High-Precision Fitting Measurements of the Kinetics of Size Changes during Germination of Individual Bacillus Spores

Abstract

High-precision measurements of size changes of individual bacterial spores based on ellipse fitting to bright-field images recorded with a digital camera were employed to monitor the germination of Bacillus spores with a precision of ∼5 nm. To characterize the germination of individual spores, we recorded bright-field and phase-contrast images and found that the timing of changes in their normalized intensities coincided, so the bright-field images can be used to characterize spore size and refractility changes during germination. The major conclusions from this work were as follows. (i) The sizes of germinating B. cereus spores were nearly unchanged until T_release, the time of the completion of CaDPA (a 1:1 chelate of Ca²⁺ and dipicolinic acid [DPA]) release after addition of nutrient germinants. (ii) The minor axis of germinating B. cereus spores rapidly increased by ∼50 nm in a few seconds right after T_release, while the major axis was slightly decreased or unchanged. Both the minor and major axes remained unchanged for a further 30 to 45 s and then increased by 100 to 200 nm by T_lys, the time of completion of cortex lysis. (iii) Individual spores in a population showed significant heterogeneity in the timing of germination events, such as T_release and T_lys, but also variation in size changes during germination. (iv) Bacillus subtilis wild-type spores, B. subtilis spores lacking the cortex-lytic enzyme CwlJ, and wild-type Bacillus megaterium spores showed similar kinetics of size changes during nutrient germination. The size increases in germinating spores probably result from uptake of water and cortex lysis after completion of CaDPA release.

INTRODUCTION

Spores of various Bacillus species are metabolically dormant; can survive for many years (1–3); and are extremely resistant to heat, desiccation, radiation, and many toxic chemicals (4). The spores can also rapidly return to life in germination that can be triggered by a variety of agents, including specific nutrients; cationic surfactants, such as dodecylamine, exogenous CaDPA (a 1:1 chelate of Ca²⁺ and dipicolinic acid [DPA]); and high hydrostatic pressure (HP) (1, 5). Nutrients trigger germination by binding to germinant receptors (GRs) located in the inner spore membrane (1, 2). Stimulation of these GRs triggers the release of the spore core's large (∼10% of the spore dry weight) depot of CaDPA and its replacement by water in stage I of germination, and this triggers activation of the cortex-lytic enzymes (CLEs) CwlJ and SleB, either of which can initiate hydrolysis of the spore's peptidoglycan (PG) cortex, leading to completion of spore germination in stage II. Concomitant with cortex hydrolysis, the spore core becomes fully hydrated, and this allows resumption of enzyme activity and initiation of metabolism and macromolecular synthesis in the core, and thus spore outgrowth (1, 2). It is normally considered that in stage II, swelling of the spore core through further water uptake, together with the hydrolysis of the spore's peptidoglycan spore cortex and expansion of the germ cell wall, results in an increase in the spore's size (1). However, high-resolution noninvasive measurement of spore size during germination is challenging, and direct quantitative measurements of the size changes of germinating spores in real time have not been reported.

High-resolution measurements with optical imaging beyond the diffraction limit have undergone remarkable development in recent years (6–12). Indeed, high-precision single-molecule or particle localization based on the statistical fitting of point-spread functions (PSF) to measured fluorescence images of individual molecules can allow the determination of the position of a single molecule with an accuracy up to 1 nm or less (6, 7). The precise localization of photoswitchable single fluorescent molecules further allows the reconstruction of superresolution fluorescence images (8, 9). On the other hand, high-precision size measurement of an object or a cell can also be performed by fitting the image of each object to an ellipse so that the size (major and minor axes) and the position (x and y positions) and orientation of each object can be determined with an accuracy beyond the diffraction limit (10–12). Size changes (∼50 nm) of individual Bacillus thuringiensis spores in response to changes in relative humidity have been observed (10). It was found that the spores swell in response to increased relative humidity and shrink to near their original size on reexposure to dry air (10). Here, we report on high-precision measurements of size changes of individual Bacillus spores based on ellipse fitting to the optical images to monitor the dynamics of Bacillus spore germination with subwavelength precision. The objective of the work was to link size changes in germination with other germination events, including the release of a spore's CaDPA and cortex lysis.

