FEBS Letters 585 (2011) 2749–2754

journal homepage: www.FEBSLetters .org

Identification and modelling of a PPM protein phosphatase foldin the Legionella pneumophila deAMPylase SidD

Daniel J. Rigden ⇑University of Liverpool, Institute of Integrative Biology, Crown St., Liverpool L69 7ZB, UK

a r t i c l e i n f o

Article history:Received 7 July 2011Revised 26 July 2011Accepted 3 August 2011Available online 11 August 2011

Edited by Robert B. Russell

Keywords:AMPylationadenylylationPPM protein phosphataseDistant homologyPost-translational modificationORFan

0014-5793/$36.00 � 2011 Federation of European Biodoi:10.1016/j.febslet.2011.08.006

⇑ Fax: +44 151 795 4414.E-mail address: drigden@liv.ac.uk

a b s t r a c t

The intracellular parasitic bacterium Legionella pneumophila subverts host vesicle transportthrough reversible AMPylation of Rab1. The effector enzyme for deAMPylation is SidD. Here a com-plete PPM protein phosphatase fold catalytic domain in SidD is identified and modelled. The SidDmodel reveals insertions and deletions near the metal ion containing catalytic site which presum-ably determine its novel activity. It also sheds light on possible substrate binding residues and high-lights the lack of an obvious group to act as general acid during reaction. Assignment of a PPM foldto SidD offers an important pointer towards identification of further deAMPylases.� 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

Many proteins are post-translationally modified, the resultantchemical changes being important for diverse reasons – structure,stability, localization and regulation among them [1]. The bestknown of these is phosphorylation in which the modification stateof sites can be controlled by the balance of activity between a kinase,transferring a phospho group from ATP to the target amino-acid, anda phosphatase which can hydrolyse off the phospho group restoringthe original side chain structure. In human cells, a large majority ofcovalently bound phosphate on proteins is attached to Ser and Thrresidues [2] with only around 2% on Tyr. Ser and Thr phosphoryla-tion are also more abundant than Tyr phosphorylation on bacteria[3], where phosphorylation in His and Asp residues as part of two-component signalling pathways is also well-characterised.

Protein Ser/Thr phosphatases belong predominantly in threeclasses generally known as phosphoprotein phosphatases (PPPs),metal-dependent protein phosphatases (PPMs) and a group employ-ing Asp residues in catalysis [4]. Although unrelated, the first twoclasses both binding catalytically essential metal divalent cationsin geometrically similar fashion at their active sites [5]. The PPMfamily includes important signalling proteins in human (e.g., PP2C[5]), plants (e.g., HAB1 [6]) and bacteria (e.g., PphA [7]). A PPM do-main is also found in the signalling protein adenylate cyclase in

chemical Societies. Published by E

some species where it regulates the activity of the separate catalyticdomain [8]. Structures from various sources are now available andinvariably contain at least two metal divalent cations in conservedpositions, numbered 1 and 2. A third site is sometimes occupied ina way that correlates with closure of a ‘flap’ region over the catalyticsite and substrate recognition [7,9,10]. Further metals are occasion-ally seen in variable positions but are considered unlikely to be func-tionally relevant [9]. The fullest picture of PPM catalytic mechanismcomes from a series of structures of Mycobacterium smegmatis PstP[11] where, serendipitously, different tetrahedral ligands occupieddifferent positions and were convincingly argued to represent thesubstrate complex (cacodylate-bound form), product complex(phosphate) and incoming substrate (sulphate). No experimentalstructure with substrate has yet been obtained, but inhibitor in thecase of HAB1 [6] and the crystal contact of a symmetry relatedmolecule [10] have each been suggested to be indicative of substratebinding mode.

The essential roles of post-translational modifications in cellu-lar processes offer inviting prospects for pathogens to interfereand subvert host networks, thereby facilitating their growth andvirulence [12]. For example, both prokaryotic [13] and eukaryotic[14] pathogens introduce their own kinases and phosphatases intohost cells and bacteria interfere with host (de)ubiquitination inseveral ingenious ways [15]. A less known post-translational mod-ification, protein AMPylation (or adenylylation) [16,17], is alsostrongly associated with host-pathogen interaction. AlthoughAMPylation was initially characterised as a means to metabolically

lsevier B.V. All rights reserved.

