Main

The N-myc gene plays an essential role in organogenesis, and overexpression due to genomic amplification has been observed in many human tumors. N-myc belongs to a small family of oncogenes that also include c- and L-myc. Transcription factors that modulate the expression of target genes involved in cell cycle control and cell differentiation are encoded by myc genes(1). In contrast with c-myc, the N-myc gene displays a restricted expression profile with expression being observed in the epithelial component of multiple organs during early development(2–4). The importance of N-myc in organogenesis is emphasized by experiments showing that targeted gene disruption results in embryonic lethality due to organ hypoplasia(4,5). Expression is markedly decreased after birth with only transient expression in pre-B cells(6).

N-myc gene amplification is observed in the common childhood tumor neuroblastoma, and this property is associated with an extremely poor prognosis(7,8). Amplification remains the most adverse prognostic feature, and patients whose tumors harbor this genetic phenotype continue to do poorly even with aggressive treatment, including bone marrow transplantation(9,10). In an effort to identify repression domains that could be used to develop therapeutic strategies to decrease expression in amplified tumors, we have analyzed the human N-myc promoter.

Like c-myc, the N-myc gene is composed of three exons, the first of which is not translated. We and others have noted previously that promoter reporter genes containing N-myc sequences 5′ to exon 1 were expressed in all cell types, even those without expression of the endogenous gene(11–14). However, inclusion of a 910-bp region composed of 3′ exon 1 and 5′ intron 1 resulted in an appropriate pattern of reporter gene expression; reporter activity was observed only in cells with endogenous N-myc transcripts(11).

To determine the role of this TSE in mediating the temporal and spatial pattern of N-myc expression during development, we generated a series of transgenic animal lines carrying the human N-myc promoter with and without the associated TSE linked to the bacterial β-galactosidase (lacZ) reporter gene.

We show that the 1.8-kb 5′ N-myc promoter directed reporter gene expression only in the CNS in newborn animals. The restricted distribution of transgene expression is in contrast with the ubiquitous expression of the endogenous murine gene. Expression in the CNS was observed in the olfactory system and was more restricted than noted previously for the endogenous gene. No reporter activity was observed in transgenic animals carrying the 910-bp 3′ exon 1/intron 1 TSE. Therefore, the TSE represses expression from immediate 5′ promoter elements. These results, not surprisingly, indicate a complex mode of developmental regulation of the human N-myc gene. Further experiments are now directed at defining the subregions within exon 1/intron 1 that direct gene silencing and the mode of this negative regulation.

METHODS

Transgene construction. N-myc promoter transgenes were generated by cloning N-myc promoter regions into the β-galactosidase reporter plasmid pnlacF. The pSI/B N-myc lacZ and pSI/SpI N-myc lacZ minigenes were constructed by ligation of a SacI/BamHI 5′ N-myc promoter fragment (-1682-151, relative to the major N-myc transcription start site) and a larger SacI/SpeI fragment (which included 3′ exon 1 and 5′ intron 1; -1682-1061), respectively, into pnlacF (Fig. 1). N-myc promoter fragments were blunt-ended using Klenow polymerase and cloned into the SmaI site of pnlacF. Provided by Dr. Jacques Peschon, pnlacF is a PUC18-based vector containing the Escherichia coli lacZ gene, a eukaryotic ATG initiation codon, an SV40 nuclear transport signal, and a mouse protamine structural gene providing an artificial intron(15).

Figure 1
figure 1

N-myc transgenes. The human N-myc genomic locus (top) and the N-myc promoter reporter transgenes are illustrated. Exons are represented by boxes. Untranslated regions are illustrated in black, translated regions in white. Restriction enzyme sites: E, EcoRI; SI, SacI; B, BamHI.