Phase-contrast microscopy (13–15) and differential interference contrast (DIC) microscopy (16, 17) have been applied extensively to characterize the kinetics of germination of single Bacillus and Clostridium spores. Dormant spores appear bright in phase-contrast and DIC microscopy because of the high refractive index in the spore core due to its high levels of CaDPA and relatively low water content. However, when spores germinate, CaDPA is released and the peptidoglycan cortex is hydrolyzed so that the spore core's refractive index decreases and spores become phase dark. Raman spectroscopy and fluorescence microscopy have also been combined with phase-contrast microscopy and DIC microscopy to monitor the kinetics of the germination of individual Bacillus spores (18–20). This work has shown that the beginning and end of the rapid drop in the intensity of phase-contrast or DIC images (defined as T_lag [the lag time] and T_release [the CaDPA release time], respectively) precisely correspond to the beginning and completion of the release of CaDPA as revealed by Raman spectroscopy; the entry of fluorescent stains into the spore core also begins at T_release and continues until cortex hydrolysis is complete at T_lys (15, 20). However, due to the artifacts of Zernike phase-contrast and DIC images, precise measurements of size changes in germinating spores have not been reported.

In this work, we measured the sizes of single spores by fitting the bright-field images recorded with an electron magnifying charge-coupled-device (EMCCD) camera to an ellipse, so that the size of each spore could be measured. In order to increase the precision of size measurements, we averaged size measurements over 100 sequential images of the same object in ∼3 s and reached a precision of ∼5 nm. The advantages in choosing bright-field images for size measurements are that (i) the images have no artifacts of Zernike phase-contrast and DIC images, so they are suitable for size measurements, and (ii) the image intensities of single spores can also be used to characterize the germination process. We first established a kinetic coincidence between intensity changes in phase-contrast and bright-field images, showing that the normalized intensities of either type of image can be used to characterize germination. We then measured the kinetics of size changes of multiple individual Bacillus cereus, Bacillus subtilis (wild-type and cwlJ strains), and Bacillus megaterium spores during nutrient-triggered germination.

MATERIALS AND METHODS

Bacillus strains used and spore preparation.

The Bacillus species used in this work were (i) B. cereus T (originally obtained from H. O. Halvorson); (ii) B. megaterium QM B1551 (originally obtained from H. S. Levinson); (iii) B. subtilis PS533 (wild type), a prototrophic strain derived from strain 168 and carrying plasmid pUB110, encoding resistance to kanamycin (21); and (iv) B. subtilis FB111 (cwlJ), isogenic with strain PS533 but lacking plasmid pUB110 and also the cwlJ gene encoding the CLE CwlJ (22). Spores of these species and strains were prepared and stored as described previously (23–25). All spores used in this work were free (>98%) of growing or sporulating cells and germinated spores, as determined by phase-contrast microscopy.

Spore germination with a nutrient germinant.

Prior to germination, spores were heat activated in distilled water for 30 min at 70°C (B. subtilis), 30 min at 65°C (B. cereus), or 15 min at 60°C (B. megaterium) and then cooled on ice for >15 min prior to germination experiments. The germination conditions were as follows: B. subtilis, 37°C with 10 mM l-alanine in 25 mM potassium-(4-[2-hydroxyethyl]-1-piperazineethanesulfonic acid) (K-HEPES) buffer (pH 7.4); B. cereus, 37°C with 10 mM l-valine in 25 mM K-HEPES buffer (pH 7.4); and B. megaterium, 37°C with 10 mM d-glucose in 25 mM potassium phosphate buffer (pH 7.4). Briefly, 1 μl of heat-activated spores (10⁸ spores/ml in water) was spread on the surface of a glass coverslip glued to a clean and sterile sample container. The spores on the container were quickly dried in a vacuum chamber at room temperature so that they adhered to the coverslip. The spore container was then mounted on a microscope heating stage kept at 37°C. Preheated germination solution was then added to the container to start the germination (15, 18).

Bright-field and phase-contrast microscopy.

