Comprehensive epitope analysis of monoclonal anti-proenkephalin antibodies using phage display libraries and synthetic peptides: revelation of antibody fine specificities caused by somatic mutations in the variable region genes

Gene therapy for trigeminal pain in mice

The aim of this study was to test the efficacy of a single direct injection of viral vector encoding for encephalin to induce a widespread expression of the transgene and potential analgesic effect in trigeminal behavioral pain models in mice. After direct injection of HSV-1 based vectors encoding for human preproenkephalin (SHPE) or the lacZ reporter gene (SHZ.1, control virus) into the trigeminal ganglia in mice, we performed an orofacial formalin test and assessed the cumulative nociceptive behavior at different time points after injection of the viral vectors. We observed an analgesic effect on nociceptive behavior that lasted up to 8 weeks after a single injection of SHPE into the trigeminal ganglia. Control virus injected animals showed nociceptive behavior similar to naïve mice. The analgesic effect of SHPE injection was reversed/attenuated by subcutaneous naloxone injections, a μ-opioid receptor antagonist. SHPE injected mice also showed normalization in withdrawal latencies upon thermal noxious stimulation of inflamed ears after subdermal complete Freund’s adjuvans injection indicating widespread expression of the transgene. Quantitative immunohistochemistry of trigeminal ganglia showed expression of human preproenkephalin after SHPE injection.

Direct injection of viral vectors proved to be useful for exploring the distinct pathophysiology of the trigeminal system and could also be an interesting addition to the pain therapists’ armamentarium.

Epitope mapping using phage display peptide libraries

Phage libraries displaying millions of peptides with randomized sequences are extremely useful tools for mapping antibody epitopes. In many cases, antibodies are able to select peptides with reasonable affinity for their combining sites (paratopes) from these libraries. Ideally, consensus motives can be deduced from multiple peptide sequences and matched to areas of the antigen against which the antibody was raised. That way, critical components of the antibody epitope can be defined. This chapter focuses on technical details of epitope mapping employing pre-made filamentous phage peptide display libraries. Examples are given for illustration.

Identification of cell proliferation-associated epitope on CD98 oncoprotein using phage display random peptide library

CD98 is known as a cell surface antigen expressed in proliferating normal tissues and in almost all tumor cells. Although the function of CD98 is not yet fully elucidated, it is suggested that CD98 is concerned functionally in lymphocyte activation, cell proliferation, and malignant transformation. Monoclonal antibody against human CD98 heavy chain (h.c.), termed HBJ127, shows inhibition of lymphocyte activation and tumor cell growth in vitro. These observations suggest that the epitope recognized by HBJ127 may be crucial for CD98 function. In the present study, the authors investigated the epitope recognized by HBJ127 using a phage display random heptapeptide library. Approximately 2.4 x 10(4)-fold amplification of eluted phage titer was obtained after three rounds of panning of the phage library against HBJ127. Seven different heptapeptide sequences were isolated from 30 randomly selected clones of the post-panning phage population. A homology search using ClustalW identified the peptide sequence corresponding to (442)AFS(444) of human CD98 h.c. It was also found that (443)F is a human-specific amino acid by comparing sequences of human, rat, and mouse origin. Reduced reactivity of HBJ127 was detected against the phenylalanine-substituted peptide but not detected against the alanine or serine-substituted one. It has been identified that HBJ127 reacts only with human species and the HBJ127 epitope position is predicted in 418-529 of human CD98 h.c. From these results and observations, it was estimated that (442)AFS(444) of human CD98 h.c. may be the HBJ127 epitope. Moreover, (443)F may be critical for the binding of HBJ127 against human CD98 h.c.

Human monoclonal rheumatoid synovial B lymphocyte hybridoma with a new disease-related specificity for cartilage oligomeric matrix protein

Joint-specific self-Ags are considered to play an important role in the induction of synovial T and B cell expansion in human rheumatoid arthritis (RA). However, the nature of these autoantigens is still enigmatic. In this study a somatically mutated IgG2 lambda B cell hybridoma was established from the synovial membrane of an RA patient and analyzed for its Ag specificity. A heptameric peptide of cartilage oligomeric matrix protein (COMP) could be characterized as the target structure recognized by the human synovial B cell hybridoma. The clonotypic V(H) sequences of the COMP-specific hybridoma could also be detected in synovectomy material derived from five different RA patients but in none of the investigated osteoarthritis cases (n = 5), indicating a preferential usage of V(H) genes closely related to those coding for a COMP-specific Ag receptor in RA synovial B cells. Moreover, the COMP heptamer was preferentially recognized by circulating IgG in RA (n = 22) compared with osteoarthritis patients (n = 24) or age-matched healthy controls (n = 20; both p < 0.0001). Hence, the COMP-specific serum IgG is likely to reflect local immune responses toward a cartilage- and tendon-restricted Ag that might be crucial to the induction of tissue damage in RA.

