Alkylation of Staurosporine to Derive a Kinase Probe for Fluorescence Applications

Abstract The natural product staurosporine is a high‐affinity inhibitor of nearly all mammalian protein kinases. The labelling of staurosporine has proven effective as a means of generating protein kinase research tools. Most tools have been generated by acylation of the 4′‐methylamine of the sugar moiety of staurosporine. Herein we describe the alkylation of this group as a first step to generate a fluorescently labelled staurosporine. Following alkylation, a polyethylene glycol linker was installed, allowing subsequent attachment of fluorescein. We report that this fluorescein–staurosporine conjugate binds to cAMP‐dependent protein kinase in the nanomolar range. Furthermore, its binding can be antagonised with unmodified staurosporine as well as ATP, indicating it targets the ATP binding site in a similar fashion to native staurosporine. This reagent has potential application as a screening tool for protein kinases of interest.


Introduction
The natural product staurosporine, first described in 1977, [1] has been demonstrated to be ah igh-affinity inhibitoro fn early all mammalian protein kinases. [2,3] Staurosporineb inds to the ATPb inding site of the kinase, making multiple contacts with the hinge region and the N-and C-terminal lobes of the catalytic domain. Staurosporine itselfi sn ot used as at herapeutic agent, but has proveni nvaluable in the discoveryo fn ovel anticancerd rugs based on kinasei nhibition. [4,5] Furthermore, it has been modified to generate functionalised molecular tools such as immobilisedstaurosporine to capture kinases, [6] astaurosporine-based photoaffinity probe, [7] ac ell-permeable affinitybased probe for kinome labelling, [8] ah ighly selective bivalent staurosporine-tethered peptide ligand, [9] and af luorescent staurosporine conjugate. [10] The molecular interactions between staurosporine and numerous protein kinase catalytic domains have been determined in great detail in co-crystallisation experiments. [11][12][13][14][15][16] The main polar interaction points on staurosporine involvet he lactam oxygen and nitrogen atoms, the tetrahydropyran moiety,a sw ell as the 3'-methoxy and 4'-methylamine side groups (see Supporting Information for numbering). Interactions at the 4'-methylamine contribute to binding, and the number of interactions taking place at this group appears to correlatew ith affinity.B oth hydrogen bond and ionic interac-tions of this group have been implicated in binding, with different kinases engaging it in ad ifferent fashion. [17] For example, cAMP-dependent protein kinase (PKA) is predicted to make ah ydrogen bond as well as an ionic interaction, whilst the EGF receptor kinase relies on an ionic interaction and Fyn kinase on ahydrogen bond interaction with this group.
To derivatise staurosporine, most strategies have reliedu pon modification of this 4'-methylamine group, usually involving an acylation step. The resulting amide may be predicted to display decreased basicity with respect to the nitrogen atom, affectingi ts ability to act as hydrogen bond acceptor. As such, the acyl-based chemical modification may result in some loss of affinity of the probe relative to staurosporinei tself. This has indeedb een suggested as one of the reasonsw hy,c ompared with staurosporine, ap robe obtained by acylation of the secondary amine showed decreased affinity to the kinase ASK1. [10] However,a cylated tools have been shown to retain kinase binding and in somecases even showedaslight increase in affinity (e.g.,R ef. [8]). As an alternative to the acylation modification, we have soughtt od erivatise staurosporineu sing alkylation of the 4'-methylamine, which does not lead to formation of an amide. We report ah igh-affinity fluorescent ligand, consistingo fs taurosporine linked to fluoresceinv ia ap olyethylene glycoll inker and show that this can be used to interrogate the ATPb inding site by simple fluorescence polarisation detection. We show that the ligand is useful to predict compound interactions at the ATPb inding site of PKA and that the ligand is viable for the exploration of aw ide range of different kinases.