2750 D.J. Rigden / FEBS Letters 585 (2011) 2749–2754

regulate Escherichia coli glutamine synthase (GS) [18], most atten-tion has focussed on pathogens’ AMPylation of host regulatoryGTPases and the resultant interference with phagocytosis [16,17].AMPylation of GS is carried out by the N-terminal adenylyl trans-ferase domain of a dedicated enzyme, GlnE, the homologous C-terminal domain of which carries out the reverse deAMPylationreaction through a phosphorolytic mechanism, forming ADP andreleasing unmodified GS [19,20]. The first characterised pathogenAMPylation enzymes were Fic domains [21,22] from the Fido do-main superfamily [23], a broadly distributed domain whose singlehuman representative also has activity although uncertain in vivosubstrates [24]. Most recently, Legionella pneumophila effectorenzymes catalysing (de)AMPylation of host Rab1 have beencharacterised for their role in the hijacking of host intracellulartransport mechanisms by this intravacuolar parasite. The AMPy-lase SidM (also known as DrrA) contains an adenylyl transferasefold [25]. SidD, recently identified as the deAMPylase [26,27], lackscharacterised close sequence homologues in databases, although alimited, local resemblance between it and two bacterial phospha-tases has been reported [27]. Here we show that SidD shares a dis-tant but unambiguous relationship with protein phosphatases ofthe PPM family covering the entire catalytic domain and enablingmodelling of the SidD structure. SidD is the first non-phosphataseactivity characterised for the PPM group and suggests that thesearch for further suspected deAMPylases should explore phospha-tase families as well as the phosphodiesterases shown to catalysethe reaction in vitro [24].

2. Materials and methods

The sequence of the L. pneumophila strain Philadelphia-1 SidD(code Q5ZSQ2) was retrieved from Uniprot [28]. Sequence data-base searches were carried out locally with the HMMER 3 package(http://hmmer.org; [29]) and domain searches with RPS-BLAST[30]. SidD was then submitted to the HHPRED server [31,32]. Thecatalytic domain located by the initial results was also submittedseparately to the server and to the GeneSilico Metaserver [33], aswere N- and C-terminal flanking regions. Alignments for modelconstruction were obtained from the HHPRED server and from aPcons5 [34] meta analysis at the Metaserver. A structural align-ment of the templates and other related structures was obtainedusing MUSTANG [35] and post-processed to optimise structurallyunaligned sequence with STACCATO [36]. Secondary structure intemplates was defined by STRIDE [37] and a consensus secondarystructure prediction for SidD taken from the Metaserver. Modelswere constructed using MODELLER, using PPM phosphatase struc-tures from Streptococcus agalactiae (PDB code 2pk0; [10]) and Ther-mosynechococcus elongatus (PDB code 2j82; [7]) as templates. Thethree bound manganese ions, cacodylate ion and catalytic watermolecule from the structure of M. smegmatis phosphatase (PDBcode 2jfs; [11]), representing the enzyme-substrate complex [11],were also included in the modelling with metal-Asp carboxylateoxygen distances restrained to 2.2 Å. A final favoured model wasselected on the basis of DOPE scores [38] and stereochemical anal-ysis with PROCHECK [39]. Regional validation of the model wasdone by profile analysis with PROSA [40] and VERIFY_3D [41]. Jal-view [42] and PyMOL (http://pymol.org) were used for manipula-tion and presentation of sequences and structures respectively.

3. Results and discussion

3.1. SidD contains a PPM fold

In the present databases the only apparent sequence relatives ofL. pneumophila strain Philadelphia-1 SidD, even by sensitive

iterative Hidden Markov Model (HHM) searches (http://hmmer.org; [29]), are near-identical (P96% sequence identity) sequencesfrom other strains of L. pneumophila. Nor can any informative hitsbe obtained by searches in secondary sequence databases such asPfam [43] and CDD [44]. Nevertheless, significant hits were ob-tained at the HHPRED server which employs matching betweenquery- and target-derived HMM-profiles and includes a scoringof compatibility of (predicted) secondary structure. The top hit,with a probability of 72%, was for a match of around 110 residuesbetween a central portion of SidD and the structure of a PPM familyphosphatase from T. elongates (PDB code 2j82; [7]). The second hit,although scoring slightly less well – probability of 70% – matched amuch larger region of SidD (residues 90–330) to almost the fulllength of another PPM phosphatase from Sphaerobacter thermophi-lus (PDB code 3pu9; unpublished). The probability scores increasedto 84% and 73% when the matching region alone was resubmitted.Assignment of the central portion of SidD as the catalytic domainleft unmatched portions at both N- and C-termini of around 70and 165 residues, respectively. Although both are predicted to con-tain regular secondary structure, no match could be obtained ineither case by sequence comparisons or fold recognition.