Delineation of a linear epitope by multiple peptide synthesis and phage display

Two different approaches, the phage display technique and the Spot peptide synthesis on cellulose membranes, were used to identify sequences recognized by Fab 57P, specific for tobacco mosaic virus protein (TMVP), and define the preferred chemical composition of a functional epitope. Kinetic measurements of the interaction between peptide variants and the antibody fragment were used to further refine the molecular basis of binding activity. Our results show that the functional epitope of Fab 57P requires precise physico-chemical properties at a limited number of positions, and that residues flanking these key residues can influence binding affinity. The phage display and Spot synthesis methods allowed the straightforward localization of the binding region and the identification of residues that are essential for recognition. However, these methods yielded slightly different views of accessory factors that are able to influence antibody binding. The influence on binding activity of these factors can only be assessed through quantitative affinity measurements.

Selection of peptidic mimics of digoxin from phage-displayed peptide libraries by anti-digoxin antibodies

Since the initial report of the development of methodology to generate high-affinity digitalis-specific (digoxin) antibodies, these antibodies have proven extremely useful tools to monitor digoxin levels in digitalized patients and, as Fab fragments, to reverse toxic digoxin effects in life-threatening digoxin overdoses. These antibodies (both digoxin-specific and ouabain-specific) have been used extensively by investigators for the identification and characterization of putative endogenous digitalis-like factors. In this study, we used two well-characterized mouse anti-digoxin monoclonal antibodies (mAbs), designated 26-10 and 45-20, as binding templates with which to select short bacteriophage-displayed (pIII protein inserted) peptides that are capable of binding to these mAbs and mimicking the conformational structure of digoxin. Selective enrichment from two phage-displayed random peptide libraries enabled us to isolate and identify distinct 15 and 26 amino acid residue peptide inserts that bind with high avidity and idiotypic specificity to the selecting mAbs. Among these displayed inserts a subset was identified whose mAb binding is inhibited by digoxin and whose corresponding synthetic peptides inhibit phage binding. They, therefore, appear to bind at the mAbs digoxin-binding sites. These data provide the first clear evidence that short polypeptides can serve as surrogates for the low molecular mass hapten digoxin.

Characterization of a differential immunoscreen epitope of Plasmodium falciparum using combinatorial agents

A differential serological screening of a lambdagt11 cDNA expression library has identified several clones, which react exclusively to sera samples from persons clinically immune to malaria but not to acute malaria patient sera. One such clone, IPf9, has a 315-bp cDNA insert, which was found to be conserved in different strains of the human and rodent malarial parasite Plasmodium falciparum and Plasmodium berghei, respectively. The induced expression product of IPf9 was used to generate polyclonal sera in rabbits. The IPf9 expression product was also screened with phage surface display combinatorial libraries to isolate reagents that specifically bound to the IPf9 product. The polyclonal antisera and the combinatorial reagents recognized a 50-kDa protein from P. falciparum, and a 53-kDa product from P. berghei. Immunofluorescence studies using asexual and sexual stages of P. falciparum showed the protein to be present within the parasite in each of the asexual and sexual stages. The combinatorial reagents showed a partial inhibition in the growth of P. falciparum in vitro. Mice infected with the P. berghei showed the presence of T-cells that exhibited lymphoproliferation when stimulated with the IPf9 protein. It is suggested that IPf9 protein is a conserved protein epitope, and may be relevant for a protective immune response to malaria.

