define at least four different subsets of olfactory hairs and suggest that the ...... the 0. volvulus protein is also part of the same group of proteins, although in this ...
Neuron,
Vol. 12, 35-49,
January,
1994, Copyright
0 1994 by Cell Press
Members of a Family of Drosophila Putative Odorant-Binding Proteins Are Expressed in Different Subsets of Olfactory Hairs C. W. Pikielny,* G. Hasan,+ F. Rouyer,* and M. Rosbash* *Howard Hughes Institute and Department of Biology Brandeis University Waltham, Massachusetts 02254 +National Centre for Biological Sciences TIFR Centre, I. I. SC. Campus P. 0. Box 1234 Bangalore 560012 India
Summary A polymerase chain reaction-based method was used to generate a Drosophila melanogaster antenna1 cDNA library from which head cDNAs were subtracted. We identified five cDNAs that code for antenna1 proteins containing six cysteines in a conserved pattern shared with known moth antenna1 proteins, including pheromone-binding proteins. Another cDNA codes for a protein related to vertebrate brain proteins that bind hydrophobic ligands. In all, we describe seven antenna1 proteins which contain potential signal peptides, sug gesting that, like pheromone-binding proteins, they may be secreted in the lumen of olfactory hairs. The expression patterns of these putative odorant-binding proteins define at least four different subsets of olfactory hairs and suggest that the Drosophila olfactory apparatus is functionally segregated. Introduction Olfaction, the ability to sense and to discriminate among a large number of airborne molecules, often at minute concentrations, plays a major role in the feeding and mating behaviors of most terrestrial animals. Although the morphology, organization, and complexity of olfactory organs are quite different between insects and vertebrates, several aspects of the physiology appear to be similar (Schneider, 1984; Kaissling, 1986; Snyder et al., 1988; Reed, 1992; Ronnett and Snyder, 1992; Breer and Boekhoff, 1992; Firestein, 1992). Odorants act on olfactory neurons present in the olfactory mucosa in vertebrates and in olfactory sensilla primarily located on the antennae of insects. In both cases, theolfactory neuron membrane is separated from the air by an aqueous phase, the mucus in vertebrates (Snyder et al., 1988) and the sensillar fluid in insects (Kaissling, 1986). Detection of an odorant causes changes in the membrane potential of olfactory neurons (Firestein, 1992), which occurthrough G protein-mediated second messenger pathways, involving the generation of cyclic AMP in some cases and inositol 1,4,5-trisphosphate (IP3) in others (Reed, 1992; Ronnett and Snyder, 1992; Breer and Boekhoff, 1992).
In recent years, many proteins that are specific to vertebrate olfactory tissue and participate in signal transduction have been characterized at the molecular level (Reed, 1992; Ronnett and Snyder, 1992). Olfactory-specific adenylate cyclase and G protein have been isolated by virtue of their sequence similarity with proteins involved in other sensory processes. These molecules are thought to participate in steps of the signal transduction cascade common to many or all olfactory neurons. More recently, a large family of genes likely to code for odorant receptors has been cloned from rats, humans, mice, and catfish (Buck and Axel, 1991; Parmentier et al., 1992; Ngai et al., 1993a; Ressler et al., 1993). These proteins belong to the superfamily of G protein-coupled, seven transmembrane domain receptors, and they are specificallyexpressed in theolfactorytissueof all vertebrates tested. Interestingly, they contain variable regions that might be involved in the recognition of a large number of chemically diverse odorants. Generation of IP3 in response to odorants in a cell line expressing one of these proteins has provided direct evidence that they are indeed olfactory receptors (Raming et al., 1993). Earlier biochemical approaches led tothecharacterization of proteins capable of binding known odorants, namely, pyrazine odorants in vertebrates (Pevsner et al., 1985, 1986) and pheromones in moths (Vogt and Riddiford, 1981; Vogt et al., 1988,1989). When odorantbinding protein (OBP) from rat (Pevsner et al., 1988) and a frog homolog (Lee et al., 1987) were cloned, they were found to be related to a family of proteins that bind lipophilic compounds such as retinol, cholesterol, and biliverdin. As a consequence, OBPs are now thought to playthe role of carrier, allowing the solubilization and concentration of hydrophobic molecules in the aqueous mucus (Snyder et al., 1988; Ronnett and Snyder, 1992). More recently, cDNAs for four proteins that appear to code for abundant, secreted proteins expressed in rat olfactory tissue have been cloned and suggested to be novel OBPs (Dear et al., 1991a, 1991 b). Although odorant binding has not been demonstrated for any of them, three have sequence similarity to proteins that bind either chlorinated biphenyls or lipopolysaccharides, whereas a fourth (OBPII) shows sequence similarity to the original OBP. In moths, pheromone-binding proteins (PBPs) are expressed on the antennae of males (Vogt and Riddiford, 1981; Vogt et al., 1989, 1991a, 1991b) in specialized sensilla that are responsible for the detection of female pheromones (Steinbrecht et al., 1992). PBPs have been shown to bind directly to pheromones in vitro (Vogt et al., 1988) and are present at high concentrations in the sensillar fluid (Vogt and Riddiford, 1981). Thus, they are thought to play a role similar to that of OBPs in vertebrates despite the lack of sequence similarity (Gyorgyi et al., 1988; Raming et al., 1990; Krieger et
NWKNl 36
A 4) Subtractive
I a ae a d small mRNAs, I &a ofntotal RNA.
hybridization.
DRIVER B~AAAAAAAA
,biotm z I
1)
Partial
base
hydrolysis
of mRNAs.
A B
cDNA
~-
G-tailing
blotin
---_, biotin/-~
-I-
,blotin
biotin’----
and DENATURE HYBRIDIZE
and size
aTdTdTdTdTdldTdTdCdCdCdCdCdC
-----cVdTdTdTdTdTfldTdCdCdCdCdCdC
All cDNAs
3)
~-.--
synthesis
-z =
===I=
I
Reoea t on til tracer driver are almost entire/v different. I 2) First strand purification.
=
b&ah’-----
are of similar
size.
and PCR.
dCdCdCdCdCdt+A-dAdAdAdAdAdAdAdGdGdGdGdGdG dGdGdGdGdG+-dTdTOTdTdTfldTdTdCdCdCdCdCdC
B
dCdCdCdCdCdedAdAdAdAdAdAdAdGdGdGdGdGdG dGdGdGdGdG+-dTdTdTdTfldTdTdTdCdCdCdCdCdC
A mo,ifica strinaent
tie” with UC II or&&., hvbridization conditiorp.
EiaaarY BEADS
Figure
1. An Antennae-Minus-Head
cDNA
Library
(A) A PCR-based method for the construction cDNA libraries. For details, see Experimental Results. SA, avidin. (B) Analysis of the abundance of three specific successive subtractions. Southern blot analyses loading equal amounts of the cDNAs indicated The probes used are indicated below each panel.
rp49
of subtracted Procedures and cDNAs through were performed above the lanes. Ant., antennae.
IPBrec.
