Design and Synthesis of AX7574: A Microcystin-Derived ...

794 Bioconjugate Chem., Vol. 15, No. 4, 2004 Shreder et al.Scheme 1.Synthesis of the Microcystin-TAMRA Probe AX7574Downloaded by WUHAN UNIV on October 25, 2009 | http://pubs.acs.orgPublication Date (Web): June 18, 2004 | doi: 10.1021/bc04995804,6-dioxoheptanoic acid to form a carboxylic acid bearingderivative. However, this acid was not further elaboratedupon. In the end, they chose a synthetic route in which4,6-dioxoheptanoic acid was carbodiimide coupled tocephalexin and the resulting dione conjugated to pyoverdinin 20% yield. Because microcystin modification with4,6-dioxoheptanoic acid would yield a tricarboxylic acid,any resulting amidation with an amine bearing fluorophorederivative would be difficult to accomplish selectively.Consequently, we started our approach, analogousto the case of cephalexin, with the synthesis of a fluorophore-dionederivatization reagent. In particular, wechose to use tetramethylrhodamine (TAMRA) as thefluorophore, because of its previous success as a sensitiveand quantitative tag for ABPs.The requisite dione was prepared from the mixedsuccinimidyl esters of 5-(and 6)-carboxytetramethylrhodamine(TAMRA-OSu). Reaction of TAMRA-OSu with(Boc-4-aminomethyl)piperidine followed by treatment ofthe resulting product with 1:1 TFA/DCM produced thesecondary amine 1 (Scheme 1). When this amine wastreated with an excess of N,N′-carbonyldiimidazoleactivated4,6-dioxoheptanoic acid, compound 1 was quantitativelyconverted into the desired dione 2. Afterpurification via C 18 reverse phase HPLC, the TAMRAdione2 was isolated in 66% yield. As expected, 1 H NMRanalysis showed the presence of resonances for both theenol and dione form of compound 2. Furthermore, doublingof the enol, dione, and aromatic resonances wasobserved from the use of the mixed TAMRA isomerstarting material.The probe AX7574 was produced from the condensationof compound 2 with microcystin-LR under basic conditions.Microcystin-LR (580 µg) was heated with a 10-foldexcess of dione in a mixture of 20% dioxane/1 M NaHCO 3(pH 9.0) for 2 days at 65 °C. LC/MS analysis indicatedthe consumption of microcystin-LR and the presence ofa new TAMRA-containing peak (as evidenced by anabsorption at 550 nm) with the desired product mass(MALDI-FTMS (DHB): m/z 1625.8402 - C 87 H 112 N 14 O 17+ H requires 1625.8464). AX7574 was isolated from otherimpurities using C 18 reverse phase chromatography in4% yield.AX7574 was characterized structurally using tandemmass spectrometry. Fragmentation of the parent ion (MS/MS) resulted in two major ions, 1099.5 and 527.4. Theseions were assigned to the fragments that result from theN-CO cleavage of the tertiary amide bond found in thelinker. Further isolation and fragmentation of these ions(MS/MS/MS) were consistent with the expected structureof AX7574 (see Figure 2). For the former ion, fragmentationscorresponding to cleavage of the cyclic peptidebackbone, decarbonylation, and loss of the pyrimidinemoiety were seen. For the latter ion, fragments correspondingto cleavage about the amide bond of the linker-TAMRA moiety were observed.Once the structure of AX7574 was secured, attentionwas turned to demonstrating its ability to label PP-1.PP-1 (25 nM) was incubated with AX7574 (0.5 µM) in 50mM HEPES, 5% DMSO, 0.01% Triton x-100, pH 7.4, atambient temperature. At various timepoints, samplealiquots were removed, quenched with DTT-containingsample buffer, and then heat-denatured. The resultingsamples were subjected to denaturing SDS-PAGE, andlabeled enzyme was visualized using a flatbed laserinducedfluorescence scanner. Importantly, the TAMRAtag provided a sensitive fluorescent handle by which thedegree of labeling could be quantified (see ExperimentalProcedures). Consistent with the results of Holmes et al.(vide supra), the kinetics of PP-1 labeling was on theorder of hours (see Figure 3) with 65% of the labelingoccurring at ∼16 h. When the time course data were fitto a single exponential, maximal labeling was determinedto approach a ratio of 1:1 PP-1 to AX7574.Several additional experiments were run to determineif the observed labeling was specific for the active site of