Although the sequence identity of the central domain matchwas low – 15% to 18% – there was an excellent correspondence be-tween the predicted secondary structure of SidD and the secondarystructures of the phosphatases (Fig. 1). More striking was conser-vation of the binding residues of the catalytically essential divalentmetal cations characteristic of the PPM phosphatase family (Fig. 1).Catalysis is discussed in more detail below: here we simply notethat five Asp residues whose side chains coordinate at least onemetal cation are conserved in SidD, as is a Gly residue contributinga main chain carbonyl to ligation of one metal. As previously, thedispersal of these conserved residues along the alignment providesa series of fixed reference points which help improve the overallquality of the alignment [45,46]. Taken together with the clear par-allel between the reactions catalysed by PPM phosphatases and bySidD (Fig. 2), these results demonstrate that SidD and the phospha-tases share a common evolutionary origin. Since PPM phospha-tases have a very broad distribution among eukaryotes andbacteria while SidD appears confined to L. pneumophila, it is clearthat phosphatase activity arose first and that L. pneumophila con-scripted and adapted a phosphatase to serve as a deAMPylationenzyme.

3.2. Possible origins of SidD

Interestingly, some other Legionella species, but not L. pneumo-phila, contain sequences with PPM domains although they are notannotated as phosphatases. They are found in L. drancourtii (Uni-prot C6N5H6; locus LDG_3539) and L. longbeachae (UniprotD3HME8; locus LLO_3182 [47]). However, there is no reason tothink that these bear any particularly recent evolutionary relation-ship with SidD. For example, none of the characteristic insertionsor deletions of SidD with respect to templates (Figs. 1 and 4) arefound in the regular Legionella PPM sequences. Neither PPM en-zyme has an N-terminal extension like SidD. Both have C-terminalextensions, in the case of the L. longbeachae enzyme appearing torepresent a partial duplication of the catalytic domain, but in nei-ther case do they bear any detectable similarity to the C-terminaldomain of SidD.

Another possible origin for SidD would be horizontal genetransfer from the host – amoeba or human. This would be particu-larly plausible since some Legionella genes have clear eukaryoticorigins [48,49]. Furthermore, many of these are substrates of theIcm-Dot Type IV secretion system indicating host-related func-tions. However, these substrates, although including products ofseven other named Sid (substrates of Icm-Dot) genes, do not

Fig. 1. Alignment of L. pneumophila SidD with selected PPM enzymes of known structure. The structures were aligned and SidD added based on profile matching and foldrecognition alignments. Each structure is labelled with PDB code, species name, protein name, and numbering according to the PDB file. Abbreviated species names are Sa forStreptococcus agalactiae, Te (Thermosynechococcus elongatus), Ms (Mycobacterium smegmatis), Hs (Homo sapiens) and At (Arabidopsis thaliana). The numbering of SidD is that ofthe native sequence. Insertions and deletions located near the catalytic site are labelled immediately above the alignment (see also Fig. 4). Predicted secondary structure ofSidD and actual secondary structure of templates are indicated below the alignment. Conserved amino-acids making direct contacts to bound metals in the modelledconformation are shown in white on pink and an additional conserved Asp near the catalytic site as white on brown. These Asp residues are also labelled above the alignment.Other positions are coloured according to conservation and chemical type. The numbers at the beginning and end of the human PP2C (PDB code 1a6q) sequence representstructurally-determined terminal extensions not included in the alignment. The sequences of the superimposed structures include stretches of native sequence not defined insome structures (PDB codes 2j82 and 3kb3).

Fig. 2. Comparison of reactions catalysed by SidD (above) and the archetypal PPM phosphatase PP2C (below).

D.J. Rigden / FEBS Letters 585 (2011) 2749–2754 2751

include SidM or SidD. Eukaryote PP2C enzymes characteristicallycontain C-terminal helical domains (represented in Pfam by thePP2C_C domain PF07830) but no match between these and theC-terminal extension of SidD could be obtained by fold recognition.More tellingly, eukaryote PP2C structures were entirely absentfrom the HMM-profile:HMM-profile matching and fold recognitionresults suggesting that a prokaryotic PPM was the progenitor ofSidD. This may, nevertheless, have entered L. pneumophila via theIcm-Dot Type IV secretion pathway since that system may alsomediate conjugation between bacteria [50].

3.3. The SidD catalytic site

Insertions and deletions in SidD compared to templates – la-belled in Figs. 1 and 4 – could readily be accommodated in the loopsbetween regular secondary structure elements so that, as expected,the SidD model preserves the layered a+b fold of the PPM proteins(Fig. 3). Only a 15 residue insertion from residues 284-298 was ex-cluded from the model since it is too large to build accurately: thefinal model otherwise represents SidD from position 72 to 337.The final chosen model, of 50, had the best normalised DOPE score

Fig. 3. The model of SidD shown in cartoon view, coloured according to secondarystructure. Bound metal ions are shown in purple, the catalytic water in cyan andbound cacodylate in gold. Conserved Asp residues ligating metal ions in thisconformation, representing substrate complex (see text), are shown as pink sticks.An additional conserved Asp near the catalytic site that binds metal in otherconformations is shown as brown sticks. Hydrogen bonds are shown as blue dottedlines.