The germination of multiple individual spores was monitored by both bright-field and phase-contrast microscopy. An inverted microscope with an external phase-contrast system with a long-working-distance objective (Nikon 50×; 0.55 numerical aperture [NA]; 10-mm working distance) was used to image the individual spores (18, 19). A digital CCD camera (16 bit; 1,600 by 1,200 pixels) was used to record phase-contrast images at a rate of one frame per 15 s. These images were analyzed with a computation program in Matlab to locate each spore's position and to calculate the average pixel intensity of an area of 20 by 20 pixels that covered the whole individual spore on the phase-contrast image (5, 16). The phase-contrast image intensity of each individual spore was plotted as a function of the incubation time (with a resolution of 15 s), and the initial intensity (the first phase image recorded after the addition of germination solution) was normalized to 1 and the intensity at the end of the measurements was normalized to 0.

An EMCCD camera (Princeton Instruments; ProEM512B) was used to record the bright-field images of the individual objects at a rate of 34 frames per second. It took about 3 s to record 100 frames of bright-field images that were used to measure the size of one object at each time point, and the measurements of 100 frames of images were repeated at intervals of 15 s for 60 to 120 min.

For long-term (1- to 2-h) measurements of spore germination on a microscope slide, slow drift of the microscope stage in axial directions is a major concern. The axial drift was corrected by a homemade autolocking system based on the detection of slide-objective separation and feedback control of the movement of the objective using a piezo actuator (PiezoSystem Jenna, Inc., Hopedale, MA) to lock the distance between the objective and the slide. The measured long-term stability of the locking in the axial direction was less than 20 nm, and the automatic locking could last for more than 12 h (26).

Ellipse-fitting algorithm and relative size measurement.

In contrast to the phase-contrast images, in which dormant spores appear to be bright against a dark background, the bright-field images of spores appear to be dark against a bright background because the cells that have a slightly higher index of refraction than the surrounding medium deflect the illumination light off its original path and thus lead to less light arriving at the detector. The raw data for each frame of the bright-field images from the EMCCD camera were loaded and processed by a computation program in Matlab. The Matlab algorithm first inverted a spore's image (a dark object with a bright background) into a bright image (a bright object with a dark background) in order to compare these images with phase-contrast images. The Matlab algorithm then determined the center location of each individual object from the first inverted image frame based on a pixel intensity that was above a preset threshold. An area of 50 by 50 pixels that covered the whole individual object on the bright-field image was then fed into an ellipse-fitting algorithm (10, 12), which computed image brightness gradients of the object and selected the image's edge pixels from the brightness gradients that exceeded a given threshold value set at 50% of the maximum gradient value, thus producing a list of the edge pixel coordinates. This list was then organized and fitted to an ellipse so that five parameters (x and y positions, minor and major axes, and orientation) were recorded for each detected object (10). In addition, the summed intensity of each object can also be computed, and the size of each object is represented as the minor- and major-axis lengths. The absolute precision of a single measurement from one frame of bright-field image is ∼50 nm, as each measurement is computed from many pixels. By averaging over 100 image measurements on the same object, the precision in size measurement can be improved to be better than 5 nm (10).

Polystyrene spheres 0.99 μm in diameter (Bangs Laboratories Inc., Fishers, IN; PS03N/5750; mean diameter, 0.99 μm; standard deviation, 0.05 μm) were used for testing the ellipse-fitting algorithm and relative size measurements (see Results). The sizes of spores that were more elliptical in shape were also measured with the ellipse-fitting algorithm.

RESULTS

Ellipse image fitting of polystyrene spheres and dormant B. cereus spores.

