Sesquiterpene lactones inhibit luciferase but not β-galactosidase activity in vitro and ex vivo

Analytical Biochemistry 328 (2004) 147–154

www.elsevier.com/locate/yabio

Sesquiterpene lactones inhibit luciferase but not �-galactosidase activity in vitro and ex vivo

Maja T. Lindenmeyer,a Alfonso J. García-Piñeres,a Victor Castro,b and Irmgard Merforta,¤

a Institut für Pharmazeutische Wissenschaften, Lehrstuhl für Pharmazeutische Biologie, Albert-Ludwigs-Universität Freiburg,Stefan-Meier-Str. 19, 79104 Freiburg, Germany

b Escuela de Quimica, CIPRONA, Universidad de Costa Rica, San Jose, Costa Rica

Received 29 October 2003

Abstract

Reporter enzymes such as WreXy luciferase or �-galactosidase of Escherichia coli are frequently used to study transcriptionalactivity of genes and to investigate the eVects of novel compounds on gene or transcription factor activity. It is generally assumedthat the activity of these enzymes is unaVected by the treatment conditions. Therefore, this factor is not considered when interpretingthe data obtained. Biologically active compounds such as sesquiterpene lactones (SLs) have also been tested in reporter gene assaysfor their inXuence on gene expression. Here we show in in vitro and ex vivo experiments that SLs inhibit WreXy luciferase activityprobably by direct targeting of the enzyme while �-galactosidase remains almost completely unaVected. The loss of luciferase activityafter SL treatment could be an eVect of their sulfhydryl-modifying potency and the subsequent alteration of the enzyme’s tertiarystructure. These results demonstrate that the eVect of the test substance on the reporter enzyme used should be taken into consider-ation when the transcriptional eVect of novel compounds is investigated. 2004 Elsevier Inc. All rights reserved.

Keywords: Reporter gene assay; Luciferase; �-Galactosidase; Sesquiterpene lactones; Parthenolide

Chronic inXammatory diseases, such as rheumathoidarthritis, asthma, and inXammatory and autoimmunediseases, are characterized by an increased expression ofinXammatory genes. These diseases result from an inter-play of genetic and environmental factors. One of thesefactors is the ubiquitous transcription factor NF-�Bwhich plays a crucial role in immune and inXammatoryresponses and cell growth by regulating the expressionof speciWc cellular genes [1,2]. NF-�B is a dimeric tran-scription factor formed by the hetero- or homodimeriza-tion of proteins of the rel family [3]. It has been shownthat activation of NF-�B that is consistent with I�B deg-radation, NF-�B nuclear translocation, or binding to theDNA does not imply transcriptional activation.Depending on the activating stimulus, NF-�B itself issubject to posttranslational modiWcations that can

enhance transcriptional activation of NF-�B-dependentgenes [2,4].

As NF-�B plays a pivotal role in inXammation,antagonizing NF-�B activity is a potential therapeutictarget to treat immune and inXammatory diseases. It hasbeen shown that numerous synthetic and natural com-pounds block NF-�B activation through multiple mech-anisms by interfering with various steps of the NF-�Bactivation cascade [1,5,6]. Thus, there are diVerent meth-ods for screening the ability of a compound to blockNF-�B activation. To examine the eVect of inhibitors onNF-�B nuclear translocation and DNA binding, electro-phoretic mobility shift assays (EMSA)1 are used andreporter gene assays are performed to investigate eVectson �B-dependent transcription [7].

¤ Corresponding author. Fax: + 49-761-203-8383.E-mail address: [email protected] (I.

Merfort).

1 Abbreviations used: EMSA, electrophoretic mobility shift assay;SLS, sesquiterpene lactones; CMV, cytomegalovirus; HSV-TK, Herpessimplex virus thymidine kinase; PBS, phosphate-buVered saline; BSA,bovine serum albumin.

0003-2697/$ - see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ab.2004.01.021

148 M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154

Two of the commonly used reporter enzymes areWreXy luciferase and �-galactosidase from Escherichiacoli. FireXy luciferase is a 62-kDa molecular weight oxy-genase with known crystal structure [8] that generates abioluminescent signal by catalyzing the oxidation of itssubstrate luciferin in a two-step process [7,8]. �-Galacto-sidase from E. coli is a glycosidase which catalyzes thehydrolysis and transgalactosylation of �-D-galactosides.It is a well-characterized tetrameric enzyme which con-sists of identical subunits with molecular masses of116 kDa and possesses four binding sites for galactosides[9–11]. In most cases, it is assumed that the activity ofthese reporter enzymes is unaVected by the treatmentconditions, and interference of the speciWc treatmentwith the reporter enzyme is normally not consideredwhen interpreting the data obtained from these assays.But both enzymes, WreXy luciferase and �-galactosidase,can be inactivated under certain conditions such as heatshock [12] or in the presence of proteasome inhibitors[13], alkylating substances [14,15], substrate analogues[9,16], or other compounds [17,18].

