Chem. Senses 28: 479-489,
2003
© Oxford University Press 2003
Structure–Activity Studies with Pheromone-binding Proteins of the Gypsy Moth, Lymantria dispar
? Department of Chemistry, Simon Fraser University, 8888 University Drive, Burnaby BC, Canada V5A 1S6 1 US Department of Agriculture, Chemicals Affecting Insect Behavior Laboratory, Beltsville, MD 20705 2 Department of Medicinal Chemistry, The University of Utah, 30 South 2000 East, Room 307, Salt Lake City, UT 84112–5820, USA
Correspondence to be sent to: Erika Plettner, Department of Chemistry, Simon Fraser University, 8888 University Drive, Burnaby BC, Canada V5A 1S6. e-mail: plettner{at}sfu.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
Pheromone olfaction in the gypsy moth, Lymantria dispar, involves accurate distinction of compounds with similar structure and polarity. The identified sex pheromone is (7R,8S)-2-methyl-7,8-epoxyoctadecane, 1a, and a known antagonist is (7Z)-2-methyloctadec-7-ene, 4a. The first step in pheromone olfaction is binding of odorants by small, soluble pheromone-binding proteins (PBPs), found in the pheromone-sensing hairs. We have studied the molecular determinants recognized by the two PBPs found in the gypsy moth, using three pheromone/PBP binding assays. Results indicate that (i) PBPs bind analogs of the pheromone with some discrimination; (ii) PBPs experience enhancement of binding when presented with 1a or its enantiomer and 4a simultaneously; and (iii) the binding enhancement is also seen at high ligand:PBP ratios. We found no evidence of allostery, so the synergistic binding effects and the concentration effect may only be explained by multimerization of PBPs with each other, which leads to more than one population of binding sites. We suggest that the enhanced ligand binding at high ligand:PBP ratios may serve to sequester excess ligand and thereby attenuate very strong signals.
Key words: binding assay, fluorescence, insect, odorant, olfaction, signal attenuation, synergy, transport
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
In many species of moth, pheromones play a central role in reproduction. The female releases the pheromone, and the males follow the pheromone plumes upwind. The main sex pheromone component of the gypsy moth (Lymantria dispar) is (7R,8S)-2-methyl-7,8-epoxyoctadecane 1a (Miller et al., 1977
). The antipode 1b is believed to be a major component
of the nun moth, Lymantria monacha, pheromone blend
(Hansen, 1984). However, while
1b strongly antagonizes the effect of 1a in the gypsy moth, it
has no effect on nun moth behavior (Grant
et al., 1996). The alkene 4a is a strong
antagonist in the gypsy moth (Gries et
al., 1996), but a synergist in the nun moth
(Grant et al., 1996).
The nun moth also has the unbranched analogs of the epoxide 1c and the
alkene 4b as synergistic components. Again, in the gypsy moth the
unbranched compounds are antagonistic
(Gries et al., 1996).
Thus, even though both species strongly respond to 1a, the additional
components of the nun moth blend are believed to be responsible for
reproductive isolation between these two species. Because pheromones are
gaining more widespread use as monitoring and control agents
(Wyatt, 1997), it is of great
interest to understand the molecular basis of the pheromone olfactory system
(Plettner, 2002).
The pheromone olfactory system in male moths is housed in antennal sensory
hairs (sensilla trichodea). Two or three olfactory neurons project their
dendrites into the hollow space of one hair
(Kaissling and Torson, 1980;
Keil, 1982). The dendrites are
bathed by sensillar lymph, an aqueous solution which contains
pheromone-binding protein (PBP) (Vogt and
Riddiford, 1981). The pheromone enters the sensilla through pores
in the cuticle, and the abundant PBPs transport the hydrophobic pheromone
through the aqueous lymph to the dendritic membrane
(Vogt et al., 1985).
There, the odorants interact with G-protein-coupled transmembrane receptors,
which initiate a nerve potential through several signal transduction steps
(Krieger and Breer, 1999). The
pheromone sensory hairs have two remarkable properties. (i) They exhibit a
very high selectivity towards particular pheromone blend components or towards
antagonists from related species
(Mustaparta, 1997). (ii) They
have a very broad range in sensitivity, from 
1.6 x 10-17
M/s (Kaissling and Priesner,
1970) to 5 x 10-9 M/s
(Kaissling, 1977) of pheromone
in the plume (see supplementary material). At this point it is not known how
much each of the molecular components outlined above contributes to the
selectivity and sensitivity of an olfactory hair.
The PBPs are not the only proteins that function in the transport of
hydrophobic ligands through aqueous media. The general odorant binding
proteins (GOBPs) (Maibeche-Coisne et
al., 1998; Vogt et
al., 1991), which are related in sequence to the PBPs
(Vogt et al., 1999)
and the chemosensory specific proteins (CSPs), which are in a different
protein family (Briand et al.,
2002) are either associated with chemosensory structures or have
been shown to bind small ligands (Briand et al., 2003;
Mosbach et al.,
2003). It is not clear whether PBPs only act as pheromone
transporters between the cuticle and the dendritic receptors, or whether PBPs
are also scavengers, helping to clear the pheromone from the sensillar
lymph.
Gypsy moth PBPs exhibit 2- to 10-fold discrimination ability towards
structurally related odorants, when only the dissociation constants are
compared (Plettner et al.,
2000; Kowcun et al.,
2001). However, two ligands that bind to a protein with the same
equilibrium constant may be exploiting very different molecular interactions,
which may influence the conformational properties of the protein–ligand
complex. In this paper we have studied a set of pheromone components and
analogs, to address the molecular interactions that contribute to binding
PBP–ligand complex. The functional role of molecular recognition by PBPs
is not known. One possibility may be the exclusion of compounds that are
irrelevant to pheromonal communication, while still providing flexibility to
accommodate a variety of structurally related compounds. Understanding the
structure–function relationships of the PBP–odorant interaction is
an important step towards understanding the function of PBPs.
