- Email: [email protected]
Differential display and gene arrays to examine auditory plasticity
Hearing Research 147 (2000) 293^302 www.elsevier.com/locate/heares
Di¡erential display and gene arrays to examine auditory plasticity Margaret I. Lomax a
a;b;
*, Li Huang a , Younsook Cho Richard A. Altschuler a;b
a;b
, Tzy-Wen L. Gong a ,
Kresge Hearing Research Institute, Department of Otolaryngology/Head-Neck Surgery, University of Michigan Medical School, 9301E Medical Sciences Research Building III, Box 0648, Ann Arbor, MI 48109, USA b Department of Cell and Developmental Biology, University of Michigan Medical School, Ann Arbor, MI 48109, USA Received 11 October 1999; accepted 5 April 2000
Abstract Differential gene expression forms the basis for development, differentiation, regeneration, and plasticity of tissues and organs. We describe two methods to identify differentially expressed genes. Differential display, a PCR-based approach, compares the expression of subsets of genes under two or more conditions. Gene arrays, or DNA microarrays, contain cDNAs from both known genes and novel genes spotted on a solid support (nylon membranes or glass slides). Hybridization of the arrays with RNA isolated from two different experimental conditions allows the simultaneous analysis of large numbers of genes, from hundreds to thousands to whole genomes. Using differential display to examine differential gene expression after noise trauma in the chick basilar papilla, we identified the UBE3B gene that encodes a new member of the E3 ubiquitin ligase family (UBE3B). UBE3B is highly expressed immediately after noise in the lesion, but not in the undamaged ends, of the chick basilar papilla. UBE3B is most similar to a ubiquitin ligase gene from Caenorhabditis elegans, suggesting that this gene has been conserved throughout evolution. We also describe preliminary experiments to profile gene expression in the cochlea and brain with commercially available low density gene arrays on nylon membranes and discuss potential applications of this and DNA microarray technology to the auditory system. ß 2000 Elsevier Science B.V. All rights reserved. Key words: Noise trauma ; Ubiquitin ligase; UBE3B; Inferior colliculus; Auditory nerve
1. Introduction Di¡erential gene expression provides the driving force for most biological processes, including development, di¡erentiation, regeneration, and plasticity of all tissues and organs. The function of sensory organs such as the ear depends on the expression of subsets of genes in the speci¢c cells and tissues that comprise both the peripheral sensory organ and the central auditory pathway. Stress or other environmental stimuli such as excessive noise may induce changes in gene expression. Thus, analysis of di¡erential gene expression in regenerating tissues after stress can identify genes for proteins required to re-establish homeostasis and restore normal function.
* Corresponding author. Tel.: +1 (734) 647-0952; Fax: +1 (734) 647-2563; E-mail: [email protected]
Molecular biologists have historically been interested in analyzing changes in gene expression during the many biological processes noted above. Gene expression has been analyzed with nucleic acid probes at the mRNA level (mRNA expression) or with antibodies to speci¢c proteins (protein expression). These molecular biology and cell biology experiments have been limited to the analysis of individual genes, or at best small groups of genes, and have provided crucial information on where (which cell or tissue) and when (what stage of development or regeneration) a gene is expressed. Clearly, to relate changes in gene expression to changes in function, it would be desirable to examine di¡erential expression of groups of genes and, ideally, to examine the entire repertoire of genes expressed in a tissue or organ, i.e., the `gene pro¢le' or `transcriptome' for that tissue or organ. Two major experimental approaches have been used to identify and clone genes that are di¡erentially ex-
0378-5955 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 0 0 ) 0 0 1 3 9 - 8
HEARES 3515 18-8-00
294
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
pressed. The ¢rst method, subtractive hybridization (Robertson et al., 1994; Bonaldo, 1996; Soares et al., 1994), involves a hybridization step (subtractive hybridization) that physically removes mRNAs for genes expressed in both tissues or experimental conditions, thus enriching for genes expressed preferentially in one tissue vs. the other. Subtractive hybridization has been used in the auditory system to generate a cDNA library of genes preferentially expressed in human fetal cochlea (Robertson et al., 1994; Skvorak et al., 1999). This approach has identi¢ed two novel and important ear-speci¢c genes: antiquitin (Skvorak et al., 1997) and COCH (Robertson et al., 1997). Mutations in the COCH gene cause a human non-syndromic deafness with vestibular dysfunction, DFNA9 (Robertson et al., 1998; de Kok et al., 1999), demonstrating that this tissue-speci¢c gene plays an important role in hearing. The second method, di¡erential display of mRNA (Liang and Pardee, 1992; Liang et al., 1993; Welsh et al., 1992; McClelland et al., 1995), compares subsets of genes to identify those that are di¡erentially expressed. Di¡erential display polymerase chain reaction (DDPCR) can be applied to small amounts of total RNA, a major advantage in the auditory system where small amounts of tissue are the norm. A signi¢cant disadvantage of both methods is that complete molecular characterization of each di¡erentially expressed gene requires a substantial amount of e¡ort, since both methods generate only partial cDNAs for each di¡erentially expressed gene. Furthermore, the identi¢cation of the function of novel proteins and/or the biochemical pathways in which they participate is problematic. The advent of gene arrays, also called DNA microarrays (Schena et al., 1995; Brown and Botstein, 1999 ; Duggan et al., 1999; Gerhold et al., 1999), will supplant both of these methods for analysis of di¡erential gene expression in mammals. Soares (1997) suggested that the resources available through the human, mouse and rat genome projects should expedite and facilitate the analysis of di¡erential gene expression in mammals. The principle of the gene array technique is that large numbers of cDNAs from both known genes and/or novel genes of unknown function can be arrayed with an automated robotic spotting device (Cheung et al., 1999) on a solid support, either a nylon membrane or a glass slide, as small dots or microdots. RNAs from di¡erent tissues or di¡erent experimental conditions are reverse transcribed to cDNA and simultaneously labeled with either radioactive (32 P- or 33 P-labeled) nucleotides for nylon membranes or £uorochrome-tagged nucleotides for glass slides. The cDNA arrays (probes) are then hybridized with labeled cDNA target and the hybridization signals are analyzed. The signal intensity at each array element is proportional to the expression level of that gene in the sample. For radiolabeled sig-
nals, membranes are exposed to a phosphorimager screen and laser-scanned to determine the signal intensity. For £uorescent signals, a £uorescent confocal microscope reader is used to read out and digitize signals. In this article we describe and evaluate two molecular genetic approaches we have used to examine changes in gene expression in the auditory system after noise trauma: di¡erential display and gene arrays. DD-PCR was applied to the problem of identifying genes that are di¡erentially expressed after noise trauma in the chick basilar papilla (Gong et al., 1996, 1997, 1999 ; Adler et al., 1999). We illustrate the power of DD-PCR by describing the molecular cloning of a novel E3 ubiquitin ligase, UBE3B, that may play an important role in regeneration and recovery. We also describe preliminary experiments with gene arrays on nylon membranes to pro¢le gene expression in subregions of cochlea and brain as a prelude to examining changes in gene expression after noise trauma. The potential applications of both low density gene arrays and high density DNA microarrays (Duggan et al., 1999) to problems in the auditory system will be evaluated. 2. Materials and methods 2.1. Di¡erential display The complete experimental details of the noise exposure conditions and our di¡erential display analysis of gene expression after noise trauma in the chick basilar papilla have been described previously (Gong et al., 1996). The partial cDNA designated KH125 was isolated in this experiment. There were three experimental groups for the noise exposure experiments. The control group (C) was not exposed to noise; the acoustic trauma group (A) was exposed to noise and sacri¢ced immediately after the noise exposure; and the recovery group (R) was similarly exposed to noise but allowed to recover for 2 days following noise exposure, then sacri¢ced. 2.2. Isolation of chick and human cDNAs for KH125 A BLAST search of the non-redundant GenBank database of known genes with the DNA sequence of a partial chick cDNA (KH125) isolated in the di¡erential display experiment (Gong et al., 1996) failed to identify the gene. To identify and characterize the gene, therefore, we isolated both chick cochlear and human cDNAs as follows. Screening cDNA libraries from chick cochlea by plaque hybridization yielded two overlapping partial cDNAs designated KH411 (2.5 kb) and KH412 (1.2 kb). We sequenced the 2.5-kb chick cDNA,
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
295
KH411; the cDNA sequence has been submitted to GenBank and received the accession number AF251045. We used the DNA sequence of KH411 in a BLAST search of dbEST, the database of expressed sequence tags (EST; Adams et al., 1991), and identi¢ed a 2.5-kb partial human cDNA (KH438; IMAGE Clone ID 592621) (Lennon et al., 1996). This plasmid was obtained from Research Genetics and sequenced; the cDNA sequence has been submitted to GenBank and received the accession number AF251046. The chick KH411 cDNA was also used as a probe to screen a human pituitary gland cDNA library to identify additional human cDNA sequences.
er's protocol (http://www.clontech.com/clontech/Manuals/index.html) with [32 P]dATP-labeled cDNA generated from 2 Wg total RNA of each rat tissue. Male Sprague^Dawley rats (250 g) were used in all cases. Hybridization solution (5 ml) containing 6.9^7.2U106 cpm of 32 P-labeled cDNA probe was added to the membrane and incubated at 68³C overnight. After washing, the membranes were exposed overnight to a phosphorimager screen (Molecular Dynamics) and scanned with PhosphorImager SI 445 (Molecular Dynamics).
2.3. Northern blot and RNA dot-blot hybridization
Puri¢ed plasmid DNAs were sequenced in the University of Michigan DNA Sequencing Core. DNA sequences were aligned to generate longer cDNA sequences with the SeqMan program and protein sequences from di¡erent species were aligned using the CLUSTAL program of LASERGENE (DNA Star, Madison, WI). Database searches with either DNA or protein sequences were performed with the BLAST search program on NCBI databases via internet servers.