Opioids intrinsically inhibit the genesis of mouse cerebellar granule neuron precursors in vitro: differential impact of mu and delta receptor activation on proliferation and neurite elongation

Although opioids are known to affect neurogenesis in vivo, it is uncertain the extent to which opioids directly or indirectly affect the proliferation, differentiation or death of neuronal precursors. To address these questions, the intrinsic role of the opioid system in neurogenesis was systematically explored in cerebellar external granular layer (EGL) neuronal precursors isolated from postnatal mice and maintained in vitro. Isolated neuronal precursors expressed proenkephalin-derived peptides, as well as specific mu and delta, but negligible kappa, opioid receptors. The developmental effects of opioids were highly selective. Morphine-induced mu receptor activation inhibited DNA synthesis, while a preferential delta2-receptor agonist ([D-Ala2]-deltorphin II) or Met-enkephalin, but not the delta1 agonist [D-Pen2, D-Pen5]-enkephalin, inhibited differentiation within the same neuronal population. If similar patterns occur in the developing cerebellum, spatiotemporal differences in endogenous mu and delta opioid ligand-receptor interactions may coordinate distinct aspects of granule neuron maturation. The data additionally suggest that perinatal exposure to opiate drugs of abuse directly interfere with cerebellar maturation by disrupting normal opioid signalling and inhibiting the proliferation of granule neuron precursors.

Identification of peptide mimotopes for the fluorescein hapten binding of monoclonal antibody B13-DE1

Using 6mer and 12mer phage peptide libraries three unique phage clones were identified which specifically bind to a monoclonal anti-FITC antibody, B13-DE1. The two 6mer and one 12mer peptide insert sequences are clearly related to each other and contain a high proportion of hydrophobic amino acids. The peptides are bound by the antibody combining site of B13-DE1 probably in a similar manner to FITC and represent therefore true peptidic mimics of the fluorescein hapten. No reactivity of the peptides could be demonstrated with another monoclonal anti-fluorescein antibody or with polyclonal anti-fluorescein antibodies. Immunization of mice with the peptides resulted in the production of antibodies cross-reacting with all peptides but not with fluorescein. The results show that phage peptide libraries can be used to isolate mimotope peptides which can mimic low molecular weight structures seen by a specific antibody and probably other recognition molecules.

Antihyperalgesic effects of infection with a preproenkephalin-encoding herpes virus

To test the utility of gene therapeutic approaches for the treatment of pain, a recombinant herpes simplex virus, type 1, has been engineered to contain the cDNA for an opioid peptide precursor, human preproenkephalin, under control of the human cytomegalovirus promoter. This virus and a similar recombinant containing the Escherichia coli lacZ gene were applied to the abraded skin of the dorsal hindpaw of mice. After infection, the presence of beta-galactosidase in neuronal cell bodies of the relevant spinal ganglia (lacZ-containing virus) and of human proenkephalin (preproenkephalin-encoding virus) in the central terminals of these neurons indicated appropriate gene delivery and expression. Baseline foot withdrawal responses to noxious radiant heat mediated by Adelta and C fibers were similar in animals infected with proenkephalin-encoding and beta-galactosidase-encoding viruses. Sensitization of the foot withdrawal response after application of capsaicin (C fibers) or dimethyl sulfoxide (Adelta fibers) observed in control animals was reduced or eliminated in animals infected with the proenkephalin-encoding virus for at least 7 weeks postinfection. Hence, preproenkephalin cDNA delivery selectively blocked hyperalgesia without disrupting baseline sensory neurotransmission. This blockade of sensitization was reversed by administration of the opioid antagonist naloxone, apparently acting in the spinal cord. The results demonstrate that the function of sensory neurons can be selectively altered by viral delivery of a transgene. Because hyperalgesic mechanisms may be important in establishing and maintaining neuropathic and other chronic pain states, this approach may be useful for treatment of chronic pain and hyperalgesia in humans.

Comparative study of the p53-mdm2 and p53-MDMX interfaces

Mdm2 and MDMX are two structurally related p53-binding proteins which show the highest level of sequence similarity in the N-terminal p53-binding domains. Apart from its ability to inhibit p53 mediated transcription, a feature it shares with mdm2, very little is known about the physiological functions of MDMX. It is clearly distinct from mdm2 since its expression appears not to be regulated by p53 and it cannot compensate for lack of mdm2 in early development. We present data on the structural similarity between the p53 binding pockets of mdm2 and MDMX using p53- and phage-selected peptides. From the results we conclude that our recently devised innovative approach to reverse the mdm2-mediated inhibition of p53's transactivation function in vivo would probably target MDMX as well. Strategies for selectively targeting mdm2 and MDMX are suggested and a possible mechanism for regulating the p53-mdm2/MDMX interactions by protein phosphorylation is discussed.

