Fos: An Immediate-Early Transcription Factor in Neurons

Tom Curran* and James 1. Morgan

Roche Institute of Molecular Biology, Roche Research Center, Nutley, New Jersey 071 10

SUMMARY

In the past several years a great deal of evidence has ac- cumulated linking neurons1 activation events to the regu- lation of gene expression. We have pursued an analysis of c-fos regulation in the nervous system to elucidate the molecular mechanisms involved in stimulus-transcrip- tion coupling. The c-fos gene can be viewed as a n arche- type of the set of cellular immediate-early genes encoding transcription factors. These genes are believed to func- tion in coupling short-term signals elicited by extracellu- lar to long-term changes in cellular phenotype by orches- trating alterations in target gene expression. Several ani- mal seizure models have been used to demonstrate the activation of gene expression in specific populations of neurons. Using a transgenic mouse approach, based on a

foslacz fusion gene, we now demonstrate an association between c-fos expression and cell death in the nervous system. A delayed and protracted induction was observed following surgical lesion and in response to neurotoxin exposure. This system allows us to determine, for the first time, the DNA regulatory sequences that are re- sponsible for the induction of gene expression in neurons in vivo. Furthermore, fosiucz transgenic mice provide a unique resource for identifying cell populations that re- spond to specific stimuli or that are susceptible to partic- ular toxins. Keywords: oncogenes, transcription factors, neurotoxic- ity, seizures.

0 1995 John Wiley & Sons, Inc.

ONCOGENES AND THE BRAIN

The analysis of c-fos expression in neurons was in- itiated following a chance encounter between an English neurobiologist and a Scottish molecular oncologist. Ethnic predjudice notwithstanding, the resulting collaboration led to a series of studies that neither would have pursued independently. Here we will review some ofthe results of this joint effort, up to and including recent experiments concerning the association of c-fos expression with neuronal cell death in a transgenic mouse system.

What Are Oncogenes? Retroviral oncogenes were first defined as the ge- netic information responsible for the induction of

Received July 1, 1994; accepted October 27, 1994 Journal ofNeurobiology, Vol. 26, No. 3, pp. 403-412 (1995) 0 1995 John Wiley & Sons, Inc. CCC 0022-3034/95/030403- 10

* To whom correspondence should be addressed.

cellular transformation by RNA tumor viruses. Subsequent studies revealed that they were derived from normal cellular genes, termed proto-onco- genes, by recombination with viruses (Bishop, 1985). The oncogenic potential of these cellular genes is realized either by rnutagenesis or by dereg- ulation of expression. Thus, oncogenesis can result from the aberrant or inappropriate expression of normal cellular genes. The definition of oncogenes has now been expanded to include many classes of genes that are mutated or otherwise activated dur- ing tumorigenesis. A great deal of effort has been invested over the past few years to identify and characterize retroviral oncogenes and their cellular counterparts, proto-oncogenes. The main driving force behind this research was the hope that by studying the molecular basis of virus-induced tu- mors we would learn something about human can- cer. Indeed, this has proven to be the case, as sev- eral viral oncogenes, for example rus, have been found to be mutated in human tumors (for review

403

404 Curran and Morgan

see Brugge et al., 1991). However, a second, but equally important, rationale argued that proto-on- cogenes were likely to represent cellular genes that played critical roles in normal growth regulation and would, therefore, be rewarding to study in their own right. This led to a detailed analysis of the properties of the proteins encoded by proto-onco- genes. Such investigations revealed that proto-on- cogenes are involved in several critical aspects of the regulation of cell growth and differentiation. They were found to encode proteins that function as extracellular growth factors, cell surface recep- tors, G-proteins, protein kinases, hormone recep- tors, and transcription factors (for review see Reddy et al., 1988). Although the exact molecular function of each of these proteins can be varied, they share a common feature in that they partici- pate in signal transduction processes, that is, in the transmission of information between and within cells. Thus, oncogenesis can be viewed as a break- down of the molecular mechanisms involved in in- tracellular and intercellular communication. The cancer cell either fails to respond or responds inap- propriately to its environment.

The interest in the proteins encoded by proto- oncogenes and related genes extends well beyond the field of oncology. Proto-oncogenes can be viewed simply as components of signal transduc- tion pathways that operate in several cellular contexts. Indeed, they have been shown to function in critical aspects of differentiation and develop- ment in several species (Reddy et al., 1988). In the past few years it has become apparent that they may also have a function in the adult nervous sys- tem. One of the most obvious recent examples was the finding that the trk proto-oncogene encodes a protein tyrosine kinase that normally functions as the receptor for nerve growth factor (NGF) (Chao, 1992). The connection between the fields of molec- ular oncology and neurobiology is not limited to the importation of reagents such as DNA probes and antibodies. Several of the concepts developed during the study of cellular transformation and the regulation of cell proliferation by oncogenes have also been adapted to the neurosciences. Although the nervous system has several unique features, it also uses some of the same regulatory molecules and signaling mechanisms that operate in other cell types for the transmission and storage of informa- tion. Therefore, in discussing the possible role of the,fos proto-oncogene in the nervous system, it is useful to review its molecular properties and its proposed role in growth regulation.