Full size image

The N-myc fragments used in these constructs contained all or part of nontranslated exon 1 and, thus, did not contain the N-myc translational initiation site. Thus, N-myc promoter sequences did not disrupt protein expression from the lacZ gene. In preliminary control experiments, we documented appropriate β-galactosidase expression from both pSI/B N-myc lacZ and pSI/SpI N-myc lacZ using the murine neuroblastoma lines NB41A3 and Neuro 2a (ATCC CCL 133 and 147, data not shown).

Transgenic mouse production. The pSI/B N-myc lacZ and pSI/SpI N-myc lacZ minigenes were linearized with KpnI and purified by gel elution (Qiagen gel purification kit) and elution from an Elutip minicolumn (Schleicher and Schuell). Transgenic founders were produced by pronuclear injection of the linearized DNA into BCF1 zygotes as described(16). Animals were generated through the University of Utah Transgenic Core Facility. Tail DNA was harvested from pups after weaning, and founders were identified by PCR using primers specific for the human N-myc promoter and the β-galactosidase gene. The sequences of sense oligonucleotide primers used were the following: (5′)GGCCCAAGCTTAGACACCCGCGCAGAA(3′) (N-myc exon 1, nt 121-141 relative to major transcription start site) (17) or (5′)AATCTGGATCCCGGCTGCTCCAGCTTGGAG(3′) (N-myc intron 1, nt 426-456). A single antisense primer (5′)AAAGCGCCATTCGCCATTCA(3′) (nt 269-289 in pnlacF) was used that hybridized to the lacZ gene. PCR amplification was performed using 400 ng of tail DNA in a 50-µL reaction containing 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 5 mM MgCl2, 200 mM each dNTP, 200 nM each primer, and 0.05 units of AmpliTaq DNA polymerase (5U/µL, Promega, GIBCO-BRL). Cycle times and conditions were piloted and were found to be optimal at 94°C (denaturation) for 1 min, 60°C (annealing) for 30 s, and (extention) at 72°C for 90 s. A total of 30 cycles were used. To estimate transgene copy number, a simultaneous PCR reaction was performed on a control sample in which a known copy/cell equivalent of the transgene was diluted into 400 ng of normal mouse genomic DNA.

Transgenic founder mice were crossed with C57BL/6BKL hybrid mates to give rise to F1 progeny. Transgene-positive F1 animals were identified again as described above. All animal experimentation was performed according to guidelines established by the National Institutes of Health, and protocols were approved by the Animal Care and Use Committee of the University of Utah.

RT-PCR amplification. A semiquantitative RT-PCR was used to detect expression of the human N-myc lacZ transgene as well as the endogenous murine N-myc gene in organs from transgenic animals(18). Total RNA was isolated from organs by acid guanidine thiocyanate-phenol-chloroform extraction with minor modifications(19). Reverse transcription was carried out using 5 µg of total RNA in a 50-µL reaction mixture containing RT buffer (GIBCO-BRL), 10 mM DTT, 0.125 mM each deoxyribonucleotide triphosphate (GIBCO-BRL), 0.5 µg of random primers (Boehringer Mannheim), and 400 units of SuperScript II RNase H- RT (GIBCO-BRL). After a 60-min incubation at 37°C, 2 µg of DNase-free RNase was added and the reaction was incubated for an additional 5 min at 37°C. The reaction was purified by phenol/chloroform, chloroform extraction, and the cDNA was precipitated with ethanol at -20°C overnight. cDNA concentration was measured spectrophotometrically.

Semiquantitative PCR amplification was done with primers specific for the transgene and endogenous murine N-myc transcript using a rapid air thermocycler as described(18). cDNA amplification was accomplished in a 10-µL reaction that contained 70 pmol each primer, reaction buffer (50 mM Tris, pH 8.3, 3 mM MgCl2, 20 mM KCl, and 0.5 mg/mL BSA), 0.8 mM dNTP, 2.5 µCi [32P]dCTP (3000 Ci/mmol; New England Nuclear, Boston, MA), 0.72 U of AmpliTaq DNA polymerase (5 U/µL; Promega, GIBCO-BRL), and 200 ng of cDNA. β-actin was used as an internal standard to equalize cDNA concentration and to compare lacZ expression between samples. PCR reactions were loaded into glass microcapillary tubes (#1705, Idaho Technology, Idaho Falls, ID), and amplification was done using the 1605 Air Thermocycler (Idaho Technology).

