Identification of preferentially expressed cochlear genes by systematic sequencing of a rat cochlea cDNA library

Molecular Brain Research 47 Ž1997. 1–10

Research report

Identification of preferentially expressed cochlear genes by systematic sequencing of a rat cochlea cDNA library Ana Soto-Prior ) , Mireille Lavigne-Rebillard, Marc Lenoir, Chantal Ripoll, Guy Rebillard, Philippe Vago, Remy ´ Pujol, Christian P. Hamel INSERM U254 and UniÕersites Saint Charles, 34295 Montpellier Cedex ´ de Montpellier 1 et 2, CHU Hopital ˆ ´ 5, France Accepted 26 November 1996

Abstract 107 expressed sequence tags ŽESTs. from a rat cochlea cDNA library were identified by systematic sequencing coupled to database selection and RT-PCR analysis of novel sequences. This approach led us to select a clone, pCO8, showing no significant homology with any database sequence, that corresponds to a mRNA whose expression is restricted to the cochlea, except for traces detected in brain. Additional clones with novel sequences enriched in the cochlea were also found. ESTs bearing significant homologies with database sequences Ž63 out of 107. were classified according to the putatively encoded protein. They include tissue-specific genes not previously described in the cochlea as well as known genes from other species. We performed in situ hybridization in cochlear tissues to localize the pCO8 mRNA and that of clone pCO6 which is 100% homologous to the delayed rectifier potassium channel drk1. We found that both mRNAs were exclusively expressed in the cellular body of the primary auditory neurons from the spiral ganglion of the cochlea. These results indicate that this approach is an efficient way to identify novel genes that could be of importance in cochlear function. q 1997 Elsevier Science B.V. All rights reserved. Keywords: Cochlea cDNA library; Systematic sequencing; RT-PCR; Hybridization, in situ; Novel gene; pCO8; drk1

1. Introduction The mammalian auditory organ, the cochlea, is composed of highly specialized and precisely assembled tissues which provide proper sensitivity and selectivity to stimulatory sounds Žsee w36,43x.. In the cochlea, the auditory message originates in the organ of Corti, a complex epithelial structure comprised of two types of sensory hair cells Žinner and outer hair cells. and various types of supporting cells. Inner and outer hair cells are connected to type I and type II primary auditory neurons, respectively, whose cellular bodies form the spiral ganglion of the cochlea. The mechano-electrical transduction occurring in hair cells is dependent on a high potassium concentration in the endolymph, generated by another specialized epithelium, the stria vascularis. Given the unique features of

) Corresponding author. INSERM U254, Neurobiologie de l’AuditionPlasticite´ Synaptique, CHU Hopital Saint Charles, 34295 Montpellier ˆ Cedex 5, France. Fax: q33 6752-5601. ´

these tissues and the complexity of the auditory mechanisms, it can be anticipated that many genes are preferentially or specifically expressed in the cochlea. Recent reports support this hypothesis. Several genes, including some homeobox genes, thought to play a role in cochlear development, appear to be predominantly expressed in the cochlear progenitor cells w17,37x. The relatively high incidence Ž1 out of a 1000. of children born with a hearing impairment and the finding that mutations at many different loci in both mice and human cause deafness w49x, are also suggestive of genes exerting specific roles in the cochlea. However, very little is known of the molecular mechanisms responsible for the uniqueness of the cochlea, in part due to the limited amount of cochlear tissues. In fact, only a few studies have successfully addressed the m o le c u la r c o m p o sitio n o f c o c h le a r tissu e s w12,22,32,38,39,41x. Partial sequencing of randomly selected clones directly from a cDNA library has been shown to be an excellent way of identifying new genes and describing the transcriptional activity of a tissue or cell line w1,2,9,22,33x. Re-

0169-328Xr97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 9 - 3 2 8 X Ž 9 7 . 0 0 0 3 3 - 8

2

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

cently, cDNA libraries have been constructed either from the entire cochlea w24,32,39x or from the organ of Corti w55x, allowing the identification of cochlea-specific genes. In an attempt to gain rapid access to novel cochlear genes implicated in auditory function and to simplify the process of isolating genes, we used the approach of systematic sequencing of a rat cochlea cDNA library. Randomly isolated cDNA clones from the library were partially sequenced and compared to database sequences. The expression pattern of the unknown genes was analyzed by RTPCR. We report the isolation of novel sequences preferentially expressed in the cochlea, as well as those of several known genes that we have classified, a number of which encode structural proteins or enzymes previously associated with hearing loss.

2. Materials and methods 2.1. Construction of a rat cochlear cDNA library 100 cochleas from 24-day-old Wistar rats were dissected out. Cochleas were homogenized with a potter homogenizer in a denaturing solution containing 4 M guanidine thiocyanate ŽFluka, Switzerland., 25 mM sodium citrate pH 7.0, 0.5% N-lauryl sarcosine, 0.1 M 2-mercaptoethanol and RNA extracted according to the single-step method of Chomczynski and Sacchi w13x. PolyŽA.q RNAs were selected from the total RNA obtained Ž440 m g. and used to construct a directional cDNA library in lZAP II ŽStratagene, USA.. The cDNA library, which contained - 1% non-recombinant phages in 5.8 = 10 6 pfusrml, was amplified to a titer of 3 = 10 10 pfusrml. 2.2. Isolation of cDNA clones XL1 Blue E. coli were infected with the library phages and plated out at low density to allow for separation of individual plaques. To evaluate the size of inserts, 54 plaques were randomly isolated, inserts were PCR-amplified using T3 and T7 primers and the product from each clone was run on a 1% agarose gel. Subsequently, 107 individual plaques, including the 54 PCR-amplified ones, were randomly cored, the phagemid ‘in vivo excised’ ŽStratagene. and purified. 2.3. Sequence analysis Sequences at the 5X-end of the coding strand of the cloned cDNAs were obtained by manual sequencing using PUCrM13 reverse primer. Partial nucleotide sequences of clones were compared to the nucleotide sequences deposited in Genbank andror EMBL using the Fasta programme w35x. Clones showing no significant homology with known cDNAs were selected, and their tissue expres-