Figure 1a shows one frame of bright-field images of a polystyrene sphere 0.99 μm in diameter acquired by the EMCCD camera, in which the sphere appears dark due to its larger index of refraction than that of the surrounding water. Figure 1b shows the extracted image edge pixels computed by the image process algorithm as described in Materials and Methods, which are fitted to an ellipse, as shown in Fig. 1c. As the result of the goodness of fit, Fig. 1d shows the ellipse fitting to the bright-field image of the polystyrene sphere, with major- and minor-axis lengths of 1.321 and 1.279 μm, respectively. By averaging over measurements of 100 image frames of the same sphere, the precision in the measurements of the major- and minor-axis lengths can be better than 5 nm. Figure 1e shows the measured mean major- and minor-axis lengths averaged over 100 image frames as a function of repeat measurement times, and the standard deviations of the mean major and minor axes of the sphere were 3.7 and 5.5 nm, respectively. The slight difference (<0.05 μm) in the measured major and minor axes suggests that the polystyrene bead is not perfectly spherical. The difference between the fitting size (1.32 μm in the major axis) of the bright-field images and the actual diameter of the polystyrene sphere (0.99 ± 0.05 μm) is most likely due to the diffraction limit, so that the convolution integral of the Airy disc function of a point object and the actual object makes the size of the bright-field image appear larger than the actual physical size of the object (10). However, the image size is dependent on the actual size of the object. The result in Fig. 1e indicates that repeated measurements of the image size of an object have a precision (∼5 nm) that is well below the diffraction limit (∼500 nm with our system), which could be very suitable for measuring changes in the size of biological objects during experiments (10).

FIG 1.

Ellipse image fitting of a 0.99-μm polystyrene sphere. (a) Bright-field image of the sphere. (b) Extracted image edge. (c) Ellipse fitting to the image edge. (d) Overlay of the ellipse on the bright-field image. (e) Major- and minor-axis values of 100 consecutive measurements of the same sphere. Each measurement of major and minor axes was the average of the fitting values of 100 images, all of which were acquired in 3 s. The scale bars are 1.0 μm.

To test the applicability of the ellipse image-fitting algorithm to biological cells, we measured the image size of a dormant B. cereus spore in water for 10 to 30 min, in which time one would expect no change in size. Each measurement took about 3 s to record the 100 frames of bright-field images, and the measurement was repeated every 15 s. Figure 2a to d shows a bright-field image of a B. cereus spore, the extracted image edge, ellipse fitting to the image edge, and the goodness of fit to one frame of the spore's images, respectively. Figure 2e shows the mean major- and minor-axis lengths averaged over 100 image frames as a function of repeat measurement times. The results indicate that the mean major- and minor-axis lengths of this B. cereus spore were 1.411 ± 0.016 μm and 0.907 ± 0.009 μm, respectively. The size change of the bright-field images over 25 min was less than ∼15 nm, suggesting that dormant B. cereus spores are relatively static, with no swelling or shrinking during incubation in distilled water.

FIG 2.

Ellipse image fitting of a dormant B. cereus spore. (a) Bright-field image of a B. cereus spore. (b) Extracted image edge. (c) Ellipse fitting to the image edge. (d) Overlay of the ellipse on the bright-field image. (e) Major- and minor-axis values of 100 consecutive measurements of the same spore. Each measurement of the major and minor axes was the average of the fitting values of 100 images, all of which were acquired in 3 s.

Effects of microscope defocusing on the size of image fitting.

Since the bright-field image of an object is sensitive to the focusing of the microscope, the size of ellipse image fitting may be affected by the focusing distance between the object position in the axial direction and the focusing plane of the microscope. Figure 3 shows the change in the major- and minor-axis lengths of a dormant B. subtilis spore adhered on a microscope coverslip as a function of the focusing distance that was adjusted by changing the voltage applied on the piezo. The results indicated that both the major and minor axes increased by >50 nm as the focusing distance was changed by ± 1.5 μm from the focus. However, when the focus distance was changed to within ± 0.75 μm from the focus, the change in size of bright-field images of a B. cereus spore was less than 20 nm. The focusing drift due to the size change in a germinating spore is estimated to be 0.1 to 0.2 μm (see Table 2), and this drift would not affect the size measurements. In our analyses of spore germination over 1 to 2 h, the long-term axial drift of the focus distance was locked at less than 20 nm with an active piezo-objective locking system (26).

FIG 3.

Major-axis (●) and minor-axis (◼) lengths of a dormant B. subtilis spore adhered on a microscope coverslip versus the distance from the focus without locking the focus. With an active piezo-objective locking system, the drift of the focus distance can be less than 20 nm.

TABLE 2.

