Yersinia Protein Kinase YopO Is Activated by A Novel G-actin Binding Process*

Pathogenic bacteria of the genus Yersinia employ a type III secretion system to inject effector proteins (Yops) into host cells. The Yops down-regulate host cell functions through unique biochemical activities. YopO, a serine/threonine kinase required for Yersinia virulence, is activated by host cell actin via an unknown process. Here we show that YopO kinase is activated by formation of a 1:1 complex with monomeric (G) actin but is unresponsive to filamentous (F) actin. Two separate G-actin binding sites, one in the N-terminal kinase region (amino acids 89–440) and one in the C-terminal guanine nucleotide dissociation inhibitor-like region (amino acids 441–729) of YopO, were identified. Actin binding to both of these sites was necessary for effective autophosphorylation of YopO on amino acids Ser-90 and Ser-95. A S90A/S95A YopO mutant was strongly reduced in substrate phosphorylation, suggesting that autophosphorylation activates YopO kinase activity. In cells the kinase activity of YopO regulated rounding/arborization and was specifically required for inhibition of Yersinia YadA-dependent phagocytosis. Thus, YopO kinase is activated by a novel G-actin binding process, and this appears to be crucial for its anti-host cell functions.


Pathogenic bacteria of the genus Yersinia employ a type III secretion system to inject effector proteins (Yops) into host cells. The Yops down-regulate host cell functions through unique biochemical activities. YopO, a serine/threonine kinase required for Yersinia virulence, is activated by host cell actin via an unknown process. Here we show that YopO kinase is activated by formation of a 1:1 complex with monomeric (G) actin but is unresponsive to filamentous (F) actin. Two separate G-actin binding sites, one in the N-terminal kinase region (amino acids 89 -440) and one in the C-terminal guanine nucleotide dissociation inhibitor-like region (amino acids 441-729) of
YopO, were identified. Actin binding to both of these sites was necessary for effective autophosphorylation of YopO on amino acids Ser-90 and Ser-95. A S90A/S95A YopO mutant was strongly reduced in substrate phosphorylation, suggesting that autophosphorylation activates YopO kinase activity. In cells the kinase activity of YopO regulated rounding/arborization and was specifically required for inhibition of Yersinia YadA-dependent phagocytosis. Thus, YopO kinase is activated by a novel G-actin binding process, and this appears to be crucial for its anti-host cell functions.
Assembly of filamentous (F) 3 -actin from monomeric (G)-actin drives cell motility, cytokinesis, and phagocytosis (1). Vesicle transport and intracellular movement of pathogenic bacteria are also dependent on actin polymerization (2,3). The turnover of actin filaments and their organization into complex structures is controlled by a plethora of actin-binding proteins, which often are regulated by specific signal transduction pathways (4). Besides governing motile cellular processes, actin has been recognized as an important cofactor for regulators of gene transcription and cell proliferation (5). The myocardin-related serum response factor (SRF) coactivator (MAL) accumulates in the nucleus dependent upon G-actin binding and activates transcription of SRF. It has been speculated that G-actin binding activates MAL for nuclear translocation (6). A recent report suggested that actin associates with RNA polymerase I in the nucleus and in cooperation with nuclear myosin I supports transcription (7). Furthermore, c-Abl tyrosine kinase, which is involved in cell responses to DNA damage and apoptosis signaling, becomes inhibited upon interaction with F-actin (8). Hence, a number of crucial cellular regulators are controlled by G-or F-actin, but the biochemical basis for this has not been determined.
Pathogenic Yersinia species encompassing Yersinia enterocolitica, Yersinia pseudotuberculosis, and Yersinia pestis employ a type III secretion system to translocate six effector proteins (Yersinia outer proteins, Yops) into target cells (9). Interestingly, at least four of these Yop effectors, namely YopE, -T, -H, and -O, impinge on the actin cytoskeleton (10 -15). YopO from Y. enterocolitica (YpkA in Y. pseudotuberculosis) is a serine/threonine kinase that disrupts F-actin structures in host cells, causes cell rounding, and contributes to antiphagocytosis of yersiniae (16 -18). YopO associates with the Rho GTP-binding proteins RhoA and Rac via a region showing homology to guanine nucleotide dissociation inhibitors (GDIs) and was found to block activation of RhoA in Yersinia-infected cells (19 -21). Interestingly, the kinase activity of YopO is activated by actin (22). It has remained unclear whether kinase activity contributes to F-actin disruption and antiphagocytic activity of YopO or whether these effects are solely due to inactivation of Rho GTPases. Clearly the kinase activity of YopO is relevant in vivo given that yersiniae expressing kinase-deficient YopO mutants are reduced in mouse virulence by at least 2 orders of magnitude (16,23).
