• Abstract and Figures
• Public Full-text

Computational modeling and experimental validation PNAS PLUS of the Legionella and Coxiella virulence-related type-IVB secretion signal Ziv Lifshitza,1, David Bursteinb,1, Michael Peerib, Tal Zusmana, Kierstyn Schwartzc, Howard A. Shumanc, Tal Pupkob,2, and Gil Segala,2 Departments of aMolecular Microbiology and Biotechnology and bCell Research and Immunology, George S. Wise Faculty of Life Sciences, Tel-Aviv University, Tel-Aviv 69978, Israel; and cDepartment of Microbiology, University of Chicago, Chicago, IL 60637 Edited by Thomas J. Silhavy, Princeton University, Princeton, NJ, and approved December 28, 2012 (received for review September 5, 2012) Legionella and Coxiella are intracellular pathogens that use the pathogens are unique to one of them, but a few were shown to virulence-related Icm/Dot type-IVB secretion system to translocate contain similar protein motifs, such as ankyrin domains (13, 14). effector proteins into host cells during infection. These effectors In order for an effector to be translocated into the host cell were previously shown to contain a C-terminal secretion signal re- cytoplasm via the Icm/Dot secretion system, the effector must be quired for their translocation. In this research, we implemented recognized by components of the secretion system. The first ef- a hidden semi-Markov model to characterize the amino acid com- fector identified, RalF, was shown to harbor a C-terminal secre- position of the signal, thus providing a comprehensive computa- tion signal (15), and additional analyses of this effector indicated tional model for the secretion signal. This model accounts for that one of its three C-terminal amino acids should be hydro- dependencies among sites and captures spatial variation in amino phobic in order for it to translocate (16). Further analysis using acid composition along the secretion signal. To validate our model, several effectors indicated that in addition to the hydrophobic we predicted and synthetically constructed an optimal secretion amino acid described, the C-terminal secretion signal is enriched signal whose sequence is different from that of any known effector. with tiny, polar, and charged amino acids (e.g., alanine, serine, We show that this signal efficiently translocates into host cells in an threonine, glutamic acid) (17). When the number of known Icm/Dot-dependent manner. Additionally, we predicted in silico and L. pneumophila effectors increased to about 100, an analysis of the experimentally examined the effects of mutations in the secretion C-terminal region of effectors indicated there are amino acids that signal, which provided innovative insights into its characteristics. are enriched and/or depleted in this region. Glutamic acid was Some effectors were found to lack a strong secretion signal accord- found to be enriched at positions −17 to −10 (from the C-terminal ing to our model. We demonstrated that these effectors were highly end) and depleted from the five C-terminal amino acids, which dependent on the IcmS-IcmW chaperons for their translocation, un- were enriched with hydrophobic amino acids (e.g., isoleucine, like effectors that harbor a strong secretion signal. Furthermore, our leucine, valine). In addition, serine and threonine were enriched at model is innovative because it enables searching ORFs for secretion positions −10 to −5 (18). A more recent study (19) described the fi signals on a genomic scale, which led to the identi cation and ex- importance of the glutamic acid stretch (E-block) for effector perimental validation of 20 effectors from Legionella pneumophila, translocation and used this feature to identify several novel Legionella longbeachae,andCoxiella burnetii. Our combined com- effectors. Notably, all these studies were performed only on putational and experimental methodology is general and can be fi L. pneumophila. applied to the identi cation of a wide spectrum of protein fea- In addition to the C-terminal secretion signal described above, tures that lack sequence conservation but have similar amino the IcmS and IcmW proteins were found to function as a chap- acid characteristics. eron complex that assists in the translocation of effector proteins into host cells (20). Many effectors were shown to have a reduced pathogenomics | type-IV secretion | translocated substrates | level of translocation in the absence of this chaperon complex, but Legionnaire disease | Q-fever others were not affected (16). The current information suggests that both the IcmS–IcmW chaperon complex and the C-terminal he bacteria Legionella pneumophila, the causative agent of secretion signal jointly contribute to the translocation efficiency of TLegionnaire disease, and Coxiella burnetii, the causative agent the Icm/Dot effectors (21, 22). of Q-fever, are both human intracellular pathogens that multiply The purpose of this research was to characterize the secretion in alveolar macrophages (1, 2). In nature, L. pneumophila mul- signal of Icm/Dot effectors. To