A Novel Photoaffinity-Based Probe for Selective Detection of Cathepsin L Active Form

A Novel Photoaffinity-Based Probe for Selective Detection of Cathepsin L Active Form
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  DOI: 10.1002/cbic.201200389 A Novel Photoaffinity-Based Probe for Selective Detectionof Cathepsin L Active Form Ana Torkar, [a] Sarah Bregant, [b] Laurent Devel, [b] Marko Novinec, [c] Brigita Lenarcˇicˇ, [c] Tamara Lah, [a, c] and Vincent Dive* [b] Introduction The functional roles of proteases in cancer progression are dif-ficult to study by traditional genomic and proteomic ap-proaches that focus on measuring changes in protein abun-dance. This parameter does not necessarily correlate with pro-tease activity because most proteases are expressed as inactivezymogens or exist in complexes with their endogenous inhibi-tors.The development of activity-based probes (ABPs) for papain-like cysteine proteases, however, has been particularly success-ful. This is mainly due to the availability of a large number of covalently reactive electrophilic functional groups that canreact with the conserved nucleophilic cysteine residue that isactivated by the nearby histidine residue in the cysteine pro-tease active site. [1,2] The development of selective ABPs thattarget individual members of these cysteine proteases eitherexclusively or preferentially, though, has been less successful.These cysteine proteases comprise the C1A clan, the subfamilyof eleven human cysteine cathepsins: B, C, H, F, K, L, O, S, V, Wand X/Z. They share a conserved active site, as well as broadand similar specificities of their substrate binding pockets, [3] sothe selective inhibition of individual members remains chal-lenging. Indeed, to the best of our knowledge, no cathepsin L-selective (CatL-selective) ABPs have been developed to date.The currently available ABPs for CatL are class-wide reactiveprobes for these cysteine proteases based on the electrophilicepoxide or acetyloxymethyl ketone warheads. [4,5] The lysosomal proteolytic enzyme CatL is a ubiquitous eu-karyotic endopeptidase that participates in protein turnover ina number of important cellular processes, such as the immunesystem responses, bone resorption, regulation of the cell cycle,development of the central nervous system, and skin homeo-stasis. CatL knock-out mice suffer from, for example, periodichair loss, epidermal hyperplasia, acanthosis, hyperkeratosis andabnormal spermatogenesis. [6] A combined deficiency of CatBand CatL in mice is lethal due to brain atrophy. [6] Human andanimal studies have also identified overexpression of CatL, to-gether with several other cysteine cathepsins, in various diseas-es, [7] including malignant cancer progression in breast, lung,gastric, colon, and head and neck carcinomas and melano-mas. [6,7] In gliomas, CatL overexpression correlates with pro-gression from astrocytoma to high-grade (WHO IV) malignantglioblastoma, the most common and fatal type of braintumour. [8,9] The role of CatL in glioblastoma might be in alter-ing the activation of transcription factors, which would haveimpact on cell senescence, chemoresistance and apopto-sis. [8,10–13] Unlike that of CatB, which is a marker of an invasivecell phenotype, [14–16] however, the role of CatL in glioblastomaremains incompletely understood [17] and so tools for its detec-tion in biological samples are needed.To progress further in the development of more selectiveCatL ABPs, in this study we evaluated the potential for probesbased on photoaffinity labelling. This approach has alreadyDetecting the active forms of proteases by using activity-basedprobes in complex proteomes has become an intensively in-vestigated field of research over the past years because manypathogenic conditions involve alterations in protease activities.The detection of lysosomal cysteine proteases, the cathepsins,has mostly relied on the use of probes that incorporate reac-tive electrophilic moieties to modify a cysteine in the activesite covalently. Here we report the first example of an activity-based probe that targets the cathepsins and incorporatesa photoactivatable