Revisiting the Specificity of Mamestra brassicaeand Antheraea polyphemus Pheromone-binding Proteins with a Fluorescence Binding Assay*

Pheromone-binding proteins (PBPs), located in the sensillum lymph of pheromone-responsive antennal hairs, are thought to transport the hydrophobic pheromones to the chemosensory membranes of olfactory neurons. It is currently unclear what role PBPs may play in the recognition and discrimination of species-specific pheromones. We have investigated the binding properties and specificity of PBPs fromMamestra brassicae (MbraPBP1), Antheraea polyphemus (ApolPBP1), Bombyx mori (BmorPBP), and a hexa-mutant of MbraPBP1 (Mbra1-M6), mutated at residues of the internal cavity to mimic that of BmorPBP, using the fluorescence probe 1-aminoanthracene (AMA). AMA binds to MbraPBP1 and ApolPBP1, however, no binding was observed with either BmorPBP or Mbra1-M6. The latter result indicates that relatively limited modifications to the PBP cavity actually interfere with AMA binding, suggesting that AMA binds in the internal cavity. Several pheromones are able to displace AMA from the MbraPBP1- and ApolPBP1-binding sites, without, however, any evidence of specificity for their physiologically relevant pheromones. Moreover, some fatty acids are also able to compete with AMA binding. These findings bring into doubt the currently held belief that all PBPs are specifically tuned to distinct pheromonal compounds.

In the context of molecular recognition, the perception of pheromones by male moths is a system that shows both high specificity and significant affinity, resulting in mating within the same species and with the response being elicited by relatively low concentrations of pheromone. In a generally recognized scheme, the hydrophobic pheromone molecules enter the antennal sensilla (sensory hairs) through pores in the cuticle and traverse the aqueous sensillum lymph transported by low molecular mass (15)(16)(17) pheromone-binding proteins (PBPs), 1 finally reaching receptors located in the dendritic membranes of the olfactory neurons, where they are recognized either free or in association with the PBP (1)(2)(3)(4)(5).
Although molecular cloning has rapidly increased the information available on moth PBPs primary structures, cloned genes have not greatly increased our knowledge of their binding affinities and specificities. Heterologously expressed PBPs from Antheraea polyphemus (ApolPBP1), Antheraea pernyi (AperPBP1 and -2), and Lymantria dispar (LdispPBP1) have already been used to develop quantitative binding assays with photoactivable derivatives of the pheromone or radioactive pheromones (8,(21)(22)(23)(24). However, synthesis of photoactivable or radioactive pheromones is difficult and limits the candidate compounds tested, usually to the main components of pheromone blends identified in behavioral studies.
In the present study, we have examined by fluorescence the binding affinity of 1-aminoanthracene (AMA) to recombinant PBPs of A. polyphemus (ApolPBP1), Mamestra brassicae (MbraPBP1), Bombyx mori PBP (BmorPBP), and a mutant of MbraPBP1 (Mbra1-M6) mutated at residues in the internal cavity to mimic the cavity found in B. mori PBP. The results show that AMA binds with high affinity to ApolPBP1 and MbraPBP1, but neither BmorPBP nor Mbra1-M6. The comparison of the binding behavior of MbraPBP1 and Mbra1-M6 indicates that AMA binds in the internal cavity of PBPs. The fluorescence of bound AMA was used to follow and quantify the binding of several pheromonal compounds, taking advantage of the ability of pheromones to displace AMA from the binding site. The dissociation constants estimated demonstrate a high affinity of MbraPBP1 and ApolPBP1 both for pheromones and fatty acids. The binding data reveal that the interaction of pheromones with binding sites of the two PBPs examined by AMA fluorescence shows little specificity.

Pheromones and Analogues
Pheromones and analogues were purchased from Chemtech B.V. (The Netherlands) and Sigma, and stored as specified by the manufacturers. AMA was from Fluka. All alcoholic solutions were freshly prepared.

