- Email: [email protected]
Helical Packing of Needles from Functionally Altered Shigella Type III Secretion Systems
doi:10.1016/j.jmb.2005.09.062
J. Mol. Biol. (2005) 354, 206–211
C OMMUNICATION
Helical Packing of Needles from Functionally Altered Shigella Type III Secretion Systems Frank S. Cordes1,2, Sarah Daniell3, Roma Kenjale4, Saroj Saurya5 Wendy L. Picking4, William D. Picking4, Frank Booy3, Susan M. Lea1,2 and Ariel Blocker5* 1 Laboratory of Molecular Biophysics, University of Oxford, Oxford OX1 3QU UK
Sir William Dunn School of Pathology, University of Oxford Oxford, Oxon OX1 3RE, UK
Gram-negative bacteria commonly interact with eukaryotic host cells using type III secretion systems (TTSSs or secretons), which comprise cytoplasmic, transmembrane and extracellular domains. The extracellular domain is a hollow needle-like structure protruding 60 nm beyond the bacterial surface. The TTSS is activated to transfer bacterial proteins directly into a host cell only upon physical contact with the target cell. We showed previously that the monomer of the Shigella flexneri needle, MxiH, assembles into a helical structure with parameters similar to those defining the architecture of the extracellular components of bacterial flagella. By analogy with flagella, which are known to exist in different helical states, we proposed that changes in the helical packing of the needle might be used to sense host cell contact. Here, we show that, on the contrary, mutations within MxiH that lock the TTSS into altered secretion states do not detectably alter the helical packing of needles. This implies that either: (1) host cell contact is signalled through the TTSS via helical changes in the needle that are significantly smaller than those linked to structural changes in the flagellar filament and therefore too small to be detected by our analysis methods or (2) that signal transduction in this system occurs via a novel molecular mechanism.
*Corresponding author
Keywords: type III secretion; contact-mediated activation; structure– function relationships; helical supramolecular assemblies; electron microscopy
2
Department of Biochemistry University of Oxford, Oxford OX1 3QU, UK 3
Department of Biological Sciences, Imperial College London SW7 2AZ, UK 4
Department of Molecular Biosciences, University of Kansas, Lawrence, KS60045–7534, USA 5
q 2005 Elsevier Ltd. All rights reserved.
Nature often uses helical symmetry to assemble large machineries from a limited number of components. One such assembly is the bacterial flagellum where the rod segments, hook and filament are each constructed from single subunits arranged in helical arrays. Flagellar assembly and function have been studied in depth,1 revealing that their gross physical properties can be altered by switching of their helical architecture between minutely different structural states.2,3 Flagella are also specialised secretion systems, optimised for efficient secretion and assembly of their components via the 2–3 nm central, axial channel that
Abbreviations used: TTSS, type III secretion system; CR, Congo red; TMV, tobacco mosaic virus. E-mail address of the corresponding author: [email protected]
traverses them. Perhaps unsurprisingly therefore, ten TTSS proteins share distant homologies with flagellar proteins. Additional TTSS proteins show “functional conservation” (mutations in the genes for such components lead to similar defects in assembly or function of either apparatus4), despite showing no significant sequence similarity with flagellar proteins. We determined the structure of the Shigella TTSS needle by X-ray fibre-diffraction and electron ˚ resolution. We found that its microscopy at 16 A structure was defined by similar parameters (w5.6 ˚ helical pitch, 2–3 nm subunits per turn, w24 A central channel) to those of the flagellar rod, hook and filament.5–7 This was remarkable because the needle protein monomer, MxiH (Figure 1), is up to five times smaller and displays no primary sequence homology with any flagellar axial component. Others have shown that the
0022-2836/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
Helical Analysis of Mutant Shigella TTSS Needles
207
Figure 1. Sequence of MxiH and phenotypes of mutants selected for structural studies. Upper part: the amino acid sequence for MxiH is shown with a predicted secondary structure beneath (JPRED; H, helix, E, sheet, (*) unstructured; see Cuff et al. 19). Lower part: the secretion phenotypes13 of the alanine point mutants selected for structural study (boxed in alignment above) are summarised beneath the alignment.