al., 1991; Vogt et al., 1991b). Other abundant moth antenna1 proteins called general odorant-binding proteins (COBPs) are highly similar in sequence to PBPs, but equally present in both sexes (Vogt et al., 1991a, 1991b). A direct interaction of GOBPs with odorants has not been demonstrated; however, their expression in antennae and their sequence similarity with PBPs have led to the proposal that they may also be OBPs. Drosophila has a sensitive but relatively simple olfactory system located on the third antenna1 segment (funiculus) (Siddiqi, 1987; Carlson, 1991). Approximately 500 sensory hairs (sensilla) can be subdivided according to three morphological types: trichoid (long, with a sharp tip), coeloconic (small, cone shaped), and
basiconic (club shaped). Sensilla of the three types contain a total of only about 2500 neurons and are distributed in a stereotyped fashion on the funiculus. Almost all funicular sensilla contain multiple pores in their cuticle (Venkatesh and Singh, 1984; ltoh et al., 1991; Stocker et al., 1992), a characteristic that in other insects is always associated with olfactory function (Altner and Prillinger, 1980). Other sensilla that may have an olfactory function can also be found on the maxillary palp (Singh and Nayak, 1985; Ayer and Carlson, 1992). To identify molecules involved in Drosophila olfaction, we have developed a method for the construction of subtracted cDNA libraries from very small amounts of starting material and used it to generate a library containing antenna1 cDNAs from which
A Family 37
of Drosophila
Olfactory
Hair
Proteins
head cDNAs have been subtracted. Using this library, we have found 7 clones with several properties that suggest they may be Drosophila OBPs. Results Subtraction Cloning of Antenna1 cDNAs The construction of subtractive cDNA libraries, in which ubiquitous cDNAs are removed from a collection of cDNAs from the tissue of interest, has proven useful in the identification of genes expressed in a tissue-specific manner (e.g., Wang and Brown, 1991). We have developed a polymerase chain reaction (PCR)-based method for the generation of subtracted cDNA libraries that combines two useful properties (Figure IA). First, it is designed to use very small amounts of tissue or RNA as starting material; we have routinely used as little as 10 ng of total RNA. Second, the amplified cDNA can be used as either component in multiple rounds of subtractive hybridizations, allowing for the enrichment of cDNAs specifically present or present at higher abundance in only one of the two populations. The method we used for cDNA synthesis is based on the one previously described by Belyavsky and colleagues (Belyavsky et al., 1989) with modifications designed to cope with two major difficulties. First, the efficiency of the PCR is very dependent on the size of a DNA molecule, such that long cDNAs tend to be out-competed by smaller ones (Lisitsyn et al., 1993). To overcome this limitation, our method leads to the specific amplification of cDNAs representing only the 400-600 nucleotides at the 3’ end of mRNAs, regardless of their size. RNA is partially base hydrolyzed to an average size of 500 nucleotides. After reverse transcription with an oligo(dT) primer, cDNAs of 400600 nucleotides are purified on a polyacrylamide gel. Second, theearliermethodcallsfortheGtailingofthe first strand cDNA, then allowing PCR amplification of the cDNAs with both an oligo(dT) primer (complementary to the natural poly[A] tail) and an oligo(dC) primer (complementary to the G tail). However, we have found that conditions compatible with annealing of the oligo(dT) primer tend to allow nonspecific priming bytheoligo(dC)primerasseen bytheamplification of nontailed products (data not shown). This is probably due to large differences in the melting temperatures of the two homopolymers and results in a lack of reproducibility in the sequence representation in the amplified cDNA population. To solve that problem, we designed an oligo(dT) primer with a string of deoxycytidines at its 5’ end. This allows the amplification of the G tailed first strand cDNA with a single oligo(dC) primer under stringent hybridization conditions, during which nonspecific priming is minimized. The reproducible representation in the amplified cDNA of sequences from the 3’ ends of a variety of large, as well as small, mRNAs was confirmed by
hybridizing Southern blots with specific probes. We found less than a 2-fold variation between duplicate samples that were processed in parallel (data not shown). We used this method to amplify cDNAs made from total RNA extracted from a dozen hand-dissected antennae (IO-20 ng) and from an equivalent amount of RNA extracted from heads without antennae. The amplified cDNAswere then used in two symmetrical subtractive hybridizations. One of the cDNAs (tracer) was radioactively labeled and hybridized to a IO- to 20-fold excess of the other cDNA (driver), which contained covalently linked biotin residues. The driver and hybrid cDNAs were then removed by avidin chromatography. The resulting subtracted cDNAs are designated antennae-minus-heads (antenna1 cDNA tracer, head cDNA driver) or heads-minus-antennae (head cDNA tracer, antenna1 cDNA driver). Since we expected the first round of hybridization to subtract preferentially abundant sequences (Wang and Brown, 1991), we proceeded to a second and third round using antennaeminus-heads as tracer and an equal mixture of heads and heads-minus-antennaeasdriver.ThecDNAsfrom one and two additional rounds of subtraction are designated (antennae-minus-heads)* and (antennaeminus-heads)3 cDNAs, respectively. After each round of hybridization, the efficiency of the subtraction was estimated from the percentage of radioactively labeled tracer that did not bind to the avidin beads. Whereas the first hybridization was greater than 90% effective in removing the tracer, hybridization to the driver decreased to 85% for the second round and to less than 50% for the third round. This suggested that most of the antenna1 sequences present after two rounds of subtractive hybridization were absent or rare in head cDNA. To confirm this interpretation, we measured by Southern blotting the representation of several wellcharacterized mRNAs in the various cDNAs (Figure IB). The level of the abundant and ubiquitous ribosomal protein 49 message (O’Connell and Rosbash, 1984), equal in antenna1 and head cDNA, decreased by a factor of 25 after two rounds of subtraction. The message for the much less abundant, neuronalspecific protein embryonic lethal abnormal visual system (elav) (Campos et al., 1987) increased 2-to 3-fold in thefirst round-subtractedcDNAs.Theopsin message, very abundant in the head and absent from antennae (Zuker et al., 1985), increased 2-to 3-fold in the headsminus-antennae cDNA (data not shown). We find that abundant sequences such as opsin (which is absent from antennae) and several abundant, antennalspecific cDNAs (data not shown) are only slightly enriched in heads-minus-antennae and antennae-minusheads, respectively, even though they are absent from the corresponding driver. This could be due to the fact that abundant sequences tend to form networks which are more likely to be lost through nonspecific
interactions with the driver. In contrast, sequences corresponding to the IP3 receptor, which has been shown to be encoded by a relatively rare mRNA 5 times more abundant in antennae than heads (Hasan and Rosbash, 1992), were further enriched by a factor of 25 after two subtractions (Figure IB). In summary, subtractive hybridizations seem to be having two different and sometimes contrary effects on the relative abundance of individual cDNAs. First, rare sequences are favored at the expense of abundant ones. Second, sequences present in the tracer but not in the driver are enriched after each round of hybridization, as exemplified by the case of the IP? receptor.
, >
Figure 2. Five Drosophila Proteins
Members
of a PBP-Related
Family
of
(A) Multiple sequence alignment of the Drosophila and moth members of the PBP-related family of antenna1 proteins. The PILEUP program was used, except for the final alignment of PBPRP-4 (near cysteine residue I), which was done manually. For PBPRP-4, the 22 amino acids between positions 40 c/) and 41 (V) are written on the next line. Two known moth sequences, Antheraea polyphemus PBP and Antheraea pernyi GOBPl, are left out of this comparison because they are too similar to A. pernyi PBP and M. sexta COBPI, respectively. Cysteines are boxed. Residues in common between three or more members of thefamilyareshaded;when~ochoicesarepossible,onlyone is shown. Putative signal peptides were defined as hydrophobic regions in a Kyte-Doolittle profile (Kyte and Doolittle, 1982) and obey criteria for eukaryotic signal peptide sequences (van Heijne, 1985); they are underlined. Sequences are from (Vogt et al., 199lb) or references therein. PBPAper, A. pernyi PBP; PBPMs, M. sexta PBP; COBPlMs, M. sexta GOBPI; GOBP2Ms, M. sexta GOBPL. (8) Plot of the overall sequence similarity in a multiple sequence alignment such as the one shown in (A), except that PBPRP4was left out. The PLOTSIMIlARITY program was used. The position of the cysteines in the multiple sequence alignment is indicated, as well as the minimal and maximal distances between any two successive cysteines.