As mentioned above, NF-�B has become an impor-tant target in drug development as inhibitors of thistranscription factor could prove to be eYcacious antiin-Xammatory agents. Among the NF-�B-targeting naturalcompounds, sesquiterpene lactones (SLs) deserve specialinterest. They are the active constituents of many medic-inal plants from the Asteraceae family, whose prepara-tions are used for the treatment of inXammation intraditional medicine. It has been shown in variousassays that plant extracts and the puriWed SLs possessantiinXammatory properties [19]. The antiinXammatoryactivity of SLs is most often chemically mediated by �,�-unsaturated carbonyl structures, such as an �-methy-lene-�-lactone or an �,�-unsubstituted cyclopentenone.These functional groups are known to react with nucleo-philes, especially with cysteine sulfhydryl groups, in aMichael-type addition [19]. We have recently demon-strated that SLs inhibit activation of NF-�B by alkylat-ing its p65 subunit at Cys38, which participates in DNAbinding [20,21]. As not only inhibition of DNA bindingby SLs is of interest, several studies have been carriedout to examine their eVect on NF-�B-driven geneexpression by using reporter gene assays, especially withluciferase as a reporter [21–25]. However, in all thesestudies it was not assessed whether the WreXy luciferaseitself is targeted by SLs, which cannot be excludedbecause of their alkylating properties.

For this reason we investigated the eVects of twodiVerent SLs on WreXy luciferase ex vivo and in vitro.Our investigations show that WreXy luciferase is inhib-ited by SLs and loses its activity in a concentration-dependent manner, whereas �-galactosidase, which wasalternatively studied, remains almost completelyunaVected. Thus, use of SLs in combination with theluciferase reporter enzyme may lead to a reduction of

enzyme activity and to a misinterpretation of quantita-tive data.

Materials and methods

Test compounds

Parthenolide was purchased from Sigma. 4�,15-Epoxy-miller-9E-enolide was isolated from leaves ofMilleria quinqueXora [26].

Plasmids and constructs

pNF�B-Luc (Clontech) contains four tandem copiesof the NF-�B consensus sequence fused to a TATA-likeregion from Herpes simplex virus thymidine kinase(HSV-TK) promoter in front of a luciferase gene (luc).p�gal-Basic (Clontech) is a mammalian reporter vectorthat lacks eukaryotic promoter and enhancer sequencesand serves as a promoter-cloning vehicle for strong pro-moters. pRC/CMV, a mammalian cloning and expres-sion vector containing a cytomegalovirus promoter, waspurchased from Invitrogen. pM1-�-Gal expression vec-tor (Roche Diagnostics) contains E. coli �-galactosidaseunder the control of the human cytomegalovirus imme-diate-early promoter/enhancer. pCL encodes WreXyluciferase under the control of a cytomegalovirus pro-moter. It was constructed by using pNF-�B-Luc (Clon-tech) as a template for PCR ampliWcation of theluciferase gene (luc) with the forward primer 50-CAGAAG CTT GGC ATT CCG GTA C-30 (HindIII site)and reverse primer 50-GCG ACT CTA GAA TTA CACGGC GAT C-30 (XbaI site). The forward and reverseprimers contain a recognition site for HindIII and XbaI,respectively, which are underlined. The PCR productwas digested with HindIII and XbaI and ligated intopRc/CMV (Invitrogen). pN�G encodes �-galactosidasefrom E. coli under the control of four tandem copies ofthe NF-�B consensus sequence fused to a TATA-likepromotor region from HSV-TK promoter. It was con-structed by using pNF-�B-Luc (Clontech) as a templatefor PCR ampliWcation of the NF-�B promotor regionwith the forward primer 50-GTA GGC TGT CTC GAGTGC AAG TG-30 (XhoI site) and the reverse primer 50-CGT TAT GCC AAG CTT CTG CTT CAT C-30 (Hin-dIII site). The forward and reverse primers contain arecognition site for XhoI and HindIII, respectively,which are underlined. The PCR product was digestedwith XhoI and HindIII and ligated into p�-gal-Basic(Clontech).

Cell culture

293 cells were maintained in Dulbecco’s modiWedEagle’s medium (Gibco), supplemented with 10% fetal

M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154 149

bovine serum (Sigma), 100 IU/ml penicillin, and 100 �g/ml streptomycin (Roche Diagnostics). TNF-� was pur-chased from Roche Diagnostics.

