Visual arrestins in olfactory pathways of and the vector gambiae

C. E. Merrill*, J. Riesgo-Escovar†‡, R. J. Pitts*, F. C. Kafatos§, J. R. Carlson†, and L. J. Zwiebel*§¶

*Department of Biological Sciences, Program in Developmental Biology and Center for Molecular Neuroscience, VU Station B35-1812, Vanderbilt University, Nashville, TN 37235-1812; †Department of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT 06520; ‡Department of Developmental Neurobiology, Centro de Neurobiologı´a,Apdo. Postal 1-1141, Quere´taro, Qro 76001, Mexico; and §European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69117 Heidelberg, Germany

Edited by John H. Law, University of Arizona, Tucson, AZ, and approved November 7, 2001 (received for review September 24, 2001) Arrestins are important components for desensitization of G protein- transduction systems. Taken together this report establishes that coupled receptor cascades that mediate neurotransmission as well as particular arrestins are important components in both visual- olfactory and visual sensory reception. We have isolated AgArr1,an and nonvisual-signaling pathways. arrestin-encoding cDNA from the malaria vector mosquito, Anopheles gambiae, where olfaction is critical for vectorial capacity. Analysis of Experimental Procedures AgArr1 expression revealed an overlap between chemosensory and Generation and Screening of a Subtracted cDNA Library. Total RNA photoreceptor neurons. Furthermore, an examination of previously was prepared from hand-dissected olfactory tissue (antennae and identified arrestins from Drosophila melanogaster exposed similar maxillary palps) or carcasses (bodies stripped of heads including bimodal expression, and Drosophila arrestin mutants demonstrate antennae, maxillary palps, and all other appendages) of A. gambiae impaired electrophysiological responses to olfactory stimuli. Thus, we by using RNeasy kit (Qiagen, Chatsworth, CA). Selective amplifi- show that arrestins in Drosophila are required for normal olfactory cation via biotin and restriction-mediated enrichment (SABRE) physiology in addition to their previously described role in visual (11) was used to prepare an ‘‘antennae minus carcass’’ subtracted signaling. These findings suggest that individual arrestins function in library. The original protocol was modified by using SuperScript II both olfactory and visual pathways in Dipteran ; these genes reverse transcriptase (GIBCO͞BRL) for first-strand cDNA syn- may prove useful in the design of control strategies that target thesis and by redesigning PCR primers. Purified antennal cDNAs olfactory-dependent behaviors of insect disease vectors. were amplified and subtracted an additional time or subcloned directly into pBluescript II-KS (Stratagene). he events that initiate and terminate chemosensory signal Initial sampling of 53 insert-containing plasmids by DNA se- Ttransduction have received considerable attention in recent quencing showed that eight cDNA fragments had significant sim- years. Olfactory transduction is mediated by G protein-coupled ilarly to arrestins by BLAST (12) analysis. One fragment was chosen receptor (GPCR) second messenger pathways (1, 2). Olfaction for further study based on its strong similarly to a 94-aa stretch of exhibits not only activation but also adaptation, whereby a progres- antennal-specific arrestins. Specific primers were designed for both Ј Ј sively weaker response is generated to repeated or persistent 5 and 3 rapid amplification of cDNA ends (RACE; ref. 13) by stimuli. At the molecular level, desensitization of GPCR signaling using a Marathon cDNA amplification kit (CLONTECH). These Ј Ј is brought about by reduced coupling between the receptor and were designated Arr5 -RACE (5 -CTAGTCTCCAGCGATGC- Ј Ј Ј heterotrimeric G proteins. After receptor phosphorylation by G CACTGTGTT-3 ) and Arr3 -RACE (5 -CAGCTGGGTGTGT- Ј protein-receptor kinases, binding of an arrestin uncouples GPCRs GGATGTGGTGC-3 ). Finally, the complete cDNA representing Ј Ј from the signaling cascade (3). Arrestins also have been shown to this A. gambiae arrestin gene was engineered by joining 5 and 3 be involved in GPCR internalization, an integral component of RACE products before final ligation into pBlueScript II (KS) GPCR resensitization in many systems (4). (Stratagene) and was designated AgArr1. Most arrestins characterized to date have been divided into two broad categories that are largely reflective of their presump- Sequence Analysis. Sequence comparisons with the DNA Data Base ͞ ͞ tive functional roles (5). Visual arrestins have been reported to in Japan European Molecular Biology Laboratory (EMBL) display nearly exclusive expression in photoreceptors and inter- GenBank and SWISSPROT databases were performed by using act with rhodopsin to terminate visual stimulation (6, 7). The the GCG software (14), and protein alignment was performed by second general arrestin subtype category comprises the non- using the CLUSTAL W software package (15). visual arrestins that are expressed in a wide variety of tissues (excluding photoreceptors) where they are presumed to interact Reverse Transcription–PCR (RT-PCR). RNA for developmental pro- with a diverse array of GPCRs (8). files was made as above by using whole bodies at various stages: (i) In light of the rapid progress in understanding the mechanisms embryos refer to freshly laid fertilized eggs; (ii) early larvae are 1–2 that govern olfaction (9), we have initiated molecular studies to instars, (iii) late larvae are 3–4 instars, and (iv) pupae refers to extend this understanding to insect disease vectors. Olfaction pupae taken at least 6 h before eclosion. Total adult A. gambiae principally mediates host preference, a trait that strongly contrib- RNA was prepared as described above from either female or male utes to the ability of Anopheline mosquitoes to act as vectors for malaria or other serious human diseases (10). We now report the This paper was submitted directly (Track II) to the PNAS office. cloning and characterization of AgArr1, a novel arrestin family Abbreviations: GPCR, G protein-coupled receptor; RACE, rapid amplification of cDNA ends; member from Anopheles gambiae sensu stricto (hereafter A. gam- EAG, electroantennogram; EPG, electropalpogram; EA, ethyl acetate; SH3, Src homology 3; biae). Additionally, we show its expression in both olfactory and BU, butanol; RT-PCR, reverse transcription–PCR. visual systems. Such an overlap in sensory system expression is Data deposition: The sequence reported in this paper has been deposited in the GenBank verified by studies of two arrestins in Drosophila melanogaster database (accession no. AY017417 for AgArr1). previously thought to be exclusively visual in nature; both Arrestin1 See commentary on page 1113. and Arrestin2 (DmArr1 and DmArr2) are expressed not only in ¶To whom reprint requests should be addressed. E-mail: [email protected]. photoreceptors, but also in olfactory neurons. By using well- The publication costs of this article were defrayed in part by page charge payment. This