The gypsy moth has two different pheromone binding proteins, PBP1 and PBP2
(Vogt et al., 1989;
Merritt et al.,
1998). We have addressed the enantiomer selectivity of these two
PBPs, and found that at pH 7.5 and low ionic strength PBP1 prefers 1b
while PBP2 prefers 1a (Plettner
et al., 2000). Because pheromone olfaction has been
studied extensively by electroantennography (EAG) in the gypsy moth, this
species is a good model to elucidate the extent to which the PBPs contribute
to the olfactory selectivity and sensitivity. In this study, we have performed
three groups of experiments.
Experiment I
We addressed the question of PBP contribution to olfactory selectivity by
examining binding of various analogs of epoxides 1a and 1b to
PBPs, using three different pheromone/PBP binding assays. EAG data exist in
the literature for all compounds examined
(Schneider et al.,
1977; Hansen,
1984; Grant et al.,
1996; Dickens et al.,
1997). To assess the importance of the epoxide moiety we have
studied analogs with NH2+ (2a, 2b) and
CX2 (5, 6) (Scheme
1) instead of the epoxide oxygen, or analogs lacking a
three-membered ring between positions 7 and 8 (3a, 3b,
4a, 4b). The protonated aziridines 2a and 2b have
the opposite polarization to the epoxide O in 1a, 1b and
1c: the aziridines have a full positive charge on the N, while the
epoxides have a partial negative charge on the O. The alkenes 4a and
4b have a 
-system instead of the epoxide and, consequently, a
planar geometry around positions 7 and 8. To determine recognition of the
2-methyl group we have compared epoxides and alkenes with (1a,
1b, 4a) and without (1c, 4b) this substituent. In
an effort to understand the ligand binding interactions, we performed homology
modeling based on the crystal structure of Bombyx mori PBP
(Sandler et al.,
2000).
|
Experiment II
Because blend composition is so important in pheromone recognition, we have investigated the binding of binary combinations of 1a or 1b with 2–6 to PBPs.
Experiment III
To gain insights whether PBPs might act as scavengers, we measured and compared ligand binding at low ligand:PBP ratios to binding at high ligand:PBP ratios.
For experimental group I, our results indicate that the ligand binding selectivity of the PBPs studied here does not parallel the selectivity documented in EAG studies: ligands that bind strongly to PBPs do not necessarily evoke strong EAG responses. Importantly, differences in the dissociation constant do not necessarily reflect a lack of ligand discrimination. Our binding and molecular modeling studies suggest that different molecular interactions are exploited by different ligands. For experimental group II we have observed binding synergy between certain combinations of ligands, and for experimental group III, we have observed significantly stronger binding at high ligand:PBP ratios than at low ratios. No cooperativity was observed in these binding studies, so we explain the observed ligand-mediated effects by the existence of multiple binding equilibria between the ligand, PBP monomers and PBP multimers. We propose that the effects observed herein explain the dual role of PBP as both odorant transporter and scavenger.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
Materials
PBPs were expressed in Escherichia coli (with the pHN1+ vector,
supplied by L. Chen and G.L. Verdine, Harvard, via G.D.P.), reconstituted from
inclusion bodies and purified as described elsewhere
(Plettner et al.,
2000). The PBPs had the correct mass by MALDI-TOF MS and the
expected pI. Purification of ditritiated (at C-5 and C-6) 1a and
1b (from G.D.P.) was as described
(Plettner et al.,
2000); ditritiated (at C-7 and C-8) 4a and 4b were
purified by chromatography on 5% AgNO3 silica gel with
heptane:toluene 3:1. Synthesis of the analogs 2, 5 and 6
have been described in (Dickens et
al., 1997). A sample of
(7R,8S)-epoxyoctadecane (1c) was obtained from Ms R.
Gries (SFU).
We used three types of binding assays: (i) with radiolabeled pheromone: (a)
with only radiolabeled ligand and (b) competition assays with radiolabeled
ligand and non-radiolabeled analogs; (ii) with non-radiolabeled analogs by
taking advantage of changes in the fluorescence of tryptophan 37 upon addition
of ligand (Bette et al.,
2002); and (iii) by attaching a dansyl group to the PBP and
monitoring changes in fluorescence upon addition of ligand.
Radioassays
In the radioassays, labeled pheromone was incubated at six different
concentrations between 0.02 and 0.40 nM with 2 µM PBP in 10 mM Tris at pH
7.5. Protein-bound ligand was separated from unbound ligand by passage through
a small column of Bio-Gel P-2 (Bio Rad, CA, 2 kDa exclusion limit). The
protein was eluted from the mini-columns with assay buffer and the unbound
ligand remained behind. The assay and its validation have been described in
detail (Plettner et al.,
2000). Alkene (4a, 4b) Kd values
were determined with 80 µM CHAPS or 13 µM SDS (both 1% of the critical
micelle concentration). Both detergents gave the same result.
Competition binding radioassays
Because the analogs 3–6 were not radiolabeled, we performed
competition binding assays using radiolabeled 1a or 1b as the
reporter ligand. Assays were performed as described previously
(Plettner et al.,
2000), with the following modifications. PBP and non-radiolabeled
analog were first equilibrated for 1 h, then radiolabeled ligand was added and
the solution incubated for an additional hour. At that time, samples were
passed through the P2 mini-columns to separate protein-bound from unbound
ligand. Unless indicated otherwise, competitor concentration was 20 µM.
