The YeeK and GerQ substrates control spore encasement by Tgl-CFP

Four Tgl substrates are known, YeeK, GerQ, SafA^FL, and C30 [31–35]. With the exception of YeeK, which is under σ^K control, all other known Tgl substrates are produced early in development, under the control of σ^E. In previous work we have shown that safA is important, but not the exclusive determinant, for the assembly of Tgl [35]. Therefore, we wondered whether other substrates played redundant roles in the assembly of Tgl. To test this possibility, we analyzed the localization pattern of a functional Tgl-CFP fusion in single or multiple mutants unable to produce the known Tgl substrates. The functional Tgl-CFP fusion was previously described; in it, Tgl is separated from the CFP moiety by a 20Å-long, rigid, α-helical linker [35]. In wild-type (WT) sporangia, and in line with previous results [35], Tgl-CFP initially localizes to phase dark spores, as two caps, simultaneously to both the mother cell proximal (MCP) and the mother cell distal (MCD) poles of the engulfed forespore (localization class A); Tgl-CFP then encases the spore forming a continuous ring of fluorescence (class B) [35] (Fig 2A and 2B). Thus, Tgl belongs to class V of coat proteins as defined by McKenney and Eichenberger, which are produced mainly under σ^K control and localize to both poles of phase dark spores before encasing the forespore [17]. The accumulation of Tgl at the spore poles, which may correspond to our two-cap (class A) pattern, has been described [37]. Here, and as a reference for the localization of Tgl-CFP in different mutants, the localization of the fusion protein was assessed by scoring sporangia according to classes A and B (as defined above; with one exception, described below, other localization patterns were not considered in this study). Since the main period of tgl expression is under σ^K control [31, 37–39], we analyzed the localization of Tgl-CFP 6 and 8 hours after sporulation initiation, when σ^K is active and the majority of the sporangia have visible phase dark or phase bright spores (Fig 2A). For the WT, classes A and B correspond to ≥ 98% of the sporangia with phase dark or phase bright spores scored, in agreement with previous work [35] (Fig 2B and 2C). In both the yeeK and gerQ mutants, the sum of classes A and B corresponded to >97% of the sporangia with either phase dark or phase bright spores, similar to the WT (Fig 2B and 2C). Thus, Tgl-CFP is recruited to the forespore surface in nearly all yeeK or gerQ sporangia. The main difference found for the yeeK mutant when compared to the WT was that for sporangia of phase dark spores the representation of class A (49%) and class B (51%) was nearly the same, whereas class A represented 61% of WT sporangia with phase dark spores (Fig 2B). Deletion of gerQ also increased the representation of class B in sporangia of phase dark spores (to about 60%). The increase in class B in sporangia of phase dark spores in either mutant suggests that the absence of YeeK or GerQ facilitates encasement by Tgl-CFP. This effect seems mainly mediated by gerQ, because in a yeeK gerQ double mutant, class B in sporangia of phase dark spores represents 69% (as compared to 62% for the gerQ single mutant) (Fig 2A and 2B). These results suggest that YeeK and GerQ, per se, have no major impact on the recruitment of Tgl to the forespore but somehow control the encasement step, which seems to be more efficient in their absence. In agreement with the inference that recruitment is largely unaffected in the gerQ or yeeK mutants, or in the gerQ yeeK double mutant, quantification of the forespore/mother cell fluorescence ratio in sporangia of phase bright spores revealed a median value of 1.00 for the WT and yeeK strains, a median value of 1.02 for the gerQ mutant and of 0.95 for the yeeK gerQ double mutant (Fig 2D). Thus, in what concerns the recruitment step, essentially no difference was found between the WT and yeeK or gerQ mutants, or the double yeeK gerQ double mutant.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The localization of Tgl depends on its substrates.

A: Samples of different strains expressing tgl-cfp were collected 6 and 8 hours after induction of sporulation by resuspension and processed for phase contrast (PC) and fluorescence microscopy. All strains bear an insertional tgl::sp allele and tgl-cfp at the amyE locus, in addition to the indicated mutations. The tgl::sp mutant producing Tgl-CFP from amyE serves as the reference strain and is referred to in the figure as the WT. The ΔC30 strain bears the safA^M161/164A allele and is unable to produce C30 by internal translation; the ΔsafA_FL strain bears the safA^F155STOP allele and produces C30 but not SafA^FL. Representative images of the patterns of fluorescence detected for the different strains are shown. The CFP and PC channels are shown. The arrowheads indicate the position of the MCP or MCD forespore poles. The yellow arrows point to partially retractile material which is detected in the strains unable to produce SafA. Scale bar, 1 μm. B and C: The graphs show the % of cells that display a normal Tgl-CFP localization pattern in the different strains analyzed in A. In the majority of the population for the WT (≥ 98%), at hour 6 or 8 of sporulation, Tgl-CFP localizes to the forespore either as 2 caps of fluorescence at the forespore poles (class A), or as a ring of fluorescence around the forespore (class B). These two classes, representing the prevailing Tgl-CFP localization patterns and were scored for sporangia carrying phase dark (panel B) or phase bright spores (panel C). The numbers above the bars correspond to the sum of the two classes. A minimum of 175 cells were scored for each strain over two independent experiments. The mean and standard deviation are represented. D: The graph displays the individual forespore/cell fluorescence ratios, normalized by the median value obtained for the wild type strain, grown on the same day (for details see materials and methods section). The median and interquartile range, for the strains analyzed, is also represented. The analysis was only conducted for sporangia with phase bright spores. n indicates the number of cells analyzed. The WT values represented are from 8 independent experiments, while the values for the other strains were obtained from 2 independent experiments.