Origin of Fos

The fos proto-oncogene was first described as the gene responsible for induction of bone tumors by the Finkel-Biskis-Jinkins murine sarcoma virus (Curran and Teich, 1982). It encodes a DNA-bind- ing protein (Fos) that functions as a component of the mammalian transcription factor activator pro- tein-1 (AP-1) (Curran and Franza, 1988). Histori- cally, AP-1 was thought to be involved in the regu- lation of gene expression in response to the treat- ment ofcells with phorbol esters (Angel et al., 1987; Lee et al., 1987). Initially, it was suggested to con- sist of the product of the jun proto-oncogene (Boh- mann et al., 1987), which is the cellular counter- part of the oncogene carried by avian sarcoma vi- rusl7 (Maki et al., 1987). However, it is now clear that AP- 1 corresponds to several proteins related to Fos and Jun that form homodimeric and hetero- dimeric complexes through a leucine zipper struc- ture (Fig. 1) (Curran and Franza, 1988). These pro- teins regulate the expression of many genes con- taining AP-1 DNA sequence elements in a wide variety of cell types (see later). Oncogenesis in- duced by Fos and Jun is the consequence of the presence of either transcription factor in an inap- propriate context, that is, in the wrong cell at the wrong time.

Cellular Irnmediate-Early Response

The concept of the cellular immediate-early re- sponse arose from observations made in the field of growth regulation. It was noted that growing cells could be rendered quiescent by deprivation of growth factors. Entry into the cell cycle could be triggered by resupplying a cocktail of growth fac- tors that conveyed cells through a series of defined steps culminating in mitosis. A critical early phase, termed “competence,” was defined in which cells treated with platelet-derived growth factor (PDGF) were competent to proceed through the cycle if they were supplied with additional factors (Stiles et al., 1979). The use of protein synthesis inhibitors, such as cycloheximide, revealed that, during acqui- sition of competence, there was a critical early pe- riod in which PDGF was required to stimulate ex- pression of a set of genes, termed “competence genes” for progression through the cell cycle to oc- cur (Cochran et al., 1983). Thus, this set of genes, later named immediate-early genes (see later), were induced rapidly by extracellular stimuli, even in the presence of protein synthesis inhibitors, and en-

Fos Regitlution in Neurons 405

Figure 1 Dimerization and DNA binding by Fos and Jun. The Fos and Jun proteins are members of gene families that form an array of dimeric complexes. Dimerization occurs through a leucine zipper interaction and serves to juxtapose basic regions in each protein that form a bimolecular DNA binding surface. The protein dimers can have distinct properties in terms of DNA binding and transcriptional activity. Fos-Jun complexes can be phosphorylated by several protein kinases. Their DNA binding activity can also be modulated by an unusual reduction/oxidation (redox) modification.

coded proteins that were required, in combination with other signals, for the cellular growth response. This is the general concept that was borrowed from the field of growth regulation and applied to the study of signal transduction in neurons (Fig. 2).

Two major strategies were pursued to identify and characterize genes regulated by growth factors. Candidate competence genes were isolated by differential cDNA cloning of genes whose expres- sion was increased following treatment of cells with growth factors (Cochran et al., 1983). One of these revealed a partial sequence similarity with c-fos (Cochran et al., 1984). Unfortunately, this cDNA was subsequently lost. Its nucleotide sequence is similar but not identical to the fos-related genes that were cloned later. Alternatively, potential competence genes, such as the c-myc and c-fos proto-oncogenes, were examined for their ability to be regulated by growth factors (Greenberg and Ziff, 1984; Muller et al., 1984). These efforts were fol- lowed by several studies that identified a large class of rapidly induced genes in many different circum- stances, such as following treatment of cells with mitogens, other growth factors, or phorbol esters in

the presence and absence of protein synthesis in- hibitors (Lau and Nathans, 1985; Lim et al., 1987; Almendral et al., 1988; Sukhatme et al., 1988).

The term “cellular immediate-early gene” was coined when it became clear that these inducible genes were expressed in many circumstances that were not related to competence or to the cell cycle (Curran and Morgan, 1987; Lau and Nathans, 1987). The name was borrowed from the field of virology. Viral genes have been defined as early or late, depending on whether they are expressed be- fore or after replication of the viral genome. How- ever, a distinct class of viral genes was identified that is expressed extremely rapidly after infection of a cell, even in the presence of protein synthesis inhibitors. These genes were referred to as “viral immediate-early genes” to reflect these properties. Several viral immediate-early genes encode tran- scription factors that serve to activate expression of other viral genes that are required at later stages of the life cycle. Cellular immediate-early genes share many properties with their viral counterparts and so this useful term has caught on.