We report here that YopO binds to G-actin via two separate binding sites, which leads to its autophosphorylation. Autophosphorylation up-regulates YopO capability to phosphorylate external substrates. Cellular analysis indicated that YopO kinase specifically controls cell rounding and inhibits a Yersinia-specific internalization mechanism. We propose that these YopO effects are connected to the G-actin-triggered activation process described herein.

EXPERIMENTAL PROCEDURES
Materials and Cell Culture-Chemical reagents were from Sigma-Aldrich, Roche Diagnostics, or Merck if not otherwise indicated. Oligonucleotides were from Metabion (Martinsried, Germany) or MWG (Ebersberg, Germany). The pGEX-6P-2 plasmids encoding GST-YopOwt and GST-YopOK269A were a generous gift of J. Dixon and S. Juris (University of Michigan). Human brain microvascular endothelial cells (HBMEC) and human umbilical vein endothelial cells were grown in ECGM from PromoCell (Heidelberg, Germany), 10% fetal bovine serum in plastic culture flasks at 37°C, 5% CO 2 in a humidified atmosphere. Human monocytes were isolated and cultivated as described (30).
Purification of Actin-Skeletal muscle actin was prepared according to Spudich and Watt (24) and further purified on a S200 gel filtration column, and the concentration of actin was measured as described by Wegner (25). For the preparation of actin from Dictyostelium discoideum, the soluble extract was chromatographed on a DE52 column (5 ϫ 30 cm) using a linear salt gradient of 0 -350 mM NaCl (2 ϫ 750 ml) essentially as described (26). Human platelet G-actin was purchased from tebu-bio (Offenbach, Germany).
Fluorescence Spectroscopy-Actin was labeled with N-(1pyrenyl)iodoacetamide (pyrene) according to Kouyama and Mihashi (27). Aliquots of pyrene-labeled actin were frozen in liquid nitrogen and stored at Ϫ70°C. Before use, the pyrenelabeled G-actin was thawed quickly and dialyzed in buffer containing 2 mM Tris/HCl, pH 8.0, 0.2 mM ATP, 0.5 mM DTT, 0.2 mM CaCl 2 overnight. For the measurement of actin polymerization a mixture of pyrene-labeled and unlabeled G-actin in a molar ratio of 1:10 was used at a final concentration of 2.2 M with or without YopO constructs. The measurements were performed at room temperature in a luminescence spectrometer (Aminco Bowman). For kinetic measurements the excitation wavelength was 365 nm, and the emission wavelength was 386 nm.
Construction of Plasmids and Expression of Proteins-Plasmid DNA preparations and isolation of DNA fragments from agarose gels were performed with Qiagen kits (Hilden, Germany). Restriction endonuclease digestions and PCR were performed according to standard protocols. The 5Ј and 3Ј deletions of yopO were constructed by PCR with the respective primers denoted in supplemental Table 1 using Y. enterocolitica O:8 yopO as template (28). Identity to the published YopO sequence on pYV8081 (accession number NP_863565) was verified. PCR products were subcloned using EcoRI and XhoI restriction sites and in-frame fusion with GST into pGEX-4T-1 or -2. The vector for expression of GST-YopO-(700 -729) was constructed by inserting an oligonucleotide corresponding to amino acids 700 -729 of yopO into pGEX-4T-2.