Synthesis
Staurosporine (1)( Scheme1)w as 4'-N alkylated with av ariety of alkyl halideso fw hichm ethyl bromoacetate was found to The naturalp roduct staurosporine is ah igh-affinity inhibitor of nearly all mammalian protein kinases.T he labellingo fs taurosporine has proven effective as am eanso fg enerating protein kinase research tools. Most tools have been generatedb ya cylation of the 4'-methylamine of the sugar moiety of staurosporine. Herein we describe the alkylation of this group as af irst step to generate af luorescently labelled staurosporine. Following alkylation,apolyethylene glycol linker was installed, allow-ing subsequenta ttachment of fluorescein. We report that this fluorescein-staurosporinec onjugateb inds to cAMP-dependent protein kinase in the nanomolarr ange. Furthermore, its binding can be antagonisedw ith unmodified staurosporine as well as ATP, indicatingi tt argetst he ATPb inding site in as imilar fashion to native staurosporine. This reagent has potentiala pplicationa sas creening tool for protein kinaseso fi nterest. afford the N-acetyl methyl ester 2 in quantitative yield. Compound 2 was sufficiently pure to use without furtherp urification so was aminolysed directly using ethylenediamine to give compound 3 in quantitativey ield. Compound 3 provided as uitable moietyf or further derivatisation of staurosporine with dyes, linkers or potentially other functionalities.
Compound 6 was deprotected using diethylamine, releasing the free amine 7,w hichw as again purifiedb yR P-HPLC. The overall yield of 43 %w as somewhat disappointing, so the final step was repeated as at wo-step one-pot procedure. Compound 6 was deprotected and then immediately coupled with fluorescein isothiocyanate to generate compound 8 with aslightly improvedyield of 52 %. The overall yield for the preparation of compound 8 from staurosporine( 1)w as 24 %. NMR analysisd emonstratedt hat the tethering point for the linker was indeed the 4'-methylamine.

Opticalproperties
Absorption and emission spectra of compound 8 and fluorescein isothiocyanate were compared. Figure 1d emonstrates that 8 displayed similara bsorbance and emission profiles to those of fluoresceini sothiocyanate itself. Thus the basic spectral behaviour of the fluorophore wasn ot affected by attachment of the modified staurosporine. Next we evaluated the anisotropic properties of 8 and compared thesea gain to those of fluoresceini sothiocyanate. Fluoresceini sothiocyanate provoked af ull depolarisation of incident polarised light at ac oncentration of 10 nm (Figure 1c). Compound 8 also depolarised the incident polarisedl ight;h owever,t he concentrationr equired was higher.T he minimum concentration of 8 resulting in near complete depolarisation was 50 nm (Figure 1c). This concentration was used to measure kinase binding in subsequent assays.

Evaluation of kinase binding
It is predicted that if compound 8 can bind to PKA, the depolarisationo bserved above (Figure 1c that the kinase bound the probe. Because this PKA preparation is semi-pure, it could be hypothesised that the reversal was not due to binding to PKA, but to nonspecific binding to non-PKA proteins in the preparation. To assess this, ap arallel incubation was carried out with excessu nmodified ('cold')s taurosporine which would prevent specific binding of the probe.A s expected, excess staurosporine did not affect the polarisation signal in the absence of PKA (as this is as ignal derived from the probe itself). In the presence of PKA, some reversal of the polarisation signal was observed even in the presence of excess 'cold' staurosporine ( Figure 2a, 'nonspecific'), which we attributed to ac ertain level of nonspecific interaction of the probe.T he differenceb etween the total and nonspecific was then taken as specific binding to PKA. This was used throughout the study to assess probe binding. The specific binding of 8 to PKA was time dependent( Figure2b). Increased incubation times were associated with ad ecrease in the variability of the signal. To estimate the binding constant,t he probe was incubated with increasing concentrations of PKA. Specific binding of compound 8 to PKA was saturable in aw ay that is compatible with ao ne-site binding event at the kinase ( Figure 2c). The binding constant observed for binding of compound 8 was 44 nm.T his is compatible with the K d value recorded previously for unmodified staurosporine binding to PKA. [3,17,19] To verify that 8 could be displaced from PKA, it was incubated at ac oncentration of 50 nm with PKA in the presence of increasingc oncentrationso fu nmodified staurosporine. As showni nF igure 2d,s taurosporinei nhibits the binding of compound 8 to PKAi nt he low nanomolar range with ap IC 50 value of 8.0 AE 0.1 (IC 50 = 11 nm). Similarly,b inding of 8 to PKA was antagonised by ATPi nt he presence of MgCl 2 ,i ndicating it interacted with the ATPb inding site of PKA ( Figure 2e).