It is now clear that the cellular immediate-early

406 Curran and Morgan

Figure 2 The cellular immediate-early response. Induction of cellular immediate-early genes can be elicited by a great variety of extracellular stimuli. Cell surface events trigger second messenger signals that lead to a transient elevation of immediate-early gene transcription. The cellular immediate-early transcription factors Fos and Jun are translocated to the nucleus where they dimerize and bind to DNA sequences containing AP- 1 and CAMP responsive ele- ment sites to regulate target gene expression.

response is complex. Many types of genes, not just transcription factors, can be immediate-early genes. Different subsets of immediate-early genes may be induced in each cell type and in response to particular stimuli. A rather general view is emerg- ing in which immediate-early genes can be re- garded as part of a common cellular mechanism that modifies gene expression in response to alter- ations in the extracellular environment. Immedi- ate-early transcription factors function in coupling short-term second messenger-mediated events to long-term alterations in cellular phenotype in a cell- and stimulus-specific manner. Thus, their targets must be defined by the cell type in which they are expressed.

Immediate-Early Genes in Neurons

The general features of immediate-early gene func- tion and regulation just described also apply to neurons. The first indication that immediate-early genes play a role in neuronal cells was a series of studies demonstrating that NGF and several other

agents could induce c-fos expression in PC12 cells (Curran and Morgan, 1985; Greenberg et al., 1985; Kruijer et al., 1985). Two major independent sig- naling pathways were defined that mediate activa- tion of clfos expression in PC i 2 cells. One involves stimulation of the NGF or fibroblast growth factor receptors and presumably is mediated by a tyrosine kinase phosphorylation cascade. The other re- quires extracellular calcium and is triggered by de- polarization (Bartel et al., 1989; Morgan and Cur- ran, 1986). Subsequent studies have revealed that the signaling pathways linking cell surface stimuli to activation of c-fos expression are complex (Sheng and Greenberg, 1990; Morgan and Curran, 1991). Indeed, it is likely that multiple signaling pathways operate in concert to induce c-fos expres- sion.

To extend the in vitro analysis of c-fos regulation to neurons, we adopted a pharmacological model of seizure induction in mice (Morgan et al., 1987). The rationale behind this analysis was that, if c-fos expression could be induced in neurons, the largely synchronized and extensive activation of neurons

Fos Regulation in Neurons 407

foslacZ

kb

Figure 3 Schematic representation of the fos-lucZ transgene. The lacZ gene encoding P-ga- lactosidase was fused in frame into the C-terminus region ofc-fos in the fourth exon. The fusion gene contains all of the known transcriptional control elements required for the regulation of c$os expression. The positions of the sis-conditioned medium-responsive element (SCM-RE), serum-responsive element (SRE), AP- 1 site, and calcium/cAMP responsive element are indi- cated. The arrow represents the site of transcription initiation. Note that the scales for upstream and downstream DNA sequences are different. The sequences associated with mRNA turnover (A), and polyadenylation were also retained

that occurs during seizure should elicit a robust re- sponse. Indeed, this turned out to be the case; fur- thermore, the seizure paradigm has proved to be useful in dissecting the events associated with im- mediate-early gene regulation in vivo, for review see (Morgan and Curran, 1991). A great many studies have now been published on the regulation of immediate-early gene expression in neurons. Nucleic acid and antibody probes to these genes have proved to be of utility in activity-mapping studies, for example, in the analysis of gluta- matergic (Cole et al., 1989; Sonnenberg et al., 1989a; Popovici et al., 1990), dopaminergic (Gray- biel et al., 1990; Young et al., 1991), nociceptive (Hunt et al., 1987; Bullitt, 1989), and photic (Rea, 1989; Aronin et al., 1990; Kornhauser, et al., 1990; Rusak, et al., 1990) stimuli. However, it should be kept in mind that immediate-early genes are not universal markers of activity. In addition, the re- sponse is not always identical. Different subsets of immediate-early genes can be induced indepen- dently (Bartel et al., 1989; Sonnenberg et al., 1989b; Smeyne et al., 1992a). Thus, a complex pic- ture is emerging in which the immediate-early re- sponse must be viewed in its cellular context. To obtain insights into the molecular events responsi- ble for the exquisite cell type- and stimulus-depen- dent regulation of gene expression in the nervous system, we have used a transgenic mouse approach to the analysis of c-fos induction in the adult ner- vous system.

TRANSGENIC APPROACH TO C - ~ O S REGULATION

A fos-lacZ fusion gene was constructed that re- tained all of the known regulatory elements associ- ated with transcription of the c-fos gene (Fig. 3).

This fusion gene allows c-fos expression to be mon- itored by a simple and quantitative o-galactosidase enzyme assay. Analysis of the expression of this construct in transfected cells in culture revealed that P-galactosidase activity could be used to mon- itor c-fos induction with single-cell sensitivity (Schilling et al., 199 1). This provides a useful rapid assay for agents that induce c-fos expression in cul- tured cells.