Total RNA was isolated from either the central, noise-damaged region of the chick basilar papilla (lesion), or the undamaged region (ends), by the acidic phenol method (Chomczynski and Sacchi, 1987), subjected to electrophoresis (1% agarose^formaldehyde gel), transferred to and immobilized on a Nytran membrane (Schleicher and Schuell, Keene, NH). A master dot blot (Clontech, Palo Alto, CA) contained poly(A) RNAs from 50 human fetal and adult tissues `dotted' on a nylon membrane; each RNA sample has been normalized to the amount of ubiquitin mRNA in the tissue. The chick Northern blot was probed at 65³C overnight with a [32 P]UTP-labeled riboprobe derived from the 3P untranslated region (UTR) of the UBE3B cDNA (KH411). The most stringent post-hybridization wash was with 0.1USSC^0.5% SDS at 65³C. The membranes were exposed to Kodak X-OMAT ¢lm with intensifying screens for 2 weeks at 380³C. 2.4. Slot-blot and gene array hybridization Both slot-blot and gene array hybridization experiments are essentially `reverse' Northern blots in which several gene-speci¢c cDNAs (probes) are ¢xed to the membrane or solid support. The RNA population (target) is converted to a mixture of radiolabeled cDNAs, which are hybridized to the unlabeled gene-speci¢c probes spotted on the membrane. The amount of radiolabeled cDNA that binds to each gene probe is linearly related to the concentration of that cDNA in the target RNA population. For slot-blot analysis of gene expression after noise trauma in the chick basilar papilla, we generated the radiolabeled cDNA by reverse transcription of 1 Wg total RNA, followed by random primer labeling of ¢rst strand cDNA with two radiolabeled deoxynucleotides. Additional experimental details can be found in Gong et al. (1996). For Clontech Rat Atlas1 cDNA Array membranes, we hybridized membranes according to the manufactur-
2.5. DNA and protein sequence analysis
2.6. Analysis of gene array data The 16-bit phosphorimager ¢le (.gel) for each membrane was converted to a .tif ¢le with Adobe Photoshop and analyzed with the AtlasImage V. 1.0 software program (Clontech). The signal in each cell of the membrane was determined from the pixel density and corrected for background signal. Although it is highly desirable when using gene arrays for gene pro¢ling to hybridize membranes simultaneously with labeled cDNA from two di¡erent tissues, this is often not possible ; therefore, signals from two di¡erent experiments must be normalized to each other. The Atlas Gene Array membranes contain nine housekeeping genes for normalization : L- and Q-actin, GAPDH, HPRT, myosin I, ribosomal protein L19, K-tubulin, and ubiquitin. Based on the results of several di¡erent hybridization experiments, we chose three genes (L19, L29, and ubiquitin) for normalization because these genes gave signals signi¢cantly above background with target RNA samples from all cochlear and brain tissues. The ratio of the signals for each of the three genes was determined, then the ratios were averaged. This average ratio was then applied to the signals from the second membrane to correct for experimental di¡erences due to amount of input radioactivity or length of exposure in the phosphorimager. 2.7. Animal care Procedures for the care and use of animals reported
HEARES 3515 18-8-00
296
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
on in this study were approved by the University of Michigan Committee on the Care and Use of Animals. (supported by NIH Grants DC02492, `Molecular Analysis of Ear Development in the Chick', and DC02982 `Molecular Mechanisms of Cochlear Function and Dysfunction' : Project III, `Molecular Genetics of Acoustic Trauma and Response to Trauma in Mammals'; and Project IV, `Molecular Mechanisms of Protection'.) 3. Results 3.1. Di¡erential gene expression in the regenerating chick auditory epithelium Birds can regenerate the auditory epithelium follow-
ing damage due to noise or ototoxic drugs (reviewed in Stone et al., 1998; Staecker and Van de Water, 1998). Raphael and colleagues (Raphael and Altschuler, 1992; Raphael, 1993; Adler et al., 1995) have shown morphologically that a limited noise exposure produces a lesion in the central region of the chick basilar papilla in which all hair cells have been destroyed. New stereocilia bundles, presumably from new hair cells, appear by 4 days following the noise exposure. Evidence has accumulated supporting several di¡erent mechanisms of new hair cell production, including mitosis, transdi¡erentiation, and repair (reviewed in Staecker and Van de Water, 1998); however, there are few molecular data underlying any of these mechanisms. To begin to examine the process of hair cell regeneration in the chick basilar papilla at the molecular level, we performed a di¡erential display experiment designed
Fig. 1. Slot-blot hybridization analysis of gene expression after acoustic trauma. Triplicate membranes were generated with a slot-blot apparatus (Schleicher and Schuell, Keene, NH). Plasmid DNA (2.5 Wg) was denatured with alkali, then ¢ltered through each slot. The KH129 plasmid contained a partial cDNA derived from the KIF1B gene (Gong et al., 1999). Plasmid DNAs in the next three rows contained di¡erential display cDNAs for known genes (KH150, KH161, KH167) (Gong et al., 1996). Plasmid DNA alone was included as negative (vector) control (pCR-Script, pGEM-3Zf). The chick L-actin plasmid contained part of the coding region plus part of the 3P UTR. Plasmids KH125, KH143, KH158, KH159, KH164, and KH166 contained di¡erential display cDNAs for genes that in a preliminary screen appeared to be di¡erentially expressed after noise trauma. Each membrane (column) was hybridized separately to a labeled cDNA probe containing all mRNA sequences in the target RNA population from that experimental group (C, control; A, acoustic trauma; R, recovery; see Methods in Gong et al., 1996). KH125 is one of three genes that appear to be up-regulated after acoustic trauma. KH143 has been shown to be WDR1 (Adler et al., 1999). KH158 has not been characterized further. KH159, KH164, and KH166 did not have signals signi¢cantly higher than the negative controls (plasmid DNAs).
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
297
to identify genes that are di¡erentially expressed after noise trauma (Gong et al., 1996). The 40 partial cDNAs isolated were analyzed by a slot-blot hybridization experiment, which is essentially a `reverse' Northern blot. Individual cDNA `probes' from the di¡erential display experiment were ¢xed to a membrane and hybridized with radiolabeled cDNAs generated from the `target' RNA population, i.e., RNA from each experimental condition. This experimental approach allows one to verify results of di¡erential gene expression simultaneously for several genes. Slot-blot hybridization (Fig. 1) with RNA from basilar papillae of either control (C), acoustic trauma (A) or recovery (R) groups con¢rmed di¡erential expression of several genes. Three of these genes (KH150, KH161, KH167) were known genes and have been described previously (Gong et al., 1996). Four genes (KH129, KH125, KH143, KH158) were novel. To illustrate the process of gene identi¢cation in a DD-PCR experiment, we describe the partial cloning and identi¢cation of the UBE3B gene that encodes the KH125 partial cDNA. 3.2. Di¡erential expression of UBE3B after noise trauma in lesions Northern blot analysis of RNA from chick basilar papilla con¢rmed that the UBE3B gene, from which the KH125 partial cDNA was derived, is di¡erentially expressed after noise trauma. UBE3B is highly expressed in the lesioned area of the chick basilar papilla immediately after noise, but not in the undamaged ends nor in the basilar papillae from control chicks (Fig. 2). We hybridized a multiple-tissue human RNA dot blot with the human UBE3B cRNA probe to identify the overall expression pattern of the human gene (Fig. 3). The highest signals were seen in RNA from pituitary gland and several other tissues, including heart. Based on these results, we chose to screen a pituitary gland library to isolate additional human cDNAs. Northern blot analysis of total RNA from normal chick tissues probed with a chick cRNA for UBE3B demonstrated that the gene encodes a single transcript of approximately 7.5 kb that is present at high levels in heart and brain, but not in liver or intestine (data not shown). 3.3. UBE3B encodes a novel ubiquitin ligase Analysis of partial chick and human cDNAs isolated with the chick partial cDNA (KH125) provided su¤cient information to identify the type of protein encoded by the gene, even though the full-length UBE3B cDNA has not yet been isolated (Fig. 4). Both chick and human partial cDNAs contained a HECT (homology to E6-AP C-terminus) domain (Sche¡ner, 1998), which identi¢ed these proteins as
Fig. 2. Northern blot analysis of gene expression after acoustic trauma. (A) RNA was isolated from either the central lesioned areas or the ends (base and apex) of the three experimental groups (see Section 2). Each lane contained 5.7 Wg total RNA. The hybridization probe was an antisense riboprobe from the partial chick cDNA (KH411). Lane 1: control RNA; Lane 2: RNA from the acoustic trauma group; Lane 3: RNA from the recovery group; Lane 4: RNA from the ends of control basilar papillae; Lane 5: RNA from the ends of basilar papillae from the acoustic trauma group; Lane 6: RNA from the ends (base and apex) of basilar papillae of the recovery group. (B) Photograph of ethidium bromide-stained gel prior to transfer to visualize the 28S and 18S rRNAs as a control for RNA loading.
novel E3 ubiquitin ligases. Ubiquitin is a small polypeptide that, when transferred to substrate or target proteins, marks those proteins for degradation by proteasomes. The ubiquitin pathway requires three enzymes : E1, ubiquitin activating enzyme; E2, ubiquitin conjugating enzyme ; and E3, ubiquitin ligase. The E3 enzyme transfers (ligates) ubiquitin to the ^NH2 groups of lysine residues on the target protein to yield polyubiquitin chains. 3.4. Gene pro¢ling with low density gene arrays on nylon membranes The advent of gene array (or cDNA microarray) technology has enabled us to use a more powerful method for analyzing di¡erential gene expression after noise trauma in the mammalian auditory system. We are initially using low density arrays of known rat genes on nylon membranes (Atlas Gene Arrays, Clontech).
HEARES 3515 18-8-00
298
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
Fig. 3. Human RNA dot-blot analysis. The blot of mRNA from human tissues was hybridized with an antisense riboprobe derived from the 3P UTR region of the human partial cDNA (KH438). The human gene is clearly expressed in many tissues, consistent with the identi¢cation of human ESTs from many cDNA libraries.
The Atlas Gene Arrays contain short cDNAs from 588 genes grouped according to biochemical function. We used these nylon membranes ¢rst for gene pro¢ling studies in the mammalian auditory system. The purpose of these ¢rst experiments is to determine which mRNAs for various genes or biochemical pathways can be detected in whole cochlea, subregions of cochlea and re-
gions of the auditory pathway in the brain. Atlas Gene Array membranes were hybridized with total RNA from two neuronal regions, the modiolus of the cochlea and the inferior colliculus (IC) of the brain, scanned in a phosphorimager, and analyzed with Clontech's AtlasImage software (see Section 2 for details). We focused on the genes for apoptosis-related pro-
Fig. 4. The chick and human UBE3B proteins contain a HECT domain with a conserved active site. BLAST search of the non-redundant GenBank database of known proteins identi¢ed several proteins with high sequence identity to either chick or human UBE3B. Alignment of the Ctermini of the UBE3B ubiquitin ligases with a known ubiquitin ligase (E6-AP) reveals high sequence conservation in the active site region, which includes the invariant Cys residue (C) to which the ubiquitin is transferred. The sequence preceding the active site is quite variable among these proteins and may de¢ne di¡erent subfamilies. The sequence of chick and human UBE3B is most similar to the Caenorhabditis elegans protein, both in the active site and the £anking region. Gaps (^) have been inserted to optimize the alignment. The GenBank accession numbers of the sequences used are: E6-AP (2340821); C. elegans (3880812); Schizosaccharomyces pombe Pub1 (3218406); mouse Itch (AF037454); mouse Nedd-4 (1709250); human NEDD-4 (1171682); KIAA0010 (265983); KIAA0312 (2224565); KIAA0322 (2224585). Several HECT domain E3 ligases have been identi¢ed in the databases through their high sequence identity to the C-terminus of E6-AP; however, the function of these proteins and the identity of their targets are unknown. The poly-ubiquitinated protein targets are transported to the 26S proteasome, where the target protein is degraded and the ubiquitin monomers are released (Hershko, 1998; Hershko and Ciechanover, 1998).
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
Fig. 5. Gene pro¢ling of RNA from normal adult rat inferior colliculus and modiolus. The Atlas1 cDNA expression array (catalog number 7738-1, Clontech, Palo Alto, CA) contains species-speci¢c PCR products derived from the cDNAs of 588 known rat genes. These PCR fragments are spotted in duplicate (10 ng cDNA/dot) on nylon membranes. The genes are grouped into six blocks, labeled A^E. Each block contains functionally related genes (for more detail, see the Clontech webpage at http://www.clontech.com). For hybridization experiments, radiolabeled cDNA was synthesized according to the vendor's protocol with a slight modi¢cation (Clontech, Palo Alto, CA). Reverse transcriptase reaction was performed on 2 Wg of total RNA from either inferior colliculus (IC) or modiolus (MOD) in 10 Wl of a reaction mixture containing 20 nM each of gene-speci¢c primers, 500 WM each of dCTP, dGTP and dTTP, 1.2 WM of [K-32 P]dATP (3000 Ci/mmol), and 100 U of MMLV reverse transcriptase Superscript II (Life Technologies Inc., Rockville, MD) for 25 min at 50³C. After purifying the probe with CHROMA SPIN-200 DEPC-H2 O columns (Clontech, Palo Alto, CA), 1^2U106 cpm/ml of the 32 P-labeled ¢rst-strand cDNA probe in 10 ml of ExpressHyb1 hybridization solution (Clontech, Palo Alto, CA) was hybridized to Atlas1 Rat cDNA Expression Array at 68³C overnight. The membranes were washed with 2USSC/1% SDS four times for 30 min each and two more times with 0.1USSC/0.5% SDS for 30 min each at 68³C and exposed to a Storage Phosphor screen for phosphorimaging overnight. The screen was scanned with phosphorimager 445 SI (Molecular Dynamics, Sunnyvale, CA). Block C of the array containing apoptosis (C1a^C3e)- and nervoussystem (C3f^C7n)-related genes from the resulting image ¢le is shown above. (A) The signals on block C when hybridized with labeled cDNA from IC. (B) The signals on block C when hybridized with labeled cDNA from MOD.
