Germinant receptor proteins cluster in ‘germinosomes’ in the inner membrane of Bacillus subtilis spores. We describe a protocol using super resolution microscopy and fluorescent reporter proteins to visualize germinosomes. The protocol also identifies spore inner membrane domains that are preferentially stained with the membrane dye FM4-64.
The small size of spores and the relatively low abundance of germination proteins, cause difficulties in their microscopic analyses using epifluorescence microscopy. Super-resolution three-dimensional Structured Illumination Microscopy (3D-SIM) is a promising tool to overcome this hurdle and reveal the molecular details of the process of germination of Bacillus subtilis (B. subtilis) spores. Here, we describe the use of a modified SIMcheck (ImageJ)-assistant 3D imaging process and fluorescent reporter proteins for SIM microscopy of B. subtilis spores’ germinosomes, cluster(s) of germination proteins. We also present a (standard)3D-SIM imaging procedure for FM4-64 staining of B. subtilis spore membranes. By using these procedures, we obtained unsurpassed resolution for germinosome localization and show that >80% of B. subtilis KGB80 dormant spores obtained after sporulation on defined minimal MOPS medium have one or two GerD-GFP and GerKB-mCherry foci. Bright foci were also observed in FM4-64 stained spores’ 3D-SIM images suggesting that inner membrane lipid domains of different fluidity likely exist. Further studies that use double labeling procedures with membrane dyes and germinosome reporter proteins to assess co-localization and thus get an optimal overview of the organization of Bacillus germination proteins in the inner spore membrane are possible.
Spores of the orders Bacillales and Clostridiales are metabolically dormant and extraordinarily resistant to harsh decontamination regimes, but unless they germinate, cannot cause deleterious effects in humans1. In nutrient germinant triggered germination of Bacillus subtilis (B. subtilis) spores, the initiation event is germinant binding to germinant receptors (GRs) located in the spore’s inner membrane (IM). Subsequently, the GRs transduce signals to the SpoVA channel protein also located in the IM. This results in the onset of the exchange of spore core pyridine-2,6-dicarboxylic acid (dipicolinic acid; DPA; comprising 20% of spore core dry wt) for water via the SpoVA channel. Subsequently, the DPA release triggers the activation of cortex peptidoglycan hydrolysis, and additional water uptake follows2,3,4. These events lead to mechanical stress on the coat layers, its subsequent rupture, the onset of outgrowth and, finally, vegetative growth. However, the exact molecular details of the germination process are still far from resolved.
A major question about spore germination concerns the biophysical properties of the lipids surrounding the IM germination proteins as well as the IM SpoVA channel proteins. This largely immobile IM lipid bilayer is the main permeability barrier for many small molecules, including toxic chemical preservatives, some of which exert their action in the spore core or vegetative cell cytoplasm5,6. The IM lipid bilayer is likely in a gel state, although there is a significant fraction of mobile lipids in the IM5. The spore’s IM also has the potential for significant expansion5. Thus, the surface area of the IM increases 1.6-fold upon germination without additional membrane synthesis and is accompanied by the loss of this membrane’s characteristic low permeability and lipid immobility5,6.
While the molecular details of the activation of germination proteins and organization of IM lipids in spores are attractive topics for study, the small size of B. subtilis spores and the relatively low abundance of germination proteins, pose a challenge to microscopic analyses. Griffiths et al. compelling epifluorescence microscope evidence, using fluorescent reporters fused to germination proteins, suggests that in B. subtilis spores the scaffold protein GerD organizes three GR subunits (A, B and C) for the GerA, B and K GRs, in a cluster7. They coined the term ‘germinosome’ for this cluster of germination proteins and described the structures as ~300 nm large IM protein foci8. Upon initiation of spore germination, fluorescent germinosome foci ultimately change into larger disperse fluorescent patterns, with >75% of spore populations displaying this pattern in spores germinated for 1 h with L-valine8. Note that the paper mentioned above used averaged images from dozens of consecutive fluorescent pictures, to gain statistical power and overcome the hurdle of low fluorescent signals observed during imaging. This visualization of these structures in bacterial spores was at the edge of what is technically feasible with classical microscopic tools and neither an evaluation of the amount of foci in a single spore nor their more detailed subcellular localization was possible with this approach.