2752 D.J. Rigden / FEBS Letters 585 (2011) 2749–2754

of -0.2 and a single Ramachandran-disallowed residue located farfrom the catalytic site. Profile analysis of the final model with VER-IFY_3D [41] and PROSA [40] is shown in Supplementary Fig. 1. Theoverall Z-score by PROSA was -4.74, within the range of scores forexperimental structures of the same size. Three regions score poorly(<0) by VERIFY_3D – 82-90, 222-234 and 262-278. Since the sameregions are unfavourably positive by PROSA profile analysis theyshould be considered the least reliable portions of the model.

The alignments used in model construction had already indi-cated conservation of the metal-binding sites between PPMs andSidD but the model confirmed that there were no steric impedi-ments such as nearby insertions to formation of a putative catalytic

Fig. 4. Cross-eye stereo close-up view of the SidD model catalytic site coloured as in Fig.and the main chain path of one of the templates (2pk0) shown as purple cartoon for compto make p–p stacking interactions with the bound adenine ring of AMPylated-Rab1 submolecular surface is shown to illustrate the positioning of the catalytic site at the botto

site in SidD (Figs. 3 and 4). Two of the Asp residues shown in Figs. 3and 4 – Asp92 and Asp110 – have been mutated to Ala with accom-panying loss of catalytic activity [27]. A third mutation Asp60Alahad no effect on activity: this residue lies outside of the assignedcatalytic domain. Outside the environs of the bound metal thereare limitations on the degree to which the model can be used topredict details of SidD substrate binding and catalysis. First, thedistant relationship between SidD and phosphatase templates re-stricts the accuracy of the model, particularly in the vicinity ofthe two insertions near the catalytic site (Figs. 1 and 3). Secondly,the limited number and diversity of SidD homologues precludesthe use of conservation mapping to highlight more important res-idues and place them in a structural context. Nevertheless somebroad conclusions can be drawn.

It is clear that the SidD lies at the bottom of a crevice whosewalls are not formed solely from insertions with respect to thetemplates, but also from regions confidently aligned with and con-served between PPM structures. The AMPylated Tyr77 on the sub-strate Rab1 must therefore detach from its crystallised position, inwhich the adenine ring lies flat on top of a surface Phe residue [25],in order to insert into the relatively enclosed SidD catalytic site.Such insertion may require localised conformational change ofthe main chain in the vicinity of the AMPylated Tyr. Interestingly,the model reveals that two aromatic side chains – Phe238 andTyr113, not generally present in phosphatases (Fig. 1), lie suffi-ciently close to the catalytic site to potentially form p-p stackinginteractions with the adenine group of the AMPylated Rab1 sub-strate (Fig. 4). Such interactions are commonly seen at adeninebinding sites [51] and are also present in the substrate structure[25].

Of the insertions and deletions found in SidD relative to PPMenzymes, three insertions, the larger two I1 and I3 being neigh-bours, and a single deletion D1 lie at the catalytic site (Fig. 4). Itseems reasonable to suppose that these are responsible for theswitch in activity from Ser/Thr protein phosphatase to Tyr proteindeAMPylase. This necessitates two distinct changes (Fig. 2), firstlyin the region binding Ser/Thr in the phosphatases to accommodatethe larger Tyr and secondly to recognise the nucleoside componentof the SidD substrate. There are precedents for the specific correla-tion of loop regions with activity in the PPM superfamily. In pyru-vate dehydrogenase phosphatase, for example, a large insertion

3. Nearby insertions and deletions, also indicated on Fig. 1, are labelled as I1-I3, D1arison in those areas. Also shown and labelled are aromatic residues suitably placed

strate (see text). The three bound metal ions are numbered and a semi-transparentm of a cleft (see text).

D.J. Rigden / FEBS Letters 585 (2011) 2749–2754 2753

fashion a lipoyl group binding pocket [52]. In the case of SidD,however, any attempt to map substrate binding sites in further de-tail seems excessively speculative. Reliable data regarding sub-strate binding to PPM phosphatases would help, but arepresently unavailable. Two rather different proposals about PPMsubstrate binding mode have been made, one based on a crystallo-graphic symmetry contact [10] and another on the structure of aninhibitor [6]. However, modelling of a Tyr residue into the regionsin question shows that large structural shifts would be required forits accommodation into SidD (not shown), even if either of the sug-gestions proves to be correct for the PPM enzymes in question.