Pilot PCR reactions were performed to optimize conditions for each transcript. Each cycle entailed denaturation for 0 s at 94°C, annealing for 0 s at an optimal temperature, and elongation for 6 s for mouse β-actin and 10 s for mouse N-myc and transgene mRNA at 72°C. The annealing temperatures for these reactions were 59, 60, and 66°C, and a total of 16, 24, and 30 cycles were used, respectively, to give a semiquantitative estimate of mRNA levels.

The sequences of oligonucleotide primers used for PCR amplification were the following: β-actin sense (5′)TCCATCGTGGGCCGCTCTAG(3′) (exon 1, nt 159-179) and antisense (5′)GTAACAATGCCATGTTCAAT(3′) (exon 2, nt 273-293); for the endogenous mouse N-myc gene, sense (5′)AGCCCCTCCGGATCCCCGGC(3′) (exon 1, nt 137-157) and antisense (5′)AGCCCAATTCGAAGGCTCCGG(3′) (exon 2, nt 1552-1572); and for the transgene, sense (5′)TCGGCGGAATTCCAGCTGA(3′) (lacZ, pnlacF nt 2019-2038) and antisense (5′)TGTTCCTTAGCAGGCTCCTG(3′) (mouse protamine, pnlacF nt 2733-2755).

After PCR, the ends of the microcapillary tubes were scored, and samples were removed with a Captrol microaspirator. All 10 µL of the sample was added to an equal volume of stop solution (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, and 0.05% xylene cyanol), heated to 95°C for 5 min, and 4 µL was electrophoresed in a 6% acrylamide gel. Radiolabeled PCR products were detected by autoradiographic exposure overnight at -70°C. Product sizes were estimated by the migration of 32P-end-labeled MspI digest of pBR322. Quantitation of autoradiographic signal was performed on a Series 300 scanning densitometer with ImageQuant 3.2 software (Molecular Dynamics, Sunnyvale, CA).

β-galactosidase staining. Mice were euthanized by cervical dislocation after anesthesia with Metofane (Mallinckrodt Veterinary, Inc., Mundelein, IL) and perfused with 2% formaldehyde [(HCHO, ultrapure)/PBS]. Brains were then removed and postfixed in 2% HCHO/0.1 M PIPES pH 6.9, 2 mM MgCl2, 1.25 mM EGTA for 1 h at 4°C. Brains were then rinsed three times (20 min each) in PBS, 1 mM MgCl2, 0.02% NP-40, and 0.01% NaDOC at room temperature. Brains were rinsed again in KFeCN solution [5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6·3H2O, and 1 mM MgCl2 in PBS (pH 7.5)] with 0.02% NP-40, 0.01% NaDOC for 15 min at room temperature and were then stained in KFeCN solution, 0.02% NP-40, 0.01% NaDOC, 1 mg/mL X-gal (5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside) in a humidified atmosphere at 37°C overnight. Samples were fixed again in 4% HCHO, PBS at 4°C overnight, then dehydrated in the series of ethanol solutions and embedded in paraffin. Paraffin-embedded brains were cut in 5-µm sections, and the slides were counterstained with eosin.

Northern blot. Twenty micrograms of total RNA was fractionated by electrophoresis on a 1% agarose-formaldehyde gel, transferred to an activated nylon membrane, and hybridized with a 32P radiolabeled 169-bp EcoRI BamHI mouse N-myc cDNA fragment that contained exons 1 and 2. After hybridization, filters were washed successively in 2X SSC, 0.1% SDS, 0.2X SSC, 0.1% SDS and subjected to autoradiography overnight.