enteropathogenic Escherichia coli EspA filament (a specialised extension of the E. coli TTSS) has a similar structure.8 This suggests that these helical parameters represent central structural constraints in the assembly and function of flagella and TTSSs. To allow proper assembly, flagella and TTSSs require a mechanism to control the order in which their component proteins are secreted.9–12 Flagella switch secretion specificity at the “hook-to-filament transition” while TTSSs arrest secretion once the needle is assembled and are later reactivated to secrete new classes of proteins following contact with a host cell. The mechanisms that control these changes in secretion activity are unclear. Similarities in quaternary structure between needle and flagellar filament7 and in secondary structure between MxiH and the key regions of flagellin,13 led us to propose that, for TTSSs, the activation signal was transmitted mechanically via alterations in the helical structure of the needle.4 Such switching between helical states is seen in flagellar filaments. ˚ These switch their helical architecture (a 0.65 A alteration in the helical pitch) to adapt to changes in the direction of motor rotation by subtle rearrangements of the N and C termini of the flagellin subunits within filaments.2,14 Our model seemed to be supported by structure–function analysis of needle components from two different TTSS of animal pathogens,13,15 which indicated that mutations in the needle protein can lock the needles into altered functional states (Figure 1; see Kenjale et al.13). To investigate whether a change in needle function is accompanied by changes in its structure, we performed a structural analysis of these mutant needles. We generated a three-dimensional density map of the needle using electron microscopy and helical image reconstruction. Here, we have used refinements of these methods to investigate whether
changes in the functional state of the secreton are associated with structural transitions within its needle. We wanted to look for subtle changes in the spacing of layer-lines (due to the helical symmetry of the needle) visualised by Fourier transformation needle images and ask whether they correlated with changes in secreton function. The Shigella needle has a smooth surface structure and therefore the layer-line intensities are relatively weak. We are therefore only able to visualise the layer-lines by averaging the signal from many needles. We considered study of the needle within the secreton base, the “needle complex” (activated either by host–cell contact or by an artificial inducer, Congo red, (CR), as in our previous work. 7 However, the maximum resolution of our published three-dimensional reconstruction derived ˚ from isolated needle complexes was only w24 A because of the limited purity of these samples. This would not permit a sufficiently accurate determination of the location of the layer-line spacings to detect changes in the helical pitch of the magnitude of those seen to control the flagellar filament function. In our earlier work, we also studied “superlong” needles generated by overexpression of the needle subunit MxiH. These superlong needles led to a substantially better three-dimen˚) sional image reconstruction (resolution w16 A with well-resolved layer-lines. Since the helical pitch of the needle in the context of the needle complex was not significantly different from that of longer needles, here we chose to analyse superlong needles throughout. Host–cell contact would provide the most physiologically relevant means of obtaining activated needles for analysis. However, additional sample heterogeneity would also be introduced, since it is not then possible to guarantee activation of all needle complexes on all bacteria used as
208 the source of needles. Since we are unable to see layer-lines reliably in the power spectrum derived from an image of a single needle (but only in the averaged power spectra derived from images of many needles) there is no mechanism by which we can sort the needles into activated and unactivated populations. Activation with CR was also rejected because needles stained with CR were extremely sensitive to electron irradiation (data not shown). For both of these methods of activation there may be further problems in producing a homogenous sample, since we do not know whether the activation signal is irreversible or decays back to an unactivated state over time. Thus, to ensure that all the needles in the preparation were in a homogenous state we chose to work with needles where the activation state is permanently and irreversibly altered by mutagenesis. We have recently completed a detailed mutagenesis study of MxiH (see Kenjale et al.13 and Figure 1). This work revealed that many mutations in MxiH lead to a loss-of-function often due to gross changes in the morphology of the needle. However, three mutations locked the secreton into altered secretion states without leading to obvious alterations in the
Figure 2. Illustration of the steps involved in helical analysis of the needles. Section of an electron micrograph showing filaments of TMV (thick filament) and WT Shigella needles (thin filaments). Long needles were purified by shearing from the surface of Shigella as described7 except that, after overnight induction, 2% (not 10%) weight per volume of polyethylene glycol MW 6000 Da is added to the culture to precipitate free needles at the beginning and end of the purification. Needles are stored in 10 mM Tris (pH 7.4), 150 mM NaCl, 0.02% (w/v) sodium azide at 4 8C and used within two days. Needles and TMV (a gift from J. P. Carr, Cambridge, UK) are diluted appropriately in water and prepared for electron microscopy using a modification of the Valentine staining technique.20 Two microliters of diluted needles are pipetted under a thin carbon film evaporated onto a mica support. The carbon film is washed by partial floatation on water and allowed to detach on a drop of 1% (w/v) uranyl acetate stain in a Teflon block. The carbon film is then collected on a lacey grid (Ted Pella, USA). The needles are photographed under minimal exposure ˚ 2) with a defocus value that placed conditions (w45 eK/A the first zero of the contrast transfer function in the range ˚ K1 using a Philips CM 200 FEG, operating at of 16–18 A 200 kV and magnification 66,000!, and recorded on new Kodak SO-163 films.
Helical Analysis of Mutant Shigella TTSS Needles
morphology of the needle. One mutant, Q51A, was strongly activated for secretion prior to induction with CR (termed constitutively “on”) and showed a further increase in secretion after exposure to CR. The other two mutants, D73A and the double mutant P44ACQ51A, were also constitutively “on” but showed no increase in secretion upon induction (termed “onCuninducible”). These three mutants, which appear morphologically normal but are functionally activated, were analysed here by overexpression of the mutant MxiH sequences to generate “superlong” needles. To ensure that subtle changes in helical state were not obscured by differences in microscope magnification and image processing, we used TMV particles to normalise the scale of all images (Figure 2). We compared the helical packing in wild-type needles from bacteria that had not been induced to secrete with the packing of mutant needles. Figure 3 shows how an averaged power spectrum is generated by averaging all the needles from a single field. In Figure 3(a) (i), the helical layer-lines are clearly visible, although there is some asymmetry between the intensities of the layer-lines on the left