Previously, His62 found at the catalytic site of human PP2C hasbeen proposed as a general acid donating a proton to the leavinggroup [53]. In the absence of this residue in mycobacterial PPM,an alternative, His153, was mutated to Ala to see if it fulfilled theproton donor role [11]. However, the lack of a significant effecton activity led the authors to conclude that a water molecule ora group on the protein substrate acted as proton donor [11]. Inthe case of SidD there is again no suitable ionisable group well-positioned to act as a general acid. Furthermore, inspection of thestructure of AMPylated Rab1 [25] shows that there is no His resi-due – the most suitable amino-acid due to its pKa value near neu-trality – close to the site of modification of the SidD substrate. Thisargues against substrate-assisted protonation and leaves a watermolecule as the most likely general acid for the SidD reaction.

4. Conclusions

Lacking readily identifiable sequence homologues, SidD is anORFan protein [54]. Furthermore, it is not even found in every se-quenced L. pneumophila strain, only in three of five compared bySchroeder et al. [55]. Such ORFan sequences have been suggestedlargely to be unrecognised members of known superfamilies [54]and we show here that this is indeed the case for SidD. A majorconclusion of the present work is in demonstrated catalytic reac-tion diversity in the PPM family previously characterised as beingcomposed exclusively of Ser/Thr protein phosphatases. In this re-spect, the PPM family now joins others which harbour both proteinphosphatase activities and other hydrolase functions [56,57].

Since a SidD deletion mutant replicates normally in host cells[26,58] there may be limited interested in developing small mole-cule SidD inhibitors for pharmacological purposes. However, suchcompounds would still be useful as aids to further experimentalcharacterisation of SidD. Thus, the distant homology between SidDand PPM phosphatases, coupled with conservation of the catalyticsite, suggests that it may well be worth mining inhibitors of thelatter – both small molecules [59] and peptides [60,61] – for com-pounds that modulate SidD activity in vitro and in vivo.

Despite the limited distribution of SidD, the complete foldassignment reported here, which extends the observation of Tanand Luo [27], is an important precedent for deAMPylating activityin families of protein phosphatases. Families of phosphodiester-ases, such as that previously shown to contain an in vitro deAMP-ylase activity [24], have hitherto been considered more likely toharbour deAMPylases. The human genome contains around 17PPM(-like) sequences, according to UniProt, some still largelyuncharacterised. Might one or more of these catalyse the elusivedeAMPylase activity expected to counteract the AMPylation activ-ity of the human FID domain protein ?[24]

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.febslet.2011.08.006.

References

[1] Walsh, C.T. (2005) Posttranslational Modification of Proteins: ExpandingNature’s Inventory, Roberts and Co., Greenwood Village, Colorado, USA.

[2] Olsen, J.V., Blagoev, B., Gnad, F., Macek, B., Kumar, C., Mortensen, P. and Mann,M. (2006) Global, in vivo, and site-specific phosphorylation dynamics insignaling networks. Cell 127, 635–648.

[3] Macek, B., Gnad, F., Soufi, B., Kumar, C., Olsen, J.V., Mijakovic, I. and Mann, M.(2008) Phosphoproteome analysis of E. coli reveals evolutionary conservationof bacterial Ser/Thr/Tyr phosphorylation. Mol. Cell. Proteomics 7, 299–307.

[4] Shi, Y. (2009) Serine/threonine phosphatases: mechanism through structure.Cell 139, 468–484.

[5] Das, A.K., Helps, N.R., Cohen, P.T. and Barford, D. (1996) Crystal structure of theprotein serine/threonine phosphatase 2C at 2.0 A resolution. EMBO J. 15,6798–6809.

[6] Melcher, K., Ng, L.M., Zhou, X.E., Soon, F.F., Xu, Y., Suino-Powell, K.M., Park, S.Y.,Weiner, J.J., Fujii, H., Chinnusamy, V., Kovach, A., Li, J., Wang, Y., Li, J., Peterson,F.C., Jensen, D.R., Yong, E.L., Volkman, B.F., Cutler, S.R., Zhu, J.K. and Xu, H.E.(2009) A gate-latch-lock mechanism for hormone signalling by abscisic acidreceptors. Nature 462, 602–608.

[7] Schlicker, C., Fokina, O., Kloft, N., Grune, T., Becker, S., Sheldrick, G.M. andForchhammer, K. (2008) Structural analysis of the PP2C phosphatase tPphAfrom Thermosynechococcus elongatus: a flexible flap subdomain controlsaccess to the catalytic site. J. Mol. Biol. 376, 570–581.