The fos-lucZ fusion gene was introduced into the mouse germ line by microinjection to generate a transgenic mouse strain. In this strain of mice 0- galactosidase activity can be used to determine the sites of basal and stimulated c-fos expression and to follow c-fos expression during development (Smeyne et al., 1992a, 1992b). Expression of fos- lucZ was readily detected in sections using the sub- strate, X-Gal. This allowed determination of the cell types expressing c-fos continuously and after induction by several stimuli. In addition, these ex- periments defined the DNA sequences required for c-fos regulation in vivo. In unstimulated mice Fos- lacZ was observed in several regions of adult brain at relatively low levels, with the exception of the raphe nuclei (Smeyne et al., 1992a). However, some areas in which Fos-like immunoreactivity (FLI) had been reported were not positive for Fos- lacZ. The use of Fos-specific antibodies and immu- noblotting procedures has revealed that, in many cases, FLI does not represent authentic Fos, as the majority of antibodies used for these studies also recognize Fos-related proteins (Sonnenberg et al., 1989b). Thus, Fos-lacZ provides a more accurate representation of c-fos expression than FLI.

Induction of Fos-lac2 by Seizures Previously, we documented the patterns of FLI staining induced following induction of seizures in

408 Curvan and Morgan

mice (Morgan et al., 1987). Similar analysis was performed using fus-ZacZ mice using the convul- sants pentylenetetrazole (PTZ) and kainic acid (KA). The patterns of expression obtained were similar but not identical (Smeyne et al., 1992a). Predominant sites of Fos-lacZ expression were as- sociated with the limbic system. Both agents in- duced Fos-lac2 within 30 min, expression reached peak levels in about 2 to 3 h, declined within 5 h and returned to basal values by 24 h. KA elicited a pattern of Fos-lacZ that was similar to the reported FLI distribution induced in the rat (Popovici et al., 1990). PTZ, which induces clonic-tonic seizures, generated a more widespread distribution of Fos- lacZ than KA, particularly in the cerebral cortices and diencephalon. However, there was a relatively poor induction of Fos-lacZ in the hippocampal dentate gyrus (Fig. 4). In contrast, previously we reported extensive FLI induction in the dentate gy- rus after PTZ treatment (Morgan et al., 1987). We confirmed that Fos-lacZ represented the distribu- tion of authentic Fos expression using a Fos-spe- cific antibody. Therefore, the majority of FLJ in the dentate gyrus after a PTZ seizure is contributed by Fos-related antigens. Fos-lacZ can be induced in the dentate gyrus, as is evident from the pattern ob- tained after administration of K A (Fig. 4). Further- more, at least in a subpopulation of dentate gyrus granule neurons Fos-lac2 was induced by PTZ. These results indicate that the signaling pathways controlling immediate-early genes are complex and that they may be regulated independently in similar cell types.

Association between Fos Expression and Cell Death

The most surprising finding obtained to date from thefbs-facZ mice was the observation that Fos was expressed in cells undergoing terminal differentia- tion and programmed cell death. In our initial study, we reported that Fos-lacZ was expressed in cell populations undergoing terminal differentia- tion in skin, hair follicle, and bone (Smeyne et al., 1992a). In skin, continuous expression was ob- served in keratinized epithelium, particularly in the flattened cells of the stratum corneum and to a lesser extent the cells of the strata granulosum and spinosum. We also observed expression of Fos- lacZ in skin at around embryonic day 17 (E17) in the degenerating peridermal layer of developing skin. At this time the peridermal layer is shed and keratinization is initiated, both of which are associ-

Figure 4 Induction of Fos-lacZ in the dentate gyrus by seizures. Sections were taken from similar regions of the dentate gyrus of control transgenic mice (A), 2 hours af- ter a seizure induced by PTZ (B), and 2 hours after a seizure induced by KA (C). Positive cells are distin- guished by the dark nuclear stain. (Adapted from Smeyne et al., 1992a).

ated with cell death. In agreement with others (Dony and Gruss, 1987), the interdigital web was also found to express Fos-lacZ at the time of initia- tion of programmed cell death, which in the mouse occurs at around E 18.

In addition to the cells of the epithelium, Fos- lacZ was also continuously expressed in two dis- tinct populations of cells in adult bone; presump- tive proliferating osteoblasts and hypertrophic chondrocytes (Smeyne et al., 1992a). During devel- opment of the long bones, the initial bone model is

Fos Regulation in Neurons 409

formed by a cartillaginous chondrocytic structure. At later stages of development, blood vessel inva- sion triggers chondrocyte hypertrophy in the core of the bone, resulting in cell death and migration of osteoblasts. Analysis of developing bone demon- strated that Fos-lacZ was expressed in these hyper- trophic chondrocyte populations prior to death.