Changes of minor and major axis lengths of Bacillus spores during germination^a

Species	Change in minor axis (change in major axis) (nm)
	Tlag	Trelease	Tlys
B. cereus	5.3 ± 24.5 (3.2 ± 12.0)	83.6 ± 68.6 (−18.6 ± 52.5)	271.2 ± 36.2 (157.2 ± 36.5)
B. megaterium	0.7 ± 2.5 (1.0 ± 2.0)	52.4 ± 45.6 (7.1 ± 2.7)	217.1 ± 58.3 (193.8 ± 62.7)
B. subtilis PS533 (wt)	−1.1 ± 3.8 (−5.5 ± 20.3)	82.1 ± 34.1 (2.6 ± 2.8)	194.0 ± 49.1 (149.8 ± 33.3)
B. subtilis FB111 (cwlJ)	−1.1 ± 7.8 (6.6 ± 17.1)	12.7 ± 76.2 (23.4 ± 35.0)	147.9 ± 49.0 (177.3 ± 58.3)

^aThe nutrient germination of heat-activated spores of B. cereus, B. megaterium, and B. subtilis strains PS533 (wild type), and FB111 (cwlJ) was followed for 60 min as described in Materials and Methods. The values for the changes in major- or minor-axis lengths are the difference between the major- or minor-axis lengths at various times in germination and the major- or minor-axis lengths at T₀. Data from >20 individual spores were used to calculate the germination parameters shown.

Kinetics of size changes in multiple individual B. cereus spores during l-valine germination.

The nutrient germination of multiple individual B. cereus spores was observed by both bright-field and phase-contrast microscopy. Figure 4a to d shows typical phase-contrast images and bright-field images of germinating B. cereus spores adhered on a coverslip, which were recorded 0 and 15 min after the addition of 10 mM l-valine in 25 mM K-HEPES buffer (pH 7.4). The arrows mark two germinated spores at each time point. All spores were measured so that their germination could be tracked, which allowed direct comparison of the intensities of phase-contrast and bright-field images of each spore at different times.

FIG 4.

Phase-contrast (a and b) and bright-field (c and d) images of germinating B. cereus spores adhered on a microscope coverslip. The images were recorded at 0 and 15 min after the addition of 10 mM l-valine in 25 mM K-HEPES buffer (pH 7.4). The arrows mark the germinated spores at each time point. The scale bars are 5 μm, and all images are at the same scale.

Figure 5a and b shows time-lapse phase-contrast and bright-field images of a single B. cereus spore at different times after the addition of 10 mM l-valine. The bright-field and phase-contrast image intensities were calculated by averaging the pixel intensities over a 20- by 20-pixel region that covered the whole spore. Figure 5c shows the normalized intensities of phase-contrast and bright-field images of the germinating spore as a function of incubation time. It was clear that changes in phase-contrast and bright-field image intensities were coincident. Therefore, although the phase-contrast image intensity has generally been used to characterize spore germination (13–15, 18), the bright-field image intensity (or darkness) can also be used to characterize the germination parameters T_lag, T_release, and T_lys. As a consequence, bright-field microscopy is appropriate for monitoring germination of individual spores and is very convenient to use. Figure 5d shows the major- and minor-axis lengths of the germinating individual B. cereus spore as a function of the germination time. The data in Fig. 5d indicate that (i) the size of the germinating spore was nearly constant until T_release, even though the normalized bright-field image intensity decreased by ∼70% and (ii) right after T_release, the minor axis of the germinating spore increased by ∼50 nm in a few seconds, while the major axis was slightly decreased or nearly unchanged. The minor and major axes remained unchanged for 30 to 45 s and then started to increase by 100 to 200 nm until T_lys.

FIG 5.

Changes in phase-contrast and DIC image intensities during B. cereus spore germination. (a and b) Time-lapse phase-contrast images (a) and bright-field images (b) of a single B. cereus spore at different times after the addition of l-valine. (c) Normalized intensities of phase-contrast and bright-field images of the germinating spore as a function of incubation time. (d) Major- and minor-axis lengths of the germinating spore as a function of incubation time.