and right hand sides of the meridian, (e.g., the 1-start layer-line is weak but visible on one side of the meridian, but not on the other). This asymmetry probably results from uneven staining, which is a known artefact of negative staining.16,17 It cannot result from out of plane tilt, since the layer-line peak(s) in all the powerspectra corresponding to the 1-start helix are well separated from the meridian and at constant positions (data not shown). These peaks, even when of uneven intensity, would be expected to merge at the meridian at a tilt of 58 and disappear entirely at a tilt in excess of 108. Thus, our data indicate that the out of plane tilt is small enough to validate the spacing measurements. Since we make no use of the intensities of the layer-line reflections, but only measure their spacing, the observed asymmetry in layer-line intensities has no implications for our analysis. The mutants examined display different degrees of functional alteration in comparison to the wildtype with the D73A and P44ACQ51A mutations resulting in the most extreme activation phenotype. In contrast, the L46A mutant is essentially wildtype in its phenotype and the Q51A mutant has an intermediate phenotype. The full results of the structural analyses are presented in Table 1 and demonstrate that none of the mutations lead to a significant alteration of the helical packing of needles. We were unable to collect data which allowed measurement of layer-line positions from the double mutant, P44ACQ51A, which has the same phenotype as D73A, since the superlong needles generated from this mutant were less abundant and less stable than those of any of the other needle populations, suggesting that their overall structure may be more severely affected. The close similarity of the helical parameters defining these functionally different mutants to
209
Helical Analysis of Mutant Shigella TTSS Needles
Figure 3. Analysis of electron micrographs to locate layer-line positions. Negatives are scanned on an Imacon Flextight 848 scanner at 2000 dpi. Individual filaments are selected using EmTool,21 and further image analysis carried out using SPIDER/WEB.22 Isolated, straight sections of filaments of the needles and TMV are extracted from each micrograph and divided into partially overlapping segments (each 512 pixels and either 60 pixels wide for needles, or 100 pixels wide for ˚ on the film. The average length of the needles in our TMV), avoiding any ends. One pixel corresponds to w2 A preparations is 250–300 nm. The offset between the segments was five pixels. A power spectrum is computed for each individual segment and added to an average power spectrum derived from all the needles selected from that field (w50 needles). Once powerspectra have been calculated for all images, these are averaged, and the average powerspectrum is used to generate a binary mask that is applied to each individual image. TMV filaments are treated similarly. (a) Average power spectrum generated from the (i) needle and (ii) TMV filaments in image 10 with the intense equator removed for clarity (left). (b) Powerspectra after application of a 3 sigma filter and masking. (c) Integration of intensities along each layer-line plotted against the altitude (red, right hand side of the powerspectrum; black; left-hand side). The altitude of the layer-lines corresponding to the 6-start, 5-start and 1-start helices for needle filaments or the 1-start helix for TMV are measured and averaged from the reflections seen on either side of the equator. For the needle populations, not all layerlines are seen in all images (e.g., in this image, the layer arising from the 1-start helix is not seen in some quadrants). Table 1. Helical parameters for Shigella needle populations studied
WT D73A Q51A L46A
No. of micrographs
˚) Mean helical pitchGSD (A
˚) Mean axial riseGSD (A
7 14 6 11
24.18(G0.19), nZ12 24.18(G0.34), nZ19 24.21(G0.25), nZ9 24.22(G0.23), nZ18
4.31(G0.03), nZ12 4.33(G0.05), nZ15 4.32(G0.05), nZ9 4.31(G0.03), nZ18
Powerspectra are calculated from micrographs as described in the legend for Figure 3. Each powerspectrum (derived from all needles in a single field, generally in the range of 20–50 needles) is carefully examined and the positions of all observed layer-lines noted. The layer-line positions are corrected for subtle differences in magnification between different fields using a powerspectrum derived from TMV particles in the same field. The helical pitch is then determined using both the position of the 1-start layer-lines and the 5 and 6start layer-lines (where seen), as independent measurements (n). The axial rise is then calculated by combining the helical pitch with the altitude of the layer-line arising from either the 5 or 6-start helices. A Student’s t-test indicates that none of the means differs significantly from any of the others. The standard deviations for the parameters describing the needle populations are of the order of 1–1.5% and are ˚ helical of a similar magnitude to the standard deviation calculated when using one-half of the TMV particles as a reference set (23 A ˚ ). pitch) to derive a helical pitch for the other TMV particles present (23.04(G0.15) A
210 those of the wild-type population may be interpreted in two different ways: (i) the structural alterations are simply too small to be resolved with the current techniques or (ii) the phenotypic alterations seen in these mutants do not relate to changes in the architecture of the helix and may instead relate to some subtle effect caused by the specific amino acid substitutions. We cannot distinguish between these hypotheses at this time. However, the maximal alteration in the helical pitch seen in bacterial flagella (from straight-L to straightR form, allowing the complete change in handed˚ (or a 2.5% ness of the flagellar filament) is 0.65 A change in helical pitch) as measured by X-ray fiber diffraction of highly ordered liquid crystalline sols of the filaments.18 Since the standard deviations associated with our estimates of pitch are at the level of 1–1.5% of the mean pitch (Table 1), we are confident that we would be able to detect structural alterations of a similar magnitude. We cannot formally exclude the hypothesis that changes in helical pitch of a much smaller magnitude than those known to be linked to changes in flagellar activity are occurring in our mutant needles and are responsible for the differing activities of the Shigella needle mutants. However, we now think it more likely that the process of signal transduction in this system does not occur by a large-scale shift in the helical structure of the assembly. Instead it could occur by a novel, as yet undiscovered switch mechanism, which may only be revealed once the atomic structure of the protein is solved.
Helical Analysis of Mutant Shigella TTSS Needles
3.
4.
5. 6.
7.
8.
9.
10.
11.
Acknowledgements A.B. thanks Chi Aizawa for initial discussions and an introduction to Keiichi Namba, with whose invaluable aid this entire work was developed. David DeRosier is also acknowledged for key comments on an earlier version of this paper. The electron microscopy was performed at the CBEM, Imperial College London. F.S.C. was funded by a Wellcome Trust Studentship, S.D. and the microscopy by Wellcome Trust Grant (069235) to F.B. and R.K. by a Barbara Johnson Bishop Scholarship. W.P.’s laboratory was supported by PHS grants AI034428 and RR017708 and the University of Kansas Research Development Fund. A.B.’s laboratory was supported by the Guy G. F. Newton Senior Research Fellowship. We dedicate this paper to Kaoru Komoriya.
