Protein Modification: Bacterial Effectors Rewrite the Rules of Ubiquitylation

Current Biology

Dispatches known among animals, a measure often correlating with morphological complexity [18] and thus highlighting Intoshia’s simplicity. Unlike in most other animals, developmentally important Hox genes are not organized in a cluster and many are missing. Until recently it had been questioned whether orthonectids had a true nervous system [7], despite being able to coordinate muscular movement. It is now known that a simple nervous system exists, and this is also in part reflected in their genome, which contains a large repertoire of neural genes, although certain gene families and pathways have been lost. Therefore, despite being highly simplified, the orthonectid genome still retains elements of the genetic toolkit that characterizes bilaterians. The new study [1] has not solved all mysteries surrounding orthonectids but provides a solid step forward towards a better understanding of one of the most peculiar clades of animals and the first genetic data in two decades. Future phylogenetic analyses could be directed towards subsets of genes with low evolutionary rates [19] to be able to pinpoint more precisely their position in the tree of life. The available genome will also facilitate much needed evo-devo work in both the plasmodial stage and the worm-like adults. There is little doubt that the future of orthonectid research looks brighter than it ever has before. REFERENCES 1. Giard, A. (1878). On the Orthonectida, a new class of animals parasitic on Echinodermata and Turbellaria. Ann. Mag. Nat. Hist. Zool. Bot. Geol. 1, 181–183. 2. Mikhailov, K.V., Slyusarev, G.S., Nikitin, M., Logacheva, M.D., Penin, A.A., Aleoshin, V.V., and Panchin, Y.V. (2016). The genome of Intoshia linei affirms orthonectids as highly simplified spiralians. Curr. Biol. 26, 1768–1774. 3. Brusca, R.C., Moore, W., and Shuster, S.M. (2016). Invertebrates, 3rd ed. (Sunderland, MA: Sinauer Associates). 4. Kozloff, E.N. (1992). The genera of the phylum Orthonectida. Cah. Biol. Mar. 33, 377–406. 5. Kozloff, E.N. (1969). Morphology of the orthonectid Rhopalura ophiocomae. J. Parasitol. 55, 171–195. 6. Slyusarev, G.S. (1994). Fine structure of the female Intoshia variabili (Alexandrov and Sljusarev) (Mesozoa: Orthonectida). Acta Zool. 75, 311–321.

7. Slyusarev, G.S., and Starunov, V.V. (2016). The structure of the muscular and nervous systems of the female Intoshia linei (Orthonectida). Org. Divers. Evol. 16, 65–71. 8. Slyusarev, G.S. (2008). Phylum Orthonectida: morphology, biology, and relationships to other multicellular animals [in Russian]. Zh. Obshch. Biol. 69, 403–427. 9. Deheyn, D., Watson, N.A., and Jangoux, M. (1998). Symbioses in Amphipholis squamata (Echinodermata, Ophiuroidea, Amphiuridae): geographical variation of infestation and effect of symbionts on the host’s light production. Int. J. Parasitol. 28, 1413–1424. 10. Slyusarev, G.S., and Kristensen, R.M. (2003). Fine structure of the ciliated cells and ciliary rootlets of Intoshia variabili (Orthonectida). Zoomorphology 122, 33–39. 11. Maslakova, S.A., Martindale, M.Q., and Norenburg, J.L. (2004). Fundamental properties of the spiralian developmental program are displayed by the basal nemertean Carinoma tremaphoros (Palaeonemertea, Nemertea). Dev. Biol. 267, 342–360. 12. Hanelt, B., Van Schyndel, D., Adema, C.M., Lewis, L.A., and Loker, E.S. (1996). The phylogenetic position of Rhopalura ophiocomae (Orthonectida) based on 18S ribosomal DNA sequence analysis. Mol. Biol. Evol. 13, 1187–1191. 13. Pawlowski, J., Montoya Burgos, J.I., Fahrni, J.F., Wuest, J., and Zaninetti, L. (1996).

