ENZYMOLOGY Bacterial pseudokinase catalyzes protein … · ENZYMOLOGY Bacterial pseudokinase catalyzes protein polyglutamylation to inhibit the SidE-family ubiquitin ligases Miles

ENZYMOLOGY

Bacterial pseudokinase catalyzesprotein polyglutamylation to inhibitthe SidE-family ubiquitin ligasesMiles H. Black1, Adam Osinski1, Marcin Gradowski2, Kelly A. Servage1,3,Krzysztof Pawłowski2,4, Diana R. Tomchick5,6, Vincent S. Tagliabracci1,7,8*

Enzymes with a protein kinase fold transfer phosphate from adenosine 5′-triphosphate(ATP) to substrates in a process known as phosphorylation. Here, we show that theLegionella meta-effector SidJ adopts a protein kinase fold, yet unexpectedly catalyzesprotein polyglutamylation. SidJ is activated by host-cell calmodulin to polyglutamylate theSidE family of ubiquitin (Ub) ligases. Crystal structures of the SidJ-calmodulin complexreveal a protein kinase fold that catalyzes ATP-dependent isopeptide bond formationbetween the amino group of free glutamate and the g-carboxyl group of an active-siteglutamate in SidE. We show that SidJ polyglutamylation of SidE, and the consequentinactivation of Ub ligase activity, is required for successful Legionella replication in a viableeukaryotic host cell.

In search of divergent members of the pro-tein kinase superfamily, we analyzed thetype 4 secretion system (T4SS) effector rep-ertoire from the genus Legionella, a group ofGram-negative bacteria that includes Legionella

pneumophila, the causative agent of Legionnaires’disease. L. pneumophila secretes more than 300proteins via the T4SS to target host-cell signal-ing and establish a replicative niche known asthe Legionella-containing vacuole (LCV) (1–3).Using the fold and function assignment system

(FFAS) sequence profile method (4), we deter-mined that the bacterial effector SidJ (lpg2155)bears borderline sequence similarity to proteinkinases. SidJ acts as a “meta-effector” to antago-nize four functionally redundant effector paralogs,which share >40% sequence identity (SdeA, SdeB,SdeC, and SidE; collectively known as the SidEfamily) (5, 6). The SidE effectors are required forLegionella replication in host cells and SdeA canfunctionally complement a deletion of all fourSidE family members (7).The SidE effectors successively employ aden-

osine 5′-diphosphate (ADP)–ribosyltransferase(ART) and phosphodiesterase (PDE) activitiesto catalyze covalent ligation of ubiquitin (Ub)to Ser residues on substrate proteins indepen-dently of E1 or E2Ub-activating enzymes (8–10).

The ART domain of SidE ADP-ribosylates Ub onArg42, which is subsequently converted by thePDE domain of SidE to phosphoribosyl (pR) Ubor linked to serine residues of target proteinsthrough a Ser-pR-Ub linkage (fig. S1). Ubiquiti-nation of host proteins, including Rab33 (8) andRtn4 (9), is required for proper formation of theLCV (11). However, unrestrained SidE activity isharmful to the host, poisoning the cellular Ubpool by phosphoribosylation and possibly block-ing the action of other Legionella effectors thatmanipulate the host Ub machinery (5, 6, 10).Because it would be detrimental to the bacte-rium to damage its host, Legionella uses SidJ tomodulate the potent SidE family. Whereas theSidE effectors are translocated into the host cellduring the early stages of infection, SidJ levelsrise slowly, permitting temporal control of SidEactivity as the LCV matures (6, 12).SidJ resides in a genomic locus with SdeA,