https://doi.org/10.1371/journal.pgen.1007912.g002

Morphogenetic proteins SafA and C30 are the main recruiters of Tgl to the developing spore

We then examined the assembly of Tgl-CFP in a strain where SafA^FL is produced, but because the internal Met codons 161 and 164 were changed to Ala codons, the independent production of C30 through internal translation of the safA mRNA is eliminated [28, 35]. In this strain, herein termed ΔC30 (carrying the safA^M161A/M164A allele) the sum of classes A and B still represent the majority (>95%) of the sporangia of phase dark spores, but this number dropped to 89% for sporangia of phase bright spores. C30 thus plays a role in recruitment of Tgl. In agreement with this conclusion, the analysis of the forespore/mother cell fluorescence ratio for the ΔC30 mutant strain shows that Tgl-CFP accumulates to slightly higher levels in the mother cell (median value of 0.95, compared to 1.00 for the WT strain) (Fig 2D).

In sharp contrast, in the ΔsafA in frame-deletion mutant, unable to produce both SafA^FL and C30, the sum of classes A and B represented only 5% of the sporangia of phase dark spores and 28% of those with phase bright spores (Fig 2A, panels k-q and Fig 2B and 2C). Thus, in the absence of SafA^FL and C30, recruitment of Tgl-CFP to the spore surface is severely impaired. Note that in spite of the strong localization defect, the accumulation of Tgl-CFP does not differ significantly from the WT, as shown by immunoblot analysis of whole cell extracts (S1A Fig). Also, that localization improves in sporangia of phase bright spores, further suggests a delay in the assembly of Tgl. In keeping with these inferences, the median value of the forespore/mother cell fluorescence ratio for the ΔsafA strain (0.69) is considerably lower than for the WT (1.00), implying that the elimination of safA causes accumulation of Tgl-CFP in the mother cell (Fig 2D). The ΔsafA mutation, however, does not completely eliminate the localization of Tgl-CFP, suggesting that YeeK and GerQ may partially compensate for the absence of SafA^FL and C30, and that their assembly is partly independent of safA. Alternatively, additional, as yet unknown factors, may control the localization of Tgl.

Partial redundant roles for the Tgl substrates in assembly of the enzyme

We then analyzed the localization of Tgl-CFP in a ΔsafA yeeK gerQ triple mutant, that is, in the absence of all the known substrates of Tgl (Fig 2A, panels r-b´). In line with the conclusion that SafA^FL and C30 are the main factors involved in the recruitment of Tgl, the sum of classes A and B in sporangia of phase dark spores did not differ much from the ΔsafA single mutant (Fig 2B). Surprisingly, however, in sporangia of phase bright spores, not only the sum of classes A and B represented 45% of the sporangia scored (Fig 2C), but the forespore/cell fluorescence ratio for the triple mutant (0.86) is significantly higher than that of the safA mutant (0.69) (Fig 2D). One possible explanation is that the simultaneous absence of SafA, C30, YeeK and GerQ unmasks an alternative SafA-independent pathway that can recruit Tgl-CFP to the spore surface, albeit inefficiently. If so, one possible implication is that additional, as yet unknown Tgl substrates may exist. In previous work we presented evidence indicating that synthesis of the spore cortex peptidoglycan facilitates assembly of Tgl [35].

Finally, we examined the assembly of Tgl-CFP in a strain bearing a safA allele with a stop codon replacing the Phe155 codon; in this strain (termed ΔsafA^FL, carrying the safA^Phe155STOP allele) C30 is still formed through internal translation but production of SafA^FL is eliminated (Fig 2A, panels c’-l’). Tgl-CFP was not detected in sporangia of phase dark spores, and classes A and B represented 16% of the sporangia of phase bright spores, a value lower than that found for the ΔsafA mutant (Fig 2C). Thus, even though the forespore/mother cell fluorescence ratio of ΔsafA^FL and ΔsafA sporangia with phase bright spores is very similar (Fig 2D), assembly of Tgl-CFP is more compromised when SafA^FL (but not C30) is absent (ΔsafA^FL) than when both SafA and C30 are absent (ΔsafA). Because the localization of C30 is dependent on SafA^FL [28, 29] one possible explanation is that C30 retains Tgl-CFP in a mislocalized pattern. Consistent with this supposition, in the majority of the ΔsafA^FL sporangia carrying phase dark spores, Tgl-CFP accumulated as a dot of fluorescence in the MCP forespore pole (Fig 2A, panel c’) (NB: this pattern will be further discussed below). Together, the results suggest that the known Tgl substrates, as well as additional factors, play partially redundant roles in the assembly of the enzyme, with GerQ and YeeK mainly involved in enforcing the normal pattern of encasement, and SafA and C30 primarily involved in recruitment. This partial redundancy in the recruitment of Tgl-CFP to the surface of the developing spore explains why assembly of the enzyme cannot be completely abolished in any of the single mutants studied.