SDS/PAGE, Densitometry, and Western Blot Analysis-Proteins were separated by 10% SDS/PAGE followed by Coomassie Blue staining. For determination of the YopO-actin binding stoichiometry, the corresponding protein bands from gel filtration fractions containing the actin-YopO complex were quantified by two dimensional densitometry using a Gel Doc EQ System and Quantity One software (Bio-Rad). For Western blot, proteins were transferred by semidry blotting onto polyvinylidene fluoride (PVDF) membranes for 1 h at 1.2 mA/cm 2 . The PVDF membrane was blocked for 1 h or overnight at room temperature or 4°C, respectively, in phosphate-buffered saline, pH 7.4, 3% bovine serum albumin, 0.05% Tween 20. Membranes were developed with one of the following primary antibodies: anti-YpkA/YopO rabbit polyclonal (1:100,000; 20), anti-phosphoserine/threonine ATM/ATR rabbit polyclonal (1:500; Cell Signaling, New England Biolabs, Frankfurt, Germany). Secondary antibodies were goat anti-mouse or anti-rabbit IgG coupled to horseradish peroxidase (1:2,000 to 1:10,000; Amersham Biosciences). Detection was performed with an enhanced chemiluminescence kit (ECL; Amersham Biosciences).
Phosphorylation Assays-Kinase reactions were performed with 3-10 g of GST-YopO, YopOwt, or the indicated fragments in 20 l of kinase buffer containing 20 mM Hepes pH 7.4, 1.0 mM ATP, 1 mM DTT, 10 mM MgCl 2 , 2 mM MnCl 2 supplemented with 5 Ci of [␥-32 P]adenosine 5Ј-triphosphate (Amersham Biosciences) and incubated with 3-10 g of D. discoideum G-actin for 30 min at 30°C. Where indicated, human platelet G-actin, D. discoideum G-actin, or rabbit skeletal muscle G-or F-actin were added. The reaction was terminated by the addition of 4 ϫ SDS/PAGE sample buffer and heating for 5 min at 95°C. Radioactivity in dried SDS/PAGE gels or on blot membranes was visualized using phosphorimaging (Fujifilm FLA 3000) and Aida 4.0 Image Analyzer software or by exposure to x-ray film.
Mass Spectrometry Analysis of Phosphorylated YopO-YopOwt or mutant proteins were subjected to a phosphorylation reaction (see above) in the presence of 1 mM cold ATP, separated by 10% SDS/PAGE, and stained with Coomassie Blue. Protein bands were excised and in-gel-digested using modified trypsin (Promega, Mannheim, Germany) according to standard procedures (29). Peptides were desalted using ZipTips (Millipore, Schwalbach, Germany) and either eluted using a saturated ␣-cyano-4-hydroxycinnamic acid solution in 50% acetonitrile, 0.3% trifluoroacetic acid for matrix-assisted laser desorption ionization time-of-flight analysis or by spray negative ion mode spray solution (50% methanol, 5% ammonium hydroxide) for MS/MS analysis (MS/MS is tandem mass spectrometry (MS)) and precursor ion scans. Precursor ion scans were recorded in negative mode on a Q-STAR XL tandem mass spectrometer detecting all precursor ions that resulted in a product ion with a m/z value of 79 atomic mass units (PO 3 Ϫ ). The charge state of the precursor ions was resolved in a negative ion mode time-offlight spectrum of the same sample.
Chromatography and Purification of YopO-Actin Complex-For preparation of monomeric YopO-(89 -729), 500 g of protein in 50 l of buffer containing 50 mM Tris/HCl, pH 8.0, 150 mM NaCl, 1 mM DTT, 5 mM MgCl 2 , 2.5 mM CaCl 2 were loaded onto a Superdex 200 PC 3.2/30 column equilibrated with 30 mM Tris/HCl, pH 7.5, 150 mM NaCl, 1 mM DTT in a Smart system (Amersham Biosciences). The flow rate was 40 l/min, and the fraction size was 50 l. Fractions eluting shortly before the 66-kDa marker contained monomeric YopO and were combined. Freshly dialyzed D. discoideum actin (50 l, 1 g/l) in 2 mM Tris/HCl, pH 8.0, 0.2 mM ATP, 0.5 mM DTT, 0.2 mM CaCl 2 was centrifuged for 30 min at 4°C and 100,000 ϫ g to remove residual F-actin. For preparation of the YopO-actin complex, monomeric YopO-(89 -729) and G-actin (1 g/l, each) were mixed in gel filtration buffer, incubated for 15 min at room temperature, and then loaded onto the gel filtration column at a flow rate of 40 l/min and a fraction size of 25 l. Fractions were analyzed by SDS/PAGE, Coomassie Blue staining or by Western blot analysis as indicated. For dephosphorylation, purified YopO-(89 -729)-actin complex was incubated with or without protein serine/threonine phosphatase 1 (2500 units/ml; New England Biolabs) at 30°C for 1 h and then refractionated on gel filtration columns. Molecular mass markers were sweet potato ␤-amylase (200 kDa), bovine serum albumin (66 kDa), and bovine erythrocyte carbonic anhydrase (29 kDa).