[8] Feger, G., De Vendittis, E., Vitelli, A., Masturzo, P., Zahn, R., Verrotti, A.C.,Kavounis, C., Pal, G.P. and Fasano, O. (1991) Identification of regulatoryresidues of the yeast adenylyl cyclase. EMBO J. 10, 349–359.

[9] Wehenkel, A., Bellinzoni, M., Schaeffer, F., Villarino, A. and Alzari, P.M. (2007)Structural and binding studies of the three-metal center in two mycobacterialPPM Ser/Thr protein phosphatases. J. Mol. Biol. 374, 890–898.

[10] Rantanen, M.K., Lehtio, L., Rajagopal, L., Rubens, C.E. and Goldman, A. (2007)Structure of Streptococcus agalactiae serine/threonine phosphatase. Thesubdomain conformation is coupled to the binding of a third metal ion.FEBS J. 274, 3128–3137.

[11] Bellinzoni, M., Wehenkel, A., Shepard, W. and Alzari, P.M. (2007) Insights intothe catalytic mechanism of PPM Ser/Thr phosphatases from the atomicresolution structures of a mycobacterial enzyme. Structure 15, 863–872.

[12] Ribet, D. and Cossart, P. (2010) Pathogen-mediated posttranslationalmodifications: a re-emerging field. Cell 143, 694–702.

[13] DeVinney, R., Steele-Mortimer, O. and Finlay, B.B. (2000) Phosphatases andkinases delivered to the host cell by bacterial pathogens. Trends Microbiol. 8,29–33.

[14] Blader, I.J. and Saeij, J.P. (2009) Communication between Toxoplasma gondiiand its host: impact on parasite growth, development, immune evasion, andvirulence. APMIS 117, 458–476.

[15] Rytkonen, A. and Holden, D.W. (2007) Bacterial interference of ubiquitinationand deubiquitination. Cell. Host Microbe 1, 13–22.

[16] Itzen, A., Blankenfeldt, W. and Goody, R.S. (2011) Adenylylation: renaissanceof a forgotten post-translational modification. Trends Biochem. Sci. 36, 221–228.

[17] Woolery, A.R., Luong, P., Broberg, C.A. and Orth, K. (2010) AMPylation:something old is new again. Front. Microbiol. 1, 113.

[18] Kingdon, H.S., Shapiro, B.M. and Stadtman, E.R. (1967) Regulation of glutaminesynthetase. 8. ATP: glutamine synthetase adenylyltransferase, an enzyme thatcatalyzes alterations in the regulatory properties of glutamine synthetase.Proc. Natl. Acad. Sci. USA 58, 1703–1710.

[19] Xu, Y., Zhang, R., Joachimiak, A., Carr, P.D., Huber, T., Vasudevan, S.G. and Ollis,D.L. (2004) Structure of the N-terminal domain of Escherichia coli glutaminesynthetase adenylyltransferase. Structure 12, 861–869.

[20] Xu, Y., Carr, P.D., Vasudevan, S.G. and Ollis, D.L. (2010) Structure of theadenylylation domain of E. coli glutamine synthetase adenylyl transferase:evidence for gene duplication and evolution of a new active site. J. Mol. Biol.396, 773–784.

[21] Garcia-Pino, A., Christensen-Dalsgaard, M., Wyns, L., Yarmolinsky, M.,Magnuson, R.D., Gerdes, K. and Loris, R. (2008) Doc of prophage P1 isinhibited by its antitoxin partner Phd through fold complementation. J. Biol.Chem. 283, 30821–30827.

[22] Luong, P., Kinch, L.N., Brautigam, C.A., Grishin, N.V., Tomchick, D.R. and Orth, K.(2010) Kinetic and structural insights into the mechanism of AMPylation byVopS Fic domain. J. Biol. Chem. 285, 20155–20163.

[23] Kinch, L.N., Yarbrough, M.L., Orth, K. and Grishin, N.V. (2009) Fido, a novelAMPylation domain common to fic, doc, and AvrB. PLoS One 4, e5818.

[24] Worby, C.A., Mattoo, S., Kruger, R.P., Corbeil, L.B., Koller, A., Mendez, J.C.,Zekarias, B., Lazar, C. and Dixon, J.E. (2009) The fic domain: regulation of cellsignaling by adenylylation. Mol. Cell 34, 93–103.

[25] Muller, M.P., Peters, H., Blumer, J., Blankenfeldt, W., Goody, R.S. and Itzen, A.(2010) The Legionella effector protein DrrA AMPylates the membrane trafficregulator Rab1b. Science 329, 946–949.

[26] Neunuebel, M.R., Chen, Y., Gaspar, A.H., Backlund Jr, P.S., Yergey, A. andMachner, M.P. (2011) De-AMPylation of the small GTPase Rab1 by thepathogen Legionella pneumophila. Science 333, 453–456.