Subcloning in pET22b(ϩ) Vector and Expression
Escherichia coli strain XL1-Blue was used for DNA subcloning and propagation of the recombinant plasmid. The MbraPBP1 gene was amplified by PCR using pQE30/PBP1 as template (25) with the following primers: Mbra1-MscI 5Ј-CGAGTAAAGAACTGATCACG-3Ј; Mbra1-NotI 5Ј-CCGGGCGGCCGCCTACACGGCCGTCATGATCTC-3Ј. The amplified PCR fragment was purified and digested with MscI and NotI before being cloned between the same restriction sites of the pET22b(ϩ) vector (Novagen). Recombinant MbraPBP1 was produced by growing E. coli BL21(DE3) transformed with the recombinant pET22b(ϩ)/ MbraPBP1 plasmid at 37°C in LB medium supplemented with 50 g/ml carbenicillin. Cultures were induced with 0.5 mM isopropyl-1thio-␤-D-galactopyranoside, and the temperature was decreased to 28°C when A 600 reached the value of 0.8.
The gene of ApolPBP1 was amplified, subcloned, and expressed in a similar way. PCR was performed using the primers: ApolPBP1-MscI 5Ј-CTTCGCCAGAGATCATGAAGAAT-3Ј and ApolPBP1-XhoI 5Ј-AGA-CACTCGAGCAATCTAAACTTCAGCTA-3Ј. An ApolPBP1 cDNA clone (Apo-3 (7)) was used as template. The fragment obtained by PCR was purified, digested by XhoI, and cloned into the MscI and XhoI restriction sites of the pET22b(ϩ) vector. The recombinant pET22b(ϩ)/Apol-PBP1 plasmid was transformed into E. coli BL21(DE3) and expression was performed following the protocol of Wojtasek and Leal (26) by growing the cells without induction at 28°C in LB medium supplemented with 50 g/ml carbenicillin until A 600 reaches a value Ͼ2.5. The gene of BmorPBP was amplified, subcloned, and expressed as described by Wojtasek and Leal (26).

Site-directed Mutagenesis of MbraPBP1
Six residues of MbraPBP1, Ile-5, Met-8, Ala-56, Leu-61, Ile-62, and Met-68 were replaced by Met, Leu, Ser, Met, Leu, and Leu, respectively, by PCR (Fig. 1). 5Ј and 3Ј regions of the cDNA of MbraPBP1 were amplified separately using pET22b(ϩ)/MbraPBP1 as template and the following mutated primers (mutations are underlined, NruI site is in italic and encoded amino acids are indicated between parentheses). 5Ј After amplification and purification, the PCR fragments were used to reconstruct a single and continuous cDNA using M6-NruI and Mbra1-NotI primers. After purification and digestion by NruI and NotI, the mutated cDNA was cloned between the MscI and NotI restriction sites of the pET22b(ϩ) vector (as described above) and the final construct was referred as pET22b(ϩ)/Mbra1-M6. The mutant, Mbra1-M6, was produced by growing E. coli BL21(DE3) transformed by the recombinant pET22b(ϩ)/Mbra1-M6 plasmid at 37°C in LB medium supplemented with 50 g/ml carbenicillin. When A 600 reached the value of 0.8, the cultures were induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside, and the temperature was decreased to 28°C.

Purification Procedure
The purification of recombinant MbraPBP1 and Mbra1-M6 mutant was initiated by centrifugation harvesting of 0.5-4 liters of induced cultures after 20 -24 h of induction. The periplasmic proteins were released by osmotic shock as described in the pET system manual.
Periplasmic proteins, obtained from induced cultures of BL21(DE3) (pET22b(ϩ)/MbraPBP1) were submitted to a 100% ammonium sulfate precipitation, dialyzed overnight against 50 mM Tris-HCl, 50 mM NaCl, pH 8.0, and purified by anion exchange on a ResourceQ column (Amersham Pharmacia Biotech, Ä kta FPLC) pre-equilibrated in the same buffer. Elution was carried out with a linear gradient of 0 -1 M NaCl in 50 mM Tris, pH 8.0. Fractions (1 ml each) were collected and analyzed by SDS-PAGE. Fractions containing the MbraPBP1 were pooled and concentrated by a 10-kDa Microsep (Filtron), and loaded onto a Superdex 200 gel filtration column (Amersham Pharmacia Biotech) pre-equilibrated with 50 mM Tris, 150 mM NaCl, pH 8.0; the elution was carried out at 0.25 ml/min. Purification of Mbra1-M6 followed the same procedure, with the exception of the ammonium sulfate precipitation step. The periplasmic fraction was directly dialyzed overnight against 50 mM Tris-HCl, 25 mM NaCl, pH 8.0, before purification by anion exchange on a ResourceQ column followed by gel filtration (see above). A second anion exchange on a ResourceQ column was performed with a linear gradient of 0 -1 M NaCl in 10 mM Tris, pH 8.0. Fractions (0.5 ml each) were analyzed by SDS-PAGE. Both proteins were shown to be Ͼ95% pure in SDS and native PAGE. The integrity of the hexamutant was checked by mass spectrometry and N-terminal sequencing.
Purification of the recombinant ApolPBP1 and BmorPBP proteins was performed by adjusting the periplasmic fraction to 10 mM Tris, pH 8.0, and loading it onto a 20-ml DEAE HR16/10 column (Tyopearl 650S, TOSOH). Proteins were eluted with a linear gradient of 0 -300 mM NaCl in 10 mM Tris, pH 8.0. The collected fractions (1 ml) were analyzed by SDS-PAGE. Fractions containing PBP were pooled, concentrated in a Centriprep-10 (Amicon), and separated on a Superdex 75 gel filtration column (Amersham Pharmacia Biotech) pre-equilibrated with 10 mM Tris, 150 mM NaCl, pH 8.0. PBP containing fractions (1 ml each) were determined by SDS-PAGE, pooled, and desalted by dialysis against 20 mM Tris, pH 8.0. Purified ApolPBP1 and BmorPBP protein was shown to be Ͼ95% pure in SDS and native PAGE.