GST Pulldown Assays-D. discoideum G-actin (ϳ30 g in 200 l of 2 mM Tris/HCl, pH 8.0, 0.2 mM Na-ATP, 0.5 mM DTT, 0.2 mM CaCl 2 , 0.01% NaN 3 ) was incubated with 50 l of glutathione-Sepharose beads and loaded with 30 g of the appropriate GST fusion protein at 4°C for 1.5 h. Beads were washed 4ϫ with lysis buffer containing 1% Triton-X100 and then subjected to SDS/PAGE and Western blot analysis.
Analytical Ultracentrifugation-Sedimentation equilibrium experiments were performed with an analytical ultracentrifuge (Optima XL-I; Beckman Instruments) equipped with absorbance and interference optics. Aluminum double sector cells with sapphire windows were used throughout. Protein solutions containing 0.3 mg of YopO-actin complex per ml were dialyzed against 30 mM Tris/HCl, pH 7.5, 150 mM NaCl. The partial specific volume was estimated from the amino acid composition of YopO and actin yielding a value of 0.737 ml/g. The protein solution was centrifuged at 8064 ϫ g and 4°C for 72 h.
Cell Rounding, Actin Disruption, and Phagocytosis Assays-Cells (HBMEC, human umbilical vein endothelial cells, human monocytes) were transiently transfected with vectors expressing GFP or GFP-YopO constructs using Metafectene (Biontex, Martinsried, Germany) or Amaxa nucleofector technology (Amaxa, Cologne, Germany). Cells showing GFP fluorescence were scored for rounding/arborization using phase contrast microscopy and for actin filament disruption using Alexa Fluor 568 phalloidin staining. Only cells with completely rounded phenotype or complete disruption of actin filaments were scored as positive. IgG-or complement coated sheep red blood cells were added at a ratio of 15:1 to human monocytes (30) transfected with vectors expressing GFP or GFP-YopOwt. Intra-and extracellular sheep red blood cells were distinguished by phase contrast microscopy as described (31). Yersinia invasin-coated fluorescent latex beads were attached to HBMEC (50:1) expressing GFP, GFP-YopOwt, or GFP-N17Rac. The percentage of internalized beads was determined as described (30). Yersinia YadA expressing E. coli were attached to cells expressing the indicated GFP-YopO constructs (ratio 30:1), and internalized bacteria were quantified with a double immunofluorescence staining method as described (30,32). Internalization was normalized to control cells that were set to 100%. Statistical evaluation was performed using Student's t test. To verify equal expression of GFP fusion proteins, the respective transfected cells were recorded with a spot Pursuit 1.4MP monochrome camera and Spot software (Diagnostic Instruments, Inc, Sterling Heights, MI) using identical exposure settings. Average pixel intensities of three outlined regions per cell were obtained using ImageJ software (Research Services Branch, National Institute of Mental Health) and imported into graphics program for statistical analysis. Fig. 1. The N-terminal region mediating bacterial secretion/translocation and membrane localization in host cells is followed by the predicted kinase domain (amino acids 150 -400). The C-terminal half of YopO contains a Rho GTPase binding GDI-like domain (within amino acids 431-612) and a stretch of amino acids (amino acids 710 -729) displaying homology to the actin bundling protein coronin. This domain model is a synopsis of all published structural and biochemical data of YopO and YpkA (9,16,19,21,22,33).