SdeB, and SdeC and can suppress the toxicityof each of them, as well as that of SidE, whichresides in a distinct genomic locus (Fig. 1A). Weidentified a putative protein kinase domain (KD)in SidJ spanning residues 336 to 593 (Fig. 1B).Although we could tentatively assign the active-site ion-pair residue K72 (PKA nomenclature,SidJ; K367) and the metal-binding residues N171and D184 (SidJ; N534 and D542) (Fig. 1C), thelocations of the Gly-rich loop and the catalyticloop were ambiguous, suggesting that SidJ maybe a pseudokinase. The putative metal-bindingD542 lies within a 542DXXD motif (where “X” isany amino acid), previously shown to be requiredfor successful Legionella replication in amoeba(6) and in mammalian cells (12).Expression of SdeA or SdeC caused a strong

growth-inhibition phenotype in yeast that waspartially suppressed by coexpression of SidJ, butnot the K367A and D542A mutants (Fig. 1D andfig. S2A). Host-cell E1 enzymes activate Ub fortransfer by adenylating its C-terminal glycine

residue (13); however, this residue is not re-quired for SidE-catalyzed ubiquitination (8). Wemutated the C-terminal GG motif of Ub to AA(UbGG/AA) for expression in mammalian cells tointerrogate SidE-catalyzed ubiquitination. Ex-pression of hemagglutinin (HA)–tagged wild-type(WT) Ub, but not UbGG/AA, resulted in the mod-ification of several host proteins by the endoge-nous ubiquitination machinery (Fig. 1E, lanes1 and 2). When UbGG/AA was coexpressed withMyc-tagged SdeA, but not the E860A (ART do-main) or the H407A (PDE domain) mutants, wedetected abundant ubiquitination of host pro-teins (Fig. 1E, lanes 3 to 5). Coexpression of SidJ,but not the K367A or D542A mutants, markedlyinhibited SdeA-catalyzed ubiquitination (Fig.1E, lanes 6 to 8). Likewise, deletion of SidJ inLegionella led to an accumulation of pR-linkedUb in infected mammalian cells, which was ame-liorated by complementation with WT SidJ, butnot the K367A mutant (fig. S2B).We expressed Flag-epitope–tagged fusions of

SidJ (Flag-SidJ) in yeast. Analysis of anti-Flagimmunoprecipitates by liquid chromatography–tandem mass spectrometry (LC-MS/MS) identi-fied calmodulin (CaM) in SidJ and SidJK367A, butnot in control immunoprecipitates (fig. S3A).A region within SidJ (residues 841 to 851) rough-ly follows the consensus sequence of a CaM-interacting “IQ” motif (fig. S3B). We generatedmutations that we predicted would disrupt CaMbinding by substituting I841, Q842, R843, andR846 with acidic residues (fig. S3B, SidJIQ). SidJimmunoprecipitated endogenous CaM when ex-pressed in mammalian cells, but the SidJIQ mu-tant did not (Fig. 1F). A stable truncation of SidJ(residues 59 to 851, SidJDNC), which was suitablefor purification in Escherichia coli, suppressedSdeA-catalyzed ubiquitination (fig. S3C), formeda stable complex with CaM during gel-filtrationchromatography (fig. S3D), and bound CaMwitha KD of ~30 nM (Fig. 1G). CaM binding was re-quired for SidJ to inhibit SdeA-catalyzed ubiq-uitination in mammalian cells (Fig. 1H). Thus,SidJ requires kinase catalytic residues and CaMbinding to counteract SidE.Although we were unable to observe phos-

phorylation of SdeA by SidJ in vitro, we observedan electrophoreticmobility shift of SdeA residues178 to 1100 (SdeADNC, PDE and ART domains)that formed only when WT SidJ and CaM werecoexpressed inE. coli (Fig. 2A). Intactmass analysisrevealed a series of mass increases in incrementsof ~128.99 Da (Fig. 2B). Searching a database ofmass tags (14), we observed a match to gluta-mylation, a posttranslational modification of theg-carboxyl of a glutamate side chain modifiedwith the amino group of free glutamate formingan isopeptide bond (Fig. 2C). LC-MS/MS analysisidentified peptides of SdeA modified by one ortwo glutamates (Fig. 2D and fig. S4A). Using[U-14C]Glu, we reconstituted glutamylation invitro. The reaction required CaM, ATP/Mg2+,and the kinase residues in SidJ (Fig. 2E).Glutamylation of SdeADNC was time dependent(fig. S4B) and preferred Glu over Gln, Asp, Lys,or Gly (fig. S4C). SidJ also glutamylated SdeB,