benzophenone group for covalent labelling.When tested on a set of five cathepsins (B, K, L, S and V), thisprobe selectively labelled the active site of cathepsin L. Fur-thermore, when tested on crude cell extracts, the probe specif-ically detected cathepsin L quantities as low as a few pico-moles. This study suggests that photoaffinity labelling is apromising approach for developing highly selective and usefulcathepsin L probes. In particular, this probe might allow thedetection of small amounts of the secreted active cathepsin Lform in the cellular microenvironment in vitro and ex vivo. [a]  A. Torkar, Dr. T. LahDepartment of Genetic Toxicology and Cancer Biology National Institute of Biology Vecˇ na pot 111, 1000 Ljubljana (Slovenia) [b]  Dr. S. Bregant, Dr. L. Devel, Dr. V. DiveCEA, Service d’Ingnierie Molculaire des Protines (SIMOPRO)iBiTec-S, 91191 Gif-sur-Yvette (France)E-mail: vincent.dive@cea.fr  [c]  Dr. M. Novinec, Dr. B. Lenarcˇ icˇ  , Dr. T. LahFaculty of Chemistry and Chemical Technology, University of LjubljanaCesta v Mestni log 88a, 1000 Ljubljana (Slovenia) ChemBioChem  0000  , 00, 1–7    2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  & 1 & These are not the final page numbers!   been proven to be valuable in drug discovery for performingselective covalent modifications of protein targets. [18–20] Results and Discussion Probe design The photoaffinity-based probe  1  (Scheme 1) was based on theCbz-Phe-Lys-NH 2  sequence found in a reported acetyloxymeth-yl ketone (AOMK) inhibitor, a class-wide reactive inhibitor of cysteine proteases. [21] The lysine residue was used as theanchor to functionalise the inhibitor with the detectionmoiety: a fluorescent cyanine-3 (Cy3) group. [21,22] Bulky hydro-phobic and aromatic residues are accepted well in the S 2  sub-site of CatL, [22,23] so a benzoylphenylalanine amino acid wasplaced at this position as the photoactivatable group. Carbo-benzyloxy (Cbz) capping of the N terminus of the moleculewas selected in preference to the acetoxy moiety because CatLcan accommodate bulkier groups at this site than can CatB.Finally, a short di(ethylene glycol) moiety was added at the C-terminal position to aid in the probe’s water solubility. The syn-thesis route to this benzophenone probe  1  is outlined inScheme 1A.To compare the performance of this new probe, two ABPscontaining the electrophilic AOMK as a warhead group werealso prepared (Scheme 1B) as reported in the literature, [24] withslight protocol modifications (see the Supporting Information).Analogues of probe  2  with BODIPY and cyanine-5 fluorescent Scheme 1.  Chemical structures of the inhibitors designed for the cysteine cathepsins. A) Synthesis and chemical structure of probe  1 . a) SPPS; b) 20% piperi-dine, DMF, CbzCl, DIEA, DMF; c) 95% TFA; d) Cy3/DIEA/DMSO. The core molecule with no fluorescent group was obtained by standard solid-phase peptidesynthesis. B) Chemical structures of acetyloxymethyl ketone probes  2  and  3 . & 2 &  www.chembiochem.org   2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  ChemBioChem  0000  , 00, 1–7   These are not the final page numbers! V. Dive et al.  groups had previously been characterised as unselective elec-trophilic inactivators of cysteine proteases that can be used tolabel both CatL and CatB. [4,21] To achieve better selectivity forCatL, probe  3 , with a bulky biphenyl group at its C terminus,was also developed. CatB contains a 20-residue insertion loopnot present in CatL, and this blocks the active-site cleft on theprimed binding side. The presence of a bulky group at the C-terminal position in probe  3  should thus induce a steric clashwith CatB but not with CatL, promoting selectivity towardsCatL. Characterisation of benzophenone probe 1 towards CatLand CatB Probe  1  was initially characterised with the recombinant CatLand CatB enzymes in the absence of light activation, to deter-mine the probe’s potency for the formation of the noncovalentEI (enzyme–inhibitor) complex. The probe showed micromolarpotency and was about eight times more potent as an inhibi-tor of CatL ( K  i = 3.6  m  m ) than of CatB ( K  