YopO Autophosphorylation Is Stimulated by G-actin-The modular protein organization of YopO is depicted in
To determine which form of actin stimulates YopO kinase, we assayed autophosphorylation of recombinant wild type YopO (YopOwt) in the presence of G-actin or F-actin. As described before (22), no autophosphorylation of YopO occurred in the absence of actin. By comparison, YopO autophosphorylation was detectable already at a G-actin:YopO molar ratio of 0.1:1 and was maximal at a 1:1 ratio. In contrast, even a 5-fold molar excess of F-actin versus YopO did not significantly stimulate YopO autophosphorylation (Fig. 2a).
We also investigated whether muscle and nonmuscle actins differ in their ability to stimulate YopO. These experiments showed that human platelet actin and D. discoideum actin were equally effective in activating YopO autophosphorylation, whereas muscle actin was about a 1 ⁄ 10 as active (supplemental Fig. 1S). Yet at high enough concentrations muscle G-actin could also cause full activation of YopO. 4 Hence, YopO kinase is best activated by nonmuscle G-actin, which fits with the idea that it mainly acts in immune cells such as neutrophils and macrophages (9,10).
YopO and Actin Form a Phosphorylation-independent 1:1 Complex-To get first insights into the mechanism of YopO activation by G-actin, we sought to determine the stoichiometry and molecular mass of the YopO-G-actin complex. Because YopOwt tends to degrade and requires detergents such as Triton X-100 for purification and stability (22), we produced a construct that lacks the N-terminal amino acids 1-88 that serve as the bacterial secretion and translocation signal, giving rise to YopO-(89 -729) (Fig. 1). As anticipated, YopO-(89 -729) retained all the in vitro activities of YopOwt (see below).
To isolate an YopO-actin complex, prepurified monomeric YopO-(89 -729) was incubated with D. discoideum G-actin. Formation of a complex was verified by gel filtration chromatography. As shown in Fig. 2b, upper and lower panel, YopO-(89 -729) and G-actin individually eluted as monomeric proteins around the 66 kDa and between the 66-and 29-kDa markers, respectively. Upon coincubation, YopO-(89 -729) and actin could be detected in fractions eluting between the FIGURE 1. Scheme of YopO protein organization and YopO constructs used in this study. The first and the last amino acids of YopOwt or its fragments and mutants are indicated. cDNAs encoding these constructs were cloned into vectors, allowing expression of GST fusion proteins in E. coli or GFP fusion proteins in host cells. Sec/Trans, secretion and translocation region used by the Yersinia type III secretion system. Kinase, region with homology to serine/threonine kinases. Asp-267 and Lys-269 are amino acids critical for catalytic activity. GDI-like, region with homology to Rho GDI. CH, coronin homology region (16,19,21,22).  200-and 66-kDa markers, indicating formation of an YopOactin complex (Fig. 2b, middle panel). To determine the YopOactin stoichiometry, we quantified the two proteins within the complex by densitometry of Coomassie Blue-stained gels using separate dilution series for calibration. The molar ratio of YopO to actin was determined as 1:1.
To directly measure the mass of the YopO-actin complex, sedimentation equilibrium ultracentrifugation was performed (see supplemental Fig.  2S for a representative experiment). These experiments yielded a molecular mass of 126 Ϯ 6 kDa (mean Ϯ S.D.; n ϭ 2) for the YopO-actin complex purified by gel filtration chromatography (fractions corresponding to the boxed lanes in Fig.  2b were used). This clearly indicated that the YopO-actin complex consists of one molecule of actin and one molecule of YopO.
We also tested whether dephosphorylation of YopO has an effect on the stability of the YopO-actin complex. By using an anti-phosphoserine/threonine antibody (see supplemental Fig. 3S) it was first verified that YopO in the complex with actin was phosphorylated (Fig. 2c). The purified complex was treated with the nonspecific protein serine/ threonine phosphatase 1 and again loaded onto the gel filtration column. As documented in Fig. 2c, the complex of dephosphorylated YopO and actin remained stable. We, therefore, conclude that YopO is phosphorylated upon forming a 1:1 complex with G-actin, but phosphorylation is not required for the stability of the complex.

