Barandun et al. BMC Biology 2012, 10:95 http://www.biomedcentral.com/1741-7007/10/95 REVIEW Open Access The pupylation pathway and its role in mycobacteria Jonas Barandun, Cyrille L Delley and Eilika Weber-Ban* Abstract (Figure 1). However, like ubiquitination, tagging with Pup can render proteins as substrates for proteasomal Pupylation is a post-translational protein modification degradation [5, 6, 10]. The existence of a depupylation occurring in actinobacteria through which the small, activity in actinobacteria [11, 12] and the fact that some intrinsically disordered protein Pup (prokaryotic members harbor the pupylation gene locus without ubiquitin-like protein) is conjugated to lysine encoding proteasomal subunits suggest that pupylation residues of proteins, marking them for proteasomal might fulfill a broader role in regulation and cellular degradation. Although functionally related to signaling. The purpose of the pupylation system in ubiquitination, pupylation is carried out by different actinobacteria is still a matter of investigation. In Mtb, enzymes that are evolutionarily linked to bacterial the Pup-proteasome system (PPS) has been linked to the carboxylate-amine ligases. Here, we compare the bacterium’s survival strategy inside the host macrophages mechanism of Pup-conjugation to target proteins with [13, 14]. ubiquitination, describe the evolutionary emergence of pupylation and discuss the importance of this pathway An ubiquitin-like modification pathway in bacteria for survival of Mycobacterium tuberculosis in the host. marks proteins for proteasomal degradation Actinobacteria form a large and diverse phylum with many members living in close association with eukaryotic Post-translational protein modification is a prevalent hosts as either pathogens (Mycobacterium spp.) or means of diversification and regulation in all cells [1]. The symbionts (nitrogen-fixing or gastrointestinal species) functional consequences range from immediate effects [15, 16]. Phylogenetic analysis identified actinobacteria as like changes in protein conformation or stability, one of the earliest prokaryotic lineages. They are known regulation of enzymatic activities to the determination of to share traits with eukaryotes [17]. For example, like subcellular localization. Tags marking substrates for eukaryotes they encode single-chain eukaryotic-like degradation by energy-dependent protease complexes fatty-acid synthase (FASI; in addition to the dissociated exist in pro- and eukaryotes, as exemplified by eukaryotic bacterial FASII enzymes) [18], actinomycetes form ubiquitination [2, 3] or bacterial co-translational ssrA- exospores and mycobacteria produce sterols [17]. tagging [4]. However, until recently, the use of small- Another eukaryotic-like feature is the existence of protein modifiers such as ubiquitin was considered a proteasomes in actinobacteria in addition to the typical feature exclusive to eukaryotic cells. The discovery of bacterial-like compartmentalizing protease complexes pupylation, the covalent modification of protein lysines (Clp proteases [19], FtsH [20], Lon [21], but not with prokaryotic, ubiquitin-like protein Pup, in HslUV) [22]. These bacterial proteases are architecturally Mycobacterium tuberculosis (Mtb) and Mycobacterium related to the proteasome but of only very distant smegmatis [5, 6] and the detection of conjugates between homology [23]. It is still a matter of debate how small archaeal modifier proteins (SAMPs) and substrate actinobacteria came by their proteasomes. One theory lysines in archaea [7, 8] show that prokaryotes also proposes horizontal transfer of the corresponding employ macromolecular tags. It has been demonstrated proteasomal genes from archaea or eukaryotes [22]. In that modification of target proteins with Pup occurs by contrast to that, others suggest that the actinobacterial a chemical pathway distinct from ubiquitination [9] proteasome represents an ancestral form, based on their hypothesis that eukaryotes and archaea derived from *Correspondence: [email protected] actinobacteria [24]. Irrespective of the suggested ETH Zurich, Institute of Molecular Biology & Biophysics, CH-8093 Zurich, Switzerland evolutionary scenarios, the fact remains that no bacterial © 2012 Barandun et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Barandun et al. BMC Biology 2012, 10:95 Page 2 of 9 http://www.biomedcentral.com/1741-7007/10/95 (a) K Pup-Substrate Proteasomal PafA Degradation GGE GGE- K Depupylation PupE (b) Ub- Substrate K48-polyUb Proteasomal E1/E2/E3 Cascade GG G G -K Degradation Deubiquitination Ubiquitin K Figure 1. Bacterial pupylation, like