Mass Spectrometry and Circular Dichroism Analysis
Mass analysis of recombinant PBPs was performed with a Voyager-DE RP spectrometer (PerSeptive Biosystems). Samples (0.7 l containing 15 pmoles) were mixed with an equal volume of sinapinic acid matrix solution and spotted on the target, then dried at room temperature for 10 min. The mass standard was apomyoglobin.
The CD spectra were measured on a CD6 spectropolarimeter (Jobin Yvon). Spectra were recorded in 10 mM Na phosphate, pH 7.0, at 20°C between 178 and 260 nm, with a 30-s averaging.

Fluorescence Assays
In the case of MbraPBP1, BmorPBP, and Mbra1-M6, the fluorescence spectra were recorded on a Jobin-Yvon FL 3-21 spectrofluorimeter at 20°C using a front face fluorescence accessory. Slit widths of 1 and 10 nm were used for excitation and emission, respectively. The spectra were processed with DataMax software. In the case of Apol-PBP1, the spectra were recorded in a right angle configuration on a PerkinElmer LS 50B spectrofluorimeter using a 1-cm light path fluorimeter quartz cuvette. Slit width of 5 nm were used for both excitation and emission.

Intrinsic Fluorescence of PBPs
The interaction of AMA and bombykol with PBPs was monitored by following the quenching of the intrinsic protein fluorescence (excitation at 280 nm and emission 290 -420 nm, slits as described above). Spectra were recorded with 1 M protein in 20 mM Tris buffer, pH 8.0, 0.3% methanol and under the same conditions in the presence of different concentrations of AMA (0 -10 M).

Fluorescence Emission Spectra Using AMA
The effects of solvents, such as methanol, ethanol, and dimethyl sulfoxide on AMA binding have previously been tested with rat OBP (27). In that study, methanol was demonstrated to be less effective than ethanol in displacing the bound AMA from the binding site. In our study, the displacement of AMA by ethanol and methanol has been determined by successively adding aliquots of solvent to the AMA⅐MbraPBP1 complex solution. As with rat OBP, ethanol was found to compete significantly with AMA bound to MbraPBP1, and methanol was found to have a lesser effect (data not shown). For this reason, methanol was used for dissolving AMA and its competitors in the fluorescence titrations.

MbraPBP1, Mbra1-M6, and BmorPBP
All the fluorescence experiments were carried out in 20 mM Tris, pH 8.0, at 20°C (excitation 298 nm, emission 400 -575 nm). The binding affinity for AMA was titrated by adding to the protein sample (1 M) aliquots of a stock solution of AMA (10 mM) solubilized in 100% methanol. The fluorescence of AMA was recorded after stabilization of the signal (5-6 min). In competition experiments, PBP (1 M) was incubated in the presence of AMA (5 M) for 1 h at room temperature, the fluorescence signal of AMA was then monitored and the decrease in fluorescence intensity upon addition of different test compounds was recorded. The competition experiments were carried out at constant methanol concentration (0.3%) and with a test compound concentration between 0 and 10 M.