sion patterns evaluated by RT-PCR. The cellular localization of cDNAs expressed within cochlear tissues was subsequently analyzed by in situ hybridization. 2.4. Tissue expression analysis by RT-PCR Total RNAs from various tissues – brain, cerebellum, eye, lung, kidney, cochlea and liver – were isolated as described above. The cDNA was synthesized by random priming using 2 m g of total RNA from each tissue and 9.6 U of a MuMLV reverse transcriptase ŽEurogentec, Belgium. in a 10-m l volume. 1r10 of the RT product volume was amplified by PCR in a 20-m l volume, using 20-merspecific primers for each clone, as follows: 958Cr2 min followed by 35 cycles Ž958Cr30 s, Tm -38Cr30 s, 728Cr1 min. and 728Cr5 min. 35 PCR cycles are sufficient to detect cDNA fragments from 20 fg of a specific low-abundance mRNA in 2 m g of total RNA on an ethidium bromide-stained gel w30x. In these conditions, abundant mRNAs will lead to early saturated amplified cDNAs preventing proper relative estimate of the expression levels of these mRNAs. However, an intense cDNA amplification from only one or a few tissues is enough to assign tissue-predominant expression to the highly amplified, corresponding RNA, providing that cDNAs from ubiquitously expressed mRNAs could be amplified in all tissues. This latter requirement was done by amplifying a 176-bp fragment from the rat ATPase subunit 6 ŽGenbank accession number M27315. with forward Ž5X-AGAAGGGTGAATACATAGGC-3X . and reverse Ž 5X -CGACTAA CAGCAAACATTAC-3X . primers using the same PCR programme. For each clone, the absence of contaminating cDNA in the buffers and reagents was verified in a PCR reaction omitting the cDNA. To ensure that RNA did not contain any detectable genomic DNA, RNAs from each tissue were subjected to PCR amplification without previous reverse transcription. This latter control was performed for each of the 5 clones exhibiting a differential tissue distribution. The presence of the desired PCR product was ascertained by electrophoresis on an ethidium bromidestained 4% agarose gel in 1 = TAE. 2.5. Cochlear expression analysis by in situ hybridization 2.5.1. Tissues Rats were deeply anesthesized with pentobarbital Ž60 mgrkg. and perfused via the aortic arch with 50 ml of 0.1 M sodium phosphate buffer pH 7.4, followed by 200 ml of cold Ž48C. 4% paraformaldehyde in the same buffer. The cochleas were dissected from the temporal bones, postfixed at 48C for 1 h in 4% paraformaldehyde and immersed for 5 min in the decalcifying solution DC3 ŽLabonord, France.. The eyes and brains were similarly processed except for the decalcification step. Then, the tissues were incubated overnight in the phosphate buffer

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

containing 20% sucrose, placed in OCT compound ŽMiles, USA., frozen and sectioned at 15 m m on a Reichert-Jung 2800 cryostat. The sections were mounted on gelatin-coated slides and either used immediately or stored at y708C until use.

3

the DIG nucleic acid detection kit ŽBoehringer Mannheim, Germany. according to the manufacturer’s protocol. Sections were examined with a Reichert Polyvar microscope.

3. Results 2.5.2. Probes Linearized drk1 and pCO8 cDNA templates were used to synthesize digoxigenin-labeled sense Žlinearized with XhoI. and antisense Žlinearized with EcoRI. riboprobes according to the manufacturer’s protocol ŽBoehringer Mannheim, Germany.. The length of drk1 and pCO8 riboprobes was 2900 and 426 nucleotides, respectively. The drk1 riboprobe was reduced to 400-nucleotide fragments by alkaline hydrolysis w4x. The probes were finally denatured at 858C for 2 min immediately before hybridization. 2.5.3. Hybridization procedure The tissue sections were permeabilized with 15 m grml of Proteinase K ŽBioprobe Systems, France. in 0.1 M Tris–HCl pH 8.0 and 50 mM EDTA at 378C for 15 min, post-fixed with 4% paraformaldehyde, acetylated, soaked at room temperature in the pre-hybridization buffer Ž4 = SSC, 1 = Denhart’s solution, 1% sodium N-lauryl sarcosine. for 1 h, dehydrated and air-dried w42x. They were then incubated in 20 m l of hybridization buffer Ž50% deionized formamide, 4 mM EDTA, 0.2% sodium N-lauryl sarcosine, 600 mM NaCl, 0.05% disodiumpyrophosphate, 0.05% tetrasodiumpyrophosphate, 80 mM Tris–HCl pH 7.5. containing 4 to 20 ngrm l of the labeled riboprobe. Cochlear sections were hybridized overnight at 508C in a moist chamber, washed at room temperature in 4 = SSC for 15 min and then treated with ribonuclease A Ž100 m grml, 30 min at 378C. in 2 = SSC. They were then successively washed in 2 = SSC for 1 h at room temperature, in 0.1 = SSC for 15 min at 508C and again in the same buffer for 10 min at room temperature. They were finally processed for digoxigenin immunodetection using

In order to evaluate the insert sizes of our amplified rat cochlea cDNA library, we initiated a pilot study by random selection of 54 clones. The sizes of the inserts were between 400 and 4300 nucleotides, among which 1.8% were - 0.5 kb, 27.7% were 0.5–1 kb and 70.4% were ) 1 kb. 3.1. Classification and database selection of sequenced clones 107 clones were randomly isolated. On average, 130 nucleotides of sequence were obtained from each clone. A comprehensive view of the strategy used and of the results obtained is illustrated in Fig. 1. The nucleotide sequences were analyzed and compared to those in Genbank and EMBL databases. Clones showing either nucleotide identities to database sequences ) 85% or FASTA scores ) 150 were considered as identified genes; those showing nucleotide identities - 85% or scores - 150 were classified as unknown genes. Following this first screening, 63 clones Ž58.8%. were classified as identified genes and 40 clones Ž37.4%. represented unknown rat genes. Four clones were not interpretable due to a poly-T tail at both the 3X- and 5X-ends of the insert. Among the 63 identified clones, 58 corresponded to previously characterized genes and 5 to sequence tags from various tissues that we classified as uncharacterized genes Žsee Table 1 for summary of results.. Among the 58 characterized clones, 37 corresponded to previously described rat genes and 21 were homologous to genes from other species, including mouse, bovine, monkey and human, therefore, probably representing their rat counterpart

Fig. 1. Distribution and expression analysis of clones obtained from the systematic sequencing of the rat cochlea cDNA library.