RESULTS

The N-myc 3′ exon 1/5′ intron region has a negative regulatory role in early development. Transgenic founder lines were generated with two N-myc promoter constructs, each directing expression of the E. coli β-galactosidase reporter gene(15). pSI/B N-myc lacZ contains 1.8 kb of the N-myc promoter extending 5′ to the BamHI site in exon 1. pSI/SpI contains the TSE, an additional 910-bp region including 3′ exon 1 and a portion of intron 1 (Fig. 1). Multiple independent founder lines were generated for each construct, and copy number was estimated by PCR (Fig. 2). Copy number ranged from 1 to 100 in the lines analyzed (Table 1). Because expression may be influenced by integration site and copy number, we examined transgene expression in multiple lines in newborn animals by semiquantitative PCR. Two of four lines containing the pSI/B N-myc lacZ minigene (without 3′ exon 1/5′ intron 1) showed strong expression in the brain, an organ showing strong expression of the endogenous gene (Fig. 3)(2–4,20). No expression was observed in other organs despite widespread expression of the endogenous N-myc gene, including organs like the kidney, another organ previously shown to express large amounts of N-myc. Two lines failed to show any expression, likely reflecting integration effect (data not shown)(21).

Figure 2
figure 2

Transgene copy number. Gene copy number was estimated by semiquantitative PCR by comparing product intensity (left) to a known amount of copy/cell equivalent of plasmid diluted into normal mouse genomic DNA (right).

Full size image
Table 1 Characterization of N-myc promoter-reporter transgenes
Full size table
Figure 3
figure 3

Spatial pattern of transgene expression. Transgene expression was assayed by RT-PCR (top panel) and compared with the expression of the endogenous mouse gene (middle panel). An actin control (lower panel) was run simultaneously to verify equal cDNA substrate concentration among samples harvested from different organs. A comparison with a normal littermate (left) was run for each transgene founder line (right). Organs: B, brain; S, spleen; BM, bone marrow; K, kidney; L, lung; Sm, small intestine; Lg, large intestine; H, heart; Lu, lung; T, thymus. Positive control, RT-PCR on testes mRNA harvested from a transgenic animal expressing β-galactosidase from a human aquaporin-2 promoter (generously provided by Dr. Raoul Nelson).

Full size image

In contrast, in four separate founder lines analyzed, no expression from the pSI/SpI N-myc lacZ transgene was detected (Fig. 3, data not shown). These results indicate that the exon 1 intron 1 TSE exerts a negative regulatory influence. This repression domain seemed to be dominant over the immediate 5′ promoter regions (i.e. the 1.8-kb 5′ N-myc promoter) and was capable of silencing transgene expression in the CNS.

pSI/B N-myc transgene expression in the CNS is restricted to a subset of regions that show expression of the endogenous gene. Previous studies using in situ hybridization have shown that expression of the murine N-myc gene is restricted to a subset of cells within the CNS including those of the olfactory system, the neocortex, and the cerebellum(3,4). To determine whether pSI/B N-myc transgenes also directed expression in these same areas, we examined transgene expression in multiple litters of pSI/B N-myc lacZ. 116 animals by β-galactosidase staining. Whole brain staining of newborn and early adult (3 wk) animals showed strong expression in the olfactory system (Fig. 4A). This was confirmed by staining of tissue sections (Fig. 4B). In the olfactory bulb, lacZ staining was seen in the mitral and internal granular layers, regions previously shown to express N-Myc(3). However, unlike the robust and generalized expression reported for the endogenous gene in the neo-cortex, the maximum expression of pSI/B N-myc was predominantly restricted to the anterior insular cortex (Fig. 4), and only limited staining of sporadic neuronal-like bodies of the neocortex (predominantly olfactory cortex) was observed. At maximum transgene expression (3-5 wk), some limited staining of neuronal-like cells in the lateral hypothalamus and interpeduncular nucleus was present (not shown). Strong staining of cells within the trigeminal ganglia (Fig. 4A and diagram, Fig. 4B) was present from d 10 through 5 wk. Occasional staining of cells in the lateral hypothalamus persisted into the 10-wk-old adult. Rare staining was observed in the entorhinal cortex, superior colliculus, medial habenula, and prepositus hypoglossal nucleus and was limited to the 3-5-wk-old animal group. Also, during this time, stained cells could be observed in continuous sections from the anterior insular cortex to the cortical amygdala. Transgene expression was rarely seen in the posterior insular (granular) cortex nucleus. There was no detectable expression at any time from pSI/B N-myc lacZ in the cerebellum, another site of prominent endogenous N-myc expression. These results indicate a complex mode of N-myc regulation in the CNS. The immediate 1.8-kb human N-myc promoter is sufficient to direct N-myc expression in some but not all regions of the CNS. Sequences that direct expression in the olfactory network are present in our transgenes, but other important cis-acting regulatory regions that direct expression in the neocortex and cerebellum apparently lie elsewhere in the N-myc promoter.

