Eur. J. Biochem. 258, 768�774 (1998) FEBS 1998

Molecular cloning and bacterial expression of a general odorant-binding proteinfrom the cabbage armyworm Mamestra brassicaeMartine MAIBECHE-COISNE1, Sonia LONGHI2, Emmanuelle JACQUIN-JOLY 1, Carole BRUNEL2, Marie-Pierre EGLOFF2,Louis GASTINEL2, Christian CAMBILLAU2, Mariella TEGONI2 and Patricia NAGNAN-LE MEILLOUR11 INRA, Unite de Phytopharmacie et des Mediateurs Chimiques, Versailles, France2 CNRS, Laboratoire Architecture et Fonction des Macromolecules Biologiques, Marseille, France

(Received 3 July 1998) � EJB 98 0853/3

A cDNA clone encoding a general odorant-binding protein (GOBP2) was isolated from antennal RNAof Mamestra brassicae by reverse transcription-PCR (RT-PCR) and RACE-PCR. The cDNA encodingthe GOBP2 was further used for bacterial expression. Most of the recombinant GOBP2 (�90%) wasfound to be insoluble. Purification under denaturing conditions consisted of solubilisation of inclusionbodies, affinity chromatography, refolding and gel filtration. The refolded rGOBP2 was cross-reactivewith a serum raised against the GOBP2 of the Lepidoptera Antheraea polyphemus. The purified refoldedrGOBP2 was further characterised by native PAGE, IEF, N-terminal sequencing, and two-dimensionalNMR. A functional characterisation of the rGOBP2 was carried out by testing its ability to bind phero-mone compounds. The yields of production and purification fulfil the requirements of structural studies.Keywords: olfaction; general odorant-binding proteins; Mamestra brassicae; protein expression; ligandbinding.

Odour reception in insects is mediated by specific olfactory-receptor neurons (ORNs) housed in different types of antennalsensilla. The transfer of the airborne hydrophobic odorant mole-cules through the fluid sensillar lymph bathing the dendrites ofORNs is ensured by odorant-binding proteins (OBPs) [1, 2].These small proteins (molecular mass ~16 kDa) are synthesisedby accessory cells and accumulated to high concentrations in thesensillar lymph [3]. In Lepidoptera, three groups were defined,the pheromone-binding proteins (PBPs) [4] and the generalodorant-binding proteins (GOBPs) [5, 6], which were dividedinto two groups, GOBP1 and GOBP2, on amino acid-sequencecriteria. The PBPs are predominantly expressed in male anten-nae and constitute a heterogeneous group in which the similaritybetween and within species is low [6�9]. The PBPs are locatedin the pheromone-sensitive sensilla, the sensilla trichodea, whileGOBPs are located in the sensilla basiconica, tuned to the detec-tion of plant volatiles [10�12]. Evidence of specific bindingwith pheromone compounds has clearly been shown for thePBPs [4, 7, 13�16]. The GOBP2s are able to bind a large rangeof odorants, including pheromone components, with a broadspecificity [16�18].

Correspondence to P. Nagnan-Le Meillour, INRA, Unite de Phyto-pharmacie et des Mediateurs Chimiques, Route de Saint-Cyr, F-78026Versailles Cedex, France

Fax: �33 1 30833119.E-mail: nagnan@versailles.inra.frAbbreviations. ORNs, olfactory-receptor neurons; PBP, pheromone-

binding protein; GOBP, general odorant-binding protein; OBP, odorant-binding protein; Z11�16:Ac, cis-11-hexadecenyl acetate ; Z11�16:OH,cis-11-hexadecenol; 16:Ac, hexadecanyl acetate; PVDF, poly(vinyl-difluoride) ; DTNB, 5,5ʹ′-dithiobis(2-nitrobenzoic acid).

Note. The novel nucleotide sequence presented here has been sub-mitted to the GenBank database and is available under accession numberAF051144. Both S. Longhi and M. Maibeche-Coisne contributed equallyto the work described in this paper.

Molecular-cloning approaches have provided the completeor partial primary structure of several moth OBPs [5, 8, 19�23]which, in parallel with biochemical studies, showed that thereexists great diversity within this multigene family (for review,see [24]). The cDNAs encoding the PBPs of Antheraea pernyi[25], A. polyphemus [26] and the GOBP2 of Manduca sexta[18] have been expressed in heterologous systems. Recombinantproteins were produced with features similar to the native pro-teins, including the ability to bind pheromone compounds [18].

In the cabbage armyworm Mamestra brassicae, three maleand two female PBPs have been purified by means of reverse-phase HPLC and identified by their N-terminal sequences [9].One GOBP2 was also found in both sexes. The male-purifiedPBPs (MbraPBP1, MbraPBP1ʹ′ and MbraPBP2) were used inbinding experiments with the tritiated analogue of the majorpheromonal compound, cis-11-hexadecenyl acetate (Z11�16:Ac), and were shown to have opposite binding affinities [16].Except for the binding of the major compound of the phero-mone, ligand-binding properties of the GOBP2 have not yetbeen elucidated. Recently, Feng and Prestwich [18] describedphotoaffinity-labelling experiments between a recombinantGOBP2 and leaf odorants. These experiments support the hy-pothesis of a broad specificity of the GOBP2 class, contrary tothe high specificity of the PBP class. However, the proper li-gands of the GOBPs have still to be characterised. As substantialquantities of proteins are required for structural and extensivebinding studies, we isolated and sequenced the gene encodingthe GOBP2 in M. brassicae and we report herein its bacterialexpression.