teins (C1a^C3e) and nervous-system-related proteins (C3f^C7n), which are contained in one subregion of the membrane (panel C). Comparing the nervous-system-related proteins (C3f^C7n) in two neuronal regions, the modiolus and IC, we observed that only three genes were di¡erentially expressed based on the threshold sensitivity de¢ned (200% above background) (Fig. 5). Myelin proteolipid protein (PLP; C3f) is the major protein component of CNS myelin, whereas peripheral myelin protein (PMP22 ; C3h), as the name indicates, is the major protein component of myelin in peripheral nerves. As expected, PLP was expressed at higher levels in IC than in modiolus, although there are signi¢cant amounts of PLP mRNA in modiolus. In contrast, the PMP22 mRNAs levels were higher in modiolus than IC. The mRNAs for both genes are fairly abundant and generate signals that are signi¢cantly above the sensitivity threshold. The PLP signal in IC was 2.5
299
times that in modiolus; conversely, the PMP22 signal was 5.8 times higher in modiolus than IC. Both genes were detected in random sequencing of cDNAs from a human fetal cochlear cDNA library, again con¢rming that the PLP gene for CNS myelin was also expressed in peripheral tissues like the cochlea. Mutations in both genes lead to neurological disorders. A third gene with signi¢cant di¡erences between IC and modiolus encodes the plasma membrane calcium transporting ATPase isoform 2 (ATP2B2 or PMCA ; C6g), one of four di¡erent ATP2B/PMCA isoforms in mammals. ATP2B2 mRNA was above the threshold sensitivity in both neuronal tissues, and was about twice as high in IC as in modiolus. PMCA2 exhibits a highly restricted tissue distribution, suggesting that it serves more specialized physiological functions than some of the other calcium transporting ATPase isoforms. Mutation analysis suggests that the PMCA2 gene may play a unique role in hearing. Street et al. (1998) showed that the deafwaddler mouse mutant (dfw), which is deaf and has vestibular/motor imbalance, has a mutation in the Atp2b2 gene. Furthermore, Kozel et al. (1998) assessed the physiological role of PMCA2 in the inner ear by producing PMCA2-de¢cient mice through gene targeting. These mice were also deaf and had balance problems. Histological analysis of both mouse mutants demonstrated high level expression of Atp2b2 in the cochlea, primarily in outer hair cells and spiral ganglion cells (Kozel et al., 1998 ; Street et al., 1998). Thus, a gene with a limited expression pattern (only ¢ve cDNAs identi¢ed by random sequencing of brain cDNA libraries), but a well-documented role in the ear can be detected by the gene array technique. 4. Discussion Di¡erential display analysis of regeneration in the chick basilar papilla suggests that proteins involved in actin polymerization, and in signaling through the actin cytoskeleton, may play important roles. Lee and Cotanche (1995) detected higher levels of L-actin mRNA in the chick basilar papilla after noise trauma by using a RT-PCR assay speci¢c for L-actin. We con¢rmed this observation in our slot-blot hybridization experiment, which included L-actin cDNA. Furthermore, we identi¢ed genes involved in actin signaling, including CDC42 (Gong et al., 1997) and WDR1 (Adler et al., 1999). CDC42 is a member of the Rho family of small GTPases known to be involved in actin signaling. CDC42 is required for the later, elongation steps of actin ¢ber formation (Kollmar, 1999). WDR1, a 68kDa protein with nine WD40 repeats, is the vertebrate homologue of yeast actin-interacting protein 1 (AIP1), which was ¢rst identi¢ed in a two-hybrid screen for
HEARES 3515 18-8-00
300
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
actin binding proteins. Recent studies de¢ne the function of yeast AIP1p ; it interacts with co¢lin to disassemble actin ¢laments (Rodal et al., 1999). DAip1, a Dictyostelium homologue of the yeast AIP1, is involved in endocytosis, cytokinesis, and motility (Konzok et al., 1999). It is not known whether vertebrate WDR1 serves similar functions in disassembling actin ¢laments, or whether it serves a structural role in its interactions with actin. These experiments identi¢ed two genes that were clearly di¡erentially expressed (upregulated) after noise trauma and that encoded novel proteins. The partial cDNA designated KH129 is the 3P end of a previously unrecognized 10-kb mRNA that encodes a 204-kDa kinesin of the UNC-104/KIF1 family (Gong et al., 1999). The 10-kb transcript is derived from the Kif1b gene by alternative splicing. The second gene (UBE3B) was partially cloned in this study and identi¢ed as a new member of a family of functionally related proteins de¢ned by a conserved C-terminal 350-amino acid `HECT' domain. These proteins are all related to human E6-AP (UBE3A), the ¢rst known E3 ubiquitin ligase (Sche¡ner, 1998). E6-AP was initially identi¢ed as a cellular protein that mediates the in vitro association of the human papillomavirus E6 oncoprotein with p53, leading to the ubiquitin-dependent turnover of p53, a tumor suppressor protein (Huibregtse et al., 1991, 1998; Sche¡ner et al., 1993). This leads to immortalization and oncogenesis. Ubiquitin-mediated protein turnover is central to many cellular processes, including di¡erentiation, proliferation, oncogenesis, and apoptosis. The HECT domain contains the conserved active site of the enzyme and is a hallmark of this family of E3 ubiquitin ligases (Hochstrasser, 1996; Hershko, 1998; Hershko and Ciechanover, 1998). UBE3B is thus a member of this important class of regulatory proteins. Mutations in the human E6-AP gene (UBE3A), the ¢rst E3 ubiquitin ligase identi¢ed, cause Angelman syndrome (Matsuura et al., 1997), which involves severe developmental and neurological problems. This was the ¢rst example of a genetic disorder of the ubiquitin-dependent proteolytic pathway. E6-AP targets the tumor suppressor p53 for destruction when papilloma viruses infect cells, thus leading to cell proliferation and tumorigenesis (Sche¡ner, 1998). Ubiquitination, the addition of poly-ubiquitin chains to proteins, targets many additional important transcriptional regulators for rapid turnover, e.g., NF-UB, AP1 (c-Fos and c-Jun), and GCN4 in yeast (Hochstrasser and Kornitzer, 1998). The UBE3B gene exhibits the greatest di¡erential expression after acoustic trauma of any gene that we have analyzed to date. The novel ubiquitin ligase identi¢ed in our experiments may target an important regulatory protein for rapid turnover and a¡ect important physio-
logical processes in the cell. Further experiments are required to identify the protein or proteins that the UBE3B ubiquitin ligase targets for rapid turnover, and to determine the roles of both the UBE3B protein and its target in the chick cochlea following noise trauma. Di¡erential display (and subtractive hybridization) clearly has the power to identify genes that exhibit the desired pattern of expression ; however, only a few genes can be analyzed because of the considerable time and e¡ort required to completely characterize each gene. It became clear, as we prepared to analyze di¡erential gene expression in mammals after noise trauma, that limiting the analysis initially to well-characterized genes might be faster and more productive. Soares (1997) suggested that the resources available through the human, mouse and rat genome projects should expedite and facilitate the analysis of di¡erential gene expression in mammals. In fact, this insight led to the development of gene arrays, which are essentially slot-blot experiments where the slots have been reduced to dots and the number of genes increased from 10 to hundreds and thousands. In the following section we discuss our preliminary results on `gene pro¢ling' in the auditory system using low density gene arrays. In gene pro¢ling experiments, the concentration of the `probe', i.e., the amount of unlabeled cDNA for the gene of interest, is determined by the size of the spot or microdot on the array. The concentration of the labeled cDNA in the `target' is determined by the relative abundance of the mRNA in that tissue, which determines whether or not a su¤cient amount of labeled cDNA will bind to the cDNA probe during the hybridization experiment. Thus, it is important to understand that, in general, gene arrays and microarrays are sampling primarily the moderately abundant mRNAs, and not necessarily the important low and rare abundance mRNA classes. The fact that we detected preferential expression of CNS myelin (PLP) in brain regions and peripheral myelin (PMP22) in the cochlea was to be expected, as these myelin proteins are very abundant proteins and their mRNAs are in the moderately abundant class. The fact that we could detect signi¢cant levels of mRNA for the plasma membrane calcium ATPase (PMCA2 or ATP2B2) was not anticipated. In fact, very few cDNAs derived from this gene were found by random sequencing of brain cDNAs (mouse EST sequencing project), and none was found in the Morton fetal cochlear cDNA library (Skvorak et al., 1999). Nevertheless, using the Clontech gene arrays we detected signi¢cant levels of mRNA for this gene in cochlea and all auditory brain regions examined (Cho and Lomax, unpublished data). This exciting and gratifying result suggests that gene arrays may have a profound impact on the analysis of di¡er-
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
ential gene expression in the mammalian auditory system. Gene arrays will certainly provide an additional important method for determining which genes are expressed in the inner ear, even when expression in other tissues is quite low. DNA microarrays on glass slides are a major technological advance and the ultimate way to analyze di¡erential gene expression. An automated robotic spotting device developed by Patrick Brown, Stanford University (Schena et al., 1995, 1996, 1998) can generate many replicates of the gene arrays quickly and reproducibly. This technological advance enables researchers to analyze simultaneous changes in expression of many genes, from hundreds to thousands to whole genomes. Brown applied this technology to di¡erential gene expression studies of the entire yeast genome under di¡erent physiological conditions (Brown and Botstein, 1999) and also to mammalian culture systems (Iyer et al., 1999). This technology employs £uorescently labeled cDNAs from target RNA populations. cDNAs from two target populations can be labeled separately with di¡erent £uorochromes, combined in equal amounts, and hybridized in one experiment to the microdots on the glass slide. The relative signal intensity of the £uorochromes at each spot is a measure of relative gene expression and provides a more accurate measure than the nylon membranes, which must be analyzed in parallel. The relative abundance of a mRNA in the target RNA population will a¡ect the amount of signal obtained. Increasing the sensitivity of the gene pro¢ling approach will require methods to amplify the signal for the low and rare abundance classes, such as linear RNA ampli¢cation (Van Gelder et al., 1990; Eberwine et al., 1992). There is considerable expense in establishing this technology, including the cost of purchasing all human or mouse ESTs, the cost of the robotic spotting device and the laser confocal reader. Furthermore, the large amount of data generated from these arrays requires sophisticated bioinformatics approaches for analysis (Ermolaeva et al., 1998; Bassett et al., 1999; Iyer et al., 1999). Although this technology is not currently available to many researchers, it should become more accessible in the near future (see Cheung et al., 1999). In summary, gene arrays hold great promise for differential gene expression in the mammalian auditory system. Potential applications might include analysis of gene expression in response to various stimuli, in disease states, and in mouse mutants to detect downstream targets of transcription factors or changes in biochemical pathways in response to mutant proteins. Unfortunately, there is no systematic attempt to identify and map all avian expressed genes, i.e., an Avian Genome Project, therefore, we cannot apply gene arrays to one of the most interesting questions in the auditory system, the regenerating chick auditory epithelium.
301
Analysis of di¡erential gene expression in birds will require those molecular approaches summarized in Section 1 and illustrated in this article. Acknowledgements Supported by NIH Grants DC02492 (M.I.L.) and DC02982 (M.I.L. and R.A.A.).