Figure 4
figure 4

Transgene expression by β-galactosidase staining. lacZ staining in whole mount newborn brain (A, dorsal left, ventral right) and in coronal sections (B, 3-wk-old animal). A diagram of the mouse brain in sagittal section indicating approximate location of sections is shown. Abbreviations: AON, anterior olfactory nucleus; Cb, cerebellum; CAmg, cortical amygdala; GC, granular cells; IC, insular cortex; IPN, interpeduncular nucleus; LH, lateral hypothalamic nucleus; Mi, mitral cells; OB, olfactory bulb; PkJ, Purkinje cells; TGN, trigeminal nucleus. Bar corresponds to 100 µm.

Full size image

Temporal expression of the pSI/B N-myc transgene is extended beyond the newborn period. Under normal circumstances, N-myc expression is greatest during organogenesis in the embryo, and transcript levels diminish significantly at birth(2,4,20). To determine the temporal expression pattern of the pSI/B N-myc lacZ transgene, we analyzed expression in the CNS in pSI/B N-myc lacZ.116 mice at birth and at 3, 5, and 10 wk of age using β-galactosidase staining (Figs. 4B and 5). Although expression did decrease during postnatal life, this regression was significantly delayed compared with that reported for the endogenous gene(2). In fact, expression seemed to be optimal at 3-5 wk, a time when endogenous myc expression is diminished greatly (Figs. 4B and 5)(20). In all cases, expression from the pSI/B N-myc transgene diminished rapidly by 10 wk, with expression limited largely to only occasional cells of the olfactory bulb, anterior olfactory nucleus, and the insular cortex.

Figure 5
figure 5

Expression of pSI/B N-myc lacZ as a function of age. β-galactosidase staining in olfactory bulb/anterior olfactory nucleus in transgenic animals at birth and at 5 and 10 wk. Bar corresponds to 100 µm.

Full size image

To be certain that regression of endogenous N-myc expression was not influenced by the presence of the transgene, we analyzed murine N-myc transcript levels in newborn and adult (3 wk) from wild-type and transgenic animals. As reported previously, abundant N-myc expression was observed in the brain of newborn animals, and expression was absent in the adult brain as detected by Northern blot (Fig. 6)(2). The regression of endogenous N-myc expression was identical in wild-type and transgenic animals. Thus, it seems that N-myc expression during later stages of development are absent from the immediate 5′ promoter elements studied in this investigation.

Figure 6
figure 6

Expression of the endogenous mouse N-myc gene in transgenic animals. Northern blot (top) of RNA was harvested from brains of normal littermates (N) and transgenic (T) animals at birth (newborn) and 3 wk (adult). Ethidium-stained gel before transfer (bottom).