Transient transfection and luciferase assay

293 cells were plated 12–16 h prior to transfection at adensity of 1 £ 105 cells per each well in a six-well tissueculture dish and transfected with the newly constructedvector pCL. Transfections were performed by usingSuperFect (Qiagen) according to the manufaturer’s rec-ommendations. Two days after transfection, cells weretreated with the test compounds for 1 h prior to harvest-ing. Cells were then harvested in lysis buVer (supple-mented with the Luciferase Reporter Gene System fromRoche Diagnostics) and WreXy luciferase activity wasmeasured using the Luciferase Reporter Gene AssaySystem from Roche Diagnostics. Luciferase activity wasthen normalized to total cellular protein using the Pro-tein Assay ESL (Roche Diagnostics). Experiments wereperformed in triplicate and all experiments wererepeated at least three times.

Transient transfection and �-galactosidase assay

293 cells were plated on a 12-well tissue culture dish12–16 h prior to transfection at a density of 5 £ 104 cellsper each well and transfected with pM1-�-Gal (RocheDiagnostics). Transfections were performed by usingSuperFect (Qiagen) according to the manufacturer’s rec-ommendations. Forty-eight hours after transfection cellswere treated with the test compounds for 1 h prior toharvesting. Cells were then harvested in lysis buVer (sup-plemented with the �-Gal Reporter Gene System fromRoche Diagnostics) and �-galactosidase activity was mea-sured using the �-Gal Reporter Gene System from RocheDiagnostics. �-Galactosidase activity was then normalizedto total cellular protein using the Protein Assay ESL(Roche Diagnostics). Data represent means§ standarddeviation from at least three independent measurementsin triplicate. For monitoring the NF-�B-dependent geneexpression, 293 cells were plated on a 12-well tissue culturedish 12–16 h prior to transfection at a densitiy of1£105 cells per each well and transfected with the newlyconstructed vector pN�G. Forty-eight hours after trans-fection cells were treated with the test compounds for 1hand subsequently stimulated for 3 h with 5 ng/ml TNF-�,after which lysates were prepared and measured for �-galactosidase activity as described above. All experimentswere performed in triplicate and were repeated at leastthree times.

Determination of the transfection rate by X-gal staining

293 cells were plated on a 12-well tissue culture dish12–16 h prior to transfection at a density of 5 £ 104 cells

per each well and transfected with pM1-�-Gal (RocheDiagnostics). Transfections were performed by usingSuperFect (Qiagen) according to the manufacturer’s rec-ommendations. Forty-eight hours after transfection cellswere washed twice with PBS and then Wxed with 0.25%(v/v) glutaraldehyde solution for 10 min at room temper-ature. After washing three times with PBS cells werestained with X-gal reagent (100 mg/ml 5-bromo-4-chloro-3-indolyl-�-D-galactoside in dimethyl-formam-ide), diluted 1:100 with X-gal solution (5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, 2 mM MgCl2), and incubatedin the dark at 37 °C for 3 h. Analysis of X-gal stainingdemonstrated a transfection eYciency of 40–50%.

In vitro luciferase assay

To study the direct inhibition of the enzyme, recombi-nant WreXy luciferase (Promega) was diluted in LuciferaseReporter Gene Assay lysis buVer (Roche Diagnostics),supplemented with 1mg/ml BSA (Sigma), to obtain aconcentration of 1 ng/15 �l. This luciferase solution wastreated with the test compounds for 1 h at room temper-ature and then the activity was assayed with the Lucifer-ase Reporter Gene Assay. Experiments were performedin triplicate and were repeated at least two times.

In vitro �-galactosidase assay

To investigate the direct inhibition of the �-galactosi-dase enzyme, recombinant �-galactosidase from E. coli(Roche Diagnostics), diluted in �-Gal Reporter GeneAssay lysis buVer (Roche Diagnostics), was used at aconcentration of 0.05 ng/�l and treated with the testcompunds for 1 h at 37 °C; then �-galactosidase activitywas assayed with the �-Gal Reporter Gene Assay.Experiments were performed in triplicate and wererepeated at least two times.

Statistical analysis

Statistical analysis was performed using the Origin 7.0software. Data are reported as means § SD and analyzedusing an independent t test (two groups); p 0 0.05 is con-sidered statistically signiWcant (*p 0 0.05, **p 0 0.005versus positive control, op 0 0.05, oop 0 0.005 compari-son between in vitro and ex vivo data.