12.
13.
14. 15.
16.
17.
References 1. Macnab, R. (2003). How bacteria assemble flagella. Annu. Rev. Microbiol. 57, 77–100. 2. Samatey, F. A., Imada, K., Nagashima, S., Vonderviszt, F., Kumasaka, T., Yamamoto, M. & Namba, K. (2001).
18.
Structure of the bacterial flagellar protofilament and implications for a switch for supercoiling. Nature, 410, 331–337. Samatey, F. A., Matsunami, H., Imada, K., Nagashima, S., Shaikh, T. R., Thomas, D. R. et al. (2004). Structure of the bacterial flagellar hook and implication for the molecular universal joint mechanism. Nature, 431, 1062–1068. Blocker, A., Komoriya, K. & Aizawa, S.-I. (2003). Type III secretion systems and bacterial flagella: insights into their function from structural similarities. Proc. Natl Acad. Sci. USA, 100, 3027–3030. Thomas, D., Morgan, D. G. & DeRosier, D. J. (2001). Structures of bacterial flagellar motors from two FliFFliG gene fusion mutants. J. Bacteriol. 183, 6404–6412. Kato, S., Okamoto, M. & Asakura, S. (1984). Polymorphic transition of the flagellar polyhook from Escherichia coli and Salmonella typhimurium. J. Mol. Biol. 173, 463–476. Cordes, F. S., Komoriya, K., Larquet, E., Yang, S., Egelman, E. H., Blocker, A. & Lea, S. M. (2003). Helical structure of the needle of the type III secretion system of Shigella flexneri. J. Biol. Chem. 278, 17103–17107. Daniell, S. J., Kocsis, E., Morris, E., Knutton, S., Booy, F. P. & Frankel, G. (2003). 3D structure of EspA filaments from enteropathogenic Escherichia coli. Mol. Microbiol. 49, 301–308. Chilcott, G. S. & Hughes, K. T. (2000). Coupling of flagellar gene expression to flagellar assembly in Salmonella enterica serovar typhimurium and Escherichia coli. Microbiol. Mol. Biol. Rev. 64, 694–708. Karlinsey, J. E., Pease, A. J., Winkler, M. E., Bailey, J. L. & Hughes, K. T. (1997). The flk gene of Salmonella typhimurium couples flagellar P- and L-ring assembly to flagellar morphogenesis. J. Bacteriol. 179, 2389–2400. Kutsukake, K. (1997). Hook-length control of the export-switching machinery involves a double-locked gate in Salmonella typhimurium flagellar morphogenesis. J. Bacteriol. 179, 1268–1273. Magdalena, J., Hachani, A., Chamekh, M., Jouihri, N., Gounon, P., Blocker, A. & Allaoui, A. (2002). Spa32 regulates a switch in substrate specificity of the type III secreton of Shigella flexneri from needle components to Ipa proteins. J. Bacteriol. 184, 3433–3441. Kenjale, R., Wilson, J., Zenk, S. F., Saurya, S., Picking, W. L., Picking, W. D., & Blocker, A. (2005) The needle component of the type III secreton of Shigella regulates the activity of the secretion apparatus. In the press. Calladine, C. (1982). Construction of bacterial flagellar filaments, and aspects of their conversion to different helical forms. Symp. Soc. Exp. Biol. 35, 33–51. Torruellas, J., Jackson, M. W., Pennock, J. W., & Plano, G. V. (2005). The Yersinia pestis type III secretion needle plays a direct role in the regulation of Yop secretion. Mol. Microbiol. 57, 1719–1733. Boekema, E. J., Van Heel, M. G. & Van Bruggen, E. F. (1984). Three-dimensional structure of bovine NADH: ubiquinone oxidoreductase of the mitochondrial respiratory chain. Biochim. Biophys. Acta, 787, 19–26. Fujiyoshi, Y., Morikawa, K., Uyeda, N., Ozeki, H. & Yamagishi, H. (1983). Electron microscopy of tRNA crystals. I. Thin crystals negatively stained with uranyl acetate. Ultramicroscopy, 12, 210–212. Yamashita, I., Hasegawa, K., Suzuki, H., Vonderviszt, F., Mimori-Kiyosue, Y. & Namba, K. (1998). Structure and switching of bacterial flagellar filaments studied by X-ray fiber diffraction. Nature Struct. Biol. 5, 125–132.
Helical Analysis of Mutant Shigella TTSS Needles
19. Cuff, J. A., Clamp, M. E., Siddiqui, A. S., Finlay, M. & Barton, G. J. (1998). JPred: a consensus secondary structure prediction server. Bioinformatics, 14, 892–893. 20. Valentine, A. F., Chen, P. K., Colwell, R. R. & Chapman, G. B. (1965). Structure of a marine bacteriophage as revealed by the negative-staining technique. J. Bacteriol. 91, 819–822.
211 21. Ludtke, S. J., Baldwin, P. R. & Chiu, W. (1999). EMAN: semiautomated software for high-resolution single-particle reconstructions. J. Struct. Biol. 128, 82–97. 22. Frank, J., Radermacher, M., Penczek, P., Zhu, J., Li, Y., Ladjadj, M. & Leith, A. (1996). SPIDER WEB: processing and visualization of images in 3D electron microscopy and related fields. J. Struct. Biol. 116, 190–199.