Full size image

DISCUSSION

The purpose of the present study was to establish the role of the human N-myc exon 1/intron region in directing expression in vivo. Previous work using established cell lines in vitro indicated that immediate 5′ promoter regions functioned promiscuously(12,13). However, association with a 910-bp region containing 3′ exon 1/5′ intron 1 restored authentic expression to promoter reporter genes because expression was observed only in cell lines in which expression of the endogenous gene was observed(11). One prediction from these studies would be that 5′ promoter regions were responsible for directing the ubiquitous expression observed in many organs during early development and that the exon 1/intron region directed the temporal extinction of N-myc expression postnatally. Not surprisingly, our study looking at in vivo regulation of the gene reveals a more complex mode of regulation.

The temporal and spatial patterns of expression of N-myc are restricted in comparison with the well-studied c-myc gene(2). The two promoters differ substantially. Expression of c-myc is governed by two major promoters, P1 and P2, each of which is preceded by a TATA box(22–24). The N-myc promoter, in contrast, contains a nonfunctional TATA box and has multiple transcription start sites(17). These differences suggested that mechanisms that regulate N-myc would differ from those that operate at the c-myc locus.

N-myc is expressed in the epithelial component of many organs during development, and targeted gene disruption leads to early embryonic lethality(4,5). Organs of such mice show correct formation of organ anlagen but profound cellular hypoplasia. These observations suggested that N-myc may be responsible for the maintenance (by enhancing proliferation and/or preventing apoptosis) of a common progenitor cell population(5). The N-myc gene may contain cis-regulatory elements that are responsible for directing expression in progenitor cells in all organs. We had speculated that the immediate 5′ promoter may contain such an element and that the 3′ exon 1/intron 1 TSE may be responsible for temporal regulation. Although our experiments do not exclude such a model, regulation seems to be more complicated.

Our results do suggest that exon 1/intron 1 contains negative regulatory subregions that may influence developmental expression. These results are in agreement with previous work. The role of exon 1/intron 1 elements in controlling myc expression was initially discovered by examination of the c-myc gene. Control of RNA polymerase II processivity (transcription attenuation) across c-myc coding sequences is an important mode of regulating steady state levels of myc transcript(22,25). Loss of attenuation is observed in Burkitt's lymphoma and leads to the dramatically increased c-myc mRNA levels in these tumors(22). An attenuation site has been mapped at the exon/intron junction, and attenuation is influenced by promoter use; transcripts initiated at the P1 promoter do not terminate at transcription attenuation signals, whereas P2 initiated transcripts may either terminate or read through(26). In Burkitt's lymphoma, point mutations either disrupt the attenuation signal (endemic Burkitt's lymphoma) or the translocation breakpoint actually disassociates coding portions of the gene (exon 1 and exon 2) from the exon 1/intron 1 regulatory region (sporadic Burkitt's lymphoma)(27). Deregulated expression of L-myc in some small-cell lung cancer lines is also associated with loss of attenuation in the absence of gene amplification(28).

These studies led investigators to examine the role of the exon 1/intron 1 region in controlling N-myc expression. Indeed, early studies showed that transcription attenuation accounted for the dramatic decrease in steady state N-myc mRNA levels in adult brain in the mouse(29). Nuclear run-on analysis localized the site of transcriptional blocking to a site of potential stem loop structure associated with a thymine tract located within exon 1 just 3′ to the BamHI site. Deletion of this region resulted in an increase in steady state N-myc mRNA levels and an increase in oncogenic potency as assessed by an N-myc/ras rat embryo fibroblast transformation assay(30,31). We and others also noted a role for the 3′ exon 1/intron 1 region in controlling expression of the human N-myc gene(11–13).