References Adams, M., Kelley, J., Gocayne, J.D., Dubnick, M., Polymeropoulos, M.H., Xiao, H., Merril, C.R., Wu, A., Olde, B., Moreno, R.F., Kerlavage, A.R., McCombie, W.R., Venter, J.C., 1991. Complementary DNA sequencing: Expressed sequence tags and the human genome project. Science 252, 1651^1656. Adler, H.J., Niemec, A.J., Moody, D.B., Raphael, Y.R., 1995. Tectorial membrane regeneration in acoustically damaged birds: An immunocytochemical technique. Hear. Res. 86, 43^46. Adler, H.J., Winnicki, R.S., Gong, T-W.L., Lomax, M.I., 1999. A gene upregulated in the acoustically damaged chick basilar papilla encodes a protein with WD40 repeats. Genomics 56, 59^69. Bassett, D.E., Jr., Eisen, M.B., Boguski, M.S., 1999. Gene expression informatics ^ it's all in your mine. Nat. Genet. 21, 51^55. Bonaldo, F.M., 1996. Normalization and subtraction: Two approaches to facilitate gene discovery. Genome Res. 6, 791^806. Brown, P.O., Botstein, D., 1999. Exploring the new world of the genome with DNA microarrays. Nat. Genet. 21, 33^37. Cheung, V.G., Morley, M., Aqilar, F., Massimi, A., Kucherlapati, R., Childs, G., 1999. Making and reading microarrays. Nat. Genet. 21, 15^19. Chomczynski, P., Sacchi, N., 1987. Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156^159. de Kok, Y.J., Bom, S.J., Brunt, T.M., Kemperman, M.H., van Beusekom, E., van der Velde-Visser, S.D., Robertson, N.G., Morton, C.C., Huygen, P.L., Verhagen, W.I., Brunner, H.G., Cremers, C.W., Cremers, F.P., 1999. A Pro51Ser mutation in the COCH gene is associated with late onset autosomal dominant progressive sensorineural hearing loss with vestibular defects. Hum. Mol. Genet. 8, 361^366. Duggan, D.J., Bittner, M., Chen, Y., Meltzer, P., Trent, J.M., 1999. Expression pro¢ling using cDNA microarrays. Nat. Genet. 21, 10^ 14. Eberwine, J., Yeh, H., Miyashiro, K., Cao, Y., Nair, S., Finnell, R., Zettell, M., Coleman, P., 1992. Analysis of gene expression in single live neurons. Proc. Natl. Acad. Sci. USA 89, 3010^3014. Ermolaeva, O., Rastogi, M., Pruitt, K.D., Schuler, G.D., Bittner, M.L., Chen, Y., Simon, R., Meltzer, P., Trent, J.M., Boguski, M.S., 1998. Data management and analysis for gene expression arrays. Nat. Genet. 20, 19^23. Gerhold, D., Rushmore, T., Caskey, C.T., 1999. DNA chips: promising toys have become powerful tools. Trends Biochem. Sci. 24, 168^173. Gong, T.-W.L., Hegeman, A.D., Shin, J.J., Adler, H.J., Raphael, Y., Lomax, M.I., 1996. Identi¢cation of genes expressed after noise exposure in the chick basilar papilla. Hear. Res. 96, 20^32. Gong, T.-W.L., Shin, J.J., Burmeister, M., Lomax, M.I., 1997. Complete cDNAs for CDC42 from chicken cochlea and mouse liver. Biochim. Biophys. Acta 1352, 282^292. Gong, T.L., Winnicki, R.S., Kohrman, D.C., Lomax, M.I., 1999. A
HEARES 3515 18-8-00
302
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
novel mouse kinesin of the Unc-104/KIF1 subfamily encoded by the Kif1b gene. Gene 239, 117^127. Hershko, A., 1998. The ubiquitin system: Past, present, and future perspectives. In: Peters, J.-M., Harris, J.R. and Finley, D. (Eds.), Ubiquitin and the Biology of the Cell. Plenum Press, New York, pp. 1^13. Hershko, A., Ciechanover, A., 1998. The ubiquitin system. Annu. Rev. Biochem. 67, 425^479. Hochstrasser, M., 1996. Ubiquitin-dependent protein degradation. Annu. Rev. Genet. 30, 405^439. Hochstrasser, M., Kornitzer, D., 1998. Ubiquitin-dependent inactivation of transcription regulators. In: Peters, J.-M., Harris, J.R. and Finley, D. (Eds.), Ubiquitin and the Biology of the Cell. Plenum Press, New York, pp. 279^302. Huibregtse, J.M., Sche¡ner, M., Howley, P.M., 1991. A cellular protein mediates association of p53 with the E6 oncoprotein of human papillomavirus types 16 or 18. EMBO J. 10, 4129^4135. Huibregtse, J.M., Maki, C.G., Howley, P.M., 1998. Ubiquitination of the p53 tumor suppressor. In: Peters, J.-M., Harris, J.R. and Finley, D. (Eds.), Ubiquitin and the Biology of the Cell. Plenum Press, New York, pp. 323^343. Iyer, V.R., Eisen, M.B., Ross, D.T., Schuler, G., Moore, T., Lee, J.C.F., Trent, J.M., Staudt, L.M., Hudson, J., Jr., Boguski, M.S., Lashkari, D., Shalon, D., Botstein, D., Brown, P.O., 1999. The transcriptional program in the response of human ¢broblasts to serum. Science 283, 83^87. Kollmar, R., 1999. Who does the hair cell `do'? Rho GTPases and the hair-bundle morphogenesis. Curr. Opin. Neurobiol. 9, 394^398. Konzok, A., Weber, I., Simmeth, E., Hacker, U., Maniak, M., Muller-Taubenberger, A., 1999. DAip1, a Dictyostelium homologue of the yeast actin-interacting protein 1, is involved in endocytosis, cytokinesis, and motility. J. Cell Biol. 146, 453^464. Kozel, P.J., Friedman, R.A., Erway, L.C., Yamoah, E.N., Liu, L.H., Riddle, T., Du¡y, J.J., Doetschman, T., Miller, M.L., Cardell, E.L., Shull, G.E., 1998. Balance and hearing de¢cits in mice with a null mutation in the gene encoding plasma membrane Ca2 -ATPase isoform 2. J. Biol. Chem. 273, 18693^18696. Lee, K.H., Cotanche, D.A., 1995. Detection of L-actin mRNA by RTPCR in normal and regenerating chicken cochlae. Hear. Res. 87, 9^15. Lennon, G., Au¡ray, C., Polymeropoulos, M., Soares, M.B., 1996. The I.M.A.G.E. Consortium: an integrated molecular analysis of genomes and their expression. Genomics 33, 151^152. Liang, P., Pardee, A.B., 1992. Di¡erential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science 257, 967^971. Liang, P., Averboukh, L., Pardee, A.B., 1993. Distribution and cloning of eukaryotic mRNAs by means of di¡erential display: re¢nements and optimization. Nucleic Acids Res. 21, 3269^3275. Matsuura, T., Sutcli¡e, J.S., Fang, P., Galjaard, R.-J., Jiang, Y., Benton, C.S., Rommens, J.M., Beaudet, A.L., 1997. De novo truncating mutations in E6-AP ubiquitin-protein ligase gene (UBE3A) in Angelman syndrome. Nat. Genet. 15, 74^77. McClelland, M., Mathieu-Daude, F., Welsh, J., 1995. RNA ¢ngerprinting and di¡erential display using arbitrarily primed PCR. Trends Genet. 11, 242^246. Raphael, Y., 1993. Reorganization of the chick basilar papilla following acoustic trauma. J. Comp. Neurol. 330, 521^532. Raphael, Y., Altschuler, R.A., 1992. Early micro¢lament reorganization in injured auditory epithelia. Exp. Neurol. 115, 32^36. Robertson, N.G., Khetarpal, U., Gutierrez-Espeleta, G.A., Bieber, F.R., Morton, C.C., 1994. Isolation of novel and known genes from a human fetal cochlear cDNA library using subtractive hybridization and di¡erential screening. Genomics 23, 42^50.
Robertson, N.G., Skvorak, A.B., Yin, Y., Weremowicz, S., Johnson, K.R., Kovatch, K.A., Battey, J.F., Bieber, F.R., Morton, C.C., 1997. Mapping and characterization of a novel cochlear gene in human and in mouse: a positional candidate gene for a deafness disorder, DFNA9. Genomics 46, 345^354. Robertson, N.G., Lu, L., Heller, S., Merchant, S.N., Eavey, R.D., McKenna, M., Nadol, J.B.Jr., Miyamoto, R.T., Linthicum, F.H.Jr., Lubianca Neto, J.F., Hudspeth, A.J., Seidman, C.E., Morton, C.C., Seidman, J.G., 1998. Mutations in a novel cochlear gene cause DFNA9, a human nonsyndromic deafness with vestibular dysfunction. Nat. Genet. 20, 299^303. Rodal, A.A., Tetreault, J.W., Lappalainen, P., Drubin, D.G., Amberg, D.C., 1999. Aip1p interacts with co¢lin to disassemble actin ¢laments. J. Cell Biol. 145, 1251^1264. Sche¡ner, M., 1998. Ubiquitin, E6-AP, and their role in p53 inactivation. Pharmacol. Ther. 78, 129^139. Sche¡ner, M., Huibregtse, J.M., Vierstra, R.D., Howley, P.M., 1993. The HPV-16 E6 and E6-AP complex functions as a ubiquitinprotein ligase in the ubiquitination of p53. Cell 75, 495^505. Schena, M., Shalon, D., Davis, R.W., Brown, P.O., 1995. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 270, 467^470. Schena, M., Shalon, D., Heller, R., Chai, A., Brown, P.O., Davis, R.W., 1996. Parallel human genome analysis: Microarray-based expression monitoring of 1000 genes. Proc. Natl. Acad. Sci. USA 93, 10614^10619. Schena, M., Heller, R.A., Theriault, T.P., Konrad, K., Lachenmeier, E., Davis, R.W., 1998. Microarrays: biotechnology's discovery platform for functional genomics. Trends Biotechnol. 16, 301^306. Skvorak, A.B., Robertson, N.G., Yin, Y., Weremowicz, S., Her, H., Bieber, F.R., Beisel, K.W., Lynch, E.D., Beier, D.R., Morton, C.C., 1997. An ancient conserved gene expressed in the human inner ear: identi¢cation, expression analysis, and chromosomal mapping of human and mouse antiquitin (ATQ1). Genomics 46, 191^199. Skvorak, A.B., Weng, Z., Yee, A.J., Robertson, N.G., Morton, C.C., 1999. Human cochlear expressed sequence tags provide insight into cochlear gene expression and identify candidate genes for deafness. Hum. Mol. Genet. 8, 439^452. Soares, M.B., 1997. Identi¢cation and cloning of di¡erentially expressed genes. Curr. Opin. Biotechnol. 8, 542^546. Soares, M.B., Bonaldo, M.F., Jelene, P., Su, L., Lawton, L., Efstratiadis, A., 1994. Construction and chracterization of a normalized cDNA library. Proc. Nat. Acad. Sci. USA 91, 9228^9232. Staecker, H., Van de Water, T.R., 1998. Factors controlling hair-cell regeneration/repair in the inner ear. Curr. Opin. Neurobiol. 8, 480^487. Stone, J.S., Oesterle, E.C., Rubel, E.W., 1998. Recent insights into regeneration of auditory and vestibular hair cells. Curr. Opin. Neurol. 11, 17^24. Street, V.A., McKee-Johnson, J.W., Fonseca, R.C., Tempel, B.L., Noben-Trauth, K., 1998. Mutations in a plasma membrane Ca2 -ATPase gene cause deafness in deafwaddler mice. Nat. Genet. 19, 390^394. Van Gelder, R.N., Von Zastrow, M.E., Yool, A., Dement, W.C., Barchas, J.D., Eberwine, J.H., 1990. Ampli¢ed RNA synthesized from limited quantities of heterogeneous cDNA. Proc. Natl. Acad. Sci. USA 87, 1663^1667. Welsh, J., Chada, K., Dalal, S.S., Cheng, R., Ralph, D., McClelland, M., 1992. Arbitrarily primed PCR ¢ngerprinting of RNA. Nucleic Acids Res. 20, 4965^4970.