achieve this goal, we have imple- MICROBIOLOGY tiplies in a broad range of amoebae, whereas C. burnetii infects mented a hidden semi-Markov model (HSMM). The HSMM was various ruminants, such as cattle and sheep (3, 4). The in- trained on all 283 L. pneumophila known effectors and a set of tracellular vacuole formed by these two bacteria has been shown noneffectors to identify sequence properties that are unique to to be completely different; L. pneumophila inhibits phagosome- lysosome fusion and resides in a vacuole at almost neutral pH, whereas C. burnetii multiplies in an acidic vacuole (5–8). How- Author contributions: Z.L., D.B., M.P., T.Z., K.S., H.A.S., T.P., and G.S. designed research; Z.L., ever, these two pathogens have been shown to use a similar Icm/ D.B., M.P., T.Z., and K.S. performed research; Z.L., D.B., M.P., T.Z., K.S., H.A.S., T.P., and G.S. – contributed new reagents/analytic tools; Z.L., D.B., M.P., T.Z., K.S., H.A.S., T.P., and G.S. Dot type-IVB secretion system for pathogenesis (9 12). analyzed data; and Z.L., D.B., T.P., and G.S. wrote the paper. The Icm/Dot secretion systems of these bacteria have been The authors declare no conflict of interest. shown to translocate a large number of effector proteins from This article is a PNAS Direct Submission. the bacterial cytoplasm into the host cell during infection (the 1Z.L. and D.B. contributed equally to this work. current number of effectors in L. pneumophila and C. burnetii is 2To whom correspondence may be addressed. E-mail: [email protected] or talp@tauex. estimated as 300 and 100, respectively). The effector proteins tau.ac.il. subvert a wide repertoire of functions in the host cells and direct See Author Summary on page 2709 (volume 110, number 8). the establishment of the unique phagosome formed by these two This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. pathogens (5, 6). Most of the effectors translocated by these two 1073/pnas.1215278110/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1215278110 PNAS | Published online February 4, 2013 | E707–E715 Downloaded by guest on September 25, 2021 the C terminus of effectors. We have used the trained model to strain in comparison to the icmS-icmW double-deletion mutant. predict and experimentally examine a synthetic “optimal” secre- The results obtained show that the translocation efficiency of all tion signal (OSS) that was found to translocate into host cells these effectors was strongly reduced in the icmS-icmW double- efficiently in an Icm/Dot-dependent manner. Furthermore, we deletion mutant (Fig. 1) but that their level of expression was predicted several mutations according to the HSMM and validated similar to the one in the WT strain (Fig. S1). For comparison, we their effect on the translocation of the secretion signal experi- also examined the effect of the IcmS–IcmW chaperon complex on mentally. Finally, the HSMM was used to identify effectors in the translocation efficiency of the two top-scoring effectors L. pneumophila, Legionella longbeachae, and C. burnetii. (lpg1144-CegC3 with a signal score of 16.5 and lpg1290-Lem8 with a signal score of 13.8). The translocation of these two effectors was Results not affected by the absence of the IcmS–IcmW chaperon complex To characterize the C-terminal secretion signal of Icm/Dot ef- (CegC3 and Lem8 in Fig. 1). Altogether, this analysis reveals fectors, we have implemented an HSMM. Based on a set of known the involvement of both the C-terminal secretion signal and the L. pneumophila effector proteins and a set of noneffectors, two IcmS–IcmW chaperon complex in the translocation of Icm/Dot HSMMs were trained: one that best characterizes the C terminus effectors (Discussion). of effectors and one that best characterizes the C terminus of noneffectors. Contrasting these two models allowed character- A Protein Carrying a Synthetic Signal Predicted by the Model Is ization of the physicochemical properties of the secretion signal of Efficiently Translocated into Host Cells in an Icm/Dot-Dependent effectors. We chose the HSMM over position-specific score ma- Manner. One of the advantages of the computational approach trices (23) because the HSMM is more general; it accounts for used in this study is its ability to predict the amino acid sequence dependencies among sites, and thus can better capture spatial of the OSS by inferring the probability of occurrence of each variation in amino acid composition along the secretion signal amino acid in each position of the signal (Materials and Methods). (24, 25). Examination of the computational analysis revealed that amino acids with similar properties were found at high probability in the HSMM Analysis for Identification of the Effectors’ Secretion Signal. same positions (Fig. S2 and Dataset S3). Therefore, we grouped The HSMM we implemented consists of different states, each together amino acids that have similar physicochemical properties representing one or more positions of the C-terminal signal of (ILVF, ED, RK, QN, and TS).

Load more

  9

Copy Link