At least three subregions have been identified within the N-myc exon 1/intron 1 region used in these experiments that could account for the negative regulation that we have observed in our experiments. We have used promoter deletion analysis of the TSE to define elements that could modulate expression from the 5′ N-myc promoter. One region located within exon 1 between nt 151 to 273 was associated with a 2-fold reduction in promoter activity in nonneuroblastoma cell lines. Nuclear run-on analysis showed partial transcript attenuation across this region in these cell lines similar to what has been observed for the murine N-myc gene(14). However, attenuation could not account for the total absence of N-myc transcripts in nonneuroblastoma lines because equal polymerase density was seen at 3′ translated portions of the N-myc locus. A 116-bp region within intron 1, the tissue-specific element, or TSE, was sufficient to completely abrogate expression in N-myc-nonexpressing cells while maintaining expression in neuroblastoma cell lines. In contrast, nuclear run-on experiments indicated that the TSE functioned posttranscriptionally, possibly by influencing N-myc pre-mRNA stability. Finally, a consensus binding site for RORα1 (related orphan receptor) and RVR (ErbAα-related receptor), orphan members of the superfamily of nuclear hormone receptors, is located within intron 1 just 5′ to the TSE(32). Experiments show that RORα1 stimulates N-myc expression, and this effect is antagonized by RVR. We do not know whether one or more of these regions were responsible for the negative regulation associated with 3′ exon 1/intron 1 that we observed in vivo.

All four founder lines containing pSI/SpI N-myc lacZ transgenes failed to show transgene expression, indicating that the exon 1/intron 1 region functions in a negative regulatory manner. It is possible that lack of transgene expression in some animals could be due to integration site effect(21). We examined expression in mice generated from four separate founder lines to verify a negative regulatory effect of exon 1/intron 1 on transgene expression. Two of these founder lines showed a low copy number of integrated minigenes (the copy number of the third founder was not examined). Although we cannot completely exclude the possibility that examination of additional pSI/SpI N-myc lacZ founders with higher transgene copy numbers may have shown expression, it is actually common for expression per copy to decrease with higher copy number(21). Because nuclear run-on analysis could not distinguish bona fide silencing from down-regulation due to an integration site effect, further studies looking directly at the influence of the exon 1 attenuation site and the intron TSE on transgene expression will be undertaken to determine the mechanism of negative regulation.

The focal nature of pSI/B N-myc lacZ transgene expression in the CNS was unexpected. Although it is possible that promoter elements that guide expression in precursor cells of diverse organs lie elsewhere, outside the immediate 2.9-kb N-myc promoter fragment analyzed here, our results are also consistent with a model in which specific promoter elements guide expression in each organ type. The expression of the pSI/B N-myc transgene showed the remarkable characteristic of being present almost exclusively in structures of the olfactory system. This was particularly evident at 3-5 wk postnatally, when prominent expression in mitral cells and other cells of the main olfactory bulb as well as other basal telencephalic structures directly receiving olfactory efferents was observed. The occasional staining of the amygdaloid cortex and lateral hypothalamic nucleus at this time is also consistent with pSI/B N-myc lacZ directing transient expression of a gene product to regions of the brain involved primarily in olfactory sensory acquisition and processing. Consequently, relative to its native counterpart, this promoter must lack key cis elements that respond to many spatially restricted factors in the CNS that regulate expression of the endogenous gene, but contains a regulatory domain that is targeted by a transcription regulatory factor limited predominantly to the olfactory system. The 1.8-kb SI/B N-myc promoter responds only partially to the temporal cues that regulate expression of N-myc, again indicating absence of key regulatory components within this region. These results are consistent with previous suggestions that factors in addition to those that regulate cell division must be associated with regulating N-myc expression in the brain(4). In any case, this highly restricted expression of pSI/B N-myc lacZ may offer the opportunity to further dissect these suspected regulatory mechanisms through the novel ability to target genes of interest for expression specifically to cells of the olfactory sensory system.

In summary, our data suggest that the N-myc gene is regulated by a complex system dependent on spatial, temporal, and physiologic cues. Although our studies show a more complex mode of N-myc regulation in vivo compared with our previous studies using established cell lines, the 3′ exon 1/5′ intron region does seem to play a negative regulatory role in both systems.