas S, Chapot-Chartier M-P, Navet B, Deghorain M, Bernard E, Hols P& Dufrêne YF . 2010Imaging the nanoscale organization of peptidoglycan in living Lactococcus lactis cells. Nat. Commun. 1, 27. (doi:10.1038/ncomms1027). Crossref, PubMed, ISI, Google Scholar - 32
Beaussart A, Alsteens D, El-Kirat-Chatel S, Lipke PN, Kucharikova S, Van Dijck P& Dufrêne YF . 2012Single-molecule imaging and functional analysis of Als adhesins and mannans during Candida albicans morphogenesis. ACS Nano 6, 10 950–10 964. (doi:10.1021/nn304505s). Crossref, ISI, Google Scholar - 33
Dufrene YF, Martinez-Martin D, Medalsy I, Alsteens D& Muller DJ . 2013Multiparametric imaging of biological systems by force–distance curve-based AFM. Nat. Methods 10, 847–854. (doi:10.1038/nmeth.2602). Crossref, PubMed, ISI, Google Scholar - 34
Hornstra LM, de Vries YP, de Vos WM& Abee T . 2006Influence of sporulation medium composition on transcription of Ger operons and the germination response of spores of Bacillus cereus ATCC 14579. Appl. Environ. Microbiol. 72, 3746–3749. (doi:10.1128/aem.72.5.3746-3749.2006). Crossref, PubMed, ISI, Google Scholar - 35
Yadavalli VK, Forbes JG& Wang K . 2006Functionalized self-assembled monolayers on ultraflat gold as platforms for single molecule force spectroscopy and imaging. Langmuir 22, 6969–6976. (doi:10.1021/la060320h). Crossref, PubMed, ISI, Google Scholar - 36
Zhang XJ& Yadavalli VK . 2009Functionalized self-assembled monolayers for measuring single molecule lectin–carbohydrate interactions. Anal. Chim. Acta 649, 1–7. (doi:10.1016/j.aca.2009.07.027). Crossref, PubMed, ISI, Google Scholar - 37
Dufrene YF . 2004Using nanotechniques to explore microbial surfaces. Nat. Rev. Microbiol. 2, 451–460. (doi:10.1038/nrmicro905). Crossref, PubMed, ISI, Google Scholar - 38
Francius G, 2008Detection, localization, and conformational analysis of single polysaccharide molecules on live bacteria. ACS Nano 2, 1921–1929. (doi:10.1021/Nn800341b). Crossref, PubMed, ISI, Google Scholar - 39
Gillis A, Dupres V, Delestrait G, Mahillon J& Dufrene YF . 2012Nanoscale imaging of Bacillus thuringiensis flagella using atomic force microscopy. Nanoscale 4, 1585–1591. (doi:10.1039/c1nr11161b). Crossref, PubMed, ISI, Google Scholar - 40
Fernandes JC, Eaton P, Gomes AM, Pintado ME& Malcata FX . 2009Study of the antibacterial effects of chitosans on Bacillus cereus (and its spores) by atomic force microscopy imaging and nanoindentation. Ultramicroscopy 109, 854–860. (10.1016/j.ultramic.2009.03.015). Crossref, PubMed, ISI, Google Scholar - 41
Pan JH, Ge XP, Liu RX& Tang HX . 2006Characteristic features of Bacillus cereus cell surfaces with biosorption of Pb(II) ions by AFM and FT-IR. Colloid Surf. B 52, 89–95. (10.1016/j.colsurfb.2006.05.016). Crossref, PubMed, ISI, Google Scholar - 42
Plomp M, Leighton TJ, Wheeler KE& Malkin AJ . 2005Architecture and high-resolution structure of Bacillus thuringiensis and Bacillus cereus spore coat surfaces. Langmuir 21, 7892–7898. (doi:10.1021/la050412r). Crossref, PubMed, ISI, Google Scholar - 43
Chen X, Mahadevan L, Driks A& Sahin O . 2014Bacillus spores as building blocks for stimuli-responsive materials and nanogenerators. Nat. Nanotechnol. 9, 137–141. (doi:10.1038/Nnano.2013.290). Crossref, PubMed, ISI, Google Scholar - 44