Here, we demonstrate the use of Structured Illumination Microscopy (SIM) to obtain a detailed visualization and quantification of the germinosome(s) in spores of B. subtilis, as well as of their IM lipid domains9. The protocol also contains instructions for the sporulation, slide preparation and image analysis by SIMcheck (v1.0, an imageJ plugin) as well as ImageJ10,11,12.
1. B. subtillis Sporulation (Timing: 7 Days Before Microscopic Observation)
2. Decoating
3. Coverslip and Slide Preparation11 (Timing: 1 H Before Observation)
4. Sampling Fluorescent Microspheres or Spores in the Gene Frame Slide10 (Timing 15 Min)
5. Imaging11,17(Timing: 1 H)
6. Reconstruct 3D-SIM Raw Images of FM4-64 Stained PS4150 Spores
7. Image Analysis
The current protocol presents a SIM microscope imaging procedure for bacterial spores. The sporulation and slide preparation procedures were carried out as shown in Figure 1 before imaging. Later, the imaging and analysis procedures were applied both for dim (fluorescent protein labeled germination proteins) and bright (lipophilic probe stained IM) spore samples as shown in the following text.
Localization of germinosomes in B. subtilis spores
Levels of GerD and GerKB are reported to be ~3,500 and ~700 molecules per spore, respectively, based on Western blot analyses of extracts from spores prepared in a rich sporulation medium19. Both the gerD-gfp and gerKB-mCherry genes in the KGB80 strain are under the control of their native promotor. The relative low abundance of fusion proteins led to a low fluorescent signal during imaging, so it was difficult to reconstruct such dim raw SIM images by the SIM reconstruction algorithm. However, the SIM microscope was still applied for the germinosome image acquisition, although the raw SIM images were converted into Pseudo–Widefield images by SIMcheck (ImageJ plugin). In addition, a seven stack 3D imaging was implemented to get a better overview of this IM focus. As shown in the left hand panel of Figure 2, two foci of GerD-GFP appeared in different stacks. The, in total, three GerD-GFP foci are indicated by the white arrows in the compositive column’s Z3 stack. The right hand panel of Figure 2 shows a spore with only one GerD-GDP focal point in the spore as evidenced by the white arrow in the composite column’s Z4 stack. In total, around 40% and 50% of spores had two or one GerD-GFP and GerKB-mCherry cluster, respectively (Table 1). Among the 346 spores examined, two had 4 GerD-GFP foci, and one spore even had 5 GerD-GFP foci. Noticeably, in our hands, the number of GerD-GFP and GerKB-mCherry foci in the same spore were not always the same18. As the SIM microscope had no phase contrast option, we used transmission light microscopy for spore localization. Thus, spores appear with a dark and dense core surrounded by a brighter halo.
The integrated fluorescence intensity of KGB80 spores was measured by ImageJ. Spores, which had 0, 4, or 5 foci were not included into our statistical analysis due to their low frequency. There was a high positive correlation between GerD-GFP and GerKB-mCherry integrated intensities (Spearman rank correlation coefficient = 0.73). While the integrated intensity of the GerD-GFP scaffold protein was different between different populations (Figure 3C), the integrated intensity of GerKB-mCherry was about the same in different populations (Figure 3D). The maximum fluorescence intensity of GerD-GFP and GerKB-mCherry foci tended to decrease, when the spore had multiple foci (Figure 3A, B). The maximum fluorescence of all bright spots, regarded as germinosome foci, was higher than the maximum auto-fluorescence of PS4150 spores (spores from the background strain KGB80; Figure 3A, B).