Edited by I. B. Holland (Received 13 July 2005; received in revised form 19 September 2005; accepted 20 September 2005) Available online 7 October 2005
J. Mol. Biol. (2005) 354, 206–211
C OMMUNICATION
Helical Packing of Needles from Functionally Altered Shigella Type III Secretion Systems Frank S. Cordes1,2, Sarah Daniell3, Roma Kenjale4, Saroj Saurya5 Wendy L. Picking4, William D. Picking4, Frank Booy3, Susan M. Lea1,2 and Ariel Blocker5* 1 Laboratory of Molecular Biophysics, University of Oxford, Oxford OX1 3QU UK
Sir William Dunn School of Pathology, University of Oxford Oxford, Oxon OX1 3RE, UK
Gram-negative bacteria commonly interact with eukaryotic host cells using type III secretion systems (TTSSs or secretons), which comprise cytoplasmic, transmembrane and extracellular domains. The extracellular domain is a hollow needle-like structure protruding 60 nm beyond the bacterial surface. The TTSS is activated to transfer bacterial proteins directly into a host cell only upon physical contact with the target cell. We showed previously that the monomer of the Shigella flexneri needle, MxiH, assembles into a helical structure with parameters similar to those defining the architecture of the extracellular components of bacterial flagella. By analogy with flagella, which are known to exist in different helical states, we proposed that changes in the helical packing of the needle might be used to sense host cell contact. Here, we show that, on the contrary, mutations within MxiH that lock the TTSS into altered secretion states do not detectably alter the helical packing of needles. This implies that either: (1) host cell contact is signalled through the TTSS via helical changes in the needle that are significantly smaller than those linked to structural changes in the flagellar filament and therefore too small to be detected by our analysis methods or (2) that signal transduction in this system occurs via a novel molecular mechanism.
*Corresponding author
Keywords: type III secretion; contact-mediated activation; structure– function relationships; helical supramolecular assemblies; electron microscopy
2
Department of Biochemistry University of Oxford, Oxford OX1 3QU, UK 3
Department of Biological Sciences, Imperial College London SW7 2AZ, UK 4
Department of Molecular Biosciences, University of Kansas, Lawrence, KS60045–7534, USA 5
q 2005 Elsevier Ltd. All rights reserved.
Nature often uses helical symmetry to assemble large machineries from a limited number of components. One such assembly is the bacterial flagellum where the rod segments, hook and filament are each constructed from single subunits arranged in helical arrays. Flagellar assembly and function have been studied in depth,1 revealing that their gross physical properties can be altered by switching of their helical architecture between minutely different structural states.2,3 Flagella are also specialised secretion systems, optimised for efficient secretion and assembly of their components via the 2–3 nm central, axial channel that
Abbreviations used: TTSS, type III secretion system; CR, Congo red; TMV, tobacco mosaic virus. E-mail address of the corresponding author: [email protected]
traverses them. Perhaps unsurprisingly therefore, ten TTSS proteins share distant homologies with flagellar proteins. Additional TTSS proteins show “functional conservation” (mutations in the genes for such components lead to similar defects in assembly or function of either apparatus4), despite showing no significant sequence similarity with flagellar proteins. We determined the structure of the Shigella TTSS needle by X-ray fibre-diffraction and electron ˚ resolution. We found that its microscopy at 16 A structure was defined by similar parameters (w5.6 ˚ helical pitch, 2–3 nm subunits per turn, w24 A central channel) to those of the flagellar rod, hook and filament.5–7 This was remarkable because the needle protein monomer, MxiH (Figure 1), is up to five times smaller and displays no primary sequence homology with any flagellar axial component. Others have shown that the
0022-2836/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
Helical Analysis of Mutant Shigella TTSS Needles
207
Figure 1. Sequence of MxiH and phenotypes of mutants selected for structural studies. Upper part: the amino acid sequence for MxiH is shown with a predicted secondary structure beneath (JPRED; H, helix, E, sheet, (*) unstructured; see Cuff et al. 19). Lower part: the secretion phenotypes13 of the alanine point mutants selected for structural study (boxed in alignment above) are summarised beneath the alignment.