Results

Sesquiterpene lactones aVect luciferase activity

To investigate whether SLs interfere with luciferaseactivity, 293 cells were transiently transfected with aluciferase reporter gene driven by the cytomegaloviruspromoter. These transfected cells were left untreated or

150 M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154

treated for 1 h at 37 °C with diVerent concentrations oftwo SLs, parthenolide or 4�,15-epoxy-miller-9E-enolide(Fig. 1). Cells were subsequently harvested and luciferaseactivity was determined in total cell extracts and normal-ized to total protein concentration (Figs. 2A and B).Luciferase activity was inhibited by both SLs in a con-centration-dependent manner. Compared with theuntreated cells, which produce luciferase constitutively,treatment with 40 �M parthenolide or 10�M 4�,15-epoxy-miller-9E-enolide led to a loss of »80% of lucifer-ase activity. To exclude cytotoxic eVects of the SLs, cellviability after treatment with parthenolide or 4�,15-epoxy-miller-9E-enolide tested with trypan blue stainingwas more than 90% (data not shown).

To conWrm that inactivation of luciferase activity wasdue to a direct inhibition of the enzyme, and not to anyother phenomenon, recombinant WreXy luciferase wasincubated for 1 h at room temperature with the same con-centrations of the SLs and then its activity was assayed(Figs. 2A and B). Recombinant luciferase was also inhib-ited in a concentration-dependent manner. In contrast tothe ex vivo data, the loss of luciferase activity in the invitro experiments was lower at the same SL concentra-tion. This could be due on the one hand to the lower incu-bation temperature (20 °C instead of 37 °C with all cellcultures), given that the enzyme loses its activity at highertemperatures [12]. On the other hand, SLs could bind tocomponents of the reaction mixture such as BSA, whichis added in the in vitro experiment at a 1 mg/ml concen-tration as a stabilizer for recombinant luciferase. Dataobtained in our group show that SLs bind to BSA(S. Wagner, current thesis). This would lead to smalleramount of SLs available for reaction with luciferase.

Interestingly, 4�,15-epoxy-miller-9E-enolide showsbimodal behavior in the ex vivo data (Fig. 2B). Toexplore this phenomenon a time course using the twolowest concentrations was carried out. Transiently trans-fected cells were incubated with 1 or 2.5�M 4�,15-epoxy-miller-9E-enolide for various time intervals at37 °C and subsequently harvested, and luciferase activitywas determined in total cell extracts (Figs. 3A and B).The same procedure was performed with the recombi-nant enzyme at 20 °C. At a 1 �M concentration, a

decrease of luciferase activity was observed only for theex vivo data. For the higher concentration of 2.5 �M aslight activity loss with increasing time was observed inthe in vitro experiments, while the ex vivo data showedbimodal behavior.

�-Galactosidase activity remains nearly unaVected bysesquiterpene lactones

To Wnd an alternative reporter enzyme for monitoringNF-�B-dependent gene expression we investigatedwhether SLs would interfere with �-galactosidaseactivity. For the ex vivo experiments, 293 cells wereFig. 1. Structures of the investigated sesquiterpene lactones.

Fig. 2. Reduction of luciferase activity ex vivo ( ) and in vitro (�) byparthenolide (A) and 4�,15-epoxy-miller-9E-enolide (B). For the exvivo experiments, 293 cells were transiently transfected with pCL.Transfected cells were left untreated (+) or treated with the indicatedconcentrations of parthenolide [5–40 �M] (A) or 4�,15-epoxy-miller-9E-enolide [1–10 �M] (B) for 1 h. Subsequently, cells were harvestedand WreXy luciferase activity was measured using the LuciferaseReporter Gene Assay System from Roche Diagnostics. Mean valuesfrom at least three independent experiments performed in triplicate areshown. In in vitro test, recombinant luciferase at a concentrationof 1 ng/15 �l was incubated with the same concentrations of partheno-lide (A) or 4�,15-epoxy-miller-9E-enolide (B) for 1 h at room tempera-ture or left untreated (+). Mean values from at least two independentexperiments performed in triplicate are shown.

M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154 151

transiently transfected with pM1-�-Gal plasmid, whichcontains E. coli �-galactosidase under the control of thehuman cytomegalovirus immediate-early promoter/enhancer. The transfected cells were left untreated ortreated for 1 h at 37 °C with various concentrations ofparthenolide or 4�,15-epoxy-miller-9E-enolide. �-Galac-tosidase activity was determined in the total cell extractsand normalized to total protein concentration (Figs. 4Aand B). Both parthenolide and 4�,15-epoxy-miller-9E-enolide showed negligible inhibiting eVects on �-galacto-sidase. To corroborate these data recombinant E. coli �-

galactosidase was treated with the two substances for 1 hat 37 °C. The same results as those in the ex vivo experi-ments were obtained.