Figure 6 shows the bright-field image intensities and the minor- and major-axis lengths of six individual B. cereus spores as a function of incubation time during germination. The arrows in each plot indicate T_lag, T_release, and T_lys for each individual spore. Although T_lag and T_release are very heterogeneous for individual spores, as observed previously (16–18), the variability in the sizes and size changes of the individual germinating B. cereus spore was relatively small. Thus, the average size of 20 individual spores was relatively uniform at T₀ (1.020 ± 0.057 μm [minor axis] and 1.539 ± 0.027 μm [major axis]) (Table 1), and the size change of individual B. cereus spores at T_lag was minimal (less than 5 nm, on average). However, the size change at T_release averaged approximately 84 nm in the minor axis and approximately −19 nm in the major axis, and at T_lys, the size changes in the minor and major axes averaged ∼271 and 157 nm, respectively (Tables 1 and 2).

FIG 6.

Changes in bright-field image intensities and minor- and major-axis lengths during germination of six individual B. cereus spores. Shown are the bright-field image intensities (squares) and the minor-axis (triangles) and major-axis (circles) lengths of six individual germinating B. cereus spores as a function of incubation time during germination with 10 mM l-valine in 25 mM K-HEPES buffer (pH 7.4) at 37°C. The arrows in each plot indicate, in temporal order, the lag time (T_lag), the CaDPA release time (T_release), and the cortex lysis time (T_lys) for each individual spore.

TABLE 1.

Mean values and standard deviations of minor-axis and major-axis lengths of Bacillus spores during germination^a

Species	Minor axis (major axis) (μm)				Trelease (min)
	T0	Tlag	Trelease	Tlys
B. cereus	1.020 ± 0.052 (1.539 ± 0.027)	1.026 ± 0.054 (1.542 ± 0.027)	1.104 ± 0.065 (1.521 ± 0.053)	1.292 ± 0.059 (1.697 ± 0.032)	4.5 ± 0.7
B. megaterium	1.394 ± 0.075 (1.593 ± 0.072)	1.395 ± 0.076 (1.594 ± 0.072)	1.443 ± 0.085 (1.599 ± 0.073)	1.611 ± 0.086 (1.786 ± 0.095)	5.6 ± 1.4
B. subtilis PS533 (wt)b	0.797 ± 0.033 (1.198 ± 0.060)	0.795 ± 0.033 (1.193 ± 0.070)	0.879 ± 0.049 (1.201 ± 0.060)	0.991 ± 0.068 (1.348 ± 0.062)	4.7 ± 1.6
B. subtilis FB111c (cwlJ)	0.788 ± 0.035 (1.115 ± 0.105)	0.787 ± 0.033 (1.121 ± 0.106)	0.801 ± 0.074 (1.138 ± 0.115)	0.936 ± 0.060 (1.292 ± 0.124)	42.5 ± 7.0

^aHeat-activated spores of various species were germinated for 60 min, and spore sizes were measured by bright-field microscopy as described in Materials and Methods. Data from >20 individual spores were used to calculate the germination parameters shown.

^bwt, wild type.

^cFor FB111 spores, T_lag was selected as the time at which the image intensity was dropped to 0.9 of its initial value and T_lys as the time at which the image intensity no longer changed.

Size changes of B. megaterium and B. subtilis spores during germination.

To determine if the size changes observed during B. cereus spore germination also took place during germination of spores of other species, we examined germinating B. megaterium and B. subtilis spores. Figure 7 shows the bright-field image intensities and the minor- and major-axis lengths of six individual germinating B. megaterium spores. The image size of multiple individual B. megaterium spores at T₀ was 1.394 ± 0.075 μm in the minor axis and 1.593 ± 0.072 μm in the major axis (Table 1), indicating that B. megaterium spores are more spherical than B. cereus spores. As seen with the germinating B. cereus spores, the size change of individual B. megaterium spores at T_lag was minimal (Table 2). However, at T_release, the minor axis increased by ∼52 nm while the major axis increased only ∼7 nm, but by T_lys, the minor and major axes had increased ∼217 nm and 194 nm, respectively (Tables 1 and 2).

FIG 7.

Changes in bright-field image intensities and minor- and major-axis lengths during germination of six individual B. megaterium spores. Shown are the bright-field image intensities (squares) and the minor-axis (triangles) and major-axis (circles) lengths of six individual germinating B. megaterium spores as a function of incubation time during germination with 10 mM glucose in 25 mM potassium phosphate buffer (pH 7.4) at 37°C. The arrows in each plot indicate T_lag T_release, and T_lys for each individual spore.