Five Antenna1 Proteins Have Sequences Similar to Those of Moth PBPs After cloning the (antennae-minus-head+ cDNA into a plasmid vector, we picked 8 insert-containing clones at random and analyzed their expression by Northern blotting using RNA extracted from bodies (without heads, wings, or legs), heads (without antennae), and appendages (legs, antennae, and wings; data not shown). We found that 5 of the 8 clones show either specific or enriched expression in appendages. Partial sequence data gave clues to the identity of 2 of the clones, a2 and a5 (see below). In the case of a2, although noobvious match could be found with known sequences in computerized searches, we were struck by the fact that it codes for a small protein (148 amino acids) containing a putative signal peptide and six cysteine residues, as do several moth antenna1 proteins, PBPs and GOBPs (Raming et al., 1990; Vogt et al., 1991b). The spacing of the cysteine residues in A2 is also very similar to the spacing found in the moth proteins: (X36.42-Cys-X&ys-X,-Cys-X42-Cys-X,0-Cys-XSCys-XZC27, in which X, stands for any p amino acids) for the PBPs and GOBPs, and (X41-Cys-X2rCys-X3-CysX3,-Cys-Xe-Cys-&-Cys-Xls) for A2. In a multiple sequence alignment, other similarities can be found between A2 and the moth PBPs and GOBPs (see below). This led us to conclude that A2 is structurally related to this family of moth antenna1 proteins implicated in olfaction. Although A2 is likely one of the Drosophila counterparts of moth GOBPs (Vogt et al., 1991a, 1991b), we prefer the name PBPRPs for PBP-relatedproteins because it makes no assumption as to their a2 has thus been renamed pbprp-1. properties. PBPRP-1 is much less similar to any of the moth proteins than they are to each other (Figure 2A). In particular, the introduction of several small deletions and insertions is required to optimize the alignment to the moth proteins that are perfectly collinear between the cysteines (Vogt et al., 1991 b). The fact that the moth Manduca sexta has at least three related proteins (GOBPI, GOBP2, and PBP; Vogt et al., 1991a, is only ex1991b), coupled with the fact that pbprp-7 pressed in a fraction of olfactory hairs (see below), pbprp suggested the existence of other Drosophila clones. However, the low level of sequence similarity of PBPRP-1 to its moth counterparts and the absence
A Family
of Drosophila
Olfactory
Hair Proterns
39
Figure 3. Northern tive OBPs
Blot
Analysis
of Puta-
Five micrograms of total RNA from appendages, bodies, and heads was loaded in each lane as indicated. Probes are indicated under each panel. Number of counts and hybridization times were normalized for filters in same panel (except for rp49 in [B]). (A) PBP-related clones pbprp-7 to pbprp-5. (B) Two other antenna1 clones, a5 and a70 (pbprp-7 and rp49 are included for comparison).
Pbm-2
PbPrP-
PWP-3
1
P&w-5
rp49
of any cross-hybridizing band on low stringency Southern blots (data not shown) suggested that strategies based on sequence similarity might not succeed in finding other members of this family in Drosophila. As an alternative, we reasoned that as the moth genes are both expressed at relatively high and pbprp-7 levels in antennae and since pbprp-7 was discovered after sampling only 8 random clones, proteins of this familymightcomprisealargeproportionoftheclones in our subtractive library. Sixty-two clones were picked at random, and their inserts were analyzed on Southern blots using probes made from antenna1 and head cDNAs. Using this assay, 17 clones were found to have a higher level of expression in antennae than in heads. Cross-hybridization showed that these clones correspond to 11 different sequences, and partial seand quence data indicated that 2 clones (pbprp-2 pbprp-3) code for members of the PBPRP family. Partial sequencing of 33 clones, which were insufficiently abundant to be detected in the differential hybridiand pbprp-5) zation, showed that 2 more (pbprp-4 are part of the same family, bringing the total number to 5. In addition to the sequence similarity with moth PBPs and GOBPs, all five members of the family have putative signal sequences at their amino termini, suggesting that, like their moth relatives (Vogt et al., 199la; Steinbrecht et al., 1992), they are secreted in the lumen of olfactory hairs (Figure 2A). A multiple
sequence alignment shows that there is much more sequence divergence among the Drosophila members of this family of proteins than is the case among their moth counterparts. In particular, although all the moth proteins show identical spacing between the cysteines, each of the Drosophila PBPRPs has its own unique pattern except for PBPRP-2 and PBPRP5. The higher degree of similarity existing between PBPRP-2 and PBPRP-5 on one hand and PBPRP-1 and PBPRP-3 on the other may correspond to subfamilies of OBPs. A graphic representation of the similarity of all the proteins in the multiple sequence alignment (Figure 28) shows that, in addition to the hydrophobic leader peptide, three regions of the proteins, near cysteine 1, between cysteines 2 and 3, and near cysteines 4, 5, and 6, are more conserved and may thus take part in a structurally important motif. The low level of sequenceconservation among Droclones was confirmed by the fact that sophila pbprp low stringency genomic Southern blots with any of clones gave no hint of cross-hybridizing the pbprp genes (data not shown), making it impossible to estimatedirectlythetotal number ofpbprpclones in Drosophila. Members of the Drosophila PBP-Related Family Are Expressed in Different Subsets of Sensilla Northern blot analysis was carried out using RNAs extracted from bodies, heads, and appendages (see
Neuron 40
A
Anterior
A Family 41
of Drosophila
Olfactory
Hair
Proteins
al0
a5
Df rsal
Dqrsal
Anterior
-
Posterior
Anterior
-
Posterior Figure OBPs
Sagittal
Section
Figure 3A). The members of this family are expressed at different levels that are consistent with the initial Southern blot screen. pbprp-2 is expressed at the highest level, mostly in appendages but also in heads. The other four cDNAs, pbprp-7, pbprp-3, pbprp-4, and pbprp-5, only show detectable expression in appendages, with pbprp-3 and pbprp-4 showing the highest and lowest levels of expression, respectively. To delineate better the expression pattern of these 5 clones, we performed in situ hybridizations on frozen
4. Differential
Antenna1
Expression
Patterns
of Putative
(A) Horizontal sections of heads at approximately the same level were hybridized to probes made from all 5 PBP-related clones. In these sections, only the third segment of the antennae (the most distal from the head) is visible. For each of thepbprpclones, a view of the right antenna is shown. Note that the number of cells stained with each probe in these sections is not representative of the total number of stained cells in the third antenna1 segment. Ant, anterior; Med, medial; Post, posterior. In the lower right-hand corner, a horizontal section of a whole head (hybridized with an a70 probe) is shown for orientation. A, antenna; E, eye. (B) Sagittal sections of heads were hybridized with a5 and a70 probes. Views of the right antenna are shown in each case. S, sacculus. (C) Schematic frontal view of the right antenna. Area 1 corresponds approximately to expression of pbprp-I and pbprp-3. Similarly, pbprp-5 expression is approximated by area 2. A horizontal section is pictured to illustrate the three surfaces of the antenna: Ant, anterior; Post, posterior; Med, medial. Note that this horizontal section is upside down compared with the ones shown in (A).
sectionsof adult Drosophila heads. Exceptforpbprp-2 (see below), and in agreement with the Northern blot data, no staining could be detected in any part of the head other than the third antenna1 segment (data not shown). It remains possible that one or more of these genes is expressed in small numbers of cells in other parts of the body. The staining patterns for each of the pbprp genes on the funiculus suggest expression in various subsets of sensilla (Figure 4A). We found that the fraction of sensilla in which a given mRNA is present
NellrOll 42
C Figure
5. pbprp-2
Is Expressed
in Taste
Organs
in the Head
Sagittal sections of the head at the level of the proboscis were (A and B) Two successive sections through the proboscis show and at that of the support cells (B), which appear to be the site of sense organ. (C) Taste bristles at the tip of the proboscis. TB, taste bristles;
hybridized with a pbprp-2 probe. the ventral cybarial sensory organ at the level pbprp-2 expression (see text). OE, oesophagus; N, neuron;
SC, sheath
cells.