EXPERIMENTAL PROCEDURES

Insects. Animals were reared in Domaine du Magneraud(INRA, France) on a semi-artificial diet [27] at 20°C, 60% rela-

769Maibeche-Coisne et al. (Eur. J. Biochem. 258)

tive humidity, exposed to a 16-h light :8-h dark photoperiod andsexed as pupae. Antennae from 3-day-old male and femaleadults were dissected and stored at �80°C until use.RNA extraction and RT. Total RNA was extracted from

200 antennae of both sexes by the RNAzol B method (Wak-Chemie). Single-stranded cDNA was synthesised from 1 µg totalRNA with M-MLV reverse transcriptase (USB), using the bufferand protocol supplied with the enzyme. The reaction containeda dNTP mix (Pharmacia), RNasin (Promega), OligodT 12�18(Pharmacia), sterile water and template RNA to a final volumeof 50 µl. The mix was heated at 68°C for 5 min and chilled onice before adding the enzyme (500 U) and incubating for 1 h at37°C.PCR amplification. PCR amplification was carried out on

a fraction (1/25) of the RT mixture in a solution containing Taqreaction buffer (50 mM KCl, 10 mM Tris/HCl, pH 9.0, 0.1%Triton X-100), 1.5 mM MgCl2, 0.2 mM of each dNTP, Taq poly-merase (1 U) (Promega), sense GA(G/A)GTGATGAGCCA(T/C)GT(C/T)AC and anti-sense TTGAAGCA(G/A)GG(T/A)-GC(C/A)AC(C/T)TT primers (25 pmol of each) to a final vol-ume of 25 µl. Primers were synthesised by Isoprim S. A. (Sable-sur-Sarthe). The sense primer was synthesised according to theGOBP2 N-terminal sequence obtained by gas-phase microse-quencing [9]. The anti-sense primer was deduced from consen-sus sequences of GOBP cDNAs from other lepidopteran speciesretrieved from GenBank. The reaction cycles were performed asfollows : 94°C for 5 min, followed by 35 cycles of 30 s at 94°C,40 s at 40°C and 30 s at 72°C, and finally 5 min at 72°C. The3ʹ′ end of the MbraGOBP2 cDNA was obtained using a RACEprocedure [28]. Amplification was performed under the sameconditions as above, with the specific sense primer and an oli-go(dT)20 as anti-sense primer.

The PCR-amplified cDNAs were gel purified (GenElute, Su-pelco) and ligated into the plasmid vector pCRTM II using theTA Cloning Kit from Invitrogen. Recombinant clones were di-gested with EcoRI (Biolabs) to screen for the presence of insertDNA. Positive clones were subjected to automated sequencing(ESGS).Southern-blot and Northern-blot analysis. Partial Mbra-

GOBP2 cDNA (RT-MbraGOBP2) was used as a probe forSouthern-blot and Northern-blot analysis. The probe was la-belled by random priming using 32P dATP with the Rediprimekit (Amersham).

Genomic DNA was isolated from males and females of M.brassicae according to Ausubel et al. [29]. Aliquots containing10 µg genomic DNA were digested with EcoRI and the resultingfragments were separated by means of electrophoresis in a 1%agarose gel. The gel was pretreated by soaking in 0.25 M HClbefore the denaturation (1.5 M NaCl, 0.5 M NaOH) and neutrali-sation (1.5 M NaCl, 0.5 M Tris/HCl, pH 7.0) steps. DNA wasthen transferred to a Hybond-N membrane (Amersham) using20�SSC (Na3Citrate, pH 7.1, NaCl 3 M). The membrane wasbaked for 2 h at 80°C before prehybridisation at 65°C in6�SSC, 0.5% SDS, 5�Denhardt’s solution and 100 µg/ml soni-cated salmon-sperm DNA. Hybridisation was performed over-night at 65°C in the prehybridisation solution with the labelledprobe. The filter was washed at 65°C for 20 min in 2�SSC,0.5% SDS, then for 20 min in 2�SSC, 0.1% SDS and2�20 min in 0.1�SSC, 0.5% SDS at 65°C (high stringency).Autoradiography was performed at �80°C using Biomax MSKodak films (Amersham).

Total RNA from male and female antennae (10 µg of each)were separated on a 0.8% formaldehyde gel and transferred ontoa Hybond-N membrane [30] which was then baked at 80°C for2 h. The filter was prehybridised for 2 h at 42°C in 45% form-amide without probe, then hybridised overnight at 42°C in the

same solution with the labelled probe. Following hybridisation,the blot was washed as described in the case of the Southern-blot procedure and autoradiography was performed at�80°C as above.PCR primers for synthesis of MbraGOBP2 expression

cassette and PCR amplification. The two missing codons atthe 5ʹ′ end of the MbraGOBP2 gene, as well as 5ʹ′ BamHI and3ʹ′ HindIII restriction sites, were introduced by PCR using thefollowing primers (Genosys) : sense primer GGCC GGA TCCACT GCA GAA GTG ATG AGC CAT GTC; anti-sense primerGGCC AAG CTT TCA GTA CTT CTC CAT GAC GGC TTC.The floating parts of the primers are underlined and the BamHIand HindIII sites are in bold.