Origin of the Mesozoa inferred from 18S rRNA gene sequences. Mol. Biol. Evol. 13, 1128– 1132. 14. Hejnol, A., Obst, M., Stamatakis, A., Ott, M., Rouse, G.W., Edgecombe, G.D., Martinez, P., Bagun˜a`, J., Bailly, X., Jondelius, U., et al. (2009). Assessing the root of bilaterian animals with scalable phylogenomic methods. Proc. Biol. Sci. 276, 4261–4270. 15. Laumer, C.E., Bekkouche, N., Kerbl, A., Goetz, F., Neves, R.C., Sørensen, M.V., Kristensen, R.M., Hejnol, A., Dunn, C.W., Giribet, G., et al. (2015). Spiralian phylogeny informs the evolution of microscopic lineages. Curr. Biol. 25, 2000–2006. 16. Struck, T.H., Wey-Fabrizius, A.R., Golombek, A., Hering, L., Weigert, A., Bleidorn, C., Klebow, S., Iakovenko, N., Hausdorf, B., Petersen, M., et al. (2014). Platyzoan paraphyly based on phylogenomic data supports a noncoelomate ancestry of Spiralia. Mol. Biol. Evol. 31, 1833–1849. 17. Philippe, H., Zhou, Y., Brinkmann, H., Rodrigue, N., and Delsuc, F. (2005). Heterotachy and long-branch attraction in phylogenetics. BMC Evol. Biol. 5, 50. 18. Levine, M., and Tjian, R. (2003). Transcription regulation and animal diversity. Nature 424, 147–151. 19. Ferna´ndez, R., Hormiga, G., and Giribet, G. (2014). Phylogenomic analysis of spiders reveals nonmonophyly of orb weavers. Curr. Biol. 24, 1772–1777.

Protein Modification: Bacterial Effectors Rewrite the Rules of Ubiquitylation Jason M. Berk and Mark Hochstrasser* Department of Molecular Biophysics & Biochemistry, Yale University, New Haven, CT 06520, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cub.2016.05.032

A family of virulence factors from the bacterial pathogen Legionella pneumophila has been discovered to modify human Rab GTPases with ubiquitin. Surprisingly, this modification occurs via a noncanonical mechanism that uses nicotinamide adenine dinucleotide as a cofactor. Humans can be infected by the bacterium Legionella pneumophila when macrophages within lung alveoli phagocytose bacteria introduced by contaminated aerosolized water [1,2]. The

bacteria proliferate in the macrophages, and in severe cases infection can lead to Legionnaires’ disease, a potentially fatal form of pneumonia [3,4]. The processes of infection and propagation in

Current Biology 26, R523–R542, July 11, 2016 ª 2016 Elsevier Ltd. R539

Current Biology

Dispatches A

E1

E3

+

Ub

ATP

PPi Mg²+

Ub

E1

SH

AMP AMP SH

E2 E1 S

Substrate

SH

E2

Ub

B

S

E2

S

Ub

E3

Ub

Substrate

Ub

Substrate

Rab33b

SdeA

+

Ub

NAD

Nicotinamide

?

SdeA Ub

Ub

AMP

Rab33b

Ribosyl-ADP

? AMP Ub

SdeA Current Biology

Figure 1. Comparison of the canonical enzymatic cascade and novel SdeA ubiquitylation mechanisms. (A) Classical ubiquitylation mechanism. An E1 enzyme adenylates the carboxy-terminal glycine of ubiquitin (Ub) using ATP and Mg2+ as cofactors. The adenylated Ub is attacked by the E1 active site cysteine creating a thioester bond with Ub and releasing AMP. E1 and E2 interaction allows for transthioesterification of Ub to the E2 active site cysteine. Finally, an E3 binds to a substrate and an E2–Ub intermediate to catalyze the transfer of Ub to the substrate. (B) SdeA-mediated ubiquitylation. Using NAD as the only cofactor, Ub is covalently attached to substrates (for example, Rab33b or SdeA itself) in a two-step reaction with Ub–ribosyl-ADP on Arg42 as an intermediate, and AMP and nicotinamide as known reaction products. The mechanism for Ub transfer to the substrate, the chemical linkage and residues required of Ub and the substrate for conjugation remain unknown.