RESEARCH

Black et al., Science 364, 787–792 (2019) 24 May 2019 1 of 6

1Department of Molecular Biology, University of TexasSouthwestern Medical Center, Dallas, TX 75390, USA.2Warsaw University of Life Sciences, Warsaw, Poland.3Howard Hughes Medical Institute, Dallas, TX 75390, USA.4Lund University, Lund, Sweden. 5Department of Biophysics,University of Texas Southwestern Medical Center, Dallas, TX75390, USA. 6Department of Biochemistry, University ofTexas Southwestern Medical Center, Dallas, TX 75390, USA.7Harold C. Simmons Comprehensive Cancer Center,University of Texas Southwestern Medical Center, Dallas, TX75390, USA. 8Hamon Center for Regenerative Science andMedicine, University of Texas Southwestern Medical Center,Dallas, TX 75390, USA.*Corresponding author. Email: [email protected]

on July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from

Black et al., Science 364, 787–792 (2019) 24 May 2019 2 of 6

Fig. 1. A putative KD and CaM binding are required for SidJsuppression of SdeA toxicity and noncanonical ubiquitination.(A) Organization of the SidE family (red) and SidJ (teal) effectors in thegenome of L. pneumophila. (B) Domain architecture of L. pneumophilaSidJ depicting the location of the predicted KD. (C) Sequence logos(weblogos) highlighting conserved kinase active-site residues in 106 SidJhomologs and 3998 homologs of typical protein kinases (Pfam domainPF00069). The height of the amino acid stack is proportional to thesequence conservation at that position. (D) Growth-inhibition assaydepicting the growth of Saccharomyces cerevisiae expressing SdeA-GFPand Flag-SidJ or the predicted inactive K367A and D542A mutants.EV, empty vector. (E) Protein immunoblotting of total extracts fromHEK293A cells expressing HA-Ub, HA-UbGG/AA, Myc-SdeA, andSidJ-V5 or the indicated mutants. GAPDH is shown as a loading control.

(F) Protein immunoblotting of V5 immunoprecipitates and cellextracts from HEK293A cells expressing V5-tagged SidJ, SidJD542A,or the SidJIQ mutant. Immunoprecipitates and cell extracts wereanalyzed for CaM and SidJ. GAPDH is shown as a loading control.(G) Isotherms depicting the binding of human CaM to SidJD542A DNC.The top panel shows the SVD-reconstructed thermograms (DP, differ-ential power); the bottom panel shows the isotherms. Resultsare reported as best fit with boundaries of 68.3% confidence interval.(H) Protein immunoblotting of HEK293A total cell extracts expressingHA-tagged UbGG/AA, Myc-SdeA, V5-tagged SidJ, SidJD542A, or theSidJIQ mutant. GAPDH is shown as a loading control. Single-letterabbreviations for the amino acid residues are as follows: A, Ala;C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met;N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

RESEARCH | REPORTon July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from

SdeC, and SidE (fig. S4D). Whereas both yeastand human CaM activated SidJ to glutamylateSdeA, a relatedprotein, troponinC, didnot (fig. S4E).Furthermore, SidJ purified from Legionella alsoglutamylated SdeADNC (fig. S4F).LC-MS/MS andmutagenesis revealed the cat-

alytic E860 of SdeA as the major site of gluta-mylation in vitro (Fig. 2, F and G). Likewise,Myc-tagged SdeA isolated from mammalian

cells coexpressing WT SidJ, but not the D542Amutant, was polyglutamylated on E860 (Fig. 2H).We coexpressed Myc-tagged SdeA with FcgRIIaand challenged human embryonic kidney (HEK)293 cells with antibody-opsonized Legionella (15).SdeA isolated from cells challenged with the WTstrain (Lp02), but not the DsidJ (fig. S5, A to C) orT4SS mutants, was glutamylated on E860, asdetermined by LC-MS/MS (fig. S5D). Thus, SidJ

is a CaM-dependent protein polyglutamylase thatmodifies a catalytic glutamate residue in the ARTactive site of the SidE effectors in vitro and duringinfection.Glu860 is part of the conserved catalytic R-S-