of the sensory VCSO, ventral
hairs (A) cybarial
A Family
of Drosophila
Olfactory
Hair
Proteins
43
corresponds generally to the relative expression levels detected by Northern blotting.pbprp-2,the most highly expressed gene of the family, also exhibits the most widespread expression, corresponding to regions of the antennae containing sensilla of all three types, basiconic, coeloconic, and trichoid (Figure 4A). However, pbprp-2 expression does not seem to occur in all sensilla (data not shown). In addition to its antenna1 expression and consistent with the Northern blot analysis, pbprp-2 mRNA is also detected in the maxillary palps (data not shown) and in cells at the bases of the taste hairs on the proboscis and internal taste organs of the head (Figure 5; see Stocker and Schorderet, 1981). In the case of the taste bristles on the proboscis, the larger size of the sensilla and a better tissue preservation allow the resolution of the two types of cells present (Figure 5C). It appears that, as has been shown for PBP (Steinbrecht et al., 1992), pbprp-2 is expressed in the sheath cells close to the surface of the cuticle and not in the more deeply located neurons. pbprp-1 andpbprp3 mRNAsareobserved in ventrolateral regions mostly on the anterior surface of the third antenna1 segment (Figure 4A; Figure 4C, area 1); those regions contain almost exclusively trichoid sensilla (Venkatesh and Singh, 1984). Although the resolution of this technique does not allow us to det’ermine whether the two genes are coexpressed in the same hairs, neither of them is expressed in all regions containing trichoid hairs. Direct counting of staining cells in successive sections suggests that about 80 cells, corresponding to close to half the number of trichoid sensilla (150) present on the funiculus (Venkatesh and Singh, 1984), express each of the two genes. Sensilla showing pbprp-5 expression are restricted to a different area mostly on the medial and posterior surface of the antennae with little or no overlap with pbprp-7 and pbprp-3 expression patterns (Figure 4A; Figure 4C, area 2). This expression pattern correlates well with the distribution of basiconic sensilla (Venkatesh and Singh, 1984). pbprp-4 mRNA is only observed in a small number of hairs scattered over the surface of the funiculus, consistent with the fact that it has a lower level of expression than any other gene in this family. Note that a single section is shown for each probe in Figure 4A and that the number of sensilla staining may underrepresent (pbprp-l and pbprp-3) or overrepresent (pbprp-4) the number of staining sensilla on the whole funiculus. Because of the male-specific expression patterns observed for was moth PBPs, expression of each Drosophilapbprp examined by in situ hybridization on sections from male as well as female animals; no difference was observed. Two Other Antenna1 Proteins Are Expressed in Different Subsets of Olfactory Hairs and Have Putative Signal Peptides at Their Amino Termini Since members of the multigene family are expressed
in different sensilla, we considered the possibility that this might also be true for antennal-specific cDNAs in our subtracted librarywhich are not related to PBP. Four additional clones of this type were analyzed by in situ hybridization. The 2 most abundant ones, a5
A I
B.PZl R.p23 OvAg A5 TFSl
. -KISA _.
50
.PV
..I..... fiK -lCLARSODND
HA ENANCKKCTPtlL EN ..llNOAIDFA OASID
51
TFSI
100
NAYVPODOOC
CDLR LLNEATHETS
151
2Do
251 8. P21 .a&x R. P23 I&w OvAg A5 TLYGVSE TFSl
B A10
10
20
30
40
50
so
IIGOPGFRRAI WESLYVAU~CTICFOVEGL PHPPATSPSP,,“ERWEGAY DDKFDNVDLO 70
SD
90
IOD
110
120
ElLNOERLLl NYIKCLEGTG PCTPDAKMK EILPDAIGTD CTKCTEKGRYGAEKVTRHL, 130
140
150
DNRPTDWERLEKlYDPEGTYRlKYGE,,KSKANEEP Figure 6. A5 and A10 Have Amino-Terminal Ends
Candidate
Signal
Peptides
at Their
(A) A5 is similar to a family of proteins from yeast, nematode, rat, and bovine. The sequences of the five members of this family were aligned using the PILEUP program in the CCC70 sequence package. Shaded residues are those in common between two or more members of the family. The putative signal peptides of A5 and the 0. volvulus antigen were defined as hydrophobic regions in aKyte-Doolittleprofile(Kyteand Doolittle, 1982),obey criteria for eukaryotic signal peptide sequences (van Heijne, 1985), and are underlined. To maximize the sequence similarity, a T was deleted at position 453 of the published sequence (Lobes et al., 1990), changing the frame after Tyr-134 (at position 171 of thealignment shown). B. ~21, bovine brain ~21 protein (Schoentgen et al., 1987); R. ~23, rat brain p23 protein (Grandyet al., 1990); OvAg, 0. volvulus antigen (Lobes et al., 1990); TFSl (Robinson and Tatchell, 1991). (B) Translated sequence of AIO. The putative signal peptide is underlined.
and a?0 (Figure 3B), have detectable and distinctive expression patterns (Figure 4B). The expression of a5 is restricted to cells at the bases of a few scattered sensilla on the posterior surface of the antenna (Figure 4B). The distribution of these as-positive cells suggests that they may be part of coeloconic sensilla. This attribution is confirmed by the lack of effect of the lz mutation, which selectively removes basiconic sensilla (Stocker and Cendre, 1989), on the a5 staining pattern (data not shown). The expression of a70 shows the most spatially restricted pattern of the seven we have characterized (Figure 4B). The a70 probe hybridizes primarily to cells within the largest cavity of the sacculus. The sacculus is a sac-like indentation of the posterior surface of the antennae, and its largest cavity has been shown to contain several types of coeloconic sensilla (Venkatesh and Singh, 1984; ltoh et al., 1991). In addition, a?0 is expressed in a few cells that lie outside of the sacculus on the anterior and posterior surfaces of the antennae, in sensillathatare probablyalsocoeloconic (Venkatesh and Singh, 1984). As mentioned above, no similarity exists between the sequences of A5 and A10 and those of the PBPrelated family of genes described here. However, the presence in their open reading frames of candidate signal peptides (Figures 6A and 6B) (von Heijne, 1982, 1985) and the absence of any potential transmembrane domain suggest that they too may be secreted in the lumen of the sensilla. In addition, we found significant similarities between A5 and sequences in the databases that may give clues to its function. The A5 Protein Is Highly Similar Both to Vertebrate Brain Proteins That Bind Hydrophobic Ligands and to Yeast and Nematode Proteins The open reading frame present in the a5 clone is highly similar to members of a previously described group of proteins (Figure 6A), including two vertebrate brain proteins and one yeast protein (TFSI), the latter implicated in the raslcdc25 signal transduction pathway (Tripp et al., 1989; Robinson and Tatchell, 1991). Both vertebrate proteins have been shown to bind with some specificity to hydrophobic ligands. One is a rat 23 kd protein that binds specifically to the correct stereoisomer of morphine, levorphanol (Grandy et al., 1990). The other, a bovine brain 21 kd protein, binds to phospholipids with a preference for phosphatidyl-ethanolamine (Schoentgen et al., 1987). In addition, A5 also has sequence similarity to a major antigen expressed by the nematode Onchocerca volvulus, theetiologicagent of onchocerciasis in humans (Lobos et al., 1990). A multiple alignment shows that the 0. volvulus protein is also part of the same group of proteins, although in this case optimal alignment requires the introduction of an arbitrary frameshift in the published sequence (see Figure 6A, legend). Although neither the rat brain 23 kd protein nor the yeast TFSI seems to have a signal peptide (as determined by their nucleotide sequence), both the A5 and
0. volvulus open reading frames start with stretches of highly hydrophobic amino acids. In the case of the 0. volvulus antigen, the secreted nature of this protein has been directly demonstrated by immunocytochemistry (Lobos et al., 1990). To explore the possibility that any of the cDNAs described above correspond to known mutations in olfaction-related genes, we mapped their locationson the cytological genetic map by in situ hybridization to larval polytene salivary gland chromosomes. pbprp-7, a70, pbprp-3, and pbprp-4 map on chromosome III at positions 698, 73F, 83CID, and 84D, respectively. a5 and pbprp-5 map to 22AlB and 288, respectively, on chromosome II. pbprp-2 is at 19D on the X-chromosome. No relevant mutation has been mapped to any of these locations. Discussion Drosophila melanogaster olfaction has been studied from the genetic (Ayyu b et al., 1990; Carlson, 1991) and behavioral (Tompkins et al., 1980; Venard and Jallon, 1980; Cailey et al., 1986; Ferveur et al., 1989) points of view; however, there have been relatively few detailed molecular studies. Cloning and characterization of previously characterized olfactory mutants are ongoing (Hasan, 1990; Carlson, 1991), and recently, the Drosophila IPs receptor gene has been cloned and shown to have enhanced expression in the antennae (Hasan and Rosbash, 1992; Yoshikawa et al., 1992). The enhancer trap method allows for the generation of mutants in genes screened for their expression in antennae, and this approach is being actively pursued to identify new olfactory genes (Riesgo-Escovar et al., 1992; Vijayraghavan et al., 1992). As an alternative method to identify genes involved in olfaction in Drosophila, we have generated a subtracted cDNA library enriched for clones expressed at higher levels in antennae than in the rest of the head. In a preliminary characterization of this library, we find that approximately half of the clones (7 of 12; data not shown) may be expressed specifically or preferentially in antennae. These results validate the method described here and suggest that it should be useful in other situations in which the amount of available tissue is a limiting factor.