established olfactory paradigms in the Drosophila model system, we article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. NEUROBIOLOGY directly demonstrate the function of arrestins in olfactory signal §1734 solely to indicate this fact.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.022505499 PNAS ͉ February 5, 2002 ͉ vol. 99 ͉ no. 3 ͉ 1633–1638 Downloaded by guest on October 2, 2021 head appendages (antennae, palps, and proboscis), heads stripped SPOT camera system (Diagnostic Instruments, Sterling Heights, of appendages, or carcasses (bodies stripped of heads). PCRs were MI). Drosophila stainings were done as in ref. 19 by using Abs carried out with AgArr1 oligonucleotide primers ARR5.0 (5Ј- directed against DmArr1 and DmArr2 kindly provided by P. Dolph TTAAGGCCATGGTCCAGCAGGGTG-3Ј) and ARR-R5 (5Ј- (Dartmouth University, Hanover, NH) and C. Zuker (University of CGTTGCGTCGTCTATTCAAA-3Ј). Expression of ribosomal California, San Diego) (20). Sections and whole mounts were protein s7 (16) was assayed in parallel reactions by using sequence- incubated with horseradish peroxidase-conjugated secondary Abs, specific primers: S7a (5Ј-GGCGATCATCATCTACGTGC-3Ј) developed, mounted, and photographed by using a Zeiss Axiophot and S7b (5Ј-GTAGCTGCTGCAAACTTCGG-3Ј). D. melano- microscope. gaster RT-PCR reactions were performed on adult RNA generated as above from acetone-fixed, hand-dissected third antennal seg- Electrophysiology. Electrophysiological recordings of antennae and ments, heads stripped of antennae, and bodies with heads removed. maxillary palps in D. melanogaster were done with two different The reactions to amplify DmArr1 used primers DmArra1 (5Ј- setups, as described in refs. 19 and 21. To generate recordings, glass CAACTCCAACAAGGTGGTGA-3Ј) and DmArra3 (5Ј-GCT- microelectrodes were positioned on the surfaces of these olfactory TCCAGTTGGGCCTTG-3Ј); DmArr2 used primers DmArrb1 organs until electrical contact was made. Reference electrodes were (5Ј-ATGGTGAACGCCCAGTTTAG-3Ј) and DmArrb2 (5Ј- inserted into the head capsule. Most experiments were carried out GGCGAAGTCCTCGAATACAA-3Ј). PCR for DmKrz cDNA independently in both setups and gave similar results. Dilutions of was performed with krzL1 (5Ј-GGTGGTGGTGGAGGAAGTG- odorants were in paraffin oil. 3Ј) and krzR1 (5Ј-GCTGCTCACCGACTTTGGAT-3Ј) (17). In all reactions, primers for the ribosomal protein 49 gene (rp49a, 5Ј- Insect Stocks and Culture. A. gambiae G3 strain embryos were kindly GTATCGACAACAGAGTCGGTCGC-3Ј and rp49b, 5Ј-TTG- provided by Mark Benedict (Centers for Disease Control and GTGAGCGGACCGACAGCTGC-3Ј) were included to simulta- Prevention, Atlanta) and reared to adult stage under standard neously amplify a constitutive message. conditions. Adult mosquitoes were maintained in plastic containers at 27°C with 75% relative humidity under a 12:12 photoperiod. For Ab Production. Full-length AgArr1 cDNA was cloned into pET15b Drosophila, the DmArr1 and DmArr2 mutant alleles and the (Novagen), transformed into BL21 (DE3)pLysS bacteria (No- Df(2L)I131 stock were obtained from P. Dolph and C. Zuker and vagen), and purified by using His-Bind affinity resin (Novagen). are described in ref. 20. Dp(2:Y)H3 was obtained from the Indiana Polyclonal Abs were generated by using standard immunization University, Bloomington, stock center and is a duplication that protocols. Affinity-purified antisera was obtained by passing a covers the DmArr1 locus. Controls were heterozygous for the Protein G (Sigma)-purified IgG fraction over an affinity column mutations in the cases in which the effects were recessive, or flies constructed from purified soluble rAgARR1 immobilized on heterozygous for the mutation (in the case of arr12) or deficiency, resin (AffiGel 10͞15, Bio-Rad) according to the manufacturer’s and carrying an extra copy of the gene in a duplication, and the instructions. wild-type stock Canton S. All other stocks and balancers used are described in ref. 22. Immunoblots. Frozen male and female adult A. gambiae (G3 strain) antennae͞palps, heads, and carcasses were hand-dissected over dry Results ice for preparation of crude total protein extracts by homogeniza- A putative arrestin gene, AgArr1, was isolated as part of a broad tion in 50 ␮l of buffer HmB (PBS͞0.1% SDS͞0.1 M 2-mercapto- screen for olfactory components in A. gambiae, whereby a pool ethanol); samples were mixed with 10 ␮lofSDS͞PAGE sample of cDNA fragments representing enhanced antennal expression buffer (62.5 mM Tris⅐HCl,pH6.8͞10% glycerol͞2% SDS͞5% were subcloned and surveyed by random sequencing. We sub- 2-mercaptoethanol͞0.05% bromophenol blue) and boiled for 5 sequently used 5Ј and 3Ј RACE to complete the sequence of this min. Samples were then electrophoresed on a 10% SDS͞PAGE gel gene. The total length of AgArr1 encompasses 1,964 bp, of which and transferred to poly(vinylidene difluoride) by using semidry 1,152 bp constitutes a 383-aa residue ORF followed by a 553-bp techniques (Bio-Rad). Blots were probed with polyclonal rabbit 3Ј untranslated region, including a 40-bp polyadenosine anti-AgARR1 sera (1:5,000) or rabbit anti-actin sera (1:250, Sig- [(poly)Aϩ] tract. Multiple 3Ј RACE clones were sequenced and ma). Detection was accomplished with alkaline phosphatase- found to be identical throughout the region preceding their conjugated secondary Abs and stained with AP color substrate per (poly)Aϩ tract, supporting the hypothesis that the reported the manufacturer’s instructions (Bio-Rad). Duplicate immunoblots AgArr1 3Ј untranslated region sequence is complete. The ORF were prepared simultaneously for use with either antisera. is preceded by a 5Ј untranslated region of at least 259 bp, identical in the three independent 5Ј RACE clones that were Immunohistochemistry. Female A. gambiae adults were cold- sequenced. Although this result provides a high degree of anaesthetized, and heads plus antennae were dissected and fixed 30Ј confidence that the common 5Ј end represents the true start of in 4% paraformaldehyde͞PBS, washed in PBST (PBS͞0.2% the AgArr1 mRNA, we cannot exclude the possibility that the Tween-20), and cryoprotected in 12% sucrose overnight. Heads actual start is further upstream. The predicted AgARR1 protein and antennae were embedded in Tissue-Tek (Lab-Tek) and frozen has a molecular mass of 42.8 kDa and a pI of 8.0, consistent with in liquid nitrogen. Tissue was collected in 9-␮m sections onto all known arrestin subtypes. pretreated slides (VWR Scientific), dried at least 30 min, fixed in 4% paraformaldehyde͞PBS 30 min, and washed in PBST. Tissue Sequence Comparisons. A comparison of the deduced amino acid was blocked with PBSG5 (PBS, 0.2% Tween-20͞3% BSA͞5% sequence of AgARR1 with other invertebrate and all human normal goat serum) for 1 h under coverslips. Either preimmune or arrestin subtypes is shown in Fig. 1. The highest (72%) identity is affinity-purified polyclonal serum against full-length rAgARR1 seen with an arrestin isolated from an antennal cDNA library from was diluted 1:50 in PBSG5. Antisera were added under a coverslip the moth Heliothis virescens (HvARRH; ref. 23), followed by 68% and incubated overnight at 4°C in humid chambers. Detection was identity to the visual subtype DmARR1 (24) and 65% to CvARRA done with VectaStain ABC reagents (Vector Laboratories) per the (25), a visual arrestin from the blowfly Calliphora vicina. Following manufacturer’s instructions and visualized with either a 5-bromo- this cluster of homology, the identity between AgARR1 and other 4-chloro-3-indolyl phosphate͞nitroblue tetrazolium substrate invertebrate arrestins ranges from 40% to 48%. When compared to (Boehringer Mannheim) for a dark blue stain or with VectorRed all human arrestin subtypes, AgARR1 remains 38–42% identical. for a red precipitate (Vector Laboratories). Microscopy was per- The AgARR1-deduced amino acid sequence contains several formed by using an Olympus BX60 microscope connected to the domains implicated in arrestin function. These motifs include