E860XEmotif present in Arg-specific ARTs. Struc-tures of SdeA and SidEwithUb revealed that E860plays a key role in catalysis (Fig. 2F) (16–18). Weused the SdeA H407A mutant, which inactivates

Black et al., Science 364, 787–792 (2019) 24 May 2019 3 of 6

Fig. 2. SidJ polyglutamylates the SidE effectors. (A) SDS–polyacrylamidegel electrophoresis (SDS-PAGE) and Coomassie staining of SdeADNC

isolated from E. coli after coexpression with CaM, SidJ, or the indicatedmutants. (B) Intact mass LC/MS spectra of SdeADNC from (A).(C) Structures depicting glutamylation and polyglutamylation. (D) MS/MSspectrum of monoglutamylated (Glu) SdeADNC peptide ion HGEGTE(Glu)SEFSVYLPEDVALVPVK. The precursor ion, mass/charge ratio (m/z)878.10 (3+) and labeled with “x,” was subjected to higher-energy collisionaldissociation fragmentation to generate the MS/MS spectrum shown.Fragment b-ions containing the modified glutamate residue show a massshift consistent with the addition of one Glu group (+129.043 Da) (red

labels). Peaks labeled with a single asterisk (*) correspond to loss ofwater (–18 Da) from fragment ions. (E) Incorporation of 14C-Glu intoSdeADNC by SidJDNC. Reaction products were separated by SDS-PAGEand visualized by Coomassie staining (top) and autoradiography(bottom). (F) Magnification of the NAD+-binding pocket of SdeA [PDB5YIJ (16)]. (G) Incorporation of 14C-Glu into WT SdeADNC, but notthe E860A mutant. Reaction products were analyzed as in (E).(H) Histogram depicting the MS/MS spectral matches to unmodified,glutamylated, and polyglutamylated E860-containing tryptic peptides ofMyc-tagged SdeA immunopurified from HEK293 cells expressingSidJ-V5 or the D542A mutant.

RESEARCH | REPORTon July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from

the PDE domain and allows for specific analysis ofARTactivity (10). Glutamylation of SdeAH407A DNC

inhibited its ability to ADP-ribosylate Ub (Fig. 3A).Inactivation of SdeAH407A DNC required CaM,ATP/Mg2+, Glu, K367, and D542 of SidJ. Sub-stituting Glu with Gln, Asp, Lys, or Gly had noeffect on SdeAH407A DNC ART activity (Fig. 3B).Glutamylated SdeADNC behaved identically to

SdeADNC E860A in ubiquitination assays, losingboth its ability to ladder and to phosphoribosylateUb, yet was unaffected when ADP-ribosylatedUb was used as a substrate to bypass the ARTreaction (Fig. 3C). Moreover, SidJ suppressedthe toxicity of SdeAH407A and SdeCH416A in yeast(Fig. 3D and fig. S6), indicating that SidJ targetsthe SidE effectors at the point of ART activity.Thus, SidJ inhibits SidE by inactivating ARTactivity.We solved the structure of SidJDNC bound to

yeast CaM (yCaM) to a resolution of 2.1 Å (Fig. 4,A and B; fig. S7; and table S1). The structureresolves the N-terminal domain (NTD, black),the KD (teal), and the C-terminal domain (CTD,orange) of SidJ. The a-helical NTD and CTD arenestled beneath the C lobe of the KD and yCaMis bound to the “back” of SidJ (Fig. 4B). yCaMadopts a closed conformation with Ca2+ ionscoordinated in EF1 and EF3 (fig. S8A). The SidJKD exhibits a b-strand–rich N lobe containingthe regulatory aC helix and an a-helical-rich C

lobe. Structural homology searches (19) on theSidJ KD identify the histone H3 kinase Haspin(20) and the AMPylating selenoprotein-O (SelO)(21) as the closest structural homologs (fig. S8B).The SidJ KD is bound to AMP, Mg2+, and pyro-phosphate (PPi). In typical kinases, the nucleo-tide is positioned in a pocket between the N andC lobes. Surprisingly, in the SidJ structure, AMP,PPi, and Mg2+ are bound in a migrated pocketformed by an insertion in the KD catalytic loop(fig. S8C). This insertion contains highly con-served residues that bury the adenosine ringin a tight cleft near the base of the SidJ C lobe.The SidJ canonical kinase active site contains