PBPRPs from Drosophila and Moths Contain Six Cysteines in a Conserved Distribution We have described a group of Drosophila proteins related in sequence to moth PBPs and COBPs. The greater degree of sequence variation displayed by the Drosophila proteins than by their moth counterparts allows for a better analysis of the conserved features of this family than was previously possible (Raming et al., 1990; Vogt et al., 1991b). In particular, although the moth proteins are almost perfectly collinear, the alignment of the Drosophila members requires the inclusion of several small deletionsor inser-
A Family 45
of Drosophila
Olfactory
Hair Proteins
tions. The most striking feature shared by all these proteins, already noted in the case of the moth proteins (Raming et al., 1990; Vogt et al., 1991b), is the presence of six cysteines. Although the intercysteine spacing is generally conserved among all members of this family, the intervals between cysteines 1 and 2, 3 and 4, and 4 and 5 show a limited amount of variation (27-37, 33-43, and 9-11 residues, respectively), whereas those between cysteines 2-3 and 5-6 are strictly conserved (4 and 9 residues, respectively; Figure 2A). This suggests that some domains of the proteins may be more critical than others to a common secondary structure. Indeed, when the overall similarity of the nine proteins (Drosophila and moths combined) is graphically displayed, three domains appear to be more highly conserved (near cysteine 1, between cysteines 2 and 3, and between cysteines 4, 5, and 6; Figure 26). This general pattern of conservation in the vicinity of the cysteine residues suggests that all the proteins in this group have a similar structure, in which the cysteines play a critical role. A good candidate for such a role would be the establishment of interor intramolecular disulfide bonds (Raming et al., 1990), although the apparent requirement for a strict spacing between cysteine residues is more reminiscent of metal-binding domains (O’Halloran, 1993). Interestingly, three vertebrate OBPs (rat OBP, OBPII, and frog BG) and a fourth protein secreted by salivary glands and possibly involved in taste are part of a group of carrier proteins that share a general structure, unrelated to that of PBPRPs, in which intramolecular disulfide bonds play a critical role (Lee et al., 1987; Pevsner et al., 1988; Dear et al., 1991b; Schmale et al., 1990). We have found 5 pbprp clones through random sampling of a small number of clones (62) in our subtracted antenna1 library. Three of these clones were found only once and are present at frequencies of approximately one in a thousand (data not shown) in this library, suggesting that we have analyzed only a small fraction of all such cDNAs. Extrapolating from our results leads to the estimate that there should be at least 15 such clones in our library (p < 0.05; data not shown). The fact that only three PBPRPs have so far been found in any given moth may be due to such studies having been initiated through the biochemical purification of proteins, which would greatly bias the search toward the most abundant molecules. This bias may also explain the relatively small amount of sequence variation observed in the moth proteins compared with Drosophila PBPRPs. Varied Expression Patterns of the Drosophila PBPRPs Are Consistent with a Function in Olfaction The moth PBPs have been implicated in olfaction by their specific expression in pheromone-sensing hairs and by the fact that they bind pheromone. We have now shown that the Drosophila pbprp clones are expressed in hairs which vary in morphology, location,
and even primary sensory function, but are all consistent with the chemosensory detection of volatile molecules. Except for pbprp-2, expression of all thepbprp clones appears to be limited to cells on the surface of the third antenna1 segment, which is almost exclusively devoted to olfaction, although it may also contain some contact chemoreceptors (Stocker et al., 1992). pbprp-2 is also expressed in regions of the second antenna1 segment, on the maxillary palps, and in taste hairs in the head, both on the proboscis and in the internal taste organs. Expression in the maxillary palps is observed at the base of sensilla, which may correspond to olfactory hairs present in this secondaryolfactoryorgan (data not shown; Ayer and Carlson, 1992). Although the function of the pbprp-Zexpressingcellson the second antenna1 segment is uncertain, blowfly taste organs homologous to the ones showing pbprp-2 expression have been shown to be sensitive to hydrophobic molecules in two ways. First, behavioral studies have shown that, after ablation of all known olfactory organs, the presence of even low concentrations of hydrophobic molecules in a mixture whose taste is otherwise attractive to the fly will either prevent proboscis extension or promote regurgitation (Dethier, 1955). Second, electrophysiological recordings have shown that concentrated vapors of hydrophobic substances will provoke specific responses by some taste hair neurons (Dethier, 1972). The sensitivity of hairs primarily involved in taste to hydrophobic volatile molecules has been suggested to be important in preventing insects from ingesting food contaminated with toxic substances. In summary, the expression of the Drosophila PBPRPs, although more widespread and diversified than that of the moth PBPs, is confined to chemosensory organs and is consistent with a function in olfaction or closely related chemosensory processes. Spatial Segregation of the Olfactory Apparatus Is Revealed by the Expression Patterns of Putative OBPs It has been pointed out that, unlike the situation in vertebrates in which all olfactory neurons bathe in a common aqueous medium (mucus), the segregation of insect olfactory neurons in sensilla affords the possibility of differentiation of the sensillar fluid as exemplified by the restriction of PBPs to male-specific sensilla (Vogt et al., 1991a; Steinbrecht et al., 1992). Our data provide the first evidence that sensilla present in animals of both sexes also differ in the composition of their sensillar proteins. First, the presence of putative signal peptides and the overall sequence similarity suggest that, like PBPs, PBPRPs are secreted in the sensillar fluid. Since they also have potential putative signal peptides, this may also be the case for A5 and AIO. Second, our analysis of the spatial expression patterns of each of these proteins reveals a surprising diversity in expression patterns, both for the five PBPRPs and for the other two putative sensillar proteins. Althoughpbprp-2expression is widespread and
probably covers sensilla of all three morphological types, it seems not to be expressed in some sensilla (data not shown). At least four different, more restricted and partially overlapping expression patterns are displayed by the other 6 clones. The expression patterns of pbprp-I and pbprp3 on the one hand and pbprp5 on the other are confined to regions containing trichoid or basiconic sensilla, respectively(Figure 4C, areas 1 and 2). In both cases, however, only part of the area containing sensilla of the corresponding type is stained. pbprp-4 and a5 are each expressed in small numbers of cells scattered over a large fraction of the antenna1 surface. Although their expression patterns may overlap with each other and with those of pbprp-7, pbprp3, and pbprp-5, they are also expressed in regions from which the previous three genes are excluded and, for a5 in particular, containing mostly coeloconic sensilla. a70 expression is mainly confined to the sacculus. Since we have only sampled a small number of clones in our library, it seems likely that many other sensillar proteins exist which display other expression patterns. Finally, the variety in sequence and the similarityto proteins (PBP and the 21 kd and 23 kd vertebrate brain proteins) which bind to hydrophobic ligands with some specificity suggest that these sensillar proteins bind to different odorants and that their presence in a particular sensillarfluid may be related to theodorant specificity of that hair (Vogt et al., 1991a). The spatially segregated expression of putative OBPs may therefore correspond to a functional segregation of the peripheral olfactory apparatus. Acorrelation between the restricted expression patterns of putative Drosophila OBPs and a functional segregation is consistent with previous morphological and physiological studies. In insects as well as vertebrates, the first synaptic relay of olfactory information occurs in anatomically separate structures in the antenna1 lobe, the glomeruli. In both cases, specific odorants activate one or a small number of glomeruli, suggesting the existence of a topographic map of odors in the brain (Lancet et al., 1982; Rodrigues, 1988; Hansson et al., 1992; Shepherd, 1992). In addition, a relationship between specific glomeruli in the antennal lobe and groups of sensilla of similar morphology and location on the antennae has been suggested by retrograde labeling studies that correlate well with our findings (Stocker et al., 1983). In particular, retrograde filling of trichoids in an area showing pbprp-7 and pbprp3 expression has shown that all neurons project to the same two glomeruli (VA1 and DAI; Stocker et al., 1983). Similarly, all neurons from basiconit sensilla in an area expressing pbprp-5 have projections on the V, VMI, and DMI glomeruli. Finally, single hair recordings (Siddiqi, 1987), as well as electroantennograms from different regions of the antennae (Ayer and Carlson, 1992), have shown a relationship between position on the antenna1 surface and odorant specificity. The organization of the Drosophila antennae can