1634 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.022505499 Merrill et al. Downloaded by guest on October 2, 2021 Fig. 1. Arrestin alignment. The CLUSTAL W alignment (12) of all invertebrate and human arrestins. Proteins included in alignment: AgARR1 (AY017417); HvARRH [P55274; (23)]; CvARRA [P51486; (25)]; DmARR1 [P15372; (24)]; DmARR2 [P19107; (37)]; DmiARRB [P19108; (38)]; LmARRH [P32122; (23)]; LpARRA [P51484; (39)]; DmKRZ [AAF32365; (17)]; HsARRC [XP010397; (40)]; HsARRS [P10523; (41)]; HsARRB1 [AAH03636; (18)]; and HsARRB2 [AAH07427; (18)]. Overall homology is indicated by the degree of shading where darker regions denote more conserved domains. Three characters are used above the alignment to mark strongly conserved positions: * indicates a single, fully conserved residue; : indicates one of these strong groups is conserved: STA; NEQK; NHQK; NDEQ; QHRK; MILV; MILF; HY; and FYW whereas a ⅐ indicates one of these weak groups is conserved: CSA; ATV; SAG; STNK; STPA; SGND; SNDEQK; NDEQHK; NEQHRK; FVLIM; and HFY. SH3, SH3 consensus-binding sequences; CB, clathrin-binding domain; arrows, AP-2 basic residues.