A Microcystin-Derived Phosphatase Probe Bioconjugate Chem., Vol. 15, No. 4, 2004 795Figure 2. Mass spectrometric (MS) structural identification of AX7574. Shown are some of the observed fragmentation patternsfrom the two major parent ions observed during the MS/MS analysis of the doubly charged parent ion peak. Dual letter fragmentmasses (vide infra) are derived from a counterclockwise trace along the cyclic peptide backbone of AX7574 from the designated pointof fragmentation to the next indicated point of fragmentation: ea - H 2O ) 592, ga + NH 2-H 2O ) 521, fb ) 842, fb - CO ) 814, dc) ih ) 971, gc ) ai ) 786, hc ) 703, ic ) ag ) 574, ad ) 390, af ) 503, ah ) 657, jh ) 674, hj - H 2O ) 408.Downloaded by WUHAN UNIV on October 25, 2009 | http://pubs.acs.orgPublication Date (Web): June 18, 2004 | doi: 10.1021/bc0499580Figure 3. Kinetics of AX7574 labeling of PP-1. PP-1 (25 nM)was labeled with AX7574 (0.5 µM). At the indicated timepoints,sample aliquots were quenched. The quenched samples wereelectrophoresed, the gel was scanned, and the fraction of PP-1labeled was determined as described in the ExperimentalProcedures. The curve drawn is the best fit of the data to a singleexponential.Figure 4. Inhibition of AX7574 labeling of PP-1 with phosphataseinhibitors. PP-1 (25 nM) was labeled with AX7574 (0.5µM) (A) in the absence of phosphatase inhibitors or in thepresence of (B) microcystin-LR, (C) calyculin A, (D) proteinphosphatase inhibitor 2, and (E) fostriecen at 100 nM each. Anadditional AX7574 (0.5 µM) labeling experiment using denaturedPP-1 (25 nM, F) is also shown. Labeling reactions werequenched after 3 h. The samples were electrophoresed, the gelwas scanned, and the scanned image was quantified as describedin the Experimental Procedures.PP-1 (see Figure 4). In the first set of experiments,inhibition of AX7574 labeling with known inhibitors ofFigure 5. PP-1 IC 50 determination for microcystin (b, IC 50 )0.3 nM) and AX7574 (O, IC 50 ) 4.0 nM). The curves drawn arethe best fit of the data to a variation of the Hill equation (seeExperimental Procedures section).PP-1 was tested. The inhibitors selected included microcystin,calyculin A (PP-1 IC 50 ) 1.2 nM) (29), and proteinphosphatase inhibitor 2 (PP-1 IC 50 ) 1 nM) (42). Theseparticular inhibitors were selected because they interactwith phosphatases through diverse mechanisms. Forexample, microcystin is an irreversible inhibitor, and bothcalyculin A and protein phosphatase inhibitor 2 arereversible ones. Furthermore, while microcystin andcalyculin A are small molecules, protein phosphataseinhibitor 2 is a protein macromolecule (MW ) 31 kDa).In addition, fostriecin, a weak inhibitor of PP-1 (PP-1 IC 50) 131 µM, PP-2A IC 50 ) 3.2 nM), was included as anegative control. PP-1 (25 nM) was preincubated for 10min with microcystin (100 nM), calyculin A (100 nM),protein phosphatase inhibitor 2 (100 nM), or fostriecin(100 nM) or allowed to stand for 10 min in the absenceof added phosphatase inhibitors. Afterward, AX7574 (0.5µM) was added to each sample and the labeling reactionstopped after 3 h using DTT-containing sample buffer.When compared to the same labeling reaction done inthe absence of phosphatase inhibitors, significantly lesslabeling with AX7574 was observed in the presence ofmicrocystin, calyculin A, or protein phosphatase inhibitor2. In the case of microcystin and calyculin A, the resultsindicate that AX7574 competes for the active site wherethese inhibitors are known to bind. As expected, the PP-2A specific inhibitor, fostriecin, did not cause a significantdecrease in labeling of PP-1 by AX7574.