ApolPBP1
Binding and competition experiments were carried out in 20 mM Tris, pH 8.0, at 20°C (excitation 256 nm, emission 440 -600 nm). All values reported were obtained from three independent measurements. The curves for binding of AMA to ApolPBP1 were obtained by titration of 2 M protein with increasing concentrations of chromophore dissolved in methanol. In competition assays, we monitored the fluorescence signal of AMA (2 M) equilibrated with ApolPBP1 (2 M) upon addition of increasing amounts of competitor. Fluorescence intensities at the maximum of emission (487 nm) were determined for different concentrations of competitor and were corrected before further data analysis by the extent of AMA fluorescence decrease due to the methanol present in the cuvette.

Data Analysis
The affinity of different compounds for MbraPBP1 and ApolPBP1 was estimated by plotting the decrease of intensity of AMA fluorescence at the emission maximum, calculated as (I Ϫ I min )/(I 0 Ϫ I min ) against the competitor concentration; I 0 is the maximum of fluorescence intensity of the complex AMA⅐MbraPBP1 and AMA⅐ApolPBP1, at 492 and 487 nm, respectively, I is the fluorescence intensity after addition of an aliquot of competitor, and I min the fluorescence intensity at saturating concentration of the competitor. The IC 50 values were estimated on the direct plot by non-linear regression with equation corresponding to a single binding site using Prism 3.02 (GraphPad software, Inc). K diss values were calculated according to K diss ϭ [IC 50 ]/(1 ϩ [AMA]/K AMA ), in which [AMA] ϭ free AMA concentration and K AMA ϭ dissociation constant for MbraPBP1/AMA and ApolPBP1/AMA, respectively.

Design of the Mutant of MbraPBP1
Residues were chosen using two criteria: first, they are localized in the binding pocket defined in the three-dimensional structure of B. mori PBP (17) and second, they are not conserved among the PBPs. Six residues of MbraPBP1 were thus replaced by their counterparts in B. mori PBP (Fig. 1). Three primers were designed to introduce these mutations by PCR. The final DNA sequence, subcloned in pET22b(ϩ), was verified by automated DNA sequencing to check the introduction of mutations and PCR fidelity.

Subcloning, Expression, and Purification
After PCR amplification, gel purification, digestion, and precipitation, the MbraPBP1, hexamutant Mbra1-M6, BmorPBP, and ApolPBP1 cDNAs were subcloned into the pET22b(ϩ) vector. This vector leads to expression of a pelB leader/PBP fusion protein in E. coli, which targets the recombinant proteins from the cytoplasm to the periplasm of the bacteria. Upon translocation, the pelB signal peptide is cut off and the PBP is released into the oxidative environment of the periplasm, favorable to the formation of disulfide bonds. This system has been successfully used for the expression of the pheromone-binding protein of Bombyx mori (26), which contains 3 disulfide bridges (28), and in the case of a chemosensory protein from M. brassicae. 2 PBP1 and Hexamutant of M. brassicae  Fig. 2A, lanes 3 and 4). N terminus sequencing and mass spectrometry confirmed the identity of the recombinant MbraPBP1 and the mutant Mbra1-M6.

PBP1 of A. polyphemus and PBP of B. mori
In the case of ApolPBP1, the expression was slightly better than MbraPBP1 and up to 5 mg of pure PBP per liter of culture could be purified. Two bands at about 16 and 18 kDa (Fig. 2B,  lane 1), likely corresponding to recombinant ApolPBP1 (15,783 Da) and the pelB/ApolPBP1 fusion protein (15,783 Da ϩ 2,228 Da) can be observed in the SDS-PAGE of whole cell lysates of BL21(DE3) cells transformed with the pET22b(ϩ)/ApolPBP1 construct. In contrast, only the 16-kDa band can be seen in the SDS-PAGE of the periplasmic fraction (Fig. 2B, lane 2). This size is identical to that of the purified ApolPBP1 (Fig. 2B, lane  3). The identity of the isolated recombinant ApolPBP1 protein was confirmed by mass spectrometry and Western blot analysis using an antiserum directed against the wild type protein. Results similar to those described in Wojtasek and Leal (26) were obtained for BmorPBP.