PPi and two Mg2+ ions, which are coordinatedby standard kinase catalytic residues (Fig. 4C).In addition to the K367A and D542A mutants,Ala substitutions of R352 and N534 inactivatedSidJ (Fig. 4D). Within the migrated nucleotide-binding pocket, Ala substitutions of the invariantH492 and Y506 also inactivated SidJ (Fig. 4, Eand F). Arg500 andN733 also coordinate the phos-phate group of AMP and face into solvent; theirmutagenesis to Ala markedly inhibited glutamyl-ation activity.The interactions between SidJ and yCaM span

more than 2000 Å2 of surface area and involveboth lobes of yCaM (Fig. 4, G and H). Within theIQ helix, I841 is buried in a hydrophobic cleft ofthe yCaM C lobe. R843 and R846 contact Glu’s,

pinching the yCaMmolecule around the IQ helix(Fig. 4H). The aA and aB helices of the yCaM Nlobe are engaged in a groove formed by the CTDof SidJ (Fig. 4G). To disrupt the binding of theN lobe of CaM to the CTD of SidJ, we mutatedF763, F801, and E812 to Ala, Glu, and Ala, re-spectively, to generate the mutant “SidJFFE.”SidJFFE bound to CaM with 100-fold lower af-finity thanWT SidJ, and SidJIQ+FFE failed to bindCaM entirely (fig. S9, A to D). The CaM-bindingmutants were also significantly impaired in theirability to glutamylate SdeADNC in vitro (Fig. 4I).The SidJmutants that lost glutamylation activityand CaM binding in vitro also lost the ability tosuppress SdeA-catalyzed ubiquitination of hostproteins in mammalian cells (fig. S9E).SidJ is one of only a few T4SS effectors that

produce a growth-inhibition phenotype whenablated from the L. pneumophila genome (22, 23).We complemented the DsidJ Legionella strainwith SidJ mutants (fig. S5, A to C) and mon-itored growth within the environmental hostAcanthamoeba castellanii. Deletion of SidJ re-sulted in amarked growth defect, consistent withprevious reports (6) (Fig. 4J). Complementationof the DsidJ Legionella strain with WT SidJ, butnot the K367A, D542A, or the SidJIQ+FFEmutants,restored replication. Thus, SidJ requires kinaseand CaM-binding residues to facilitate Legionellainfection by inactivating SidE.

Black et al., Science 364, 787–792 (2019) 24 May 2019 4 of 6

Fig. 3. SidJ inactivates SdeA Ub ligaseactivity by polyglutamylating anactive-site residue in the ART domain.(A) SDS-PAGE and autoradiographydepicting the incorporation of32P-ADPR into HA-Ub from[a-32P]NAD+ by SdeAH407A DNC.SdeAH407A DNC or the E860A mutantwas pretreated in glutamylationassays with SidJDNC or the indicatedmutants (+/– ATP/Mg2+ or CaM)and then SdeAH407A DNC activity wasanalyzed. Reaction productswere analyzed as in Fig. 2E.(B) SDS-PAGE and autoradiographydepicting the incorporation of32P-ADPR into HA-Ub from[a-32P]NAD+ by SdeAH407A DNC.SdeAH407A DNC or the E860A mutantwas pretreated with SidJDNC, CaM,or ATP/Mg2+ or Glu, Gln, Asp, Lys,or Gly and SdeAH407A DNC activitywas analyzed as in (A). (C) Proteinimmunoblotting after in vitroubiquitination assays with the indicatedmutants of SdeADNC. SdeADNC andmutants were pretreated in glutamylationreactions in the absence (lanes 1 to 5)or presence (lanes 6 to 8) of SidJDNC/CaMor SidJD542A DNC/CaM (lanes 9 to 11).Ubiquitination reactions were started by theaddition of NAD+ and HA-Ub or HA-ADPRUb. The reaction components were resolved by SDS-PAGE and SidJDNC, SdeADNC, and Ub (using anti-HA or Abcam antibodies, which donot recognize Arg42-modified Ub) were detected by immunoblotting. (D) Growth-inhibition assay depicting the growth of S. cerevisiae expressingSdeAH407A-GFP and Flag-SidJ or the predicted catalytically inactive K367A and D542A mutants. EV, empty vector.