Sesquiterpene lactones inhibit NF-�B-dependent geneexpression

Given that luciferase is directly attacked by SLs, thestrength of inhibition of NF-�B-driven gene expressionafter incubating with these substances can be misinter-preted. Luciferase is not the most appropiate enzyme in

Fig. 3. Time course of the reduction of luciferase activity ex vivo ( )and in vitro (�) by 1 �M (A) or 2.5 �M (B) 4�,15-epoxy-miller-9E-eno-lide. For the ex vivo experiments, 293 cells were transiently transfectedwith pCL. Transfected cells were left untreated (+) or treated with1 �M (A) or 2.5 �M (B) of 4�,15-epoxy-miller-9E-enolide for the indi-cated time (15–120 min). Subsequently, cells were harvested and WreXyluciferase activity was measured using the Luciferase Reporter GeneAssay System from Roche Diagnostics. Mean values from at leastthree independent experiments performed in triplicate are shown. Inthe in vitro test, recombinant luciferase at a concentration of 1 ng/�lwas incubated with 1 �M (A) or 2.5 �M (B) of 4�,15-epoxy-miller-9E-enolide at room temperature for the same incubating times or leftuntreated (+). Mean values from at least two independent experimentsperformed in triplicate are shown.

Fig. 4. InXuence of parthenolide (A) and 4�,15-epoxy-miller-9E-eno-lide (B) on �-galactosidase activity ex vivo ( ) and in vitro (�). Forthe ex vivo experiments, 293 cells were transiently transfected withpM1-�Gal. Transfected cells were left untreated (+) or treated with theindicated concentrations of parthenolide [5–40 �M] (A) or 4�,15-epoxy-miller-9E-enolide [1–10 �M] (B) for 1 h. Subsequently, cells wereharvested and �-galactosidase activity was measured using the �-GalReporter Gene System from Roche Diagnostics. Mean values from atleast three independent experiments performed in triplicate are shown.In the in vitro test, recombinant �-galactosidase at a concentrationof 0.05 ng/�l was incubated with the same concentrations of partheno-lide (A) or 4�,15-epoxy-miller-9E-enolide (B) for 1 h at room tempera-ture or left untreated (+). Mean values from at least two independentexperiments performed in triplicate are shown.

152 M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154

a reporter gene assay for alkylating compounds, such asSLs. As a consequence, previous analyses that used lucif-erase as a reporter gene to assign an inhibitory role toSLs on NF-�B-dependent transcription must be reas-sessed [23,27]. �-Galactosidase was not aVected by SLs.For this reason, we studied their inXuence on NF-�B-driven gene expression with a �-galactosidase reportersystem. We transiently transfected 293 cells with a �-galactosidase reporter gene under the control of fourtandem copies of the NF-�B consensus sequence fusedto a TATA-like promotor region from Herpes simplexvirus thymidine kinase promoter. These cells were incu-bated with various concentrations of parthenolide or4�,15-epoxy-miller-9E-enolide for 1 h and stimulated for3 h with 5 ng/ml TNF-�. Cells were subsequently har-vested; �-galactosidase activity was determined in total

cell extracts and normalized to total protein concentra-tion (Figs. 5A and B). The induction of the NF-�B tran-scription by stimulating cells with TNF-� wassuppressed in a concentration-dependent manner. Nocytotoxic eVects were observed. Cell viability was higherthan 90% after pretreatment with SLs and stimulationwith TNF-� (data not shown). Furthermore, whetherconstitutively expressed �-galactosidase produced frompM1-�-Gal remains unaVected after treatment with20 �M parthenolide or 10�M 4�,15-epoxy-miller-9E-enolide was tested. Even after 4 h of treatment, �-galac-tosidase was not aVected (data not shown).

Discussion

Reporter enzymes such as chloramphenicol acetyl-transferase, WreXy luciferase, and �-galactosidase fromE. coli are frequently used for studies of gene regulation,gene activity, and expression in eukaryotic cells [7]. Inparticular WreXy luciferase has been extensively applied inmolecular and cell biology studies because of its negligi-ble background and high sensitivity and the relative sim-plicity of the assay [28,29]. These reporter enzymes areused with diVerent cell types and treatment conditionsand, in general, the interference of the experimental treat-ment on the reporter enzyme activity is not considered.

The biologically active SLs belong to those naturalproducts that have also been tested in reporter geneassays for their inXuence on gene expression [21,22,24,25,30]. Although their alkylating potency is known, untilnow studies on the eVect of SL treatment on the activityof reporter enzymes such as luciferase or �-galactosidasehave not been conducted. Here we show in in vitro andex vivo experiments that two diVerent SLs, parthenolideand 4�,15-epoxy-miller-9E-enolide, inhibit WreXy lucifer-ase activity by directly targeting the enzyme, while �-galactosidase from E. coli remains almost completelyunaVected (Figs. 2A, B and 4A, B). Because of the shorthalf-life of luciferase (t1/2 D 3–4 h in mammalian cells)[31] an incubation time of 1 h was used. A higher inhibi-tion will probably be observed with longer incubationtimes.