macrophages generally mirror those seen in environmental amoebas, which are the natural hosts of L. pneumophila. Engulfment of a nonpathogenic bacterium by an amoeba leads to its encapsulation in a phagosome, which travels through the endosomal pathway to eventually fuse with a lysosome. The acidic and protease-rich environment of the resulting phagolysosome serves to break down the bacterium and release simple nutrients for the amoeba. By contrast, when the pathogenic bacterium L. pneumophila is engulfed, it triggers an elaborate mechanism to impair lysosomal fusion, allowing for intravacuolar replication in the host [1,2]. Within minutes of uptake, the Legionella-containing vacuole (LCV) becomes surrounded by membranes derived from the endoplasmic reticulum (ER) and mitochondria. Coating of the vacuole with these membranes results in expansion of the vacuole to accommodate bacterial propagation and also helps to prevent host detection. Intravacuolar replication eventually leads to host-cell rupture and death, and release of the bacteria. Formation of the LCV is brought about by bacterial effector proteins, which hijack the host-membrane trafficking system. L. pneumophila encodes nearly

300 effector proteins [2], which are translocated across the doublemembrane bacterial envelope and into the host cytoplasm by the Dot/Icm type IV secretion system. This translocation machinery is required for L. pneumophila virulence. Of the many effectors, most remain biochemically uncharacterized, often because they lack detectable similarity to known proteins. Functional redundancy also frequently masks the contribution to virulence of individual effector proteins because single effector-gene deletions rarely impair bacterial replication. A recent study from Qiu and colleagues [5] has now uncovered a previously unappreciated mechanism of action for a known set of Legionella effectors. In analyzing the SidE effector family, the authors made the unexpected discovery that these large multidomain proteins catalyze covalent modification of human host proteins with the protein ubiquitin. Remarkably, SidE-catalyzed protein ubiquitylation uses a mechanism entirely different from the highly conserved mechanism described in textbooks (Figure 1A). The SidE proteins are required for virulence, but their complete array of functions has not been fully elucidated. Qiu et al. [5] used protein-sequence

R540 Current Biology 26, R523–R542, July 11, 2016

comparisons to uncover a conserved mono-adenosine diphosphateribosyltransferase (mART) motif within the central domain of all members of this effector family (SdeA, SdeB, SdeC and SidE) [5]. The mART motif normally catalyzes mono-ADP-ribosylation of arginine residues in protein substrates; the ADP-ribose is derived from nicotinamide adenine dinucleotide (NAD). While ADP-ribosyltransferase activity is required for the virulence of other bacterial effectors [6], no such activity was initially detected for SdeA. Nevertheless, Qiu et al. [5] found that mutating the mART motif prevented two previously described SdeA activities: causing cellular toxicity when expressed in yeast [7,8] and promoting proficient bacterial replication in an amoeboid host [9]. The mutation also protected mammalian cells from SdeA-induced protein secretion defects. These results all point to a mART-motif-dependent enzymatic activity required for virulence. Formation of the LCV depends on modulation of host Rab GTPases, molecular switches that regulate membrane trafficking [10]. Given that the LCV is composed of host-ER-derived membrane and the SidE family associates with the cytoplasmic face of the vacuole during Legionella infection [8], Qiu et al. [5] analyzed multiple Rab GTPases for potential post-translational modifications. Transient co-transfection of human cells with SdeA and a panel of Rab GTPases revealed that SdeA specifically altered ER Rab GTPases, as seen by changes in their mobility on denaturing SDSpolyacrylamide gels, and did so in a mART-motif-dependent manner. Of the four modified Rab GTPases, Rab33b appeared to be the preferred target of SdeA. Mass spectrometric analysis of modified Rab33b yielded peptide sequences derived from ubiquitin. Importantly, ubiquitin-modified Rab33b was also detected after Legionella infection and this ubiquitylation required an intact mART motif in SdeA. Ubiquitylation is a common posttranslational modification in all eukaryotes and is required for cellular processes ranging from cell division to DNA repair. The highly conserved 76-residue ubiquitin protein is covalently attached to target proteins through a well-characterized enzymatic cascade [11] (Figure 1A). In