Organization of the inner membrane
As mentioned in the introduction, germinosome proteins are located in the spore’s IM. However, few details are known about the biophysical properties of this largely immobile membrane. Exploring more details, such as the IM’s local organization, would promote the understanding of the organization of IM proteins, in particular GRs and channel proteins. B. subtilis spores have a structure comprising multiple concentric layers, and a lipophilic probe cannot easily pass through these multiple layers to stain the IM surrounding the spore core. The passage of such probes is most likely hampered by the protein-rich coat layers and perhaps also the outer membrane20,21. To overcome this problem, the lipophilic membrane dye FM4-64 was added to a PS4150 culture during sporulation. By doing so, the PS4150 vegetative cell membrane was stained by FM4-64, and thus membranes in forespores obtained from this culture at the asymmetric sporulation cell division and subsequent forespore engulfment are well stained5. Consequently, the mature spore’s membranes can be visualized. A previous study indicated that most if not all of the FM4-64 is in the IM in cleaned spores5. During an approximately 2 week period of incubation and spore purification treatments, the washing procedures applied remove any FM4-64 from the outer membrane, the latter effectively being removed following the decoating treatment and extensive washing steps5. However, the decoating procedure removes no FM4-64 from the IM, nor has any effect on the germinosome foci5,6. What excited us is that brighter FM4-64 spots similar to germinosome foci appeared in both intact (Figure 4A, B) and decoated spores (Figure 4C, D) of PS4150 spores. These brighter FM4-64 spots might be involved in the clustering of germinosome proteins in the IM.
Figure 1: Overview of sporulation and slide preparation. An overview displaying the initial steps required before imaging. Detailed information is given in the protocol. (A) The schematic of the sporulation procedure in defined minimal MOPS medium. A B. subtilis PS4150 (PS832 ΔgerE::spc ΔcotE::tet) or KGB80 (PS4150 gerKA gerKC gerKB-mCherry cat gerD-gfp kan) single colony was cultured in 5 mL of LB rich medium, and adapted in 5 mL and 20 mL of MOPS medium in turn, and finally sporulated in 250 mL of MOPS medium. An early exponential phase culture (OD600, 0.3-0.4) is used in all intermediate cultures. FM4-64 (2 µg/mL) was added to the PS4150 sporulation medium for spore membrane staining 1 or 2 h after reaching the peak OD600 value. (B) Method of harvesting spores from MOPS sporulation culture and purifying spores by density gradient centrifugation. (C) Procedure of stabilizing spores on 1 % agarose pad in a gene frame chamber. Please click here to view a larger version of this figure.
Figure 2: Representative Pseudo-Widefield (PWF) 3D images of GerD-GFP and GerKB-mCherry foci in two KGB80 (PS4150 gerKA gerKC gerKB-mCherry cat gerD-gfp kan) dormant spores’ 3D-SIM raw images were taken with dual channel excitation (561nm, 30% laser power, 1 s, and 488nm, 60% laser power, 3 s) using 7 steps from top to bottom Z-stacks. Subsequently, the raw SIM images were converted into Pseudo-Widefield 3D images by the ImageJ plugin SIMcheck. From left to right, 3D images (Z2-Z5) of GerD-GFP (green), GerKB-mCherry (red) and the corresponding composite images of two KGB80 spores (i and ii) are shown in the panel. Transmission (Trans) light images of two spores (i and ii) indicated the location of spores that appear as dark dense images surrounded by a brighter halo. Scale bar = 1 µm and all panels are at the same magnification. Please click here to view a larger version of this figure.
Figure 3: The maximum fluorescence intensity of GerD-GFP in KGB80 (PS4150 gerKA gerKC gerKB-mCherry cat gerD-gfp kan) dormant spores, and the maximum auto-fluorescence intensity of PS4150 spores in arbitrary units. All spores were illuminated by the settings indicated the protocol. Panels (A) and (B) show the maximum fluorescence intensity of the GerD-GFP and GerKB-mCherry foci, respectively, in KGB80 dormant spores as well as in both cases the maximum auto-fluorescence intensity of the parent PS4150 spores. Panels (C) and (D) show the integrated fluorescence intensity of the GerD-GFP foci and the integrated fluorescence intensity of the GerKB-mCherry foci, respectively, in B. subtilis KGB80 dormant spores. We used one-way ANOVA-tests for significance determination of differences in maximum focal point intensity and integrated spore fluorescence intensities with software Origin 9.0 considering P values <0.05 as significant. Spores with 4 or 5 foci were excluded from the analysis because of their low abundance. The data is represented in notched boxplots. The notches in the plots are around the median values observed with their width proportional to the interquartile range (IQR). The whiskers shown represent a maximum of 1.5 the IQR. Asterisks indicate a significant difference of median values. GerD-GFP and GerKB-mCherry integrated fluorescence intensities have a strong positive correlation (Spearman correlation coefficient = 0.73)18. Please click here to view a larger version of this figure.