796 Bioconjugate Chem., Vol. 15, No. 4, 2004 Shreder et al.Downloaded by WUHAN UNIV on October 25, 2009 | http://pubs.acs.orgPublication Date (Web): June 18, 2004 | doi: 10.1021/bc0499580Figure 6. AX7574 can detect serine/threonine phosphatases in a proteome. (A) Soluble proteome derived from Jurkat cellspreincubated with or without microcystin (1 µM) or calyculin a (1 µM) for 10 min and then treated with AX7574 (0.5 µM) for 60 min.Reactions were quenched with standard 2× SDS/PAGE loading buffer (reducing), separated from unreacted AX7574 using SDS-PAGE, and visualized in-gel using a flatbed laser-induced fluorescence scanner. (B) Coomassie blue stained gel from Figure 6Aconfirming that all three lanes contained approximately equal amounts of protein. (C) AX7574-labeled proteins identified using MSprotein identification.Figure 7. AX7574 can detect the decrease of serine/threonine protein phosphatase activities in calyculin A-treated Jurkat T cells.(A) Jurkat cells treated with calyculin A (25 nM) for 60 min or control DMSO, and soluble proteome derived from these cells weretreated with AX7574 (0.5 µM) for 60 min. Reactions were quenched with standard 2× SDS/PAGE loading buffer (reducing), separatedby SDS/PAGE, and visualized in-gel using a flatbed laser-induced fluorescence scanner. (B) Side-trace analysis of the gel data fromFigure 7A. (C) The scanned gel from A was transferred to nitrocellulose and immunoblotted with anti-PP1 and anti-PP2A.A second kind of experiment was carried out toexamine whether the labeling of PP-1 by AX7574 requiredPP-1 to retain its native tertiary structure (seeFigure 4). To address this question, PP-1 (25 nM) wasfirst denatured by heating the protein at 65 °C for 10min. The subsequent addition of AX7574 (0.5 µM) resultedin no significant labeling of PP-1. This resultindicates that the recognition and subsequent attachmentof AX7574 to PP-1 does in fact require the nativestructure of PP-1.To investigate the impact of TAMRA conjugation onthe affinity of the microcystin scaffold for PP-1, the IC 50values of both microcystin-LR and AX7574 for PP-1 weredetermined and compared (see Figure 5). Both compoundsat various concentrations (0-30 nM) were incubatedwith PP-1 (5 nM) for 10 min. Postincubation,p-nitrophenyl phosphate (pNPP, 20 mM) was added andthe degree of substrate hydrolysis measured spectroscopicallyat 450 nm. For each inhibitor, A 450 was plottedversus inhibitor concentration, and the points were fitto a variation of the Hill equation (43) to determine IC 50values. Such an analysis yielded IC 50 values of 0.3 and4.0 nM for microcystin and AX7574, respectively. Bycomparison, a similar pNPP-based determination yieldeda PP-1 IC 50 value of 0.25 nM for microcystin-LR (44). Asevidenced by these values, TAMRA conjugation had onlya modest impact on binding and the resulting probeAX7574 remained a potent PP-1 inhibitor.To test whether AX7574 would react with serine/threonine protein phosphatases in a complex proteome,we incubated soluble fractions of Jurkat cells with thisprobe (0.5 µM) for 60 min. Protein separation of the