Current Biology

Dispatches a reaction dependent upon both ATP and Mg2+, an activating enzyme (E1) adenylates the ubiquitin carboxy-terminal glycine, and this reaction is followed by formation of a thioester bond between ubiquitin and a cysteine side chain of the E1. Ubiquitin is then transferred to the active-site cysteine of a second enzyme (E2), again forming a highenergy thioester intermediate. Finally, a ubiquitin ligase (E3) catalyzes transfer of ubiquitin to a substrate, usually forming an amide bond with a lysine side chain on the substrate. SdeA was initially tested as an E3 ligase for Rab33b using several human E2s and an E1 in vitro, but none of the E2–SdeA combinations tested was capable of ubiquitylating the GTPase. When in vitro ubiquitylation reactions were performed with mammalian cell lysates, Rab33b ubiquitylation was readily detected, suggesting a different E2 might be required. However, two key experiments indicated that SdeA catalyzes ubiquitylation of Rab33b independently of any E1 and E2. First, if the added cell lysate was first boiled, which denatures most proteins, the in vitro ubiquitylation of Rab33b was unaffected. Second, lysates from Escherichia coli, which lack all components of the classical ubiquitylation machinery, were sufficient for Rab33b in vitro ubiquitylation. Further analysis revealed the only component of the lysate required for SdeA-catalyzed Rab33b ubiquitylation was NAD, the known cofactor for mART proteins (Figure 1B). Unlike canonical ubiquitin–protein ligation, SdeA-mediated ubiquitylation did not require the diglycine motif at the carboxyl terminus of ubiquitin or any of its seven lysines. Instead, Arg42 of ubiquitin was required for formation of ADP-ribosylated ubiquitin, a reaction intermediate only detected when sequences required for the selfubiquitylation of SdeA were deleted. The ADP-ribosylated intermediate might be subject to subsequent nucleophilic attack by a reactive residue of either SdeA or a substrate protein, although the complete mechanism of ubiquitylation remains unknown, including the nature of the chemical bond between ubiquitin and substrate. Determining which residues of ubiquitin and substrate are covalently linked may help clarify the

reaction mechanism. A novel linkage could provide a potential tool in screening for additional NAD-dependent ubiquitinligase substrates by mass spectrometry. SdeA-mediated ubiquitylation only modestly affected Rab33b GTP loading and hydrolysis, and thus probably does not account for toxicity of the SidE protein family. Instead, there are probably additional host targets and/or downstream effects of SdeA-mediated Rab ubiquitylation. Qiu et al. [5] do not address the possibility that ubiquitylation of Rab33b might affect interactions with its normal binding targets [12,13]. Notably, the SidE effectors all contain active deubiquitylating enzyme (DUB) domains [14,15]. Although the aminoterminal DUB domain is dispensable for both bacterial virulence and SidE protein toxicity in yeast, this region appears to be important for the ability of another Legionella effector, SidJ, to suppress SidE protein toxicity in yeast and mammalian cell culture [7,8]. How these two enzymatic domains of SidE and SidJ function and interact during infection appears to be quite complex and will be an interesting challenge to resolve. The unexpected discovery of a ubiquitinconjugation reaction that is independent of ATP, E1 or E2 will motivate future work into its enzymology and biological consequences, such as in host–pathogen interactions. Structural characterization of the linkage and reaction intermediates will be crucial for understanding this unusual ubiquitin–protein modification. Both in vitro and in vivo experiments revealed a clear laddering of ubiquitin– Rab33b bands on SDS-PAGE gels, suggesting that multiple ubiquitins can be attached to a single Rab33b molecule. These species could represent multiple monoubiquitin or polyubiquitin chain attachments to Rab33b or, less likely, covalent linkage of a ubiquitin(s) to multiple Rab33 polypeptides. An exciting possibility is that other NAD-dependent ubiquitin ligases exist, potentially in organisms other than bacteria. Bioinformatic analyses have identified a plethora of potential ADPribosyltransferase bacterial effectors [6]. Some of these may also have NADdependent ubiquitin ligase activity. In humans, two of the four annotated mART-containing proteins (ART3 and