HEARES 3515 18-8-00
Di¡erential display and gene arrays to examine auditory plasticity Margaret I. Lomax a
a;b;
*, Li Huang a , Younsook Cho Richard A. Altschuler a;b
a;b
, Tzy-Wen L. Gong a ,
Kresge Hearing Research Institute, Department of Otolaryngology/Head-Neck Surgery, University of Michigan Medical School, 9301E Medical Sciences Research Building III, Box 0648, Ann Arbor, MI 48109, USA b Department of Cell and Developmental Biology, University of Michigan Medical School, Ann Arbor, MI 48109, USA Received 11 October 1999; accepted 5 April 2000
Abstract Differential gene expression forms the basis for development, differentiation, regeneration, and plasticity of tissues and organs. We describe two methods to identify differentially expressed genes. Differential display, a PCR-based approach, compares the expression of subsets of genes under two or more conditions. Gene arrays, or DNA microarrays, contain cDNAs from both known genes and novel genes spotted on a solid support (nylon membranes or glass slides). Hybridization of the arrays with RNA isolated from two different experimental conditions allows the simultaneous analysis of large numbers of genes, from hundreds to thousands to whole genomes. Using differential display to examine differential gene expression after noise trauma in the chick basilar papilla, we identified the UBE3B gene that encodes a new member of the E3 ubiquitin ligase family (UBE3B). UBE3B is highly expressed immediately after noise in the lesion, but not in the undamaged ends, of the chick basilar papilla. UBE3B is most similar to a ubiquitin ligase gene from Caenorhabditis elegans, suggesting that this gene has been conserved throughout evolution. We also describe preliminary experiments to profile gene expression in the cochlea and brain with commercially available low density gene arrays on nylon membranes and discuss potential applications of this and DNA microarray technology to the auditory system. ß 2000 Elsevier Science B.V. All rights reserved. Key words: Noise trauma ; Ubiquitin ligase; UBE3B; Inferior colliculus; Auditory nerve
1. Introduction Di¡erential gene expression provides the driving force for most biological processes, including development, di¡erentiation, regeneration, and plasticity of all tissues and organs. The function of sensory organs such as the ear depends on the expression of subsets of genes in the speci¢c cells and tissues that comprise both the peripheral sensory organ and the central auditory pathway. Stress or other environmental stimuli such as excessive noise may induce changes in gene expression. Thus, analysis of di¡erential gene expression in regenerating tissues after stress can identify genes for proteins required to re-establish homeostasis and restore normal function.
* Corresponding author. Tel.: +1 (734) 647-0952; Fax: +1 (734) 647-2563; E-mail: [email protected]
Molecular biologists have historically been interested in analyzing changes in gene expression during the many biological processes noted above. Gene expression has been analyzed with nucleic acid probes at the mRNA level (mRNA expression) or with antibodies to speci¢c proteins (protein expression). These molecular biology and cell biology experiments have been limited to the analysis of individual genes, or at best small groups of genes, and have provided crucial information on where (which cell or tissue) and when (what stage of development or regeneration) a gene is expressed. Clearly, to relate changes in gene expression to changes in function, it would be desirable to examine di¡erential expression of groups of genes and, ideally, to examine the entire repertoire of genes expressed in a tissue or organ, i.e., the `gene pro¢le' or `transcriptome' for that tissue or organ. Two major experimental approaches have been used to identify and clone genes that are di¡erentially ex-
0378-5955 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 5 9 5 5 ( 0 0 ) 0 0 1 3 9 - 8
HEARES 3515 18-8-00
294
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
pressed. The ¢rst method, subtractive hybridization (Robertson et al., 1994; Bonaldo, 1996; Soares et al., 1994), involves a hybridization step (subtractive hybridization) that physically removes mRNAs for genes expressed in both tissues or experimental conditions, thus enriching for genes expressed preferentially in one tissue vs. the other. Subtractive hybridization has been used in the auditory system to generate a cDNA library of genes preferentially expressed in human fetal cochlea (Robertson et al., 1994; Skvorak et al., 1999). This approach has identi¢ed two novel and important ear-speci¢c genes: antiquitin (Skvorak et al., 1997) and COCH (Robertson et al., 1997). Mutations in the COCH gene cause a human non-syndromic deafness with vestibular dysfunction, DFNA9 (Robertson et al., 1998; de Kok et al., 1999), demonstrating that this tissue-speci¢c gene plays an important role in hearing. The second method, di¡erential display of mRNA (Liang and Pardee, 1992; Liang et al., 1993; Welsh et al., 1992; McClelland et al., 1995), compares subsets of genes to identify those that are di¡erentially expressed. Di¡erential display polymerase chain reaction (DDPCR) can be applied to small amounts of total RNA, a major advantage in the auditory system where small amounts of tissue are the norm. A signi¢cant disadvantage of both methods is that complete molecular characterization of each di¡erentially expressed gene requires a substantial amount of e¡ort, since both methods generate only partial cDNAs for each di¡erentially expressed gene. Furthermore, the identi¢cation of the function of novel proteins and/or the biochemical pathways in which they participate is problematic. The advent of gene arrays, also called DNA microarrays (Schena et al., 1995; Brown and Botstein, 1999 ; Duggan et al., 1999; Gerhold et al., 1999), will supplant both of these methods for analysis of di¡erential gene expression in mammals. Soares (1997) suggested that the resources available through the human, mouse and rat genome projects should expedite and facilitate the analysis of di¡erential gene expression in mammals. The principle of the gene array technique is that large numbers of cDNAs from both known genes and/or novel genes of unknown function can be arrayed with an automated robotic spotting device (Cheung et al., 1999) on a solid support, either a nylon membrane or a glass slide, as small dots or microdots. RNAs from di¡erent tissues or di¡erent experimental conditions are reverse transcribed to cDNA and simultaneously labeled with either radioactive (32 P- or 33 P-labeled) nucleotides for nylon membranes or £uorochrome-tagged nucleotides for glass slides. The cDNA arrays (probes) are then hybridized with labeled cDNA target and the hybridization signals are analyzed. The signal intensity at each array element is proportional to the expression level of that gene in the sample. For radiolabeled sig-
nals, membranes are exposed to a phosphorimager screen and laser-scanned to determine the signal intensity. For £uorescent signals, a £uorescent confocal microscope reader is used to read out and digitize signals. In this article we describe and evaluate two molecular genetic approaches we have used to examine changes in gene expression in the auditory system after noise trauma: di¡erential display and gene arrays. DD-PCR was applied to the problem of identifying genes that are di¡erentially expressed after noise trauma in the chick basilar papilla (Gong et al., 1996, 1997, 1999 ; Adler et al., 1999). We illustrate the power of DD-PCR by describing the molecular cloning of a novel E3 ubiquitin ligase, UBE3B, that may play an important role in regeneration and recovery. We also describe preliminary experiments with gene arrays on nylon membranes to pro¢le gene expression in subregions of cochlea and brain as a prelude to examining changes in gene expression after noise trauma. The potential applications of both low density gene arrays and high density DNA microarrays (Duggan et al., 1999) to problems in the auditory system will be evaluated. 2. Materials and methods 2.1. Di¡erential display The complete experimental details of the noise exposure conditions and our di¡erential display analysis of gene expression after noise trauma in the chick basilar papilla have been described previously (Gong et al., 1996). The partial cDNA designated KH125 was isolated in this experiment. There were three experimental groups for the noise exposure experiments. The control group (C) was not exposed to noise; the acoustic trauma group (A) was exposed to noise and sacri¢ced immediately after the noise exposure; and the recovery group (R) was similarly exposed to noise but allowed to recover for 2 days following noise exposure, then sacri¢ced. 2.2. Isolation of chick and human cDNAs for KH125 A BLAST search of the non-redundant GenBank database of known genes with the DNA sequence of a partial chick cDNA (KH125) isolated in the di¡erential display experiment (Gong et al., 1996) failed to identify the gene. To identify and characterize the gene, therefore, we isolated both chick cochlear and human cDNAs as follows. Screening cDNA libraries from chick cochlea by plaque hybridization yielded two overlapping partial cDNAs designated KH411 (2.5 kb) and KH412 (1.2 kb). We sequenced the 2.5-kb chick cDNA,
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
295
KH411; the cDNA sequence has been submitted to GenBank and received the accession number AF251045. We used the DNA sequence of KH411 in a BLAST search of dbEST, the database of expressed sequence tags (EST; Adams et al., 1991), and identi¢ed a 2.5-kb partial human cDNA (KH438; IMAGE Clone ID 592621) (Lennon et al., 1996). This plasmid was obtained from Research Genetics and sequenced; the cDNA sequence has been submitted to GenBank and received the accession number AF251046. The chick KH411 cDNA was also used as a probe to screen a human pituitary gland cDNA library to identify additional human cDNA sequences.
er's protocol (http://www.clontech.com/clontech/Manuals/index.html) with [32 P]dATP-labeled cDNA generated from 2 Wg total RNA of each rat tissue. Male Sprague^Dawley rats (250 g) were used in all cases. Hybridization solution (5 ml) containing 6.9^7.2U106 cpm of 32 P-labeled cDNA probe was added to the membrane and incubated at 68³C overnight. After washing, the membranes were exposed overnight to a phosphorimager screen (Molecular Dynamics) and scanned with PhosphorImager SI 445 (Molecular Dynamics).
2.3. Northern blot and RNA dot-blot hybridization
Puri¢ed plasmid DNAs were sequenced in the University of Michigan DNA Sequencing Core. DNA sequences were aligned to generate longer cDNA sequences with the SeqMan program and protein sequences from di¡erent species were aligned using the CLUSTAL program of LASERGENE (DNA Star, Madison, WI). Database searches with either DNA or protein sequences were performed with the BLAST search program on NCBI databases via internet servers.