YopO Contains Separate N-and C-terminal G-actin Binding Sites-
To delineate the regions of YopO involved in actin binding, we performed pulldown assays using purified D. discoideum G-actin and Nand C-terminal deletion mutants of YopO fused to GST (depicted in Fig.  1). As demonstrated in Fig. 3  (700 -729), containing the putative C-terminal actin binding domain and the coronin homology region, respectively, did not pull down actin (21,22). As reported before, YopO-(89 -710) was unable to bind to actin in this assay (22; Fig. 3). Consequently, YopO-(89 -440) harboring the predicted catalytic domain and lacking the C-terminal half of the protein did not present actin binding activity either (Fig. 3). These results indicate that the C-terminal portion of YopO harbors an actin binding site for which amino acids 441-729 are necessary and sufficient. In experiments aimed to narrow down the YopO catalytic domain, we noticed that YopO-(89 -440) harboring the kinase region displayed autophosphorylation in the presence but not in the absence of G-actin (data in Fig. 5b). This indicated that the N-terminal half of YopO-(89 -440) must also have the capability to associate with actin even though this could not be detected by GST pulldown assays. We, thus, employed an alternative method to determine G-actin binding of YopO- (89 -440). Proteins that sequester G-actin can inhibit polymerization of G-actin to F-actin in vitro, thus allowing the calculation of bound versus free G-actin and an approximate K d of the G-actin-sequestering protein complex. In this assay YopO-(89 -729) dose-dependently inhibited actin polymerization; the K d of the YopO-(89 -729)-G-actin complex was estimated to be ϳ1.9 M. Under these conditions similar values have been determined for the G-actin-sequestering profilins I (5.1 M) or II (1.8 M) from D. discoideum (34). As inferred from its actin-dependent autophosphorylation activity, YopO-(89 -440) also inhibited formation of F-actin in the assay. A rough calculation yielded a K d of 4.1 M for the YopO-(89 -440)-G-actin complex. In contrast, YopO-(89 -398) and YopO151-440 showed no inhibitory activity in this assay (Fig.  4). Additional actin polymerization assays indicated that YopO-(441-729) (K d 6.1 M) and YopO-(399 -729) (K d 5.5 M) also sequester G-actin, which is consistent with the GST pulldown data ( Fig. 4 and not shown). We conclude from this data that YopO harbors two independent actin binding sites. One is encompassed by amino acids 89 -440, whereby amino acids 399 -440 are essential. The other is formed by amino acids 441-729, whereby amino acids 710 -729 are essential. Because of the 1:1 stoichiometry of the YopO-G-actin complex, we also conclude that the N-and C-terminal portions of YopO bind to different regions of the actin molecule. N-and C-terminal Actin Binding Is Required for Full Activity of YopO-To characterize the interplay of kinase activation and actin binding, we tested autophosphorylation of YopO deletion mutants upon stimulation with actin. As demonstrated in Fig.  5a, YopOwt and YopO-(89 -729) showed comparable incorporation of radioactivity, consistent with their intact N-and C-terminal actin binding sites. YopO-(151-729) was kinasedead, which may be due to disruption of its N-terminal actin binding site but could also be caused by removal of autophosphorylation sites (see below). YopOK269A displayed low residual autophosphorylation. Interestingly, YopO-(89 -440), which consists of the N-terminal kinase region and actin binding site, was capable of autophosphorylating in the presence but not in the absence of actin (Fig. 5b). In comparison, YopO-(89 -398) showed no autophosphorylation (Fig. 5b), consistent with its abolished actin binding (Fig. 4). Autophosphorylation of YopO-(89 -440) was considerably diminished compared with YopO-(89 -729) and closely resembled that of YopO-(89 -710) (compare Fig. 5, a and b). Because YopO-(89 -710) essentially lacks the amino acids required for C-terminal actin binding (Fig. 3), these data strongly indicate that YopO requires simultaneous binding of G-actin to its N-and C-terminal regions for full stimulation of kinase activity.