RESEARCH | REPORTon July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from

Black et al., Science 364, 787–792 (2019) 24 May 2019 5 of 6

Fig. 4. Crystal structure of the SidJ-CaM complex provides insightinto the mechanism of SidJ-catalyzed polyglutamylation. (A) Domainarchitecture of SidJ depicting the NTD (black), the KD (teal), and theCTD (orange). Residues corresponding to SidJDNC are indicated above theschematic. (B) Overall structure of the SidJDNC-yCaM complex. yCaM isin magenta and SidJ is colored as in (A). (C) Magnified view of the canonicalkinase active site of SidJ showing the interactions (dashed lines) involvedin PPi binding. The PPi is shown as sticks and the Mg2+ ions as light bluespheres. (D) Glutamylation activity of SidJ and mutants using SdeADNC and[3H]Glu as substrates. Reaction products were resolved by SDS-PAGE,radioactive gel bands were excised, and 3H incorporation into SdeADNC

was quantified by scintillation counting. (E) Magnified view of the migratedSidJ nucleotide-binding site depicting the interactions involved in AMPand PPi binding. The AMP and PPi are shown as sticks and the Mg2+ ion as alight blue sphere. (F) Glutamylation activity of SidJ and kinase active-sitemutants using SdeADNC and [3H]Glu as substrates. Reaction products wereanalyzed as in (D). (G) Magnified view of the interaction between the N lobeof CaM and residues in the CTD of SidJ. SidJ is colored as in (A) and theyCaM N lobe and C lobes are in pink and magenta, respectively. (H) Magnified

view of the IQ helix of SidJ and its interactions with the N and C lobes ofyCaM. Color coding is as in (G). (I) Glutamylation activity of SidJ- andCaM-binding mutants. Reaction products were analyzed as in (D).(J) Replication of L. pneumophila strains in A. castellanii. Infected amoebacells were lysed at the indicated time points and bacterial replication wasquantified by plating serial dilutions of lysates. Results are representative ofthree independent experiments. (K) Glutamylation activity of SidJ usingSdeADNC and [3H]Glu as substrates with different nucleotide analogs.Reactions went to completion and products were analyzed as in (D). Thechemical structures of AMP-PNP and AMP-CPP are also shown with thenonhydrolyzable bonds in red. (L) High-performance liquid chromatography(HPLC)MS/MS quantification of AMP released during SidJ-catalyzedglutamylation of SdeADNC. (M) Quantification of acyl-adenylate formationafter reactions with SidJDNC, SdeADNC, and [a-32P]ATP. The reactionswere terminated by the addition of TCA and the SdeA acyl-adenylate wasdetected by scintillation counting of the acid-insoluble material. In lanes4 and 5, Glu was added at time 0 or after 30 min. (N) Schematicrepresentation of the proposed SidJ-catalyzed glutamylation reactionwith the acyl-AMP SidE intermediate in brackets.