A Microcystin-Derived Phosphatase Probe Bioconjugate Chem., Vol. 15, No. 4, 2004 797Downloaded by WUHAN UNIV on October 25, 2009 | http://pubs.acs.orgPublication Date (Web): June 18, 2004 | doi: 10.1021/bc0499580sample was achieved using SDS-PAGE and visualizedin-gel using a flatbed laser-induced fluorescence scanner.As seen in Figure 6A, AX7574 strongly labeled only twosets of proteins (36 and 38 kDa) and the labeling of theseproteins by AX7574 was strongly inhibited by bothmicrocystin and calyculin A. Protein staining revealedthat all lanes contained approximately equal amountsof protein (Figure 6B), an indication that the labeling wasspecific for the proteins in those bands and not merelythe result of labeling highly abundant proteins nonspecifically.To confirm that the proteins labeled by AX7574 inJurkat soluble fractions were serine/threonine proteinphosphatases, mass spectrometric (MS) protein identificationwas performed (see Experimental Procedures):The two fluorescent gel bands were excised (Figure 6A),labeled proteins isolated using anti-TAMRA antibodybeads, and the captured proteins digested using trypsin.Subsequent MS sequence analysis of the resulting peptideswas used to identify the protein composition of eachband. The 38-kDa protein band was shown to be PP-1and the 36-kDa protein band was found to contain PP-2A, PP-4, and PP-6 (Figure 6C). Both PP-2A and PP-4, aphosphatase closely related to PP-2A (45), have previouslybeen shown to be inhibited by microcystin-LR withIC 50 values of 0.15 nM for both enzymes (46, 47). Inaddition, PP-2A, PP-4, and PP-6 have been capturednoncovalently using a microcystin-sepharose column aspart of an effort to purify these proteins from solublebovine testes extracts. Our data, which relies on thecovalent labeling of these phosphatases, is the firstaccount, to the best of our knowledge, of the irreversiblebinding of PP-4 and PP-6 by a microcystin derivative.Ultimately, and perhaps most importantly, the MSidentification results demonstrate that AX7574 is ableto react with and detect numerous serine/threoninephosphatases in a crude cell sample.To test whether AX7574 could record changes in serine/threonine protein phosphatase activity levels, we treatedJurkat cells with calyculin A dissolved in a minimalamount of DMSO (2% v/v) such that the final concentrationof inhibitor was 25 nM. Calyculin A was chosen notonly for its cell permeability but also its potent activityagainst both PP1 (IC 50 ) 1.2 nM) and PP2A (IC 50 ) 7.3nM) (29). As a control, Jurkat cells were treated with anequivalent volume of DMSO alone. After labeling withAX7574 (0.5 µM) for 60 min, AX7574-reactive serine/threonine protein phosphatases showed significantlylower labeling intensities in the calyculin A-treatedJurkat cells versus the control Jurkat cells (see Figure7A and 7B). These results indicated that serine/threonineprotein phosphatase activities were strongly inhibited bycalyculin A. Critically, western blot analysis using anti-PP1 and anti-PP2A antibodies identified the same quantityof PP1 and PP2A in both control and calyculinA-treated Jurkat cells (see Figure 7C). These resultsfurther confirm that AX7574 labels serine/threonineprotein phosphatases in an activity-dependent manner,recording variations in protein phosphatase activityrather than protein level. Without the AX7574 probe,western blot analysis and/or standard proteomics studieswould have difficulties in distinguishing active frominactive protein phosphatases in these samples.In conclusion, we describe herein the synthesis ofAX7574, a microcystin derived fluorescent probe forserine/threonine phosphatases. The conjugation of a 1,3-diketone-containing fluorophore to the arginine-containingscaffold of microcystin was critical in maintaining thespecific, high affinity, and