Identification of Autophosphorylated Amino Acids in YopO That Regulate Its Kinase Activity-For identification of autophosphorylated amino acid residues in YopO, we performed MS analysis of in vitro phosphorylated YopO-(89 -729). By employing precursor ion scanning, a peptide corresponding to amino acids 89 -108 of YopO (peptide 89 -108) containing two phosphate groups was detected. In addition, peptides 89 -94 and 95-108, each containing one phosphate group, were found (summarized in Table 1). This indicates that there are two phosphorylated residues within amino acids 89 -108 of YopO whereby one lies within peptide 89 -94 (Thr-89 or Ser-90) and the other one within peptide 95-108 (Ser-95 or Ser-102). To identify the relevant site within amino acids 95-108, we analyzed the single point mutants YopOS95A and YopOS102A. Two phosphorylated residues could be detected within peptide 89 -108 of YopOS102A, implicating Ser-95 as a phosphorylation site. No signals were detected with YopOS95A. We next analyzed the following double point mutants of YopO-(89 -108): T89A/S95A, T89A/S102A, S90A/S95A, and S90A/ S102A. The results obtained with YopOS90A/S102A were revealing given that peptides 89 -108 and 95-108 were detected and that each contained one single phosphorylated residue (Table 1). This result not only confirmed Ser-95 as a phosphorylation site (by exclusion of Ser-102) but also sug-gested that Ser-90 is the second site; otherwise, a doubly phosphorylated peptide 89 -108 should have been detected. No conclusive results could be obtained with the other double point mutants. The reason might be that some of these mutations reduce autophosphorylation/incorporation of radioactivity into YopO and/or alter the chemical properties of the generated peptides.
To verify the phosphorylation sites putatively identified by MS, we subjected the following single or double point mutants of Yop-(89 -729) to kinase assays: S95A, S102A, S90A/S95A, and T89A/S95A. Consistent with the MS data, incorporation of radioactivity was unchanged in YopOS102A and reduced by about 50% in YopOS95A and YopOT89A/S95A. Importantly, autophosphorylation was greatly diminished by about 90% in YopOS90A/S95A (Fig. 5c). These data clearly establish that amino acid residues Ser-90 and Ser-95 are autophosphorylation sites within YopO.
Next we asked whether autophosphorylation of YopO on Ser-90/Ser-95 also regulates phosphorylation of external substrates. The data presented in Fig. 5d reveal that YopO-(89 -729)S90A/S95A displayed greatly diminished phosphorylation of the artificial substrates myelin basic protein and histone. This indicates that autophosphorylation of YopO on Ser-90/ Ser-95 activates its kinase activity. In line with this, YopO-(151-729) lacking the autophosphorylation sites did not present any substrate phosphorylation.
Here we could also identify cellular effects of YopO that are directly related to its kinase activity. The kinase-dead mutant YopOD267A/K269A produced much less cell rounding/arborization but a similar extent of F-actin disruption as YopOwt, indicating that the kinase activity specifically acts back on the cytoskeleton. The YopO kinase activity was also required for inhibition of YadA-mediated phagocytosis, whereas other routes of phagocytosis were unaffected by YopO. This points to a unique role for YopO kinase in modulating cell shape/motility and a Yersinia-specific subtype of phagocytosis.
A major question is which cellular substrates are responsible for the YopO kinase effects. Recently ovarian tumor domain ubiquitin aldehyde binding 1 (otubain-1) was identified as a phosphorylation substrate of YopO (36). Otubain-1 is a putative ubiquitinase and associates with the E3 ubiquitin ligase GRAIL, implicated in regulating anergy in lymphocytes (37). However, the consequences of otubain-1 phosphorylation and its putative role in cytoskeletal regulation are elusive.
Because no bacterial or mammalian kinases activated by G-actin have been characterized so far, YopO represents a prototype. It will be interesting to find out which advantage Yersinia takes of exploiting cellular G-actin homeostasis for regulating its activity. Once thought to be merely a structural protein of the cytoskeleton, it has now become evident that actin can directly regulate the function of numerous signaling proteins. Some of these may turn out to act as G-actin regulated kinases like YopO.