El-Kirat-Chatel S, Beaussart A, Alsteens D, Sarazin A, Jouault T& Dufrene YF . 2013Single-molecule analysis of the major glycopolymers of pathogenic and non-pathogenic yeast cells. Nanoscale 5, 4855–4863. (doi:10.1039/C3nr00813d). Crossref, PubMed, ISI, Google Scholar - 45
Murazumi N, Araki Y& Ito E . 1986Biosynthesis of the wall neutral polysaccharide in Bacillus cereus AHU-1356. Eur. J. Biochem. 161, 51–59. (doi:10.1111/j.1432-1033.1986.tb10123.x). Crossref, PubMed, Google Scholar - 46
Plomp M, Leighton TJ, Wheeler KE& Malkin AJ . 2005The high-resolution architecture and structural dynamics of Bacillus spores. Biophys. J. 88, 603–608. (doi:10.1529/biophysj.104.049312). Crossref, PubMed, ISI, Google Scholar - 47
Wunschel D, Fox KF, Black GE& Fox A . 1995Discrimination among the Bacillus cereus group, in comparison to Bacillus subtilis, by structural carbohydrate profiles and ribosomal-RNA spacer region PCR. Syst. Appl. Microbiol. 17, 625–635. (doi:10.1016/S0723-2020(11)80085-8). Crossref, ISI, Google Scholar - 48
Boylen CW& Ensign JC . 1968Ratio of teichoic acid and peptidoglycan in cell walls of Bacillus subtilis following spire germination and during vegetative growth. J. Bacteriol. 96, 421–427. Crossref, PubMed, ISI, Google Scholar - 49
Lequette Y, Garenaux E, Tauveron G, Dumez S, Perchat S, Slomianny C, Lereclus D, Guerardel Y& Faille C . 2011Role played by exosporium glycoproteins in the surface properties of Bacillus cereus spores and in their adhesion to stainless steel. Appl. Environ. Microbiol. 77, 4905–4911. (doi:10.1128/aem.02872-10). Crossref, PubMed, ISI, Google Scholar - 50
Sylvestre P, Couture-Tosi E& Mock M . 2002A collagen-like surface glycoprotein is a structural component of the Bacillus anthracis exosporium. Mol. Microbiol. 45, 169–178. (doi:10.1046/j.1365-2958.2000.03000.x). Crossref, PubMed, ISI, Google Scholar - 51
Daubenspeck JM, Zeng HD, Chen P, Dong SL, Steichen CT, Krishna NR, Pritchard DG& Turnbough CL . 2004Novel oligosaccharide side chains of the collagen-like region of BclA, the major glycoprotein of the Bacillus anthracis exosporium. J. Biol. Chem. 279, 30 945–30 953. (doi:10.1074/jbc.M401613200). Crossref, ISI, Google Scholar - 52
Alsteens D, Dupres V, Yunus S, Latge JP, Heinisch JJ& Dufrene YF . 2012High-resolution imaging of chemical and biological sites on living cells using peak force tapping atomic force microscopy. Langmuir 28, 16 738–16 744. (doi:10.1021/la303891j). Crossref, ISI, Google Scholar - 53
Pfreundschuh M, Alsteens D, Hilbert M, Steinmetz MO& Muller DJ . 2014Localizing chemical groups while imaging single native proteins by high-resolution atomic force microscopy. Nano Lett. 14, 2957–2964. (doi:10.1021/nl5012905). Crossref, PubMed, ISI, Google Scholar - 54
Ando T, Uchihashi T& Scheuring S . 2014Filming biomolecular processes by high-speed atomic force microscopy. Chem. Rev. 114, 3120–3188. (doi:10.1021/Cr4003837). Crossref, PubMed, ISI, Google Scholar - 55
Cappella B& Dietler G . 1999Force–distance curves by atomic force microscopy. Surf. Sci. Rep. 34, 1–104. (doi:10.1016/S0167-5729(99)00003-5). Crossref, ISI, Google Scholar - 56
Lo YS, Zhu YJ& Beebe TP . 2001Loading-rate dependence of individual ligand–receptor bond-rupture forces studied by atomic force microscopy. Langmuir 17, 3741–3748. (doi:10.1021/la001569g). Crossref, ISI, Google Scholar