Competition pH profiles
The pH profiles required 0.8 nM of tritium labeled 1a or 1b and 3 µM cold analog. Buffers with pH varying from 4.0 to 9.0 were of equivalent ionic strength (adjusted with KCl). For the buffer at pH 4.0, a 50 mM glycine–HCl solution was used. For buffers of pH 5.0, 5.5, and 6.0, 50 mM MES–Tris solutions were prepared. For buffers of pH 7.0, 8.0, and 9.0, 50 mM Tris–HCl solutions were prepared.
Fluorescence assays: tryptophan
Titrations were performed as described
(Bette et al., 2002)
in 10 mM Tris, pH 7.5, presaturated with 0.8% ethanol, in a total volume of
1000 µl. The concentration of PBP1 was 1 µM, which gave a total of 1
nmol of PBP in the cuvette. Aliquots of ligand stock solutions in distilled
ethanol (300 µM) were added, such that titrations proceeded in
increments of 0.15 nmol (=0.15 µM concentration increments), until
1–2 nmol had been added. After each addition, the fluorescence emission
was monitored on a PTI (Quantum Master Model QM-1) fluorimeter. Excitation was
at 295 nm and emission was monitored at 340 nm. Ligands 2a, 2b,
4a and 4b caused a decrease in fluorescence upon addition and
ligand 1a caused an increase in fluorescence.
Dansylation of PBP
PBP1 and PBP2 were coupled with dansyl chloride to produce the fluorescent
protein. Six equivalents of DTT were added to 35 µM of PBP (which was in 10
mM ammonium acetate pH 7.5), under argon, and incubated for 1 h on ice. We
have found that under these conditions only one disulfide bond is reduced
(Kowcun et al.,
2001). The protein was then separated from unreacted DTT by
filtration through P2 gel (Bio Rad) and elution with 30 mM potassium phosphate
pH 7.7. The reduced protein was reacted with 18 equivalents of dansyl chloride
and incubated for 20 min on ice. At the pH of the reaction mixture (7.6),
the lysine residues are protonated, but the newly formed thiols are largely
deprotonated and are the strongest nucleophile on the protein. Unreacted
dansyl chloride was removed by filtering twice through the gel column with 100
mM potassium phosphate, pH 5.0. MALDI-TOF MS and thiol titration revealed that
only one of the two thiol groups coupled with a dansyl moiety. A number of
experiments indicate that a cysteine thiol and not an amino group has been
dansylated (see Appendix). Structural details of the dansyl modified PBP will
be published elsewhere.
Dansyl-PBP titrations
Titrations were performed in 10 mM Tris, pH 7.5, in a total volume of 1000 µl. Two series of titrations were done: (i) with 3 µM dansylated PBP and 1a, 1c, 2a, 2b, 4a and 4b and (ii) with 1 µM dansylated PBP and 1a, 1c, 4a and 4b. Stock solutions of the ligands were prepared in distilled ethanol and were in the range of 300 µM. Aliquots were added to the cuvette such that titrations proceeded in 0.1 µM increments. Excitation was at 340 nm and emission was at 506 nm. Fluorescence intensity increased significantly upon addition of ligand. An ethanol control titration was performed where 1 µl aliquots of ethanol were added to a 3 µM dansylated PBP solution. The presence of ethanol had a minimal effect on the change in fluorescence intensity (about a 10% increase of the total fluorescence change after 7 µl ethanol was added).
Homology modeling
Models of L. dispar PBP1 and PBP2 were obtained using the X-ray
structure of the Bombyx mori PBP
(Sandler et al.,
2000) as a template. The 143 amino acid sequence of PBP1 and the
145 amino acid sequence of PBP2 were submitted to the fully automated
SWISS-MODEL server using the `Iterative magic fit' algorithm available within
the Swiss-PDB Viewer interface
(http://www.expasy.ch/spdbv/)
(Guex and Peitsch, 1997).
Comparative protein modeling was used to generate a protein model since both
proteins exhibit significant sequence identity with B. mori PBP (61%
and 48% with PBP1 and PBP2, respectively).
Models of the complexes of pheromones 1a and 1b, as well as
aziridine analogs 2a and 2b, with PBP1 and PBP2 were constructed
as follows. The ligands were built using Insight/Discover software (Accelerys,
Inc.) implemented on a Silicon Graphics O2 platform. Ligands were constructed
assuming all-trans conformations for the alkyl chains and were
subjected to brief energy minimizations to obtain reasonable structures before
docking. The aziridine analogs were constructed in protonated form, consistent
with the biochemical data. Energy minimizations were performed using the CVFF
force field (Dauber-Osguthorpe et
al., 1988). Ligands were docked by manual positioning within
the binding site, at a similar location to the position of bombykol within the
binding site of B. mori PBP
(Sandler et al.,
2000). These complexes were then optimized by energy minimizations
(200 steps, using the conjugate gradient algorithm) before examination of the
intermolecular interactions (hydrogen bonds, hydrophobic interactions and
cation– interactions).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
I a. Importance of the functional group at position 7 and 8
The alkene 4a has a -system in place of the epoxide.