irreversible-bond formingnature of the microcystin-phosphatase interaction. Theability to form a covalent bond within the phosphataseactive site allowed this tagged chemical probe to captureand visualize the activity of this important class ofenzymes in a complex proteome. Importantly, AX7574expands the range of activity-based probes and providesa means to investigate the role that serine/threoninephosphatases play in physiological and/or pathologicalevents.ACKNOWLEDGMENTThe authors thank Drs. John Kozarich, Katrin Szardenings,and David Campbell for helpful suggestionsconcerning this work and/or the manuscript.LITERATURE CITED(1) Anderson, N. L., and Anderson, N. G. (1998) Proteome andproteomics: New technologies, new concepts, and new words.Electrophoresis 19, 1853-1861.(2) Pandey, A., and Mann, M. (2000) Proteomics to study genesand genomes. Nature 405, 837-846.(3) Burbaum, J., and Tobal, G. M. (2002) Proteomics in drugdiscovery. Curr. Opin. Chem. Biol. 6, 427-433.(4) Cravatt B. F., and Sorensen E. J. (2000) Chemical strategiesfor the global analysis of protein function. Curr. Opin. Chem.Biol. 4, 663-668.(5) Patricelli, M. P. (2002) Activity-based probes for functionalproteomics. Briefings Funct. Genomics Proteomics 1 (2), 151-158.(6) Liu, Y., Patricelli, M. P., and Cravatt, B. F. (1999) Activitybasedprotein profiling: The serine hydrolases. Proc. Natl.Acad. Sci. U.S.A. 96(26), 14694-14699.(7) Jessani, N., Liu Y., Humphrey, M., and Cravatt B. F. (2002)Enzyme activity profiles of the secreted and membraneproteome that depict cancer cell invasiveness. Proc. Natl.Acad. Sci. U.S.A. 99 (16), 10335-10340.(8) Leung, D., Hardouin, C., Boger, D. L., and Cravatt, B. F.(2003) Discovering potent and selective reversible inhibitorsof enzymes in complex proteomes. Nat. Biotechnol. 21 (6),687-691.(9) Patricelli, M. P., Giang, D. K., Stamp, L. M., and BurbaumJ. J. (2001) Direct visualization of serine hydrolase activitiesin complex proteomes using fluorescent active site-directedprobes. Proteomics 1, 1067-1071.(10) Kidd, D., Liu, Y., and Cravatt, B. F. (2001) Profiling serinehydrolase activity in complex proteomes. Biochemistry 40,4005-4015.(11) Taylor, W. P., and Widlanski, T. S. (1995) Charged withmeaning: The structure and mechanism of phosphoproteinphosphatases. Chem. Biol. 2(11), 713-718.(12) van Huijsduijnen, R. H., Bombrun, A., and Swinnen, D.(2002) Selecting protein tyrosine phosphatases as drugtargets. Drug Discov. Today 7 (19), 1013-1019.(13) Honkanen, R. E., and Golden, T. (2002) Regulators ofserine/threonine protein phosphatases at the dawn of aclinical era? Curr. Med. Chem. 9(22), 2055-2075.(14) Lo, L.-C., Wang, H.-Y., and Wang, Z.-J. (1999) Design andsynthesis of an activity probe for protein tyrosine phosphatases.J. Chin. Chem. Soc. 46(5), 715-720.(15) Lo, L.-C., Pang, T.-L., Kuo, C.-H., Chiang, Y.-L., Wang, H.-Y., and Lin, J.-J. (2002) Design and synthesis of class-selectiveactivity probes for protein tyrosine phosphatases. J. ProteomeRes. 1 (1), 35-40.(16) Myers, J. K., and Widlanski, T. S. (1993) Mechanism-basedinactivation of prostatic acid phosphatase. Science 262, 1451-1453.(17) Cohen, P. T. W. (2002) Protein phosphatase 1sTargetedin many directions. J. Cell. Sci. 115 (2), 241-256.(18) Parsons, R. (1998) Phosphatases and tumorigenesis. Curr.Opin. Oncol. 10 (1), 88-91.(19) Haugland, R. P. Handbook of Fluorescent Probes andResearch Chemicals (2003) Molecular Probes, Inc., Eugene,Oregon.