Non-active site residues at both entrances of the Tgl tunnel are important for enzyme activity

Since the assembly of Tgl is largely controlled by proteins that are also Tgl substrates, then, it should be possible to identify residues in Tgl, involved in enzyme-substrate interactions, important for the assembly of the enzyme. In addition, mutations in those residues of Tgl could possibly have more pronounced effects on the assembly of the enzyme than the absence of the known substrates individually. Tgl has a distinctive catalytic dyad composed by Cys116 and Glu187, the latter of which can be non-reciprocally substituted by Glu115 [14] (Fig 3A). The catalytic residues of Tgl are located within a tunnel that traverses the molecule from side to side and thus are not surface exposed (Fig 3B–3F). As has been seen for the animal-type TGases, Gln (acceptor) and Lys (donor) proteins that participate in cross-linking reactions (Q and K substrates) approach the active site from opposite sides of a tunnel [40, 41] (Fig 3B). Thus, both entrances of the Tgl tunnel act as substrate docking ports. Accordingly, previous work has shown that His200, located at one of the entrances of the tunnel (in the so-called back side of the enzyme) (Fig 3D and 3F), while not playing an essential role in catalysis, is important for the interaction with substrates and in that way for the overall activity of Tgl [14]. Thus, in searching for additional residues with a role in enzyme-substrate interactions, we focused our attention on four residues at the front entrance of the tunnel (Trp149, Tyr171, Arg185 and Asn188; Fig 3C and 3E), and two residues at its back entrance (Phe69 and Trp184; Fig 3D and 3F). With the exception of Arg185, all of the selected residues are well conserved among Tgl homologues [14]. Arg185 appeared interesting, however, because with other residues in its vicinity, it shows high thermal displacement parameters (B factors) suggesting flexibility [14]. Flexibility in this region of the protein could be important for substrate interactions and for displacement of the ceiling of the tunnel following catalysis, allowing release of a cross-linked product [14] (Fig 3E).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Residues at the front and back entrances of the Tgl tunnel are important for activity.

A: Cartoon and molecular surface representation of the structure of Tgl. The side chains and surface occupancy of the catalytic residues are shown in green. The active site of Tgl sits in a tunnel that traverses the molecule from side to [14]. B-F: Zoomed view of the tunnel region observed from the top (B), front (C, E) or back (D, F). In B and D secondary structure elements and side chains of relevant residues are shown, while in E and F only the surface of the molecule is represented. Relevant residues are highlighted in color: front (in blue), back (red), and catalytic residues (green). The free space of the tunnel is represented in orange, either as a mesh (B) or by surface occupancy (E, F). The models in A-F were drawn using the published structure of Tgl (PDB accession code 4P81). G: Tgl (WT type or mutant forms) was incubated at 50°C in the presence of BSA and dansylcadaverine (Ds-cd). Samples were taken at the indicated times (in min) and resolved by SDS-PAGE. The gels were exposed to UV light to reveal labeled BSA (BSA*), and then stained with Coomassie brilliant blue for the visualization of the amount of BSA and Tgl loaded. H: shows representative assays for the wild type (WT) enzyme and mutant forms with the indicated single Ala substitutions of residues in the front or back entrances of the Tgl tunnel. I: Quantification of the activity of Tgl^W149A. Note that Tgl^W149A did not label BSA, but showed auto-labeling activity (producing Tgl*). Thus, Tgl^W149A shows an altered substrate specificity.In H and I, the fluorescence values obtained for BSA* (H) of for Tgl* (I) were normalized to the value obtained for the wild type enzyme after 120 min of reaction and plotted for all the assays where fluorescence could be detected (see also [14]).

https://doi.org/10.1371/journal.pgen.1007912.g003

To examine the contribution of the selected residues on the overall activity of Tgl in vitro, each was independently substituted by Ala. Tgl^WT and its variants were overproduced with a C-terminal His₆ tag in E. coli by auto-induction [15]. All the proteins accumulated in E. coli and we have shown before that the His₆ tagged-Tgl is functional, both in vitro and in vivo [15]. The activity of Tgl^WT and its variants was then evaluated using BSA as the Q substrate, and the fluorescent primary amine dansylcadaverine as the K substrate. Enzymatic activity can be assessed in this manner by monitoring the formation of fluorescent BSA over time [14]. These assays were conducted at 50ºC, which we showed before to be the optimal temperature for labeling of BSA; presumably, increased conformational flexibility of the substrate makes surface exposed Q residues more prone to labeling [14].

For two of the mutant enzymes tested, Tgl^W149A and Tgl^Y171A, no labeled BSA could be detected (Fig 3G and 3H; not plotted). The activity of Tgl^R185A and Tgl^W184A was significantly reduced compared to the WT, while Tgl^F69A and Tgl^N188A showed only residual activity (Fig 3G and 3H). In previous work we showed that no activity could be detected for Tgl^H200A,Tgl^E115A and Tgl^C116A, while Tgl^E187A retained considerable activity (about 60% when compared to Tgl^WT), as Glu115 acts as a substitute in the absence of the E187 side chain [14]. In all, these results are in line with the view that residues at both entrances of the tunnel are involved in enzyme-substrate interactions and in that way important for the overall activity of Tgl.

The Q and K sides of Tgl are located at the front and back ends of its tunnel, respectively

While we could not detect accumulation of fluorescent BSA during the time of the amine incorporation assay with the Tgl^W149A variant, we detected the accumulation of fluorescent Tgl^W149A (Fig 3G). The auto-catalytic activity of Tgl [15] and other TGases has been described [42–46], and during our labeling assay, we could also detect formation of fluorescent Tgl^wt concurrently with the labeling of BSA (Fig 3G). Both Tgl^W149A and Tgl^wt show similar auto-labeling activity during the first 60 minutes of reaction (Fig 3I). Thus, Tgl^W149A is essentially functional although it may be somewhat unstable over long incubation times. In any case, and importantly, Tgl^W149A shows an altered substrate specificity when compared to Tgl^WT, as it no longer uses BSA but labels itself. This is in line with the view that residues at the two entrances of the Tgl tunnel, as seen for other TGases, are important for enzyme-substrate interactions (above). More importantly, however, because BSA or Tgl are the Q substrate in the BSA/Tgl-dansylcadaverine labeling reactions, Trp149 must be involved in the interaction with the Q substrate. It follows that the front side of the tunnel, where Trp149 is located, corresponds to the Q substrate docking site, while the back side of Tgl is the site of interaction with K substrates (Fig 3C and 3E).