i = 28  m  m ).Covalent labelling of cathepsins by probe  1  was first testedwith CatL and CatB. Probe  1  (50  m  m ) was incubated with0.5  m  m  recombinant enzyme in the dark for a short time andthen irradiated for 40 min at 365 nm. After SDS-PAGE gel reso-lution to remove excess probe, the enzyme–probe fluorescentcovalent adduct (1.2 pmol) was visualised by fluorescenceimaging (Figure 1). Under these conditions, more intense label-ling of CatL than of CatB was observed with probe  1  (Fig-ure 1A), and integration of the band intensities demonstratedthat the fluorescent signal detected for CatL was approximate-ly 85% stronger than that for CatB. Competitive experimentswith active-site-directed inhibitors of CatL and CatB were thencarried out with epoxysuccinyl E-64 and AOMK GB111-NH 2 , re-ported by Bogyo and colleagues. [4,21] The data demonstratedthat the observed fluorescent complex resulted from a covalentmodification of residues present in the CatL active site byprobe  1  (Figure 1A). With CatB the competition experimentswith these inhibitors were less clear-cut than with CatL, whichsuggests possible labelling of CatB outside of its active site. Comparison of probe 1 with AOMK probes 2 and 3 Labelling of CatL by probe  1  was compared with that observedwhen the more conventional AOMK probes  2  and  3  wereused. As shown in Figure 1B, probes  2  and  3  covalently modi-fied both CatL and CatB, with probe  3  showing better selectivi-ty in favour of CatL labelling. Better reactivity of probe  2  withCatB was also shown by kinetic measurements (see Table S1 inthe Supporting Information). Unexpectedly, the band intensityobserved for the labelling of CatL by probe  1  was the same asthat observed for CatL labelling by the AOMK probes. Therewas no information in the literature on the cross-linking yieldsof AOMK inhibitors or derived probes with papain-like proteas-es. On the assumption that the mechanism of the reactions of AOMK probes with CatL is irreversible, the cross-linking effi-ciency would be expected to be near 100% after a prolongedtime of incubation with the enzyme. Even if very high yields of cross-linking have been reported in exceptional cases, [10,25] inmost cases a few per cent of the target protein are covalentlymodified with benzophenone photolabelling. Our results thatshow similar band intensities with probe  1  and with the twoAOMK probes thus suggest a high yield of cross-linking of CatLby probe  1 . Labelling of other cysteine cathepsins with probe 1 The most abundantly expressed human cysteine cathepsin isCatB, which is also the most intensely investigated in variousdiseases. This is followed by CatK, which is found in bone, andCatS, which is expressed in lymphoid tissues and cells. Allthree of these cathepsins are upregulated in various types of cancers. CatL, CatS, CatK and CatV have very high sequencehomology, with CatV having the highest sequence identitywith CatL at 80%. The affinities of probe  1  towards these cath-epsins were determined by measuring the dissociation con-stants of the enzyme–inhibitor complexes ( K  i ) in the absence of light activation. The  K  i  values of probe  1  towards all of thetested recombinant cathepsins (Table 1) were in the micromo-lar range and differed by less than two orders of magnitude;this is a common observation for reversible inhibitors of  Figure 1.  CatL labelling by probe  1  in comparison with other recombinant cathepsins and AOMK probes. A) The intensity of the band shows potent andselective labelling of CatL relative to CatB with probe  1 . Pretreatment of the recombinant enzymes (0.5  m  m ) with the general cathepsin inhibitors (50  m  m ),epoxysuccinyl E-64 and AOMK GB111-NH 2[4,21] for 20 min before the addition of probe  1  (50  m  m ) with incubation for 10 min prior to irradiation for 40 minat 365 nm. Data indicate that probe  1  covalently modifies residues within the CatL active site. B) Labelling of recombinant enzymes (0.5  m  m ) with probe  1 (50  m  m ) with incubation for 10 min prior to irradiation for 40 min at 365 nm, relative to labelling with AOMK probes  2  and  3  (50  m  m , 50 min incubation).C) Comparative labelling of various recombinant cysteine cathepsins (0.5  m  m ) by probe  1  (50  m  m , irradiation 40 min at 365 nm). ChemBioChem  0000  , 00, 1–7    2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  www.chembiochem.org  & 3 & These are not the final page numbers!  