Neuronal death occurs naturally in the spinal cord and dorsal root ganglia (DRG) during devel- opment through a process that involves target-de- rived trophic factors (Levi-Montalchini and Angel- etti, 1966). Indeed, the observation was made by Levi-Montalchini that injection of NGF antibodies into mice resulted in the destruction of the sympa- thetic ganglia (Levi-Montalchini and Booker, 1960). The initial observations of Fos-lacZ expres- sion in dying cells during development prompted us to examine the association of Fos-lacZ expres- sion and cell death in the nervous system in this now classic model (Smeyne et al., 1993). A few Fos- lacZ-positive cells were detected in the sympathetic ganglia in transgenic mice. Since cell death occurs over several days in the mouse, and the elimination of dead cells requires only a few hours, it is difficult to establish whether the Fos-lacZ-positive cells rep- resent degenerating neurons. To resolve this issue, we induced a synchronous wave of cell death in the spinal cord and DRG by a lesion of the sciatic nerve. On the day of birth, transgenic animals were given unilateral scissor-cut lesions of the sciatic nerve and the mouse was examined at 8 and 24 hours postsurgery. After 8 hours, extensive Fos- lacZ expression was observed throughout the ipsi- lateral limb and trunk, including skeletal muscle, bone, joint capsule, and periosteum [Fig. 5(A,B)]. Since little Fos-lacZ expression was observed con- tralateral to the lesion [Fig. 5 (A)], it is most likely that this expression is neurogenic in origin. In the central nervous system (CNS), 8 hours following sciatic nerve lesion, Fos-lacZ expression was de- tected in laminae 1/2 and 4/5 of the dorsal horn of the spinal cord [Fig. 5(C)]. This induction was specific both with regard to laterality and level in the spinal cord. Fos-lacZ expression was only ob- served in the ipsilateral spinal cord between L3 and S2, which is consistent with the denervation of the sensory zones supplied by the sciatic nerve. Lami- nae 1/2 at this level contain neurons that receive afferent innervation primarily from cutaneous no- ciceptive zones, whereas neurons in laminae 4/5 respond to a wide range of sensory stimuli and form the proper sensory nucleus. Although affer-

ents to laminae 4/5 arise in the DRG, no Fos-lacZ staining was noted in this structure at this time.

Later, at 24 hours postsurgery, the extensive Fos-lacZ expression in the ipsilateral skeletal mus- cle, bone cells, and dorsal horn of the spinal cord largely disappeared. Thus, the initial pattern of ex- pression of Fos-lacZ was transient in nature. In the spinal cord ipsilateral to the lesion, Fos-lacZ ex- pression was detected in approximately half of the somatic alpha motor neurons of lamina 9 in a level- specific manner [Fig. 5(D,E)]. On the side contra- lateral to the lesion, only a small number (0 to 21 section) of motor neurons expressed Fos-lacZ. At 24 hours, Fos-lacZ was also observed in the ipsilat- era1 DRG from L2-L4 [Fig. 5(F,G)]. Cells of the DRG are known to die following sciatic nerve le- sion because of the absence of a target-derived tro- phic influence. Indeed, NGF has been shown to rescue DRG neurons following sciatic nerve lesion (Yip and Johnson, 1984).

The correlation between Fos expression and cell death was observed in several circumstances, in- cluding in response to KA 1 toxicity and in degen- erating cerebellar granule neurons in the weaver strain of mice (Smeyne et al., 1992a). Fos-1acZ was also expressed in cell populations undergoing nat- urally occurring cell death during development and following induction of apoptotic cell death in culture using the drug etoposide. This resulted in a prolonged increase in Fos-lacZ expression prior to the appearance of the apoptosis phenotype or nucleosome laddering (Smeyne et al., 1992a).

Implications

The development of a multicellular organism in- volves a delicate balance among the processes of proliferation, differentiation, and death. Naturally occumng cell death serves to facilitate tissue remod- eling, to eliminate supernumerary cell populations and to provide structural elements, such as hair and skin. In the nervous system, selective cell death con- tributes to the formation and organization of the spinal cord, sympathetic ganglia, and corpus callo- sum. However, cell death also occurs in several neu- ropathological conditions, such as amyotrophic lat- eral sclerosis and Alzheimer’s disease. The func- tional role of Fos in cell death is not yet clear. It is possible that Fos is part of a programmed response pathway that leads to cell death. At least in some circumstances, protein synthesis inhibitors can in- hibit cell death. Alternatively, activation of c-fos ex- pression may simply be a side affect of the onset of