eukaryotic ubiquitination, targets proteins for proteasomal degradation. (a,b) Pupylation (a) or ubiquitination (b) of a target protein is shown. Both small protein modifiers (red) are attached to a lysine side chain of a substrate protein (grey). A random coil model of Pup (red) represents its intrinsically disordered state in solution. In contrast, ubiquitin (Ub) adopts a stable β-grasp fold (PDB 1aar). Note that Ub is linked to the substrate lysine via its carboxy-terminal di-glycine-motif (‘GG’), while Pup is attached via its carboxy-terminal glutamate (‘GGE’). proteasomes were found outside the actinobacterial Ligation of Pup to target lysines on the other hand is phylum beyond a few sporadic cases in other lineages carried out by a single enzyme, the Pup ligase PafA like, for example, nitrospirae [25]. e pupylation (proteasome accessory factor A) [9]. In all mycobacteria machinery of nitrospirae, in fact, was speculated to and many other actinobacteria, preparation of Pup by originate in Acidimicrobiales by horizontal gene transfer another enzyme (Dop, deamidase of Pup) must, however, [26], which seems to be supported by the recent occur before the actual ligation [9]. is can be likened to availability of such a genome [27] (Figure 2). the processing of the Ub-precursor to reveal the carboxy- e post-translational modification Pup that recruits terminal di-glycine motif. proteins for degradation by bacterial proteasomes is Ub adopts a defined three-dimensional structure in functionally related to the eukaryotic ubiquitin (Ub) tag solution referred to as the β-grasp fold [28]. In contrast, without showing any sequence or structural homology Pup is mostly unstructured in its free, unbound form [29- (Figure 1). Both proteins are small (below 10 kDa), both 31]. It has been noted that the carboxy-terminal half of carry a di-glycine motif either at the very carboxyl Pup exhibits a pattern of hydrophobic and hydrophilic terminus (Ub) or at the penultimate position (Pup) and residues typical of coiled-coil formation, and NMR both are attached to the amino group of lysine side chains analysis revealed signals from weak helix formation in in target proteins via an isopeptide bond [5, 6, 9]. that part of the protein [29]. It was therefore suggested However, the enzymatic pathways for attachment are that Pup interacts with the coiled-coil domains that different. Ub is conjugated to substrates in a multi-step extend from the surface of the proteasomal ATPase ring reaction involving a cascade of three enzymes [2], the Ub to form a shared coiled-coil. e crystal structure of a activating enzyme E1, the Ub conjugating enzyme E2 and carboxy-terminal Pup fragment with a fragment of the one of the many Ub-protein ligase E3s that form the Mpa (mycobacterial proteasomal ATPase) coiled-coil isopeptide-bond between a substrate lysine and Ub. domain confirmed this hypothesis, demonstrating that, Barandun et al. BMC Biology 2012, 10:95 Page 3 of 9 http://www.biomedcentral.com/1741-7007/10/95 (a) III I M. smegmatis M. tuberculosis C. diphtheriae Rhodococcus sp. M. leprae N. farcinica C. glutamicum S. erythraea Nocardioides sp. S. coelicolor II K. rhizophila T. fusca A. cellulolyticus M. luteus Frankia sp. S. tropica K. radiotolerans Janibacter sp. A. aurescens R. salmoninarum 0.1 B. mcbrellneri Corynebacterineae A. odontolyticus A. ferrooxidans B. adolescentis Actinomycetales Actinobacteria Nitrospirae Leptospirillum ferrooxidans L1 M. leprae I (b) M. tuberculosis 9 orfs M. smegmatis Rhodococcus sp. N. farcinica 220 orfs S. erythraea Nocardioides sp. II S. coelicolor 10 orfs T. fusca 9 orfs A. cellulolyticus 13 orfs Frankia sp. 1228 orfs S. tropica K. radiotolerans 14 orfs Janibacter sp. 92 kbp L. ferrooxidans A. ferrooxidans A. aurescens R. salmoninarum B. mcbrellneri A. odontolyticus III B. adolescentis M. luteus K. rhizophila C. glutamicum Leptospirillum ferrooxidans C. diphtheriae L2 Mpa / Arc Dop Pup PrcB PrcA Paf A 9 orfs pseudo genes unrelated genes space 1500 bp Figure 2. Occurrence, genomic organization and relatedness of the pupylation gene locus. (a) Phylogenetic analysis of the combined Arc, Dop, Pup and PafA amino acid sequences reveals tight clustering of proteasome-harboring members (clusters I and II), whereas members without proteasomal genes in the pupylation locus exhibit much greater sequence variation (cluster III). The pupylation enzymes of Leptospirillum ferrooxidans, a Nitrospirae exponent, likely originate in a member of the acidimicrobiales, a subclass of the actinobacteria. (b) Genomic context of the pupylation-relevant enzymes. The genomes are listed counter clockwise as they appear