Selective Detection of Cathepsin L Active Form  papain-family cathepsins, which have very similar substratebinding pockets.As shown in Figure 1C, apart from CatL, only CatV showedlabelling after light activation, which is in agreement with theirclosest similarity in sequence. The average intensities of label-ling for each of the studied cathepsins were determined fromthree independent experiments. The labelling percentages (rel-ative to CatL) were calculated as 4% for CatB and CatK, 1% forCatS and 27% for CatV. Unlike that of the ubiquitously ex-pressed CatL, CatV localisation is restricted to thymus, testis,corneal epithelium and some colorectal and breast carcino-mas, [26] so probe  1  could be used elsewhere because labellingof CatV in other tissues would not be expected. Labelling of CatL with probe 1 in a complex proteome The presence of polar groups, necessary for good water solu-bility, in the structure of probe  1  prevents entry of this probeinto cells, as was confirmed by fluorescent microscopy after invitro incubation of the probe with the glioblastoma U87-MGcell line. Therefore, to evaluate the properties of probe  1  in acomplex proteome and to check for possible unselective label-ling of abundant proteins, the protein content from the glio-blastoma U87 cell line was isolated and incubated withprobe  1  (1  m  m ) for one hour prior to UV irradiation at 365 nm(Figure 2B). Under these conditions, only nonspecific labellingwas observed on SDS-PAGE gels, with no band at the expectedmolecular weight of CatL. When the unselective AOMK probe  2  was added to the same protein extract, one band withno background labelling was observed (Figure 2A). The use of selective inhibitors—CLIK148 for CatL and CA074 for CatB—in-dicated that this band corresponded to CatB expression in thecell extract. That only active CatB was present in this extract in-dicates why no labelling of CatL was observed with probe  1 .The lack of CatL detection might be due to inactivation of CatL during the preparation of cell extracts, as reported byYang and co-workers. [27] However, when recombinant CatL wasadded to the cell extract, we were able to label it with probe  1 (1  m  m ) under the same conditions as those described above.As shown in Figure 2B, a band corresponding to that expectedfor CatL was observed after irradiation of probe  1  with a mix-ture of 1.5 pmol (36 ng) recombinant CatL added to the cell-extract protein. With addition of 7.5 pmol (181 ng) of CatL, theunspecific labelling disappeared; this suggests that the pres-ence of CatL reduces nonspecific binding of the probe.Benzophenone is well known for its preferential reactionwith methionine, [28] so we investigated whether this residuewas present in the CatL active site. Interestingly, two methio-nines (Met70 and Met161) are present in the active site of CatL(Figure 3). Met70 is conserved in all of the cathepsins with theexception of CatB, so this methionine would not be expectedto be the target of the benzophenone core in probe  1 , a pro-posal also supported by the location of Met70 far from the S 2 subsite. Met161, on the other hand, is only observed in CatLand is located near to the S 2  subsite; this makes this residuea potential target for covalent modification by probe  1 . If true,this implies that CatV should be modified at another position,which might explain the lower yield of labelling observed forCatV. Despite some attempts to identify the target residue(s)by probe  1  in CatL, by approaches that had previously led usto identify the site of MMP-12 covalent modification by a pho- Table 1.  K  i  values of probe  1  towards cathepsins.CathepsinCatB CatL CatK CatS CatV K  i  [ m  m ] 28  6 3.6  0.5 11  1.5 1.10  0.2 0.54  0.04 Figure 2.  