The primers (15 pmol each) were mixed with 20 ng templateDNA (pCRTM II plasmid with RACE-MbraGOBP2 insert), 2 µleach of 10 mM dNTPs (Perkin-Elmer), 1 µl Pfu polymerase(2.5 U/µl, Stratagene), 5 µl of 10�Pfu buffer (Stratagene) andwater up to a final volume of 50 µl. The reaction cycles werecarried out as follows: 94°C for 5 min, followed by 25 cyclesof 1min at 94°C, 1 min at 45°C and 1min at 72°C, and finally8 min at 72°C. The 438-bp product (MbraGOBP2) was gel puri-fied (QIAEX II Gel Extraction Kit, Qiagen).Subcloning of MbraGOBP2 in expression vectors. The

438-bp MbraGOBP2 fragment was digested with BamHI andHindIII (Biolabs), ethanol precipitated, then ligated into theBamHI and HindIII sites of the pQE30 (Qiagen) expression vec-tor. The ligation mixture was used to transform Escherichia coliXL1 Blue (Stratagene) cells by the CaCl2 method [27]. Trans-formants were selected onto Luria-Bertani-medium�ampicillinagar plates. Plasmids extracted from randomly selected trans-formant clones were screened for the presence of the insert byPCR, using the expression-cassette primers. Positive candidateswere further characterised by restriction-map analysis. Auto-mated DNA sequencing (ESGS) was used to confirm the pres-ence of the MbraGOBP2 gene and to check the PCR fidelity.The resulting construct is referred to as pQE30/GOBP2.Expression in E. coli. Cell culture and induction conditions.

Expression was carried out in E. coli strain M15 [pREP4] (Qia-gen). Cultures of M15 [pREP4], transformed by either the paren-tal or the recombinant pQE30 plasmid, were grown in both Lu-ria-Bertani medium and minimal M9 media supplemented withampicillin and kanamycin. Cultures were induced by the addi-tion of 0.5 mM isopropylthio-β-D-galactoside. Before induction,cultures were grown at 37°C to an A600 of either 0.6�0.8 or0.8�1.0 depending on the temperature at which the inductionwas carried out : 37°C (Luria-Bertani medium) or 27°C (mini-mal M9 medium).Preparation of crude cellular extracts. For analytical

studies, culture aliquots (A600� 1) were taken at different timeintervals (0, 1, 3, 5 and 16�20 h) after the induction and ana-lysed for protein production. The cellular pellets were resus-pended in 100 µl 50 mM Tris/HCl, 2 mM EDTA, pH 8, 1 mMphenylmethylsulfonyl fluoride and sonicated (three cycles, 10 seach) at 4°C. The total fraction, soluble fraction and the insolu-ble pellet were analysed by means of SDS/PAGE.

For large-scale protein production, 0.5�1 l induced cultureswere harvested at the stationary phase. The cellular pellets wereresuspended in lysis buffer (50 mM NaPO4, pH 8, 300 mMNaCl, 10 mM imidazole, 8 M urea, 1 mM phenylmethylsulfonylfluoride with a ratio of 5 ml/g wet cells). Solubilisation of theinclusion bodies was done by stirring (2 h) at room temperature.Cellular debris were removed by centrifugation (30000�g,30 min at room temperature).Affinity chromatography. Large-scale denatured prepara-

tions of crude cellular extracts, obtained from induced culturesof M15 [pREP4] [pQE30/GOBP2] were purified by affinity

770 Maibeche-Coisne et al. (Eur. J. Biochem. 258)

Fig. 1. Primary structure of MbraGOBP2 and alignment with full-length sequences of other lepidopteran GOBP2s. (A) Nucleotide anddeduced amino acid sequence of RACE-MbraGOBP2 clone. The inferred amino acids are written below the corresponding codons and begin withthe first codon of the sense-primer. The asterisk marks the translation-termination codon and the polyadenylation signal is underlined. (B) Alignmentof mature MbraGOBP2 with deduced amino acid sequences of GOBP2s from different species of Lepidoptera. The two first uncoded amino acidsobtained by microsequencing of M. brassicae GOBP2 are underlined and the six conserved cysteines are marked in bold. MsexGOBP2 [21] fromM. sexta ; HvirGOBP2 [22] from H. virescens ; AperGOBP2 [5] from A. pernyi ; BmorGOBP2 [23] from B. mori.

chromatography (Fast Flow Ni-NTA resin, Qiagen). The resin,which selectively binds His-tagged proteins, was pre-equili-brated with 12 vol. of lysis buffer. The binding was carried outusing an ‘in-batch’ procedure. The samples were incubated for16�20 h with gentle agitation at 20°C. After washing with lysisbuffer (5 vol.), the elution was carried out using 8 vol. of elutionbuffer (50 mMNaPO4, pH 8, 300 mMNaCl, 250 mM imidazole,8 M urea, 1mM phenylmethylsulfonyl fluoride). Fractions (1�2 ml each) were collected and analysed on SDS/PAGE. Frac-tions containing the rGOBP2 were pooled and further subjectedto the refolding procedure.Protein refolding. Affinity-purified denatured rGOBP2 was

refolded by carrying out a cystine�cysteine redox protocol [31,32] essentially as described by Prestwich [26]. The basic ap-proach consisted of, first, ensuring that the protein was com-pletely reduced and denatured and, second, inducing folding bydiluting in the presence of cystine�cysteine, which should allowdisulfide-bridge formation. 2-mercaptoethanol (10 mM final)was added to the purified denatured rGOBP2 in 8 M urea; themixture was stirred for 1 h at room temperature. Cystine in0.5 M NaOH was added to the final concentration of 16 mMand stirred for 10 min at room temperature. The mixture wasthen diluted into 10 vol. 100 mM Tris, pH 8, containing 5 mMcysteine. The reaction was allowed to proceed at room temper-ature for 48 h.Gel filtration. The renaturation mixture was concentrated