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

4

Table 1 Rat cochlea ESTs matched to known genes in GenbankrEMBL databases Clone

Accession number

Putative identified protein and its accession number

Metabolism pCO3 AA108320 ATP synthase g-subunit ŽgbL19927. pCO9 AA108321 Mitochondrial genome, ORF 41 ŽgbJ01435. pCO21 AA108322 UV-damaged DNA-binding protein ŽemL20216. pCO24 AA108323 Mitochondrial genome, ORFa6 ŽgbM27315. pCO30 AA108262 Protective protein ŽgbJ05261. pCO32 AA108263 Cytochrome oxidase subunit II ŽgbJ01435. pCO38 AA108264 Stearyl-CoA desaturase ŽgbJ02585. pCO41 AA108265 Naq, Kq-ATPase a Žq. ŽgbM14512. pCO44 AA108266 NADq-dependent isocitratedehydrogenase ŽemU07980. pCO46 AA108267 Peroxisomal 3-ketoacyl-CoA thiolase ŽgbD90055. pCO55 AA108268 Prostaglandin-H-2 D-isomerase ŽgbM94134. pCO62 AA108269 Histone-1 ŽgbM29260. pCO75 AA108270 Ubiquinone oxidoreductase ŽemT07882. pCO78 AA108271 L-Arg: Gly amidinotransferase ŽgbU07971. pCO86 AA108272 Cytochrome oxydase III ŽgbJ01435. pCO89 AA108273 Aldolase A ŽgbM12919. pCO91 AA108274 Carbonyl reductase ŽgbX84348. pCO92 AA108275 a-Globin ŽgbM17083. pCO96 AA108276 a-Globin ŽgbM17083. pCO100 AA108277 Heat-shock protein Žhsp-E7I. ŽgbL40406. pCO102 AA108278 Cytochrome b ŽgbJ01436. pCO107 AA108279 Cytochrome b ŽgbX14848. pCO113 AA108280 Cytochrome b ŽgbX14848. pCO123 AA108281 Phosphatase 2A-b protein ŽemM23591. Ž S .-Malate NADPq oxidoreductase ŽemM26594. pCO130 AA108282 pCO133 AA108283 b-Globin ŽemX67613. pCO136 AA108284 Carbonic anhydrase III ŽemM22413. Cell signaling and transporters pCO6 AA108285 Drk1 potassium channel ŽgbX16476. pCO42 AA108286 Guanine nucleotide regulatory protein ŽemU01147. pCO49 AA108287 Chloride channel RCL1 ŽgbD13985. pCO51 AA108288 Ras-related protein ŽgbJ02998. pCO52 AA108289 Stathmin ŽgbJ04979. pCO63 AA108290 Calretinin ŽgbX66974. pCO87 AA108291 Mitochondrial Hqrphosphate symporterŽgbM23984. pCO93 AA108292 ADP ribosylation factor ŽARF.-like protein ŽgbX16476. pCO112 AA108293 a-Platelet-derived growth factor receptor ŽgbM63837. pCO120 AA108294 14-3-3 Protein b-subtype ŽemD17446. pCO121 AA108295 GDP-dissociation inhibitor ŽemX79353. Structural pCO43 AA108296 Schwann cell peripheral myelin protein ŽP0. gbK03242 pCO53 AA108297 Myelin proteolipid protein ŽgbM11185. pCO70 AA108298 Microtubule-associated protein 1A ŽgbM83196. pCO82 AA108299 b-Tubulin ŽgbM28730. pCO85 AA108300 a-Collagen ŽgbU16789. pCO90 AA108301 b-Tropomyosin ŽgbL00372. pCO116 AA108302 Peripheral myelin protein ŽP0. ŽemM62857. pCO129 AA108303 Cell-binding bone sialoprotein ŽemJ04215. Transcription factors and translation machinery pCO27 AA108304 Ribosomal protein L7 ŽgbM17422. pCO29 AA108305 Primase large subunit ŽgbD13545. pCO50 AA108306 Transcription factor P45 NF-E2 ŽgbL09600. pCO94 AA108307 Ribosomal protein L7 ŽgbM17422. pCO97 AA108308 DNA-binding protein mdm2 ŽgbX58876. pCO104 AA10830 9 Ribosomal protein S10 ŽgbX13549. pCO109 AA108310 Testis-specific transcription elongation factor ŽemD12927. Neuron-specific pCO16 AA108311 Neuron-specific MAP kinase ŽgbL35236. pCO71 AA108312 Latexin ŽgbX76985.

Species

Score

%

Nt

Rat Rat Mon Rat Mou Rat Rat Rat Bov Rat Rat Mou Hum Rat Rat Rat Rat Rat Rat Mou Rat Rat Rat Rat Rat Rat Rat

272 453 365 453 303 543 222 274 358 301 614 233 302 454 862 685 613 176 729 446 448 125 733 117 641 421 476

97.3 99.1 93.5 93.5 87.9 97.2 84.1 97.2 83.8 97.4 98.7 79.1 76.8 91.6 98.2 93.5 100 96.1 99.5 86.8 98.3 91.1 96.0 96.8 98.2 95.9 89.6

72 115 106 132 115 114 83 76 134 79 157 103 127 149 225 213 183 50 185 158 118 42 200 38 168 123 205

Rat Hum Rat Rat Rat Rat Rat Rat Rat Rat Hum

296 342 153 323 371 320 884 468 620 526 246

100 93.8 80.0 89.5 94.2 92.5 96.7 99.2 100 96.6 85.1

74 96 61 130 122 137 241 215 155 145 142

Rat Rat Rat Mou Mou Rat Hum Rat

316 364 664 116 178 673 404 290

100 98.9 100 87.5 95.8 98.9 95.5 88.9

97 97 166 39 55 173 127 100

Rat Mou Mou Rat Mou Rat Rat

230 198 309 557 320 609 482

94.2 85.1 87.6 98.0 78.9 96.9 94.0

93 118 107 147 186 161 145

Mou Rat

409 441

91.8 97.5

125 118

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

5

Table 1 Žcontinued. Clone

Accession number

Secreted factors pCO80 AA108313 pCO95 AA108314 Cell surface pCO106 AA108315 Not characterized pCO4 AA108316 pCO79 pCO81 pCO118 pCO122

AA108317 AA108318 AA108319 AA109476

Putative identified protein and its accession number

Species

Score

%

Nt

Chromogranin B ŽgbX53028. Apolipoprotein D ŽgbX55572.