potential consensus Src homology 3 (SH3)-binding sites (Fig. 1) clonal Abs against full-length rAgARR1. Immunoblots (Fig. 2b) (26). Because these regions mediate binding of c-Src to nonvisual revealed that this serum recognizes an Ϸ45-kDa protein in crude arrestins (27), intriguing possibilities have been raised regarding extracts of both female and male A. gambiae antennae and in heads interactions with the mitogen-activated protein kinase pathway. As devoid of olfactory tissue. The apparent size of both the antennal is the case for most insect arrestins characterized to date, AgARR1 lacks a putative clathrin-binding domain (Fig. 1) thought to be important in aiding receptor endocytosis. Additionally, an interac- tion with the endocytic adaptor protein AP-2 also has been demonstrated for vertebrate nonvisual arrestins; this association was shown to require specific basic residues in the C terminus, some of which are present in AgARR1 (Fig. 1).

Tissue and Developmental Specificity of AgArr1 Expression. To con- firm that AgArr1 is present in olfactory tissues (i.e., antennae and maxillary palps) of A. gambiae, as well as to characterize its overall expression profile, RT-PCR analysis was carried out by using embryonic, larval, and pupal developmental life stages of the mosquito, as well as several adult tissues from both female and male A. gambiae (Fig. 2a). AgArr1 mRNA, absent from embryos, can be detected from early larval stages into mature adulthood. Exami- nation of adult tissues shows that AgArr1 mRNA is expressed in adult female and male olfactory appendages. No AgArr1 cDNA was detected in the bodies of either sex (confirmed by Southern blotting, data not shown), whereas substantial AgArr1 template also was detected in heads devoid of olfactory tissues. In these studies, the constitutively expressed ribosomal protein s7 (16) was used as a control for the level of template cDNA. Both AgArr1 and s7 primer Fig. 2. AgArr1 molecular analysis in A. gambiae.(a) RT-PCR analysis. AgArr1 is sets were designed to span an intron so that mRNA-derived expressed in early larval stages through mature adulthood in specific tissues. products could be distinguished from contaminating genomic DNA Antennal- and head-derived RNA demonstrates amplification of AgArr1 in both products (Fig. 2a). This analysis (confirmed by sequencing, data not female and male adult A. gambiae whereas RNA from adult bodies does not. shown) verifies the specific expression of AgArr1 in the developing Parallel reactions were performed for the constitutive ribosomal protein s7 gene as a template control. A, antennae; H, heads; B, bodies; *, genomic DNA carried organism as well as in adult olfactory tissues and heads devoid of over during RNA preparation. (b) Immunoblot. Polyclonal sera detects AgARR1 olfactory tissue. This result raised the prospect that AgArr1 encodes protein from crude protein extracts of adult female and male A. gambiae an arrestin with a broader spectrum of action than a gene with olfactory tissues (antennae and palps) or heads but not bodies. Parallel blots

expression exclusively in olfactory tissues. probed with sera against actin (Sigma, 1:250) is a positive control for loading and NEUROBIOLOGY AgArr1 expression was examined at the protein level with poly- transfer.

Merrill et al. PNAS ͉ February 5, 2002 ͉ vol. 99 ͉ no. 3 ͉ 1635 Downloaded by guest on October 2, 2021 Fig. 3. In vivo localization of AgArr1 in A. gambiae.Affinity purified polyclonal sera against rAgARR1 specifically labels antennal cell bodies (a and b). AgARR1 antisera also labels retinal photoreceptors (d); preimmune sera does not give comparable signal (c and e). c, cuticle; scb, subsensillar cell bodies; s, sensilla; and pr, photoreceptors. [Scale bars represent 25 ␮m(a–c)or10␮m(d–f).]