Figure 4: Representative Pseudo-Widefield (PWF) images (A and C) and reconstructed SIM images (B and D) of the FM4-64 stained IM of B. subtilis PS4150 (PS832 ΔgerE::spc, ΔcotE::tet) spores. FM-464 was incorporated into spores during sporulation. 3D-SIM raw images of intact spores (A and B) and decoated spores (C and D) were taken with one channel excitation (561 nm laser, 20% laser power, 400 ms) using a 25 step top to bottom Z-stack. Subsequently, the raw SIM data was reconstructed by the microscope imaging software (see Table of Materials) into 3D-SIM images, or converted into PWF by SIMcheck (ImageJ plugin). The cyan arrows point to FM4-64 foci in the IM in panel B and D. Scale bar = 1 μm and all panels are at the same magnification. Please click here to view a larger version of this figure.
Strain | Spores counted | Foci | Foci per spore (%) | |||||||
0 | 1 | 2 | 3 | 4 | 5 | |||||
KGB80 | 346 | GerD-GFP | 2 | 46 | 40 | 11 | 1 | 0 | ||
GerKB-mCherry | 2 | 53 | 40 | 5 | 0 | 0 |
Table 1: Presence of foci in KGB80 spores. The germinosome foci number per spore in a population of dual labeled B. subtilis KGB80 (PS4150 gerKA gerKC gerKB–mCherry cat gerD–gfp kan) dormant spores. Fluorescence in spores was counted as germinosome foci when a focus’ maximum intensity was higher than the auto-fluorescence intensity, which was the maximum intensity of PS4150 (PS832 ΔgerE::spc ΔcotE::tet) dormant spores excited by the same illumination settings as the KGB80 spores.
The protocol presented contains a standard 3D-SIM procedure for analysis of FM4-64 stained B. subtilis spores that includes sporulation, slide preparation and imaging processes. In addition, the protocol describes a modified SIMcheck (ImageJ)-assisted 3D imaging process for SIM microscopy of B. subtilis spore germinosomes labeled with fluorescent reporters. The latter procedure allowed us to observe this dim substructure with enhanced contrast. By coupling two imaging procedures, it is possible to visualize discrete sub-structures in the same spore with the same SIM microscope, thus improving our basis for a mechanistic understanding of the germination process. Note that the procedure operates at a lateral resolution of ~100 nm and an axial resolution of ~200-250 nm. This is better than the Differential Interference Contrast (DIC) wide-field microscopy approach used by Griffith7. Time resolved analysis of germinosome appearance upon initiation of germination would be a desired next step. Unfortunately, though SIM microscopy is in principle compatible with live-imaging, due to the dim nature of the germinosome signals such time-resolved SIM, analyses are not feasible because of rapid bleaching of the samples during image acquisition. In order to obtain sufficient spores for analysis, it is crucial to make sure that sporulation is efficiently taking place. Researchers must therefore check sporulation efficiency meticulously with 90% efficiency as the target. In the representative results, in dormant spores, respectively ~50% and 40% of all spores have one or two GerD-GFP and GerKB-mCherry foci (Table 1). The percentage of spores with two foci is much higher than that reported by Griffiths previously7. There are several reasons that could explain the different result in the current work. First, the 3D imaging process could facilitate the detection of more foci. Different foci in the same spore are located in different locations in the vertical direction as shown in Figure 1. Second, the CCD camera (Table of Materials) and laser unity equipped to the SIM microscope contribute significantly to the imaging results. Third, similar to Griffiths’s approach7 to average dozens of consecutive images for better image analysis, the Pseudo-Widefield image of the germinosome was also an average image from raw SIM images (5 phases and 3 orientations images). Finally, the sporulation medium and sporulation conditions, an important variable in determining spore properties, are different in our work from that used previously. Griffiths et al.7 used rich 2x Schaeffer’s-glucose (2x SG) medium for sporulation, while a defined minimal MOPS buffered medium was employed here. Several papers have demonstrated that sporulation medium and conditions have significant effects on the protein composition, resistance, and germination of B. subtilis spores22,23,24,25. Indeed, it has been shown that levels of GR subunits are 3- to 8-fold lower in spores obtained on a poor medium versus those obtained on rich-medium. GerD levels were also around 3.5-fold lower in poor medium spores, and these spores took longer to start spore germination26. However, it is not clear whether sporulation conditions also influence the number of observed germinosome foci.