Circular Dichroism
Analytical methods such as CD have been used to define structural features of insect PBPs. The CD spectra of the four PBPs were recorded and showed that they are all well folded and have similar secondary structures. The two minima in the spectra around 208 and 222 nm are typical of a fold with a majority of ␣-helical secondary structure, in agreement with

Fluorescence Binding Assays
Fluorescence Emission of AMA Upon Binding to PBPs-When excited at 256 or 298 nm, AMA in aqueous buffer shows a weak fluorescence emission with a maximum at 563 nm. In a hydrophobic environment, such as the binding pocket of MbraPBP1 or ApolPBP1, there is a blue shift of the fluorescence emission maximum and a large increase of intensity (Fig.  3A), resulting from the modification in the AMA environment. The fluorescence emission spectra show a maximum at 492 and 487 nm for MbraPBP1 and ApolPBP1, respectively (Fig. 3,B and C, inset). The concentration dependence of AMA binding to MbraPBP1 and ApolPBP1 can be described by a hyperbolic curve as expected for a one-site binding model (Fig. 3, B and C), and K diss values of 4.5 and 0.95 M, respectively, were calculated. The affinity of ApolPBP1 for AMA is thus greater than that of MbraPBP1.
In the case of BmorPBP, no increase in fluorescence emission and no saturation are observed with increasing AMA concentration. Interestingly, the Mbra1-M6 mutant behaves like BmorPBP. In these cases, the fluorescence increase observed is not significantly greater than the experimental error, furthermore, the lack of saturation suggests either some nonspecific interaction between AMA and the proteins, not related to binding, or the formation of superstructures by AMA.

Intrinsic Fluorescence Quenching of PBPs
AMA-In order to check whether AMA binds within the internal cavity of the two PBPs, we measured the influence of AMA on the PBP intrinsic fluorescence. The intrinsic fluorescence spectra of MbraPBP1 decreases in intensity upon addition of AMA, the same behavior is observed with ApolPBP1 (data not shown). A dissociation constant of 6 Ϯ 1.7 M was obtained from intrinsic fluorescence quenching data for AMA binding to MbraPBP1 (compared with 4.5 M from AMA measurements). The decrease of Trp fluorescence intensity may be due to a ligand induced quenching or to Förster resonance energy transfer (31), between the tryptophan (being the donor) and AMA (the acceptor). Upon excitation at 280 nm, the tryptophan fluorescence is quenched by AMA (acceptor quenched donor emission), however, in the case of Förster resonance energy transfer, the energy is transferred from the protein to AMA, and we should also observe an emission of fluorescence (donor sensitized acceptor emission) when AMA is bound to the protein. The excitation spectrum of bound AMA showed no evidence for tryptophan-sensitized fluorescence, which argues against a Förster resonance energy transfer mechanism, but rather for local environmental effects.
Bombykol-To confirm the proper folding of BmorPBP, we have used bombykol as an intrinsic fluorescent probe, exploiting its conjugated double bond. Bombykol produced an increase in the intrinsic fluorescence of BmorPBP, and saturation was reached after addition of 5 equivalents. A very similar result was observed with Mbra1-M6 (Table I).