Consequently, the geometry around positions 7 and 8 is planar, which is
different from the geometry in an epoxide
(Figure 1). Also, the electron
density in an alkene is extended from C-7 to C-8, above and below the plane of
the carbon bond framework. In the epoxide, electron density is located in the
oxygen lone pairs, which are more diffuse (i.e. point further away from the
atom) and are located 55° relative to the plane of the epoxide. A
similarity between the epoxides 1a/1b and alkene 4a is
the rigidity around the C7–8 bond. Because of the subtle geometric and
electronic differences, as well as the biological function of 4a, we
were interested in the PBP binding of this alkene.
|
Radioassays revealed that both PBPs bind alkene 4a in the 4–5
µM range (Table 1), 2x
more weakly than their respective preferred epoxide enantiomer
(Plettner et al.,
2000). Binding of 1a and 4a was also examined by
endogenous tryptophan fluorescence titrations. Titration of PBP1 with
1a led to an increase in tryptophan fluorescence intensity. However,
titration with 4a led to a decrease in fluorescence intensity. In both
cases, PBP1 reached saturation after about 1 equivalent (1 µM) of the
corresponding ligand had been added to solution. In all cases the number of
binding sites per polypeptide was one and, as expected for a protein with one
binding site, no significant cooperativity (Hill coefficients 1) was
detected. Surprisingly, binding of both ligands under high ligand:PBP
conditions was 9–10x stronger
(Table 2) than observed with
the radioassay (Table 1). The
same observation was made with both dansylated PBPs and 4a
(Table 2).
|
|
To test whether the neutral or protonated forms of aziridines 2a and
2b bind to PBP, we determined the pH dependence of the competition
between radiolabeled 1a or 1b and cold 2a or 2b.
The pH profile of labeled 1a with cold 2a and PBP2 reveals
strongest competition in the neutral to basic pH ranges
(Figure 2A). A similar result
was obtained with radiolabeled 1b and cold 2b
(Figure 2B). The
pKa of aziridinium ion is 8.3
(O'Rourke et al.,
1956) and the strongest binding of 2a or 2b is at pH
< 8. These results suggest that PBP preferentially binds 2a and
2b in their protonated forms, which is surprising. Electronically,
protonated 2a and 2b are opposite to epoxides 1a and
1b (Figure 1): the O
carries a partial negative charge (
–) and the N of protonated
2a and 2b carries a full positive charge.
|
To examine the strength of aziridine binding we performed competitive
radioassays, in which a fixed amount of 1a was displaced with
increasing amounts of non-labeled 2a or 2b. The aziridines
displaced 1a from PBP2 with a saturation around 0.1 µM. Fluorimetric
titrations of 1 µM dansylated PBP1 or PBP2 with increasing concentrations
of 2a or 2b revealed similar affinities, one binding site and no
cooperativity (Figure 3,
Table 2). Very similar
Kd values were obtained by titration of 1 µM PBP1 with
either 2a or 2b and by monitoring the decrease in endogenous
tryptophan fluorescence with increasing ligand concentrations, suggesting that
PBP dansylation does not affect the binding properties significantly. The
advantage of monitoring dansyl fluorescence over tryptophan fluorescence is
that the dansyl group is a stronger fluorophore and the change in fluorescence
on ligand binding is larger (Figure
3). The dissociation constants obtained show that 2a bound
to PBP1 twice as strongly as 2b
(Table 2). Thus, some
enantiomer discrimination is taking place even though the protonated
aziridines are of opposite polarity to the epoxides. PBP2 bound 2a and
2b equally, consistent with previous observations that PBP2 is less
discriminating than PBP1 (Plettner et
al., 2000).
|
I b. Importance of the 2-methyl substituent
Three different assays were performed using 4a and 4b, PBP1 and PBP2. First, radioassays using labeled 4a and 4b revealed dissociation constants in the 6 µM range (Table 1). At this low ligand concentration relative to PBP, 4a appeared to bind more strongly than 4b to both PBP1 and PBP2. Second, fluorimetric titrations using 3 µM dansylated PBP1 and PBP2 with these ligands at saturating concentrations yielded Kd values 10-fold smaller than in the radioassay (Table 2). In addition, at this high ligand:PBP concentration ratio, 4b had a higher affinity for both PBP1 and PBP2 than 4a. Again, the same binding selectivity was obtained with endogenous tryptophan fluorescence titrations (Table 2). Thus, at low alkene concentration relative to PBP we observed the opposite ligand binding preference than at high alkene:PBP concentration ratio.
The dissociation constants obtained using the radioassay were significantly different from values obtained by fluorescence measurements. The radioassays with 4a and 4b were performed using detergents, either SDS or CHAPS, at 1% critical micelle concentration (CMC). To determine if these detergents were interfering with binding we performed experiments with 1 µM dansylated PBP1 in the presence of either SDS or CHAPS (data not shown). No major difference to detergent-free preparations was observed, showing that the detergents used in the radioassays were not responsible for the much weaker binding observed. The large difference between the Kd obtained with 4a and 4b in the radioassay and in fluorescent assays can only be explained by the difference in ligand concentration. In radioassays the ligand concentration is in the low nanomolar range with a large (µM) excess of protein. In fluorescent assays the ligand is present in stoichiometric quantities relative to the protein. We hypothesize that ligand binding to PBP influences the multimerization of protein and this, in turn, perturbs binding of further ligand.
II. Competition between pairs of ligands
In the assay with PBP1, radiolabeled 1b and non-labeled analogs 3–6, we detected a 2- to 3-fold increase in 1b binding in the presence of analogs (Figure 4A). In the competition binding assay using PBP2 and 1a, only 4a and 5 gave a slight but significant enhancement of 1a binding (Figure 4B). Similar results were obtained in two other assays with different lots of PBP. Compound 4a gave the strongest enhancement with both PBPs. Since this is of potential behavioural significance, we investigated binding of alkenes to the PBPs further. First, the affinity of PBP for radiolabeled 1b was nearly twice as large in the presence of cold 4a as in the absence, again suggesting an enhancement of binding (Table 1). Second, the pH profiles of radiolabeled 1a (Figure 5A) and 1b (Figure 5B) competing with cold 4a show that the enhancement occurs only above pH 6.5 for both PBPs.