Use of the phage display technique for detection of epitopes recognized by polyclonal rabbit gliadin antibodies

A random phage heptapeptide library was screened with rabbit antibodies against wheat flour proteins comprising gliadins and a small amount of low molecular weight glutenins (gli/glu). Gli/glu antibodies isolated from the sera selected different consensus sequences (CS). All CS contained tri- to pentapeptide stretches homologous to gli/glu sequences (proposed epitopes). In alpha- and gamma-type gliadins, these sequences are clustered in the N-terminal region recently suspected to be toxic for humans with celiac disease. Peptides with CS were synthesized and checked for reactivity. Only immune and no control rabbit sera reacted with synthetic peptides. One of eight human sera containing gliadin antibodies was reactive as well (4/8 peptides) but control sera were negative. Thus the phage display technique is useful for epitope screening of polyclonal antibodies even in the case of a group of homologous but diverse antigens.

The three-dimensional structures of a polysaccharide binding antibody to Cryptococcus neoformans and its complex with a peptide from a phage display library: implications for the identification of peptide mimotopes

The three-dimensional structure of 2H1, a protective monoclonal antibody to Cryptococcus neoformans, has been solved at 2.4 A resolution, in both its unbound form and in complex with the 12 amino acid residue peptide PA1 (GLQYTPSWMLVG). PA1 was previously identified as a potential mimotope of the cryptococcal capsular polysaccharide by screening of a phage display peptide library. Peptide binding is associated with only minor rearrangements of some side-chains and a small shift in the H2 loop of the antibody. The peptide assumes a tightly coiled conformation consisting of one inverse gamma-turn and one type II beta-turn that serves to place the entire peptide motif, consisting of ThrP5, ProP6, TrpP8, MetP9 and LeuP10, into a depression in the antibody combining site. A small number of H-bonds between peptide and antibody contribute to the affinity and specificity. Poor steric complementarity between PA1 and the antibody heavy chain along with the fact that the majority of the interactions between 2H1 and PA1 involve van der Waals interactions with the light chain may explain why this peptide acts as only a partial mimotope of the capsular polysaccharide epitope.

Epitope and mimotope for an antibody to the Na, K-ATPase

The epitope of a monoclonal antibody specific for the alpha 2 isoform of the Na,K-ATPase was determined and its accessibility in native enzyme was examined. Protein fragmentation with N-chlorosuccinimide, formic acid, trypsin, and leucine aminopeptidase indicated binding near the Na,K-ATPase N-terminus but did not unambiguously delineate the extent of the epitope. The ability of the antibody to bind to denatured enzyme made it a good candidate for screening a random peptide library displayed on M13 phage, but the consensus sequence that emerged was not found in the Na,K-ATPase, Full-length cDNA for the Na,K-ATPase was randomly fragmented and cloned into beta-galactosidase to create a lambda gt11 expression library; screening with the antibody yielded a set of overlaps spanning 23 amino acids at the N-terminus. Chimeras of Na,K-ATPase alpha 1 and alpha 2 narrowed down the epitope to 14-19 amino acids. The antibody did not recognize fusion proteins constructed with shorter segments of this epitope. It did recognize a fusion protein containing the M13 library consensus sequence, however, indicating that this sequence, which is rich in proline and hydrophobic amino acids (FPPNFLFPPPP), was a mimotope. The natural epitope, unique to the Na,K-ATPase alpha 2 isoform, was GREYSPAATTAENG. Reconstitution of antibody binding in a foreign context such as M13 PIII protein or beta-galactosidase thus required a relatively large number of amino acids, indicating that antibody mapping approaches must allow for epitopes of significant size. The epitope was accessible in native enzyme and exposed on the cytoplasmic side, documenting the surface exposure of a stretch of amino acids at the N-terminus, where the Na,K-ATPase isoforms differ most.