Q-side residues are more important than K-side residues for the localization of Tgl

Since the substrates of Tgl are important for the localization of the enzyme to the surface of the forespore, Tgl mutants with Ala substitutions of residues located at both entrances of the tunnel should exhibit impaired localization. We tested this prediction by introducing point mutations in tgl-cfp and transferring the resulting alleles to the non-essential amyE locus of the tgl::sp insertional mutant. We then analyzed the localization of the different forms of Tgl during sporulation using fluorescence microscopy (Fig 4A). The residues analyzed included not only those located at both entrances of the tunnel that were also tested for a role in enzyme activity (above), but also the residues that form the catalytic center of Tgl, hidden inside the tunnel, that we defined previously [14] (Fig 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Residues at the front entry of the Tgl tunnel are more important than back residues for the localization of the enzyme.

A: Forms of Tgl-CFP bearing the indicated single Ala substitutions (color coded according to their role in activity or position in the structure of Tgl; as in Fig 3) were produced from the amyE locus in a strain bearing a tgl::sp insertional allele. Sporulation was induced by resuspension, samples collected 8 hours thereafter, and the assembly of the WT and different forms of Tgl-CFP monitored by phase contrast and fluorescence microscopy. Scale bar, 1 μm. B: Cells bearing phase bright spores were analyzed and the normalized forespore/cell fluorescence ratio, (F_FS/F_cell)_norm, was determined (see also Fig 2). For cells of each strain, the individual normalized ratio values, the median, and interquartile range are shown. n, number of sporangia scored. C: The fluorescence signal detected in whole cells was normalized using the median value of the wild type strain grown on the same day, to give (F_cell)_norm (see materials and methods for details). The same cells were analyzed in B and C.

https://doi.org/10.1371/journal.pgen.1007912.g004

We determined the forespore/mother cell fluorescence ratio for all strains, producing the WT and mutant forms of Tgl-CFP (Fig 4B). Of the four front side residues analyzed in this way, only the R185A substitution did not seem to impair assembly of Tgl-CFP (median values of the forespore/cell fluorescence ratio of 0.99 and 1.00, for Tgl^R185A-CFP and Tgl^WT-CFP, respectively) (Fig 4B). Assembly of Tgl^W149A, Tgl^Y171A and Tgl^N188A, however, was severely impaired (median values of 0.44, 0.47 and 0.51, for the W149A, Y171A, and N188A substitutions, respectively). Moreover, for these three proteins, fluorescence dispersed throughout the mother cell cytoplasm was seen (Fig 4A, panels e, f, h). Note that all these three residues, W149, Y171 and N188, are located at the Q-acceptor side of the tunnel, and that the surface of N188, in particular, forms the bottom of the front entrance of the tunnel. Forms of Tgl with single Ala substitutions of K-acceptor side (back) residues, Tgl^F69A, Tgl^W184A, and Tgl^H200A, also show impaired assembly, but judging from the quantification of the forespore/mother cell fluorescence ratios, to a lesser degree than front side substitutions (median values between 0.73–0.94) (Fig 4B). Lastly, the assembly of Tgl^C116A, Tgl^E115A and Tgl^E187A, with Ala substitutions of the catalytic residues buried within the Tgl tunnel, was also reduced as compared to the WT (median values between 0.80–0.91), but to a lesser degree than Q- (median values between 0.44–0.51, R185A excluded) or K-side mutants (median values between 0.73–0.94) (Fig 4C). The catalytic residues are not likely to be directly implicated in substrate docking. Q-donor substrates, however, form an acyl-enzyme intermediate with the catalytic Cys116 residue, in the first step of the reaction catalyzed by TGases. Therefore, while Q-side residues (located at the front entrance of the tunnel) make the most important contribution to the assembly of Tgl, this step may be rapidly followed by the formation of an acyl-enzyme intermediate. In any event, the results suggest that enzyme-substrate interactions are important for the assembly of Tgl. The analysis of the levels of the various forms of Tgl in spore coat extracts is in agreement with the microscopy analysis (S1B Fig; see also the S1 Text).

To test whether the recruitment defects observed for the different forms of Tgl-CFP resulted from protein instability, extracts from sporulating cells were fractionated into a mother cell and a forespore fraction and the presence of the various forms of Tgl assessed by immunoblotting (S1C Fig). We found that all forms of the enzyme showed reduced levels in the forespore fraction, but correspondingly higher levels in the mother cell fraction (S1C Fig; see also the S1 Text). Thus, the mutations did not seem to grossly alter the folding of Tgl. In addition, the total cell fluorescence was measured for the strains expressing the various Tgl-CFP fusions (Fig 4C; see also the material and methods section). Note that we have reported before that the functional Tgl-CFP fusion shows minimal processing and release of the CFP moiety [35]. Some of the point mutations introduced in tgl-cfp, for instance E115A and F69A, lead to whole cell fluorescence lower than that found for the WT enzyme. While this analysis may indicate that some of the mutations may slightly alter the folding of Tgl-CFP, we note that there was no direct correlation between the levels at which the different mutant forms of Tgl-CFP accumulate in the cell and their recruitment to the forespore: the total cell fluorescence values were the lowest for Tgl^E115A-CFP (0.58) and Tgl^F69A-CFP (0.67), but the recruitment of these forms was not among the most affected, with a median value of forespore/cell fluorescence ratio of ~0.80, while the three lowest values are between 0.44 and 0.51 (Fig 4B and 4C). Thus, the defects observed in the recruitment of the different mutant forms of Tgl-CFP do not seem a direct result of protein instability and likely reflect impaired enzyme-substrate interactions, with Q-side residues (located at the front entrance) making the most important contribution to the assembly of the enzyme.