Total RNA was isolated from either the central, noise-damaged region of the chick basilar papilla (lesion), or the undamaged region (ends), by the acidic phenol method (Chomczynski and Sacchi, 1987), subjected to electrophoresis (1% agarose^formaldehyde gel), transferred to and immobilized on a Nytran membrane (Schleicher and Schuell, Keene, NH). A master dot blot (Clontech, Palo Alto, CA) contained poly(A) RNAs from 50 human fetal and adult tissues `dotted' on a nylon membrane; each RNA sample has been normalized to the amount of ubiquitin mRNA in the tissue. The chick Northern blot was probed at 65³C overnight with a [32 P]UTP-labeled riboprobe derived from the 3P untranslated region (UTR) of the UBE3B cDNA (KH411). The most stringent post-hybridization wash was with 0.1USSC^0.5% SDS at 65³C. The membranes were exposed to Kodak X-OMAT ¢lm with intensifying screens for 2 weeks at 380³C. 2.4. Slot-blot and gene array hybridization Both slot-blot and gene array hybridization experiments are essentially `reverse' Northern blots in which several gene-speci¢c cDNAs (probes) are ¢xed to the membrane or solid support. The RNA population (target) is converted to a mixture of radiolabeled cDNAs, which are hybridized to the unlabeled gene-speci¢c probes spotted on the membrane. The amount of radiolabeled cDNA that binds to each gene probe is linearly related to the concentration of that cDNA in the target RNA population. For slot-blot analysis of gene expression after noise trauma in the chick basilar papilla, we generated the radiolabeled cDNA by reverse transcription of 1 Wg total RNA, followed by random primer labeling of ¢rst strand cDNA with two radiolabeled deoxynucleotides. Additional experimental details can be found in Gong et al. (1996). For Clontech Rat Atlas1 cDNA Array membranes, we hybridized membranes according to the manufactur-
2.5. DNA and protein sequence analysis
2.6. Analysis of gene array data The 16-bit phosphorimager ¢le (.gel) for each membrane was converted to a .tif ¢le with Adobe Photoshop and analyzed with the AtlasImage V. 1.0 software program (Clontech). The signal in each cell of the membrane was determined from the pixel density and corrected for background signal. Although it is highly desirable when using gene arrays for gene pro¢ling to hybridize membranes simultaneously with labeled cDNA from two di¡erent tissues, this is often not possible ; therefore, signals from two di¡erent experiments must be normalized to each other. The Atlas Gene Array membranes contain nine housekeeping genes for normalization : L- and Q-actin, GAPDH, HPRT, myosin I, ribosomal protein L19, K-tubulin, and ubiquitin. Based on the results of several di¡erent hybridization experiments, we chose three genes (L19, L29, and ubiquitin) for normalization because these genes gave signals signi¢cantly above background with target RNA samples from all cochlear and brain tissues. The ratio of the signals for each of the three genes was determined, then the ratios were averaged. This average ratio was then applied to the signals from the second membrane to correct for experimental di¡erences due to amount of input radioactivity or length of exposure in the phosphorimager. 2.7. Animal care Procedures for the care and use of animals reported
HEARES 3515 18-8-00
296
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
on in this study were approved by the University of Michigan Committee on the Care and Use of Animals. (supported by NIH Grants DC02492, `Molecular Analysis of Ear Development in the Chick', and DC02982 `Molecular Mechanisms of Cochlear Function and Dysfunction' : Project III, `Molecular Genetics of Acoustic Trauma and Response to Trauma in Mammals'; and Project IV, `Molecular Mechanisms of Protection'.) 3. Results 3.1. Di¡erential gene expression in the regenerating chick auditory epithelium Birds can regenerate the auditory epithelium follow-
ing damage due to noise or ototoxic drugs (reviewed in Stone et al., 1998; Staecker and Van de Water, 1998). Raphael and colleagues (Raphael and Altschuler, 1992; Raphael, 1993; Adler et al., 1995) have shown morphologically that a limited noise exposure produces a lesion in the central region of the chick basilar papilla in which all hair cells have been destroyed. New stereocilia bundles, presumably from new hair cells, appear by 4 days following the noise exposure. Evidence has accumulated supporting several di¡erent mechanisms of new hair cell production, including mitosis, transdi¡erentiation, and repair (reviewed in Staecker and Van de Water, 1998); however, there are few molecular data underlying any of these mechanisms. To begin to examine the process of hair cell regeneration in the chick basilar papilla at the molecular level, we performed a di¡erential display experiment designed
Fig. 1. Slot-blot hybridization analysis of gene expression after acoustic trauma. Triplicate membranes were generated with a slot-blot apparatus (Schleicher and Schuell, Keene, NH). Plasmid DNA (2.5 Wg) was denatured with alkali, then ¢ltered through each slot. The KH129 plasmid contained a partial cDNA derived from the KIF1B gene (Gong et al., 1999). Plasmid DNAs in the next three rows contained di¡erential display cDNAs for known genes (KH150, KH161, KH167) (Gong et al., 1996). Plasmid DNA alone was included as negative (vector) control (pCR-Script, pGEM-3Zf). The chick L-actin plasmid contained part of the coding region plus part of the 3P UTR. Plasmids KH125, KH143, KH158, KH159, KH164, and KH166 contained di¡erential display cDNAs for genes that in a preliminary screen appeared to be di¡erentially expressed after noise trauma. Each membrane (column) was hybridized separately to a labeled cDNA probe containing all mRNA sequences in the target RNA population from that experimental group (C, control; A, acoustic trauma; R, recovery; see Methods in Gong et al., 1996). KH125 is one of three genes that appear to be up-regulated after acoustic trauma. KH143 has been shown to be WDR1 (Adler et al., 1999). KH158 has not been characterized further. KH159, KH164, and KH166 did not have signals signi¢cantly higher than the negative controls (plasmid DNAs).
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
297
to identify genes that are di¡erentially expressed after noise trauma (Gong et al., 1996). The 40 partial cDNAs isolated were analyzed by a slot-blot hybridization experiment, which is essentially a `reverse' Northern blot. Individual cDNA `probes' from the di¡erential display experiment were ¢xed to a membrane and hybridized with radiolabeled cDNAs generated from the `target' RNA population, i.e., RNA from each experimental condition. This experimental approach allows one to verify results of di¡erential gene expression simultaneously for several genes. Slot-blot hybridization (Fig. 1) with RNA from basilar papillae of either control (C), acoustic trauma (A) or recovery (R) groups con¢rmed di¡erential expression of several genes. Three of these genes (KH150, KH161, KH167) were known genes and have been described previously (Gong et al., 1996). Four genes (KH129, KH125, KH143, KH158) were novel. To illustrate the process of gene identi¢cation in a DD-PCR experiment, we describe the partial cloning and identi¢cation of the UBE3B gene that encodes the KH125 partial cDNA. 3.2. Di¡erential expression of UBE3B after noise trauma in lesions Northern blot analysis of RNA from chick basilar papilla con¢rmed that the UBE3B gene, from which the KH125 partial cDNA was derived, is di¡erentially expressed after noise trauma. UBE3B is highly expressed in the lesioned area of the chick basilar papilla immediately after noise, but not in the undamaged ends nor in the basilar papillae from control chicks (Fig. 2). We hybridized a multiple-tissue human RNA dot blot with the human UBE3B cRNA probe to identify the overall expression pattern of the human gene (Fig. 3). The highest signals were seen in RNA from pituitary gland and several other tissues, including heart. Based on these results, we chose to screen a pituitary gland library to isolate additional human cDNAs. Northern blot analysis of total RNA from normal chick tissues probed with a chick cRNA for UBE3B demonstrated that the gene encodes a single transcript of approximately 7.5 kb that is present at high levels in heart and brain, but not in liver or intestine (data not shown). 3.3. UBE3B encodes a novel ubiquitin ligase Analysis of partial chick and human cDNAs isolated with the chick partial cDNA (KH125) provided su¤cient information to identify the type of protein encoded by the gene, even though the full-length UBE3B cDNA has not yet been isolated (Fig. 4). Both chick and human partial cDNAs contained a HECT (homology to E6-AP C-terminus) domain (Sche¡ner, 1998), which identi¢ed these proteins as
Fig. 2. Northern blot analysis of gene expression after acoustic trauma. (A) RNA was isolated from either the central lesioned areas or the ends (base and apex) of the three experimental groups (see Section 2). Each lane contained 5.7 Wg total RNA. The hybridization probe was an antisense riboprobe from the partial chick cDNA (KH411). Lane 1: control RNA; Lane 2: RNA from the acoustic trauma group; Lane 3: RNA from the recovery group; Lane 4: RNA from the ends of control basilar papillae; Lane 5: RNA from the ends of basilar papillae from the acoustic trauma group; Lane 6: RNA from the ends (base and apex) of basilar papillae of the recovery group. (B) Photograph of ethidium bromide-stained gel prior to transfer to visualize the 28S and 18S rRNAs as a control for RNA loading.
novel E3 ubiquitin ligases. Ubiquitin is a small polypeptide that, when transferred to substrate or target proteins, marks those proteins for degradation by proteasomes. The ubiquitin pathway requires three enzymes : E1, ubiquitin activating enzyme; E2, ubiquitin conjugating enzyme ; and E3, ubiquitin ligase. The E3 enzyme transfers (ligates) ubiquitin to the ^NH2 groups of lysine residues on the target protein to yield polyubiquitin chains. 3.4. Gene pro¢ling with low density gene arrays on nylon membranes The advent of gene array (or cDNA microarray) technology has enabled us to use a more powerful method for analyzing di¡erential gene expression after noise trauma in the mammalian auditory system. We are initially using low density arrays of known rat genes on nylon membranes (Atlas Gene Arrays, Clontech).
HEARES 3515 18-8-00
298
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
Fig. 3. Human RNA dot-blot analysis. The blot of mRNA from human tissues was hybridized with an antisense riboprobe derived from the 3P UTR region of the human partial cDNA (KH438). The human gene is clearly expressed in many tissues, consistent with the identi¢cation of human ESTs from many cDNA libraries.
The Atlas Gene Arrays contain short cDNAs from 588 genes grouped according to biochemical function. We used these nylon membranes ¢rst for gene pro¢ling studies in the mammalian auditory system. The purpose of these ¢rst experiments is to determine which mRNAs for various genes or biochemical pathways can be detected in whole cochlea, subregions of cochlea and re-
gions of the auditory pathway in the brain. Atlas Gene Array membranes were hybridized with total RNA from two neuronal regions, the modiolus of the cochlea and the inferior colliculus (IC) of the brain, scanned in a phosphorimager, and analyzed with Clontech's AtlasImage software (see Section 2 for details). We focused on the genes for apoptosis-related pro-
Fig. 4. The chick and human UBE3B proteins contain a HECT domain with a conserved active site. BLAST search of the non-redundant GenBank database of known proteins identi¢ed several proteins with high sequence identity to either chick or human UBE3B. Alignment of the Ctermini of the UBE3B ubiquitin ligases with a known ubiquitin ligase (E6-AP) reveals high sequence conservation in the active site region, which includes the invariant Cys residue (C) to which the ubiquitin is transferred. The sequence preceding the active site is quite variable among these proteins and may de¢ne di¡erent subfamilies. The sequence of chick and human UBE3B is most similar to the Caenorhabditis elegans protein, both in the active site and the £anking region. Gaps (^) have been inserted to optimize the alignment. The GenBank accession numbers of the sequences used are: E6-AP (2340821); C. elegans (3880812); Schizosaccharomyces pombe Pub1 (3218406); mouse Itch (AF037454); mouse Nedd-4 (1709250); human NEDD-4 (1171682); KIAA0010 (265983); KIAA0312 (2224565); KIAA0322 (2224585). Several HECT domain E3 ligases have been identi¢ed in the databases through their high sequence identity to the C-terminus of E6-AP; however, the function of these proteins and the identity of their targets are unknown. The poly-ubiquitinated protein targets are transported to the 26S proteasome, where the target protein is degraded and the ubiquitin monomers are released (Hershko, 1998; Hershko and Ciechanover, 1998).
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
Fig. 5. Gene pro¢ling of RNA from normal adult rat inferior colliculus and modiolus. The Atlas1 cDNA expression array (catalog number 7738-1, Clontech, Palo Alto, CA) contains species-speci¢c PCR products derived from the cDNAs of 588 known rat genes. These PCR fragments are spotted in duplicate (10 ng cDNA/dot) on nylon membranes. The genes are grouped into six blocks, labeled A^E. Each block contains functionally related genes (for more detail, see the Clontech webpage at http://www.clontech.com). For hybridization experiments, radiolabeled cDNA was synthesized according to the vendor's protocol with a slight modi¢cation (Clontech, Palo Alto, CA). Reverse transcriptase reaction was performed on 2 Wg of total RNA from either inferior colliculus (IC) or modiolus (MOD) in 10 Wl of a reaction mixture containing 20 nM each of gene-speci¢c primers, 500 WM each of dCTP, dGTP and dTTP, 1.2 WM of [K-32 P]dATP (3000 Ci/mmol), and 100 U of MMLV reverse transcriptase Superscript II (Life Technologies Inc., Rockville, MD) for 25 min at 50³C. After purifying the probe with CHROMA SPIN-200 DEPC-H2 O columns (Clontech, Palo Alto, CA), 1^2U106 cpm/ml of the 32 P-labeled ¢rst-strand cDNA probe in 10 ml of ExpressHyb1 hybridization solution (Clontech, Palo Alto, CA) was hybridized to Atlas1 Rat cDNA Expression Array at 68³C overnight. The membranes were washed with 2USSC/1% SDS four times for 30 min each and two more times with 0.1USSC/0.5% SDS for 30 min each at 68³C and exposed to a Storage Phosphor screen for phosphorimaging overnight. The screen was scanned with phosphorimager 445 SI (Molecular Dynamics, Sunnyvale, CA). Block C of the array containing apoptosis (C1a^C3e)- and nervoussystem (C3f^C7n)-related genes from the resulting image ¢le is shown above. (A) The signals on block C when hybridized with labeled cDNA from IC. (B) The signals on block C when hybridized with labeled cDNA from MOD.