(Centriprep-10, Amicon) and dialysed against 10 mM Tris/HCl,

pH 8.5. The sample was then applied onto a FPLC gel-filtrationcolumn (Superdex S200 Prep Grade Column, Pharmacia) equili-brated with a 10 mM Tris/HCl, pH 8.5, 50 mMNaCl buffer. Elu-tion was carried out in the same buffer. Fractions (1.25 ml each)were analysed by SDS/PAGE.Electrophoresis and Western blotting. Proteins were sepa-

rated by 15% SDS/PAGE [33], and 16.8% native PAGE [9].Proteins were visualised by staining with Coomassie Blue[9]. Crude cellular extracts obtained from stationary-phasecultures and purified refolded rGOBP2 were separated onboth 16.8% SDS/PAGE and native PAGE, and electroblottedto a poly(vinyldifluoride) (PVDF) membrane (Immobilon P,Millipore) for immunodetection according to the proceduredescribed by Jacquin-Joly and Descoins [34] with a 1/5000 and1/10000 dilution of primary and secondary antibodies, respec-tively.Binding with pheromone analogues. Samples (10 µg) of

purified refolded rGOBP2 were incubated with either [3H]-cis-11-hexadecenol ([3H]-Z11�16:OH) (22.2 kBq) or [3H]-hexade-canyl acetate ([3H]-16:Ac) (37 kBq) or [3H]-Z11�16:Ac(42.5 kBq) for a period of 30 min on ice. The incubation mix-tures were then separated by 16.8% native PAGE andelectroblotted onto PVDF membranes (Immobilon P, Millipore).Autoradiography was performed over 7 days at �80°C usingHyperfilm MP (Amersham).N-terminal sequencing. The N-terminal sequences of M.

brassicae GOBP2 and the refolded purified rGOBP2 were ob-

771Maibeche-Coisne et al. (Eur. J. Biochem. 258)

Fig. 2. Analysis of DNA and RNA coding for MbraGOBP2. (A)Southern-hybridisation analysis of M. brassicae male and female geno-mic DNAs, digested with EcoRI. Each lane was loaded with 10 µg geno-mic DNA and the filters were hybridised with the 32P-radiolabelled RT-MbraGOBP2 probe. (B) Northern-blot analysis of total RNA (10 µg)from male and female antennae using the 32P-radiolabelled RT-Mbra-GOBP2 probe.

tained by Guy Bezard (INRA, Tours) with an automated proteinsequencer using a standard protocol of Edman degradation.Analytical isoelectric focusing. Analytical IEF was carried

out using the separation PhastSystem (Pharmacia) with a Phast-Gel IEF in a pH range of 4�6.5. Proteins were visualised byCoomassie Blue or silver staining according to the supplier’sinstructions.Free-cysteine measurement. Free cysteines were titrated

with DTNB [5,5ʹ′-dithiobis(2-nitrobenzoic acid)] according tothe procedure described in Habeeb [35]. Free cysteines weremeasured in 200 mM Tris base, pH 8, in the presence of 1%SDS, without incubation. As a control, the same solution ofDTNB was used, under identical experimental conditions, to tit-rate a protein (gastric lipase) that had one free cysteine per mole-cule. The number of cysteines reacting with DTNB was deter-mined on the basis of the increase in absorbance at 412 nm,using the absorption coefficient of 13.6 mM�1 cm�1.Two-dimensional NMR. A sample containing purified re-

folded rGOBP2 at a concentration of 0.85 mM in a 10-mMNaPO4, pH 6.5, was used for the acquisition of a NOESYspectrum on a DRX500 Brücker spectrometer at 300 K with2048 t2 and 512 t1 data points. Solvent suppression wasachieved by the WATERGATE sequence [36]. The data wereprocessed using UXNMR; they were multiplied by a sine-squared bell and zero-filled to 1 K in ω1 dimension prior toFourier transformation.

RESULTSCharacterisation of the GOBP2 clone. The nucleotide and de-duced amino acid sequences of the RACE-MbraGOBP2 cloneare shown in Fig. 1A. The microsequencing of the purifiedMbraGOBP2 [9] showed that the two first codons encoding thetwo first N-terminal amino acids (Thr�Ala) were missing in theRACE-MbraGOBP2 clone. The mature GOBP2 protein consistsof 141 amino acids with a calculated molecular mass of15928 Da and an isoelectric point of 4.78, in agreement withthe molecular mass obtained by means of matrix-assisted laserdesorption/ionisation�time-of-flight MS [37] (15920�20 Da)and with the mobility of M. brassicae antennal proteins in two-dimensional electrophoresis [16]. The deduced amino acid se-quence of M. brassicae GOBP2 was aligned with GOBP2 se-quences from other species of Lepidoptera (Fig. 1B). Themultiple sequence alignment pointed out the conservation of six

Fig. 3. Bacterial expression of recombinant M. brassicae GOBP2 instationary-phase cultures of E. coli strain M15 [pREP4]. (A) 15%SDS/PAGE. Lanes 1�6: crude cellular extracts obtained from culturesgrown and induced at 37°C in Luria-Bertani medium. Lanes 7�9: crudecellular extracts obtained from cultures grown in M9 minimal mediumand induced at 27°C. Lane 1 : total fraction of non-induced cells trans-formed by pQE30. Lane 2: total fraction of induced cells transformedby pQE30. Lane 3: total fraction of non-induced cells transformed bypQE30/GOBP2. Lanes 4 and 7: total fractions of induced cells trans-formed by pQE30/GOBP2. Lanes 5 and 8: Soluble fractions of inducedcells transformed by pQE30/GOBP2. Lanes 6 and 9: insoluble fractionsof induced cells transformed by pQE30/GOBP2. MW: low molecular-mass markers (Pharmacia) (kDa). (B) Western-blot with the anti-GOBP2serum after SDS/PAGE (30 s exposure). Lanes 1�4: total fractions ofsamples 1�4 from A.

cysteines in all insect OBPs and, as described previously [21�23], a high sequence similarity (between 76.5% and 96.4% ofsequence identity) among different species. M. brassicaeGOBP2 shows a particularly high amino acid identity to theGOBP2 from Heliothis virescens (96.4%). The distribution ofthe hydrophobic domains is typical of moth GOBP2s with amajor hydrophobic domain in the region spanning the residues40�60, which is supposed to be involved in the binding of odor-ant ligands [14].