Mou Rat

280 526

87.5 90.3

110 165

Synaptophysin ŽgbX06388.

Rat

425

97.4

121

EST from embryonal carcinoma F9 cell cDNA, 82E07 emD21724. X cDNA clone 69812 5 EST ŽemT49922. X cDNA clone 74208 3 EST ŽemT48384. mRNA for ORF ŽemD26068. cDNA clone HEA35T EST ŽemZ36286.

Mou

184

85.9

85

Hum Hum Hum Hum

300 306 347 410

80.6 93.8 82.3 91.7

117 126 133 166

Classification of the 63 identified sequences Žout of 107 randomly sequenced clones. from the rat cochlea cDNA library. Database searches were done in Genbank Žgb. and EMBL Žem.. Species are mok for monkey, bov for bovine, mou for mouse and hum for human. Scores were established using the FASTA programme Žsee Materials and methods.. Percentages of identity Ž%. and nucleotide overlap Žnt. between clone and identified sequences are indicated.

genes. Among the 63 identified clones, three bore the same mitochondrial cytochrome b sequences, two coded for the ribosomal protein L7 and two for a-globin. From our 58 clones homologous to characterized genes, 46.5% Ž n s 27. are related to metabolism. Of these, 17 clones participate in energy metabolism, 2 are related to DNA metabolism and 8 participate to other metabolisms. Less numerous are the clones coding for proteins ranged in the categories of cell signaling and transporters Ž18.9%., structural proteins Ž13.8%. and transcription factors and translation machinery proteins Ž12.0%.. A few number of clones Ž8.6%. belong to scarcely represented categories, including neuron-specific, secretory and cell surface proteins.

ŽFig. 2B. showed their highest level of expression in the cochlea indicating a preferential expression in this tissue. Clones pCO101 and pCO135 exhibited additional strong

3.2. Expression analysis of selected clones The 40 clones showing no significant identity with any sequence in the databases were selected for tissue expression analysis. RT-PCR from seven different rat tissues Žcochlea, eye, brain, cerebellum, lung, kidney and liver. was performed with each selected clone. The expression of the ATPase subunit 6 was equivalent in the 7 tissues ŽFig. 2A., indicating proper quality of RNAs. As a result of this expression screening, 35 clones exhibited ubiquitous tissue expression whereas 5 showed a differential tissue distribution. These five latter clones, namely pCO101, pCO115, pCO119, pCO135 and pCO8, received particular attention. For each of these 5 clones, the absence of detectable cDNA fragments when RNA was directly amplified Žsee material and methods. indicated that RNAs were not contaminated with genomic DNA Žnot shown.. Clones pCO115 and pCO119 ŽFig. 2A. showed low expression levels in the cochlea compared to those in some other tissues Žbrain, cerebellum, eye and liver in the case of pCO115 and brain, eye and lung in the case of pCO119.. In contrast, clones pCO101, pCO135 ŽFig. 2A. and pCO8

Fig. 2. Expression of selected unkown genes using RT-PCR analysis in various tissues. Total RNAs were reverse transcribed, amplified by PCR using clone-specific primers Žsee Materials and methods. and the PCR products analyzed on a 4% agarose gel stained with Ethidium bromide: Br, brain; Cb, cerebellum; Ey, eye; Lu, lung; Kd, kidney; Co, cochlea; Li, liver; No, no cDNA. A: pCO115 Ž124 bp., for this clone, the lane 4 is heart ) , pCO119 Ž111 bp., clones: pCO101 Ž153 bp., pCO135 Ž142 bp., ubiquitously expressed ATPase subunit 6 Ž176 bp.. B: expression of pCO8 clone. Same legends as in A and Mu, muscle; Te, testis; He, heart. A 124-bp band corresponding to a pCO8 cDNA fragment is present in the cochlea and is hardly visible in brain.

6

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

expression in brain Žboth clones. and eye ŽpCO101., and lower expression levels in lung Žboth clones., cerebellum ŽpCO101. and eye ŽpCO135.. Clone pCO8 was the sole one to be noticeably expressed in only one tissue, that is the cochlea ŽFig. 2B.. It was not found in eight other tissues Žcerebellum, eye, kidney, liver, skeletal muscle, testis, heart, lung.. However, a slight band was apparent in brain although it was not consistently present in repeated experiments. Only traces in other tissues were detected when a 40-cycle PCR was performed Žnot shown.. Consequently, pCO8 was considered as a cochlear-restricted cDNA. The cellular expression of the unknown pCO8 clone and of the identified pCO6 clone Ž100% identical to nucleotides 823–890 of the rat drk1 potassium channel. were examined by in situ hybridization. Clone pCO8 was chosen because of its predominant expression in the cochlea and clone pCO6 because of the importance of potassium channels in cochlear physiology. We observed that both pCO8 ŽFig. 3A. and drk1 ŽFig. 4A. mRNAs were strongly expressed in the type I spiral ganglion neurons. They were localized only in the cytoplasm of these neurons, the nuclei and the axons were not stained. The intensity of the signal was uniform all along the cochlear turns. Although the distribution of the two mRNAs was similar, the signal intensity observed with the pCO8 riboprobe was higher than that observed with the drk1 riboprobe. A minor population of type I spiral ganglion neurons remained weakly stained or unstained. The glial cells surrounding the type I neuronal cell bodies appeared unstained with pCO8 ŽFig. 3B. as well as with pCO6 ŽFig. 4B.. The type II spiral ganglion neurons, characterized by a small cell body and by the absence of myelin are difficult to identify on cryostat sections. Putatively identified type II spiral ganglion neurons were not stained with the drk1 and pCO8 riboprobes. Border effects could be seen on some sections with the stria vascularis and the organ of Corti ŽFig. 3A.; they do not represent authentic hybridization signals. No specific staining was observed with the sense probe ŽFig. 4C..