teins (C1a^C3e) and nervous-system-related proteins (C3f^C7n), which are contained in one subregion of the membrane (panel C). Comparing the nervous-system-related proteins (C3f^C7n) in two neuronal regions, the modiolus and IC, we observed that only three genes were di¡erentially expressed based on the threshold sensitivity de¢ned (200% above background) (Fig. 5). Myelin proteolipid protein (PLP; C3f) is the major protein component of CNS myelin, whereas peripheral myelin protein (PMP22 ; C3h), as the name indicates, is the major protein component of myelin in peripheral nerves. As expected, PLP was expressed at higher levels in IC than in modiolus, although there are signi¢cant amounts of PLP mRNA in modiolus. In contrast, the PMP22 mRNAs levels were higher in modiolus than IC. The mRNAs for both genes are fairly abundant and generate signals that are signi¢cantly above the sensitivity threshold. The PLP signal in IC was 2.5
299
times that in modiolus; conversely, the PMP22 signal was 5.8 times higher in modiolus than IC. Both genes were detected in random sequencing of cDNAs from a human fetal cochlear cDNA library, again con¢rming that the PLP gene for CNS myelin was also expressed in peripheral tissues like the cochlea. Mutations in both genes lead to neurological disorders. A third gene with signi¢cant di¡erences between IC and modiolus encodes the plasma membrane calcium transporting ATPase isoform 2 (ATP2B2 or PMCA ; C6g), one of four di¡erent ATP2B/PMCA isoforms in mammals. ATP2B2 mRNA was above the threshold sensitivity in both neuronal tissues, and was about twice as high in IC as in modiolus. PMCA2 exhibits a highly restricted tissue distribution, suggesting that it serves more specialized physiological functions than some of the other calcium transporting ATPase isoforms. Mutation analysis suggests that the PMCA2 gene may play a unique role in hearing. Street et al. (1998) showed that the deafwaddler mouse mutant (dfw), which is deaf and has vestibular/motor imbalance, has a mutation in the Atp2b2 gene. Furthermore, Kozel et al. (1998) assessed the physiological role of PMCA2 in the inner ear by producing PMCA2-de¢cient mice through gene targeting. These mice were also deaf and had balance problems. Histological analysis of both mouse mutants demonstrated high level expression of Atp2b2 in the cochlea, primarily in outer hair cells and spiral ganglion cells (Kozel et al., 1998 ; Street et al., 1998). Thus, a gene with a limited expression pattern (only ¢ve cDNAs identi¢ed by random sequencing of brain cDNA libraries), but a well-documented role in the ear can be detected by the gene array technique. 4. Discussion Di¡erential display analysis of regeneration in the chick basilar papilla suggests that proteins involved in actin polymerization, and in signaling through the actin cytoskeleton, may play important roles. Lee and Cotanche (1995) detected higher levels of L-actin mRNA in the chick basilar papilla after noise trauma by using a RT-PCR assay speci¢c for L-actin. We con¢rmed this observation in our slot-blot hybridization experiment, which included L-actin cDNA. Furthermore, we identi¢ed genes involved in actin signaling, including CDC42 (Gong et al., 1997) and WDR1 (Adler et al., 1999). CDC42 is a member of the Rho family of small GTPases known to be involved in actin signaling. CDC42 is required for the later, elongation steps of actin ¢ber formation (Kollmar, 1999). WDR1, a 68kDa protein with nine WD40 repeats, is the vertebrate homologue of yeast actin-interacting protein 1 (AIP1), which was ¢rst identi¢ed in a two-hybrid screen for
HEARES 3515 18-8-00
300
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
actin binding proteins. Recent studies de¢ne the function of yeast AIP1p ; it interacts with co¢lin to disassemble actin ¢laments (Rodal et al., 1999). DAip1, a Dictyostelium homologue of the yeast AIP1, is involved in endocytosis, cytokinesis, and motility (Konzok et al., 1999). It is not known whether vertebrate WDR1 serves similar functions in disassembling actin ¢laments, or whether it serves a structural role in its interactions with actin. These experiments identi¢ed two genes that were clearly di¡erentially expressed (upregulated) after noise trauma and that encoded novel proteins. The partial cDNA designated KH129 is the 3P end of a previously unrecognized 10-kb mRNA that encodes a 204-kDa kinesin of the UNC-104/KIF1 family (Gong et al., 1999). The 10-kb transcript is derived from the Kif1b gene by alternative splicing. The second gene (UBE3B) was partially cloned in this study and identi¢ed as a new member of a family of functionally related proteins de¢ned by a conserved C-terminal 350-amino acid `HECT' domain. These proteins are all related to human E6-AP (UBE3A), the ¢rst known E3 ubiquitin ligase (Sche¡ner, 1998). E6-AP was initially identi¢ed as a cellular protein that mediates the in vitro association of the human papillomavirus E6 oncoprotein with p53, leading to the ubiquitin-dependent turnover of p53, a tumor suppressor protein (Huibregtse et al., 1991, 1998; Sche¡ner et al., 1993). This leads to immortalization and oncogenesis. Ubiquitin-mediated protein turnover is central to many cellular processes, including di¡erentiation, proliferation, oncogenesis, and apoptosis. The HECT domain contains the conserved active site of the enzyme and is a hallmark of this family of E3 ubiquitin ligases (Hochstrasser, 1996; Hershko, 1998; Hershko and Ciechanover, 1998). UBE3B is thus a member of this important class of regulatory proteins. Mutations in the human E6-AP gene (UBE3A), the ¢rst E3 ubiquitin ligase identi¢ed, cause Angelman syndrome (Matsuura et al., 1997), which involves severe developmental and neurological problems. This was the ¢rst example of a genetic disorder of the ubiquitin-dependent proteolytic pathway. E6-AP targets the tumor suppressor p53 for destruction when papilloma viruses infect cells, thus leading to cell proliferation and tumorigenesis (Sche¡ner, 1998). Ubiquitination, the addition of poly-ubiquitin chains to proteins, targets many additional important transcriptional regulators for rapid turnover, e.g., NF-UB, AP1 (c-Fos and c-Jun), and GCN4 in yeast (Hochstrasser and Kornitzer, 1998). The UBE3B gene exhibits the greatest di¡erential expression after acoustic trauma of any gene that we have analyzed to date. The novel ubiquitin ligase identi¢ed in our experiments may target an important regulatory protein for rapid turnover and a¡ect important physio-
logical processes in the cell. Further experiments are required to identify the protein or proteins that the UBE3B ubiquitin ligase targets for rapid turnover, and to determine the roles of both the UBE3B protein and its target in the chick cochlea following noise trauma. Di¡erential display (and subtractive hybridization) clearly has the power to identify genes that exhibit the desired pattern of expression ; however, only a few genes can be analyzed because of the considerable time and e¡ort required to completely characterize each gene. It became clear, as we prepared to analyze di¡erential gene expression in mammals after noise trauma, that limiting the analysis initially to well-characterized genes might be faster and more productive. Soares (1997) suggested that the resources available through the human, mouse and rat genome projects should expedite and facilitate the analysis of di¡erential gene expression in mammals. In fact, this insight led to the development of gene arrays, which are essentially slot-blot experiments where the slots have been reduced to dots and the number of genes increased from 10 to hundreds and thousands. In the following section we discuss our preliminary results on `gene pro¢ling' in the auditory system using low density gene arrays. In gene pro¢ling experiments, the concentration of the `probe', i.e., the amount of unlabeled cDNA for the gene of interest, is determined by the size of the spot or microdot on the array. The concentration of the labeled cDNA in the `target' is determined by the relative abundance of the mRNA in that tissue, which determines whether or not a su¤cient amount of labeled cDNA will bind to the cDNA probe during the hybridization experiment. Thus, it is important to understand that, in general, gene arrays and microarrays are sampling primarily the moderately abundant mRNAs, and not necessarily the important low and rare abundance mRNA classes. The fact that we detected preferential expression of CNS myelin (PLP) in brain regions and peripheral myelin (PMP22) in the cochlea was to be expected, as these myelin proteins are very abundant proteins and their mRNAs are in the moderately abundant class. The fact that we could detect signi¢cant levels of mRNA for the plasma membrane calcium ATPase (PMCA2 or ATP2B2) was not anticipated. In fact, very few cDNAs derived from this gene were found by random sequencing of brain cDNAs (mouse EST sequencing project), and none was found in the Morton fetal cochlear cDNA library (Skvorak et al., 1999). Nevertheless, using the Clontech gene arrays we detected signi¢cant levels of mRNA for this gene in cochlea and all auditory brain regions examined (Cho and Lomax, unpublished data). This exciting and gratifying result suggests that gene arrays may have a profound impact on the analysis of di¡er-
HEARES 3515 18-8-00
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
ential gene expression in the mammalian auditory system. Gene arrays will certainly provide an additional important method for determining which genes are expressed in the inner ear, even when expression in other tissues is quite low. DNA microarrays on glass slides are a major technological advance and the ultimate way to analyze di¡erential gene expression. An automated robotic spotting device developed by Patrick Brown, Stanford University (Schena et al., 1995, 1996, 1998) can generate many replicates of the gene arrays quickly and reproducibly. This technological advance enables researchers to analyze simultaneous changes in expression of many genes, from hundreds to thousands to whole genomes. Brown applied this technology to di¡erential gene expression studies of the entire yeast genome under di¡erent physiological conditions (Brown and Botstein, 1999) and also to mammalian culture systems (Iyer et al., 1999). This technology employs £uorescently labeled cDNAs from target RNA populations. cDNAs from two target populations can be labeled separately with di¡erent £uorochromes, combined in equal amounts, and hybridized in one experiment to the microdots on the glass slide. The relative signal intensity of the £uorochromes at each spot is a measure of relative gene expression and provides a more accurate measure than the nylon membranes, which must be analyzed in parallel. The relative abundance of a mRNA in the target RNA population will a¡ect the amount of signal obtained. Increasing the sensitivity of the gene pro¢ling approach will require methods to amplify the signal for the low and rare abundance classes, such as linear RNA ampli¢cation (Van Gelder et al., 1990; Eberwine et al., 1992). There is considerable expense in establishing this technology, including the cost of purchasing all human or mouse ESTs, the cost of the robotic spotting device and the laser confocal reader. Furthermore, the large amount of data generated from these arrays requires sophisticated bioinformatics approaches for analysis (Ermolaeva et al., 1998; Bassett et al., 1999; Iyer et al., 1999). Although this technology is not currently available to many researchers, it should become more accessible in the near future (see Cheung et al., 1999). In summary, gene arrays hold great promise for differential gene expression in the mammalian auditory system. Potential applications might include analysis of gene expression in response to various stimuli, in disease states, and in mouse mutants to detect downstream targets of transcription factors or changes in biochemical pathways in response to mutant proteins. Unfortunately, there is no systematic attempt to identify and map all avian expressed genes, i.e., an Avian Genome Project, therefore, we cannot apply gene arrays to one of the most interesting questions in the auditory system, the regenerating chick auditory epithelium.
301
Analysis of di¡erential gene expression in birds will require those molecular approaches summarized in Section 1 and illustrated in this article. Acknowledgements Supported by NIH Grants DC02492 (M.I.L.) and DC02982 (M.I.L. and R.A.A.).