Selective labelling of recombinant CatL by probe  1  in the cellularextracts. A) Incubation of unspecific electrophilic AOMK probe  2  with thecell-extracted (CE) proteins from U87 glioblastoma cells led to labelling of only one cathepsin (lane 1), identified as CatB by competitive binding withits selective inhibitor CA074 (lane 3). CatL labelling with AOMK probe  2  wasnot achieved, as shown by the lack of competitive binding with CLIK148,which is a selective inhibitor of CatL (lane 2). B) Incubation of CE proteinsfrom U87 glioblastoma cells with probe  1  resulted in unspecific labelling(lane 2) because there is no active CatL in the cell homogenate. Upon addi-tion of recombinant CatL, specific labelling of CatL increases with the con-centration added in the complex CE proteins (lanes 3–5). Figure 3.  Crystal structure of CatL active site. The CatL backbone (PDB ID:1MHW) was generated with the aid of Pymol. The catalytic residue Cys25 isdisplayed in white stick form; Met70 and Met161 are in black stick form. & 4 &  www.chembiochem.org   2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  ChemBioChem  0000  , 00, 1–7   These are not the final page numbers! V. Dive et al.  toaffinity probe, [29,30] these efforts gave inconclusive results.However, identifying the sites modified by probe  1  in bothCatL and CatV would be helpful for optimising the probe se-lectivity.We have observed that the new benzophenone probe  1 does not penetrate the cell membrane and is therefore mostlysuitable for detection of extracellular levels of the active CatLform. The release of the inactive (precursor) form of CatL wasfirst suggested three decades ago, [31] but the site of its activa-tion remains unknown. Probe  1  is therefore an ideal tool for in-vestigating and possibly blocking the extracellular activation of CatL. On the other hand, cathepsins can also be secreted ex-cessively in their active forms in inflammatory and malignantdiseases, [32] and these can be studied either in vitro in cellularmodels or ex vivo in biological fluids. The latter might evenhave diagnostic and prognostic value, particularly in samplesfrom tumour patients. [33] Choosing a cyanine dye with no neg-ative charge might favour cell penetration, but as mentionedabove this would reduce probe solubility. If desirable, in thiscase more polar groups might be used to replace the Cbzgroup of probe  1 .There remains a lot of room to manipulate the structure of probe  1  so as further to improve its selectivity or its penetra-tion into other subcellular compartments. We have shownhere that specific and strong labelling of CatL can be achievedby using a probe that incorporates a photoactivatable benzo-phenone group, a possibility not previously examined in thedevelopment of a selective CatL-activity-based probe. Experimental Section Reagents/materials : Commercial reagents were used as received,without additional purification. Solvents from commercial sourceswere of reagent grade and were used without further purification.Rink Amide AM resin was from Merck, Fmoc-Lys(Boc)-OH from No-vabiochem, and Fmoc-8-amino-3,6-dioxaoctanoic acid from Neo-system. Fmoc-Bpa-OH,  N  , N  ’ -diisopropylcarbodiimide (DIC), anhy-drous DMF, benzyl chloroformate (BCF),  N  , N  -diisopropylethylamine(DIEA), Z-Phe-OH and triisopropylsilane (TIS) were from Sigma–Al-drich. Dichloromethane, piperidine and ethyl acetate were fromVWR Prolabo. 6-Chloro-1-hydroxybenzotriazole (ClHOBt) was fromMolekula and trifluoroacetic acid (TFA) from Fisher Scientific. Cya-nine-3 (Cy3) NHS ester was from Interchim and GE Healthcare.Microwave experiments were performed with a Discover apparatus(CEMmWave) in 10 mL sealed reaction tubes, or in open vesselmode with an SPS kit.Analytical and preparative RP-HPLC separations were performedwith a Shimadzu apparatus and analytical Supelco Ascentis Ex-press C18 (10 cm4.6 mm, 2.7  m  m, gradient 0 to 10 min/0 to 100%B, flow rate 1.2 mLmin  1 ), analytical Kromasil C8 (15 cm4.6 mm,gradient 0 to 20 min/0 to 100% B, flow rate 1.2 mLmin  1 ) andsemipreparative Supelco Ascentis C18 (15 cm10 mm, 5  m  m, gradi-ent 0 to 20 min/0 to 100% B, flow rate 4 mLmin  1 ). UV detectionwas performed at 230 nm. Solvent systems