|
|
III. Effect of PBP concentration on ligand binding
Previous experiments explored the binding of pheromone (nM) at increasing
concentrations of excess PBP (µM). Surprisingly, a 3-fold increase in
pheromone binding between 2 to 3 µM was observed
(Plettner et al.,
2000). PBPs equilibrate between monomeric, dimeric and
higher-order multimeric forms. This is apparent in native electrophoresis gels
where dimers, trimers and tetramers can easily be detected
(Plettner et al.,
2000). Dimeric forms have also been observed in gel filtration
experiments (Maida et al.,
1993). The effect of these equilibria should be more noticeable at
higher PBP concentrations, and to test this, fluorimetric titrations were
performed using either 1 or 3 µM dansylated PBP1 and 1a, 1c
(Table 2), 4a and
4b (data not shown) as titrants. Again, all experiments revealed one
binding site per polypeptide and no cooperativity. Kd
values for 1 µM PBP1 were 2-fold higher than with 3 µM protein,
suggesting slightly stronger binding at higher PBP concentrations. As
presented above, the strongest binding enhancements were seen in the
fluorimetric assays, where the ligand:PBP ratio was high.
|
Discussion I. |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
Binding interactions
In both competition assays and in direct binding assays with dansylated
PBPs we found that both PBPs bind to both 2a and 2b. Our
homology models based on the B.mori PBP crystal structure suggest
that a hydrogen bond from the side chain hydroxyl group of threonine 9 to the
epoxide moiety of 1b may contribute to binding of the compound to PBP1.
This effect was not detected for 1a, because the epoxide points away
from the threonine hydroxyl group. Also, this effect was not observed for PBP2
with either 1a or 1b because PBP2 has an alanine at position 9.
By analogy, it was expected that 2a should bind with no hydrogen bond,
while 2b should hydrogen bond at basic pH. However, we found that both
aziridines bind in protonated form, which precludes hydrogen bond donation
from the PBP to the ligand. The protonated aziridines have the opposite
polarity of the epoxides 1a and 1b, and the ability of the PBPs
to accommodate such a dramatic change in the functional group that is
recognized is remarkable. Docking of protonated 2a and 2b into
the homology models suggests that the aziridines are held in place by one or
more -cation interactions in the PBP binding site
(Figure 6). These interactions
are thought to be nearly the same strength as salt bridges in solvents of
intermediate polarity and approximately twice as strong in water
(Gallivan and Dougherty, 2000).
The aziridines could intercalate between the highly conserved phenylalanines
12 and 119 in the middle of the binding pocket and could interact with the
conserved tryptophan 37 and phenylalanine 36 at the base of the binding
pocket. The same interactions were predicted for PBP2, consistent with the
binding data which shows no significant difference between PBP1 and PBP2 in
the affinity for the aziridines. Interestingly, 2a and 2b were
active in electroantennogram (EAG) studies, but much less than 1a or
1b (Dickens et al.,
1997). Thus, even though the PBPs bind the aziridines with
approximately the same strength as the pheromone 1a, the aziridines
elicit very weak EAG responses and no behavioural responses, possibly because
they do not activate a transmembrane receptor.
|
All pH profiles of PBP–ligand binding we have examined exhibit a
sharp increase in binding between pH 6.0 and 6.5.
(Wojtasek and Leal, 1999;
Damberger et al.,
2000; Oldham et al.,
2000; Kowcun et al.,
2001). Previous studies suggest that a large conformational change
occurs in that range (Wojtasek and Leal,
1999; Damberger et
al., 2000; Horst et
al., 2001) and that the monomer/dimer equilibrium is shifted
towards the dimer at pH > 6.0 (Leal,
2000). Furthermore, recent studies
(Maida et al., 2000;
Deyu and Leal, 2002) suggest
that PBPs populate more than one stable conformer at pH > 6. This is
apparent in native electrophoresis, where the monomeric forms of many insect
PBPs and OBPs run as two closely spaced bands. Ligand binding could cause a
change in the conformer population in two ways: (i) by an induced-fit
mechanism or (ii) by preferentially binding to one of the existing conformers
and shifting the conformational equilibrium towards one form by mass-action.
The changes in fluorescence intensity of tryptophan [this study and
(Bette et al., 2002)]
or of the appended dansyl group we observed in our titrations probably reflect
a change in conformation of the PBP upon ligand binding.
The studies with 1c, 4a, and 4b were conducted to
observe steric effects within the binding sites of both PBP1 and PBP2. PBPs
bound 4b slightly more than 4a in the fluorescence assays,
suggesting that the 2-methyl group makes a small contribution to PBP binding,
although it is probably too small to account for the difference in the
behavioural effect or the EAG (Gries et
al., 1996). Similarly, the Kd values of
1c and 1a with PBP1 were not significantly different, suggesting
that PBP1, which so far has been the more selective of the two PBPs, does not
recognize the presence or absence of the methyl group. Interestingly, the
gypsy moth shows virtually no behavioural response to 1c and 4b
and EAG responses are very weak (Gries
et al., 1996). Thus, the transmembrane receptors must be
tuned very precisely to the presence or absence of the 2-methyl group.
II. Synergistic binding effects as explained by multiple equilibria
Incubation of PBP1 and 1b or PBP2 and 1a with hydrophobic
analogs yielded very surprising results. Some competing ligands act as
positive effectors, enhancing the binding of 1b or 1a. The
enhanced binding was more pronounced for PBP1 than for PBP2. The two PBPs
showed a different selectivity pattern: while PBP1 showed a significant
binding enhancement with 3a, 3b, 4a, 5 and
6, PBP2 showed an enhancement only with 4a and 5.
Scatchard analysis of PBP binding with various individual ligands reveal one
binding site/PBP and no significant cooperativity. The only way to explain the
binding enhancement is that binding of hydrophobic ligands perturbs the
multiple equilibria between PBPs and PBP–ligand complexes. For a
ligand-mediated binding enhancement to occur, the affinity of one
PBP–ligand complex for another complex has to be higher with multiple
ligands than with one ligand alone. The converse must be true when an
inhibition of binding occurs, as was seen previously for a 1:1 mixture of
1a:1b (racemic 1)
(Plettner et al.,
2000).