Southern-blot and Northern-blot analysis. The results of theSouthern-hybridisation analysis of genomic DNA from male andfemale M. brassicae with the RT-MbraGOBP2 probe are shownin Fig. 2A. No significant differences were observed betweenthe hybridisation patterns of males and females, indicating thatthe gene is not sex linked. The tissue specificity of the proteinMbraGOBP2 was previously determined by Western-blot ex-periments [16]. Northern-blot analysis was carried out to com-pare expression in male and female animals (Fig. 2B). The in-tensities of hybridisation signals were approximately the same inthe male and female tracks, suggesting equal levels of GOBP2-specific mRNA in each. This result is in agreement with whathas been shown with the GOBP2 of Bombyx mori [23], but con-tradicts the result obtained with the GOBP2 of A. pernyi, inwhich the gene product was female specific [5]. As long as thecDNA sequence obtained does not correspond to the full-lengthmRNA encoding GOBP2, the band hybridising with GOBP2probe seems to be larger (500�700 bp). This full-length mRNAsize is in accordance with the GOBP2 mRNA size obtained inA. pernyi [5], M. sexta [21] and B. mori [23].

Expression of GOBP2. In order to obtain the entire codingMbraGOBP2 cDNA, the two missing codons at the 5ʹ′end of theRACE-MbraGOBP2 cDNA were introduced by PCR. AfterPCR amplification, gel purification, digestion and precipitation,the MbraGOBP2 cDNA was subcloned into the pQE30 vector,which is specifically designed for the overexpression of heterol-ogous proteins in E. coli. The pQE30 vector allows the expres-

772 Maibeche-Coisne et al. (Eur. J. Biochem. 258)

Fig. 4. Analysis of a large-scale production and purification ofrGOBP2 from induced stationary-phase cultures of M15 [pREP4][pQE30/GOBP2]. (A) 15% SDS/PAGE. Lane 1 : crude denatured bacte-rial lysate in 8 M urea. Lane 2: eluate from affinity chromatography.Lane 3: purified rGOBP2 as obtained upon renaturation and gel filtra-tion. (B) Western blot with the anti-GOBP2 serum after SDS/PAGE ofthe eluate from affinity chromatography (Lane 2 of A), 30-s exposure.

Fig. 5. Characterisation of the purified refolded rGOBP2. (A) NativePAGE of crude protein extract from 15 M. brassicae female antennaeand 3.5 µg purified refolded rGOBP2. The arrow indicates the positionof the natural GOBP2 in antennal extracts. (B) Western blot with anti-GOBP2 serum of purified refolded rGOBP2 (3.5 µg) after native PAGE,30-s exposure. (C) Autoradiography of refolded rGOBP2 (10 µg) incu-bated with [3H]-Z11�16:Ac (42.5 kBq), [3H]-Z11�16:OH (22.2 kBq),and [3H]-16:Ac (37 kBq) after separation by native PAGE andelectroblotting. Six-day exposure time.

sion of a recombinant product, bearing in its N-terminus a shortand uncleavable His-tag sequence.

The SDS/PAGE analysis of M15 [pREP4] cultures trans-formed with pQE30/GOBP2 revealed the production of a proteinof the expected molecular mass (17 kDa), which was detectedneither in cells transformed by the parental vector nor in non-induced cells (Fig. 3A). The identity of the 17-kDa band wasconfirmed by Western blotting with an anti-GOBP2 serum(Fig. 3B). A fairly high expression level was observed(~150 mg/l), but more than 90% of the recombinant GOBP2was found to be insoluble (Fig. 3A). In order to improve thesolubility of the rGOBP2, various culture conditions were tested,in which different parameters (such as the E. coli strain, theisopropylthio-β-D-galactoside concentration, the induction tem-perature and the culture medium) were changed (data notshown). The best results (~20% solubility) were obtained withcultures grown in minimal M9 medium and by lowering theinduction temperature from 37°C to 27°C (Fig. 3A). However,the amounts of purified soluble rGOBP2 obtained from large-scale cultures grown under these conditions were judged to be

Fig. 6. Two-dimensional NMR. NOESY spectrum of a 0.85-mM solu-tion of purified refolded rGOBP2 at 300 K.

insufficient in view of structural studies. Accordingly, purifica-tion under denaturing conditions, followed by a refolding pro-cedure was undertaken.

Purification and characterisation of rGOBP2. InsolublerGOBP2 was obtained from 0.5-l to 1-l stationary-phase culturesof M15 [pREP4], transformed by pQE30/GOBP2 grown in Lu-ria-Bertani medium at 37°C and induced by 0.5 mM isopropyl-thio-β-D-galactoside. After solubilisation of the inclusion bodies,the denatured recombinant protein was purified by affinity chro-matography, followed by a refolding procedure and gel filtration.The eluate from the affinity chromatography and the final puri-fied refolded rGOBP2 were analysed by SDS/PAGE (Fig. 4A).The final purified product migrates as a single band (Fig. 4A,lane 3) contrary to the denatured affinity-purified product(Fig. 4A, lane 2), which migrates as two major bands labelledby the anti-GOBP2 serum (Fig. 4B).

After native-PAGE separation, the final purified product wasshown to consist of several protein bands (Fig. 5A). These bandsare characterised by slightly different charges and cross-reactedwith the anti-GOBP2 serum (Fig. 5B). The additional 12-aminoacid sequence in the N-terminus modified the pI of the recombi-nant protein (5.69 calculated instead of 4.78 for the native pro-tein) ; accordingly, in native PAGE, the rGOBP2 migrates in aless acidic region with respect to native GOBP2 (Fig. 5A).