Figure 8 and Fig. 9 show the size changes of individual germinating wild-type and cwlJ B. subtilis spores. The cwlJ spores lack the CLE CwlJ and exhibit a very low rate of CaDPA release (the long value of ΔT_release [T_release − T_lag]), as seen previously (20) and in the current work (Fig. 9 and Table 1). The average sizes of the wild-type and cwlJ strain spores at T₀ were similar, and spores of both strains exhibited minimal size change at T_lag (Tables 1 and 2). However, the wild-type spores had a much greater size change at T_release than the cwlJ strain spores, although the axes of both wild-type and cwlJ strain B. subtilis spores increased by 100 to 200 nm by T_lys.

FIG 8.

Changes in bright-field image intensities and minor- and major-axis lengths during germination of six individual B. subtilis wild-type spores. Shown are the bright-field image intensities (squares) and the minor-axis (triangles) and major-axis (circles) lengths of six individual germinating B. subtilis wild-type (PS533) spores as a function of incubation time during germination with 10 mM l-valine in 25 mM K-HEPES buffer (pH 7.4) at 37°C. The arrows in each plot indicate T_lag, T_release, and T_lys for each individual spore.

FIG 9.

Changes in bright-field image intensities and minor- and major-axis lengths during germination of two individual B. subtilis cwlJ strain spores. Shown are the bright-field image intensities (squares) and the minor-axis (triangles) and major-axis (circles) lengths of two individual germinating B. subtilis cwlJ (FB111) spores as a function of incubation time with 10 mM l-valine in 25 mM K-HEPES buffer (pH 7.4) at 37°C. The arrows in each plot indicate T_release for each individual spore.

DISCUSSION

We have developed a method for high-precision measurement of size changes in individual bacterial spores based on ellipse fitting to bright-field images. Although the fitting size of the bright-field image is generally different from (usually larger than) the actual size of the spores due to the effect of the Airy disc, the image size is still suitable for monitoring size changes in spores or similar-size objects during experiments. These measurements of the image size of an object have high precision (∼5 nm for polystyrene beads and ∼15 nm for spores) and are well below the diffraction limit. To achieve 15-nm precision in size measurement, 100 images were acquired rapidly, and their fitting values were averaged. Application of ellipse fitting to bright-field images rather than scanned images, which were used before, allowed high-temporal-resolution measurements while achieving high precision in size measurements (10). With a frame rate of 34 Hz, a temporal resolution of ∼3 s was achieved in this work.

We employed this method to monitor the dynamics of the germination of spores of Bacillus species, and this work led to a number of conclusions, as follows. First, clearly, bright-field microscopy is appropriate for monitoring germination of individual spores. Although phase-contrast image intensity has generally been used to characterize spore germination (13–15, 18), changes in phase-contrast and inverted bright-field image intensities during germination coincided precisely. Therefore, the bright-field image intensity (or darkness) can also be used to characterize spore germination parameters, such as T_lag, T_release, and T_lys. Importantly, because of its simplicity and inexpensive instrumentation, bright-field microscopy should also be a useful methodology for quantitative size measurements in other biological systems. Second, size changes were similar in germinating spores of three Bacillus species, including the lack of size change at T_lag, the increase of 50 to 80 nm in minor-axis length approximately at T_release with minimal changes in the length of the major axis, and the further slower increases in both major- and minor-axis lengths of ∼150 to 270 nm beginning soon after T_release and ending at T_lys. Thus, similar events leading to these size changes appear to take place during germination of all three Bacillus species. A third conclusion is that the increase in minor-axis length during germination of cwlJ B. subtilis spores at approximately T_release was much smaller than in the wild-type spores examined, including B. subtilis, although increases in axis lengths between soon after T_release and T_lys were similar for both cwlJ strain and wild-type B. subtilis spores. In addition, the average times for the slower size increase between T_release and T_lys were also similar for wild-type and cwlJ strain B. subtilis spores (Fig. 8 and 9).