Reversible binding with purified PKA catalytic subunit
The above-mentioned experimentsw ere conducted using as emi-pure preparation which used the whole PKA tetramer of two catalytic subunits and two regulatory subunits bound together. The binding characteristicso fc ompound 8 indicate binding to ab ona fide ATPb inding site. However,a ss hown in Figure 2a,adegree of nonspecific binding decreased the antagonisable window.W em easured the bindingo fc ompound 8 to ap urer PKA catalytic subunit preparation.T he binding window for specific binding (as assessed using the procedure in Figure 2a)i ncreased from 10-20 mP units using semi-pure PKA tetramer to~50 mP units for the purer PKA catalytic subunit (Figure3a).
The above data indicate that compound 8 can be used to identify compounds that bind to the ATPb inding site of PKA. To confirm this further,H -7, ak nown ATPb inding site blocker unrelated to staurosporine was tested in the fluorescencep olarisation binding assay described in Figure 3a.A ss hown in Figure 3b,H -7 antagonised the binding of compound 8 to PKA in ac oncentration-dependent fashion,w ith an IC 50 value of 3.9 mm.T his value is similart ot he K i reported for antagonism of PKA enzyme activity. [20]  8 were prepared at ac oncentration of 10 mm.Absorptiona nd emission spectraw ere recorded by measuring emission at 650 nm with excitation over 300-600 nm, and by measuring emission over 400-700nmw ith excitation at 350 nm, respectively.c )Fluorescence polarisationinr esponset of luorescein isothiocyanate and compound 8.Aset of serially diluted wells were preparedint riplicate for each of these, and fluorescence polarisation was measuredasd escribed in the Experimental Section.

Interaction of compound 8w ith other kinases
Whilst the experimentsa bove indicatet hat compound 8 is au seful probe to identify binding of compounds to the ATP binding site of PKA, the promiscuousn ature of staurosporine from which compound 8 was derived suggestst hat compound 8 couldb eamore universal kinase probe. Using the ability of the staurosporine moiety of compound 8 to inhibit kinase ac-tivity,w et ested this notion by evaluating the inhibition of ar ange of kinases by compound 8 and comparing this with PKA inhibition.
Ta ble 1s hows the inhibition of 50 kinases (including PKA)b y compound 8 at both 100 nm and 1 mm.I tc an be observed that compound 8 at ac oncentration of 100 nm inhibited PKA by about6 0%.G iven that compound 8 is ag ood fluorescence  www.chemmedchem.org polarisation probe for PKA, this suggests that ak inase that is inhibited by compound 8 to asimilarextent could be detected by compound 8 using fluorescence polarisation.

Discussion
Af luorescein-labelled staurosporine derivative (8)h as been prepared by alkylation in five steps with an overall yield of 24 %. Fluorescein is widely used as fluorophore in fluorescence polarisation assays. The coupling of staurosporine to the fluorophore did not affect the basic fluorescent characteristics of the fluorophore, and the resultingp robe was capable of depolarising plane-polarised light, indicating its potential application in fluorescencep olarisation assays. Using fluorescencep olarisation,compound 8 was shown to bind to PKA in the nanomolar range. Compound 8 bindingw as successfully antagonised using unmodified staurosporine as well as ATPa nd H-7, indicating it interacted with the ATPb inding site of the kinase and could monitorinhibitor binding.
The most frequently employed method for attachment to staurosporine has been acylation of the 4'-methylamine. Althought he 4'-methylamine is oriented towards solvent-accessible space,i ti si nvolved in binding to the protein. In many crystal structures the 4'-methylamine can be seen to be binding to acidic side chain residues in the kinase;f or example, Glu127 and Glu170 in the case of PKA, [21] and Asp807 in the case of ASK1. [22] For kinases that hydrogen bond via aspartate or glutamate to the 4'-methylamine nitrogen of staurosporine, a7 -t o8 0-foldd ecrease in affinity was observed upon acylation of this moiety to introduce aP EG linker. [9] However,a cylation did not affect the inhibition of c-Src by ar ecently developed covalently bindinga ffinity probe. [8] The role of the 4'methylamine and tetrahydropyran functionality of staurosporine may be inferred when comparing the IC 50 values for staurosporine andK 252c (staurosporine aglycone) which for PKC were 9a nd 680 nm,r espectively. [23] As imilarb ut larger difference was observed forP KA (40 and > 10 000 nm). [23] This difference in IC 50 suggests the importance of the tetrahydropyran functionality that bearst he 4'-methylamine with respectt or etainingi ts bondingi nteractions with the target protein. The tetrahydropyran ring of staurosporine is documented as being flexible, andc onformational changes have been observed which include movement of the 4'-methylamine. [24] This flexibility of the tetrahydropyran may be partly responsible for the broad-ranging affinity of staurosporine.
Acylation of the amine introduces an electron-withdrawing carbonyl group that greatly decreases the basicity of the nitrogen. In alkylating rather than acylating the 4'-methylamine, the basicity of the nitrogen is predicted to be preserved, while also maintaining the nitrogen atom's ability to act as ah ydrogen bond acceptor. [25] In the work presented herein, alkylation of the 4'-methylamine has successfully generated af unctional probe that has retained high affinity.A sc ompound 8 has been demonstrated to bind reversibly to the ATPb inding site of PKA, it could be employed to demonstrate whether ak nown inhibitor binds in ac ompetitive or allosteric manner,a so nly competitive blockers are expected to displace compound 8. This wasc onfirmed using H-7, ac ompetitive blocker of the ATPb inding site on PKA. Additionally,t he fact that H-7 displaced the probe from PKA indicates that the fluorescencep o- www.chemmedchem.org larisationf ormat used for compound 8 could be readily used in the screening of ATP-competitive inhibitors.