Fig. 3. In situ hybridization analysis of the pCO8 clone in an adult rat cochlea using digoxigenin-labeled antisense riboprobes ŽA,B.. A: general view of the first Ž1. and second Ž2. turn of the cochlea showing the localization of the mRNA of this clone in the spiral ganglion neurons ŽSG.. No staining is observed in the organ of Corti Žasterisks. at the levels of the inner hair cells ŽI. and outer hair cells ŽO.. Other cochlear structures ŽSV; stria vascularis, sl, spiral limbus, osl; osseous spiral lamina. are also not stained. B: high magnification of the spiral ganglion in the first turn of the cochlea showing the localization of the pCO8 mRNA in the cytoplasm of the primary auditory neurons. Stained type I auditory neurons Žlarge arrows. are surrounded by unstained glial cells Žsmall arrows.. Some type I auditory neurons are not stained Žarrowheads. as well as possible type II auditory neurons Ždouble arrows., characterized by a small cellular body and by the absence of myelin. Scale barss100 Žin A. and s 20 m m Žin B..

Because of a slight expression of pCO8 mRNA in the brain Žsee results of RT-PCR analysis., we performed in situ hybridization in several parts of brain using the same tissue treatment and hybridization procedure than those for the cochlea. As a positive control, we used the pCO6 riboprobe since drk1 is highly expressed in the CNS w16,19,21,52x. All the brain structures examined, including hippocampus, cortex, thalamus, hypothalamus and those of

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

7

4. Discussion In order to identify genes important for inner ear function and potentially involved in hearing loss, we have examined a rat cochlea cDNA library. This library contains a sufficient number of clones to include representation of relatively scarce mRNAs, even with the diversity of tissues that make up the cochlea. 107 sequence tags were randomly collected and classified according to the type of the encoded protein. Our approach allowed us to successfully identify several new mRNAs highly expressed in the cochlea, including one, pCO8, almost exclusively expressed in this tissue. Our primary goal was to identify cochlea-specific mRNAs, based on the assumption that they must code for proteins important in cochlear function, rather than to edit an expressed sequence tag ŽEST. catalogue of the cochlea. Other approaches using cochlear cDNA libraries for the identification of cochlea-specific mRNAs have been reported w39,41x. In these works, specific mRNAs were selected by various techniques used in combination, including subtraction of libraries w38x, Northern blotting w39x and in situ hybridization w41,8x. To avoid the drawbacks of the subtraction procedure, we chose a systematic sequencing approach driven by a primary database selection of unknown sequences followed by a secondary selection based on the expression analysis by RT-PCR. Although the results of RT-PCR analysis should be interpreted with caution due to variations in the efficiency of amplification in different tissues, we believe that great differences in amounts of an amplified cDNA are the reflection of authentic differences in amounts of mRNAs. We considered that RT-PCR was the easiest and fastest technique to examine the expression of many clones. In fact, large-scale screening of expression obviates the use of time consuming Northern blots or in situ hybridization at this step. Only for those mRNAs which were selected for their highest interest was in situ hybridization applied as the final step. One of the most noticeable features of our ESTs is that mitochondrial transcripts are fairly abundant Ž15.9% of the total of the characterized clones. compared to percentages found in other tissues such as 12% in heart w29x, 1.7% in

the auditory pathway Žsuperior olivary complex, cochlear nuclei, inferior colliculus and lateral lemniscus., remained negative with pCO8 whereas various neuronal structures were labeled with pCO6 Žnot shown..

Fig. 4. In situ hybridization analysis of drk1 expression in the basal turn of the adult rat cochlea. Digoxigenin-labeled antisense ŽA,B. and sense ŽC. riboprobes were hybridized to cryostat sections. A: primary auditory neurons of the spiral ganglion ŽSG. display a high hybridization signal while inner ŽI. and outer hair cells ŽO. remain negative. B: stained type I auditory neurons Žlarge arrows. are surrounded by unstained glial cells Žsmall arrows.. Some auditory neurons are unstained: they included a few type I auditory neurons Žarrowheads. and possible type II auditory neurons Ždouble arrows.. C: with the sense riboprobe, no staining is visible in primary neurons Žarrowheads.. Scale bars s100 Žin A. and s10 m m Žin B,C..

8

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

fetal brain w1x, 10.3% in hippocampus w1x, 1% in pancreatic islet cells w50x and - 0.2% in HepG2 cells w33x. However, in EST catalogs from adult brain and skeletal muscle w26x higher frequencies of mitochondrial sequences, 50 and 24.8%, respectively, were observed. The finding of relatively high levels of mitochondrial messages in our cochlear library may be a reflection of the abundance of mitochondria in hair cells and stria vascularis w48x. This observation and recent work reporting mutations of mitochondrial genes in multisystemic disorders that often include hearing loss w6,54x indicate the importance of mitochondrial metabolism in the cochlea. The percentage Ž37.4%. of unknown genes found in the present study is in accordance with that reported in recent large-scale EST studies w5,28x. Only 1 out of 40 clones showed a highly predominant cochlear expression; if we assume that the other 63 identified clones are all expressed in the cochlea and in other tissues, this would indicate that - 1% of the clones in our sample represent mRNAs predominantly expressed in cochlea. The identities of our randomly picked sample of cochlear clones appears as a good molecular signature of the cochlea. In fact, we found proteins largely distributed in the organ of Corti such as the extracellular matrix protein a collagen w23x, the cytoskeletal proteins b tubulin w45,53x and b tropomyosin w7,46x. Proteins present in many parts of the cochlea were also identified. For example, the NaqrKqATPase, previously described in the stria vascularis, spiral limbus, spiral ligament and subpopulations of cochlear neurons of the spiral ganglion w40,57x and calretinin, a calcium-binding protein which has been detected in the inner hair cells, supporting cells and spiral ganglion neurons w14,15x, were found. Another structural protein, the Schwann cell peripheral myelin protein, PMP 0, w27x obviously involved in myelination and nerve conduction of the type I auditory neurons of the spiral ganglion, was identified. Recent assignment of mutations to genes coding for a collagen w11,25x and NaqrKq ATPase w44x that causes hearing loss indicate that many cochlear genes could be involved in genetic deafness. Even though many of the clones identified largely represent housekeeping genes, some others can be classified as tissue-specific genes. We did not examine the expression of these clones at the cellular level in the cochlea since our study was aimed at isolating new cochlea-specific genes. Yet, in situ hybridization studies of such clones might reveal interesting cell-specific expression in the cochlea. For example, a clone named pCO109 was found to code for the SII transcription factor. This gene has been reported to be specifically transcribed in testis w56x. Clone pCO16 codes for a neuron-specific MAP kinase Žaccession number gbL35236.. Clones pCO50 and pCO129 code for proteins homologous to the transcription factor P45 NF-E2 and a cell-binding bone sialoprotein, respectively, which show tissue-specific expression. The former one is restricted to haematopoietic tissues and cell lines w3x while the latter one is restricted to bone matrix