RESEARCH | REPORTon July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from

Amidoligase reactions typically proceed throughan acyl-phosphate or acyl-AMP intermediate(24). Although SidJ prefers ATP (fig. S10), it wasable to use ADP and AMP-PNP, but not AMP-CPP, as cosubstrates in glutamylation reac-tions (Fig. 4K). We detected AMP, but not ADP,release during the SidJ reaction, but only whenGlu was present, suggesting the formation of anacyl-adenylate intermediate (Fig. 4L). There-fore, wemonitored [32P]AMP incorporation from[a-32P]ATP during glutamylation reactions in theabsence of Glu. Because the intermediate con-tains an alkali-unstable acyl-adenylate, reactionswere terminated with trichloroacetic acid (TCA)and the acyl-adenylate was detected as 32P inTCA-insoluble material. SidJ adenylated SdeA,but not the SdeAE860A mutant (Fig. 4M). Addi-tion of Glu decreased the acid-insoluble label,suggesting that AMP is liberated by the forma-tion of the Glu–Glu isopeptide bond (Fig. 4M).We propose a catalytic mechanism for SidJ

glutamylation of SdeA whereby CaM binds SidJto stabilize an active conformation, which allowsthe canonical kinase–like active site of SidJ tobind ATP and transfer AMP to E860 on SdeA(Fig. 4N). Adenylated SdeA then binds the mi-grated nucleotide-binding pocket in SidJ, whichpositions the acyl-adenylate and glutamate forglutamylation and inactivation of the SidE ef-fectors. Because CaM is a eukaryote-specific pro-tein, its requirement for SidJ activity preventspremature inactivation of the SidE effectors inthe bacterium. Although our results are consist-ent with many previous conclusions regardingSidJ, others have shown that SidJ functions asa deubiquitinase (DUB) to deconjugate SidE-catalyzed Ub chains (12). However, our struc-tural and biochemical analyses failed to detect

any features that would suggest that SidJ canfunction as a DUB (fig. S11).In summary, our results underscore the di-

versity and catalytic versatility of the protein ki-nase superfamily.Wepropose that ATP-dependentligation reactions may be a common featureamong the vast diversity of eukaryotic proteinkinase–like enzymes found in nature (25). Thereare more than 500 protein kinases in humansand our results suggest that they should be ex-amined for alternative activities.

REFERENCES AND NOTES

1. E. Cornejo, P. Schlaermann, S. Mukherjee, J. Cell Biol. 216,3931–3948 (2017).

2. R. K. Sherwood, C. R. Roy, Annu. Rev. Microbiol. 70, 413–433(2016).

3. R. R. Isberg, T. J. O’Connor, M. Heidtman, Nat. Rev. Microbiol.7, 13–24 (2009).

4. D. Xu, L. Jaroszewski, Z. Li, A. Godzik, Bioinformatics 30,660–667 (2014).

5. J. C. Havey, C. R. Roy, Infect. Immun. 83, 3506–3514(2015).

6. K. C. Jeong, J. A. Sexton, J. P. Vogel, PLOS Pathog. 11,e1004695 (2015).

7. J. P. Bardill, J. L. Miller, J. P. Vogel, Mol. Microbiol. 56, 90–103(2005).

8. J. Qiu et al., Nature 533, 120–124 (2016).9. K. M. Kotewicz et al., Cell Host Microbe 21, 169–181

(2017).10. S. Bhogaraju et al., Cell 167, 1636–1649.e13 (2016).11. S. Kalayil et al., Nature 557, 734–738 (2018).12. J. Qiu et al., Cell Res. 27, 865–881 (2017).13. A. Hershko, A. Ciechanover, Annu. Rev. Biochem. 67, 425–479

(1998).14. Association of Biomolecular Resource Facilities, Delta Mass:

A Database of Protein Post Translational Modifications;https://abrf.org/delta-mass.

15. S. Mukherjee et al., Nature 477, 103–106 (2011).16. Y. Dong et al., Nature 557, 674–678 (2018).17. Y. Wang et al., Cell 173, 1231–1243.e16 (2018).18. A. Akturk et al., Nature 557, 729–733 (2018).19. L. Holm, P. Rosenström, Nucleic Acids Res. 38, W545–W549

(2010).