FireXy luciferase possesses four cysteine residues(Cys81, Cys216, Cys285, Cys391). One cysteine residue(Cys391) is located close to the active site, while the othersare further away [8]. Even though mutagenesis studiescould show that cysteine residues are not absolutelyrequired for bioluminescence activity [15], signiWcantchanges in luciferase activity were observed aftertreatment with the alkylant N-tosyl-L-phenylalaninechloromethyl ketone or cysteine mutagenesis [15], sug-gesting that an alteration in the tertiary structure of theenzyme could aVect the active site. Consequently, theloss of luciferase activity after SL treatment could alsobe an eVect of their sulfhydryl-modifying potency and

Fig. 5. Inhibition of NF-�B-dependent gene expression by partheno-lide (A) or 4�,15-epoxy-miller-9E-enolide (B). 293 cells were transientlytransfected with pN�G. Transfected cells were left untreated (¡) orstimulated with 5 ng/ml TNF-� alone for 3 h (+) or pretreated for 1 hwith the indicated concentrations of parthenolide (A) or 4�,15-epoxy-miller-9E-enolide (B) and subsequently stimulated with 5 ng/ml TNF-�for 3 h. Cells were then harvested and �-galactosidase activity wasmeasured using the �-Gal Reporter Gene System from Roche Diag-nostics. Mean values from at least three independent experiments per-formed in triplicate are shown.

M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154 153

the subsequent alteration of the enzyme’s tertiary struc-ture. However, it cannot be excluded that additionalreactions of SLs with other amino acids, such as lysine,may also contribute to the inhibitory eVect. This wouldbe in accordance with recent studies, which showed thatSLs do not solely react with the exposed sulfhydrylgroup in human serum albumin (S. Wagner, current the-sis). Concerning the bimodal behavior of the SL 4�,15-epoxy-miller-9-E-enolide, further investigations arerequired to explain this unexpected result.

Despite the fact that �-galactosidase possesses 16 cys-teine residues in the reduced form per subunit [10], anegligible eVect on its activity could be observed with thetwo SLs. This result shows that SLs diVer from thealkylating reagents such as iodoacetamide or bromoace-tamide which reversibly alkylate the �-galactosidase at amethionyl residue near the active site [32]. The X-raystructure of �-galactosidase [10] showed that threecysteine residues (Cys402, Cys500, Cys536) are situatednear the active site of the enzyme [9,10], while theremaining cysteine residues are distributed on theenzyme surface or in an inaccessible position or at somedistance from the active site. Therefore, alkylation ofresidues located on the surface should not aVect theenzyme activity, as an alteration of the tertiary structureis not expected. Consequently, only the three cysteineresidues near the active site could have eVects on thecatalytical activity after modiWcations with an alkylatingreagent. As no inhibition of the enzyme was observed, amodiWcation of these residues can be excluded or, atleast, they should not aVect the enzyme’s activity. Thisresult shows that, although SLs are unspeciWc alkylatingreagents, they still possess some selectivity for theirtarget residues.

By comparing Figs. 2–4, it becomes evident that inhi-bitions in ex vivo experiments are nearly always strongerthan those in in vitro. This might be explained by a nega-tive eVect of SLs on the CMV promoter. The strongcytomegalovirus promoter is often used in expressionplasmids for transient transfection experiments and inviral vectors for infection experiments. Its enhancerregion contains many cis-acting elements and its activityis stimulated by several cellular transcription factors,such as NF-�B/rel, AP-1, and CREB/ATF [33,34]. Wehave previously demonstrated that activation of thetranscription factors NF-�B and AP-1 is inhibited bySLs [20]. Therefore, an inhibition of these transcriptionfactors could lead to a lower CMV promoter activityand subsequently to detection of lower reporter enzymeactivities.

As �-galactosidase appeared to be a more reliablereporter than luciferase for use with SLs, we used thisreporter gene assay and showed that both SLs preventthe induction of �B-dependent gene expression in a con-centration-dependent manner (Figs. 5A and B). The datacorrelated with those obtained in the NF-�B EMSA, in

which 4�,15-epoxy-miller-9E-enolide also preventedNF-�B DNA binding at a lower concentration than par-thenolide [35].

In summary, we provided evidence that SLs directlyinterfere with luciferase while �-galactosidase remainsunaVected after treatment, suggesting that for alkylatingreagents, such as SLs, �-galactosidase rather than lucif-erase is the reporter enzyme of choice to investigate theinXuence on the transcriptional activity and to avoidmisinterpretation of results. Data obtained with SLs inthe luciferase reporter gene assay should carefully beconsidered and, if used, should be used only for qualita-tive and not quantitative analysis. Thus, IC50 valuesdeduced from experiments with SLs in the luciferasereporter gene assay should be avoided [36]. Togetherwith previous reports our studies underline the necessityto include a control that monitors the direct eVect of theexperimental conditions on the reporter enzyme beforeusing it in a reporter gene assay.