and head forms of AgARR1 conforms well to the 42.8-kDa predicted size of AgARR1. Little, if any, immunoreactivity was observed in extracts from male or female bodies that had been Fig. 4. Arrestin expression patterns in D. melanogaster.(a) RT-PCR analysis. stripped of all appendages or in parallel blots by using appropriate Considerable expression of DmArr1 and DmArr2 is noted in reactions performed preimmune serum controls (Fig. 2b and data not shown). Probing with antennal RNA; both are also highly detected in RNA derived from heads, duplicate blots with polyclonal antisera against a C-terminal frag- whereas little, if any, signal is seen in bodies. DmKrz mRNA is amplified from ment of actin (Sigma), which recognizes a 42-kDa actin protein in antennae, heads, and bodies. All reactions simultaneously amplified the ribo- A. gambiae, confirms that the lack of AgARR1 signal in bodies is somal protein 49 gene as an internal template control. A, antennae; H, heads; B, not because of insufficient material (Fig. 2b). bodies; *, genomic DNA carried over during RNA preparation. (b) Immunolocal- ization. Antisera against DmARR1 labels cells in maxillary palps (part 1) and Immunolocalization of AgARR1 Protein. To fully address the question antennae (part 5) of wild-type flies but not in mutant palps (part 2) or antennae (part 6) lacking DmArr1. Sera against DmARR2 also shows staining in wild-type of AgArr1 tissue specificity, immunolocalization studies were car- maxillary palps (part 3) but not in palps from DmArr2 mutants (part 4). ried out on cryosections prepared from heads and antennae of adult female A. gambiae. These studies directly establish the presence of AgARR1 protein in both olfactory and visual tissues (Fig. 3). expressed in both head and body tissues, consistent with its role in Indeed, significant staining for AgARR1 protein appears below the neurogenesis and fat body formation (17). Additionally, DmKrz cuticle in subsensillar cell bodies where primary olfactory neurons cDNA products are observed in antennal reactions, indicating that are expected to lie (Fig. 3 a and c). AgARR1 labeling appears it may also have an olfactory role (Fig. 4a). In all reactions, the mainly in cell bodies, presumably because of de novo protein ribosomal protein 49 gene was coamplified as an indicator of synthesis before transport to dendrites. The inability to detect relative template levels. AgARR1 in dendrites is most likely a limitation of light microscopy Immunolocalization of Drosophila visual arrestins by using Abs on the highly cuticularized antennal sensilla of A. gambiae, and against either DmARR1 or DmARR2 (20) specifically detected similar staining patterns also are observed for Drosophila arrestins these proteins in olfactory organs (Fig. 4b). In whole-mount stain- (see below). Considerable labeling of AgARR1 protein also is ing, DmARR1 and DmARR2 are localized to the ventrolateral observed in retinal tissue (Fig. 3e); this finding is consistent with the surface of maxillary palps (Fig. 4 b1 and b3), corresponding to the above RT-PCR and immunoblot results from head tissue and location of olfactory neurons (29). Reduced staining observed in implicates AgArr1 in visual transduction. The lack of substantial palps from flies mutant for DmArr1 (Fig. 4b2)orDmArr2 (Fig. 4b4) immunoreactivity in control reactions by using the appropriate demonstrates the specificity of these sera. Additionally, Abs against preimmune sera (Fig. 3 b and f) demonstrates that the localizations seen with anti-ARR1 sera are not spurious, but legitimate signals. DmARR1 label a subset of cells in antennal cryosections (Fig. 4b5); antisera for DmARR2 did not yield antennal staining (data not Arrestin Expression Patterns in D. melanogaster. The unexpected shown). Again, Drosophila carrying a mutant allele of DmArr1 show finding that AgArr1 is expressed in both visual and olfactory tissues little, if any, antennal immunoreactivity (Fig. 4b6). In a pattern of A. gambiae led us to speculate that dual sensory expression of similar to that observed for AgArr1 in mosquito antennae, the arrestins may be a characteristic of Dipteran insects. To test this Drosophila signals are distributed in a manner consistent with cell hypothesis, we conducted studies in the model system D. melano- bodies of olfactory neurons. gaster to determine whether the three Drosophila arrestins would also display overlapping sensory expression patterns and to inves- Arrestin Function in Drosophila Olfaction. Arrestin mutants in Dro- tigate the functional consequence of such a finding. RT-PCR sophila are defective in olfactory physiology, as measured in elec- analysis (Fig. 4a) demonstrates that DmArr1 and DmArr2,previ- tropalpogram (EPG) and electroantennogram (EAG) recordings, ously only shown to be visual in function (24, 28), are also detectable which detect the summed potentials of olfactory receptor neurons in RNA prepared from olfactory organs. Robust signals also are in the maxillary palp and antenna, respectively. The Drosophila observed from head tissue, owing to their presence in the photo- arr12 mutant, a missense mutation resulting in Ͻ1% of protein transduction system (Fig. 4a). As before, primers spanned an intron levels (20), shows a reduction of Ϸ60% in EPG amplitude to the to distinguish cDNA-derived bands from genomic contamination odor of butanol (BU) compared to a control strain (Fig. 5a). To by the size of the expected product. In this case, the larger bands determine whether the defect maps to the DmArr1 locus, we tested observed in DmArr1 and DmArr2 reactions (asterisks, Fig. 4a) arr12 in heterozygous condition with a deficiency chromosome that represent amplification of genomic DNA as confirmed by sequenc- uncovers the DmArr1 locus and found a similarly reduced ampli- ing (data not shown). Lastly, the nonvisual DmKrz arrestin is tude. However, an arr12͞ϩ heterozygote also shows a strong