enteropathogenic Escherichia coli EspA filament (a specialised extension of the E. coli TTSS) has a similar structure.8 This suggests that these helical parameters represent central structural constraints in the assembly and function of flagella and TTSSs. To allow proper assembly, flagella and TTSSs require a mechanism to control the order in which their component proteins are secreted.9–12 Flagella switch secretion specificity at the “hook-to-filament transition” while TTSSs arrest secretion once the needle is assembled and are later reactivated to secrete new classes of proteins following contact with a host cell. The mechanisms that control these changes in secretion activity are unclear. Similarities in quaternary structure between needle and flagellar filament7 and in secondary structure between MxiH and the key regions of flagellin,13 led us to propose that, for TTSSs, the activation signal was transmitted mechanically via alterations in the helical structure of the needle.4 Such switching between helical states is seen in flagellar filaments. ˚ These switch their helical architecture (a 0.65 A alteration in the helical pitch) to adapt to changes in the direction of motor rotation by subtle rearrangements of the N and C termini of the flagellin subunits within filaments.2,14 Our model seemed to be supported by structure–function analysis of needle components from two different TTSS of animal pathogens,13,15 which indicated that mutations in the needle protein can lock the needles into altered functional states (Figure 1; see Kenjale et al.13). To investigate whether a change in needle function is accompanied by changes in its structure, we performed a structural analysis of these mutant needles. We generated a three-dimensional density map of the needle using electron microscopy and helical image reconstruction. Here, we have used refinements of these methods to investigate whether
changes in the functional state of the secreton are associated with structural transitions within its needle. We wanted to look for subtle changes in the spacing of layer-lines (due to the helical symmetry of the needle) visualised by Fourier transformation needle images and ask whether they correlated with changes in secreton function. The Shigella needle has a smooth surface structure and therefore the layer-line intensities are relatively weak. We are therefore only able to visualise the layer-lines by averaging the signal from many needles. We considered study of the needle within the secreton base, the “needle complex” (activated either by host–cell contact or by an artificial inducer, Congo red, (CR), as in our previous work. 7 However, the maximum resolution of our published three-dimensional reconstruction derived ˚ from isolated needle complexes was only w24 A because of the limited purity of these samples. This would not permit a sufficiently accurate determination of the location of the layer-line spacings to detect changes in the helical pitch of the magnitude of those seen to control the flagellar filament function. In our earlier work, we also studied “superlong” needles generated by overexpression of the needle subunit MxiH. These superlong needles led to a substantially better three-dimen˚) sional image reconstruction (resolution w16 A with well-resolved layer-lines. Since the helical pitch of the needle in the context of the needle complex was not significantly different from that of longer needles, here we chose to analyse superlong needles throughout. Host–cell contact would provide the most physiologically relevant means of obtaining activated needles for analysis. However, additional sample heterogeneity would also be introduced, since it is not then possible to guarantee activation of all needle complexes on all bacteria used as
208 the source of needles. Since we are unable to see layer-lines reliably in the power spectrum derived from an image of a single needle (but only in the averaged power spectra derived from images of many needles) there is no mechanism by which we can sort the needles into activated and unactivated populations. Activation with CR was also rejected because needles stained with CR were extremely sensitive to electron irradiation (data not shown). For both of these methods of activation there may be further problems in producing a homogenous sample, since we do not know whether the activation signal is irreversible or decays back to an unactivated state over time. Thus, to ensure that all the needles in the preparation were in a homogenous state we chose to work with needles where the activation state is permanently and irreversibly altered by mutagenesis. We have recently completed a detailed mutagenesis study of MxiH (see Kenjale et al.13 and Figure 1). This work revealed that many mutations in MxiH lead to a loss-of-function often due to gross changes in the morphology of the needle. However, three mutations locked the secreton into altered secretion states without leading to obvious alterations in the
Figure 2. Illustration of the steps involved in helical analysis of the needles. Section of an electron micrograph showing filaments of TMV (thick filament) and WT Shigella needles (thin filaments). Long needles were purified by shearing from the surface of Shigella as described7 except that, after overnight induction, 2% (not 10%) weight per volume of polyethylene glycol MW 6000 Da is added to the culture to precipitate free needles at the beginning and end of the purification. Needles are stored in 10 mM Tris (pH 7.4), 150 mM NaCl, 0.02% (w/v) sodium azide at 4 8C and used within two days. Needles and TMV (a gift from J. P. Carr, Cambridge, UK) are diluted appropriately in water and prepared for electron microscopy using a modification of the Valentine staining technique.20 Two microliters of diluted needles are pipetted under a thin carbon film evaporated onto a mica support. The carbon film is washed by partial floatation on water and allowed to detach on a drop of 1% (w/v) uranyl acetate stain in a Teflon block. The carbon film is then collected on a lacey grid (Ted Pella, USA). The needles are photographed under minimal exposure ˚ 2) with a defocus value that placed conditions (w45 eK/A the first zero of the contrast transfer function in the range ˚ K1 using a Philips CM 200 FEG, operating at of 16–18 A 200 kV and magnification 66,000!, and recorded on new Kodak SO-163 films.
Helical Analysis of Mutant Shigella TTSS Needles
morphology of the needle. One mutant, Q51A, was strongly activated for secretion prior to induction with CR (termed constitutively “on”) and showed a further increase in secretion after exposure to CR. The other two mutants, D73A and the double mutant P44ACQ51A, were also constitutively “on” but showed no increase in secretion upon induction (termed “onCuninducible”). These three mutants, which appear morphologically normal but are functionally activated, were analysed here by overexpression of the mutant MxiH sequences to generate “superlong” needles. To ensure that subtle changes in helical state were not obscured by differences in microscope magnification and image processing, we used TMV particles to normalise the scale of all images (Figure 2). We compared the helical packing in wild-type needles from bacteria that had not been induced to secrete with the packing of mutant needles. Figure 3 shows how an averaged power spectrum is generated by averaging all the needles from a single field. In Figure 3(a) (i), the helical layer-lines are clearly visible, although there is some asymmetry between the intensities of the layer-lines on the left