w34x. Clone pCO42 codes for a guanine nucleotide regulatory protein weakly present in all tissues except brain w51x. One of our characterized clones corresponds to the delayed rectifier Kq channel drk1 w18x. This channel was not identified before in the cochlea. Since potassium is the physiological charge transporter of the mechano-electrical transduction in hair cells and because it creates the hyperpolarization current induced by acetylcholine in outer hair cells of the organ of Corti w10x, we analyzed its cellular localization in the cochlea by in situ hybridization. We have not found drk1 mRNA in the organ of Corti; instead, it was exclusively expressed in the cell body of most of the type I auditory neurons. This is in accordance with previous works reporting its localization in neuronal cell bodies of the brain w16,19,21,52x and of the retina w20x in adult rat. The cDNA of another delayed rectifier potassium channel, NGK2, was recently isolated from the chick auditory sensory epithelium. It is predominantly expressed at the apical end of the basilar papilla w31x. This latter observation suggests that certain delayed rectifier Kq channels may be involved in hair cell physiology. Most importantly, pCO8, one of the clones that did not show significant identity with any database sequence was expressed in the type I auditory neurons of the cochlea. No expression was found in any other tissues examined, including other neurons of the auditory pathway, although a low level of expression was detected in brain by RT-PCR. This implies that primary auditory neurons could have particular features. In fact, 60% of the type I neurons have a high spontaneous rate of activity which correlates with a low threshold to a stimulatory sound. They are highly active neurons, both metabolically and electrically and show specific calcium buffer systems w47,14x. However, very little is known about the molecular mechanisms implicated in the physiology of the auditory response and the elements involved in the regulation of the metabolic pathway of these neurons. Isolation and characterization of molecules predominantly expressed in neurons of the spiral ganglion, as is pCO8, should allow better understanding of their functional characteristics.

5. Conclusions Identification of ESTs using a systematic sequencing approach coupled to RT-PCR analysis is a rapid and useful way to characterize novel genes expressed in the cochlea. Using this strategy we have identified and characterized several novel cDNA clones from a rat cochlea cDNA library that exhibit differential tissue expression, including 3 clones preferentially expressed in the cochlea. One of them, pCO8, almost exclusively found in the primary auditory neurons of the cochlea except for low levels of expression in the brain, is currently under investigation. Additional known cDNAs not previously reported in the cochlea were also isolated, including the drk1 potassium

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10

channel which is highly expressed in the type I auditory neurons of the cochlear spiral ganglion. Our results demonstrate the utility of this methodology and its ability to identify preferentially expressed cochlear genes.

w14x

w15x

Acknowledgements w16x

We want to gratefully thank M. Eybalin and J.L. Puel for comments on the manuscript and J.L. Pasquier for photographic work. EST sequences from clones corresponding to unknown genes are available in Genbank Žunder accession numbers: U74018 to U74052 and U75317 to U75320..

w17x

w18x

References w19x w1x Adams, M.D., Kerlavage, A.R., Fields, C. and Venter, J.C., 3.400 new expressed sequence tags identify diversity of transcripts in human brain, Nature Genet., 4 Ž1993. 256–267. w2x Adams, M.D., Soares, M.B., Kerlavage, A.R., Fields, C. and Venter, J.C., Rapid cDNA sequencing Žexpressed sequence tags. from a directionally cloned human brain infant brain cDNA library, Nature Genet., 4 Ž1993. 373–380. w3x Andrews, N.C., Erdjument-Bromage, H., Davidson, M.B., Tempst, P. and Orkin, S.H., Erhythroid transcription factor NF-E2 is a hematopoietic-specific basic-leucine zipper protein, Nature, 362 Ž1993. 722–728. w4x Angener, L.M. and Angener, R.C., Detection of poly-A RNA in sea urchin eggs and embryos by quantitative in situ hybridization, Nucleic Acids Res., 9 Ž1981. 2819–2840. w5x Auffray, C., Behar, G., Bois, F., Bouchier, C., Da Silva, C., Devignes, M.D., Duprat, S., Houlgatte, R., Jumeau, M.N. and Lamy, B., IMAGE: molecular integration of the analysis of the human genome and its expression, CR Acad. Sci. III, 3185 Ž1995. 263–272. w6x Ballinger, S.W., Shoffner, J.M., Hedaya, E.V., Trounce, I., Polak, M.A., Koontz, D.A. and Wallace, D.C., Maternally transmitted diabetes and deafness associated with a 10.4 kb mitochondrial DNA deletion, Nature Genet. Ž1992. 11–15. w7x Beisel, K.W. and Kennedy, J.E., Identification of novel alternatively spliced isoforms of the tropomyosin-encoding gene, TMnm, in the rat cochlea, Gene, 143 Ž1994. 251–256. w8x Bettenhausen, B. and Gossler, A., Efficient isolation of novel mouse genes differentially expressed in early postimplatation embryos, Genomics, 28 Ž1995. 436–441. w9x Bettini, E., Porta, A.R., Dahmen, N., Wang, H. and Margolis, F.L., Expressed sequence tags ŽEST. identify genes preferentially expressed in catfish chemosensory tissues, Mol. Brain Res., 23 Ž1994. 285–291. w10x Blanchet, C., Erostegui, C., Sugasawa and M., Dulon, D., Acethylcholine-induced potassium current of guinea pig outer hair cells: its dependance on a calcium influx through nicotinic-like receptors, J. Neurosci., 16 Ž1996. 2574–2584. w11x Byers, P.H., Inherited disorders of collagen gene structure and expression, Am. J. Med. Genet., 34 Ž1989. 72–80. w12x Chen, H., Thalmann, I., Adams, J.C., Avraham, K., Copeland, N.G., Jenkins, N.A., Beier, D.R., Corey, D.P., Thalmann, R. and Duyk, G.M., cDNA cloning tissue distribution, and chromosomal localization of Ocp2, a gene encoding a putative transcription-associated factor predominantly expressed in the auditory organs, Genomics, 27 Ž1995. 389–398. w13x Chomczynski, P. and Sacchi, N., Single step method of RNA