410 Currun and Morgan

Figure 5 Expression of Fos-IacZ following sciatic nerve section. (A) Low-power photomicro- graph through lesion site. Fos-lacZ was only induced ipsilateral to the lesion (left side of pho- tograph). (B) Higher-power photomicrograph of the lesion site. Fos-lacZ was expressed in skel- etal muscle (sm), joint capsule (jc), and periosteum (peri). (C) Eight hours following surgery, expression of the Fos-lacZ transgene was seen in laminae 1/2 and 4/5 of the dorsal horn of the spinal cord. (D) Twenty-four hours after the sciatic nerve lesion, Fos-lacZ was seen in the degenerating motor neurons of the lumbar spinal cord, ipsilateral to the sciatic nerve lesion. Contralaterally, no increase was seen in the frequency of cells expressing Fos-lacZ, which was usually 0- I cell/section (arrowhead). A higher-power photomicrograph of the box is shown in (F). Fos-lacZ was seen in the nucleus of approximately 50% of the motor neurons. Rostrally, expression of the transgene was seen in the ipsilateral, but not the contralateral, dorsal root ganglion (E). A higher-power photomicrograph of the DRG neurons is shown in (G). (From Smeyne et al., Nutuw: 166-169, 0 1993 Macmillan Magazines Ltd., reprinted with permis- sion.)

death. In either circumstance, fi,s-facZ mice carry an endogenous marker that may be of utility as a harbinger of death. Recently, we extended this transgenic model by inserting the mouse . fos-lacZ construct into the gennline of rats (Kasof et al., sub- mitted). These transgenic rats provive a particularly useful model to investigate the association of imme- diate-early gene expression with excitotoxicity. It appears that a unique subset of immediate early- genes is induced prior to the onset of death in sus- ceptible neuronal populations in these animals. These results demonstrate the the molecular mech-

anisms involved in linking gene induction with cell death are conserved between mice and rats.

MOLECULAR BASIS OF THE SELECTIVE ACTIVATION OF GENE EXPRESSION

The fos-lacZ construct defines the genetic elements required for the regulation of cfus expression in neurons in viva This provides a unique opportunity to investigate the molecular basis of gene regulation

Fos Regulation in Neurons 41 I

in response to neuronal activation. The transcnp- tion control regions of eukaryotic genes contain multiple sequence elements that have been pro- posed to function independently to regulate tran- scriptional responses to specific signals. Therefore, we developed transgenic mice carrying fos-ZacZ fu- sion genes with clustered point mutations in each of several distinct regulatory sequences; the sis-induc- ible element, the serum response element, the fos AP- 1 site and the calcium/cyclic adenosine mono- phosphate (CAMP) response element (Robertson et al., submitted). Analysis of Fos-lacZ expression in the CNS demonstrated that all of the regulatory ele- ments tested were required in concert for seizure- induced c-fos expression in the brain. Furthermore, in cultured cells all of the elements examined were required for full responses to CAMP, polypeptide factors, and depolarization. This implies that the regulation of c-fos expression requires the concerted action of multiple control elements that direct the assembly of an interdependent transcription com- plex on the c-fos promoter. This contradicts the con- cept of a simple transcription factor response ele- ment and it suggests that tissue- and stimulus-spe- cific gene regulation occurs through the modulation of interdependent transcription complexes.

This article is based on the Rita Levi-Montalchini Fidia Award Lecture, portions of which were adapted from Smeyne et al. (l992a, 1993).

REFERENCES

ALMENDRAL, J. M., SOMMER, D., MACDONALD- BRAVO, H., BURCKHARDT, J., PERERA, J., and BRAVO, R. (1988). Complexity of the early genetic re- sponse to growth factors in mouse fibroblasts. Mol. Cell. Biol. 8:2 140-2 148.

ANGEL, P., IMAGAWA, M., CHIU, R., STEIN, B., IMBRA, R. J., RAHMSDORF, H. J., JONAT, C., HERRLICH, P., and KARIN, M. ( I 987). Phorbol ester-inducible genes contain a common cis element recognized by a TPA modulated trans-acting factor. Cell 49:729-739.

ARONIN, N., SAGAR, S., SHARP, F., and SCHWARTZ, W. ( 1990). Light regulates expression of a Fos-related pro- tein in the rat SCN. Proc. Nutl. Acad. Sci. USA 87:

BARTEL, D. P., SHENG, M., LAU, L. F., and GREENBERG, M. (1989). Growth factors and membrane depolariza- tion activate distinct programs of early response gene expression-dissociation of fos and j u n induction. Genes Dev. 3:304-3 13.

5959-5962.

BISHOP, J. M. (1985). Viral oncogenes. Cell42:23-38.

BOHMANN, D., Bos, T. J., ADMON, A., NISHIMURA, T., VOGT, P. K., and TJIAN, R. (1987). Human proto-on- cogene c-jun encodes a DNA binding protein with structural and functional properties of transcription factor AP-1. Science238:1386-1392.

BRUGGE, J., CURRAN, T., HARLOW, E., and MCCOR- MICK, F., Eds. (199 1). Origins of Human Cancer: A Comprehensive Review. Cold Spring Harbor Labora- tory Press, Cold Spring Harbor, NY.