Residues important for the localization of Tgl are also important for the cross-linking of C30 in vitro

One prediction arising from the results described in the preceding section is that at least some of the Ala-substituted forms of Tgl should show reduced cross-linking activity towards its natural substrates. We used C30 to test this prediction as this is the substrate whose absence, per se, more severely impairs assembly of Tgl (above). Thus, the cross-linking activity of Tgl mutant forms whose recruitment appears more compromised, with either Ala substitutions of residues at the Q and K entrances of the tunnel, were analyzed in parallel with Tgl^WT and two variants with Ala substitutions of catalytic residues. C30 and Tgl (WT and mutant forms) were independently over-produced in E. coli with a C-terminal His₆-tag, and the various proteins were purified and incubated together at 37°C. This is the temperature at which the microscopy experiments were conducted; note, however, that the optimal temperature for Tgl activity is 50°C [14, 15], and it was at this temperature that the BSA labeling assays were conducted (see also above). A cross-linking assay with Tgl and C30 or BSA as the substrates, conducted at 50º is shown in S3 Fig, for reference. Tgl^WT efficiently cross-linked C30 into high molecular weight species, C30_m (Fig 5A and 5B, parenthesis). For Tgl^E115A, Tgl^C116A, Tgl^Y171A and Tgl^N188A, no activity could be detected [Fig 5A; note that the C30 species presumed to be a dimer, (C30)₂, resistant to the SDS-PAGE conditions, is detected in all assays at time 0]. In contrast, Tgl^H200A displayed significant activity when compared to the other forms of the enzyme, with the formation of high molecular weight products during the time of the assay (Fig 5A). Tgl^W149A displays significant enzymatic activity but an altered specificity: it shows auto-labeling activity with dansylcadaverine but is unable to label BSA (Fig 3G and 3I; see above). Yet, our results also show that the assembly of Tgl^W149A is compromised (Fig 4 and S1B Fig). The C30 cross-linking assays, however, show that Tgl^W149A produces cross-linked species that are not detected with Tgl^wt or any of the other mutant forms of the enzyme (Fig 5A and 5B, asterisks). While these forms may be the result of intramolecular cross-linking, the observation shows that Tgl^W149A interacts with C30 differently from Tgl^WT. Presumably, then, altered interaction of Tgl^W149A with C30 does not permit efficient recruitment of the enzyme in vivo.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Non-catalytic residues, important for the localization of Tgl are also important for the cross-linking of C30 in vitro.

C30 and Tgl (either Tgl^WT or the indicated mutant forms in A and B) were incubated at 37°C, and samples were collected at the indicated times (in min). The gels were stained with Coomassie brilliant blue for protein visualization; Tgl and the different forms of C30 are identified on the right-hand side of each panel. (C30)₂ corresponds to a possible dimer of C30, while (C30)_m represents multimeric forms of C30. The asterisks in panel B indicate the site of migration of forms of C30 that are only detected in assays with Tgl^W149A. Enzymes are color coded according to the location of the residues exchanged relative to the tunnel of Tgl: green, catalytic residues; blue, front side of the tunnel; red, back side of the tunnel. The site of migration of the molecular weight markers, in kDa, is indicated on the left side of each panel. C: Production of C30 in the absence of SafA^FL causes mislocalization of Tgl-CFP. Tgl-CFP was produced in the indicated strains and the percentage of sporangia in which the fusion protein localized as a single dot of fluorescence on the MCP pole of the forespore (schematically represented on the right side) was scored. Only strains where this pattern of localization was observed are illustrated. Note that the ΔsafA^FL mutant bears the safA^F155STOP allele, and only produces C30 (see text for details). The mean and standard deviation are represented. D: Samples were collected from sporulating cultures of the indicated strains 6 hours after the onset of the process and analyzed by thin sectioning TEM. Note that the ΔsafA allele was complemented at amyE with either safA^wt or safA^F155STOP. The blue arrows points to the cortex region, the brown arrows to inner coat material and the yellow arrows to the position of outer coat material. Scale bar, 100 nm.

https://doi.org/10.1371/journal.pgen.1007912.g005

C30 retains Tgl^WT, but not forms of the enzyme with single Ala substitutions of Q- and K-side residues, in the mother cell

In the three mutants, ΔsafA, ΔsafA^FL and the triple mutant ΔsafA, yeeK gerQA, in which the class A/class B pattern of Tgl-CFP localization is more significantly reduced (quantification in Fig 2B–2D), Tgl-CFP is detected in the majority of sporangia with either phase dark or phase bright spores as a strong dot of fluorescence at the MCP pole of the forespore (Fig 2A). Quantification of the dot pattern for the three strains shows a dot in 80% of the sporangia with phase dark spores and 12% of those with phase bright spores for the ΔsafA^FL mutant (Fig 5C). The dot pattern was less represented in the ΔsafA single mutant or the yeeK gerQ ΔsafA triple mutant (Fig 5C). We used transmission electron microscopy (TEM) to characterize further the phenotype of the ΔsafA^FL mutant (Fig 5D). The TEM images of sporangia at an intermediate stage of sporulation shows the accumulation of coat material at the MCP forespore pole in ΔsafA sporangia and projecting into the mother cell cytoplasm, as previously noted [22] (Fig 5D, middle panel). This phenotype is corrected by safA^wt at amyE (Fig 5D, left panel). Strikingly, accumulation of coat material was seen at the MCP forespore pole in ΔsafA^FL sporangia (Fig 5D, right panel). In the ΔsafA^FL strain, the C30 form, but not SafA^FL, is produced [28]. Because C30 is not assembled in the absence of SafA^FL [27, 28], one possibility is that C30 retains Tgl-CFP, along with other coat proteins, in the mother cell.