AMA Displacement by Pheromones or Fatty Acids
To estimate the binding affinities of MbraPBP1 and Apol-PBP1 for a variety of different ligands, we monitored the decrease of AMA fluorescence resulting from the ability of different pheromones and fatty acids to displace AMA, and determined the K diss values for the different compounds.
For MbraPBP1, the three specific M. brassicae pheromones, Z11-C16-aldehyde, alcohol, and acetate, exhibit very similar AMA displacement properties with K diss values of 0.29, 0.17, and 0.20 M, respectively (Table II). However, these speciesspecific, and physiologically relevant, compounds are less efficient than bombykol, the pheromone specific to the moth B. mori (K diss ϭ 0.13 M; Table II; Fig. 4A). The best AMA competitor was found to be cetyl alcohol (C16-OH), the saturated equivalent of the Z11-C16-OH (K diss ϭ 0.09 M). Fatty acids also bind well to MbraPBP1, especially palmitic acid (C16-COOH) which exhibits a K diss (0.12 M) very close to the K diss of bombykol. The K diss values determined for the compounds tested in competition with AMA binding to MbraPBP1 cluster in a narrow range, between 0.09 and 0.63 M (7-fold), irrespective of their behavioral specificity (Table II).
In the case of ApolPBP1, three types of test ligands were used in AMA competition assays: the three species-specific pheromones, pheromones from other species, and fatty acids. The K diss values observed for all compounds were between 0.48 and 1.36 M (Table II, Fig. 4B). The three components of the pheromonal blend released by A. polyphemus females, E6,Z11-C16-Ald, E6,Z11-C16-Ac, and E4,Z9-C14-Ac, display similar K diss values of 0.50, 0.48, and 0.51 M, respectively (Table II). The two former compounds are the best AMA competitors of this series; however, like MbraPBP1, a low K diss value, indicative for a high affinity to ApolPBP1, has also been determined for bombykol (0.54 M). Indeed this K diss value is very close to that obtained for the components of the specific pheromonal blend. As seen for MbraPBP1, fatty acids are also very efficient in displacing AMA from ApolPBP1; K diss values in this case ranged between 0.56 and 1.36 M, indicating an affinity of fatty acids to ApolPBP1 similar to that of specific pheromones. DISCUSSION Fluorescence of probes like AMA, 1-anilino-8-naphtalenesulfonic acid, or 1,1Ј-bi(4-anilino)naphtalene-5,5Ј-disulfonic acid have already been used to study the affinity constants of ligands to lipocalins, such as rat OBP (32,33), porcine OBP (34), apolipoprotein D, insecticyanin, and bovine ␤-lactoglobulin (35). Although fluorescence spectroscopy has proved its worth in probing the binding of odorous compounds to vertebrate OBPs, this technique has rarely been applied to the study of the interaction between insect PBPs and pheromones. Intrinsic fluorescence and 1-anilino-8-naphtalenesulfonic acid fluorescence have previously been used, however, but only to check for structural changes of B. mori PBP at different pH values and for its binding to bombykol (26,36). In the present study, we successfully reproduced the effect of bombykol on the BmorPBP intrinsic fluorescence. The behavior of Mbra1-M6, designed to mimic BmorPBP, with AMA and bombykol used as fluorescence quenchers was found to be similar to that of the authentic B. mori PBP. The intrinsic fluorescence of MbraPBP1 and ApolPBP1 decreased upon increasing the concentration of AMA, which indicates an interaction between AMA and tryptophans of the protein.
The three-dimensional structure of the B. mori PBP (17) displays a buried cavity surrounded by the 6 helices constituting the PBP fold. The sequences of MbraPBP1 and ApolPBP1 are, respectively, 44 and 67% identical to B. mori PBP (Fig. 1) and show no insertions or deletions. Furthermore, the CD spectra indicate a very similar overall fold for the four PBPs under study. It seems therefore reasonable to speculate on the   three-dimensional structural properties of MbraPBP1 and ApolPBP1 based on those of the B. mori PBP. In such a model, the two conserved tryptophans (at the same position in the three sequences) have been localized relative to the cavity; Trp-37 is buried in the PBP and is an integral part of the cavity wall, whereas Trp-127 is far from the cavity, and not fully exposed to the solvent (Fig. 1). In this context, the decrease in intrinsic fluorescence in the presence of AMA for two PBPs is likely to be mediated by the interaction between the fluorophore and Trp-37 as suggested by the following arguments: first, upon binding to the protein, the maximum emission wavelength of AMA is blue-shifted and the quantum yield is substantially increased, indicating a significant modification of the environment of AMA, which inside the protein becomes protected from the solvent. Second, the intrinsic fluorescence can be quenched only if the tryptophan is close enough to AMA (a few Å) to interact electrostatically, and this excludes Trp-127. In the three-dimensional model, AMA fits well into the cavity, positioned about 5 Å from Trp-37 (Fig. 5). The K diss of AMA for MbraPBP1 was found to be 6 Ϯ 1.7 and 4.5 Ϯ 0.3 M by intrinsic fluorescence and AMA fluorescence, respectively. Finally, the AMA binding assays with MbraPBP1 and Mbra1-M6, where the first PBP binds AMA and its variant not, indicate that mutations in the internal cavity have impeded AMA binding, and, therefore, that AMA very likely binds to the internal cavity in the native protein.
With the aim of determining the specificity of MbraPBP1 and ApolPBP1, competitive binding experiments with several aliphatic ligands were performed. The data indicate that pheromones as well as fatty acids can displace AMA from its binding site. Furthermore, there was no obvious correlation between the presence, the position, and the number of double bonds, and K diss , nor were there any obvious effects of the functional group of the first carbon and the chain length (C14, -16, or 18). These results are in contrast with previous studies suggesting that pheromone discrimination might occur not only at a receptor level, but also in the first molecular step, i.e. binding of pheromones to PBPs (8,(21)(22)(23)(24)(37)(38)(39). Concerning MbraPBP1, migration on native gels and autoradiography indicated that the antennal protein was able to bind the tritiated major pheromonal compound, Z11-C16-Ac (38). However, since this was the only compound tested, the result should be taken with caution. Furthermore, autoradiography studies with the recombinant PBP indicated an affinity not only for the Z11-C16-Ac, but also for the Z11-C16-TFMK (trifluoromethyl ketone), but not for the saturated C16-Ac (25). Indeed, the literature is punctuated with reports misgiving the specificity of PBPs. AperPBP1, expressed in insect cells, has been shown to bind three different compounds, E6,Z11-C16-acetate, -alcohol, and -aldehyde (22), while the PBP of A. polyphemus has been shown to bind compounds other than the specific pheromone, namely decylthio-trifluoropropanone, an inhibitor of pheromone receptor neurons of the moth (40). On this basis it has been suggested that decyl-thio-trifluoropropanone, like pheromones, could be carried through the sensillum lymph by PBP and compete with the pheromone-PBP complex at the receptor.  proposition that two residues, Thr-44 and Asp-32, were part of the pheromones-binding site (41). However, based on the threedimensional structure of B. mori PBP, these residues and the covalently modified regions do not belong to the binding cavity (17).
Our present results raise two important structural questions: 1) does AMA bind to the same internal binding site as pheromones, the site identified in the x-ray structure (17) of BmorPBP? 2) Do the various recombinant PBPs have a threedimensional structure identical to that of the natural antennal PBPs? With regards to question 1, AMA binding to both Mbra and Apol PBPs yields a significant decrease of intrinsic fluorescence (see above), suggesting a close approach between AMA and a tryptophan. Since, as deduced from the PBP structural model, the external surface of both PBPs contains only a few hydrophobic residues and no hydrophobic patches, it is not very likely that AMA binds at the PBPs surface near the second tryptophan (Trp-127). Furthermore, Mbra1-M6, in which only residues located within the cavity have been changed to mimic the cavity of BmorPBP, loses its ability to bind AMA thus resembling BmorPBP. With respect to the second question, due to the lack of a direct comparison of three-dimensional structure or binding data for natural and recombinant PBPs, no direct evidence for the correct folding of recombinant PBPs is available. However, several lines of evidence suggest these proteins fold correctly. (i) Recombinant PBPs are expressed in E. coli periplasm, in conditions allowing disulfide bridge formation; (ii) no post-translational modifications have been identified in natural PBPs, therefore bacterial expression should be adequate for obtaining a native fold; (iii) the CD spectra are all very similar and are in agreement with the secondary structure predictions; (iv) BmorPBP generated by the same procedure resulted in crystalline PBP leading to the three-dimensional structure and binding bombykol (26).