BULLITT, E. (1 989). Induction of c-fos-like protein within the lumbar spinal cord and thalamus of the rat following peripheral stimulation. Brain Res. 493:39 I - 397.

CHAO, V. M. ( I 992). Neurotrophin receptors: a window into neuronal differentiation. Neuron 9583-593.

COCHRAN, B. H., REFFEL, A. C., and STILES, C. D. ( I 983). Molecular cloning of gene sequences regulated by platelet-derived growth factor. Cell 33:939-947.

COCHRAN, B. H., ZULLO, J., VERMA, 1. M., and STILES, C. D. (1 984). Expression of c-fos oncogene and of a fos-related gene is stimulated by platelet-derived growth factor. Science 226:1080-1083.

COLE, A. J., SAFFEN, D. W., BARABAN, J. M., and WOR- LEY, P. F. (1989). Rapid increase of an immediate- early gene messenger RNA in hippocampal neurons by synaptic NMDA receptor activation. Nature 340: 474-476.

CURRAN, T. and FRANZA, B. R. Jr. ( 1 988). Fos and Jun: the AP-I connection. Cell55395-397.

CURRAN, T. and MORGAN, J. 1. (1985). Superinduction by fos by nerve growth factor in the presence of peripherally active benzodiazepines. Science 229: 1265- 1 268.

CURRAN, T. and MORGAN, J. I. ( 1 987). Memories offos. Bioessays 7:25 5-2 5 8.

CURRAN, T. and TEICH, N. M. (1982). Candidate prod- uct of the FBJ murine osteosarcoma virus oncogene: characterization of a 5 5,000-Dalton phosphoprotein. J. Virol. 42:114-122.

DONY, C. and GRUSS, P. ( I 987). Proto-oncogene c-fos expression in growth regions of fetal bone and meso- dermal web tissue. Nature 328:7 1 1-7 14.

GRAYBIEL, A., MORATELLO, R., and ROBERTSON, H. A. ( 1 990). Amphetamine and cocaine induce drug-spe- cific activation of the c-fos gene in striasome-matrix compartments and limbic subdivisions of the stria- turn. Proc. Natl. Acud. Sci. USA 87:6912-6916.

GREENBERG, M. E. and ZIFF, E. B. (1984). Stimulation of 3T3 cells induces transcription of the c-fos proto- oncogene. Nature 311:433-438.

GREENBERG, M. E., GREENE, L. A., and ZIFF, E. G. ( I 985). Nerve growth factor and epidermal growth fac- tor induce rapid transient changes in proto-oncogene transcription in PC 12 cells. J. Biol. Chem. 260: 14 101 - 141 10.

HUNT, S. P., PINI, A., and EVAN, G. (1987). Induction of c-fos-like protein in spinal cord neurons following sensory stimulation. Nature 328:632-634.

41 2 Curran and Morgan

KORNHAUSER, J., NELSON, D. E., MAYO, K. E., and TA- KAHASHI, J. S. ( I 990). Photic and circadian regulation of clfos gene expression in the hamster suprachias- matic nucleus. Neuron 5: 127- 134.

KRUIJER, W., SCHUBERT, D., and VERMA, I. M. (1985). Induction of the proto-oncogene fos by nerve growth factor. Proc. Natl. Acad. Sci. USA 82:7330-7334.

LAU, L. F. and NATHANS, D. (1985). Identification of a set of genes expressed during the Cio-GI transition of cultured mouse cells. EMBO J. 4:3 145-3 I5 1.

LAU, L. F. and NATHANS, D. (1987). Expression of a set of growth-related immediate-early genes in BALB/c 3T3 cells: coordinate regulation with clfos or c-myc. Proc. Natl. Acad. Sci. USA 84:1182- 1 186.

LEE, W., MITCHELL, P., and TJIAN, R. (1987). Purified transcription factor AP- 1 interacts with TPA-induc- ible elements. Ce1!49:741-752.

LEVI-MONTALCHINI, R. and ANGELETTI, P. U. (1 966). Immunosympathectomy. Pharmacol. Rev. 18:6 19-628.

LEVI-MONTALCHINI, R. and BOOKER, B. (1960). De- struction of the sympathetic ganglia in mammals by an antiserum to the nerve-growth promoting factor. Proc. Natl. Acad. Sci. USA 46~384-391.

LIM, R. W., VARNUM, B. C., and HERSCHMAN HR. ( 1987). Cloning of tetradecanoyl phorbol ester-in- duced “primary response” sequences and their expres- sion in density-arrested Swiss 3T3 cells and a TPA non-proliferative variant. Oncogene 1 :263-270.

VOGT, P. K. (1987). Avian sarcoma virus 17 carries a new oncogene, jun. Proc. Natl. Acad. Sci. USA 84: 2848-2852.

MORGAN, J. I. and CURRAN, T. (1986). Role of ion flux in the control of c-fos expression. Nature 322:552-555.