20. J. Eswaran et al., Proc. Natl. Acad. Sci. U.S.A. 106,20198–20203 (2009).

21. A. Sreelatha et al., Cell 175, 809–821.e19 (2018).22. T. J. O’Connor, Y. Adepoju, D. Boyd, R. R. Isberg,

Proc. Natl. Acad. Sci. U.S.A. 108, 14733–14740 (2011).23. Y. Liu, Z. Q. Luo, Infect. Immun. 75, 592–603 (2007).24. L. M. Iyer, S. Abhiman, A. Maxwell Burroughs, L. Aravind,

Mol. Biosyst. 5, 1636–1660 (2009).25. N. Kannan, S. S. Taylor, Y. Zhai, J. C. Venter, G. Manning,

PLOS Biol. 5, e17 (2007).

ACKNOWLEDGMENTS

We thank M. Phillips, J. Goldstein, K. Orth, N. Alto, and membersof the Tagliabracci laboratory for discussions. We also thankR. Isberg for Legionella strains, C. Brautigam (UTSWMacromolecular Biophysics Resource Core Facility) for helpwith ITC, and A. Lemoff (UTSW Proteomics Core Facility)and N. Williams UTSW Preclinical Pharmacology Core Facilityfor help with mass spectrometry. Results shown in this report arederived from work performed at the Structural Biology Center,Advanced Photon Source, Argonne National Laboratory. Funding:This work was funded by NIH Grants R00DK099254 (V.S.T.)and F30HL143859-01 (M.H.B.), Welch Foundation grant I-1911,CPRIT grant RP170674 (V.S.T), and Polish National ScienceCentre grant 2014/15/B/NZ1/03559 (K.P.). V.S.T. is a MichaelL. Rosenberg Scholar in Medical Research, a CPRIT Scholar(RR150033), and a Searle Scholar. Author contributions:M.H.B. and V.S.T. designed the experiments. M.H.B., A.O., D.R.T.,and V.S.T. conducted the experiments. K.A.S. performed themass spectrometry. M.G. and K.P. performed the bioinformatics.M.H.B. and V.S.T wrote the manuscript with input from allauthors. Competing interests: The authors declare no competinginterests. Data and materials availability: All materialsdeveloped in this study will be made available upon request.The atomic coordinates have been deposited in the Protein DataBank with entry code 6OQQ.

SUPPLEMENTARY MATERIALS

science.sciencemag.org/content/364/6442/787/suppl/DC1Materials and MethodsFigs. S1 to S11Table S1References (26–53)

23 January 2019; accepted 1 May 201910.1126/science.aaw7446

Black et al., Science 364, 787–792 (2019) 24 May 2019 6 of 6

RESEARCH | REPORTon July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from

ligasesBacterial pseudokinase catalyzes protein polyglutamylation to inhibit the SidE-family ubiquitin

TagliabracciMiles H. Black, Adam Osinski, Marcin Gradowski, Kelly A. Servage, Krzysztof Pawlowski, Diana R. Tomchick and Vincent S.

DOI: 10.1126/science.aaw7446 (6442), 787-792.364Science 

, this issue p. 787Science.Acanthamoeba castellaniithis mechanism for bacterial infectivity was verified in the natural reservoir host

pseudokinase fold, it does not phosphorylate SidE proteins but polyglutamylates them instead. The in vivo relevance of discovered that SidJ is activated by host calmodulin. Furthermore, although SidJ has aet al.ubiquitination. Black

SidJ also modulates the toxicity of the SidE family ubiquitin ligases that catalyze phosphoribosyl-linked host proteinwithin the host cell. SidJ prevents lysosome fusion with the vacuole, within which the bacterium resides and replicates.

that orchestrates this intracellular pathogen's establishmentLegionella pneumophilaSidJ is a protein produced by Divergent protein kinase

ARTICLE TOOLS http://science.sciencemag.org/content/364/6442/787

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2019/05/22/364.6442.787.DC1

CONTENTRELATED http://stke.sciencemag.org/content/sigtrans/12/594/eaat9797.full

REFERENCES

http://science.sciencemag.org/content/364/6442/787#BIBLThis article cites 50 articles, 9 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2019 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on July 24, 2020

http://science.sciencemag.org/

Dow

nloaded from