798 Bioconjugate Chem., Vol. 15, No. 4, 2004 Shreder et al.Downloaded by WUHAN UNIV on October 25, 2009 | http://pubs.acs.orgPublication Date (Web): June 18, 2004 | doi: 10.1021/bc0499580(20) Patricelli, M. P., and Cravatt, B. F. (2000) Clarifying thecatalytic roles of conserved residues in the amidase signaturefamily. J. Biol. Chem. 275 (25), 19177-19184.(21) Adam G. C., Burbaum J., Kozarich J. W., Patricelli M. P.,and Cravatt B. F. (2004) Mapping Enzyme Active Sites inComplex Proteomes J. Am. Chem. Soc. 126 (5), 1363-1368.(22) Perkins, D. N., Pappin, D. J. C., Creasy, D. M., and Cottrell,J. S. (1999) Probability-based protein identification by searchingsequence databases using mass spectrometry data. Electrophoresis20 (18), 3551-3567.(23) Lam, P. K., Yang, M., and Lam, M. H. (2000) Toxicologyand evaluation of microcystins. Ther. Drug. Monit. 22 (1), 69-72.(24) Zhou, L., Yu, H., and Chen, K. (2002) Relationship betweenmicrocystin in drinking water and colorectal cancer. Biomed.Environ. Sci. 15 (2), 166-171.(25) Yu, S., Zhao, N., and Zi, X. (2001) The relationship betweencyanotoxin (microcystin, MC) in pond-ditch water and primaryliver cancer in China. Zhonghua Zhong Liu Za Zhi. 23(2), 96-99.(26) Runnegar, M. T., Kong, S., and Berndt, N. (1993) Proteinphosphatase inhibition and in vivo hepatoxicity of microcystins.Am. J. Physiol. 265 (2/1), 224-230.(27) Nishiwaki-Matsushima, R., Ohta, T., Nishiwaki, S., Suganuma,M., Kohyama, K., Ishikawa, T., Carmichael, W. W.,and Fujiki, H. (1992) Liver tumor promotion by the cyanobacterialcyclic peptide toxin microcystin-LR. J. Cancer Res.Clin. Oncol. 118 (6), 420-424.(28) Craig, M.; Luu, H.-A.; McCready, T. L.; Williams, D.;Andersen, R. J.; and Holmes, C. F. B. (1996) Molecularmechanisms underlying he interaction of motuporin andmicrocystins with type-1 and type-2A protein phosphatases.Biochem. Cell Biol. 74 (4), 569-578.(29) Gupta, V., Ogawa, A. K., Du, X., Houk, K. N., andArmstrong, R. W. (1997) A model for binding of structurallydiverse natural product inhibitors of protein phosphatasesPP1 and PP2A. J. Med. Chem. 40 (20), 3199-3206.(30) Craig, M., McCready, T. L., Luu, H. A., Smillie, M. A.,Dubord, P., and Holmes, C. F. B. Toxicon 1993, 31 (12), 1541-1549.(31) Goldberg, J.; Huang, H.; Kwon, Y.; Greengard, P.; Nairn,A. C.; and Kuriyan, J. (1995) Three-dimensional structure ofthe catalytic subunit of protein serine/threonine phosphatase-1. Nature 376, 745-753.(32) Bagu, J. R., Sykes, B. D., Craig, M. M., and Holmes, C. F.(1997) A molecular basis for the different interactions ofmarine toxins with protein phosphatase-1. Molecular modelsfor bound motuporin, microcystins, okadaic acid, and calyculinA. J. Biol. Chem. 272 (8), 5087-5097.(33) Lo, K. K.-W., Ng, D. C.-M., Lau, J. S.-Y., Wu, r. S.-S., andLam, P. K.-S. (2003) Derivatisation of microcystin with aredox -active label for high performance liquid chromatography/electrochemicaldetection. New J. Chem. 27, 274-279.