If so, Tgl^WT and C30 should co-localize in ΔsafA^FL sporangia. Conversely, Tgl variants whose assembly and the ability to cross-link C30 in vitro is impaired should not co-localize with C30. To test these predictions, we co-expressed C30-yfp along with tgl^wt-cfp in the ΔsafA^FL mutant. Because the single fluorescence dot pattern of Tgl-CFP is more easily seen in sporangia with phase dark spores, the co-localization of Tgl-CFP and C30-YFP was only scored in these cells. We found that C30-YFP localizes as a bright dot of fluorescence in the majority of the sporangia (84%) and that Tgl^wt-CFP co-localized with C30 (in 76% of the sporangia scored) (Fig 6A and 6B). Forms of Tgl with Ala substitutions of K-side residues (F69A, H200A), or the catalytically inactive Tgl^C116A-CFP form, also co-localized with C30-YFP (values varying between 69 and 70%, as indicated in Fig 6B), although a weak CFP fluorescence signal was also seen dispersed throughout the mother cell cytoplasm (Fig 6A). Strikingly, however, Tgl^Y171A-CFP and Tgl^N188A-CFP (with Ala substitutions in Q-side residues) largely accumulate throughout the mother cell cytoplasm, and only co-localize with C30-YFP in 2 and 10% of the sporangia scored (Fig 6A and 6B). Note, however, that the C30-YFP dot at the MCP forespore pole is maintained in tgl^Y171A-cfp (77% of the sporangia) or tgl^N188A-cfp (80%) sporangia. Thus, forms of Tgl with Ala substitutions of residues in the Q acceptor side of Tgl, which show impaired assembly, but much less so K-side residues, fail to co-localize with C30 in the mother cell cytoplasm. This observation lends strong support to the idea that assembly of Tgl is mainly determined by its interaction with the SafA^FL/C30 substrates and that Q-side residues make a more important contribution to the assembly of Tgl than do K-side residues.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. C30 retains Tgl^WT-CFP but not front side mutants in the mother cell cytoplasm.

A: Forms of Tgl-CFP bearing the indicated single Ala substitutions (color coded according to their role in activity or position in the structure of Tgl; as in Fig 3) were produced from the amyE locus in ΔsafA^FL strain bearing a tgl::sp insertional allele and C30-YFP produced from the thrC locus. Sporulation was induced by resuspension, samples collected 8 hours thereafter, and the localization of the CFP or YFP fusions monitored by phase contrast (PC) and fluorescence microscopy. Scale bar, 1 μm. White arrowheads indicate the position of the fluorescence signal from Tgl-CFP or C30-YFP in cells in which a forespore could be detected (red color is used for the phase contrast images). Scale bar, 1 μm. B. The graphs, aligned with the panels in A, show the percentage of sporangia with phase dark spores that showed C30-YFP localized as a dot of fluorescence in the MCP pole (class a) and Tgl-CFP with a similar pattern (class b). Sporangia were also scored for the co-localization of C30-YFP and Tgl-CFP at the MCP pole of the forespore (class c). A diagram of each localization class is shown on the right side of the panel. The graphs display the mean and standard deviation for each class. The numbers on the left side indicate the mean value obtained for each class in each strain. A minimum of 200 cells were scored for each strain in each of two independent experiments.

https://doi.org/10.1371/journal.pgen.1007912.g006

As a control for these experiments, we also wanted to test whether a SafA-dependent protein that is not a direct substrate of Tgl would still co-localize with C30 when the cells produce a form of Tgl, such as Tgl^Y171A whose localization is impaired. Since in the absence of SafA^FL accumulation of some coat material is seen by TEM at the MCP forespore pole (Fig 5D; see above), we expected that such a protein would form dots at the MCP forespore pole regardless of the form of Tgl produced by the cells. A YaaH-GFP fusion was shown before to be dependent on SafA for localization to the inner coat [17, 25, 36] and YaaH is not a Tgl substrate [30–35]. We constructed a YaaH-YFP fusion and examined its localization as well as the localization of Tgl^WT-CFP or Tgl^Y171A-CFP, in the strain producing C30 but not SafA^FL (S4A Fig). In this experiment, Tgl-CFP formed a dot in 88% of the sporangia examined, while dots were only seen in 5% of the sporangia for Tgl^Y171A, the most common pattern being the fluorescence signal dispersed throughout the mother cell cytoplasm (S4B Fig, class b; see also above). For YaaH-YFP, however, no dots of fluorescence were seen at the MCP forespore pole; instead, YaaH-YFP localized as a ring around the forespore (S4A and S4B Fig). This result shows that mislocalization of C30 is not sufficient to retain all of the known SafA-dependent proteins in the mother cell. On the other hand, it strongly reinforces the specificity of the Tgl^WT-C30 interaction and its requirement for the proper localization of Tgl.