w20x

w21x

w22x

w23x

w24x

w25x

w26x

w27x

w28x

w29x

w30x

w31x

9

isolation by acid guanidinium thiocyanate-phenol-chloroform extraction, Anal. Biochem., 162 Ž1987. 156–159. Dechesne, C.J., Winsky, L., Moniot, B. and Raymond, J., Localization of calretinin mRNA in rat and guinea pig inner ear by in situ hybridization using radioactive and non-radioactive probes, Hearing Res., 69 Ž1993. 91–97. Dechesne, C.J., Rabejac, D. and Desmadryl, G., Development of calretinin immunoreactivity in the mouse inner ear, J. Comp. Neurol., 346 Ž1994. 517–529. Drewe, J.A., Verma, S., Frech, G. and Joho, R.H., Distinc spatial and temporal expression patterns of Kq channel mRNAs from different subfamilies, J. Neurosci., 12 Ž1992. 538–548. Erkman, L., McEvilly, R.J., Luo, L., Ryan, A.K., Hooshmand, F., O’Connel, S.M., Keithley, E.M., Rapaport, D.H., Ryan, A.F. and Rosenfeld, M.G., Role of transcription factors Brn-3.1, and Brn-3.2 in auditory and visual system development, Nature (London), 381 Ž1996. 603–606. Frech, G.C., VanDongen, A.M.J., Schuster, G., Brown, A.M. and Joho, R.H., A novel potassium channel with delayed rectifier properties isolated from brain by expression cloning, Nature, 340 Ž1989. 642–645. Hwang, P.M., Glatt, C.E., Bredt, D.S., Yellen, G. and Snyder S.H., A novel Kq channel with unique localizations in mammalian brain: molecular cloning and characterization, Neuron, 8 Ž1992. 473–481. Hwang, P.M., Cunningham, A.M., Peng, Y.W. and Snyder, S.H., CDRK and DRK1 channels have contrasting localizations in sensory systems, Neuroscience, 3 Ž1993. 613–620. Hwang, P.M., Fotuhi, M., Bredt, D.S., Cunningham, A.M. and Snyder, S.H., Contrasting immunohistochemical localizations in rat brain of two novel Kq channels of the Shab subfamily, J. Neurosci., 13 Ž1993. 1569–1576. Khan, A.S., Wilcox, A.S., Polymeropoulos, M.H., Hopkins, J.A., Stevens, T.J., Robinson, M., Orpana, A.K. and Sikela, J.M., Single pass sequencing and physical and genetic mapping of human, Nature Genet., 2 Ž1992. 180–185. Khetarpal, U., Robertson, N.G., Yoo, T.J. and Morton, C.C., Expression and localization of COL2A1 mRNA and type II collagen in human fetal cochlea, Hearing Res., 79 Ž1994. 59–73. Killick, R., Legan, P.K., Malenczak, C. and Richardson, G.P., Molecular cloning of chick b-tectorine, an extracellular matrix molecule of the inner ear, J. Cell Biol., 129 Ž1995. 535–547. Kuivaniemi, H., Tromp, G. and Prockop, D.J., Mutations in collagen genes: causes of rare and some common diseases in humans, FASEB J., 5 Ž1991. 2052–2060. Lanfranchi, G., Muraro, T., Caldara, F., Pacchioni, B., Pallavicini, A., Pandolfo, D., Toppo, S., Trevisan, S., Scarso, S. and Valle, G., X Identification of 4370 expressed sequence tags from a 3 -end-specific cDNA library of human skeletal muscle by DNA sequencing and filter hybridization, Genome Res., 6 Ž1996. 35–42. Lemke, G. and Axel, R., Isolation and sequence of a cDNA encoding the major structural protein of peripheral myelin, Cell, 40 Ž1985. 501–50. Lennon, G., Auffray, C., Polymeropoulos, M. and Soares, M.B., The I.M.A.G.E. Consortium: an integrated molecular analysis of genomes and their expression, Genomics, 33 Ž1995. 151–152. Liew, C.C., Hwang, D.M., Fung, Y.W., Laurenssen, E., Cukerman, E., Tsui, S. and Lee, C.Y., A catalogue of genes in the cardiovascular system as identified by expressed sequence tags, Proc. Natl. Acad. Sci. USA, 91 Ž1994. 10645–10649. Manes, ` G., Leducq, R., Pages, ` A. and Hamel, C.P., Messenger RNA quantification of the retinal pigment epithelium-specific protein RPE65 by competitive reverse transcription-polymerase chain reaction during the development of the rat retina, Vis. Res., 35 Ž1995. S235. Navaratnam, D.S., Bell, T.J., Kenyon, L.C., Oberholtzer, J.C., Characterization and differential distribution of potassium channels in the chick basilar papilla, XIXth Midwinter Research Meeting of the