be compared with that of the vertebrate olfactory epithelium, recently revealed by the analysis of the expression patterns of individual odorant receptors (Nef et al., 1992; Ngai et al., 1993a, 1993b; Ressler et al., 1993; Vassar et al., 1993). In all cases, only one or a small number of olfactory receptors is expressed in each olfactory neuron. In catfish, neurons expressing any given receptor appear to be randomly distributed throughout the olfactory epithelium (Ngai et al., 1993a). In rats and mice, the expression of each receptor is limited to cells restricted to one of three broad bilaterally symmetrical regions, within which it also appears random (Nef et al., 1992; Ressler et al., 1993; Vassar et al., 1993). This segregation is probably maintained in the projection to the olfactory bulb and may correspond to an initial sorting of the olfactory information. The Drosophila olfactory apparatus appears to be simpler than that of vertebrates in the number of smells which can be discriminated (Siddiqi, 1987), in the number of olfactory neurons (Carlson, 1991), and in the number of glomeruli (Stocker et al., 1983). Therefore, in theabsenceof cloned odorant receptors from Drosophila (or any other invertebrate), our finding that at least four different su bsets of olfactory hairs can be defined on the basis of their putative OBPs may indicate a greater importance of the organization of the peripheral olfactory system in insects than in vertebrates. In vertebrates, one olfactory receptor has been shown to respond to odorants in the absence of any other olfactory-specific protein, strongly suggesting that OBPs are not absolutely required (Raming et al., 1993). It is likely that a similar situation is found in insects (Stengl et al., 1992). However, the existence in both cases of a variety of demonstrated or potential OBPs in the aqueous medium bathing olfactory neurons suggests that these proteins play an important role in olfaction. OBPs could increase the solubility of hydrophobic odorants in the aqueous environment bathing olfactory neurons and thus result in an increased sensitivity (Snyder et al., 1988; Vogt et al., 1991a). In an extension of this model, the OBP-odorant complex would be recognized as a whole by the membrane-bound receptor. Alternatively, binding to an OBP could be one of several processes involved in the odorant inactivation necessary for the proper temporal resolution of olfaction (Kaissling, 1986; Stengl et al., 1992). Further studies using reverse genetics and site-directed mutagenesis should allow testing of the role these putative Drosophila OBPs play in olfaction. Experimental
Procedures
Construction of a Subtracted (Antennae-Minus-Heads) cDNA Library using PCR RNA was extracted using standard protocols (Maniatis et al., 1982) from ten manually dissected antennae and a few heads (without antennae). The amount of RNA was estimated using quantitative PCR with rp49 primers (Hasan and Rosbash, 1992) and approximately 10 ng was used in each case, corresponding
A Family 47
of Drosophila
Olfactory
Hair Proteins
to either ten antennae or one-twentieth of a head. Each RNA was partially base hydrolyzed in the presence of 40 mM NaHCO,, 60 mM Na2C03 (pH 10.2) for 10 min at 60°C. Hydrolysis was stopped by the addition of one-tenth volume of 3 M sodium acetate (pH 6.5), 5% acetic acid, and the RNAwas ethanol precipitated. Reverse transcription was carried out with AMV reverse transcriptase (life Sciences) in the presence of 20 pmol of a poIyO oligonucleotide ([C]&AGCJTJ,,) for 10 min at room temperatureand45minat42°C.Thesamplewasdenatured informamide sample buffer and loaded on a 5% polyacrylamide, 7 M urea gel. The cDNA fractions corresponding to 400-600 nucleotides were eluted from the gel. Size-fractionated cDNA was G tailed with 4 units of terminal transferase (Pharmacia) in the presence of 10 mM dGTP for 30 min at 37°C. PCR was carried out in 100 ul in the presence of 3.75% dimethyl sulfoxide and 20 pmol of a poly(C) oligonucleotide (ATCAGGTCACCTCCA[C],,). To avoid mispriming, the mixtures were kept on ice until the addition of Taq polymerase and were then transferred immediately to a thermal cycler that had been preheated to 80°C. The PCR program (for a GENEAMP 9600 using thin-walled 200 ul tubes) was 1 min at 95°C; 30 sat 95OC, 30 sat 60°C, 1 min at 72°C for 3cycles; followed by 30 s at 95OC, 1.5 min at 72OC for 27 cycles; and 4 min at 72OC. Special precautions were taken to avoid contamination of the samples; whenever possible, manipulations were carried out in a laminar flow hood, and controls without RNA or cDNA were incorporated at theappropriate steps to detect any contaminant. In addition, the optimal number of cycles was determined in preliminary experiments to keep the amplification in its exponential phase. To generate driver cDNA, ten PCR reactions of 108 ul each were performed under conditions identical to the ones described above with XATCAGGTGACGTCGA(C)IJ (X is the biotinylated residue; the oligonucleotide was purchased from Midland and gel purified toeliminatecontamination with nonbiotinylated oligo) as primer. For generation of the tracer, a 100 ul reaction was performed in the presence of 10 uCi of a[“P]dCTP. A total of l-2 pg of tracer DNA and IO-20 pg of driver DNA were generated, mixed, ethanol precipitated, and purified on a Select5Lf5 Prime-3 Prime Inc.) spin column. After ethanol precipitation, the mixture was resuspended in 10 ul of water containing 10 ug of each of the two primers ([C]&IGCm,,and ATCACGTGACGTCCA[C]& The presence of a molar excess of these oligonucleotides prevents hybridization of the cDNAs through terminal sequences. Subtractive hybridization was performed essentially as described (Straus and Ausubel, 1990). The hybridization mixture was incubated at 100°C for 3 min, lyophilized, resuspended in 2.75 ul of H20, 1 PI of 10x EE (100 mM N-[2-hydroxyethyllpiperazine-N’-3-propanesulfonic acid [pH 8.251, IO mM EDTA), and 1.25 ~1 of 4 M NaCI, covered with mineral oil to prevent evaporation, and incubated at 95OC for 5 min and then at 65OC for 20 hr. The hybridization was stopped by dilution in 100 ~1 of EEN (1 x EE, NaC10.5 M) and 100 ul of a 5% suspension of avidin-coated beads (Baxter), which had been previously equilibrated in EEN. After a 30 min incubation at room temperature, the mixture was loaded onto a Millipore ultrafree MC unit (0.22 pm), and unhybridized tracer was recovered in the flowthrough after a short spin in a microfuge. The filter was rinsed once with 200 ~1 of EEN. The estimate of the fraction of tracer left after hybridization was obtained by comparing radioactivity in the 400-600 nucleotide region of a 5% acrylamide urea gel in aliquots of tracer and flowthrough. The subtracted cDNAwas gel purified on a preparative 5% acrylamide gel and ethanol precipitated before additional amplification by PCR and subsequent use in a second and third round of subtraction. In the second and third rounds of subtraction, the driver contained equal amounts of head cDNA and of head-minus-antennae subtracted cDNA to improve subtraction of rare sequences present in both heads and antennae (Wang and Brown, 1991). Cloning and Sequencing The subtracted cDNA was digested with Hindlll and Sal1 and cloned into pGEM4 (Promega), and relevant clones were sequenced using the Sequenase kit (US Biochemicals). The Send of each of the three cDNAs described in this work was cloned
using RACE (Frohman et al., 1988). Briefly, specific primers were designed for each mRNA and were Send labeled with $*P]ATP and polynucleotide kinase. Primer extension was carried out with poly(A)+ RNA from appendages, heads, or bodies. In each case, a unique appendage-specific extension product was observed of a size consistent with the one calculated from the size of the mRNA and the position of the primer on the 3’ partial clone. ThecDNA was then gel purified and G tailed as described above and amplified by PCR using the specific primer and the (QIJ primer. In each case, a PCR fragment of the appropriate length was obtained, cloned, and sequenced. Once the fulllength sequence was known, two primers flanking the coding sequence were synthesized and used for PCR on appendage cDNA. This PCRfragment was directly sequenced using the fmol DNA Sequencing System (Promega) to correct any errors introduced during Taq amplification of the initial PCR product. Direct sequencing of the PCR product avoids sequence errors that can occur in the analysis of single clones, Sequence Analysis Sequence analysis was performed using the CCC70 package from Genetics Computer Group, Inc. Database searches were performed at NCBI using the BLAST network service (Altshul et al., 1990). Northern and Southern Blot Analyses and In Situ Hybridizations to Adult Head Sections and to larval Polytene Chromosomes Frozen Drosophila adults were fractionated into bodies, heads, and appendages as described (Oliver and Philips, 1970). Northern and Southern blot analyses were performed using standard protocols (Maniatis et al., 1982). In situ hybridizations to adult head sections and to larval polytene chromosomes were as described (Hasan and Rosbash, 1992). Acknowledgments Address correspondence to C. W. P. We are grateful to R. Axel, 1. Hall, K. White, J. Rutila, H. Colot, and Y. Ahmed for comments on the manuscript, to E. Perkins for the photography, and to K. Sherman and Y. Ahmed for experimental help. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 USC Section 1734 solely to indicate this fact. Received
April
22, 1993; revised
October
20, 1993.