Conclusions
Staurosporine was derivatised successfully via af ive-step alkylation route to yield af luorescein-based fluorescent tool. The linking and labelling strategy could be readily modified to introduceawide range of other fluorophores or other moieties to staurosporine. The tool developedh ere interactedw ith PKA at the ATPb inding site and was shown to be useful to monitor inhibitorb inding at this site. We established for al arge number of kinases whether or not the tool developed in this study is predicted to interact. Thiss uggestst hat it is useful as probe for the identification of ATP-competitive blockers for numerous kinases.

Experimental Section
General:Solvents were obtained from Fisher and used without further purification. Reagents were obtained from Sigma-Aldrich unless otherwise stated. 1 Ha nd 13 CN MR spectra were obtained with Bruker AV(I) 400 MHz, Bruker AV(III) 400 MHz, and Bruker AV(III) 500 MHz instruments. Mass spectra (ToF ES AE)w ere recorded on am icromass LCT.R eversed-phase HPLC was performed on aW aters 2525 gradient module coupled with aW aters 2487 dual absorbance detector set at l 254 and 366 nm. Al inear gradient was run over 35 min, from 100/0 phase 1( deionised and degassed H 2 Ow ith 0.05 %t rifluoroacetic acid) to 100/0 phase 2( 90 %M eCN, 10 %H 2 Ow ith 0.05 %t rifluoroacetic acid). The analytical column used was aP henominex Luna C 8 ,1 50 4.6 mm, 5 mm, at af low rate of 1mLmin À1 .T he semipreparative column was aY MC C 8 150 10mm, 5 mm, at af low rate of 3.00 mL min À1 .T he preparative column was aP henominex Luna C 8 ,1 50 30mm, 5 mm, at af low rate of 20 mL min À1 .R etention times (t R )are given in minutes.
PKA was used in two forms:s emi-pure PKA from bovine heart as aw hole tetramer (Sigma-Aldrich P5511), and PKA catalytic subunit purified from P5511b yc olumn chromatography (Sigma-Aldrich P2645). PKA was added to the assay at the required number of enzyme units, and this was converted into am olar value using the specific activity of the P2645 preparation as ag uide (10 units per mgp rotein, one unit of PKA being equivalent to 2.3 pmol of kinase).
Fluorescence polarisation assays were performed in 384-well plates in af inal volume of 50 mLa tR Ti n0 .01 m phosphate buffer, 0.0027 m potassium chloride, and 0.137 m sodium chloride, pH 7.4. Fluorescence polarisation was measured on aP erkinElmer Wallac Envision 2104 Multilabel plate-reading spectrophotometer using 485 nm excitation and 535 nm emission filters, suitable for measurement of fluorescein. Fluorescence polarisation was determined by measuring the parallel and perpendicular fluorescence emission intensity with respect to the polarised excitation light and is expressed in millipolarisation (mP) units following Equation (1): in which S is the parallel and P is the perpendicular fluorescence signal, and G is the gain factor,w hich was set prior to the assay so that 1nm fluorescein isothiocyanate gives ar eading of 27 mP units. For fluorescence spectra in Figure 1, all solutions tested were made up in phosphate-buffered saline (PBS) 0.01 m phosphate buffer,0 .0027 m potassium chloride, and 0.137 m sodium chloride, pH 7.4, at 25 8Cw ith 2% DMSO final well concentration. The wells were prepared in triplicate or quadruplicate.