(34) Murata, H., Shoji, H., Oshikata, M., Harada, K.-I., Suzuki,M., Kondo, F., and Goto H. (1995) High-performance liquidchromatography with chemiluminescence detection of derivatizedmicrocystins. J. Chromatograph. A 693, 263-270.(35) Campos, M., Fadden, P., Alms, G., Qian, Z., and Haystead,T. A. J. (1996) Identification of protein phosphatase-1-bindingproteins by microcystin-biotin affinity chromatography. J.Biol. Chem. 271 (45), 28478-28484.(36) Pflugmacher, S., Wiegand, C., Oberemm, A., Beattie, K.A., Krause, E., Codd, G. A., and Steinberg, C. E. W. (1998)Identification of an enzymatically formed glutathione conjugateof the cyanobacterial hepatotoxin microcystin-LR: Thefirst step of detoxification Biochem. Biophys. Acta 1425, 527-533.(37) Shimizu, M., Takahashi, T., Uratsuka, S., and Yamada,S. (1990) Synthesis of highly fluorescent dienophiles fordetecting conjugated dienes in biological fluid. J. Chem. Soc.Chem. Commun. 1416-1417.(38) Harada, K., Oshikata, M., Shimada, T., Nagata, A., Ishikawa,N., Suzuki, M., Kondo, F., Shimizu, M., and Yamada,S. (1997) High-performance liquid chromatographic separationof microcystins derivatized with a highly fluorescentdienophile. Natural Toxins 5, 201-207.(39) Yankeelov, J. A., Jr. (1972) Modification of arginine bydiketones Methods Enzymol. 25, 566-579.(40) Gilbert, H. F., and O’Leary, M. H. (1975) Modification ofarginine and lysine in proteins with 2,4-pentanedione. Biochemistry14, 5194-5199.(41) Kinzel, O.; and Budzikiewicz, H. (1999) Synthesis andbiological evaluation of a pyoverdin-beta-lactam conjugate:A new type of arginine-specific cross-linking in aqueoussolution. J. Pept. Res. 53, 618-625.(42) Park, I. K., Roach, P., Bondor, J., Fox, S. P., and DePaoli-Roach, A. A. (1994) Molecular mechanism of the synergisticphosphorylation of phosphatase inhibitor-2. Cloning, expression,and site-directed mutagenesis of inhibitor-2. J. Biol.Chem 269, 944-954.(43) Fersht, A (1985) Enzyme Structure and Mechanism, pp272-276, W. H. Freeman and Company, New York.(44) An, J., and Carmichael, W. W. (1994) Use of a colorimetricprotein phosphatase inhibition assay and enzyme linkedimmunosorbent assay for the study of microcystins andnodularins. Toxicon 32(12), 1495-1507.(45) Huang, X., Cheng, A., and Honkanen, R. E. (1997) Genomicorganization of the human PP4 gene encoding a serine/threonine protein phosphatase (PP4) suggests a commonancestry with PP2A. Genomics 44, 336-343.(46) Hastie, C. J., and Cohen, P. T. W. (1998) Purification ofprotein phosphatase 4 catalytic subunit: Inhibition by theantitumour drug fostriecin and other tumour suppressors andpromoters. FEBS Lett. 431, 357-361.(47) Craig, M., Luu, H. A., McCready, T. L., Williams, D.,Anderson, R. J., and Holmes, C. F. B. (1996) Molecularmechanisms underlying the interaction of motuporin andmicrocystins with type-1 and type-2A protein phosphatases.Biochem. Cell Biol. 74, 569-578.BC0499580