References Altner, H., and Prillinger, L. (1980). Ultrastructure of invertebrate chemo-, thermo-, and hygroreceptors and its functional significance. Int. Rev. Cytol. 67, 69-139. Altshul, S. F., Cish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990). Basic local alignment search tool. J. Mol. Biol. 275, 403-410. Ayer, R. K., and Carlson, J. (1992). Olfactory physiology Drosophila antenna and maxillary palp: acj6 distinguishes classes of odorant pathways. J. Neurobiol. 23, 965-982. Ayyub, C., Paranjape, J., Rodrigues, V., and Genetics of olfactory behavior in Drosophila Neurogenet. 6, 243-262.
in the two
Siddiqi, 0. (1990). melanogaster. J.
Belyavsky, A., Vinogradova, T., and Rajewski, K. (1989). PCRbased cDNA library construction: general cDNA libraries at the level of a few cells. Nucl. Acids Res. 77, 2919-2932. Breer, H., and Boekhoff, in olfaction. Curr. Opin.
I. (1992). Second messenger Neurobiol. 2, 439-443.
signalling
Buck, L., and Axel, R. (1991). A novel multigene family may encode odorant receptors: a molecular basis for odor recognition. Cell 65, 175-187. Campos,A. Molecular
R., Rosen, D. R., Robinow, S. N., and White, K. (1987). analysis of the locus elav in Drosophila melanogaster:
a gene whose Carbon, analysis.
expression
is neural
j. (1991). Olfaction Trends Neurosci.
specific.
in Drosophila: 74, 520-524.
EMBO genetic
J. 6,425-431. and molecular
Dear, T. N., Boehm, T., Keverne, E. B.,and Rabbitts,T. H. (1991a). Novel genes for potential ligand-binding proteins in subregions of the olfactory mucosa. EMBO J. 70, 2813-2819. Dear, T. N., Campbell, K., and Rabbitts, T. H. (1991b). Molecular cloning of putative odorant-binding and odorant-metabolizing proteins. Biochemistry 30,10376-10382. Dethier, V. G. (1955). The physiology and histology of the contact chemoreceptor of the blowfly. Quart. Rev. Biol. 30, 348-371. Dethier, V. G. (1972). Sensitivity of the blowfly to vapors. Proc. 2192.
of the contact chemoreceptors Natl. Acad. Sci. USA 69, 2189-
Karam, M. (1990). Identification of an Onchocerca vo/vulus cDNA encoding a low-molecular weight antigen uniquely recognized byonchocerciasis patient sera. Mol. Biochem. Parasitol. 39,135146. Maniatis, T., Fritsch, Cloning: A Laboratory Cold Spring Harbor
E. F., and Sambrook, Manual (Cold Spring Laboratory Press).
J. (1982). Harbor,
Molecular New York:
Nef, P., Hermans-Borgmeyer, I., Artieres-Pin, H., Beasley, L., Dionne, V., and Heinemann, S. F. (1992). Spatial pattern of receptor expression in theolfactoryepithelium. Proc. Natl. Acad. Sci. USA 89, 8948-8952. Ngai, J., Dowling, M. M., Buck, L., Axel, The family of genes encoding olfactory catfish. Cell 72, 657-666.
R., and Chess, A. (1993a). receptors in the channel
Ferveur, J.-F., Cobb, M., and Jallon, J.-M. (1989). Complex chemical messages in Drosophila. In Neurobiology of Sensory Systems, R. N. Singh and N. J. Strausfeld, eds. (New York: Plenum Publishing Corp.), pp. 397-409.
Ngai, J., Chess, A., Dowling, M. M., Necles, N., Macagno, E. R., and Axel, R. (1993b). Coding of olfactory information: topography of odorant receptorexpression in thecatfish olfatoryepithelium. Cell 72, 667-680.
Firestein, S. (1992). Electrical signals Curr. Opin. Neurobiol. 2, 444-448
O’Connell, O., and Rosbash, M. (1984). Sequence, codon preference of the Drosophila ribosomal gene. Nucl. Acids Res. 72, 5495-5513.
in olfactory
transduction.
Frohman, M. A., Dush, M. K., and Martin, G. R. (1988). Rapid production of full-length cDNAs from rare transcripts: amplification using a single gene-specific oligonucleotide primer. Proc. Natl. Acad. Sci. USA 85, 88989002. Gailey, D., Lacaillade, R. C., and Hall, 1. C. (1986). Chemosensory elements of courtship in normal and mutant, olfaction-deficient Drosophila. Behav. Genet. 76, 375-485. Grandy, D. K., Hanneman, E., Bunzow, J., Shih, M., Machida, C. A., Bidlak, J. M., and Civelli, 0. (1990). Purification, cloning and tissue distribution of a 23-kDa rat protein isolated by morphine affinity chromatography. Mol. Endocrinol. 4, 1370-1376. Cyorgyi, T. K., Robi-Shemkovitz, A. J., and Lerner, M. R. (1988). Characterization and cDNA cloning of the pheromone-binding protein from theTobacco hornworm, Manduca sexta. Proc. Natl. Acad. Sci. USA 85, 9851-9855. Hansson, B. S., Ljundberg, Functional specialization ence 256, 1313-1315.
H., Hallberg, of olfactory
Hasan, G. (1990). Molecular Drosophila melanogaster. 9041. Hasan, C., and Rosbash, mammalian intracellular and expression patterns the ryanodine receptor
H.,and Lofsted, C. (1992). glomeruli in a moth. Sci-
cloning of an olfactory gene from Proc. Natl. Acad. Sci. USA 87, 9037-
M. (1992). Drosophila homologs of two Ca2+-release channels: identification of the inositol 1,4,5-trisphosphate and genes. Development 776, 967-978.
Itoh,T.,Yokohari,F.,Tanjmura,T.,andTominaga,Y.(1991).External morphology of sensilla in the sacculus of an antenna1 flagellum of the fruitfly Drosophila melanogaster meigen. Int. J. Insect Morphol. Embryol. 20, 235-243. Kaissling, olfactory
K.-E. (1986). Chemo-electrical receptors. Annu. Rev. Neurosci.
Krieger, J., Raming, and complementary proteins: evidence 10888277-284.
transduction 9, 121-145.
in insect
K., and Breer, H. (1991). Cloning of genomic DNA encoding insect pheromone-binding for microdiversity. Biochem. Biophys. Acta
Kyte, J., and Doolittle, R. F. (1982). A simple the hydropathic character of a protein. 132.
method J. Mol.
for displaying Biol. 157, 105-
Lancet, D., Creer, C. A., Kauer, J. S., and Shepherd, G. M. (1982). Mapping of odor-related neuronal activity in the olfactory bulb by high resolution 2-deoxyglucose autoradiography. Proc. Nat. Acad. Sci. USA 79, 679-674. Lee, K. H., Wells, R., and Reed, R. R. (1987). Isolation of an olfactory cDNA: similarity to retinol-binding protein suggests a role in olfaction. Science 235, 1053-1056. Lisitsyn, N., Lisitsyn, ences between two
N., and Wigler, M. (1993). Cloning thediffercomplex genomes. Science 259,946-950.
Lobos,
M., Mengod,
E., Altmann,
C., Weiss,
N., Rudin,
W., and
Q’Halloran, pression.
T. V. (1993). Transition Science 267, 715-725.
Oliver, D. V., and Serv. 45, 58.
Philips,
metals
structure, protein
in control
J. P. (1970). Technical
note.
and rp49
of gene exDros.
Inf.