Molecular characterization of the hdm2-p53 interaction

A number of viral oncogenes target the tumour suppressor protein p53 and inactivate its function. This is an important step in tumourogenesis. The cellular oncogene hdm2 acts through a similar mechanism. It binds the N terminus of p53, thereby interfering with the ability of p53 transcriptionally to activate genes responsible for growth arrest or apoptosis after genotoxic insults. The disruption of the interaction of the two proteins therefore comprises a promising therapeutic target for treatment of the subset of human cancers in which this pathway is active. In this paper we attempt to characterize the p53-hdm2 interaction biochemically. We analyse the potential of a series of peptide inhibitors, derived from previously described mdm2 binding peptide display phage, to disrupt this interaction in ELISA assays. We conclude that F19, W23 and L26 of p53 are critical contact points for p53 binding to hdm2. Furthermore, we show the potential of the monoclonal antibody 3G5 to interfere with binding of p53 to hdm2 in ELISA assays. Consequently, we define the binding site of 3G5 on hdm2 using overlapping peptides derived from the N terminus of hdm2 and phage display libraries. The result indicates L66, Y67 and E69 on hdm2 as critical binding points for 3G5. In electrophoretic mobility shift assay we demonstrate the formation of hdm2-p53 complexes that can be disrupted in the presence of 3G5 or inhibitory peptides. Finally, we describe the effects of NEM and DTT on the interaction between the two molecules in ELISA assays. All our results are discussed in the light of the recently published crystal structure of the mdm2-p53 complex. A striking correspondence between our findings and the crystal structure is revealed.

Identification of HCV core mimotopes: improved methods for the selection and use of disease-related phage-displayed peptides

Disease-specific epitope discovery from random peptide libraries displayed on phage using sera from patients involves a number of screening steps with many immune and non-immune sera. To rapidly identify mimotopes of the human hepatitis C virus (HCV) core protein, we have used an anti-core human monoclonal antibody (mAb; B12.F8) as a probe in screening phage that were affinity-selected using a serum from an HCV infected patient. Three different positive phage were isolated displaying low or no homology with the natural antigen, but which still efficiently bound to the antigen binding site of the B12.F8 antibody. Testing the reactivity of these phage with forty-five sera from HCV infected patients showed that antibodies recognizing them are present in more than 80% of this population. These antibodies showed distinct fine specificity, as they bound the selected phage in a mutually exclusive fashion. Co-expression of two mimotopes in the same cells led to chimeric particles which were recognized by antibodies of different specificity. These data provide novel information on the potential use of the phage display technology for the characterization of antibody specificity as well as disease diagnosis and prevention.

Structural and mechanistic determinants of affinity and specificity of ligands discovered or engineered by phage display

The scope and utility of phage display is reviewed with emphasis on medical applications and structure-based ligand and drug design, from literature mostly after 1994. General principles by which phage-displayed peptides achieve affinity and selectivity for targets are described, along with selected structural or mechanistic studies of the binding of peptides or proteins discovered or engineered by phage display. Such engineered proteins whose wild-type or mutant crystal or 2D-NMR structures yield insight about the basis for enhanced affinity or altered specificity include antibodies, zinc fingers, human growth hormone, protein A, and atrial natriuretic peptide. Structures of complexes of de novo phage-discovered peptide ligands with targets such as the Src SH3 domain, streptavidin, and erythropoietin receptor reveal the structural basis for receptor-peptide recognition in these systems.

Identification of epitope recognized by an anti-c-myc monoclonal antibody that cross-reacts with E. coli sigma factor using phage display libraries

BACKGROUND

During the epitope mapping of monoclonal antibodies specific for myc proteins, two E. coli proteins cross-reactive with an anti-c-myc monoclonal antibody (MYC-X-5/1) were identified. One of the proteins is approximately 90 kDa and the other is over 150 kDa in apparent molecular mass. The molecular masses of these cross-reactive proteins suggested that they may be subunits of E. coli RNA polymerase.

OBJECTIVES

We have investigated whether or not the proteins cross-reactive with MYC-X-5/1 are subunits of E. coli RNA polymerase. In addition, we have attempted to determine the epitope of MYC-X-5/1.

STUDY DESIGN

The reactivity of MYC-X-5/1 antibody was tested against highly purified E. coli RNA polymerase holo-enzyme preparations and the cell lysate made from E. coli carrying a multi-copy plasmid with an insert of the rpoD gene, the structural gene for the E. coli sigma subunit. The epitope of MYC-X-5/1 was determined by use of phage display of random peptide libraries.

RESULTS

On immunoblotting assays, MYC-X-5/1 reacted with the 90-kDa protein in the E. coli RNA polymerase preparations and with the 90-kDa protein over-expressed in E. coli carrying the plasmid with the rpoD insert. In addition, we have deduced the epitope of the MYC-X-5/1 antibody to be residues 235-245 of the human c-myc protein. A highly similar sequence to this was also identified in residues 62-72 of the sigma subunit of E. coli RNA polymerase.