References Adams, M., Kelley, J., Gocayne, J.D., Dubnick, M., Polymeropoulos, M.H., Xiao, H., Merril, C.R., Wu, A., Olde, B., Moreno, R.F., Kerlavage, A.R., McCombie, W.R., Venter, J.C., 1991. Complementary DNA sequencing: Expressed sequence tags and the human genome project. Science 252, 1651^1656. Adler, H.J., Niemec, A.J., Moody, D.B., Raphael, Y.R., 1995. Tectorial membrane regeneration in acoustically damaged birds: An immunocytochemical technique. Hear. Res. 86, 43^46. Adler, H.J., Winnicki, R.S., Gong, T-W.L., Lomax, M.I., 1999. A gene upregulated in the acoustically damaged chick basilar papilla encodes a protein with WD40 repeats. Genomics 56, 59^69. Bassett, D.E., Jr., Eisen, M.B., Boguski, M.S., 1999. Gene expression informatics ^ it's all in your mine. Nat. Genet. 21, 51^55. Bonaldo, F.M., 1996. Normalization and subtraction: Two approaches to facilitate gene discovery. Genome Res. 6, 791^806. Brown, P.O., Botstein, D., 1999. Exploring the new world of the genome with DNA microarrays. Nat. Genet. 21, 33^37. Cheung, V.G., Morley, M., Aqilar, F., Massimi, A., Kucherlapati, R., Childs, G., 1999. Making and reading microarrays. Nat. Genet. 21, 15^19. Chomczynski, P., Sacchi, N., 1987. Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156^159. de Kok, Y.J., Bom, S.J., Brunt, T.M., Kemperman, M.H., van Beusekom, E., van der Velde-Visser, S.D., Robertson, N.G., Morton, C.C., Huygen, P.L., Verhagen, W.I., Brunner, H.G., Cremers, C.W., Cremers, F.P., 1999. A Pro51Ser mutation in the COCH gene is associated with late onset autosomal dominant progressive sensorineural hearing loss with vestibular defects. Hum. Mol. Genet. 8, 361^366. Duggan, D.J., Bittner, M., Chen, Y., Meltzer, P., Trent, J.M., 1999. Expression pro¢ling using cDNA microarrays. Nat. Genet. 21, 10^ 14. Eberwine, J., Yeh, H., Miyashiro, K., Cao, Y., Nair, S., Finnell, R., Zettell, M., Coleman, P., 1992. Analysis of gene expression in single live neurons. Proc. Natl. Acad. Sci. USA 89, 3010^3014. Ermolaeva, O., Rastogi, M., Pruitt, K.D., Schuler, G.D., Bittner, M.L., Chen, Y., Simon, R., Meltzer, P., Trent, J.M., Boguski, M.S., 1998. Data management and analysis for gene expression arrays. Nat. Genet. 20, 19^23. Gerhold, D., Rushmore, T., Caskey, C.T., 1999. DNA chips: promising toys have become powerful tools. Trends Biochem. Sci. 24, 168^173. Gong, T.-W.L., Hegeman, A.D., Shin, J.J., Adler, H.J., Raphael, Y., Lomax, M.I., 1996. Identi¢cation of genes expressed after noise exposure in the chick basilar papilla. Hear. Res. 96, 20^32. Gong, T.-W.L., Shin, J.J., Burmeister, M., Lomax, M.I., 1997. Complete cDNAs for CDC42 from chicken cochlea and mouse liver. Biochim. Biophys. Acta 1352, 282^292. Gong, T.L., Winnicki, R.S., Kohrman, D.C., Lomax, M.I., 1999. A
HEARES 3515 18-8-00
302
M.I. Lomax et al. / Hearing Research 147 (2000) 293^302
novel mouse kinesin of the Unc-104/KIF1 subfamily encoded by the Kif1b gene. Gene 239, 117^127. Hershko, A., 1998. The ubiquitin system: Past, present, and future perspectives. In: Peters, J.-M., Harris, J.R. and Finley, D. (Eds.), Ubiquitin and the Biology of the Cell. Plenum Press, New York, pp. 1^13. Hershko, A., Ciechanover, A., 1998. The ubiquitin system. Annu. Rev. Biochem. 67, 425^479. Hochstrasser, M., 1996. Ubiquitin-dependent protein degradation. Annu. Rev. Genet. 30, 405^439. Hochstrasser, M., Kornitzer, D., 1998. Ubiquitin-dependent inactivation of transcription regulators. In: Peters, J.-M., Harris, J.R. and Finley, D. (Eds.), Ubiquitin and the Biology of the Cell. Plenum Press, New York, pp. 279^302. Huibregtse, J.M., Sche¡ner, M., Howley, P.M., 1991. A cellular protein mediates association of p53 with the E6 oncoprotein of human papillomavirus types 16 or 18. EMBO J. 10, 4129^4135. Huibregtse, J.M., Maki, C.G., Howley, P.M., 1998. Ubiquitination of the p53 tumor suppressor. In: Peters, J.-M., Harris, J.R. and Finley, D. (Eds.), Ubiquitin and the Biology of the Cell. Plenum Press, New York, pp. 323^343. Iyer, V.R., Eisen, M.B., Ross, D.T., Schuler, G., Moore, T., Lee, J.C.F., Trent, J.M., Staudt, L.M., Hudson, J., Jr., Boguski, M.S., Lashkari, D., Shalon, D., Botstein, D., Brown, P.O., 1999. The transcriptional program in the response of human ¢broblasts to serum. Science 283, 83^87. Kollmar, R., 1999. Who does the hair cell `do'? Rho GTPases and the hair-bundle morphogenesis. Curr. Opin. Neurobiol. 9, 394^398. Konzok, A., Weber, I., Simmeth, E., Hacker, U., Maniak, M., Muller-Taubenberger, A., 1999. DAip1, a Dictyostelium homologue of the yeast actin-interacting protein 1, is involved in endocytosis, cytokinesis, and motility. J. Cell Biol. 146, 453^464. Kozel, P.J., Friedman, R.A., Erway, L.C., Yamoah, E.N., Liu, L.H., Riddle, T., Du¡y, J.J., Doetschman, T., Miller, M.L., Cardell, E.L., Shull, G.E., 1998. Balance and hearing de¢cits in mice with a null mutation in the gene encoding plasma membrane Ca2 -ATPase isoform 2. J. Biol. Chem. 273, 18693^18696. Lee, K.H., Cotanche, D.A., 1995. Detection of L-actin mRNA by RTPCR in normal and regenerating chicken cochlae. Hear. Res. 87, 9^15. Lennon, G., Au¡ray, C., Polymeropoulos, M., Soares, M.B., 1996. The I.M.A.G.E. Consortium: an integrated molecular analysis of genomes and their expression. Genomics 33, 151^152. Liang, P., Pardee, A.B., 1992. Di¡erential display of eukaryotic messenger RNA by means of the polymerase chain reaction. Science 257, 967^971. Liang, P., Averboukh, L., Pardee, A.B., 1993. Distribution and cloning of eukaryotic mRNAs by means of di¡erential display: re¢nements and optimization. Nucleic Acids Res. 21, 3269^3275. Matsuura, T., Sutcli¡e, J.S., Fang, P., Galjaard, R.-J., Jiang, Y., Benton, C.S., Rommens, J.M., Beaudet, A.L., 1997. De novo truncating mutations in E6-AP ubiquitin-protein ligase gene (UBE3A) in Angelman syndrome. Nat. Genet. 15, 74^77. McClelland, M., Mathieu-Daude, F., Welsh, J., 1995. RNA ¢ngerprinting and di¡erential display using arbitrarily primed PCR. Trends Genet. 11, 242^246. Raphael, Y., 1993. Reorganization of the chick basilar papilla following acoustic trauma. J. Comp. Neurol. 330, 521^532. Raphael, Y., Altschuler, R.A., 1992. Early micro¢lament reorganization in injured auditory epithelia. Exp. Neurol. 115, 32^36. Robertson, N.G., Khetarpal, U., Gutierrez-Espeleta, G.A., Bieber, F.R., Morton, C.C., 1994. Isolation of novel and known genes from a human fetal cochlear cDNA library using subtractive hybridization and di¡erential screening. Genomics 23, 42^50.
Robertson, N.G., Skvorak, A.B., Yin, Y., Weremowicz, S., Johnson, K.R., Kovatch, K.A., Battey, J.F., Bieber, F.R., Morton, C.C., 1997. Mapping and characterization of a novel cochlear gene in human and in mouse: a positional candidate gene for a deafness disorder, DFNA9. Genomics 46, 345^354. Robertson, N.G., Lu, L., Heller, S., Merchant, S.N., Eavey, R.D., McKenna, M., Nadol, J.B.Jr., Miyamoto, R.T., Linthicum, F.H.Jr., Lubianca Neto, J.F., Hudspeth, A.J., Seidman, C.E., Morton, C.C., Seidman, J.G., 1998. Mutations in a novel cochlear gene cause DFNA9, a human nonsyndromic deafness with vestibular dysfunction. Nat. Genet. 20, 299^303. Rodal, A.A., Tetreault, J.W., Lappalainen, P., Drubin, D.G., Amberg, D.C., 1999. Aip1p interacts with co¢lin to disassemble actin ¢laments. J. Cell Biol. 145, 1251^1264. Sche¡ner, M., 1998. Ubiquitin, E6-AP, and their role in p53 inactivation. Pharmacol. Ther. 78, 129^139. Sche¡ner, M., Huibregtse, J.M., Vierstra, R.D., Howley, P.M., 1993. The HPV-16 E6 and E6-AP complex functions as a ubiquitinprotein ligase in the ubiquitination of p53. Cell 75, 495^505. Schena, M., Shalon, D., Davis, R.W., Brown, P.O., 1995. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 270, 467^470. Schena, M., Shalon, D., Heller, R., Chai, A., Brown, P.O., Davis, R.W., 1996. Parallel human genome analysis: Microarray-based expression monitoring of 1000 genes. Proc. Natl. Acad. Sci. USA 93, 10614^10619. Schena, M., Heller, R.A., Theriault, T.P., Konrad, K., Lachenmeier, E., Davis, R.W., 1998. Microarrays: biotechnology's discovery platform for functional genomics. Trends Biotechnol. 16, 301^306. Skvorak, A.B., Robertson, N.G., Yin, Y., Weremowicz, S., Her, H., Bieber, F.R., Beisel, K.W., Lynch, E.D., Beier, D.R., Morton, C.C., 1997. An ancient conserved gene expressed in the human inner ear: identi¢cation, expression analysis, and chromosomal mapping of human and mouse antiquitin (ATQ1). Genomics 46, 191^199. Skvorak, A.B., Weng, Z., Yee, A.J., Robertson, N.G., Morton, C.C., 1999. Human cochlear expressed sequence tags provide insight into cochlear gene expression and identify candidate genes for deafness. Hum. Mol. Genet. 8, 439^452. Soares, M.B., 1997. Identi¢cation and cloning of di¡erentially expressed genes. Curr. Opin. Biotechnol. 8, 542^546. Soares, M.B., Bonaldo, M.F., Jelene, P., Su, L., Lawton, L., Efstratiadis, A., 1994. Construction and chracterization of a normalized cDNA library. Proc. Nat. Acad. Sci. USA 91, 9228^9232. Staecker, H., Van de Water, T.R., 1998. Factors controlling hair-cell regeneration/repair in the inner ear. Curr. Opin. Neurobiol. 8, 480^487. Stone, J.S., Oesterle, E.C., Rubel, E.W., 1998. Recent insights into regeneration of auditory and vestibular hair cells. Curr. Opin. Neurol. 11, 17^24. Street, V.A., McKee-Johnson, J.W., Fonseca, R.C., Tempel, B.L., Noben-Trauth, K., 1998. Mutations in a plasma membrane Ca2 -ATPase gene cause deafness in deafwaddler mice. Nat. Genet. 19, 390^394. Van Gelder, R.N., Von Zastrow, M.E., Yool, A., Dement, W.C., Barchas, J.D., Eberwine, J.H., 1990. Ampli¢ed RNA synthesized from limited quantities of heterogeneous cDNA. Proc. Natl. Acad. Sci. USA 87, 1663^1667. Welsh, J., Chada, K., Dalal, S.S., Cheng, R., Ralph, D., McClelland, M., 1992. Arbitrarily primed PCR ¢ngerprinting of RNA. Nucleic Acids Res. 20, 4965^4970.
HEARES 3515 18-8-00