There is evidence to support the idea of PBPs forming reversible multimers
in a ligand-dependent manner. The X-ray structure of B.mori PBP
reveals a bundle of 
-helices with one binding site in the middle
(Sandler et al.,
2000). There is no obvious second binding site on the monomeric
PBP unit, although new binding sites could form at protein/protein interfaces
in the oligomers. The possibility of dynamic oligomerization is supported by
the direct observation of dimers and higher order multimers of PBPs. First,
the B.mori PBP crystallized as a head-to-tail dimer
(Sandler et al.,
2000). Second, dimers have been observed by gel filtration
(Leal, 2000;
Maida et al., 1993)
and native polyacrylamide gel electrophoresis
(Plettner et al.,
2000). Third, at constant pheromone and increasing PBP
concentration a sharp, 3-fold increase in pheromone binding can be seen
between 2 and 3 µM of PBP (Plettner
et al., 2000). In a gel filtration assay, significantly
more pheromone binding occurred at high PBP concentration than in dilute
preparations (Maida et al.,
1993). If PBP oligomerization had no effect on pheromone binding,
then the sharp increase would not be observed. The 2-fold increase in binding
strength seen when binding of 3 µM dansylated PBP is compared to binding of
1 µM PBP to ligand is also consistent with this concentration dependence.
Finally, observations from another group also support the existence of
ligand-mediated effects in olfactory systems. In Helicoverpa zea, a
significant increase in the frequency of action potentials of
pheromone-responsive neurons was observed when the preparations were exposed
simultaneously to (11Z)-hexadecenal and (11Z)-hexadecen-1-yl
acetate or to (11Z)-hexadecenal and linalool
(Ochieng et al.,
2002).
Binding enhancement of 1a or 1b was only seen with more
hydrophobic ligands such as 4a and at pH values above the pH
transition, parallel to individual ligand binding
(Kowcun et al., 2001;
Oldham et al., 2000)
and to dimerization (Leal,
2000). Interestingly, PBP1 exhibited this effect much more
strongly than PBP2. As discussed above, insect odorant-binding proteins and
PBPs have more than one stable conformer at pH > 6. Different ligands may
bind preferentially to one of these conformers and these complexes, in turn,
may form different multimers with each other. With the ligands studied here,
it may be possible to preferentially stabilize one of these conformers and to
obtain structural information. The functional significance of these multiple
ligand effects may be to provide a very early integration of blend
information.
III. Binding at high pheromone concentrations
The dissociation constant values obtained using the different assays
depended on the relative protein to ligand concentration. Radioassays
performed at low ligand:PBP ratios resulted in much weaker binding than
fluorescent titrations performed at high ligand:PBP ratios. The binding
enhancement seen at high pheromone concentrations is probably due to a
phenomenon similar to the dual-ligand effect discussed above. Namely, at high
pheromone concentration, multimers of PBP–ligand complexes form. If the
multimers form in a head-to-tail fashion as seen in the crystal structure
(Sandler et al.,
2000), then pheromone will be trapped in the multimeric PBP and
can leave the PBP only if the multimer dissociates first
(Scheme 2). The functional
significance of this effect may be to sequester excess pheromone at high
stimulus doses, thereby preventing an overloading of the receptors. Through
this mechanism, the PBP may provide an automatic stimulus attenuation: at low
doses the protein acts as a carrier and at high doses it acts as a carrier for
a portion of the material and sequesters the excess. The PBP is thought to
occur in the lymph in 1–10 mM concentration
(Vogt et al., 1989)
and the pheromone concentration would depend upon the dose in the airstream.
The lowest limit of pheromone in the airstream, that elicits behavior, is 1.6
x 10-17 M/s and the high limit, at which adaptation occurs,
is 2.8 x 10-5 M/s. The concentration of pheromone within
the sensillum is not known, because very little is known about transport
efficiency on cuticular structures and about pheromone accumulation rates
within sensory hairs. Assuming a 0.1% transfer efficiency from airstream to
sensory hair, the range of pheromone concentrations in the sensillum is 16
pM–28 M, and the ligand:PBP ratios would vary from 1.6 x
10-8 to 2.8 x 10-2 (Appendix). The range of
ligand:PBP ratios covered in our experiments was 10-5–1.
Thus, we have covered a meaningful portion of the potential biological range
of ligand:PBP ratios. The multimerization mechanism proposed here may explain
the remarkable concentration range to which moths respond.
|
To conclude, we have described three important findings with respect to PBPs in the gypsy moth. (i) The PBPs bind a variety of ligands that are structurally related to the pheromone. The PBPs show subtle differences in binding constants, which are not sufficient to explain the strong preferences detected for the same set of ligands in electroantennogram studies. However, the ligands studied require very different molecular interactions, in order to bind. These different binding modes may give rise to different protein conformations for the PBP-ligand complexes. (ii) We have observed binding synergy between different ligands when these are presented simultaneously. This effect may represent a very early interpretation of odor blends. (iii) By far the strongest differences in ligand binding are seen at different ratios of ligand:PBP. This dependence of the binding affinity on the ligand:PBP ratio may explain why PBP may act both as odorant transporter and as scavenger.
|
Appendix: supplementary material |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
Supplementary material can be found at http://www.chemse.oupjournals.org.