Acknowledgments

We gratefully acknowledge the Wnancial support ofthe Volkswagen-Stiftung. MTL acknowledges the Lan-desgraduiertenförderung Baden-Württemberg for ascholarship. AGP thanks funding by the Deutscher Aka-demischer Austauschdienst and the Baden-WürttembergStipendium.

References

[1] J.I. Lee, G.J. Burckart, Nuclear factor kappa B: important tran-scription factor and therapeutic target, J. Clin. Pharmacol. 38(1998) 981–993.

[2] H. Zhong, R.E. Voll, S. Ghosh, Phosphorylation of NF-kappaBp65 by PKA stimulates transcriptional activity by promoting anovel bivalent interaction with the coactivator CBP/p300, Mol.Cell 1 (1998) 661–671.

[3] M. Karin, Y. Ben Neriah, Phosphorylation meets ubiquitination:the control of NF-kappaB activity, Annu. Rev. Immunol. 18(2000) 621–663.

[4] L.F. Chen, W.C. Greene, Regulation of distinct biological activi-ties of the NF-kappaB transcription factor complex by acetyla-tion, J. Mol. Med. 81 (2003) 549–557.

[5] P. Bremner, M. Heinrich, Natural products as targeted modula-tors of the nuclear factor-kappaB pathway, J. Pharm. Pharmacol.54 (2002) 453–472.

[6] F. D’Acquisto, M.J. May, S. Ghosh, Inhibition of nuclear factorkappa B (NF-�B): an emerging theme in anti-inXammatory thera-pies, Mol. Interv. 2 (2002) 22–35.

[7] I. Bronstein, J. Fortin, P.E. Stanley, G.S. Stewart, L.J. Kricka,Chemiluminescent and bioluminescent reporter gene assays, Anal.Biochem. 219 (1994) 169–181.

[8] E. Conti, N.P. Franks, P. Brick, Crystal structure of WreXy lucifer-ase throws light on a superfamily of adenylate-forming enzymes,Structure 4 (1996) 287–298.

[9] R.E. Huber, M.N. Gupta, S.K. Khare, The active site and mecha-nism of the beta-galactosidase from Escherichia coli, Int. J. Bio-chem. 26 (1994) 309–318.

154 M.T. Lindenmeyer et al. / Analytical Biochemistry 328 (2004) 147–154

[10] R.H. Jacobson, X.J. Zhang, R.F. DuBose, B.W. Matthews, Three-dimensional structure of beta-galactosidase from E. coli, Nature369 (1994) 761–766.

[11] A. Nichtl, J. Buchner, R. Jaenicke, R. Rudolph, T. Scheibel, Fold-ing and association of beta-galactosidase, J. Mol. Biol. 282 (1998)1083–1091.

[12] V.T. Nguyen, M. Morange, O. Bensaude, Protein denaturationduring heat shock and related stress. Escherichia coli beta-galacto-sidase and Photinus pyralis luciferase inactivation in mouse cells, J.Biol. Chem. 264 (1989) 10487–10492.

[13] B.J. Deroo, T.K. Archer, Proteasome inhibitors reduce luciferaseand beta-galactosidase activity in tissue culture cells, J. Biol.Chem. 277 (2002) 20120–20123.

[14] R. Lee, W.D. McElroy, Role and reactivity of sulfhydryl groups inWreXy luciferase, Biochemistry 8 (1969) 130–136.

[15] Y. Ohmiya, F.I. Tsuji, Mutagenesis of WreXy luciferase shows thatcysteine residues are not required for bioluminescence activity,FEBS Lett. 404 (1997) 115–117.

[16] N.J. Roth, B. Rob, R.E. Huber, His-357 of beta-galactosidase(Escherichia coli) interacts with the C3 hydroxyl in the transitionstate and helps to mediate catalysis, Biochemistry 37 (1998)10099–10107.

[17] N.P. Franks, A. Jenkins, E. Conti, W.R. Lieb, P. Brick, Structuralbasis for the inhibition of WreXy luciferase by a general anesthetic,Biophys. J. 75 (1998) 2205–2211.

[18] S. Rocha, K.J. Campbell, K.C. Roche, N.D. Perkins, The p53-inhibitor piWthrin-alpha inhibits WreXy luciferase activity in vivoand in vitro, BMC Mol. Biol. 4 (2003) 9.

[19] T.J. Schmidt, Toxic activities of sesquiterpene lactones: structuraland biochemical aspects, Curr. Org. Chem. 3 (1999) 577–608.