Ramirez-Peralta et al.’s results26 indicated that rates of nutrient germination of spores at population levels are influenced significantly by the levels of germination proteins and GerD. If the integrated fluorescent intensities per spore from the fluorescent reporters are directly proportional to the levels of GerD and GerKB fusion proteins, levels of both fusion proteins differ widely in KGB80 spores, which is in agreement with previous work7. This protein level heterogeneity might be related to spore germination heterogeneity observed at the single spore level, and germinosome foci number might be another factor contributing to spore germination heterogeneity. Further experiments will focus on an analysis of the possible effect that germinosome foci number and foci germination protein composition (not all germinosomes may be equal in germination protein composition) could have on germination heterogeneity. The data gave rise to a number of current research questions including: i) what is the role of GerD in the clustering of GRs in the IM; and ii) how are the two other GRs, GerA and GerB, organized in the spore IM?
The protocol presented for dim and bright spore samples makes it possible to visualize discrete sub-structures in the same spore by SIM microscopy. The bright FM4-64 spots that were observed in spores might be due to extensive folding of the IM27. Alternatively, we hypothesize that these regions are areas of the IM where the dye could more easily gain access to due to increased local IM fluidity. Such Regions of Increased Fluidity (RIFs) may be organized by the cytoskeletal actin homologue MreB, well known for its concentration of fluid short acyl chain lipids28,29. Noticeably, applying the same procedure to wild-type B. subtilis spores also leads to a similar pattern of bright FM4-64 spots (our unpublished observations). In B. subtilis vegetative cells, a collapsed membrane potential results in the clustering of MreB and RIFs29. The inner membrane of dormant spores likely has a relative low membrane potential20,21 and contains detectable levels of MreB30 which might lead to the clustering of RIFs into larger domains of high fluidity29. Whether such domains could coincide with the presence of germinosomes is currently under investigation.
The authors have nothing to disclose.
The authors thank Christiaan Zeelenberg for his assistance during the SIM imaging. JW acknowledges the China Scholarship Council for a PhD fellowship and thanks Irene Stellingwerf for her help during the primary stage of imaging.
Air dried glass slides | Menzel Gläser | 630-2870 | |
APO TIRF N20R8 100× oil objective (NA=1.49) | |||
B. subtilis KGB80 (PS4150 gerKA gerKC gerKB-mCherry cat, gerD-gfp kan) | |||
B. subtilis PS4150 (PS832 ΔgerE::spc, ΔcotE::tet) | |||
Erlenmeyer flasks 1 L | Sigma-Aldrich | Z567868 | |
Erlenmeyer flasks 250 mL | Sigma-Aldrich | Z723088 | |
FluoSpheres carboxylate-modified microspheres | Invitrogen, 0.1 μm | F8803 | |
FM4-64 | Thermo Fisher Scientific | F34653 | |
Histodenz nonionic density gradient medium | Sigma-Aldrich | D2158 | |
Image J | |||
iXON3 DU-897 X-6515 CCD camera | Andor Technology | https://imagej.net/Welcome | |
LB Agar | Sigma-Aldrich | L2897 | |
Microfuge tubes 1.5 mL | Thermo Fisher Scientific | 3451PK | |
Microscope imaging software | Nikon, Japan | NIS-Element AR 4.51.01 | |
MilliQ Ultrapure Deminerilzed Water | Millipore | Milli-Q IQ 7003 | |
Nikon Eclipse Ti microscope | |||
Polypropylene Screw Cap Bottle 180 mL | Thermo Fisher Scientific | 75003800 | |
Precision Coverslips | Paul Marienfeld | 117650 | |
Round Bottom tubes 15 mL | Thermo Fisher Scientific | Nunc TM | |
Screw cap tubes 50 mL | Thermo Fisher Scientific | Nunc TM |