Reaction of Tgl with the C30 substrate

The results in the preceding sections suggest a model in which Tgl is recruited to the surface of the developing spore by its pre-assembled substrates, with SafA and C30 making the most important contribution. This model raises the possibility that the activity of Tgl is itself controlled by the local concentration of the pre-assembled substrates. To investigate how the activity of Tgl varied with the concentration of its substrates, we conducted cross-linking assays using purified C30. In a first series of assays, C30, at a fixed concentration (25 μM) was incubated with various concentrations of Tgl, and the formation of cross-linked species monitored by SDS-PAGE. Total cross-linking was estimated by measuring the decrease in the representation of C30 over time, which is converted into high molecular weight forms (S5A Fig; see the material and methods for details on the quantification). Incubation of C30 alone results in the formation of the (C30)₂, dimer (at about 60 kDa) that resists the denaturing conditions of SDS-PAGE (S5A Fig). Formation of this species was increased for long incubation times in the presence of the lowest concentration of Tgl tested (0.8 μM); at an intermediate concentration of Tgl (4 μM), the cross-linking activity increased linearly with time (Fig 7A). For the next concentration of enzyme tested (8 μM), activity increased linearly during the first 70 min of incubation but then reached a plateau (Fig 7A). For a higher concentration of Tgl (12.5 μM), however, activity increased linearly up to 120 minutes of incubation (Fig 7A). Together, these results suggest that the enzyme is somehow inactivated over time, in a manner that depends on the enzyme/substrate ratio.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Reaction of Tgl with C30 and ab initio model of C30.

A: Tgl, at the indicated concentrations, was incubated with purified C30 (at 25 μM; see also S6A Fig) and samples of the reaction mixture collected at the indicated times. The samples were mixed with loading dye and boiled to stop the reaction and resolved by SDS-PAGE. The level of C30 decreases over time as the result of Tgl-mediated cross-linking into higher molecular weight forms (see S5A Fig for representative SDS-PAGE gels). Tgl activity is expressed as the level of C30 in each time point, normalized to the initial amount of the protein. B: The assays were conducted as described in A, except that Tgl (at 8 μM) was incubated with different concentrations of C30, as indicated (see S5B Fig for representative SDS-PAGE gels). C: Tgl (at 4 μM) was incubated alone during 20 min and C30 added at this point, to a final concentration of 25 μM. Samples were taken along time and analyzed by SDS-PAGE (representative gels are shown in S5C Fig). The level of Tgl in each sample was quantified and normalized to the initial amount of protein. In A-C, experiments were repeated three times with different preparations of Tgl and C30. D: SAXS intensity of C30 (blue circles), I(s), is represented in logarithmic scale as a function of the momentum of transfer, s. The inset shows the SAXS-generated ab initio molecular envelope obtained by clustering and averaging models from 20 independent runs, with a Normalized Spatial Discrepancy (NSD) of 0.8 Å. The low-resolution reconstruction of C30 in solution is compatible with a high-order macromolecular assembly, most likely a hexamer, termed (C30)₆. Both mesh and solid representations of the model are shown, both centered at the origin of the cartesian axis. The red line corresponds to the scattering profile calculated from the ab initio model that best fitted the experimental data (χ² = 2.2). Point-by-point residuals of the fitting and the absolute intensities are shown in the bottom panel. E: Pair-wise distance distribution (P(r)) of SEC-purified C30 calculated from data in the range 0.010<s< 0.38Å^-1 (see the Material and Methods for details). The derived R_g and D_max values are displayed in dashed lines.

https://doi.org/10.1371/journal.pgen.1007912.g007

If so, then the plateau effect seen at 8 μM of Tgl should be eliminated by decreasing the concentration of substrate. To test this, Tgl, at a final concentration of 8 μM, was incubated with progressively lower concentration of C30. Surprisingly, we found that the activity of Tgl decreased with an increase in substrate concentration (Fig 7B; representative gels are shown in S5B Fig). When C30 was present at 25 μM, the highest substrate concentration tested, enzyme activity also reached a plateau, after 60 min of incubation (Fig 7B). Inspection of the SDS-PAGE gels used to quantify the activity of Tgl show that the enzyme is depleted over time, in a manner that is dependent on the concentration of substrate (S5A and S5B Fig). This suggests that Tgl itself becomes cross-linked. To more precisely test this idea, we incubated Tgl alone (at 4 μM) and after 20 min, C30 was added to a final concentration of 25 μM. During the first 20 min of the reaction, the intensity of the band corresponding to Tgl, at 26 kDa, is slightly reduced (S5C Fig). When C30 is added (to 25 μM), however, the intensity of the Tgl band is rapidly reduced, concomitantly with the formation of the (C30)₂ dimer (S5C Fig). The quantification of this effect is shown in Fig 7C. The decrease in Tgl when incubated alone is consistent with the previously described auto-cross-linking activity [14]) and with the auto-labelling activity detected in the presence of dansylcadaverine (above). Together, these results suggest that Tgl itself is cross-linked during the reaction, and that the ratio of Tgl/substrate will limit the duration and extent of C30 cross-linking.

C30 is an oblong hexamer in solution

C30 self-interacts in a yeast two-hybrid system [27], and it has been proposed that it forms high molecular weight complexes that are cross-linked by Tgl at the surface of the developing spore [35]. The oligomeric state of C30 in solution, however, is not know. Also, the cross-linking assays described in the preceding section indicate that Tgl is itself cross-linked to C30, which could possibly be facilitated if C30 forms an oligomeric structure. To gain insight into the oligomeric state of C30, the purified protein (S6A Fig) was analyzed by size exclusion chromatography (SEC) and small-angle X-ray scattering (SAXS). In SEC, purified C30 eluted as a main single sharp peak (ca 1 ml wide) corresponding to a molecular mass of 151.4 ± 7.5 kDa, i.e., the mass of a possible hexamer, herein termed (C30)₆ (S6B Fig). The peak fractions were then analyzed by SEC-SAXS [47]. SAXS data indicate that (C30)₆ is a globular particle in solution with a radius of gyration, R_g of 63.1 ± 0.1 Å and a maximum intramolecular distance, D_max of 220.0 ± 10.0 Å (Fig 7C). The molecular weight estimation from the SAXS data is of 160.1 ± 15 kDa, in good agreement with the SEC data, and suggesting that in the absence of Tgl the particle is a stable hexamer. The C30 oligomer in solution can accommodate six molecules of Chymotrypsinogen A, a monomeric globular protein with a molecular mass similar to a C30 monomer (25.9 kDa) ([48]; S6C Fig). The smooth asymmetrical pair-wise distance distribution, P(r), indicates that (C30)₆ is slightly aspheric (Fig 7E). The corresponding low-resolution shapes of (C30)₆ reconstructed based on the P(r) [47] (inset in Fig 7D) fit the experimental data well (χ² = 2.2), and provide an oblong envelope model (average NSD = 0.8; see also the material and methods section).