[27] Tan, Y. and Luo, Z. (2011) Legionella pneumophila SidD is a deAMPylase thatmodifies Rab1. Nature. 475, 506–509.

[28] Consortium, UniProt. (2011) Ongoing and future developments at theUniversal Protein Resource. Nucleic Acids Res. 39, D214–D219.

2754 D.J. Rigden / FEBS Letters 585 (2011) 2749–2754

[29] Johnson, L.S., Eddy, S.R. and Portugaly, E. (2010) Hidden Markov model speedheuristic and iterative HMM search procedure. BMC Bioinform. 11, 431.

[30] Wheeler, D.L., Barrett, T., Benson, D.A., Bryant, S.H., Canese, K., Chetvernin, V.,Church, D.M., DiCuccio, M., Edgar, R., Federhen, S., Geer, L.Y., Kapustin, Y.,Khovayko, O., Landsman, D., Lipman, D.J., Madden, T.L., Maglott, D.R., Ostell, J.,Miller, V., Pruitt, K.D., Schuler, G.D., Sequeira, E., Sherry, S.T., Sirotkin, K.,Souvorov, A., Starchenko, G., Tatusov, R.L., Tatusova, T.A., Wagner, L. andYaschenko, E. (2007) Database resources of the National Center forBiotechnology Information. Nucleic Acids Res. 35, D5–D12.

[31] Soding, J. (2005) Protein homology detection by HMM-HMM comparison.Bioinformatics 21, 951–960.

[32] Soding, J., Biegert, A. and Lupas, A.N. (2005) The HHpred interactive server forprotein homology detection and structure prediction. Nucleic Acids Res. 33,W244–W248.

[33] Kurowski, M.A. and Bujnicki, J.M. (2003) GeneSilico protein structureprediction meta-server. Nucleic Acids Res. 31, 3305–3307.

[34] Wallner, B. and Elofsson, A. (2005) Pcons5: combining consensus, structuralevaluation and fold recognition scores. Bioinformatics 21, 4248–4254.

[35] Konagurthu, A.S., Whisstock, J.C., Stuckey, P.J. and Lesk, A.M. (2006)MUSTANG: a multiple structural alignment algorithm. Proteins 64, 559–574.

[36] Shatsky, M., Nussinov, R. and Wolfson, H.J. (2006) Optimization of multiple-sequence alignment based on multiple-structure alignment. Proteins 62, 209–217.

[37] Heinig, M. and Frishman, D. (2004) STRIDE: a web server for secondarystructure assignment from known atomic coordinates of proteins. NucleicAcids Res. 32, W500–W502.

[38] Shen, M.Y. and Sali, A. (2006) Statistical potential for assessment andprediction of protein structures. Protein Sci. 15, 2507–2524.

[39] Laskowski, R.A., MacArthur, M.W., Moss, D.S. and Thornton, J.M. (1993)PROCHECK: a program to check the stereochemical quality of proteinstructures. J. Appl. Cryst. 26, 283–291.

[40] Sippl, M.J. (1993) Recognition of errors in three-dimensional structures ofproteins. Proteins 17, 355–362.

[41] Luthy, R., Bowie, J.U. and Eisenberg, D. (1992) Assessment of protein modelswith three-dimensional profiles. Nature 356, 83–85.

[42] Waterhouse, A.M., Procter, J.B., Martin, D.M., Clamp, M. and Barton, G.J. (2009)Jalview Version 2 – a multiple sequence alignment editor and analysisworkbench. Bioinformatics 25, 1189–1191.

[43] Finn, R.D., Mistry, J., Tate, J., Coggill, P., Heger, A., Pollington, J.E., Gavin, O.L.,Gunasekaran, P., Ceric, G., Forslund, K., Holm, L., Sonnhammer, E.L., Eddy, S.R.and Bateman, A. (2010) The Pfam protein families database. Nucleic Acids Res.38, D211–D222.

[44] Marchler-Bauer, A., Lu, S., Anderson, J.B., Chitsaz, F., Derbyshire, M.K.,DeWeese-Scott, C., Fong, J.H., Geer, L.Y., Geer, R.C., Gonzales, N.R., Gwadz,M., Hurwitz, D.I., Jackson, J.D., Ke, Z., Lanczycki, C.J., Lu, F., Marchler, G.H.,Mullokandov, M., Omelchenko, M.V., Robertson, C.L., Song, J.S., Thanki, N.,Yamashita, R.A., Zhang, D., Zhang, N., Zheng, C. and Bryant, S.H. (2011) CDD: aconserved domain database for the functional annotation of proteins. NucleicAcids Res. 39, D225–D229.