|
Acknowledgments |
|---|
We thank several colleagues at S.F.U.: Dr A. Tracey for use of the SGI, Angelica Kowcun and Nasim Morawej for help with protein purification, George Vamvounis for help with the fluorimetry experiments, and Regine Gries for providing 1c. We also thank an anonymous reviewer for very helpful critique. Funded by an award from Research Corporation (RI0519), and grants from NSERC (RGPIN222923) and SFU (PRG 99–3) to E.P.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussion I.Appendix: supplementary materialReferences |
|---|
Bette, S., Breer, H. and Krieger, J. (2002) Probing a pheromone binding protein of the silkmoth Antheraea polyphemus by endogenous tryptophan fluorescence. Insect Biochem. Mol. Biol.,32 , 241-246.[CrossRef][Web of Science][Medline]
Briand, L., Swasdipan, N., Nespoulous, C., Bezirard, V., Blon, F., Huet, J.-C., Ebert, P. and Pernollet, J.-C. (2002) Characterization of a chemosensory protein (ASP3c) from honeybee (Apis mellifera L.) as a brood pheromone carrier. Eur. J. Biochem., 269,4586 -4596.[Web of Science][Medline]
Damberger, F., Nikanova, L., Horst, R., Peng, G., Leal, W.S. and Wuethrich, K. (2000) NMR characterization of a pH-dependent equilibrium between two folded solution conformations of the pheromone-binding protein from Bombyx mori. Protein Sci.,9 , 1038-1041.
Dauber-Osguthorpe, P., Roberts, V.A., Osguthorpe, D.J., Wolff, J., Genest, M. and Hagler, A.T. (1988) Structure and energetics of ligand binding to proteins: Escherichia coli dihydrofolate reductase- trimethoprim, a drug-receptor system.Prot. Struct. Funct. Genet. , 4,31 -47.
Deyu, Z. and Leal, W.S. (2002) Conformational isomers of insect odorant-binding proteins.Arch. Biochem. Biophys. , 397,99 -105.[CrossRef][Web of Science][Medline]
Dickens, J.C., Oliver, J.E. and Mastro, V.C. (1997) Response and adaptation to analogs of disparlure by specialist antennal receptor neurons of gypsy moth, Lymantria dispar.J. Chem. Ecol. , 23,2197 -2210.[CrossRef][Web of Science]
Gallivan, J.P. and Dougherty, D.A.
(2000) A computational study of cation–instead
of the instead of the interactions vs. salt bridges in aqueous media:
implications for protein engineering. J. Am. Chem. Soc.,122
, 870-874.[CrossRef]
Grant, G.G., Langevin, D., Liska, J., Kapitola, P. and Chong, J.M. (1996) Olefin inhibitor of gypsy moth, Lymantria dispar, is a synergistic pheromone component of nun moth, L. monacha. Naturwissenschaften,83 , 328-330.[CrossRef][Web of Science]
Gries, G., Gries, R., Khaskin, G., Slessor, K. N., Grant, G. G., Liska, J. and Kapitola, P. (1996). Specificity of nun and gypsy moth sexual communication through multiple-component pheromone blends. Naturwissenschaften, 83,382 -385.[CrossRef][Web of Science]
Guex, N. and Peitsch, M. C. (1997) SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis,18 , 2714-2723.[CrossRef][Web of Science][Medline]
Hansen, K. (1984) Discrimination and production of disparlure enantiomers by the gypsy moth and the nun moth.Physiol. Entomol. , 9,9 -18.
Horst, R., Damberger, F., Luginbuehl, P., Guentert, P., Peng,
G., Nikanova, L., Leal, W. S. and Wuethrich, K. (2001)
NMR structure reveals intramolecular regulation mechanism for pheromone
binding and release. Proc. Natl. Acad. Sci. USA,98
, 14374-14379.
Kaissling, K.-E. (1977) Control of insect behavior via chemoreceptor organs. In Shorey, H.H. and McKelvey, J.J. (eds), Chemical Control of Insect Behavior. Theory and Application. Wiley, New York, pp.45 -65.
Kaissling, K.-E. and Priesner, E. (1970) Die Riechschwelle des Seidenspinners.Naturwissenschaften , 57,23 -28.[CrossRef][Web of Science][Medline]
Kaissling, K.-E. and Torson, J. (1980) Insect olfactory sensilla: structural, chemical and electrical aspects of the functional organization. In Sattelle, D.B., Hildebrand, J.G. and Hall, L.U. (eds), Receptors for Neurotransmitters, Hormones and Pheromones in Insects. Elsevier/North-Holland Biomedical Press, Amsterdam, pp. 261-282.
Keil, T.A. (1982) Contacts of pore tubules and sensory dendrites in antennal chemosensilla of a silkmoth: demonstration of a possible pathway for olfactory molecules. Tissue & Cell, 14,451 -462.[CrossRef][Web of Science][Medline]
Kowcun, A., Honson, N. and Plettner, E.
(2001) Olfaction in the gypsy moth, Lymantria
dispar: effect of pH, ionic strength and reductants on pheromone transport
by pheromone-binding proteins. J. Biol. Chem.,276
, 44770-44776.
Krieger, J. and Breer, H. (1999)
Olfactory reception in invertebrates. Science,286
, 720-723.