MORGAN, J. 1. and CURRAN, T. (199 I ) . Stimulus-tran- scription coupling in the nervous system: involvement of the inducible proto-oncogenes fos and jun. Annu. Rev. Neurosci. 14:42 1-45 1.

MORGAN, J. I., COHEN, D. R., HEMPSTEAD, J. L., and CURRAN, T. ( I 987). Mapping patterns of c-jbs expres- sion in the central nervous system after seizure. Sci- ence 237: 1 92- 197.

MULLER, R., BRAVO, R., BURCHKHARDT, J., and CUR- RAN, T. (1984). Induction of c-fos gene and protein by growth factors precedes activation of c-myc. Nulure

Popovrci, T., RESPRESA, A,, CREPEL, V., BARBIN, G., BEAUOIN, M., and BEN-ARI, Y. ( 1 990). Effects of kai- nic acid-induced seizures and ischemia on c-fos-like proteins in rat brain. Brain Rex 536:183-194.

REA, M. A. (1989). Light increases Fos-related protein immunoreactivity in the rat suprachiasmatic nuclei. Brain Res. Bull. 23557-58 I .

REDDY, E. P., SKALKA, A. M., and CURRAN, T., eds. ( 1 988). The Oncogene Handbook, Elsevier, Amster- dam.

RUSAK, B., ROBERTSON, H. A,, WISDEN, W., and HUNT,

MAKI, Y., BOS, T. J., DAVIS, c., STARBUCK, M., and

312:7 16-720.

S. P. (1990). Light pulses that shift rhythms induce gene expression in the suprachiasmic nucleus. Science 24831237-1240.

SCHILLING, K., LUK, D., MORGAN, J. I., and CURRAN, T. ( I99 1 ). Regulation of a fos-lacZ fusion gene: a par- adigm for quantitative analysis of stimulus-transcrip- tion coupling. Proc. Natl. Acad. Sci. USA 885665- 5669.

SHENG, M. and GREENBERG, M. E. (1990). The regula- tion and function of c-fos and other immediate early genes in the nervous system. Neuron 4:477-485.

SMEYNE, R. J., SCHILLING, K., ROBERTSON, L., LUK, D., OBERDICK, J., CURRAN, T., and MORGAN, J. I. (1 992a). Fos-lacZ transgenic mice: mapping sites of gene induction in the central nervous system. Neuron

SMEYNE, R. J., CURRAN, T., and MORGAN, J. I. (1 992b). Temporal and spatial expression of a fos-lacZ transgene in the developing nervous system. Mol. Brain Res. 16:158-162.

SMEYNE, R. J., VENDRELL, M., HAYWARD, M., BAKER, S. J., MIAO, Ci. G., SCHILLING, K., ROBERTSON, L., CURRAN, T., and MORGAN, J. I. ( 1 993). Fos: Contin- uous c-fos expression precedes programmed cell death in vivo. Nature363:166-169.

SONNENBERG, J.. L., MITCHELMORE, C., MACGREGOR, P. F., HEMPSTEAD, J., MORGAN, J. I., and CURRAN, T. ( 1 989a). Glutamate receptor agonists increase the expression ofFos, Fra and AP-1 DNA binding activity in the mammalian brain. J. Neurosci. Rex 24:72-80.

SONNENBERG, J . L., MACGREGOR-LEON, P. F., CUR- RAN, T., and MORGAN, J. I. (1989b). Dynamic alter- ations occur in the levels and composition of tran- scription factor AP- 1 complexes after seizure. Neuron

8: 13-23.

3~359-365. STILES, c. D., CAPONE, G. T., SCHER, C. D., ANTONI-

ADES, H. N., VANWYK, J. J., and PLEDGER, W. J. ( 1 979). Dual control of cell growth by somatomedins and platelet-derived growth factor. Proc. Natl. Acud. Sci. USA. 76:1279-1284.

RIS, C.-H., STAMENKOVICH, D., FERRIERA, P. C. P., COHEN, D., EDWARDS, S. A., SHOWS, T. B., CURRAN, T., LEBEAU, M. M., and ADAMSON, E. C. (1988). A zinc finger-encoding gene coregulated with c-fos dur- ing growth and differentiation and after cellular depo- larization. Cell 53:37-43.

YIP, H. K. and JOHNSON, E. M. Jr. (1984). Developing dorsal root ganglion neurons require trophic support from their central processes: evidence for a role of ret- rogradely transported nerve growth factor from the central nervous system to the periphery. Proc. Null.

YOUNG, S. T., PORRINO, L. J., and IADAROLA, M. J. ( 1 99 1). Cocaine induces striatal c-Fos-immunoreac- tive proteins via dopaminergic D I receptors. Proc. Natl. Acad. Sci. USA 88:1291-1295.

SUKHATME, v. P., CAO, x., CHANG, L. c., TSAI-MOR-

Acctd. Sci. USA. 81~6245-6249.