ART4) have no detectable ADPribosyltransferase activity and thus are candidates for being ubiquitin ligases [16,17]. A single mART enzyme may function as both an ADP-ribosyltransferase and a ubiquitin ligase, as seems to be the case for SdeA, which ADP-ribosylates ubiquitin and uses this intermediate to ubiquitylate substrate proteins. Notably, ubiquitin-Arg42 is conserved in ubiquitin-like proteins (UBLs) such as NEDD8 and SUMO1. Some mART enzymes might therefore ligate UBLs—or even non-UBLs—to other proteins. REFERENCES 1. Isberg, R.R., O’Connor, T.J., and Heidtman, M. (2009). The Legionella pneumophila replication vacuole: making a cosy niche inside host cells. Nat. Rev. Microbiol. 7, 13–24. 2. Xu, L., and Luo, Z.Q. (2013). Cell biology of infection by Legionella pneumophila. Microbes Infect. 15, 157–167. 3. McDade, J.E., Shepard, C.C., Fraser, D.W., Tsai, T.R., Redus, M.A., and Dowdle, W.R. (1977). Legionnaires’ disease: isolation of a bacterium and demonstration of its role in other respiratory disease. N. Engl. J. Med. 297, 1197–1203. 4. Fraser, D.W., Tsai, T.R., Orenstein, W., Parkin, W.E., Beecham, H.J., Sharrar, R.G., Harris, J., Mallison, G.F., Martin, S.M., McDade, J.E., et al. (1977). Legionnaires’ disease: description of an epidemic of pneumonia. N. Engl. J. Med. 297, 1189–1197. 5. Qiu, J., Sheedlo, M.J., Yu, K., Tan, Y., Nakayasu, E.S., Das, C., Liu, X., and Luo, Z.Q. (2016). Ubiquitination independent of E1 and E2 enzymes by bacterial effectors. Nature 533, 120–124. 6. Simon, N.C., Aktories, K., and Barbieri, J.T. (2014). Novel bacterial ADP-ribosylating toxins: structure and function. Nat. Rev. Microbiol. 12, 599–611. 7. Havey, J.C., and Roy, C.R. (2015). Toxicity and SidJ-mediated suppression of toxicity require distinct regions in the SidE family of Legionella pneumophila effectors. Infect. Immun. 83, 3506–3514. 8. Jeong, K.C., Sexton, J.A., and Vogel, J.P. (2015). Spatiotemporal regulation of a Legionella pneumophila T4SS substrate by the metaeffector SidJ. PLoS Pathog. 11, e1004695. 9. Bardill, J.P., Miller, J.L., and Vogel, J.P. (2005). IcmS-dependent translocation of SdeA into macrophages by the Legionella pneumophila type IV secretion system. Mol. Microbiol. 56, 90–103. 10. Sherwood, R.K., and Roy, C.R. (2013). A Rabcentric perspective of bacterial pathogenoccupied vacuoles. Cell Host Microbe 14, 256–268.

Current Biology 26, R523–R542, July 11, 2016 R541

Current Biology

Dispatches 11. Berndsen, C.E., and Wolberger, C. (2014). New insights into ubiquitin E3 ligase mechanism. Nat. Struct. Mol. Biol. 21, 301–307. 12. Valsdottir, R., Hashimoto, H., Ashman, K., Koda, T., Storrie, B., and Nilsson, T. (2001). Identification of rabaptin-5, rabex-5, and GM130 as putative effectors of rab33b, a regulator of retrograde traffic between the Golgi apparatus and ER. FEBS Lett. 508, 201–209. 13. Doring, T., and Prange, R. (2015). Rab33B and its autophagic Atg5/12/16L1 effector assist in hepatitis B virus naked capsid

formation and release. Cell. Microbiol. 17, 747–764. 14. Sheedlo, M.J., Qiu, J., Tan, Y., Paul, L.N., Luo, Z.Q., and Das, C. (2015). Structural basis of substrate recognition by a bacterial deubiquitinase important for dynamics of phagosome ubiquitination. Proc. Natl. Acad. Sci. USA 112, 15090–15095. 15. Ronau, J.A., Beckmann, J.F., and Hochstrasser, M. (2016). Substrate specificity of the ubiquitin and Ubl proteases. Cell Res. 26, 441–456.

R542 Current Biology 26, R523–R542, July 11, 2016

16. Okazaki, I.J., and Moss, J. (1999). Characterization of glycosylphosphatidylinositiol-anchored, secreted, and intracellular vertebrate mono-ADP-ribosyltransferases. Annu. Rev. Nutr. 19, 485–509. 17. Glowacki, G., Braren, R., Firner, K., Nissen, M., Kuhl, M., Reche, P., Bazan, F., CetkovicCvrlje, M., Leiter, E., Haag, F., et al. (2002). The family of toxin-related ecto-ADPribosyltransferases in humans and the mouse. Protein Sci. 11, 1657–1670.