Leal, W.S. (2000) Duality monomer–dimer of the pheromone-binding protein of Bombyx mori.Biochem. Biophys. Res. Commun. , 268,521 -529.[CrossRef][Web of Science][Medline]
Maibeche-Coisne, M., Longhi, S., Jacquin-Joly, E., Brunel, C., Egloff, M.-P., Gastinel, L., Cambillau, C., Tegoni, M. and Nagnan-LeMeillour, P. (1998) Molecular cloning and bacterial expression of a general odorant-binding protein from the cabbage armyworm, Mamestra brassicae. Eur. J. Biochem.,258 , 768-774.[Web of Science][Medline]
Maida, R., Steinbrecht, A., Ziegelberger, G. and Pelosi, P. (1993) The pheromone binding protein of Bombyx mori: purification, characterization and immunocytochemical localization. Insect Biochem. Mol. Biol.,23 , 243-253.[CrossRef][Web of Science]
Maida, R., Krieger, J., Gebauer, T., Lange, U. and Ziegelberger, G. (2000) Three pheromone-binding proteins in olfactory sensilla of the two silkmoth species Antheraea polyphemus and Antheraea pernyi. Eur. J. Biochem.,267 , 2899-2908.[Web of Science][Medline]
Merritt, T.J.S., LaForest, S., Prestwich, G.D., Quattro, J.M. and Vogt, R.G. (1998) Patterns of gene duplication in lepidopteran pheromone binding proteins. J. Mol. Evol., 46,272 -276.[CrossRef][Web of Science][Medline]
Miller, J.R., Mori, K. and Roelofs, W.L. (1977) Gypsy moth field trapping and electroantennogram studies with pheromone enantiomers. J. Insect Physiol.,23 , 1447-1453.[CrossRef][Web of Science]
Mosbach, A., Campanacci, V., Lartigue, A., Tegoni, M., Cambillau, C. and Darbon, H. (2003) Solution structure of a chemosensory protein from the moth Mamestra brassicae.Biochem. J. , 369,39 -44.[CrossRef][Web of Science][Medline]
Mustaparta, H. (1997). Olfactory coding mechanisms for pheromone and interspecific signal information in related moth species. In Carde, R.T. and Minks, A.K. (eds), Insect Pheromone Research: New Directions. Chapman & Hall, New York, pp.144 -163.
Ochieng, S.A., Park, K.C. and Baker, T.C. (2002) Host plant volatiles synergize responses of sex pheromone-specific olfactory receptor neurons in male Helicoverpa zea.J. Comp. Physiol. A , 188,325 -333.
Oldham, N.J., Krieger, J., Breer, H., Fischedick, A., Hoskovec, M. and Svatos, A. (2000) Analysis of the silkworm moth pheromone binding protein–pheromone complex by electrospray-ionization mass spectrometry. Angew. Chem. Int. Ed., 39,4341 -4343.[CrossRef]
O'Rourke, C.E., Clapp, L.B. and Edwards, J.O. (1956) Reactions of ethylenimines VIII. Dissociation constants. J. Am. Chem. Soc., 78,2159 -2160.[CrossRef]
Plettner, E. (2002) Insect pheromone olfaction: new targets for the design of species-selective pest control agents. Curr. Med. Chem., 9,1075 -1085.[CrossRef][Web of Science][Medline]
Plettner, E., Lazar, J., Prestwich, E.G. and Prestwich, G.D. (2000) Discrimination of pheromone enantiomers by two pheromone binding proteins from the gypsy moth Lymnatria dispar.Biochemistry , 39,8953 -8962.[CrossRef][Medline]
Sandler, B.H., Nikonova, L., Leal, W.S. and Clardy, J. (2000) Sexual attraction in the silkworm moth: structure of the pheromone-binding-protein–bombykol complex.Chem. Biol. , 7,143 -151.[CrossRef][Web of Science][Medline]
Schneider, D., Kafka, W.A., Beroza, M. and Bierl, B.A. (1977) Odor receptor responses of male gypsy and nun moths (Lepidoptera, Lymantriidae) to disparlure and its analogues.J. Comp. Physiol. A , 113,1 -15.[CrossRef]
Vogt, R.G. and Riddiford, L.M. (1981) Pheromone binding and inactivation by moth antennae.Nature , 293,161 -163.[CrossRef][Medline]
Vogt, R.G., Riddiford, L.M. and Prestwich, G.D.
(1985) Kinetic properties of a sex pheromone-degrading
enzyme: the sensillar esterase of Antheraea polyphemus. Proc.
Natl. Acad. Sci. USA, 82,8827
-8831.
Vogt, R.G., Kohne, A.C., Dubnau, J.T. and Prestwich, G.D. (1989) Expression of pheromone binding proteins during antennal development in the gypsy moth, Lymantria dispar.J. Neurosci. , 9,3332 -3346.[Abstract]
Vogt, R.G., Rybczynski, R. and Lerner, M.R. (1991) Molecular cloning and sequencing of general-odorant binding proteins GOBP1 and GOBP2 from the tobacco hawk moth Manduca sexta: comparisons with other insect OBPs and their signal peptides.J. Neurosci. , 11,2972 -2984.[Abstract]
Vogt, R.G., Callahan, F.E., Rogers, M.E. and Dickens,
J.C. (1999) Cloning and expression of LAP, an adult
specific odorant binding protein of the true bug Lygus lineolaris
(Hemiptera, Heteroptera). Chem. Senses,24
, 481-495.
Wojtasek, H. and Leal, W.S. (1999)
Conformational change in the pheromone-binding protein from Bombyx
mori induced by pH and by interaction with membranes. J. Biol.
Chem., 274,30950
-30956.
Wyatt, T.D. (1997) Putting Pheromones to Work: Paths Forward for Direct Control. In Carde, R.T. and Minks, A.K. (eds), Insect Pheromone Research. New Directions. Chapman & Hall, New York, pp. 445-459.
Accepted May 23, 2003
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Solari, R. Crnjar, A. Frongia, G. Sollai, F. Secci, M. Spiga, C. Masala, and A. Liscia Oxaspiropentane Derivatives as Effective Sex Pheromone Analogues in the Gypsy Moth: Electrophysiological and Behavioral Evidence Chem Senses, October 1, 2007; 32(8): 755 - 763. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Pophof Pheromone-binding Proteins Contribute to the Activation of Olfactory Receptor Neurons in the Silkmoths Antheraea polyphemus and Bombyx mori Chem Senses, February 1, 2004; 29(2): 117 - 125. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||