CONCLUSION

These data demonstrated that the 90-kDa protein cross-reactive with MYC-X-5/1 is the sigma subunit of E. coli RNA polymerase. Furthermore, this study shows that random peptide libraries displayed on filamentous phage are useful tools for epitope mapping and defining cross-reactivities of monoclonal antibodies.

Mapping of linear epitopes recognized by monoclonal antibodies with gene-fragment phage display libraries

Epitope mapping with mono- or polyclonal antibodies has so far been done either by dissecting the antigens into overlapping polypeptides in the form of recombinantly expressed fusion proteins, or by synthesizing overlapping short peptides, or by a combination of both methods. Here, we report an alternative method which involves the generation of random gene fragments of approximately 50-200 bp in length and cloning these into the 5' terminus of the protein III gene of fd phages. Selection for phages that bind a given monoclonal antibody and sequencing the DNA inserts of immunopositive phages yields derived amino acid sequences containing the desired epitope. A monoclonal antibody (mAb 215) directed against the largest subunit of Drosophila RNA polymerase II (RPB215) was used to map the corresponding epitope in a fUSE5 phage display library made of random DNA fragments from plasmid DNA containing the entire gene. After a single round of panning with this phage library, bacterial colonies were obtained which produced fd phages displaying the mAb 215 epitope. Sequencing of single-stranded phage DNA from a number of positive colonies (recognized by the antibody on colony immunoblots) resulted in overlapping sequences all containing the 15mer epitope determined by mapping with synthetic peptides. Similarly, we have localized the epitopes recognized by a mouse monoclonal antibody directed against the human p53 protein, and by a mouse monoclonal antibody directed against the human cytokeratin 19 protein. Identification of positive colonies after the panning procedure depends on the detection system used (colony immunoblot or ELISA) and there appear to be some restrictions to the use of linker-encoded amino acids for optimal presentation of epitopes. A comparison with epitope mapping by synthetic peptides shows that the phage display method allows one to map linear epitopes down to a size only slightly larger than the true epitope. In general, our phage display method is faster, easier, and cheaper than the construction of overlapping fusion proteins or the use of synthetic peptides, especially in cases where the antigen is a large polypeptide such as the 215 kDa subunit of eukaryotic RNA polymerase II.

Proenkephalin is a nuclear protein responsive to growth arrest and differentiation signals

Neuropeptide precursors are traditionally viewed as molecules destined to be cleaved into bioactive peptides, which are then released from the cell to act on target cell surface receptors. In this report we demonstrate nuclear localization of the enkephalin precursor, proenkephalin, in rodent and human embryonic fibroblasts (Swiss 3T3 and MRC-5 cells) and in rodent myoblasts (C2C12 cells). Nuclear proenkephalin, detected by immunofluorescence with a panel of antiproenkephalin monoclonal antibodies, is distributed predominantly in three patterns. Selective abolition of these patterns with salt, nuclease, or methanol is associated with liberation of immunoprecipitable proenkephalin into the extraction supernatant. Proenkephalin antigenic domains, mapped using phage display libraries and synthetic peptides, are differentially revealed in the three distribution patterns. Selective epitope revelation may reflect different conformational forms of proenkephalin or its existence in complexes with other nuclear proteins, forms which therefore have different biochemical associations with the nuclear substructure. In fibroblast cell populations in transition to growth arrest, nuclear proenkephalin responds promptly to mitogen withdrawal and cell-cell contact by transient, virtually synchronous unmasking of multiple antigenic domains in a fine punctate distribution. A similar phenomenon is observed in myoblasts undergoing differentiation. The acknowledgment of growth arrest and differentiation signals by nuclear proenkephalin suggests its integration with transduction pathways mediating these signals. To begin to address the mechanism of nuclear targeting, we have transfected mutated and nonmutated proenkephalin into COS (African green monkey kidney) cells. Nonmutated proenkephalin is localized exclusively in the cytoplasm; however, proenkephalin mutated at the first ATG codon, or devoid of its signal peptide sequence, is targeted to the nucleus as well as to the cytoplasm. From this we speculate that nuclear proenkephalin arises from a primary translation product that lacks a signal peptide sequence because of initiation at a different site.