and right hand sides of the meridian, (e.g., the 1-start layer-line is weak but visible on one side of the meridian, but not on the other). This asymmetry probably results from uneven staining, which is a known artefact of negative staining.16,17 It cannot result from out of plane tilt, since the layer-line peak(s) in all the powerspectra corresponding to the 1-start helix are well separated from the meridian and at constant positions (data not shown). These peaks, even when of uneven intensity, would be expected to merge at the meridian at a tilt of 58 and disappear entirely at a tilt in excess of 108. Thus, our data indicate that the out of plane tilt is small enough to validate the spacing measurements. Since we make no use of the intensities of the layer-line reflections, but only measure their spacing, the observed asymmetry in layer-line intensities has no implications for our analysis. The mutants examined display different degrees of functional alteration in comparison to the wildtype with the D73A and P44ACQ51A mutations resulting in the most extreme activation phenotype. In contrast, the L46A mutant is essentially wildtype in its phenotype and the Q51A mutant has an intermediate phenotype. The full results of the structural analyses are presented in Table 1 and demonstrate that none of the mutations lead to a significant alteration of the helical packing of needles. We were unable to collect data which allowed measurement of layer-line positions from the double mutant, P44ACQ51A, which has the same phenotype as D73A, since the superlong needles generated from this mutant were less abundant and less stable than those of any of the other needle populations, suggesting that their overall structure may be more severely affected. The close similarity of the helical parameters defining these functionally different mutants to
209
Helical Analysis of Mutant Shigella TTSS Needles
Figure 3. Analysis of electron micrographs to locate layer-line positions. Negatives are scanned on an Imacon Flextight 848 scanner at 2000 dpi. Individual filaments are selected using EmTool,21 and further image analysis carried out using SPIDER/WEB.22 Isolated, straight sections of filaments of the needles and TMV are extracted from each micrograph and divided into partially overlapping segments (each 512 pixels and either 60 pixels wide for needles, or 100 pixels wide for ˚ on the film. The average length of the needles in our TMV), avoiding any ends. One pixel corresponds to w2 A preparations is 250–300 nm. The offset between the segments was five pixels. A power spectrum is computed for each individual segment and added to an average power spectrum derived from all the needles selected from that field (w50 needles). Once powerspectra have been calculated for all images, these are averaged, and the average powerspectrum is used to generate a binary mask that is applied to each individual image. TMV filaments are treated similarly. (a) Average power spectrum generated from the (i) needle and (ii) TMV filaments in image 10 with the intense equator removed for clarity (left). (b) Powerspectra after application of a 3 sigma filter and masking. (c) Integration of intensities along each layer-line plotted against the altitude (red, right hand side of the powerspectrum; black; left-hand side). The altitude of the layer-lines corresponding to the 6-start, 5-start and 1-start helices for needle filaments or the 1-start helix for TMV are measured and averaged from the reflections seen on either side of the equator. For the needle populations, not all layerlines are seen in all images (e.g., in this image, the layer arising from the 1-start helix is not seen in some quadrants). Table 1. Helical parameters for Shigella needle populations studied
WT D73A Q51A L46A
No. of micrographs
˚) Mean helical pitchGSD (A
˚) Mean axial riseGSD (A
7 14 6 11
24.18(G0.19), nZ12 24.18(G0.34), nZ19 24.21(G0.25), nZ9 24.22(G0.23), nZ18
4.31(G0.03), nZ12 4.33(G0.05), nZ15 4.32(G0.05), nZ9 4.31(G0.03), nZ18
Powerspectra are calculated from micrographs as described in the legend for Figure 3. Each powerspectrum (derived from all needles in a single field, generally in the range of 20–50 needles) is carefully examined and the positions of all observed layer-lines noted. The layer-line positions are corrected for subtle differences in magnification between different fields using a powerspectrum derived from TMV particles in the same field. The helical pitch is then determined using both the position of the 1-start layer-lines and the 5 and 6start layer-lines (where seen), as independent measurements (n). The axial rise is then calculated by combining the helical pitch with the altitude of the layer-line arising from either the 5 or 6-start helices. A Student’s t-test indicates that none of the means differs significantly from any of the others. The standard deviations for the parameters describing the needle populations are of the order of 1–1.5% and are ˚ helical of a similar magnitude to the standard deviation calculated when using one-half of the TMV particles as a reference set (23 A ˚ ). pitch) to derive a helical pitch for the other TMV particles present (23.04(G0.15) A
210 those of the wild-type population may be interpreted in two different ways: (i) the structural alterations are simply too small to be resolved with the current techniques or (ii) the phenotypic alterations seen in these mutants do not relate to changes in the architecture of the helix and may instead relate to some subtle effect caused by the specific amino acid substitutions. We cannot distinguish between these hypotheses at this time. However, the maximal alteration in the helical pitch seen in bacterial flagella (from straight-L to straightR form, allowing the complete change in handed˚ (or a 2.5% ness of the flagellar filament) is 0.65 A change in helical pitch) as measured by X-ray fiber diffraction of highly ordered liquid crystalline sols of the filaments.18 Since the standard deviations associated with our estimates of pitch are at the level of 1–1.5% of the mean pitch (Table 1), we are confident that we would be able to detect structural alterations of a similar magnitude. We cannot formally exclude the hypothesis that changes in helical pitch of a much smaller magnitude than those known to be linked to changes in flagellar activity are occurring in our mutant needles and are responsible for the differing activities of the Shigella needle mutants. However, we now think it more likely that the process of signal transduction in this system does not occur by a large-scale shift in the helical structure of the assembly. Instead it could occur by a novel, as yet undiscovered switch mechanism, which may only be revealed once the atomic structure of the protein is solved.
Helical Analysis of Mutant Shigella TTSS Needles
3.
4.
5. 6.
7.
8.
9.
10.
11.
Acknowledgements A.B. thanks Chi Aizawa for initial discussions and an introduction to Keiichi Namba, with whose invaluable aid this entire work was developed. David DeRosier is also acknowledged for key comments on an earlier version of this paper. The electron microscopy was performed at the CBEM, Imperial College London. F.S.C. was funded by a Wellcome Trust Studentship, S.D. and the microscopy by Wellcome Trust Grant (069235) to F.B. and R.K. by a Barbara Johnson Bishop Scholarship. W.P.’s laboratory was supported by PHS grants AI034428 and RR017708 and the University of Kansas Research Development Fund. A.B.’s laboratory was supported by the Guy G. F. Newton Senior Research Fellowship. We dedicate this paper to Kaoru Komoriya.