Given the relatively similar kinetics of size changes during various periods in the germination of spores of Bacillus species, obvious questions are, what is associated with these periods and what might cause the size changes seen? The lack of any size changes prior to T_lag is perhaps not surprising, since there is likely no cortex hydrolysis prior to T_lag, T_lag values are not affected by the presence or absence of CLEs, and there is little if any CaDPA release in T_lag. The subsequent increases in spore size at approximately T_release and between T_release and T_lys are temporally associated with events in and around the spore core, including CaDPA release and cortex hydrolysis, with at least the latter event leading to significant swelling of the spore core associated with significant water uptake. Given that the sizes measured in our system are those of the whole spore, it is not clear whether increases in spore size are a reflection of changes in the spores' outer layer alone or swelling of the core causes swelling of the spore's outer layers and thus the whole spore. We cannot answer this definitively at present, but it seems most likely that at a minimum it is core swelling that drives this process, since loss of significant amounts of the spore coat by mutation of the major assembly protein CotE has no significant effect on spore wet heat resistance, which is primarily dependent on the core water content (27, 28). However, it is certainly possible that there are changes in the coat associated with germination that facilitate the swelling of spores' outer layers in response to core swelling, and there is a recent report consistent with this idea (29). In addition, whole dormant spores can swell and contract significantly in both the major and minor axes, and this has been suggested to be due at least in part to changes in spore core water content (29, 30). It should be noted that the threshold value for ellipse fitting is normalized to 50% of the maximum image intensity that is reduced during germination. The measured image size is likely attributable to the CaDPA distribution in the core for a dormant spore, since the image intensity is dominated by the CaDPA before T_release, and the image size is likely attributable to the spore's envelope after complete CaDPA release.

If, as we suggest above, the changes in spore size seen during germination reflect primarily core swelling, then what causes the two phases of this swelling: phase 1 at approximately T_release and phase 2 between T_release and T_lys? Phase 1 is largely coincident with CaDPA release but must require some cortex hydrolysis by CwlJ, since the magnitude of the phase 1 size increase was minimal in cwlJ strain B. subtilis spores. Perhaps cortex hydrolysis by CwlJ sufficiently weakens the cortex so that CaDPA release is activated and this further activates CwlJ (31), and when most CaDPA has been released and the cortex has been sufficiently weakened, the core can expand a bit through its osmotic pressure. What is surprising is that this phase 1 expansion was asymmetric, as only the minor-axis length increased significantly. Again, we have no definitive explanation for this observation; it could be due to asymmetry in CwlJ's precise distribution in spores' outer layers, although at least CwlJ-green fluorescent protein (GFP) appears to be distributed uniformly around the spore (32). An alternative explanation for the asymmetric phase 1 expansion of the spore core during germination is that the cortex structure itself is asymmetric. There is known to be asymmetry in the cross-linking of the spore cortex, as cortex PG nearest the germ cell wall is significantly less cross-linked than other cortex layers (33, 34). However, it is not known if there is asymmetry in the cortex cross-linking along spores' major and minor axes. It could well be interesting in future work to examine spore size changes during germination of spores that do not exhibit the asymmetry in cortex cross-linking (33, 34).

In contrast to phase 1, it is somewhat easier to assign a cause to the swelling in phase 2, as the latter size change takes place between T_release and T_lys, when most cortex hydrolysis takes place (initiated by the two redundant cortex-lytic enzymes [CLEs] CwlJ and SleB) (20). CwlJ has been localized primarily to the outer edge of the spore cortex (35, 36). However, CwlJ-GFP accumulation in the spore is dependent on the spore coat morphogenic protein SafA, suggesting that CwlJ is associated with a spore's inner coat (32). In contrast to the location of CwlJ exterior to the cortex, SleB is predominantly at the interior side of the cortex, with much associated with the spores' inner membrane (36). As a consequence of the different locations of spores' two CLEs, during spore germination, the cortex is likely degraded simultaneously from both its outer and inner edges, resulting in the continual increase in core size seen between T_release and T_lys. Presumably, this core swelling either causes swelling of the whole spore directly or is concomitant with changes in spores' outer layers that allow these layers to swell in response to the swelling of the spore core once the cortex has been degraded and the germ cell wall has expanded.

It is evident from the discussion above that the use of the methodology described in this work to measure spore size precisely during germination has provided a new window on events that take place at crucial times in spore germination. Further analysis of these size changes during spore germination may also provide additional information.