A functional characterisation was carried out by testing theability of the final recombinant product to bind three pheromone

773Maibeche-Coisne et al. (Eur. J. Biochem. 258)

analogues, namely the [3H]-Z11�16:Ac, the [3H]-Z11�16:OHand the [3H]-16:Ac. After native-PAGE separation, the differentprotein bands were all labelled by the three ligands, thus indicat-ing that all the isoforms were functional (Fig. 5C). The rGOBP2was finally purified by reverse-phase HPLC as described in [9]and only one band was obtained in both native and SDS/PAGE,which was labelled by the pheromone analogues in bindingassays (data not shown). The final product possesses the ex-pected N-terminus MRGSHHHHHHGSTAE.

The integrity of the refolded rGOBP2 was checked by two-dimensional NMR. The NOESY spectrum (Fig. 6) showed thatthe protein was structured. The NH chemical shifts ranged from7 ppm to 10.5 ppm and the HA chemical shifts ranged from ap-proximately 4 ppm to 4.8 ppm, thus suggesting that the proteinis folded. Besides these indications, the free-cysteine measure-ment pointed out that the six cysteines were all involved in theformation of disulphide bridges, thus definitely allowing us toconclude that a successful refolding was achieved.

DISCUSSION

Using a combination of RT-PCR and RACE-PCR techniqueson antennal RNA from M. brassicae, an almost completeGOBP2 cDNA clone was obtained. The entire coding cDNAsequence was obtained after a further PCR-amplification step.The deduced amino acid sequence shows features allowing theclassification of the corresponding protein, referred to as Mbra-GOBP2, into the subfamily of the insect GOBP2s, according tothe nomenclature proposed in [23].

As the amounts of M. brassicae GOBP2 obtained after puri-fication from antennal extracts were low [9], we undertook theproduction of a recombinant protein for functional and structuralcharacterisation. Recombinant proteins were previously obtainedin the cases of the two PBPs of A. pernyi [25] and the PBP ofA. polyphemus [26], as well as in the case of M. sexta GOBP2[18]. In the case of the recombinant GOBP2 of M. brassicae,typical yields of 15�20 mg/l of induced stationary-phase E. colicultures have been obtained.

The recombinant GOBP2 of M. brassicae showed the prop-erties of the mature GOBP2 isolated from moth antennae, in-cluding the apparent molecular mass and the N-terminal se-quence, and was recognised by a specific anti-GOBP2 serum.The micro-heterogeneity observed in the final recombinant prod-uct after native PAGE might be ascribed to small charge differ-ences due to post-traductional modifications of the product, e.g.limited proteolytic events occurring at the C-terminal region,and/or to conformational differences resulting from the refoldingprocedure. More probably, the occurrence of multiple proteinforms in the final purified product could be accounted for bynon-enzymatic deamidation events of Asn and Gln residues.Deamidation is a frequent reaction in proteins and peptides oc-curring either as an artefact during purification, sequencing andon storage of proteins, or as a physiological event related toprotein ageing [38]. Asn-X sequences, particularly where X isGly, Ser or Ala, undergo β-aspartyl-shift reactions to yield bothAsp-X and isoAsp-X peptide products [39]. The possibility thatsuch post-translational modifications might be responsible forthe micro-heterogeneity observed in the rGOBP2 is supportedby two considerations: (1) a labile Asn residue (Asn-Gly) occursat position 93 of the rGOBP2 and (2) the relaxation of confor-mational constraints through denaturation by use of chaotropicagents has been shown to favour deamidation events [40, 41].Direct evidence of these eventual chemical modifications areexpected to be provided when structural data become available.It is noteworthy to point out that this micro-heterogeneity could

also have occurred in the case of other recombinant GOBPs andPBPs without being noticed, since the authors did not use thehighly resolutive 17% native PAGE in their analytical pro-cedure.

The presence of the additional N-terminal sequence contain-ing the His tag, slightly modifies the isoelectric point of therGOBP2, but does not affect the functional property of ligandbinding. This was a prerequisite in view of further functionalstudies aimed to identify the proper ligand of this protein.

The proteins belonging to the GOBP2 class share a highsequence similarity and are supposed to bind a wide range ofodorants with a broad specificity. Accordingly, they are ex-pressed in the sensilla basiconica, the sensory cells which re-spond to the so-called general odorants [10�12]. To date, noevidence of binding specificity has been obtained for this classof proteins; only non-specific binding with pheromonal com-pounds [4, 16, 17] and plant odours [18] has been observed. Thenumber of PBPs and GOBPs characterised in Lepidoptera [9,11, 23, 42, 43] suggests a high specificity for the odorants, espe-cially between PBPs and pheromone compounds. Furthermore,the recent discovery of a novel form of OBP in B. mori [23],related to the putative OBP from Drosophila melanogaster [44],suggests that the Lepidoptera OBP family might include morethan the three sub-families previously characterised. The factthat no specific ligand has yet been identified, does not implythat the GOBP2 class has no specificity, but more likely that theproper ligand is still unknown.

In the absence of specific ligands, it is not possible to putforward a model of ‘general olfaction’ for plant volatiles in thesame detail as proposed for ‘specific olfaction’ for the phero-mones. The fine discrimination between specific odours (phero-mones) and general odours (volatiles of other chemical struc-tures) has to be reconsidered. Indeed, in M. brassicae, the 2-phenylethanol is specifically emitted by males before mating[44], but it is also a common plant odour. This raises the ques-tion of whether it is detected (and recognised) as a specific oras a general odour. In such a case, the classification of com-pounds into specific or general odours appears simplistic. Itwould seem more correct to first distinguish between intra-spe-cific and inter-specific odours, and take into account both doseand temporal effects.