This analysis suggests that C30 forms a large oligomer, whose repeating unit may be the (C30)₆ hexamer detected both in our SEC and SAXS analysis. C30 is detected as a dimer by SDS-PAGE even in the absence of Tgl (Figs 5A, 5B, 5A and S5B) suggesting that this species is highly stable. In the presence of Tgl, (C30)₂ increases rapidly before higher molecular weight forms of the protein accumulate (Fig 5A and S5 Fig; see also above). This suggests that (C30)₆ is a trimer of cross-linked dimers, although other architectures can explain the accumulation of cross-linked (C30)₂. A cross-linked (C30)₆ hexamer in turn, may be an intermediate in the formation of higher molecular weight species. The size of the complexes obtained in the presence of Tgl precluded, for the moment, a study by SAXS. In any event, we posit that (C30)₆ and higher order oligomers accumulate at the surface of the developing spore, and that these pre-assembled complexes recruit Tgl. Cross-linking of these species by Tgl may be self-limiting, in that the enzyme becomes cross-linked into the forming structure, and thereby unable to conduct further catalysis.

Bacterial strains, media and general methods

The construction of all plasmids and strains is described in the supporting material; strains, plasmids and primers used in this study are listed in S1, S2 and S3 Tables.

Fluorescence microscopy

Sporulation was induced by the resuspension method as described [60]. Briefly, cultures were grown in growth medium, and the cells collected at an OD₆₀₀ of ~0.4, and transferred to a minimal medium in which sporulation is induced. Samples were taken 6 and 8 hours after resuspension (defined as the onset of sporulation), the cells collected by centrifugation (1 min at 2.400 x g, room temperature), and washed with 1 ml of phosphate-buffered saline (PBS). Finally, the cells were resuspended in 20 μl of PBS and applied to microscopy slides coated with a film of 1.7% agarose. Images were taken with standard phase contrast, CFP and YFP filters, using a Leica DM 6000B microscope equipped with an aniXon+EM camera (Andor Technologies), and driven by Metamorph software (Meta Imaging series 7.7, Molecular Devices).

Image analysis

Image analysis was conducted with Meta Imaging series 7.7 software and the determination of the values of fluorescence in the forespore or in the mother cell cytoplasm was conducted as described [35]. The individual values of the forespore/cell fluorescence ratio, (F_FS/F_cell), or cell fluorescence, F_cell, were examined with GraphPad Prism 5 (GraphPad Software, Inc) which indicated that the values showed a non-normal distribution. Thus, for normalization purposes each individual value was divided by the median value of the wild type strain that had been grown on the same day and under the same conditions; the same normalization was conducted for the individual values of the wild type strain. The calculation of the normalized values of the forespore/cell fluorescence ratio, (F_FS/F_cell)_norm, or cell fluorescence, (F_cell)_norm, was according to Eqs 1 and 2, respectively: (Eq 1) (Eq 2)

Transmission electron microscopy

Samples were collected from DSM cultures at hour 4 of sporulation and processed from thin sectioning transmission electron microscopy essentially as described [61]. Samples were viewed on a Hitachi H-7650 microscope equipped with an AMT digital camera and operated at 100 keV.

Dansylcadaverine labeling assays

Tgl (WT and mutant forms) and C30 were introduced in a pET derivative (Novagen) fused to a C-terminal His₆-tag and protein over-production was achieved by auto-induction [15]. The different proteins were purified in 1 ml Ni²⁺-NTA Agarose columns (for details see the Supporting Material). Enzyme activity assays were conducted at 37ºC or 50ºC in Tris-HCl 0.1 M, pH 8.0, as described before [14]. The following concentrations were used: BSA (New England Biolabs), 60 μM; dansylcadaverine (Fluka), 0.5 mM; and Tgl^wt/mut (16 μM). BSA/Tgl labeling by dansylcadaverine was detected after SDS-PAGE resolution using a UV transilluminator (Chemidoc XRS, Biorad) and quantified using ImageJ 1.37v [62]. The values obtained for each time point were normalized by the fluorescence detected for the wild type enzyme after 120 min of incubation. All Tgl forms were purified and assayed at least three times independently.

C30 cross-linking assays

Tgl (wild type or mutant forms, at the indicated concentrations) was incubated with C30 (at the indicated concentrations), at 37° C, in Tris-HCl 0.1 M, pH 8.0. Samples were taken at different times and resolved in 10% SDS-PAGE gels, which were subsequently stained with Coomassie brilliant blue. For graphical representation of the data, the stained SDS-PAGE gels were scanned and the amount of C30 or Tgl at each time point was quantified using ImageJ 1.37v [62]. The values obtained for each time point were normalized according to the following equation (where t represents the different time points, in minutes, in a cross-linking reaction):

Small-angle X-ray scattering

Synchrotron SEC-SAXS data were collected on the BM29 ESRF beamline (Grenoble, France) using an in-line HPLC system and used to generate an average low-resolution shape representation of C30 oligomer in solution. Please see S1 Text for details.