[20] A.J. Garcia-Pineres, V. Castro, G. Mora, T.J. Schmidt, E. Strunck,H.L. Pahl, I. Merfort, Cysteine 38 in p65/NF-kappaB plays a cru-cial role in DNA binding inhibition by sesquiterpene lactones, J.Biol. Chem. 276 (2001) 39713–39720.

[21] G. Lyss, T.J. Schmidt, I. Merfort, H.L. Pahl, Helenalin, an anti-inXammatory sesquiterpene lactone from Arnica, selectively inhibitstranscription factor NF-kappaB, Biol. Chem. 378 (1997) 951–961.

[22] P.M. Bork, M.L. Schmitz, M. Kuhnt, C. Escher, M. Heinrich, Ses-quiterpene lactone containing Mexican Indian medicinal plantsand pure sesquiterpene lactones as potent inhibitors of transcrip-tion factor NF-kappaB, FEBS Lett. 402 (1997) 85–90.

[23] B.H. Kwok, B. Koh, M.I. Ndubuisi, M. Elofsson, C.M. Crews, Theanti-inXammatory natural product parthenolide from the medici-nal herb feverfew directly binds to and inhibits IkappaB kinase,Chem. Biol. 8 (2001) 759–766.

[24] M. Li-Weber, M. Giaisi, M.K. Treiber, P.H. Krammer, The anti-inXammatory sesquiterpene lactone parthenolide suppresses IL-4gene expression in peripheral blood T, Eur. J. Immunol. 32 (2002)3587–3597.

[25] R.L. Mazor, I.Y. Menendez, M.A. Ryan, M.A. Fiedler, H.R.Wong, Sesquiterpene lactones are potent inhibitors of interleukin8 gene expression in cultured human respiratory epithelium, Cyto-kine 12 (2000) 239–245.

[26] V. Castro, P. Rungeler, R. Murillo, E. Hernandez, G. Mora, H.L.Pahl, I. Merfort, Study of sesquiterpene lactones from MilleriaquinqueXora on their anti-inXammatory activity using the tran-scription factor NF-kappa B as molecular target, Phytochemistry53 (2000) 257–263.

[27] S.P. Hehner, T.G. Hofmann, W. Droge, M.L. Schmitz, The antiin-Xammatory sesquiterpene lactone parthenolide inhibits NF-kappaB by targeting the I kappa B kinase complex, J. Immunol. 163(1999) 5617–5623.

[28] J. Altschmied, J. Duschl, Set of optimized luciferase reporter geneplasmids compatible with widely used CAT vectors, Biotech-niques 23 (1997) 436–438.

[29] C.T. Baldari, M.M. Di Somma, M.B. Majolini, C. Ulivieri, E.Milia, J.L. Telford, NF-AT-luciferase reporter T cell lines as toolsto screen immunosuppressive drugs, Biologicals 26 (1998) 1–5.

[30] K. Fukuda, Y. Hibiya, M. Mutoh, Y. Ohno, K. Yamashita, S.Akao, H. Fujiwara, Inhibition by parthenolide of phorbolesterinduced transcriptional activation of inducible nitric oxidesynthase gene in a human monocyte cell line THP-1, Biochem.Pharmacol. 60 (2000) 595–600.

[31] G.M. Leclerc, F.R. Boockfor, W.J. Faught, L.S. Frawley, Develop-ment of a destabilized WreXy luciferase enzyme for measurementof gene expression, Biotechniques 29 (2000) 590–596, 598.

[32] F. Naider, Z. Bohak, J. Yariv, Reversible alkylation of a methionylresidue near the active site of -galactosidase, Biochemistry 11(1972) 3202–3208.

[33] J.L. Meier, J.A. Pruessner, The human cytomegalovirus majorimmediate-early distal enhancer region is required for eYcientviral replication and immediate-early gene expression, J. Virol. 74(2000) 1602–1613.

[34] M.F. Stinski, Cytomegalovirus promotor for expression in mam-malian cells, in: J.M. Fernandez, J. Fernandez, J.P. HoeZer(Eds.), Gene Expression Systems: Using Nature for the Art ofExpression, Academic Press, San Diego, USA, 1999, pp. 211–233.

[35] P. Rüngeler, V. Castro, G. Mora, N. Goren, W. Vichnewski, H.L.Pahl, I. Merfort, T.J. Schmidt, Inhibition of transcription fac-tor NF-kappaB by sesquiterpene lactones: a proposed molecu-lar mechanism of action, Bioorg. Med. Chem. 7 (1999) 2343–2352.

[36] F. Hilmi, J. Gertsch, P. Bremner, S. Valovic, M. Heinrich, O. Sti-cher, J. Heilmann, Cytotoxic versus anti-inXammatory eVects inHeLa, Jurkat T and human peripheral blood cells caused by guai-anolide-type sesquiterpene lactones, Bioorg. Med. Chem. 11(2003) 3659–3663.