CONCLUDING REMARKS
Based on a fluorophore displacement assay, two recombinant PBPs are shown not to display significant discriminatory capacity between different pheromones, while two others do not bind the fluorophore AMA, precluding displacement studies. The observed lack of discrimination brings into question the hypothesis that PBP subtypes may be specifically tuned to distinct species-specific pheromones. One way to reconcile our data with previous results, is to postulate that among the several PBPs found in a single moth species (often three different PBPs), some might be nonspecific while others are more specific. Since AMA is a bulky compound, very different from the insect pheromone structures, it might be possible that PBPs able to bind AMA are rather nonspecific. Among the three PBPs studied here, MbraPBP1 and ApolPBP1 would be nonspecific, while BmorPBP might be specific. Indeed current studies indicate that some other recombinant PBPs bind AMA and others do not. While the existence of several PBPs, each specific for a distinct component of the pheromonal blend, would be conceivable in a molecular and evolutionary context, the co-existence of several nonspecific PBP subtypes retained throughout evolution is a challenging question. Alternatively, one could imagine that compounds may be readily bound by the proteins without being able to activate them, thus reflecting the situation of receptor proteins which are able to interact with a large variety of pharmacological compounds, some of which activate the receptor (agonists) whereas other do not activate (antagonists), although the latter often display a higher binding affinity.