[45] Thomas, J., Rigden, D.J. and Cronan, J.E. (2007) Acyl carrier proteinphosphodiesterase (AcpH) of Escherichia coli is a non-canonical member ofthe HD phosphatase/phosphodiesterase family. Biochemistry 46, 129–136.

[46] Rigden, D.J. (2004) A distant evolutionary relationship between GPI-specificphospholipase D and bacterial phosphatidylcholine-preferring phospholipaseC. FEBS Lett. 569, 229–234.

[47] Cazalet, C., Gomez-Valero, L., Rusniok, C., Lomma, M., Dervins-Ravault, D.,Newton, H.J., Sansom, F.M., Jarraud, S., Zidane, N., Ma, L., Bouchier, C., Etienne,J., Hartland, E.L. and Buchrieser, C. (2010) Analysis of the Legionellalongbeachae genome and transcriptome uncovers unique strategies to causeLegionnaires’ disease. PLoS Genet. 6, e1000851.

[48] Lurie-Weinberger, M.N., Gomez-Valero, L., Merault, N., Glockner, G.,Buchrieser, C. and Gophna, U. (2010) The origins of eukaryotic-like proteinsin Legionella pneumophila. Int. J. Med. Microbiol. 300, 470–481.

[49] de Felipe, K.S., Pampou, S., Jovanovic, O.S., Pericone, C.D., Ye, S.F., Kalachikov, S.and Shuman, H.A. (2005) Evidence for acquisition of Legionella type IVsecretion substrates via interdomain horizontal gene transfer. J. Bacteriol. 187,7716–7726.

[50] Segal, G., Purcell, M. and Shuman, H.A. (1998) Host cell killing and bacterialconjugation require overlapping sets of genes within a 22-kb region of theLegionella pneumophila genome. Proc. Natl. Acad. Sci. USA 95, 1669–1674.

[51] Mao, L., Wang, Y., Liu, Y. and Hu, X. (2004) Molecular determinants for ATP-binding in proteins: a data mining and quantum chemical analysis. J. Mol. Biol.336, 787–807.

[52] Vassylyev, D.G. and Symersky, J. (2007) Crystal structure of pyruvatedehydrogenase phosphatase 1 and its functional implications. J. Mol. Biol.370, 417–426.

[53] Jackson, M.D., Fjeld, C.C. and Denu, J.M. (2003) Probing the function ofconserved residues in the serine/threonine phosphatase PP2Calpha.Biochemistry 42, 8513–8521.

[54] Fischer, D. and Eisenberg, D. (1999) Finding families for genomic ORFans.Bioinformatics 15, 759–762.

[55] Schroeder, G.N., Petty, N.K., Mousnier, A., Harding, C.R., Vogrin, A.J., Wee, B.,Fry, N.K., Harrison, T.G., Newton, H.J., Thomson, N.R., Beatson, S.A., Dougan, G.,Hartland, E.L. and Frankel, G. (2010) Legionella pneumophila strain 130bpossesses a unique combination of type IV secretion systems and novel Dot/Icm secretion system effector proteins. J. Bacteriol. 192, 6001–6016.

[56] Tonks, N.K. (2006) Protein tyrosine phosphatases: from genes, to function, todisease. Nat. Rev. Mol. Cell Biol. 7, 833–846.

[57] Rigden, D.J. (2008) The histidine phosphatase superfamily: structure andfunction. Biochem. J. 409, 333–348.

[58] Luo, Z.Q. and Isberg, R.R. (2004) Multiple substrates of the Legionellapneumophila Dot/Icm system identified by interbacterial protein transfer.Proc. Natl. Acad. Sci. USA 101, 841–846.

[59] Rogers, J.P., Beuscher 4th, A.E., Flajolet, M., McAvoy, T., Nairn, A.C., Olson, A.J.and Greengard, P. (2006) Discovery of protein phosphatase 2C inhibitors byvirtual screening. J. Med. Chem. 49, 1658–1667.

[60] Yamaguchi, H., Durell, S.R., Feng, H., Bai, Y., Anderson, C.W. and Appella, E.(2006) Development of a substrate-based cyclic phosphopeptide inhibitor ofprotein phosphatase 2Cdelta, Wip1. Biochemistry 45, 13193–13202.

[61] Chuman, Y., Yagi, H., Fukuda, T., Nomura, T., Matsukizono, M., Shimohigashi, Y.and Sakaguchi, K. (2008) Characterization of the active site and a uniqueuncompetitive inhibitor of the PPM1-type protein phosphatase PPM1D.Protein Pept. Lett. 15, 938–948.