1636 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.022505499 Merrill et al. Downloaded by guest on October 2, 2021 regulation of olfactory signaling in A. gambiae. Although undetect- able in embryonic stages, AgArr1 is expressed early in larval development through mature adulthood, suggesting that this gene may function in sensory or other pathways throughout the mos- quito’s life cycle. A surprising discovery during the course of these studies was the detection of AgArr1 in retinal tissues, a finding that suggests AgArr1 could be a more broadly distributed arrestin, which participates in the desensitization of several signaling pathways. This notion led us to question whether expression across multiple sensory systems is typical among insect arrestins or is unique to AgArr1. To address this, we examined several arrestins in a well-characterized insect system, D. melanogaster. Our hypothesis that visual arrestins also may be involved in olfactory responses is supported by the expres- sion of two visual arrestins DmArr1 and DmArr2,aswellasthe nonvisual DmKrz arrestin, in the antennae. Importantly, a similar overlap between olfactory and visual systems was previously sug- gested as phospholipase C (norpA) and a phosphatidylinositol transfer protein (rdgB)ofD. melanogaster are required for normal performance by both sensory systems (19, 21). Fig. 5. Olfactory electrophysiology in Drosophila arrestin mutants. (a) arr12 Although both insects demonstrate dual sensory expression of ϫ Ϫ3 reduces EPG amplitude in response to 0.5-s pulses of BU, diluted 2 10 in arrestins, only Drosophila offers convenient genetic tools to inves- paraffin oil, measured as described in ref. 19. Df refers to Df(2L)I131 nub b pr. The deficiency uncovers the DmArr1 locus at 36E3 (20). Dp refers to Dp(2; Y)H3 tigate functional significance. We have shown that Drosophila covering the DmArr1 locus (22). ϩ͞Ϯ Refers to Canton S, which was used arrestin mutants exhibit a decrease in the amplitudes of electro- because the parental chromosome for the DmArr1 mutants was not available. physiological responses to olfactory stimuli. Similar amplitude For each genotype, n ϭ 5 flies. Flies were tested 10 times each; mean amplitude deficits can also be seen in arrestin mutant visual system responses per fly was computed, and a mean for each genotype was calculated. Signif- (20). Although the well-established role of arrestins in receptor icance was calculated by ANOVA. Error bars indicate SEM (b) EPG amplitudes desensitization (30) might suggest that arrestin mutations would be of arr11 tested with 2 ϫ 10Ϫ3 BU, n ϭ 5 flies. (c) EPG amplitudes of arr12 tested Ϫ expected to lead to some kinetic-signaling impairment, such an with a 1 ϫ 10 2 dilution of EA, n ϭ 5 flies. (d) EAG amplitude of arr12; arr23 effect was not detected in these studies. To examine this effect, ϫ Ϫ3 ϭ 2 tested with a 2 10 dilution of BU, n 5 flies. (e) EAG amplitude of arr1 ; more sensitive means will be required including single-unit elec- arr23 tested with a 2 ϫ 10Ϫ2 dilution of EA, n ϭ 6. trophysiology (31) or patch clamp recordings (32). It will be of interest to use these methods to examine the olfactory physiology amplitude reduction, suggesting that the arr12 mutation is at least in arrestin mutants to determine the molecular mechanism(s) that partially dominant. We therefore introduced a duplication covering underlie these results, as several explanations seem plausible. 2 Faulty desensitization may elevate basal receptor activity such the arr1 locus and showed that it partially rescued the phenotype, ͞ providing evidence that the EPG defect does in fact map to the that the EPG EAG amplitude elicited from this higher baseline in DmArr1 locus. As further evidence that the EPG defect maps to mutants would appear lower than the amplitude elicited from flies DmArr1, we tested another allele, arr11, a 5-kb insertion in the with wild-type receptor activity levels. Another possibility concerns the role of arrestins in GPCR internalization and resensitization (4). second intron (20), and found that this mutation, which causes a If receptor recycling is compromised by arrestin mutations, then 90% reduction in protein levels, shows a 40% reduction in ampli- reduced levels of available resensitized receptors would lower the tude, either in homozygous or heterozygous condition with the amplitude of response to new stimuli. Recent studies also have deficiency chromosome (Fig. 5b). The reduced EPG phenotype is found a role for arrestins in mitogen-associated protein kinase not restricted to BU; the response to ethyl acetate (EA) is also 2 2͞ signaling pathways by means of SH3 domain-containing proteins affected as shown for arr1 and arr1 Df in Fig. 5c. Mutations in the (27); if such a transduction cascade is somehow required to main- other visual arrestin gene, DmArr2, showed no detectable alter- tain olfactory receptor expression or function at the membrane, ations in EPG recordings. then loss of arrestins would indirectly affect receptor responsiveness We then asked whether arrestin function also is required for by the loss of this feedback regulation. Although arrestin mutations correct olfactory physiology in the antenna. Although we did not have been shown to cause neuronal degeneration in the visual detect EAG abnormalities when either DmArr1 or DmArr2 alleles system (20, 33, 34), scanning electron microscopy and histological were tested singly, double mutants showed significant reductions in analysis have not revealed degeneration at a gross level in olfactory 2 3 EAG amplitude. Specifically, arr1 ;arr2 exhibited a reduced am- tissues (data not shown). We note that rdgB mutations of Drosophila plitude of response when tested either with BU (Fig. 5d)orEA(Fig. affect the physiology of both the visual and olfactory systems, but 3 Ͻ 5e). The arr2 allele is considered one of the strongest because 1% degeneration is seen only in the visual system (21, 35). The simplest 3 of DmARR2 is detected in arr2 mutants (20). As a test of whether interpretation of the data remains that both DmArr1 and DmArr2 1 this effect maps to the DmArr2 locus, we also tested arr2 ,an are required for normal olfactory physiology. independently isolated missense allele of DmArr2 that results in We have found defects in the physiology of the maxillary palp in Ϸ80% of wild-type protein levels (20) and found that arr12; mutants of DmArr1, but not DmArr2. Because both genes are arr21͞arr23 mutants also were defective in response to both odors expressed in the maxillary palp it is plausible that our inability to (not shown). These results suggest that DmArr1 and DmArr2 are detect a functional deficit for DmArr2 in this organ reflects limi- both required for normal olfactory physiology. tations in the sensitivity of our assay, the nature of the available DmArr2 alleles, or a form of redundancy between DmArr1 and Discussion DmArr2. Genetic redundancy appears to occur in the antenna, We have isolated AgArr1, an arrestin from the malaria vector where EAG defects are observed in double, but not single mutants. mosquito, A. gambiae and have directly observed its expression in This redundancy could arise from the ability of either arrestin to act