The availability of great amounts of recombinant GOBP2fromM. brassicae was a prerequisite for structural studies aimedto determine the three-dimensional structure of this protein, asit has been done in the case of the bovine OBP [45, 46]. Prelimi-nary crystallisation tests and the production of 15N-GOBP2 forNMR studies are in progress.

We are grateful to Dr David Tepfer for constant encouragement andsuggestions and to Dr Michael Burnet for helpful discussion and forcritical reading of this manuscript. Special thanks to Dr Gunde Ziegel-berger, who kindly gave the ApolGOBP2 antiserum and to JonathanBohbot for performing the binding experiments. The authors wish tothank Olivier Bornet for acquiring the NOESY spectra. Research wassupported by INRA and CNRS.

REFERENCES1. Kaissling, K. E. (1986) Chemo-electrical transduction in insect ol-

factory receptors, Annu. Rev. Neurosci. 9, 121�145.2. Van den Berg, M. J. & Ziegelberger, G. (1991) On the function of

the pheromone-binding protein in the olfactory hairs of Antheraeapolyphemus, J. Insect Physiol. 37, 79�85.

3. Steinbrecht, R. A., Ozaki, M. & Ziegelberger, G. (1992) Immuno-cytochemical localization of pheromone-binding protein in mothantennae, Cell Tissue Res. 270, 287�302.

4. Vogt, R. G. & Riddiford, L. M. (1981) Pheromone binding and inac-tivation by moth antennae, Nature 293, 161�163.

774 Maibeche-Coisne et al. (Eur. J. Biochem. 258)

5. Breer, H., Krieger, J. & Raming, K. (1990) A novel class of binding-protein in the antennae of the Silkmoth Antheraea pernyi, InsectBiochem. 20, 735�740.

6. Vogt, R. G., Prestwich, G. D. & Lerner, M. R. (1991) Odorant-binding protein subfamilies associate with distinct classes of ol-factory receptor neurons in insects, J. Neurobiol. 22, 74�84.

7. Vogt, R. G., Köhne, A. C., Dubnau, J. T. & Prestwich, G. D. (1989)Expression of pheromone-binding proteins during antennal devel-opment in the gypsy moth Lymantria dispar, J. Neurosci. 9,3332�3346.

8. Krieger, J., Raming, K. & Breer, H. (1991) Cloning of genomic andcomplementary DNA encoding insect PBP: evidence for microdi-versity, Biochem. Biophys. Acta 1088, 277�284.

9. Nagnan-Le Meillour, P., Huet, J. C., Maıbeche, M., Pernollet, J.C. & Descoins, C. (1996) Purification and characterization ofmultiple forms of odorant/pheromone binding proteins in the an-tennae of Mamestra brassicae (Noctuidae), Insect Biochem. Mol.Biol. 26, 59�67.

10. Laue, M., Steinbrecht, R. A. & Ziegelberger, G. (1994) Immuno-cytochemical localization of general odorant-binding proteins inolfactory sensilla of the silkmoth Antheraea polyphemus, Natur-wissenschaften 81, 178�181.

11. Steinbrecht, R. A., Laue, M. & Ziegelberger, G. (1995) Immunolo-calization of pheromone-binding protein and general odorant-binding protein in olfactory sensilla of the silk moths Antheraeaand Bombyx, Cell Tissue Res. 282, 203�217.

12. Steinbrecht, R. A., Laue, M., Maida, R. & Ziegelberger, G. (1996)Odorant-binding proteins and their role in the detection of plantodours, Entomol. Exp. Appl. 80, 15�18.

13. Maida, R., Steinbrecht, A., Ziegelberger, G. & Pelosi, P. (1993) Thepheromone binding protein of Bombyx mori : purification, charac-terization and immunocytochemical localization, Insect Biochem.Mol. Biol. 23, 243�253.

14. Du, G. & Prestwich, G. D. (1995) Protein structure encodes theligand binding specificity in pheromone-binding proteins, Bio-chemistry 34, 8726�8732.

15. Feixas, J., Prestwich, G. D. & Guerrero, A. (1995) Ligand specific-ity of pheromone binding proteins of the processionary moth, Eur.J. Biochem. 234, 521�526.

16. Maıbeche-Coisne, M., Sobrio, F., Delaunay, T., Lettere, M., Du-broca, J., Jacquin-Joly, E. & Nagnan-Le Meillour, P. (1997) Pher-omone-binding proteins of the moth Mamestra brassicae : speci-ficity of ligand binding, Insect Biochem. Mol. Biol. 27, 213�221.

17. Ziegelberger, G. (1995) Redox-shift of the pheromone binding pro-tein in the silkmoth Antheraea polyphemus, Eur. J. Biochem. 232,706�711.

18. Feng, L. & Prestwich, G. D. (1997) Expression and characterizationof a lepidopteran general odorant binding protein, Insect Biochem.Mol. Biol. 27, 405�412.

19. Györgyi, T. K., Roby-Shemkovitz, A. J. & Lerner, M. R. (1988)Characterization and cDNA cloning of the pheromone-bindingprotein from the tabacco hornworm, Manduca sexta: a tissue-specific developmentally regulated protein, Proc. Natl Acad. Sci.USA 85, 9851�9855.

20. Raming, K., Krieger, J. & Breer, H. (1990) Primary structure of apheromone-binding protein from Antheraea pernyi : homologieswith other ligand-carrying proteins, Comp. Physiol. B. 160, 503�509.