consisting of A) TFA inwater (0.1%), and B) TFA in acetonitrile (0.09%) were used. Analyti-cal RP-HPLC retention times ( t  R ) are reported in minutes.UV measurements were performed with a Shimadzu spectropho-tometer (UV 1800). From the DMSO stock solutions, each samplewas diluted 1000-fold in EtOH and measurements were carried outbetween 250 to 650 nm.HRMS data were registered with a 4800 MALDI-TOF mass spec-trometer (Applied Biosystems, Foster City, USA) in positive reflec-tron mode in the  m /  z   range of 100–700. Each spectrum was theresult of 1000 to 2000 shots (20 different positions into each spotand 50 shots per subspectrum) and internal calibration was appliedwith a 4-HCCA matrix. Chemistry : Benzophenone probe  1  was created by solid–phasesynthesis on Rink amide AM resin, by the standard Fmoc strategywith a microwave. Fmoc removal was performed with piperidine inDMF (20%, 33 min, 60 8 C, 25 W). After deprotection and coupling,the resin was washed with DMF (25 min) and CH 2 Cl 2  (25 min).Coupling of a PEG linker, lysine and benzophenone (10 equiv) wasperformed with the aid of ClHOBt (10 equiv) and DIC (10 equiv) inDMF (2coupling step, 5 min, 60 8 C, 25 W). Capping with a Cbzprotecting group was carried out under the same conditions asthe coupling, with the resin immersed in a DMF solution of BCF(6 equiv) and DIEA (12 equiv). The compound was cleaved fromthe solid support in a solution of TFA/TIS/H 2 O (95:2.5:2.5), lyo-philised from a CH 3 CN/H 2 O mixture and purified by preparativereversed-phase HPLC. After lyophilisation, the precursor was thendissolved in DMSO and the Cy3 NHS ester (2 equiv) was incorporat-ed at room temperature in the presence of DIEA (2 equiv) withgentle stirring. The completion of the reaction was monitored byanalytical HPLC (Ascentis express conditions). The resulting com-pound was then purified by semipreparative RP-HPLC and itspurity was assessed by analytical HPLC and HRMS analysis. Bythese criteria, probe  1  was > 95% pure (Table S2). Kinetic measurements : Recombinant human cathepsins CatL (EC3.4.22.15), CatB (EC, CatK (EC, CatS (EC CatV (EC were prepared at the Department of Bio-chemistry, Faculty of Chemistry and Chemical Technology, Universi-ty of Ljubljana, Slovenia. All inhibition assays were carried out inphosphate buffer (pH 6.0, 100 m m ) with EDTA (2 m m ) and dithio-threitol (DTT, 2 m m ) at 25 8 C with use of the following fluorogenicsubstrates (10  m  m ): benzyloxycarbonyl-Phe-Arg-7-amino-4-methyl-coumarin (ZFRAMC) for CatL ( K  m = 0.6  m  m ), CatV ( K  m = 4  m  m ) andCatK ( K  m = 35  m  m ), benzyloxycarbonyl-Arg-Arg-7-amino-4-methyl-coumarin (ZRRAMC) for CatB ( K  m = 248  m  m ) and benzyloxycarbonyl-Val-Arg-7-amino-4-methylcoumarin (ZVRAMC) for CatS ( K  m = 6.4  m  m ). All substrates were from Bachem, Switzerland. Finalenzyme concentrations were as follows: 0.1 n m  CatL and CatV,0.5 n m  CatK and 1 n m  CatS and CatB. All assays were performed inblack 96-well plates with a total reaction volume of 100  m  L (Corn-ing–Costar, Schiphol-RijK, the Netherlands). In all assays substrateconsumption was less than 10% substrate hydrolysis (initial rateconditions). Fluorescence signals were monitored with a SynergyMx spectrophotometer (Bio-Tek Instruments, Inc., Winoosky, USA).The formation of the fluorescent degradation product was moni-tored continuously at 380 nm excitation and 460 nm emissionwavelengths. The DMSO concentration during all measurementswas below 1% or a maximum of 5% in the whole tested concen-tration range.Various concentrations of the reversible inhibitor probe  1  were pre-incubated with cathepsins for a short time before the addition of substrate. Inhibitor concentrations were selected to provide 20%to 80% inhibition.  K  i  values were determined by the method pro-posed by Horovitz and Levitski. [34] Data were analysed by usingKaleida Graph. ChemBioChem  0000  , 00, 1–7    2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim  www.chembiochem.org  & 5 & These are not the final page numbers!  Selective Detection of Cathepsin L Active Form
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!