12.
13.
14. 15.
16.
17.
References 1. Macnab, R. (2003). How bacteria assemble flagella. Annu. Rev. Microbiol. 57, 77–100. 2. Samatey, F. A., Imada, K., Nagashima, S., Vonderviszt, F., Kumasaka, T., Yamamoto, M. & Namba, K. (2001).
18.
Structure of the bacterial flagellar protofilament and implications for a switch for supercoiling. Nature, 410, 331–337. Samatey, F. A., Matsunami, H., Imada, K., Nagashima, S., Shaikh, T. R., Thomas, D. R. et al. (2004). Structure of the bacterial flagellar hook and implication for the molecular universal joint mechanism. Nature, 431, 1062–1068. Blocker, A., Komoriya, K. & Aizawa, S.-I. (2003). Type III secretion systems and bacterial flagella: insights into their function from structural similarities. Proc. Natl Acad. Sci. USA, 100, 3027–3030. Thomas, D., Morgan, D. G. & DeRosier, D. J. (2001). Structures of bacterial flagellar motors from two FliFFliG gene fusion mutants. J. Bacteriol. 183, 6404–6412. Kato, S., Okamoto, M. & Asakura, S. (1984). Polymorphic transition of the flagellar polyhook from Escherichia coli and Salmonella typhimurium. J. Mol. Biol. 173, 463–476. Cordes, F. S., Komoriya, K., Larquet, E., Yang, S., Egelman, E. H., Blocker, A. & Lea, S. M. (2003). Helical structure of the needle of the type III secretion system of Shigella flexneri. J. Biol. Chem. 278, 17103–17107. Daniell, S. J., Kocsis, E., Morris, E., Knutton, S., Booy, F. P. & Frankel, G. (2003). 3D structure of EspA filaments from enteropathogenic Escherichia coli. Mol. Microbiol. 49, 301–308. Chilcott, G. S. & Hughes, K. T. (2000). Coupling of flagellar gene expression to flagellar assembly in Salmonella enterica serovar typhimurium and Escherichia coli. Microbiol. Mol. Biol. Rev. 64, 694–708. Karlinsey, J. E., Pease, A. J., Winkler, M. E., Bailey, J. L. & Hughes, K. T. (1997). The flk gene of Salmonella typhimurium couples flagellar P- and L-ring assembly to flagellar morphogenesis. J. Bacteriol. 179, 2389–2400. Kutsukake, K. (1997). Hook-length control of the export-switching machinery involves a double-locked gate in Salmonella typhimurium flagellar morphogenesis. J. Bacteriol. 179, 1268–1273. Magdalena, J., Hachani, A., Chamekh, M., Jouihri, N., Gounon, P., Blocker, A. & Allaoui, A. (2002). Spa32 regulates a switch in substrate specificity of the type III secreton of Shigella flexneri from needle components to Ipa proteins. J. Bacteriol. 184, 3433–3441. Kenjale, R., Wilson, J., Zenk, S. F., Saurya, S., Picking, W. L., Picking, W. D., & Blocker, A. (2005) The needle component of the type III secreton of Shigella regulates the activity of the secretion apparatus. In the press. Calladine, C. (1982). Construction of bacterial flagellar filaments, and aspects of their conversion to different helical forms. Symp. Soc. Exp. Biol. 35, 33–51. Torruellas, J., Jackson, M. W., Pennock, J. W., & Plano, G. V. (2005). The Yersinia pestis type III secretion needle plays a direct role in the regulation of Yop secretion. Mol. Microbiol. 57, 1719–1733. Boekema, E. J., Van Heel, M. G. & Van Bruggen, E. F. (1984). Three-dimensional structure of bovine NADH: ubiquinone oxidoreductase of the mitochondrial respiratory chain. Biochim. Biophys. Acta, 787, 19–26. Fujiyoshi, Y., Morikawa, K., Uyeda, N., Ozeki, H. & Yamagishi, H. (1983). Electron microscopy of tRNA crystals. I. Thin crystals negatively stained with uranyl acetate. Ultramicroscopy, 12, 210–212. Yamashita, I., Hasegawa, K., Suzuki, H., Vonderviszt, F., Mimori-Kiyosue, Y. & Namba, K. (1998). Structure and switching of bacterial flagellar filaments studied by X-ray fiber diffraction. Nature Struct. Biol. 5, 125–132.
Helical Analysis of Mutant Shigella TTSS Needles
19. Cuff, J. A., Clamp, M. E., Siddiqui, A. S., Finlay, M. & Barton, G. J. (1998). JPred: a consensus secondary structure prediction server. Bioinformatics, 14, 892–893. 20. Valentine, A. F., Chen, P. K., Colwell, R. R. & Chapman, G. B. (1965). Structure of a marine bacteriophage as revealed by the negative-staining technique. J. Bacteriol. 91, 819–822.
211 21. Ludtke, S. J., Baldwin, P. R. & Chiu, W. (1999). EMAN: semiautomated software for high-resolution single-particle reconstructions. J. Struct. Biol. 128, 82–97. 22. Frank, J., Radermacher, M., Penczek, P., Zhu, J., Li, Y., Ladjadj, M. & Leith, A. (1996). SPIDER WEB: processing and visualization of images in 3D electron microscopy and related fields. J. Struct. Biol. 116, 190–199.
Edited by I. B. Holland (Received 13 July 2005; received in revised form 19 September 2005; accepted 20 September 2005) Available online 7 October 2005