21. Vogt, R. G., Rybczynski, R. & Lerner, M. R. (1991) Molecular clon-ing and sequencing of general odorant-binding proteins GOBP1and GOBP2 from the tabacco hawk moth Manduca sexta : com-parison with other insect OBPs and their signal peptides, J. Neu-rosci. 11, 2972�2984.

22. Krieger, J., Ganssle, H., Raming, K. & Breer, H. (1993) Odorant-binding protein of Heliothis virescens, Insect Biochem. Mol. Biol.23, 449�456.

23. Krieger, J., von Nickisch-Rosenegk, E., Mameli, M., Pelosi, P. &Breer, H. (1996) Binding proteins from the antennae of Bombyxmori, Insect Biochem. Mol. Biol. 26, 297�307.

24. Pelosi, P. & Maida, R. (1995) Odorant-binding proteins in insects,Comp. Biochem. Physiol. 111B, 503�514.

25. Krieger, J., Raming, K., Prestwich, D., Frith, D., Stabel, S. & Breer,H. (1992) Expression of a pheromone-binding protein in insectcells using a baculovirus vector, Eur. J. Biochem. 203, 161�166.

26. Prestwich, G. D. (1993) Bacterial expression and photoaffinity la-beling of a pheromone binding protein, Protein Sci. 2, 420�428.

27. Poitout, S. & Bues, R. (1974) Elevage de 28 especes de Noctuidaeet de 2 especes d’Arctiidae sur milieu artificiel simplifie. Particu-larites selon les especes, Ann. Zool. Ecol. Anim. 6, 431�441.

28. Frohman, M. A., Dush, M. K. & Martin, G. R. (1988) Rapid produc-tion of full length cDNAs from rare transcripts : amplificationusing a single gene specific oligonucleotide primer, Proc. NatlAcad. Sci. USA 85, 8998�9002.

29. Ausubel, F. M., Brent, R., Kingston, R. F., Moore, D. D., Seidmann,J. G., Smith, J. A. & Struhl, K. (1987) Current protocols in molec-ular biology, Greene & Wiley, New York.

30. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular clon-ing: a laboratory manual, Cold Spring Harbor Laboratory, ColdSpring Harbor NY.

31. De Bernardez-Clark, E. & Georgiou, G. (1991) Inclusion bodies andrecovery of proteins from the aggregated state, in Protein refold-ing, pp. 1�20, American Chemical Society, Washington, D. C.

32. Seely, R. J. & Young, M. D. (1991) Large-scale refolding of secre-tory leukocyte protease inhibitor, in Protein refolding, pp. 206�216, American Chemical Society, Washington, D. C.

33. Laemmli, U. K. (1970) Cleavage of structural proteins during as-sembly of the head of the bacteriophage T4, Nature 227, 680�685.

34. Jacquin-Joly, E. & Descoins, C. (1996) Identification of PBAN-likepeptides in the brain-subesophageal ganglion complex of Lepi-doptera using Western-blotting, Insect Biochem. Mol. Biol. 26,209�216.

35. Habeeb, A. F. S. A. (1972) Reaction of protein sulphydryl groupswith Ellman’s reagent, Methods Enzymol. 25, 457�464.

36. Piotto, M., Saudek, V. & Sklenär, V. (1992) Gradient-tailored excita-tion for single-quantum NMR spectroscopy of aqueous solutions,J. Biomol. NMR 2, 661�665.

37. Blais, J. C., Nagnan-Le Meillour, P., Bolbach, G. & Tabet, J. C.(1996) MALDI-TOFMS identification of odorant-binding pro-teins electroblotted onto poly(vinylidene difluoride) membranes,Rapid Commun. Mass Spectrom. 10, 1�4.

38. Di Donato, A., Ciardiello, M. A., de Nigris, M., Piccoli, R., Mazza-rella, L. & D’Alessio G. (1993) Selective deamidation of ribo-nuclease A, J. Biol. Chem. 268, 4745�4751.

39. Wright, H. T. (1991) Sequence and structure determinants of thenonenzymatic deamidation of asparagine and glutamine residuesin proteins, Protein Eng. 4, 283�294.

40. Wearne, S. J. & Creighton, T. E. (1989) Effect of protein conforma-tion on rate of deamidation: ribonuclease A, Proteins: Struct.Funct. Genet. 5, 8�12.

41. Chazin, W. J., Kördel, J., Thulin, E., Hofmann, T., Drakenberg, T. &Forsen, S. (1989) Identification of an isoaspartyl linkage formedupon deamidation of bovine caldbindin D9k and structural char-acterisation by 2D 1H NMR, Biochemistry 28, 8646�8653.

42. Maida, R., Proebstl, T. & Laue, M. (1997) Heterogeneity of odorant-binding proteins in the antennae of Bombyx mori, Chem. Senses22, 503�515.

43. Pikielny, C. W., Hasan, G., Rouyer, F. & Rosbach, M. (1994) Mem-bers of a family of Drosophila putative odorant-binding proteinsare expressed in different subsets of olfactory hairs, Neuron 12,35�49.

44. Jacquin, E., Nagnan, P. & Frerot, B. (1991) Identification of hairpen-cil secretion from male Mamestra brassicae (L.) (Lepidoptera:Noctuidae) and electroantennogram studies, J. Chem. Ecol. 17,239�246.

45. Tegoni, M., Ramoni, R., Bignetti, E., Spinelli, S. & Cambillau, C.(1996) Domain swapping creates a third putative combining sitein bovine odorant binding protein dimer, Nat. Struct. Biol. 3,863�867.

46. Bianchet, M., Bains, G., Pelosi, P., Pevsner, J., Snyder, S., Monaco,H. & Amzel, M. (1996) The three-dimensional structure of bovineodorant-binding protein and its mechanism of odor recognition,Nat. Struct. Biol. 3, 934�939.