regions of the antenna where olfactory neurons are presumed to upon the same odor receptor. Alternatively, each arrestin may NEUROBIOLOGY reside. These findings support a potential role of AgArr1 in the associate with a different subset of odor receptors, with mutations

Merrill et al. PNAS ͉ February 5, 2002 ͉ vol. 99 ͉ no. 3 ͉ 1637 Downloaded by guest on October 2, 2021 of a single arrestin affecting the signaling of only some receptors. A Finally, this study identifies a dual sensory modality arrestin from recent study (31) of coding in the antenna has revealed at least three an insect disease vector of significant medical importance; further- types of neurons that respond to ethyl acetate with distinct spec- more, this study uses Drosophila as a model system to demonstrate trums, each presumably expressing a different receptor. Perhaps that arrestins previously thought to act solely in phototransduction mutations of a single arrestin affects the signaling of a subset of are not only expressed in other sensory systems but also are neurons, and our assay is only sensitive enough to detect the required for proper olfactory function. As the arrestin from A. combined loss of neurons associated with DmArr1 and DmArr2.In gambiae shows considerable homology and comparable expression this context, we note that our inability to detect DmArr2 immuno- patterns to the Drosophila genes, it is likely that AgArr1 may play a histochemically in the antenna could reflect either a low level of similar role in mosquito physiology. In light of the broad distribu- expression in individual neurons or a small number of neurons in tion and likelihood of involvement in a range of odorant respon- which it is expressed. siveness, arrestin genes form a particularly attractive target for This study describes visual arrestins that also show expression interference of the olfactory system. In view of the critical role for and function in nonvisual (i.e., olfactory) tissues in Dipteran olfaction in mosquito host choice and vectorial capacity (10), it is insects. In addition to previous reports that certain elements involved in activation are common across multiple sensory tempting to suggest that experimental approaches designed to alter systems (19, 21), these findings support the idea that components olfactory genes such as AgArr1 could lead to a reduction in the of desensitization are also shared between visual and olfactory ability of mosquitoes to effectively transmit malaria, West Nile pathways. One of the more interesting results in this study is the encephalitis, and other globally important medical and veterinary demonstration that although these elements are shared between diseases. pathways, they may be functioning in different manners in each system; olfactory phenotypes display primarily an amplitude We thank the entire faculty and staff at the European Molecular Biology deficit whereas the effects on the visual system appear to have Laboratory (Heidelberg, Germany) where this work was initiated as well as both kinetic and amplitude components (20). It will be intriguing A. Nicole Fox and other members of the Zwiebel laboratory for helpful to determine the way in which these proteins are functioning in discussions. We thank Peter Clyne and Audrey Hing for carrying out each sensory modality. Moreover, this study demonstrates that supplementary experiments and for discussion. We thank Dr. Claudio Pikielny (University of Medicine and Dentistry of New Jersey–Robert arrestins are required for invertebrate olfactory signaling. There Wood Johnson Medical School, Piscataway, NJ) for comments on this is considerable evidence for nonvisual arrestins acting in verte- ␤ manuscript as well as assistance with subtractive hybridization and Dr. brate olfactory signal transduction pathways. Abs against -ar- Aileen McAinsh for editorial assistance. This investigation received finan- restin-2 have been used to label rat olfactory epithelium and cial support from the United Nations Development Program͞World Bank͞ ␤ incubating -arrestin-2 Abs with olfactory cilia preparations World Health Organization Special Program for Research and Training in results in a loss of odorant-induced desensitization (and in- Tropical Diseases (Grant 980697 to L.J.Z.) and grants from the National creases in cAMP levels), supporting the idea that vertebrate Institutes of Health͞National Institute of Deafness and Other Communi- nonvisual arrestins control olfactory desensitization (36). cation Disorders (DC02174–16 to J.R.C. and DC04692–01 to L.J.Z.).