Parmentier, M., Libert, F., Schurmans, S., Schiffmann, S., Lefort, A., Eggerickx, D., Ledent, C., Mollereau, C., Gerard, C., Groote goed, A., and Vassart, C. (1992). Expression of members of the putative receptor gene family in mammalian germ cells. Nature 355, 453-455. Pevsner, J., Trifiletti, R., Strittmatter, (1985). Isolation and characterization protein for odorant pyrazines. Proc. 3050-3054.
S. M., and Snyder, S. H. of an olfactory receptor Natl. Acad. Sci. USA 82,
Pevsner, J., Sklar, P., and Snyder, S. H. (1986). Odorant-binding protein: localization to nasal glands and secretions. Proc. Natl. Acad. Sci. USA 83, 4942-4946. Pevsner, J., Reed, R. R., Feinstein, P. C., and Snyder, S. H. (1988). Molecular cloning of odorant-binding protein: member of a Iigand carrier family. Science 247, 336-339. Raming, K., Krieger, J., and Breer, H. (1990). Primary structure of a pheromone-binding protein from Antheraea pemyi: homologies with other ligand-carrying proteins. J. Comp. Physiol. (B) 760,503-509. Raming, K., Krieger, J., Strotmann, J., Boekhoff, I., Kubick, Baumstark, C., and Breer, H. (1993). Cloning and expression odorant receptors. Nature 367, 535-356. Reed, R. R. (1992). Signaling ron 8, 205-209.
pathways
in odorant
detection.
S., of Neu-
Ressler, K. J., Sullivan, S. L., and Buck, L. B. (1993). A zonal organization of odorant receptor gene expression in the olfactory epithelium. Cell 73, 597-609. Riesgo-Escovar, J., Woodward, C., Gaines, P., and Carlson, J. (1992). Development and organization of the Drosophila olfactory system: an analysis using enhancer traps. J. Neurobiol. 23, 947-964. Robinson, L. C., and Tatchell, K. (1991). TFSI: a suppressor of cdc25 mutations in Saccharomyces cerevisiae. Mol. Cen. Cenet. 230, 241-250. Rodrigues, V. (1988). Spatial coding of olfactory the antenna1 lobe of Drosophila melanogaster. 299-307. Ronnett, G. V., and of olfaction. Trends
Snyder, S. H. (1992). Molecular Neurosci. 75, 508513.
Schmale, H., Holtgreve-Grez, H., and Possible role for salivary gland protein cated by homology to lipophilic-ligand 343, 366-369.
information in Brain Res. 453, messengers
Christiansen. H. (1990). in taste reception indicarrier proteins. Nature
A Family
of Drosophila
Olfactory
Hair Proteins
49
Schneider, D. (1984). Insect olfaction: our research In Foundations of Sensory Science, W. W. Dawson, Springer-Verlag), pp. 381-418.
endeavor. ed. (Berlin:
Schoentgen, F., Saccoccio, F., Jolles, J., Bernier, I., and Jolles, P. (1987). Complete amino acid sequence of a basic 21-kDa protein from bovine brain cytosol. Eur. J. Biochem. 799, 333-338. Shepherd, 458.
G. (1992).
Modules
for
molecules.
Siddiqi, 0. (1987). Neurogenetics of olfaction nogasfer. Trends Cenet. 3, 137-142.
Nature
358, 457-
in Drosophila
me/a-
Singh, R. N., and Nayak, 5. (1985). Fine structure and primary sensory projections of sensilla on the maxillary palp of Drosophi/a melanogaster meigen. Int. J. Morphol. Embryol. 74, 291-306. Snyder, S. H., Sklar, P. B., and Pevsner,). (1988). Molecular anism of olfaction. J. Biol. Chem. 263, 13971-13974.
mech-
Steinbrecht, R. A., Ozaki, M., and Ziegelberger, G. (1992). Immunocytochemical localization of pheromone-binding protein in moth antennae. Cell Tissue Res. 270, 287-302. Stengl, M., Zufall, F., Hatt, H., and Hildebrand, J. C. (1992). Olfactory receptor neurons from antennae of developing male Manduca sexta respond to components of the species-specific sex pheromone in vitro. J. Neurosci. 72, 2523-2531. Stocker, R. F., and Cendre, N. (1989). Courtship behavior sophila genetically or surgically deprived of basiconic Behav. Genet. 79, 371-385.
of Drosensilla.
Stocker, R. F., and Schorderet, M. (1981). Cobalt filling of sensory projections from internal and external mouthparts in Drosophila. Cell Tissue Res. 276, 513-523. Stocker, R. F., Singh, R. N., Schorderet, M.,and Siddiqi,O. (1983). Projection patterns of different types of antenna1 sensilla in the antenna1 glomeruli of Drosophila melanogaster. Cell Tissue Res. 232, 237-248. Stocker, R. F., Gendre, N., Lienhard, M. C., and Link, B. (1992). Drosophila olfaction: structural, behavioral, developmental and genetic approach. In Nervous Systems: Principles of Design and Function, R. N. Singh, ed. (New Delhi, India: Wiley Eastern Limited), pp. 351-371. Straus, D., and Ausubel, ing DNA corresponding Sci. USA 87, 1889-1893.
F. (1990). Genomic subtraction to deletion mutations. Proc.
Tompkins, L., Hall, J. C., and Hall, stimulating volatile compounds from sophila. 1. Insect Physiol. 26, 689-697.
L. M. normal
for clonNat. Acad.
(1980). Courtshipand mutant Dro-
Tripp, M. L., Bouchard, R. A., and Pinon, R. (1989). Cloning and characterization of NSPI, a locus encoding a component of a CDC2Edependent nutrient-responsive pathway in Saccharomyces cerevisiae. Mol. Microbial. 3, 1319-1327. Vassar, J., Ngai, J., and Axel, R. (1993). Spatial segregation of odorant receptor expression in the mammalian olfactory epithe lium. Cell 74, 309-318. Venard, R., and Jallon, pheromone of female
J.-M. (1980). Drosophila.
Evidence for an aphrodisiac Experientia 36, 211-212.
Venkatesh, S., and Singh, R. N. (1984). Sensilla on the third antennal segment of Drosophila melanogaster. Cell Tissue Res. 232, 237-248. Vijayraghavan, K., Kaur, J., Paranjape, J., and Rodrigues, V. (1992). The east gene of Drosophila melanogaster is expressed in the developing nervous system and is required for normal olfactory and gustatory responses in the adult. Dev. Biol. 754, 23-36. Vogt, R. C., and Riddiford, L. M. (1981). Pheromone-binding inactivation by moth antennae. Nature 293, 161-163.
and
Vogt, R. C., Prestwich, G. D., and Riddiford, L. M. (1988). Sex pheromone receptor proteins;visualization using a radiolabeled photoaffinity analog. J. Biol. Chem. 263, 3952-3959. Vogt, R. G., Kohne, A. C., Dubnau, J. T., and Prestwich, G. D. (1989). Expression of pheromone-binding protein during antennal development in the gypsy moth Lymantria dispar. J. Neurosci. 9, 3332-3346.
Vogt, R., Prestwich, C. D., and Lerner, M. R. (1991a). Odorantbinding protein subfamilies associate with distinct classes of olfactory receptor neurons in insects. J. Neurobiol. 22, 74-84. Vogt, R. G., cloning and GOBPI and comparison J. Neurosci.
Rybcynski, R., and Lerner, M. R. (1991b). Molecular sequencing of general odorant-binding proteins COBPLfrom the tobacco hawk moth Manducasexta: with other insect OBPs and their signal peptides. 17, 2972-2984.
von Heijne, drophobic.
G. (1982). Signal sequences J. Mol. Biol. 759, 537-541.
von Heijne, G. (1985). Signal J. Mol. Biol. 184, 99-105.
sequences,
are
not
the limits
Wang, Z., and Brown, D. D. (1991). A gene Proc. Natl. Acad. Sci. USA 88, 11505-11509.
uniformly
hy-
of variation.
expression
screen.
Yoshikawa, S., Tanimura, T., Miyawaki, A., Nakamura, M., Yuzaki,M., Furuichi,T.,andMikoshiba,K.(1992).Molecularcloning and characterization of the inositol 1,4,5-trisphosphate receptor in Drosophila melanogaster. J. Biol. Chem. 23, 16613-16619. Zuker, C. S., Cowman, A. F., and Rubin, and structure of a rhodopsin gene from 40, 851-858.
G. M. (1985). Isolation D. melanogaster. Cell