10

w32x

w33x

w34x

w35x

w36x w37x

w38x

w39x

w40x

w41x

w42x

w43x

w44x

w45x

A. Soto-Prior et al.r Molecular Brain Research 47 (1997) 1–10 Association for Research in Otolaryngology, St. Petersburg, FL, 1996, p. 45. Oberholtzer, J.C., Cohen, E.L. and Davis, J.G., Molecular cloning of a chick cochlea cDNA encoding a subunit of DNA replication factor Cractivator 1, DNA Cell Biol., 13 Ž1994. 857–863. Okubo, K., Hori, N., Matoba, R., Niiyama, T., Fukushima, A., Kojima, Y. and Matsubara, K., Large scale cDNA sequencing for analysis of quantitative and qualitative aspects of gene expression, Nature Genet., 2 Ž1992. 173–179. Oldberg, A., Franzen, A. and Heinegard, D., The primary structure of a cell-binding bone sialoprotein, J. Biol. Chem., 263 Ž1988. 19420–19432. Pearson, W.R. and Lipman D.J., Improved tools for biological sequence comparison, Proc. Natl. Acad. Sci. USA, 85 Ž1988. 2444– 2448. Pujol, R., Anatomie et physiologie de la cochlee, ´ Arch. Int. Physiol. Biochem., 97-4 Ž1989. A51–A78. Represa, J., Leon, Y., Miner, C. and Giraldez, F., The int-2 protooncogene is responsible for induction of inner ear, Nature, 353 Ž1991. 561–353. Rivolta, M. and Wilcox, E., A novel and simple methodology to generate subtracted cDNA libraries, Nucleic Acids Res., 23 Ž1995. 256–2566. Robertson, N.G., Khetarpal, U., Gutierrez-Espeleta, G.A., Bieber, F.R. and Morton, C.C., Isolation of novel and known genes from a human fetal cochlea cDNA library using subtractive hybridization and differential screening, Genomics, 23 Ž1994. 42–50. Ryan, A.F. and Watts, A.G., Expression of genes coding for a and b isoforms of NaqrKq ATPase in the cochlea of the rat, Cell Mol. Neurosci., 2 Ž1991. 179–187. Ryan, A.F., Batcher, S., Lin, L., Brumm, D., O’Driscoll, K. and Harris, J.P., Cloning genes from a inner ear cDNA library, Arch. Otolaryngol. Head Neck Surg., 119 Ž1993. 1217–1220. Saffieddine, S. and Eybalin, M., Co-expression of NMDA and AMPArkainate receptor mRNA in cochlear neurons, NeuroReport, 3 Ž1992. 1145–1148. Santos-Sacchi, J., Cochlear physiology. In A.F. Jahn and J. SantosSacchi ŽEds.., Physiology of the Ear, Raven Press, New York, 1988, pp. 271–293. Schulte, B.A. and Steel, K.P., Expression of a and b subunit isoforms of NaqrKq-ATP-ase in the mouse inner ear and changes with mutations at the Wv or Sld loci, Hearing Res., 78 Ž1994. 65–76. Slepecky, N. and Chamberlain, C., Immunoelectron microscopic and immunofluorescent localization of cytoskeletal and muscle-like con-

w46x

w47x

w48x

w49x w50x

w51x

w52x

w53x

w54x

w55x

w56x

w57x

tractile proteins in inner ear sensory hair cells, Hearing Res., 20 Ž1985. 245–260. Slepecky, N. and Chamberlain, S.C., Tropomiosin co-localizes with actin microfilaments and microtubules within supporting cells of inner ear, Cell Tissue Res., 248 Ž1987. 63–66. Soto-Prior, A., Cluzel, M., Renard, N., Ripoll, C., Lavigne-Rebillard, M., Eybalin, M. and Hamel, C.P., Molecular cloning and expression of a parvalbumin in the guinea pig cochlea, Mol. Brain Res., 34 Ž1995. 337–342. Spector, G.J. and Carr, C., The electron transport system in the cochlear hair cell: the ultrastructural cytochemistry of respiratory enzymes in hair cell mitochondria of the guinea pig, Laryngoscope, 84 Ž1974. 1673–1706. Steel, K.P. and Brown S.D.M., Genes and deafness, Trends Genet., 10–12 Ž1994. 428–435. Takeda, J., Yano, H., Eng, S., Zeng, Y. and Bell, G.I., A molecular inventory of human pancreatic islets: sequence analysis of 1000 cDNA clones, Hum. Mol. Genet., 2 Ž1993. 1793–1798. Tan, E.C., Leung, T., Manser, E. and Lim, L., The human active breakpoint cluster region-related gene encodes a brain protein with homology to guanine nucleotide exchange proteins and GTPaseactivating proteins, J. Biol. Chem., 268 Ž1993. 27291–27298. Trimmer, J.S., Immunological identification and characterization of a delayed rectifier Kq channel polypeptide in rat brain, Proc. Natl. Acad. Sci. USA, 88 Ž1991. 10764–10768. Tucker, J.B., Paton, C.C., Richardson, G.P., Mogensen, M.M. and Russel, I.J., A cell surface-associated centrosomal layer of microtubule-organizing material in the inner pillar cell of the mouse cochlea, J. Cell Sci., 102 Ž1992. 215–226. van den Ouweland, J.M.W., Lemkes, H.H.P.J., Ruitenbeek, W., Sandkuijl, L.A., de Vijlder, M.F., Struyvenberg, P.A.A., van de Kamp, J.J.P. and Maassen, J.A., Mutation in mitochondrial tRNALeuŽUUR. gene is a large pedigree with maternally transmitted type II diabetes mellitus and deafness, Nature Genet., Ž1992. 368– 371. Wilcox, E.R. and Fex, J., Construction of a cDNA library from microdissected guinea pig organ of corti, Hearing Res., 62 Ž1992. 124–126. Xu, Q., Nakanishi, T., Sekimizu, K. and Natori, S., Cloning and identification of testis-specific transcription elongation factor S-II, J. Biol. Chem., 269 Ž1994. 3100–3103. Yao, X., ten Cate, W.J.F., Curtis, L.M. and Rarey, K.E., Expression of Naq,Kq-ATPase a 1 subunit mRNA in the developing rat cochlea, Hearing Res., 80 Ž1994. 31–37.