1. Reed, R. (1992) Neuron 8, 205–209. 21. Riesgo-Escovar, J. R., Woodard, C. & Carlson, J. R. (1994) J. Comp. Physiol. 2. Boekhoff, I., Michel, W. C., Breer, H. & Ache, B. W. (1994) J. Neurosci. 14, A 175, 687–693. 3304–3309. 22. Lindlsey, D. & Zimm, G. (1992) The Genome of Drosophila melanogaster 3. Pippig, S., Andexinger, S., Daniel, K., Puzicha, M., Caron, M. G., Lefkowitz, (Academic, London). R. J. & Lohse, M. J. (1993) J. Biol. Chem. 268, 3201–3208. 23. Raming, K., Freitag, J., Krieger, J. & Breer, H. (1993) Cell Signal. 5, 69–80. 4. Ferguson, S. S., Downey, W. E., 3rd, Colapietro, A. M., Barak, L. S., Menard, 24. Smith, D. P., Shieh, B. H. & Zuker, C. S. (1990) Proc. Natl. Acad. Sci. USA 87, L. & Caron, M. G. (1996) Science 271, 363–366. 1003–1007. 5. Craft, C. M. & Whitmore, D. H. (1995) FEBS Lett. 362, 247–255. 25. Plangger, A., Malicki, D., Whitney, M. & Paulsen, R. (1994) J. Biol. Chem. 269, 6. Kiselev, A. & Subramaniam, S. (1997) Biochemistry 36, 2188–2196. 26969–26975. 7. Gurevich, V. V. (1998) J. Biol. Chem. 273, 15501–15506. 26. Cohen, G., Sen, R. & Baltimore, D. (1995) Cell 80, 237–248. 8. Gurevich, V. V., Dion, S. B., Onorato, J. J., Ptasienski, J., Kim, C. M., Sterne-Marr, 27. Luttrell, L. M., Ferguson, S. S., Daaka, Y., Miller, W. E., Maudsley, S., Della R., Hosey, M. M. & Benovic, J. L. (1995) J. Biol. Chem. 270, 720–731. Rocca, G. J., Lin, F., Kawakatsu, H., Owada, K., Luttrell, D. K., et al. (1999) 9. Hildebrand, J. G. & Shepherd, G. M. (1997) Annu. Rev. Neurosci. 20, Science 283, 655–661. 595–631. 28. Hyde, D., R., Mecklenburg, K., L., Pollock, J., A., Vihtelic, T., S. & Benzer, S. 10. Takken, W. & Knols, B. G. (1999) Annu. Rev. Entomol. 44, 131–157. (1990) Proc. Natl. Acad. Sci. USA 87, 1008–1012. 11. Lavery, D. J., Lopez-Molina, L., Fleury-Olela, F. & Schibler, U. (1997) Proc. 29. Singh, R. & Nayak, S. (1985) Int. J. Insect Morphol. Embryol. 14, 291–306. Natl. Acad. Sci. USA 94, 6831–6836. 30. Krupnick, J. G. & Benovic, J. L. (1998) Annu. Rev. Pharmacol. Toxicol. 38, 12. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990) J. 289–319. Mol. Biol. 215, 403–410. 31. de Bruyne, M., Foster, K. & Carlson, J. R. (2001) Neuron 30, 537–552. 13. Chenchik, A., Diachenko, L., Moqadam, F., Tarabykin, V., Lukyanov, S. & 32. Dubin, A. E. & Harris, G. L. (1997) J. Neurobiol. 32, 123–137. Siebert, P. D. (1996) BioTechniques 21, 526–534. 33. Alloway, P. G., Howard, L. & Dolph, P. J. (2000) Neuron 28, 129–138. 14. Devereux, J., Haeberli, P. & Smithies, O. (1984) Nucleic Acids Res. 12, 34. Kiselev, A., Socolich, M., Vinos, J., Hardy, R. W., Zuker, C. S. & Ranganathan, 387–395. R. (2000) Neuron 28, 139–152. 15. Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994) Nucleic Acids Res. 22, 35. Woodard, C., Alcorta, E. & Carlson, J. (1992) J. Neurogenet. 8, 17–31. 4673–4680. 36. Dawson, T. M., Arriza, J. L., Jaworsky, D. E., Borisy, F. F., Attramadal, H., 16. Salazar, C. E., Mills-Hamm, D., Kumar, V. & Collins, F. H. (1993) Nucleic Acids Lefkowitz, R. J. & Ronnett, G. V. (1993) Science 259, 825–829. Res. 21, 4147. 37. Yamada, T., Takeuchi, Y., Komori, N., Kobayashi, H., Sakai, Y., Hotta, Y. & 17. Roman, G., He, J. & Davis, R. L. (2000) Genetics 155, 1281–1295. Matsumoto, H. (1990) Science 248, 483–486. 18. Strausberg, R. L., Feingold, E. A., Klausner, R. D. & Collins, F. S. (1999) 38. Krishnan, R. & Ganguly, R. (1990) Nucleic Acids Res. 18, 5894. Science 286, 455–457. 39. Smith, W. C., Greenberg, R. M., Calman, B. G., Hendrix, M. M., Hutchinson, 19. Riesgo-Escovar, J., Raha, D. & Carlson, J. R. (1995) Proc. Natl. Acad. Sci. USA L., Donoso, L. A. & Battelle, B. A. (1995) J. Neurochem. 64, 1–13. 92, 2864–2868. 40. Craft, C. M., Whitmore, D. H. & Wiechmann, A. F. (1994) J. Biol. Chem. 269, 20. Dolph, P. J., Ranganathan, R., Colley, N. J., Hardy, R. W., Socolich, M. & 4613–4619. Zuker, C. S. (1993) Science 260, 1910–1916. 41. Yamaki, K., Tsuda, M. & Shinohara, T. (1988) FEBS Lett. 234, 39–43.

1638 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.022505499 Merrill et al. Downloaded by guest on October 2, 2021